28.05.2014 Views

r - The Hong Kong Polytechnic University

r - The Hong Kong Polytechnic University

r - The Hong Kong Polytechnic University

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Proceedings of the 5th Cross-strait Conference on Structural and Geotechnical<br />

Engineering (SEG-5)<br />

VOLUME 1<br />

13-15 July 2011, <strong>Hong</strong> <strong>Kong</strong>, China<br />

Edited by<br />

Y. Q. Ni, J. H. Yin and X. W. Ye<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

Organised by<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

Co-organised by<br />

Zhejiang <strong>University</strong><br />

National Taiwan <strong>University</strong>


Copyright@2011 Faculty of Construction and Land Use, <strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong>.<br />

Authors retain all proprietary right in any process, procedure, or article of manufacture described in the<br />

Work. Authors may reproduce or authorize others to reproduce the Work, material extracted verbatim from<br />

the Work, or derivative works for the author’s personal use or for company use, provided that the source is<br />

indicated.<br />

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by<br />

any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission<br />

from the publisher.<br />

Although all care is taken to ensure the integrity and quality of this publication and the information herein,<br />

no responsibility is assumed by the publisher nor the authors for any injury and/or damage to property or<br />

persons as a result of operation or use of this publication and/or the information contained herein.<br />

ISBN: 978-988-15439-1-2<br />

Published by: Faculty of Construction and Land Use, <strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong>, <strong>Hong</strong> <strong>Kong</strong>,<br />

China.


SCIENTIFIC COMMITTEE<br />

Chairman<br />

Jan-Ming KO<br />

Vice-chairmen<br />

Shi-Lin DONG<br />

Jin-Guang TENG<br />

Yeong-Bin YANG<br />

Members<br />

Andrew, Ka-Ching CHAN<br />

Siu-Tack CHAN<br />

Kuo-Chun CHANG<br />

Kam-Tim CHAU<br />

Yun-Min CHEN<br />

Zu-Yu CHEN<br />

Moe MS CHEUNG<br />

Kin-Kuen CHOY<br />

Reuben Pui-Kwan CHU<br />

Xue-Yi FU<br />

Xiu-Run GE<br />

Ji-Ping HAO<br />

Albert Ngai-Leung HO<br />

Vai-Pan IU<br />

Wei-Liang JIN<br />

Chang-Hua KE<br />

Sritawat, KITIPORNCHAI<br />

Albert K H KWAN<br />

Kin-Kei KWAN<br />

Ching-Kwong LAU<br />

Chack-Fan LEE<br />

Liang-Jenq LEU<br />

Andrew YT LEUNG<br />

Christopher, K Y LEUNG<br />

Chien-Chung LI<br />

Ching-Lung LIAO<br />

Hung-Jiun LIAO<br />

Kim-Meow LIEW<br />

Chi-Chang LIN<br />

Xi-Liang LIU<br />

Chin-Hsiung LOH<br />

Ke-Jian MA<br />

Za-Chieh MOH<br />

Jian-Guo NIE<br />

Jui-Lin PENG<br />

Ji-Ping RU<br />

Zu-Yan SHEN<br />

Shi-Zhao SHEN<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

Zhejiang <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

National Yunlin <strong>University</strong> of Science & Technology<br />

Ove Arup & Partners<br />

Housing Department, the Government of the HKSAR<br />

National Center for Research on Earthquake Engineering<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

Zhejiang <strong>University</strong><br />

China Institute of Water Resources and Hydropower Research<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> of Science & Technology<br />

Buildings Department, the Government of the HKSAR<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> Institution of Engineers<br />

China Construction Design International (Shenzhen)<br />

Institute of Rock and Soil Mechanics, Chinese Academy of Sciences<br />

Xian <strong>University</strong> of Architecture & Technology<br />

<strong>Hong</strong> <strong>Kong</strong> Geotechnical Society<br />

<strong>University</strong> of Macau<br />

Zhejiang <strong>University</strong><br />

Beijing Institute of Architectural Design<br />

City <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

<strong>The</strong> <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

Ove Arup & Partners<br />

Maunsell Consultants Ltd.<br />

<strong>The</strong> <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

National Taiwan <strong>University</strong><br />

City <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> of Science & Technology<br />

CECI Engineering Consultants, Inc., Taiwan<br />

China Engineering Consultants, Inc.<br />

National Taiwan <strong>University</strong> of Science and Technology<br />

City <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

National Chung Hsing <strong>University</strong><br />

Tianjin <strong>University</strong><br />

National Taiwan <strong>University</strong><br />

Guizhou <strong>University</strong><br />

Moh and Associates, Inc.<br />

Tsinghua <strong>University</strong><br />

National Yunlin <strong>University</strong> of Science & Technology<br />

National Natural Science Foundation of China<br />

Tongji <strong>University</strong><br />

Harbin Institute of Technology<br />

i


SCIENTIFIC COMMITTEE<br />

James C TAI<br />

Leslie George THAM<br />

Keh-Chyuan TSAI<br />

Chi-Sing WAI, JP<br />

Chung-Yue WANG<br />

Hok-Ning WONG<br />

Shi-Lang XU<br />

You-Lin XU<br />

De-Qing YI<br />

De-Yu YIN<br />

Ji-Da ZHAO<br />

Jian ZHENG<br />

Ying-Ren ZHENG<br />

Dai ZHOU<br />

Xu-<strong>Hong</strong> ZHOU<br />

T.Y. Lin Taiwan Consulting Engineers, Inc.<br />

<strong>The</strong> <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

National Taiwan <strong>University</strong><br />

Development Bureau, the Government of the HKSAR<br />

National Central <strong>University</strong><br />

Civil Engineering and Development Department, the HKSAR<br />

Zhejiang <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

Zhejiang Provincial Institute of Architectural Design & Research<br />

Taiyuan <strong>University</strong> of Technology<br />

China Academy of Building Research<br />

Ministry of Railways, <strong>The</strong> People’s Republic of China<br />

Logistical Engineering <strong>University</strong><br />

Shanghai Jiao Tong <strong>University</strong><br />

Lanzhou <strong>University</strong><br />

ii


ORGANIZING COMMITTEE<br />

Co-chairmen<br />

Yi-Qing NI (Structural Engineering)<br />

Jian-Hua YIN (Geotechnical Engineering)<br />

Members<br />

Chih-Chen CHANG<br />

Chien-Chou CHEN<br />

Zhi-Hua CHEN<br />

Yung-Ming CHENG*<br />

Kwok-Fai CHUNG<br />

Jian-Guo DAI*<br />

Hua DENG<br />

Shang-Hsien HSIEH<br />

Eddie LAM<br />

Siu-Seong LAW<br />

Qiu-Sheng LI<br />

Zong-Jin LI*<br />

Han-Long LIU<br />

Man-Hoi LOK<br />

Yao-Zhi LUO<br />

Wai-Meng QUACH*<br />

Li-Zhong WANG<br />

Yuk-Lung WONG<br />

Wen-Hwa WU<br />

Yu-Fei WU*<br />

Yong XIA<br />

Xue-Yu XIONG<br />

Jun YANG*<br />

Ben YOUNG<br />

Quentin Z Q YUE<br />

Ka-Veng YUEN<br />

Li-Min ZHANG*<br />

Yang ZHAO*<br />

Wan-Huan ZHOU*<br />

Song-Ye ZHU*<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> of Science & Technology<br />

National Yunlin <strong>University</strong> of Science & Technology<br />

Tianjin <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

Zhejiang <strong>University</strong><br />

National Taiwan <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

City <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> of Science & Technology<br />

Hohai <strong>University</strong><br />

<strong>University</strong> of Macau<br />

Zhejiang <strong>University</strong><br />

<strong>University</strong> of Macau<br />

Zhejiang <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

National Yunlin <strong>University</strong> of Science & Technology<br />

City <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

Tongji <strong>University</strong><br />

<strong>The</strong> <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

<strong>The</strong> <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

<strong>The</strong> <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong><br />

<strong>University</strong> of Macau<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> of Science & Technology<br />

Zhejiang <strong>University</strong><br />

<strong>University</strong> of Macau<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

(* Executive Committee Members)<br />

Secretaries<br />

Xiao-Wei YE<br />

Tao YU<br />

Hua-Fei ZHOU<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong><br />

iii


PREFACE<br />

Riding on the success of the previous 4 conferences respectively held in 1994 (Hangzhou), 1997<br />

(<strong>Hong</strong> <strong>Kong</strong>), 2003 (Taipei) and 2007 (Hangzhou), we are delighted to convene the 5th Cross-strait<br />

Conference on Structural and Geotechnical Engineering from 13 to 15 July, 2011 at <strong>The</strong> <strong>Hong</strong> <strong>Kong</strong><br />

<strong>Polytechnic</strong> <strong>University</strong>.<br />

This conference is organized by <strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong> and co-organized by<br />

Zhejiang <strong>University</strong> and National Taiwan <strong>University</strong>. This event is not only to provide a forum for<br />

structural and geotechnical engineering professionals and academia from the Chinese mainland,<br />

Taiwan, <strong>Hong</strong> <strong>Kong</strong> and Macau as well Chinese scholars from other countries to meet together and<br />

share new ideas, achievements and experiences through presentations and discussions, but also to<br />

review trends in research development and engineering applications.<br />

<strong>The</strong> proceedings of the conference comprise 8 keynote papers (or lectures) and 20 invited papers (or<br />

lectures) as well as 92 regular papers. <strong>The</strong>se papers have covered a wide range of issues concerning<br />

structural and geotechnical engineering. Showcasing diversity and quality, these papers report the<br />

current state-of-the-art and point to future directions of research and applications in this exciting area.<br />

<strong>The</strong> success of the conference is due to the dedication and support of many individuals and<br />

organizations. On behalf of the Organizing Committee, we would like to thank all authors for careful<br />

preparation of their papers, and all speakers of keynote papers, invited papers, and regular papers for<br />

sharing their work, experience and insight at the conferencing. All papers submitted to the conference<br />

were reviewed by members of the Scientific Committee and the Organizing Committee. We are<br />

grateful to all of them for their important contributions to the conference. In addition to sharing the<br />

paper review work, members of the Organizing Committee have also been most generous with their<br />

time in the organization work. As chairs of the Organizing Committee, we are indebted to all of them.<br />

<strong>The</strong> financial support from Kwang-Hua Fund for College of Civil Engineering, Tongji <strong>University</strong> is<br />

acknowledged with heartfelt gratitude.<br />

On behalf of the Organizing Committee, we would like to express many thanks to colleagues from<br />

Faculty of Construction and Land Use for their secretarial support, in particular, Miss Liz Lau, Miss<br />

Cindy Li, Ms Connie Man, and Mr Jason Au, who have been involved in the production of the<br />

proceedings and preparation of the conference.<br />

One of the founders of this series of conferences, Professor Wen-Lu Jin ( 金 问 鲁 教 授 ) has passed<br />

away. We have written one Chinese article, following this preface, to memorialize Professor Jin for<br />

his contributions to this series of conferences and to his achievements in research and practices in<br />

structural and geotechnical engineering.<br />

Prof. Y.Q. Ni and Prof. J.H. Yin<br />

Chairs of the Organizing Committee of SGE-5<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong>, <strong>Hong</strong> <strong>Kong</strong>, China<br />

iv


设 计 大 师 金 问 鲁 教 授<br />

( 第 一 届 海 峡 两 岸 结 构 与 地 基 国 际 学 术 研 讨 会 发 起 人 之 一 )<br />

设 计 大 师 金 问 鲁 教 授 于 1997 去 世 , 特 写 此 文 感 谢 他 对 海 峡 两 岸 结 构 与 地 基 国 际 学 术 研 讨 会 的<br />

贡 献 和 他 在 学 术 与 实 践 工 作 上 的 成 就 。<br />

金 问 鲁 教 授 (1925-1997), 浙 江 嘉 兴 人 ,1946 年 毕 业 于 上 海 交 通 大 学 。 原 杭 州 城 建 设 计 院 院<br />

长 兼 总 工 程 师 , 第 六 、 第 七 届 全 国 人 大 代 表 , 国 家 设 计 大 师 。 曾 任 中 国 力 学 学 会 理 事 、 中 国 土 木<br />

建 筑 学 会 理 事 、 浙 江 省 力 学 学 会 理 事 长 、 浙 江 省 土 木 建 筑 学 会 副 理 事 长 、 大 连 理 工 大 学 兼 职 教 授<br />

等 。 发 表 学 术 论 文 百 余 篇 , 学 术 专 著 七 部 。 曾 获 全 国 科 技 大 会 重 大 科 技 成 果 奖 及 个 人 先 进 奖 、 全<br />

国 先 进 科 技 工 作 者 、 建 设 部 劳 模 以 及 其 他 国 家 、 省 ( 部 )、 市 级 奖 励 共 16 项 。1994 年 任 杭 州 市 科<br />

协 副 主 席 和 杭 州 结 构 与 地 基 研 究 会 理 事 长 期 间 , 金 教 授 建 议 并 参 与 发 起 了 第 一 届 海 峡 两 岸 结 构 与<br />

地 基 国 际 学 术 研 讨 会 。<br />

金 问 鲁 教 授<br />

金 教 授 一 贯 刻 苦 钻 研 、 从 严 治 学 、 学 以 致 用 。 作 为 设 计 院 总 工 程 师 , 他 善 于 从 工 程 问 题 中 发<br />

现 未 解 决 的 理 论 问 题 , 在 学 术 上 不 断 创 新 , 并 将 其 应 用 于 工 程 实 践 。 他 所 发 表 的 论 文 与 专 著 几 乎<br />

涉 及 当 时 土 木 工 程 所 有 理 论 前 沿 , 比 如 在 50 年 代 由 他 首 先 提 出 将 流 变 理 论 应 用 于 沥 青 混 凝 土 路 面<br />

的 设 计 计 算 ;60 年 代 在 预 应 力 混 凝 土 结 构 、 钢 丝 网 水 泥 结 构 、 组 合 廻 转 壳 等 方 面 完 成 了 大 量 创 造<br />

性 工 程 设 计 与 论 文 ;70 年 代 出 版 了 我 国 第 一 部 《 悬 挂 结 构 理 论 》 和 第 二 部 《 悬 挂 结 构 计 算 理 论 》<br />

专 著 , 系 统 论 述 了 各 种 悬 挂 结 构 的 几 何 非 线 性 分 析 方 法 , 广 泛 应 用 了 普 遍 变 分 原 理 , 从 理 论 上 解<br />

决 了 索 与 索 组 合 结 构 的 静 力 计 算 问 题 ;80 年 代 出 版 了 我 国 第 一 部 《 预 应 力 混 凝 土 - 徐 变 状 态 统 一<br />

计 算 理 论 》 专 著 , 统 一 了 按 弹 性 和 徐 变 两 种 状 态 的 计 算 理 论 , 考 虑 了 有 粘 结 和 无 粘 结 两 种 情 况 ,<br />

提 出 了 预 应 力 混 凝 土 板 壳 分 析 方 法 , 并 在 粘 土 三 维 固 结 和 次 固 结 问 题 、 双 层 地 基 承 载 力 、 桩 土 共<br />

同 作 用 分 析 、 桥 梁 结 构 三 维 分 析 、 弹 性 地 基 上 的 壳 体 内 力 分 析 、 高 层 建 筑 结 构 等 方 面 在 学 术 期 刊<br />

上 发 表 了 一 系 列 论 文 。90 年 代 出 版 了 《 高 层 建 筑 结 构 的 连 续 化 分 析 》 专 著 , 提 出 了 开 口 、 闭 口 薄<br />

壁 构 件 统 一 计 算 模 型 , 为 宏 观 分 析 复 杂 高 层 建 筑 及 其 抗 震 设 计 开 辟 了 思 路 。 后 期 专 心 于 随 机 振 动<br />

方 面 的 研 究 , 其 中 《 结 构 非 线 性 非 平 稳 随 机 振 动 分 析 》 一 文 发 表 于 第 十 六 届 国 际 理 论 力 学 和 应 用<br />

v


力 学 会 议 , 并 发 表 了 《 结 构 动 力 学 的 谱 分 解 变 分 原 理 及 有 限 元 计 算 》、《 随 机 振 动 的 有 限 元 分<br />

析 》 等 十 多 篇 论 文 。<br />

金 教 授 在 第 一 届 海 峡 两 岸 结 构 与 地 基 国 际 学 术 研 讨 会 上 宣 读 了 《 奇 异 摄 动 理 论 在 薄 壳 中 的 应<br />

用 》, 此 文 可 作 为 求 解 任 意 形 状 的 薄 壳 解 析 解 的 工 具 。 他 发 表 在 “ 应 用 数 学 与 力 学 ” 第 20 卷 第 3<br />

期 上 的 《 固 体 的 统 一 弹 、 粘 、 塑 性 理 论 》, 基 于 热 力 学 定 律 及 虚 弹 性 假 设 , 提 出 了 一 个 固 体 弹 、<br />

粘 、 塑 性 统 一 理 论 , 可 用 于 计 算 物 体 在 任 意 受 力 过 程 中 弹 、 粘 、 塑 性 的 变 化 情 况 , 文 中 导 出 本 构<br />

关 系 以 及 有 关 的 变 分 原 理 , 由 此 容 易 推 导 出 空 间 ― 时 间 的 有 限 元 构 式 。 该 文 代 表 了 金 教 授 对 固 体<br />

力 学 研 究 的 新 高 度 , 也 可 以 说 是 金 教 授 毕 生 学 问 的 总 结 , 说 明 他 已 经 洞 察 到 固 体 力 学 各 种 分 支 间<br />

理 论 上 的 内 在 联 系 。 金 教 授 理 论 联 系 实 际 的 学 风 给 理 论 研 究 带 来 了 活 力 , 他 的 研 究 成 果 在 行 业 内<br />

得 到 了 广 泛 应 用 。<br />

第 一 届 海 峡 两 岸 结 构 与 地 基 学 术 研 讨 会 成 员 合 影 – 1994 年 摄 于 杭 州<br />

第 一 排 ( 左 一 ) : 金 问 鲁 教 授<br />

vi


TABLE OF CONTENTS<br />

Scientific Committee<br />

Organizing Committee<br />

Preface<br />

A Memorial of Professor Wen-lu Jin<br />

Table of Contents<br />

i<br />

iii<br />

iv<br />

v<br />

vii<br />

Volume 1<br />

Keynote Lectures<br />

Conceptualization of a Bridge Crossing Taiwan Strait 1<br />

M.C. Tang<br />

Application and Development of Modern Long-Span Space Structures in Mainland China 6<br />

S.L. Dong, D. Xing & Y. Zhao<br />

Ground Vibrations Due to Underground Trains by the 2.5D Finite/Infinite Element Approach 20<br />

Y.B. Yang & H.H. Hung<br />

<strong>The</strong> Zhouqu Debris Flow 30<br />

C.F. Lee<br />

<strong>The</strong> <strong>The</strong>ory of Limit Analysis and the Method of Numerical Limit Analysis 31<br />

Y.R. Zheng, X.S. Tang & S.Y. Zhao<br />

Bai-He-Liang Ancient Hydrological Inscription - <strong>The</strong> World's First Site Class Underwater 37<br />

Museum<br />

X.R. Ge<br />

Shaking Table Tests on Full-Scale Low-Rise Cold-Formed Thin-Walled Steel Residential 70<br />

Buildings Using Light-Gauge Composite Walls<br />

Z.Y. Shen, Y.Q. Li, F. Liu, Y.F. Qing & S.D. Wu<br />

Two Kinds of New Partial Pre-Stressed Space Steel Grid Structure with Super Large Span 78<br />

(150m×150m)<br />

K.J. Ma, B. Shen & G.S. Feng<br />

Invited Lectures<br />

Innovation and Practice of China High-Speed Rail Stations 93<br />

J. Zheng<br />

Diagnosis of Track Integrity 97<br />

C.L. Liao, W.F. Chen & C.Y. Wang<br />

Recent Seismic Design and Retrofit Studies of Bridges at NCREE 106<br />

K.C. Chang, H.H. Hung & K.Y. Liu<br />

vii


Pseudo-Ductile Permanent Formwork for the Construction of Durable Concrete Structures 117<br />

Christopher K.Y. Leung & C.L. Yu<br />

Key FRP Technologies in Structural Retrofitting and Strengthening 124<br />

Z.S. Wu, X. Wang & G. Wu<br />

<strong>The</strong> Evolution of Transversely Confined Structural Columns 148<br />

Y. Xiao<br />

Spatial Ground Motion Modelling and Its Effect on Bridge Responses 156<br />

H. Hao & K.M. Bi<br />

Dynamic Mechanical Analysis of Magnetorheological Smart Nanocomposites 166<br />

L.Z. Sun<br />

A <strong>The</strong>oretical Plate End Flexural Debonding Model for Plated Beams 167<br />

J.F. Chen, V. Narayanamurthy, J. Cairns & D.J. Oehlers<br />

Smart Aggregate-Based Damage Detection of FRP-Strengthened Columns Under Reversed 177<br />

Cyclic Loading<br />

H.C. Gu, R. Howser, Y. Moslehy, H. Dhonde, G.B. Song, Y.L. Mo & A. Ayoub<br />

On the Multi-Scale Modeling of Heterogeneous Geomaterials 187<br />

J.F. Shao, A. Guery, T. Jiang, Q.Z. Zhu & D. Kondo<br />

Failure Types of Anchors and Anchored Slopes in Taiwan 197<br />

H.J. Liao & S.H. Cheng<br />

Research Progress of Liquefaction Evaluation of Sandy Soils by Shear Wave Velocity 208<br />

Y.M. Chen & Y.G. Zhou<br />

Design <strong>The</strong>ory and Application of Tubed Concrete Columns 217<br />

X.H. Zhou & L.J. Peng<br />

Research Advances of Steel-Concrete Composite Bridges 227<br />

J.G. Nie, M.X. Tao, L.L. Wu, X. Nie, F.X. Li & F.L. Lei<br />

Rate Dependence of Ultra High Toughness Cementitious Composite in Tension 241<br />

S.L. Xu & H.D. Li<br />

Structural and Geotechnical Aspects of Super-Tall Structures over 1000 m 242<br />

X.F. Chen<br />

Design of Wind and Structural Health Monitoring System for Stonecutters Bridge 243<br />

K.Y. Wong<br />

A Reliability Based Simulation, Monitoring and Code Calibration of Vehicle Effects on 290<br />

Existing Bridge Performance<br />

C.S. Cai, W. Zhang, M. Xia & L. Deng<br />

Parallel Session – Geotechnical I<br />

A Study on Displacement-Based Earthquake Loss Assessment Adopting Equivalent Stiffness 300<br />

Linearization Method<br />

viii


J.T. Shi & L. Su<br />

Impact of Spatial Variability on Soil Shear Strength 311<br />

J. Ching, K.K. Phoon & Y.G. Hu<br />

Adsorption and Desorption Behavior of Bivalent Nickel and Manganese Ions on Loess Soil 317<br />

Y. Wang, X.W. Tang, H.Y. Wang & Z.F. Sun<br />

Uncertainty in Nonlinear Seismic Ground Response Analyses 323<br />

O.L.A. Kwok<br />

An Accurate Geological Model is an Essential Requirement for Geotechnical Engineering – 327<br />

A Case Study on the Geology of Tuen Mun to Tin Shui Wai Area, <strong>Hong</strong> <strong>Kong</strong><br />

K.W. Lai & H.M.S. Chan<br />

Upper Bound Limit Analysis of Soil Slope Stability Based on Rpim Meshless Method 337<br />

F.T. Liu, J.D. Zhao, Y.F. Fan & J.H. Yin<br />

Characteristics of Damaged Dams and Influencing Factors in <strong>The</strong> Wenchuan 5.12 Earthquake 347<br />

H.A. Liang, L.P Jing, Y.Q Li & C.H. Liu<br />

Lessons Learned from 2011 Tohoku Earthquake and Tsunami 354<br />

K.T. Chau<br />

Accident Treatment on Large Differential Settlement Between Rigid and Flexible Pile 362<br />

Foundation in the Same Building<br />

Z.M. Zhang & Q.Q. Zhang<br />

Discussion of Active Earth Pressure’s Coefficient Formula about the National Standard 369<br />

X. Lu, Y.F. Wang & Y.R. Zheng<br />

Deformation Character and Control Analysis for a Large-Section Twin Tunnel in Construction 375<br />

Z.M. Li, Q. Feng, X.H. Zhu, F. Wu, L. Wan & J.L. Ou<br />

Scaling Earthquake Records for Seismic Performance Assessment of Buildings 382<br />

Y.N. Huang<br />

Model Tests Study on Cast-In-Place X-Sectional Pile 390<br />

G.Q. <strong>Kong</strong>, H.L. Liu & M.X. Zhang<br />

Performance Prediction of Lateral Response of Adjacent Single Pile Based on the Inclinometer 395<br />

Curves<br />

R.J. Zhang, J.J. Zheng, Y.T. Pan & S. Yu<br />

Parallel Session – Structural I<br />

A New Reliability Analysis Method Based on Uniform Design Method and Support Vector 399<br />

Machines<br />

X.L. Yu, J.B. Yu, H.B. Zheng & Q.S. Yan<br />

Shaking Table Test of Semi-Active Friction Tuned Mass Damper for Structural Control 407<br />

C.C. Lin, G.L. Lin, Y.B. Ho & L.Y. Lu<br />

Establishment of Taiwan Pedestrian Suspension Bridge Management System 417<br />

M.H. Chen, C.Y. Wang & C.L. Liao<br />

ix


Active Vibration Control of Cable-Strut Tensegrity Structures Under Wind Excitation 427<br />

N. Xiao, Y.Z. Miao & H.P. Chen<br />

Prediction of Residual Service Life and Through-Life Maintenance Costs for <strong>Hong</strong> <strong>Kong</strong> 435<br />

Public Rental Housing Estates<br />

H.W. Pang, C.O. Chan & W.B Chan<br />

Technological and Economic Comparison of Two Super High-Rise Structural Systems 442<br />

H.M. Zhou, J.C. Ye & Z.L. Xie<br />

Wind-Induced Structural Dynamic Response Analysis for a Large Stadium Cable-Net Roof 452<br />

W.J. Chen, T. Ren & Y.L. He<br />

Analysis of Mechanical Behavior of Half-Through Arch Beam Combination Bridge with 457<br />

Swing Construction<br />

Y.Q. Xiang, J.W. Chen, J.F. Wang, W.L. Yang & K. Cheng<br />

Finite Element Study of Inelastic Local Web Buckling Capacity of Coped Steel I-Beam 464<br />

Y. Qin, C.C. Lam, V.P. Iu & K.P. Kou<br />

A Refined Ultimate Bearing Capacity Analysis on Single Layer Latticed Domes with Welded 474<br />

Hollow Spherical Joints<br />

F. Wu, M.Q. Ding, L. Gu, X.Y. Fu & X.C. Chen<br />

Effect of Anchorage System in Identifying Modal Frequencies of Short Stay Cables 480<br />

C.C. Chen, W.H. Wu & C.Y. Liu<br />

Analysis of Shear Performance of Concrete T-Beam Bridge Strengthened by External 489<br />

Prestressed Tendon<br />

Y.Q. Xiang, X.H. Zhu, T.J. Lou & Q.Q. Wu<br />

In-Plane Pure Shear Panel Test of Self-Consolidating High Performance Fiber Reinforced 495<br />

Concrete (SCHPFRC)<br />

W.C. Liao<br />

<strong>The</strong> Structural Design of Shanxi Sports Center Stadium 499<br />

X.Y. Fu, Y.J. Zhu, Y. Zhou, X. B. Yang, Y.P. Cheng, T. Wang & J. Zhang<br />

Design on the Sub-Structure of Shenzhen North Train Station 508<br />

M.L. Meng, B. Wu, X.Y. Fu, Z.H. Chen, Y.W. Feng & J.W. Shao<br />

Scouring Evaluation of Cable-Stayed Bridges Based on Ambient Vibration Measurements 515<br />

W.H. Wu, C.C. Chen, F. Shi & S.W. Wang<br />

Volume 2<br />

Parallel Session – Structural II<br />

Revitalization of Historic Buildings – “Conversion of Yau Ma Tei <strong>The</strong>atre and the Red Brick 525<br />

Building into a Xiqu Centre”<br />

K.Y. Ma, Y.K. Chan & C.Y. Wong<br />

x


Temperature Effect on Variation of Frequency of Beams: A Comparative Study 535<br />

X. Q. Zhou, B. Chen & Y. Xia<br />

Topology Optimization Using Stochastic Search Methods 545<br />

J.Y. Guo & L.J. Leu<br />

Analysis of the Kinematic Path of Load-Bearing Cable-Bar Mechanisms 557<br />

H. Deng, Y.Z. Zu, X.S. Wu & B.W. Jiang<br />

Application of Recursive Stochastic Subspace Identification in On-Line Bridge Monitoring 569<br />

System<br />

J.H. Weng, C.H. Loh, S.S. Chao, K.C. Lu & C.H. Chen<br />

Development of Multifunctional Laminar Shear Container for Shaking Table Test 581<br />

H.F. Sun, L.P. Jing, N.W. Wang & X.C. Meng<br />

Experimental Study on Crack Mode in Reinforced Concrete Structures with Rebar Corrosion 587<br />

Y.X. Zhao, J. Yu & W.L. Jin<br />

Experimental Study on Seismic Behavior of Autoclaved Aerated Concrete Block Composite 596<br />

Walls with Structural Columns<br />

P.C. Ling, J. Zhao, B.Z. Zhou, H.G. Wu, Q.S. Miao, Y.Y. Zhu & J. Tu<br />

Seismic Resistant Design and Analysis of Vertical Boundary Elements in Steel Plate Shear Walls 602<br />

K.C. Tsai, C.H. Li, J.T. Chang & C.H. Lin<br />

Application of Singular Spectrum Analysis to Health Monitoring of Bridge Structure 611<br />

S.H. Chao, C.H. Loh, J.H. Weng, P.Y. Lin & C.H. Chen<br />

FEM Analysis of Shear Connector Behavior of Continuous Steel-Concrete Composite Girder 618<br />

Bridge<br />

H.Y. Cao, Z.J. Chen, H.P. Zhu & H.Y. Yang<br />

A Study on Damage Assessment of the Scoured Bridges 627<br />

C.Y. Wang & Y.C. Sung<br />

Wind Tunnel Test and Wind-Induced Structural Dynamic Response Analysis for a Steel 636<br />

Velarium Roof<br />

H.F. Zhu, W.J. Chen, Y.L. He & S.L. Dong<br />

Load-Deformation Analysis of Reinforced Concrete Columns Including Shear Effects 643<br />

Q. Zhang & J.X. Gong<br />

Systematic Categorization of Structural Components in Stonecutters Bridge 649<br />

K.C. Lin, X.W. Ye, Y.Q. Ni & K.Y. Wong<br />

Stability Behavior of Fixed-Roof Tapered-Wall Steel Tanks under Differential Settlement 655<br />

Z. Wang, X. Lei, Y. Zhao & X.Y. Xie<br />

Parallel Session – Geotechnical II<br />

Preliminary Study of the Influence of Underground Structures on Groundwater in the Centre 661<br />

of Guangzhou City<br />

H. Cao & G.Y. Luo<br />

xi


Reflection of Seismic Waves at Boundary of an Unsaturated Poroelastic Half-Space 670<br />

W.Y. Chen, T.D. Xia & X.B. <strong>Kong</strong><br />

Earthquake Loss Estimation of Building with Shallow Foundation due to Soil Liquefaction 677<br />

C.C. Lu, J.H. Hwang, C.R. Huang & Y.D. Lu<br />

A Quantitative Evaluation on Bridge Scouring Safety 686<br />

H. Wang, C.Y. Wang, T.R. Wu, S.C. Hsieh, Y.Y. Ko, J.S. Chiou,<br />

M.Z. Chen, T.Y. Chen, C.H. Chen & C. Lin<br />

Influence of Pore Pressure on Tunnel Face Stability 697<br />

W. Liu, X.W. Tang, H.Y. Wang & B. Huang<br />

Installation Load and Working Capacity of Jacked Piles: Some Experiences in China 701<br />

F. Yu & J. Yang<br />

Effects of Presence of a Soft Soil Layer on Liquefaction Evaluation 706<br />

C. Li, W.W. Yang, Y.C. Lo, S.H. Yung & C.H. Wong<br />

Resistance of Soil-Root System of Selected Native Plants in <strong>Hong</strong> <strong>Kong</strong> 716<br />

F.T.Y. Leung, B.C.H. Hau, L.G. Tham & W.M. Yan<br />

Guided Wave Interpretation of Impulse Responses for Deep Foundations with Emphasis on 720<br />

Cross-Sectional Profiles<br />

H. Wang, B.T. Chen & T.P. Chang<br />

Establishment of the New Failure Criterion of Soils 728<br />

R.Q. Xu & X.C. Wang<br />

Numerical Evaluation of the Seismic Performance for a Ductile Fiber Reinforced Concrete 731<br />

Coupled Wall System<br />

C.C. Hung<br />

Geotechnical Properties of Colluvial and Alluvial Deposits in <strong>Hong</strong> <strong>Kong</strong> 735<br />

K.W. Lai<br />

Parallel Session – Structural III<br />

Multi-Located Tuned Mass Dampers for Vibration Mitigation of Tower-Like Structures with 745<br />

Whipping Effect<br />

Y.F. Duan, Y.Q. Ni & J.M. Ko<br />

Nonlinear Analysis and Calculation of Second-Order Effect of Sway-Restricted Columns 756<br />

J. Xu & J.X. Gong<br />

Nonlinear Stability Analysis for a Concrete-Filled-Steel-Tubular Arch/Continuous Beam Bridge 763<br />

Including Defects in Arch Rib<br />

J.L. Hu, Q.S. Yan, Z. Chen & Y.H. Gao<br />

Numerical Simulation of Wind Pressure and Wind-Resistant Optimization on Concave Roofed 769<br />

Low-Rise Building with Eaves<br />

D. Zhou, J.H. Tu & J.L. Li<br />

xii


Trace Analysis of Mechanical Response of Defect Member During Losing Overall Structural 777<br />

Stability<br />

J.M. Guo, W.L. Xue, X.Q. Zhao & X.S. Wu<br />

Introduction for Program Saptrans to Convert SAP2000 Model to ABAQUS and NASTRAN 784<br />

X.Q. Yang, X.Y. Fu & Y.J. Huang<br />

Walking Comfort Analysis on Station Building Floor System of Shen Zhen North Station 792<br />

B. Wu, X.Y. Fu, M.L. Meng, Z.H. Chen & J.X. Qu<br />

Integrate On-Line Recursive SSA and SSI-COV Algorithms for Operational Modal Analysis 797<br />

of Structures<br />

Y.C. Liu & C.H. Loh<br />

Progressive Collapse Analysis of Multi-Storey Frame Structures 806<br />

Z. Wang, J.R. Pan & H.P. He<br />

<strong>The</strong> Approximate Formula of Vertical Vibration Fundamental Frequency of Three-Tower 810<br />

Suspension Bridge<br />

B. Liu, Y. Zhang, K.L. Chen & J.X. Shen<br />

Evaluation of the Potentiality of Bridge Scouring 818<br />

S.H. Hung, C.Y. Wang, W.F. Lee, H.P. Lien & C.K. Huang<br />

Recommending a Patent Technology of Reducing the Noise Induced by Rail Transit 828<br />

F. Lin, J.F. Gu & G.Z. Qian<br />

Development of a Digital Camera System for Monitoring of Structural Displacement 832<br />

F. Xu, Y.Q. Ni & X.W. Ye<br />

RC Moment Frame Buildings Column Loss Analysis: <strong>The</strong> Effect of Masonry-Infill Wall 838<br />

S. Li, S.P. Liu, C.H. Zhai & L.L. Xie<br />

Seismic Response Analysis of Shen Zhen North Station under Multi-Dimension and 846<br />

Multi-Support Seismic Excitation<br />

B. Wu, X.Y. Fu, M.L. Meng & J.X. Qu<br />

Axial Load Carrying Capacity of Circular Reinforced Concrete Columns 851<br />

Ivy F.Y. Ho & Eddie S.S. Lam<br />

Parallel Session – Structural IV<br />

Ultimate Strengths of Concrete Bridge Deck Slabs Reinforced with GFRP Bars 861<br />

Y. Zheng, Y.F. Pan & G.Y. Yu<br />

Seismic Upgrade of Reinforced Concrete Beam-Column Joint Under High Level Axial Load 868<br />

B. Li & Eddie S.S. Lam<br />

Structural Health Monitoring of RC Structures Subjected to Seismic Loading Using 880<br />

Piezoceramic-Based Sensors<br />

W.I. Liao, Y.C. Sung, K.C. Chang & J.S. Hwang<br />

Smart Elasto-Magneto-Electric (EME) Sensors for Stress Monitoring of Steel Bars Using 886<br />

Magneto-Electric (ME) Sensing Units<br />

R. Zhang, Y.F. Duan, Y. Zhao, S.W. Or & K.Q. Fan<br />

xiii


Some Considerations on Five Technical Specifications of Steel-Concrete Composite Structure 893<br />

Z.T. Tu, G.M. Teng & G.Z. Qian<br />

Space Analysis of Long-Span Curved Continuous Rigid Frame Bridges with High Pier Under 897<br />

Gravity Load<br />

Y.X. Yang, X.W. Hao & L. Sun<br />

Stability Analysis Method for Lattice Shells Accounting for Member Buckling 903<br />

W. Tian , Y. Zhao & S.L. Dong<br />

Structure Design of Shanxi Olympic Gymnasium 913<br />

X.B. Yang, Y. Gao, X.Y. Fu, X.M. Cui, T. Wang & J.R. Zhou<br />

Study on the Mineral Admixtures Influence on the Electric Flux and Chloride Ion Migration 925<br />

Coefficient of Concrete<br />

Y. Li, L. Qiao, C. Yan, Y. Zhang & X.L. Du<br />

Applications of Non-Contact Measurement Techniques to Bridge Health Inspection 930<br />

C.S. Wang, C.Y. Wang, Y.C. Sung, C.C. Cheng & P.Y. Hung<br />

Vortex Shedding Suppression of Cylindrical Structures Near Plane Wall 938<br />

J.S. Cui, F.P. Gao, Z.P. Zang & X.T. Han<br />

Revitalization of Existing Building Group by Using Dampers at Top Floor Level and Base 946<br />

Isolators<br />

Z.D. Yang & Eddie S.S. Lam<br />

Summary on the Structural Design of Shenzhen North Train Station 955<br />

X.Y. Fu, B. Wu, Z.H. Chen, M.L. Meng, M. Guo, C. Sun, H.B. Jiang,<br />

Y.W. Feng, J.W. Shao, J.R. Zhou, J.X. Qu & Y.L. Liu<br />

Numerical Modelling of the Cyclic Behaviour of RC Columns Retrofitted with FRP Jackets 970<br />

J.G. Teng, J.Y. Lu, L. Lam, G. Lin & Q.G. Xiao<br />

Direct Tensile Test of Normal Strength Concrete at Elevated Temperature 978<br />

Eddie S. S. Lam, & S. Fang<br />

An Innovative and Proven Solution for Repelling Water Ingress out of Concrete Structures 987<br />

E.C. Chang<br />

Static Performance Analysis of Schwedler Elliptic Suspendome 990<br />

F. Li, C.G. Wang, X.Z. Guo & L.Y. Wang<br />

Monitoring Analysis of Deep Foundation Pit for Xinguoguang Commodity Houses in 996<br />

Wenzhou<br />

C.J. Zhai & T.D. Xia<br />

<strong>The</strong> Influence of Inhomogeneous Initial Stress Field on the Propagation of Longitudinal 1002<br />

Perturbation in Elastic Continuum<br />

W.T. Hu, T.D. Xia & W.Y. Chen<br />

xiv


Keynote Lectures


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

台 湾 海 峡 大 桥 的 构 思<br />

邓 文 中<br />

( 林 同 棪 国 际 , 美 国 旧 金 山 )<br />

摘 要 : 近 年 大 家 对 台 湾 海 峡 连 线 的 兴 趣 愈 来 愈 浓 厚 。 把 台 湾 和 福 建 连 接 起 来 , 无 论 在 政 治 , 经<br />

济 和 社 会 都 有 很 明 显 的 好 处 。 目 前 连 线 有 隧 道 和 大 桥 几 个 不 同 设 想 。 本 文 仅 从 工 程 技 术 的 角 度 讨 论<br />

大 桥 连 线 。<br />

关 键 词 : 台 湾 海 峡 大 桥 海 峡 连 线 跨 海 大 桥 大 跨 径 桥 梁<br />

CONCEPTUALIZTION OF A BRIDGE CROSSING TAIWAN STRAIT<br />

Man-Chung Tang 1<br />

1 T.Y. Lin International, San Francisco, USA<br />

Abstract: People’s interest in building a bridge crossing the Taiwan Strait has been increasing in recent years.<br />

Building a bridge linking Taiwan with Fujian is an excellent idea with respect to business, politics or the<br />

harmony of the society. Currently, there are various tunnel and bridge schemes proposed. This paper explores<br />

only the technological aspect of the bridge schemes.<br />

Keywords: Cross strait bridge, Taiwan Strait Crossing, long span bridge.<br />

一 、 前 言<br />

自 从 中 国 改 革 开 放 以 来 , 经 济 日 益 发 达 , 交 通 的 基 础 建 设 进 展 十 分 快 速 , 高 速 公 路 , 高 速 铁 路<br />

的 建 造 都 在 世 界 领 先 。 尤 其 是 经 济 实 力 日 渐 雄 厚 , 许 多 从 前 只 能 想 象 , 不 敢 奢 望 的 工 程 , 如 三 峡 大<br />

坝 都 已 经 成 为 事 实 。 许 多 过 江 过 海 的 大 桥 和 隧 道 , 也 都 修 建 了 。 长 江 上 已 经 有 一 百 多 座 大 桥 , 杭 州<br />

湾 大 桥 、 东 海 大 桥 等 特 大 跨 海 连 线 也 都 成 功 地 建 造 了 。 大 家 的 目 光 正 关 注 着 更 长 更 大 的 连 线 , 渤 海<br />

湾 , 琼 洲 海 峡 都 有 相 当 深 入 的 研 究 了 。 当 然 , 横 跨 台 湾 海 峡 的 连 线 , 也 是 工 程 界 常 常 谈 及 的 大 事 情 !<br />

台 湾 海 峡 是 台 湾 岛 与 福 建 海 岸 之 间 的 海 峡 。 南 北 长 约 370km, 最 北 端 宽 约 200km, 最 南 端 宽 约<br />

410km, 最 狭 窄 的 位 置 在 台 湾 白 沙 岬 到 福 建 海 坛 岛 之 间 , 距 离 约 125km。 台 湾 海 峡 连 线 的 研 究 , 早<br />

在 上 世 纪 80 年 代 就 已 经 开 始 。 有 不 同 的 隧 道 方 案 和 大 桥 方 案 。 当 然 , 隧 道 和 大 桥 各 有 优 劣 。 日 本<br />

北 海 道 的 青 函 隧 道 和 英 吉 利 海 峡 的 隧 道 建 成 使 用 后 , 大 家 对 隧 道 的 可 行 性 和 经 济 效 益 有 了 较 深 入 的<br />

了 解 。 尤 其 是 由 民 间 集 资 兴 建 的 英 吉 利 海 峡 隧 道 , 在 技 术 上 和 经 济 上 都 可 以 借 鉴 。 英 吉 利 海 峡 隧 道<br />

长 度 只 有 38km,1986 年 开 工 ,1994 年 通 车 , 耗 资 106 亿 英 磅 ( 约 1100 亿 元 人 民 币 )。 和 台 湾 海 峡<br />

隧 道 比 较 , 只 有 30% 的 长 度 。 英 吉 利 海 峡 隧 道 有 三 个 管 道 , 中 间 管 道 是 机 械 和 安 全 管 道 , 每 侧 单 方<br />

向 一 个 管 道 , 只 通 行 火 车 , 是 背 负 式 运 输 。 汽 车 必 需 由 火 车 载 运 。 对 一 般 市 民 言 , 有 其 不 方 便 的 一<br />

面 。 而 且 , 使 用 成 本 比 较 高 。<br />

台 湾 海 峡 最 短 的 连 线 也 有 125km, 建 造 起 来 要 比 英 吉 利 海 峡 隧 道 的 难 度 高 。 而 且 东 侧 属 比 较 强<br />

烈 的 地 震 区 。 一 般 情 况 下 , 隧 道 对 地 震 比 较 不 敏 感 。 但 如 果 由 大 地 震 引 起 断 层 相 对 位 移 , 这 种 破 坏<br />

比 较 不 容 易 修 复 。 大 桥 受 地 震 破 坏 后 的 修 复 会 比 隧 道 容 易 处 理 。 所 以 , 大 桥 方 案 似 乎 比 较 适 宜 。 当<br />

然 , 对 这 样 庞 大 的 工 程 , 详 细 论 证 是 必 要 的 。 海 峡 上 强 风 和 浓 雾 是 大 桥 特 有 的 设 计 重 点 之 一 。 不 过 ,<br />

今 日 工 程 界 对 桥 梁 风 动 力 已 经 有 足 够 的 经 验 , 可 以 设 计 出 有 足 够 抗 风 能 力 的 大 桥 。<br />

二 、 大 桥 方 案<br />

大 家 的 讨 论 和 研 究 集 中 在 三 个 大 桥 线 路 方 案 : 分 别 列 名 为 北 线 , 中 线 和 南 线 〔2〕。 北 线 连 接<br />

福 建 平 潭 岛 与 台 湾 的 新 竹 , 海 上 距 离 全 长 约 125 公 里 , 是 最 短 的 连 线 。 中 线 连 接 福 建 莆 田 南 日 岛 与<br />

台 湾 的 台 中 , 海 上 距 离 全 长 约 140km, 南 线 从 福 建 的 厦 门 经 金 门 , 澎 湖 岛 到 达 台 湾 的 嘉 义 , 全 长<br />

240km, 从 金 门 到 澎 湖 海 上 距 离 全 长 约 140km。 见 图 1。<br />

-1-


图 1 海 峡 大 桥 三 个 路 线<br />

图 1 是 这 三 个 连 线 方 案 的 地 形 截 面 。 把 这 三 条 线 放 在 一 起 。 可 以 比 较 这 三 道 连 线 的 优 劣 。 北 线 ,a,<br />

长 度 比 较 短 , 但 地 形 很 不 平 均 。 南 线 ,c, 地 形 比 北 线 平 顺 , 但 比 较 长 。 只 包 括 从 金 门 到 澎 湖 的 路 段 。<br />

在 澎 湖 与 台 湾 嘉 义 间 仍 有 40 多 公 里 的 距 离 。 水 深 可 能 达 到 100m 以 上 。 中 线 ,b, 的 地 形 是 两 者 之<br />

间 。 从 这 三 个 截 面 看 ,30m 以 下 的 浅 水 区 不 多 , 对 整 座 大 桥 的 设 计 影 响 不 大 。 所 以 , 不 论 是 采 用 哪<br />

一 线 路 , 大 桥 的 建 造 还 是 以 深 水 区 为 主 。<br />

近 年 大 家 注 意 力 集 中 在 北 线 。 北 线 比 较 短 , 应 该 比 较 经 济 。2009 年 , 林 元 培 [1] 建 议 采 用 北 线 。<br />

并 首 先 建 议 建 数 跨 3500m 跨 径 的 悬 索 桥 。 在 工 程 上 北 线 有 很 大 的 优 点 。 但 是 , 在 经 济 和 社 会 方 面 是<br />

否 也 是 最 优 选 择 , 还 有 待 论 证 的 必 要 。 在 技 术 上 , 这 三 条 线 没 有 太 大 的 分 别 。<br />

图 2 北 、 中 、 南 三 线 断 面 [2]<br />

三 、 大 桥 的 交 通 用 途<br />

迄 今 , 大 部 份 工 程 师 的 注 意 力 集 中 在 公 路 上 。 由 于 造 价 昂 贵 , 台 湾 海 峡 在 将 来 相 当 长 一 段 时 间<br />

里 大 概 只 会 建 造 一 条 通 道 , 所 以 , 大 桥 的 功 能 应 该 尽 量 利 用 。 大 桥 的 货 运 功 能 很 重 要 。 而 货 运 在 今<br />

后 能 源 价 格 渐 渐 提 高 的 情 况 下 , 轨 道 将 是 比 较 便 宜 的 运 输 工 具 。 所 以 , 为 了 发 挥 最 大 功 能 , 这 条 通<br />

道 应 该 包 含 轨 道 交 通 。 建 议 大 桥 以 十 个 公 路 车 道 加 两 线 铁 路 轨 道 为 设 计 标 准 。 因 为 公 路 交 通 是 为 车<br />

辆 而 设 , 公 路 设 计 速 度 不 宜 太 低 。 建 议 以 120km/h 为 标 准 ; 这 样 , 驾 驶 汽 车 的 人 会 比 较 舒 适 。 另 一<br />

方 面 , 铁 路 设 计 对 大 桥 的 变 形 的 限 制 相 当 严 格 , 设 计 速 度 愈 高 , 允 许 变 形 的 限 制 愈 严 格 。 大 跨 度 桥<br />

梁 要 能 完 全 满 足 高 速 的 要 求 , 桥 梁 结 构 必 需 有 特 别 高 的 刚 度 , 造 价 就 可 能 很 高 。 在 这 一 段 120 多 公<br />

里 的 路 程 中 , 行 车 时 间 的 影 响 不 大 , 所 以 , 虽 然 今 日 中 国 已 经 有 380km/h 的 高 速 铁 路 , 在 这 条 海 峡<br />

连 线 反 而 不 需 过 于 高 速 。 建 议 以 160km/h 以 内 为 标 准 。<br />

由 于 跨 海 峡 大 桥 很 可 能 是 两 岸 的 唯 一 通 道 , 一 旦 建 成 , 就 必 需 有 全 天 候 使 用 的 功 能 。 使 它 在 浓<br />

雾 和 强 风 之 下 都 能 畅 通 无 阻 。 所 以 , 部 份 车 道 必 需 是 与 外 界 天 气 隔 离 的 结 构 。 在 上 世 纪 70 年 代 设<br />

计 从 美 国 的 阿 拉 斯 加 连 接 俄 国 西 伯 利 亚 的 白 令 海 峡 大 桥 的 时 候 , 已 故 的 林 同 棪 先 生 已 经 对 这 个 桥 梁<br />

全 天 候 的 要 求 有 很 好 的 提 议 , 图 3。 那 就 是 把 车 道 和 轨 道 放 在 一 个 箱 形 主 梁 里 面 。 但 是 , 这 只 能 是<br />

权 宜 之 计 。 驾 驶 员 在 一 个 隧 道 样 子 的 箱 内 行 驶 一 百 多 公 里 的 路 程 , 会 感 觉 十 分 单 调 而 容 易 疲 劳 , 引<br />

发 交 通 意 外 。 所 以 , 在 没 有 大 雾 和 强 风 的 时 候 , 公 路 交 通 仍 然 应 该 行 走 在 桥 面 上 。 近 年 来 对 保 护 车<br />

辆 在 大 风 下 行 驶 的 研 究 已 经 有 相 当 的 了 解 , 专 门 根 据 空 气 动 力 学 得 出 的 特 殊 形 式 的 护 栏 , 可 以 使 车<br />

辆 在 相 当 高 的 风 速 下 仍 然 安 全 行 驶 。 在 台 风 中 , 汽 车 可 以 在 箱 梁 内 减 速 行 驶 。<br />

-2-


图 3 林 同 棪 白 令 海 峡 大 桥 的 构 思<br />

四 、 大 桥 设 计 的 重 点<br />

大 桥 横 跨 海 峡 , 这 里 天 然 环 境 的 影 响 主 要 是 强 风 , 地 震 , 浓 雾 , 深 水 和 船 舶 撞 击 。 中 国 已 经 建<br />

成 了 许 多 大 跨 度 的 大 桥 , 对 强 风 和 地 震 的 结 构 设 计 已 经 有 相 当 深 入 的 研 究 。 如 苏 通 大 桥 , 西 堠 门 大<br />

桥 等 , 都 是 强 风 地 带 的 大 桥 。<br />

这 个 区 域 的 地 震 烈 度 不 太 高 , 而 且 , 大 跨 度 桥 梁 比 较 柔 软 , 地 震 对 结 构 的 影 响 比 较 低 。 但 是 ,<br />

由 于 这 座 大 桥 建 成 后 对 两 岸 交 通 十 分 重 要 , 大 桥 在 大 地 震 后 必 需 保 留 持 续 通 行 的 能 力 。 旧 金 山 海 湾<br />

大 桥 东 段 的 设 计 理 念 , 或 许 可 以 借 鉴 。 该 桥 位 于 美 国 最 强 烈 的 地 震 区 , 而 这 地 区 又 都 是 极 疏 松 的 沉<br />

积 层 。 基 础 主 要 是 100 多 米 长 2.5m 直 径 的 钢 管 混 凝 土 桩 。 在 Bangladesh 的 Jamuna 大 桥 我 们 也 用 了<br />

100 多 米 长 ,3.0m 和 3.50m 直 径 的 钢 管 混 凝 土 桩 。 都 很 成 功 。 这 种 大 直 径 的 长 桩 在 强 烈 的 地 震 下 的<br />

表 现 很 好 。 台 湾 海 峡 的 水 深 基 本 上 在 80m 以 下 。 可 以 考 虑 用 4.0m 直 径 的 钢 管 桩 。 通 常 , 这 种 钢 管<br />

桩 内 有 时 只 有 部 份 灌 满 钢 筋 混 凝 土 。 但 在 80m 深 水 , 大 概 有 全 长 满 灌 的 需 要 。<br />

旧 金 山 海 湾 大 桥 的 结 构 理 念 尤 其 值 得 研 究 比 较 。 那 就 是 强 烈 地 震 后 结 构 的 性 能 。 现 行 美 国<br />

AASHTO 的 地 震 设 计 的 基 本 要 求 是 在 最 强 烈 的 地 震 下 保 障 人 的 性 命 安 全 。 所 以 允 许 一 次 性 的 保 护 措<br />

施 , 例 如 , 可 以 断 裂 的 连 接 。 这 样 设 计 的 基 础 是 在 地 震 产 生 的 力 大 于 一 定 的 数 值 时 , 某 些 指 定 的 杆<br />

件 或 连 接 点 就 会 断 裂 , 使 结 构 的 频 率 降 低 , 减 低 了 结 构 对 地 震 的 反 应 。 但 这 些 连 接 点 必 需 在 下 次 地<br />

震 来 临 之 前 修 复 , 否 则 结 构 就 会 发 生 问 题 。 这 个 缺 点 在 最 近 的 纽 西 兰 大 地 震 就 表 现 出 来 。 许 多 建 筑<br />

物 在 第 一 次 地 震 时 损 伤 不 大 , 但 在 后 来 烈 度 明 显 较 低 的 余 震 时 受 到 的 破 坏 远 远 超 过 原 来 地 震 引 起 的<br />

破 坏 。 基 本 原 因 很 可 能 是 因 为 第 一 次 地 震 时 有 些 用 来 防 御 的 措 施 已 经 受 到 破 坏 , 这 些 防 御 措 施 只 能<br />

使 用 一 次 。 在 未 修 复 之 前 来 一 个 相 当 大 的 余 震 就 没 法 对 结 构 提 供 保 护 了 。 所 以 , 对 于 比 较 重 要 的 结<br />

构 , 如 台 湾 海 峡 大 桥 , 这 个 理 念 不 适 宜 应 用 。<br />

旧 金 山 海 湾 大 桥 的 设 计 要 求 在 最 大 设 计 地 震 后 几 小 时 内 安 全 通 车 。 这 个 要 求 应 该 也 适 用 于 台 湾<br />

海 峡 大 桥 上 。 为 了 满 足 这 个 要 求 , 我 们 发 展 了 剪 力 杆 的 应 用 : 这 些 连 接 四 根 垂 直 塔 柱 的 短 横 梁 , 在<br />

强 烈 地 震 下 会 产 生 塑 性 铰 , 使 垂 直 的 柱 可 以 接 受 足 够 的 水 平 位 移 而 仍 然 处 于 弹 性 状 态 。 在 海 湾 大 桥 ,<br />

我 们 有 两 个 安 全 的 保 障 : 第 一 , 钢 横 梁 发 生 塑 性 变 形 并 不 等 于 破 坏 , 它 们 仍 然 可 以 继 续 承 受 周 期 性<br />

的 变 形 ; 第 二 , 我 们 在 塔 上 放 置 了 许 多 根 短 横 梁 , 在 一 些 短 横 梁 受 到 破 坏 之 后 其 余 的 短 横 梁 仍 可 以<br />

继 续 工 作 。 所 以 , 在 大 地 震 后 就 算 不 立 即 修 复 , 结 构 的 荷 载 能 力 并 没 有 减 低 。 对 余 震 仍 然 有 足 够 的<br />

保 障 能 力 。 相 似 的 理 念 可 以 应 用 在 海 峡 大 桥 上 。<br />

现 在 一 般 大 桥 的 设 计 寿 命 是 100 年 。 重 要 的 工 程 设 计 寿 命 会 有 所 提 高 : 旧 金 山 海 湾 大 桥 的 设 计<br />

寿 命 是 150 年 , 港 珠 澳 大 桥 是 120 年 。 台 湾 海 峡 大 桥 的 投 资 和 重 要 性 都 更 高 , 设 计 寿 命 应 该 至 少 是<br />

150 年 。<br />

五 、 桥 梁 的 跨 径 的 确 定<br />

在 这 里 我 们 先 确 定 通 航 净 空 的 要 求 。 这 里 是 国 际 水 道 , 在 这 里 建 桥 应 该 提 供 足 够 的 净 空 , 让 世<br />

界 各 国 和 各 类 的 船 舶 通 行 无 阻 。 这 样 一 来 , 大 桥 的 净 空 就 应 该 与 其 他 国 际 水 道 上 的 大 桥 相 同 , 但 是 ,<br />

近 年 造 船 业 似 乎 有 建 造 更 大 的 船 的 兴 趣 , 所 以 , 建 议 应 该 把 通 航 净 空 提 高 到 72m。 这 个 净 空 只 需 要<br />

用 在 主 航 道 上 。 其 他 地 方 净 空 可 以 降 低 。 但 既 然 这 里 是 宽 阔 的 海 面 , 在 非 通 航 孔 净 空 也 不 宜 太 低 。<br />

这 个 问 题 需 要 专 家 论 证 。 我 觉 得 ,40m 也 许 是 个 适 当 的 数 字 。<br />

大 桥 主 航 道 的 通 航 宽 度 决 定 大 桥 主 跨 的 最 小 允 许 跨 径 。 国 际 水 道 一 般 要 求 65mX1500m 的 通 航<br />

孔 。 这 个 当 然 也 需 要 通 航 论 证 。 但 从 其 他 国 际 水 道 情 况 比 较 ,2000m 的 跨 径 是 应 当 足 够 的 。 到 现 在<br />

为 止 , 在 国 际 水 道 、 世 界 最 大 的 跨 径 是 日 本 明 石 大 桥 的 1991m。 其 次 是 丹 麦 的 Storebelt, 主 跨 1624m。<br />

我 国 西 候 门 大 桥 跨 径 1650m, 比 Storebelt 大 , 但 不 是 在 国 际 水 道 上 。 大 桥 的 价 格 随 跨 径 的 增 大 急 剧<br />

上 升 , 在 没 有 需 要 的 情 况 下 应 该 尽 量 把 跨 径 减 少 。<br />

-3-


非 主 通 航 孔 的 跨 径 应 该 根 据 经 济 效 率 决 定 。 这 里 , 基 础 必 需 能 够 承 受 船 舶 的 撞 击 、 台 风 地 震 侵<br />

袭 、 水 流 的 冲 击 等 , 水 愈 深 , 基 础 的 刚 度 就 要 愈 大 , 也 就 自 然 能 够 承 担 愈 大 跨 径 的 桥 梁 。 也 就 是 说 ,<br />

水 愈 深 , 相 对 应 的 经 济 效 益 最 高 的 跨 径 就 愈 大 。 根 据 上 面 图 2 的 地 形 截 面 , 真 正 的 浅 水 区 范 围 很 小 。<br />

大 部 份 水 深 在 40m 到 60m 之 间 。 在 实 际 设 计 的 时 候 , 当 然 要 仔 细 研 判 什 么 是 最 经 济 的 桥 梁 跨 径 。<br />

从 经 验 看 ,40m 到 60m 水 深 的 最 经 济 的 跨 度 应 该 在 500m 到 700m 之 间 。<br />

根 据 调 查 显 示 , 海 峡 上 船 只 航 行 主 要 集 中 在 海 峡 中 线 以 西 海 域 , 即 接 近 福 建 方 面 。 我 们 可 以 在<br />

这 里 设 置 两 个 2000m 的 主 跨 。 另 外 在 东 侧 也 设 置 一 个 2000m 的 主 跨 。 其 余 部 份 则 定 为 非 通 航 孔 ,<br />

建 500m 到 700m 跨 径 的 桥 梁 。<br />

六 、 桥 型 选 择<br />

能 够 提 供 2000m 主 跨 的 桥 型 是 斜 拉 桥 和 悬 索 桥 。 目 前 的 造 桥 技 术 对 这 两 种 桥 型 建 造 2000m 的<br />

跨 径 没 有 问 题 。 经 济 效 益 是 决 定 悬 索 桥 抑 或 斜 拉 桥 的 主 因 。 这 两 种 桥 型 最 大 的 分 别 是 锚 碇 。 斜 拉 桥<br />

主 梁 受 压 , 不 需 要 锚 碇 。 悬 索 桥 主 缆 庞 大 的 拉 力 必 需 由 锚 碇 传 到 地 下 。 由 于 这 里 水 很 深 , 这 些 在 桥<br />

面 上 的 水 平 力 要 由 锚 碇 传 递 到 海 底 岩 层 , 锚 碇 的 建 造 费 用 会 很 高 。 所 以 , 在 这 个 情 况 下 ,2000m 主<br />

跨 的 斜 拉 桥 会 比 较 适 宜 。<br />

图 4 2000m 跨 径 斜 拉 桥<br />

主 航 道 的 位 置 必 需 经 过 验 证 来 确 定 。 调 查 显 示 现 在 在 海 峡 航 行 的 船 只 没 有 规 范 。 如 果 要 建 桥 ,<br />

船 只 的 航 道 就 必 需 予 以 规 范 ; 建 议 安 排 南 向 和 北 向 各 一 个 主 航 道 。 在 2000m 的 跨 径 下 , 每 个 单 向 航<br />

道 可 以 定 为 1600m。 南 北 航 道 中 间 应 该 有 一 个 至 少 1000m 的 分 隔 带 。 根 据 这 些 要 求 , 我 们 建 议 如 图<br />

4 的 斜 拉 桥 的 结 构 。 假 如 两 个 单 向 航 道 之 间 需 要 更 大 的 隔 离 带 , 可 以 使 用 图 4b 的 结 构 形 式 。2000m<br />

跨 径 的 斜 拉 桥 上 的 拉 索 很 长 , 拉 索 垂 度 会 引 起 这 些 长 拉 索 刚 度 下 降 。 笔 者 对 这 个 问 题 曾 提 出 过 不 同<br />

的 对 策 〔3, 图 5〕。<br />

非 通 航 孔 如 果 用 500m 到 700m 的 跨 径 , 斜 拉 桥 也 应 该 是 一 个 最 适 当 的 桥 型 。500m 和 以 上 的 跨<br />

径 , 梁 桥 和 拱 桥 都 不 经 济 。 所 以 也 是 悬 索 桥 与 斜 拉 桥 之 间 的 选 择 。 基 於 与 主 跨 相 同 的 原 因 , 建 议 采<br />

用 斜 拉 桥 。 这 些 桥 梁 也 可 以 用 和 主 跨 相 同 的 结 构 型 式 。<br />

七 、 主 梁 截 面<br />

大 跨 径 桥 梁 对 主 梁 的 宽 度 有 一 定 的 要 求 。 在 横 向 力 的 作 用 下 , 主 梁 基 本 上 是 一 条 连 续 梁 , 太 狭<br />

窄 的 主 梁 会 过 于 柔 弱 , 横 向 变 形 太 大 。 在 风 力 作 用 下 不 能 满 足 稳 定 要 求 。 尤 其 是 轨 道 桥 梁 。 所 以 ,<br />

截 面 的 宽 度 必 需 根 据 最 大 跨 径 来 确 定 。 而 不 单 是 满 足 交 通 的 要 求 。 笔 者 曾 建 议 过 几 种 增 加 横 向 刚 度<br />

的 方 法 [3, 图 6] , 需 要 时 可 以 应 用 在 海 峡 大 桥 上 。 如 果 宽 度 还 不 够 , 可 以 像 旧 金 山 海 湾 大 桥 , 西<br />

候 门 大 桥 和 正 在 建 造 的 意 大 利 麦 仙 娜 大 桥 一 样 , 把 主 梁 分 成 两 个 或 三 个 用 横 梁 连 系 的 箱 梁 , 增 加 桥<br />

面 的 宽 度 〔3〕。<br />

图 5 提 高 拉 索 效 率 方 法<br />

图 6 加 强 横 向 刚 度 方 法<br />

-4-


上 面 谈 及 , 大 桥 可 以 设 置 十 个 公 路 车 道 和 两 线 轨 道 。 不 设 人 行 道 和 非 机 动 车 道 。 但 应 该 设 置 维<br />

修 用 的 通 道 。 车 道 分 为 上 下 两 层 : 上 层 6 个 公 路 车 道 , 下 层 两 个 轨 道 在 中 间 , 两 边 各 2 个 公 路 车 道 。<br />

加 上 路 肩 和 分 隔 带 , 主 梁 宽 度 约 是 35m。 图 75。 日 本 的 明 石 大 桥 跨 径 1991m, 桥 宽 也 是 35m。 如 果<br />

我 们 采 用 2000m 的 跨 径 , 采 用 和 苏 通 大 桥 相 似 的 倒 Y 形 或 者 倒 V 形 桥 塔 , 可 以 协 助 增 加 部 份 横 向<br />

刚 度 。 所 以 ,35m 的 主 梁 宽 度 应 该 可 以 满 足 横 向 刚 度 的 要 求 。<br />

图 7 主 梁 截 面<br />

图 8 加 上 保 护 外 壳<br />

图 7 是 一 个 钢 结 构 的 截 面 。2000m 跨 径 的 大 桥 , 在 今 日 的 环 境 下 。 混 凝 土 结 构 会 太 沉 重 。 其 他<br />

能 代 替 钢 的 新 材 料 的 应 用 还 不 成 熟 。 所 以 大 跨 径 桥 梁 还 必 需 用 钢 结 构 。 甚 至 500m 以 上 跨 径 的 大 桥 ,<br />

钢 梁 还 是 比 较 适 宜 。 但 不 管 是 用 钢 抑 或 用 混 凝 土 , 保 养 和 防 腐 都 是 重 要 的 课 题 。<br />

为 了 满 足 全 天 候 的 要 求 , 可 以 给 桁 架 加 上 一 个 外 壳 。 图 8。 这 个 外 壳 并 不 需 要 把 箱 梁 与 外 界 完<br />

全 隔 离 。 只 需 要 达 到 下 层 桥 面 行 车 不 受 风 雨 和 雾 的 干 扰 , 就 足 够 了 。 所 以 , 要 尽 可 能 让 光 线 和 空 气<br />

进 入 , 使 在 箱 内 的 乘 客 可 以 感 受 到 外 界 的 天 气 和 环 境 , 例 如 使 用 部 份 透 明 和 半 透 明 的 外 壳 , 当 然 更<br />

好 ! 这 个 外 壳 对 主 梁 的 防 腐 也 可 以 起 很 大 的 作 用 。 外 壳 顶 上 的 平 面 可 以 设 置 太 阳 能 板 , 补 助 部 份 电<br />

力 的 需 求 。<br />

八 、 附 属 设 施<br />

在 一 百 多 公 里 的 长 桥 上 , 会 需 要 不 同 的 附 属 设 施 如 照 明 , 卫 生 , 休 息 , 加 油 , 和 环 保 等 等 。 还<br />

有 电 能 的 供 应 。 近 年 汽 车 设 计 趋 向 电 能 , 这 座 超 百 年 大 计 的 通 道 , 也 应 该 考 虑 相 应 的 设 施 。 至 于 能<br />

源 的 供 应 , 应 当 取 之 于 当 地 环 境 , 例 如 风 能 , 太 阳 能 , 水 流 发 电 等 。 这 里 应 该 可 以 自 给 自 足 。<br />

九 、 施 工 和 估 算<br />

这 是 一 座 前 所 未 有 的 大 桥 。 大 桥 愈 长 , 就 愈 能 发 展 出 大 型 的 施 工 工 具 。 因 为 重 复 使 用 的 次 数 愈<br />

多 , 每 单 元 的 使 用 价 格 就 愈 低 。 杭 州 湾 大 桥 上 使 用 的 大 型 整 跨 吊 装 方 法 , 可 以 用 在 大 约 十 余 公 里 的<br />

浅 水 区 。 这 里 桥 的 跨 径 比 较 小 , 整 跨 或 半 跨 吊 装 不 是 问 题 。500m 到 2000m 跨 径 的 斜 拉 桥 , 整 体 吊<br />

装 就 不 可 能 了 。 但 是 , 桥 塔 仍 然 可 以 全 部 或 部 份 预 制 拼 装 。 减 短 工 期 。<br />

跨 径 对 大 桥 的 造 价 有 很 大 的 影 响 ,3000m 跨 径 的 大 桥 比 2000m 跨 径 的 大 桥 可 能 要 贵 好 几 倍 。 工<br />

程 师 应 该 避 免 竞 争 世 界 纪 录 , 浪 费 公 帑 。<br />

业 界 对 这 个 连 线 的 估 算 大 都 在 3000 亿 和 5000 亿 元 人 民 币 之 间 。 单 从 收 费 来 维 持 大 概 不 可 能 。<br />

但 我 们 相 信 , 这 样 一 座 桥 对 两 岸 经 济 、 政 治 和 社 会 的 影 响 可 是 远 远 超 过 这 个 价 值 !<br />

九 、 结 语<br />

在 台 湾 海 峡 建 一 座 大 桥 , 把 台 湾 和 福 建 连 接 起 来 , 应 该 是 工 程 师 们 和 非 工 程 师 们 历 来 的 梦 想 。<br />

但 是 , 以 今 天 中 国 造 桥 技 术 的 水 平 和 两 岸 人 民 的 经 济 能 力 , 这 个 梦 想 是 可 以 实 现 的 !<br />

参 考 文 献<br />

[1] 林 元 培 , 窦 文 俊 . 台 湾 海 峡 大 桥 - 全 天 候 通 道 方 案 . 2010 年 全 国 科 协 论 坛 , 2010 年 10 月 . 福 州 .f<br />

[2] 李 学 杰 , 张 以 , 冯 志 , 郭 连 生 , 周 昌 范 , 万 荣 胜 . 台 湾 海 峡 地 形 地 质 特 征 及 其 通 道 工 程 选 线 . 科 技 导 报 . 2008,<br />

26(22).<br />

[3] 邓 文 中 . (Man-Chung Tang). Bridge Forms and Aesthetics. IABMAS Conference, Porto, Portugal, 2006.<br />

-5-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

现 代 大 跨 空 间 结 构 在 中 国 大 陆 的 应 用 与 发 展<br />

董 石 麟 邢 栋 赵 阳<br />

( 浙 江 大 学 空 间 结 构 研 究 中 心 , 浙 江 杭 州 310027)<br />

摘 要 : 现 代 大 跨 空 间 结 构 大 致 是 在 二 十 世 纪 七 、 八 十 年 代 左 右 , 基 于 采 用 轻 质 高 强 的 膜 材 、 钢<br />

索 、 钢 棒 , 应 用 新 技 术 而 发 展 起 来 的 轻 盈 、 高 效 的 结 构 体 系 , 诸 如 气 承 式 充 气 膜 结 构 、 索 膜 结 构 、<br />

索 桁 结 构 、 张 弦 梁 结 构 、 弦 支 网 壳 结 构 、 索 穹 顶 结 构 、 索 穹 顶 ¬— 网 壳 组 合 结 构 、 张 弦 气 肋 梁 结 构<br />

等 。 另 一 方 面 , 对 于 二 十 世 纪 中 叶 获 得 大 量 应 用 的 近 代 空 间 结 构 , 如 薄 壳 结 构 、 网 架 结 构 、 网 壳 结<br />

构 、 一 般 悬 索 结 构 等 , 通 过 采 用 多 种 结 构 形 式 和 建 筑 材 料 的 组 合 、 预 应 力 技 术 、 结 构 概 念 和 形 体 的<br />

创 新 , 从 而 提 出 并 得 到 工 程 实 践 应 用 的 新 颖 空 间 结 构 体 系 也 可 认 为 是 现 代 空 间 结 构 , 诸 如 组 合 网 架<br />

结 构 、 空 腹 网 架 ( 壳 ) 结 构 、 多 面 体 空 间 刚 架 结 构 、 局 部 双 层 网 壳 结 构 、 斜 拉 网 架 ( 壳 ) 结 构 、 树<br />

状 结 构 、 预 应 力 装 配 弓 式 结 构 等 。 进 而 本 文 按 空 间 结 构 单 元 组 成 分 类 为 序 , 分 别 对 现 代 刚 性 空 间 结<br />

构 、 现 代 柔 性 空 间 结 构 、 现 代 刚 柔 性 组 合 空 间 结 构 的 特 点 及 在 中 国 大 陆 的 应 用 与 发 展 作 了 综 述 。<br />

关 键 词 : 现 代 大 跨 空 间 结 构 刚 性 空 间 结 构 柔 性 空 间 结 构 刚 柔 性 组 合 空 间 结 构 应<br />

用 和 发 展<br />

APPLICATION AND DEVELOPMENT OF MODERN LONG-SPAN SPACE<br />

STRUCTURES IN MAINLAND CHINA<br />

DONG Shi-lin 1 , XING Dong 1 and ZHAO Yang 1<br />

1 Space Structures Research Center, Zhejiang <strong>University</strong>, Hangzhou 310027, China<br />

Abstract: Modern long-span space structures, developed from 1970s and 1980s, are light and effective<br />

structures on the basis of new technologies and light-weight high-strength materials such as membranes and<br />

steel cables. <strong>The</strong>se structures include air-supported membrane structures, cable-membrane structures, cable truss<br />

structures, beam string structures, suspend-domes, cable domes, composite structures of cable dome and<br />

single-layer lattice shell, tensairity structures and so on. On the other hand, for premodern space structures<br />

widely used from the mid-twentieth century (such as thin shells, space trusses, lattice shells and ordinary cable<br />

structures), new space structures developed by the combination of different structural forms and materials, the<br />

application of prestressing technology and the innovation of structural concepts and configurations also belong<br />

to modern space structures. <strong>The</strong>se structures include composite space trusses, open-web grid structures,<br />

polyhedron space frame structures, partial double-layer lattice shells, cable-stayed grid structures, tree-type<br />

structures, prestressed segmental steel structures and so on. This paper provides a review of the structural<br />

characteristics and practical applications in mainland China of modern rigid space structures, modern flexible<br />

space structures and modern rigid-flexible combined space structures.<br />

Keywords: Modern long-span space structures, rigid space structures, flexible space structures, rigid-flexible<br />

combined space structures, application and development.<br />

一 、 引 言<br />

空 间 结 构 的 发 展 历 史 可 认 为 分 成 三 个 阶 段 , 即 古 代 空 间 结 构 、 近 代 空 间 结 构 和 现 代 空 间 结 构 [1~6] ,<br />

分 割 的 时 间 节 点 大 致 为 1925 年 、1975 年 前 后 。 这 是 基 于 1925 年 在 德 国 耶 拿 玻 璃 厂 建 成 历 史 上 第 一<br />

幢 40m 直 径 的 钢 筋 混 凝 土 薄 壳 结 构 ,1924 年 在 德 国 蔡 司 天 文 馆 建 成 世 界 上 首 个 直 径 为 15m 的 半 球<br />

形 单 层 钢 ( 生 铁 ) 网 壳 以 及 1975 年 在 美 国 庞 蒂 亚 克 建 成 很 有 代 表 性 的 巨 型 首 例 168m×220m 气 承<br />

式 充 气 膜 结 构 体 育 馆 。 可 以 说 , 现 代 大 跨 空 间 结 构 大 致 是 在 二 十 世 纪 七 、 八 十 年 代 左 右 , 基 于 采 用<br />

轻 质 高 强 的 膜 材 、 钢 索 、 钢 棒 , 应 用 新 技 术 而 发 展 起 来 的 轻 盈 、 高 效 的 结 构 体 系 , 诸 如 气 承 式 充 气<br />

膜 结 构 、 索 膜 结 构 、 索 桁 结 构 、 张 弦 梁 结 构 、 弦 支 穹 顶 结 构 、 索 穹 顶 结 构 等 。<br />

对 于 二 十 世 纪 中 叶 获 得 大 量 应 用 的 近 代 空 间 结 构 , 如 薄 壳 结 构 、 网 架 结 构 、 网 壳 结 构 、 一 般 悬<br />

索 结 构 等 , 大 致 从 二 十 世 纪 七 、 八 十 年 代 起 , 通 过 采 用 多 种 结 构 形 式 和 建 筑 材 料 的 组 合 而 协 同 工 作 、<br />

预 应 力 技 术 、 结 构 概 念 和 形 体 的 创 新 , 从 而 提 出 并 得 到 工 程 实 践 应 用 的 新 颖 空 间 结 构 体 系 , 也 可 认<br />

-6-


为 是 现 代 空 间 结 构 , 诸 如 组 合 网 架 结 构 、 斜 拉 网 格 ( 壳 ) 预 应 力 网 格 ( 壳 ) 结 构 、 局 部 双 层 网 壳 结<br />

构 、 树 状 结 构 、 预 应 力 装 配 弓 式 结 构 多 面 体 空 间 刚 架 结 构 等 。 因 此 , 现 代 空 间 结 构 是 图 1 中 的 Ⅰ、<br />

Ⅱ( 图 中 用 斜 线 表 示 ) 两 部 分 结 构 体 系 组 成 。 图 1 中 还 表 明 , 近 代 空 间 结 构 ( 图 中 用 小 黑 点 表 示 )<br />

在 当 前 与 现 代 空 间 结 构 一 起 仍 获 得 应 用 ( 其 中 薄 壳 结 构 的 应 用 较 少 )。<br />

图 1 空 间 结 构 年 代 划 分 图<br />

按 组 成 空 间 结 构 的 基 本 单 元 分 类 , 即 按 板 壳 单 元 、 梁 单 元 、 杆 单 元 三 种 刚 性 单 元 和 索 单 元 、 膜<br />

单 元 两 种 柔 性 单 元 分 类 , 现 代 空 间 结 构 又 可 分 为 现 代 刚 性 空 间 结 构 、 现 代 柔 性 空 间 结 构 和 现 代 刚 柔<br />

性 组 合 空 间 结 构 三 大 类 , 文 中 对 这 三 类 现 代 空 间 结 构 主 要 形 体 、 结 构 特 性 、 在 中 国 大 陆 的 应 用 与 发<br />

展 分 别 作 了 阐 述 。<br />

二 、 现 代 空 间 结 构 的 分 类<br />

现 代 空 间 结 构 的 分 类 必 然 要 涉 及 空 间 结 构 的 分 类 。 近 年 来 , 国 内 外 空 间 结 构 蓬 勃 发 展 , 建 筑 选<br />

型 新 颖 、 形 式 和 种 类 繁 多 , 按 传 统 的 空 间 结 构 形 式 和 分 类 方 法 , 即 把 空 间 结 构 划 分 为 薄 壳 结 构 、 网<br />

架 结 构 、 网 壳 结 构 、 悬 索 结 构 、 薄 膜 结 构 共 五 类 空 间 结 构 的 分 类 方 法 ( 图 2) 已 很 难 包 络 和 反 映 现<br />

有 各 种 形 式 的 空 间 结 构 。<br />

据 统 计 , 国 内 外 现 有 各 种 形 式 的 空 间 空 间 结 构 共 38 种 , 它 们 都 有 具 体 的 名 称 , 并 在 工 程 实 践<br />

中 获 得 应 用 。 如 采 用 组 成 空 间 结 构 的 基 本 构 件 或 基 本 单 元 即 板 壳 单 元 、 梁 单 元 、 杆 单 元 、 索 单 元 和<br />

膜 单 元 来 分 类 如 图 3 所 示 [1,7,8] , 就 可 避 免 传 统 分 类 方 法 的 局 限 性 , 而 且 具 有 鲜 明 的 直 观 性 、 实 用 性 、<br />

包 容 性 和 开 放 性 。 因 此 , 空 间 结 构 按 基 本 单 元 组 成 分 类 , 不 仅 可 确 知 各 种 形 式 空 间 结 构 的 组 成 , 而<br />

且 可 初 步 框 定 利 用 哪 些 计 算 方 法 和 程 序 进 行 结 构 分 析 ; 它 不 仅 可 包 络 当 前 所 有 各 种 形 式 的 空 间 结<br />

构 , 而 且 也 可 包 容 、 囊 括 今 后 开 发 和 创 造 的 新 型 空 间 结 构 。<br />

薄 壳 结 构<br />

( 包 括 折 板 结 构 )<br />

网 架 结 构<br />

五 大 空 间 结 构<br />

网 壳 结 构<br />

悬 索 结 构<br />

充 气 膜 结 构<br />

气 囊 式<br />

气 承 式<br />

薄 膜 结 构<br />

支 承 膜 结 构<br />

图 2 按 传 统 方 法 分 类 的 空 间 结 构<br />

刚 性 ( 骨 架 式 )<br />

柔 性 ( 张 拉 式 )<br />

-7-


图 3 空 间 结 构 按 单 元 组 成 分 类<br />

图 3 还 表 明 , 由 于 板 壳 单 元 、 梁 单 元 和 杆 单 元 可 认 为 是 刚 性 单 元 , 索 单 元 和 膜 单 元 可 认 为 是 柔<br />

性 单 元 , 因 而 各 种 具 体 形 式 的 空 间 结 构 又 可 归 属 为 由 刚 性 单 元 组 成 的 刚 性 空 间 结 构 ( 图 3 中 用 实 线<br />

框 框 住 的 结 构 名 称 , 计 7 小 类 共 17 中 刚 性 空 间 结 构 )、 由 柔 性 单 元 组 成 的 柔 性 空 间 结 构 ( 图 3 中 用<br />

虚 线 框 框 住 的 结 构 名 称 , 计 3 小 类 共 5 种 柔 性 空 间 结 构 ) 和 由 刚 、 柔 性 单 元 杂 交 组 合 而 成 的 刚 柔 性<br />

组 合 空 间 结 构 ( 图 3 中 用 半 实 线 框 半 虚 线 框 框 住 的 结 构 名 称 , 计 7 小 类 共 16 种 刚 柔 性 组 合 空 间 结<br />

构 ) 三 大 类 空 间 结 构 。<br />

由 二 十 世 纪 七 、 八 十 年 代 发 展 起 来 的 现 代 空 间 结 构 ( 由 图 1 的 Ⅰ、Ⅱ 两 部 分 组 成 ), 其 鲜 明 的<br />

特 点 是 轻 盈 、 高 效 、 创 新 和 实 用 , 可 划 分 为 现 代 刚 性 空 间 结 构 ( 共 5 种 )、 现 代 柔 性 空 间 结 构 ( 共 2<br />

种 ) 和 现 代 刚 柔 性 组 合 空 间 结 构 ( 共 10 种 ) 共 三 大 类 17 种 空 间 结 构 , 详 见 图 3 暗 色 矩 形 框 中 所 表<br />

示 的 空 间 结 构 。 它 们 的 特 点 、 应 用 与 发 展 分 别 在 以 下 各 节 中 说 明 。<br />

三 、 现 代 刚 性 空 间 结 构 的 应 用 与 发 展<br />

现 代 刚 性 空 间 结 构 是 指 空 腹 网 壳 、 树 状 结 构 、 多 面 体 空 间 刚 架 结 构 、 局 部 双 层 网 壳 与 组 合 网 架<br />

共 五 种 , 现 分 别 说 明 如 下 。<br />

3.1. MR1 空 腹 网 壳<br />

一 般 仅 由 梁 单 元 便 可 构 成 空 腹 网 壳 。 它 是 由 平 面 空 腹 桁 架 发 展 为 空 腹 网 架 [10] , 再 发 展 为 曲 面 形<br />

空 腹 网 壳 , 且 以 两 向 正 交 或 斜 交 的 空 腹 网 壳 居 多 。 空 腹 网 壳 的 上 、 下 弦 节 点 通 常 为 五 根 杆 件 相 交 。<br />

空 腹 部 分 能 改 善 受 力 性 能 、 节 省 材 料 用 量 外 , 尚 可 兼 作 设 备 层 用 , 因 而 建 筑 师 乐 于 采 用 空 腹 网 壳 作<br />

为 屋 盖 结 构 。<br />

典 型 的 也 是 世 界 上 跨 度 最 大 的 空 腹 网 壳 是 椭 圆 平 面 142m×212m 的 国 家 大 剧 院 椭 球 屋 盖 , 矢 高<br />

46m, 采 用 由 144 榀 径 向 空 腹 拱 ( 上 、 下 弦 及 腹 杆 截 面 均 选 用 60mm×200mm 厚 钢 板 ) 及 环 向 钢 管<br />

拼 装 而 成 。 通 过 分 析 计 算 , 设 计 方 接 受 了 我 们 的 建 议 , 在 对 角 方 向 设 置 四 道 大 型 交 叉 上 、 下 弦 杆 ,<br />

以 提 高 抗 扭 能 力 和 结 构 的 整 体 稳 定 性 ( 见 图 4)。<br />

-8-


3.2. MR2 树 状 结 构<br />

(a) 计 算 模 型<br />

图 4 国 家 大 剧 院<br />

(b) 在 施 工 安 装 中<br />

这 是 近 年 来 采 用 的 一 种 新 结 构 , 它 实 际 上 是 一 种 多 级 分 枝 的 立 柱 支 承 结 构 , 主 杆 和 枝 杆 均 可 由<br />

梁 单 元 集 成 , 枝 杆 顶 端 与 屋 盖 结 构 相 连 , 可 减 小 屋 盖 结 构 的 跨 度 、 降 低 结 构 内 力 。<br />

深 圳 文 化 中 心 前 厅 采 用 粗 枝 、 中 枝 、 端 枝 三 级 分 枝 的 树 状 结 构 ( 图 5)。 杭 州 奥 体 中 心 体 育 场 悬<br />

臂 屋 盖 一 方 案 采 用 一 圈 28 根 二 级 分 叉 的 树 状 结 构 ( 图 6), 台 湾 台 北 到 高 雄 的 高 铁 台 南 等 车 站 大 厅<br />

也 采 用 了 树 状 结 构 。<br />

图 5 深 圳 文 化 中 心 前 厅<br />

[9]<br />

3.3. MR3 多 面 体 空 间 框 架 结 构<br />

图 6 杭 州 奥 体 中 心 体 育 场 一 方 案<br />

二 级 分 叉 的 树 状 结 构<br />

这 是 一 种 全 新 的 结 构 体 系 , 由 12 面 体 、14 面 体 组 成 的 无 穷 个 多 面 体 与 切 割 面 ( 屋 盖 上 下 表 面<br />

与 墙 体 内 外 表 面 ) 的 相 交 线 及 屋 盖 与 墙 体 内 多 面 体 的 棱 边 , 共 同 构 成 空 间 结 构 的 骨 架 ( 图 7)。 这 种<br />

结 构 内 部 每 个 节 点 仅 有 四 根 杆 件 相 交 , 适 宜 于 用 在 以 最 少 的 节 点 数 和 杆 件 数 去 填 充 一 定 厚 度 的 厚 板<br />

或 三 维 体 结 构 。 多 面 体 空 间 刚 结 构 的 每 个 杆 件 是 空 间 梁 单 元 , 而 且 必 须 是 空 间 梁 单 元 , 致 使 仍 能 承<br />

载 和 传 递 各 方 向 的 外 力 作 用 。<br />

2008 年 北 京 奥 运 会 国 家 游 泳 中 心 “ 水 立 方 ” 采 用 了 这 种 多 面 体 空 间 框 架 结 构 , 也 是 世 界 上 的 首<br />

例 ( 图 8), 平 面 尺 寸 177m×177m, 高 30m。 表 面 杆 件 采 用 矩 形 钢 管 , 并 选 用 半 鼓 半 球 焊 接 空 心 球<br />

节 点 相 连 接 , 便 于 铺 设 充 气 枕 , 内 部 杆 件 采 用 圆 钢 管 。 选 用 焊 接 空 心 球 节 点 , 便 于 连 接 构 造 。 国 家<br />

游 泳 中 心 “ 水 立 方 ” 获 得 国 际 桥 梁 与 结 构 协 会 2009 年 度 世 界 上 唯 一 的 突 出 优 秀 结 构 的 工 程 奖 。<br />

-9-


图 7 多 面 体 空 间 框 架 结 构 成 形 图<br />

(a) 效 果 图<br />

图 8 国 家 游 泳 中 心 “ 水 立 方 ”<br />

(b) 表 面 杆 件 与 节 点 连 接 图<br />

3.4. MR4 局 部 双 层 网 壳<br />

这 是 一 种 双 层 网 壳 与 单 层 网 壳 的 组 合 结 构 , 也 是 杆 单 元 和 梁 单 元 两 种 单 元 组 合 的 刚 性 空 间 结<br />

构 。 网 壳 结 构 中 以 薄 膜 力 为 主 受 力 部 位 可 采 用 单 层 网 壳 , 以 弯 曲 内 力 为 主 的 受 力 部 位 可 采 用 双 层 网<br />

壳 。 如 建 筑 上 需 要 设 置 天 窗 和 通 风 口 也 可 采 用 点 式 单 层 网 壳 布 置 的 双 层 网 壳 。 利 用 立 体 桁 架 来 分 区<br />

并 加 强 单 层 网 壳 可 构 成 分 块 明 显 的 局 部 双 层 网 壳 。<br />

1992 年 建 成 的 圆 形 平 面 40m 跨 度 烟 台 塔 山 游 乐 场 斗 兽 场 , 便 采 用 点 式 开 窗 的 局 部 双 层 网 壳 ( 图<br />

9)。 广 西 桂 平 体 育 馆 , 采 用 结 构 中 部 为 单 层 网 壳 的 局 部 双 层 结 构 ( 图 10)。 杭 州 奥 体 中 心 花 瓣 体 育<br />

场 一 方 案 , 花 瓣 部 分 采 用 双 层 网 壳 , 花 瓣 之 间 采 用 单 层 网 壳 ( 图 11)。<br />

图 9 烟 台 塔 山 游 乐 场 斗 兽 馆<br />

图 10 广 西 桂 平 体 育 馆<br />

-10-


(a) 效 果 图<br />

图 11 杭 州 奥 体 中 心 花 瓣 体 育 场 一 方 案<br />

(b) 计 算 模 型<br />

[10]<br />

3.5. MR5 组 合 网 架 结 构<br />

在 一 般 钢 网 架 结 构 中 撤 去 上 弦 杆 , 搁 置 带 肋 的 钢 筋 混 凝 土 预 制 小 板 , 待 灌 缝 形 成 整 体 后 便 构 成<br />

上 部 钢 筋 混 凝 土 结 构 , 下 部 为 钢 结 构 的 组 合 网 架 结 构 , 是 杆 元 、 梁 元 、 板 壳 元 三 种 单 元 组 合 的 刚 性<br />

空 间 结 构 , 也 是 两 种 材 料 的 组 合 结 构 。 组 合 网 架 结 构 的 结 构 承 载 和 围 护 功 能 合 二 为 一 , 它 适 合 做 屋<br />

盖 结 构 , 也 适 合 做 楼 层 结 构 。<br />

我 国 在 1980 年 与 国 际 同 步 率 先 自 行 设 计 建 筑 平 面 为 21m×54m 的 徐 州 夹 河 煤 矿 食 堂 组 合 网 架<br />

( 图 12), 至 今 已 建 成 近 六 十 幢 组 合 组 合 网 架 结 构 , 也 是 世 界 上 组 合 网 架 用 得 最 多 的 国 家 。 如 有 跨<br />

度 最 大 、 平 面 尺 寸 为 45.5m×58m 的 江 西 抚 州 体 育 馆 。1988 年 建 成 的 新 乡 百 货 大 楼 加 层 扩 建 工 程 ,<br />

平 面 尺 寸 为 35m×35m, 是 我 国 首 次 在 多 层 大 跨 建 筑 中 采 用 组 合 网 架 楼 层 及 屋 盖 结 构 ( 共 4 层 ), 见<br />

图 13。 长 沙 纺 织 大 厦 , 平 面 尺 寸 24m×27m, 采 用 柱 网 为 10m×12m、7m×12m 的 高 层 (11 层 ) 建<br />

筑 组 合 网 架 楼 层 及 屋 盖 结 构 。 上 海 国 际 购 物 中 心 的 六 、 七 层 楼 层 采 用 平 面 尺 寸 为 27m×27m 预 应 力<br />

组 合 网 架 。<br />

图 12 徐 州 夹 河 煤 矿 食 堂<br />

图 13 新 乡 百 货 大 楼<br />

四 、 现 代 柔 性 空 间 结 构 的 应 用 与 发 展<br />

现 代 柔 性 空 间 结 构 是 指 气 承 式 充 气 膜 结 构 与 柔 性 支 承 膜 结 构 两 种 , 现 分 别 说 明 如 下 。<br />

4.1. MF1 气 承 式 充 气 膜 结 构<br />

充 气 膜 结 构 分 为 两 类 : 气 囊 式 与 气 承 式 ( 膜 面 高 斯 曲 率 半 径 K>0)。 这 里 是 指 后 者 , 充 气 压 力<br />

不 大 , 仅 是 大 气 压 的 1.003 倍 , 人 们 可 在 气 承 式 充 气 膜 结 构 内 活 动 。 膜 材 ( 织 物 基 层 + 涂 层 ) 主 要 有<br />

PVC( 聚 氟 乙 烯 )、PTFE( 聚 四 氟 乙 烯 ), 用 于 充 气 枕 的 膜 材 是 不 含 织 物 层 的 , 类 似 于 无 纺 布 , 如<br />

ETFE( 乙 烯 — 四 氟 乙 烯 )。 对 膜 材 有 四 大 基 本 要 求 , 即 强 度 、 半 透 明 性 、 自 洁 性 和 耐 燃 性 。<br />

美 国 、 加 拿 大 和 日 本 在 二 十 世 纪 七 、 八 十 年 代 共 建 成 十 多 幢 超 百 米 跨 度 大 型 气 承 式 充 气 膜 结 构<br />

体 育 馆 , 如 前 面 提 到 的 美 国 庞 蒂 亚 克 体 育 馆 和 1988 年 在 日 本 建 成 的 180m×180m 方 椭 圆 形 平 面 东<br />

京 后 乐 园 棒 球 馆 。 我 国 上 世 纪 七 十 年 代 在 上 海 展 览 馆 北 侧 建 成 一 幢 平 面 为 28m×36m 气 承 式 充 气 膜<br />

结 构 临 时 展 厅 。2010 年 在 内 蒙 古 响 沙 湾 建 成 椭 圆 形 平 面 95m×105m 气 承 式 充 气 膜 结 构 沙 雕 展 览 馆<br />

( 图 14)。<br />

-11-


图 14 内 蒙 古 响 沙 湾 沙 雕 展 览 馆<br />

4.2. MF2 柔 性 支 承 膜 结 构<br />

支 承 式 膜 结 构 分 为 两 类 , 刚 性 支 承 式 和 柔 性 支 承 式 ( 膜 面 高 斯 曲 率 半 径 K≥0), 后 者 的 支 承 杆<br />

件 往 往 都 是 索 。 柔 性 支 承 膜 结 构 也 称 张 拉 膜 结 构 , 是 索 单 元 和 膜 单 元 两 种 单 元 组 合 而 成 的 柔 性 空 间<br />

结 构 。 这 类 膜 结 构 与 支 承 索 系 的 共 同 工 作 非 常 明 显 , 设 计 计 算 时 必 须 考 虑 索 膜 的 协 同 工 作 。 实 际 工<br />

程 中 柔 性 支 承 膜 结 构 与 梁 ( 拱 ) 支 承 的 刚 性 支 承 膜 结 构 也 常 混 合 选 用 。<br />

威 海 体 育 中 心 体 育 场 轮 廓 尺 寸 209m×236m, 内 环 尺 寸 143m×205m, 由 32 个 锥 状 悬 挑 柔 性 支<br />

承 膜 结 构 单 体 组 成 ( 图 15), 类 似 的 有 深 圳 欢 乐 谷 音 乐 厅 。2010 年 上 海 世 博 会 兴 建 的 世 博 轴 [11] , 由<br />

6 个 独 立 的 “ 阳 光 谷 ” 钢 结 构 和 多 跨 连 续 的 柔 性 支 承 膜 结 构 组 成 。 其 中 膜 结 构 总 长 度 840m, 横 向 最<br />

大 跨 度 97m, 总 面 积 64000m 2 , 为 当 前 世 界 上 最 大 的 张 拉 索 膜 结 构 。 膜 结 构 的 支 承 体 系 包 括 有 脊 索 、<br />

谷 索 、 边 索 、 吊 索 、 水 平 索 、 加 劲 索 、 背 索 等 。 同 时 , 还 要 有 19 根 中 桅 杆 、31 根 外 桅 杆 和 18 个 在<br />

“ 阳 光 谷 ” 上 的 支 承 点 ( 图 16)。<br />

图 15 威 海 体 育 中 心 体 育 场<br />

(a) 结 构 全 景 图 (b) 索 膜 结 构 计 算 模 型 (54000 个 单 元 )<br />

图 16 上 海 世 博 会 世 博 轴<br />

五 、 现 代 刚 柔 性 组 合 空 间 结 构 的 应 用 与 发 展<br />

-12-


现 代 刚 柔 性 组 合 空 间 结 构 是 现 代 化 空 间 结 构 应 用 与 发 展 的 主 流 , 品 种 繁 多 , 主 要 是 指 张 弦 梁 结<br />

构 、 弦 支 网 壳 、 索 穹 顶 — 网 壳 、 张 弦 气 肋 梁 、 预 应 力 网 架 ( 壳 )、 斜 拉 网 架 ( 壳 )、 张 弦 ( 立 体 ) 桁<br />

架 、 预 应 力 装 配 弓 式 结 构 、 索 桁 结 构 和 索 穹 顶 结 构 共 十 种 , 现 分 别 说 明 如 下 。<br />

5.1. MRF1 张 弦 梁 结 构<br />

这 是 由 下 弦 索 、 上 弦 梁 、 竖 杆 三 种 单 元 组 成 的 刚 柔 性 组 合 空 间 结 构 , 二 十 世 纪 九 十 年 代 由 日 本<br />

引 进 , 当 时 称 为 BSS 体 系 (BSS=Beam string structure)。 张 弦 梁 通 过 张 拉 下 弦 索 成 形 , 并 可 构 成 自<br />

平 衡 体 系 。 由 于 本 身 为 平 面 结 构 , 要 采 用 屋 面 支 承 体 系 保 证 平 面 外 的 稳 定 性 。 设 计 与 施 工 宜 采 用 多<br />

阶 级 张 弦 及 受 荷 方 式 , 以 利 用 材 料 的 反 复 受 力 性 能 , 并 应 考 虑 结 构 几 何 非 线 性 的 影 响 。 如 在 双 向 都<br />

采 用 张 弦 梁 结 构 , 便 由 平 面 结 构 真 正 转 变 为 空 间 结 构 。<br />

我 国 采 用 大 跨 度 的 张 弦 梁 结 构 要 首 推 1999 年 建 成 的 上 海 浦 东 国 际 机 场 航 站 楼 , 其 中 最 大 跨 度<br />

是 82.6m 的 售 票 大 厅 , 纵 向 间 距 为 9m( 图 17)。2006 年 建 成 的 浦 东 机 场 T2 航 站 楼 采 用 平 面 尺 寸 为<br />

(48+89+46)m×414m 的 三 跨 连 续 张 弦 梁 结 构 ( 图 18), 间 距 为 9m, 支 承 柱 选 用 空 间 双 层 Y 形 立<br />

柱 , 也 是 最 简 单 树 状 结 构 , 柱 间 距 18m, 但 效 果 显 著 。 浙 江 大 学 紫 金 港 校 区 图 书 馆 大 厅 , 采 用 了 双<br />

向 正 交 的 张 弦 梁 结 构 。 天 津 火 车 站 改 建 的 无 站 台 柱 雨 篷 采 用 五 跨 (48.5m×2+41m+42m+39.5m) 张<br />

弦 双 肢 梁 结 构 。<br />

图 17 上 海 浦 东 机 场 航 站 楼<br />

图 18 浦 东 机 场 T2 航 站 楼<br />

5.2. MRF2 弦 支 网 壳<br />

这 是 由 弦 支 体 系 的 斜 索 与 环 索 、 竖 杆 及 单 层 网 壳 的 空 间 梁 三 种 单 元 组 成 刚 柔 性 组 合 空 间 结 构 ,<br />

具 有 单 层 网 壳 和 索 弯 顶 两 种 结 构 体 系 的 优 点 。 通 过 施 加 预 应 力 可 提 高 结 构 刚 度 , 是 一 种 自 平 衡 空 间<br />

结 构 , 可 不 产 生 或 减 少 结 构 的 水 平 推 力 。 建 筑 平 面 以 圆 形 居 多 , 也 可 适 用 于 椭 圆 形 、 多 边 形 和 矩 形<br />

平 面 。<br />

日 本 最 早 在 1993 年 建 成 35m 跨 的 弦 支 光 球 穹 顶 。 我 国 早 期 的 弦 支 网 壳 要 首 推 用 于 2001 年 建 成<br />

的 35.4m 跨 度 的 天 津 港 保 税 区 商 务 中 心 大 堂 。 此 后 在 2007 年 建 成 椭 圆 平 面 80m×120m 常 州 体 育 馆 ,<br />

在 2008 年 建 成 圆 形 平 面 跨 度 93m 的 北 工 大 体 育 馆 。2009 年 建 成 圆 形 平 面 跨 度 122m 全 运 会 济 南 体<br />

育 馆 ( 图 19), 也 是 当 前 世 上 跨 度 最 大 的 弦 支 网 壳 [13] 。2010 建 成 矩 形 平 面 28m×43m 柱 网 双 向 多 跨<br />

连 续 的 深 圳 北 站 无 站 台 柱 雨 篷 ( 图 20), 覆 盖 建 筑 面 积 为 6.8 万 m 2 , 是 国 内 外 首 创 矩 形 平 面 弦 支 圆<br />

柱 面 网 壳 结 构 。<br />

图 19 全 运 会 济 南 体 育 馆 施 工<br />

图 20 深 圳 北 站 无 站 台 柱 雨 篷<br />

-13-


5.3. MRF3 索 穹 顶 — 网 壳<br />

这 是 我 国 自 己 提 出 的 一 种 新 型 空 间 结 构 体 系 [14] , 由 索 穹 顶 的 索 杆 体 系 与 单 层 网 壳 组 合 而 成 ( 图<br />

21) 也 是 杆 、 索 、 梁 三 种 单 元 构 成 的 刚 柔 性 组 合 空 间 结 构 。 施 工 时 无 需 满 堂 红 脚 手 架 可 在 自 平 衡 的<br />

索 杆 体 系 上 安 装 单 层 网 壳 , 并 可 做 成 刚 性 屋 面 , 一 改 通 常 索 穹 顶 的 膜 面 屋 顶 , 以 扩 大 应 用 范 围 。<br />

图 21 索 穹 顶 — 网 壳 构 成 图<br />

浙 江 大 学 曾 做 过 直 径 5m 的 索 穹 顶 — 网 壳 模 型 试 验 [15] ( 图 22), 在 文 献 [16] 也 对 这 种 新 结 构 进 行<br />

了 研 究 ,2009 年 全 运 会 济 南 体 育 馆 评 审 时 , 索 穹 顶 — 网 壳 曾 是 一 个 遴 选 的 方 案 ( 图 23), 后 因 种 种<br />

原 因 未 被 采 用 , 而 选 用 了 国 内 已 有 成 熟 设 计 与 施 工 经 验 的 弦 支 网 壳 结 构 。<br />

图 22 索 穹 顶 — 网 壳 模 型 试 验<br />

图 23 全 运 会 济 南 体 育 馆 索 穹 顶 — 网 壳 方 案<br />

5.4. MRF4 张 弦 气 肋 梁<br />

张 弦 气 肋 梁 (Tensairity = Tension + air + integrity) 是 在 张 弦 梁 结 构 中 用 气 肋 来 代 替 竖 杆 而 构 成<br />

的 空 间 结 构 体 系 , 也 是 梁 、 索 、 膜 三 种 单 元 组 成 刚 柔 性 组 合 空 间 结 构 。2004 年 在 法 国 召 开 的 IASS<br />

学 术 会 议 上 首 次 对 张 弦 气 肋 梁 有 所 报 道 [17] ,2007 年 已 建 成 张 弦 气 肋 梁 结 构 试 点 工 程 , 应 用 于 瑞 士 蒙<br />

特 立 克 斯 车 站 汽 车 库 ( 图 24) [18] , 法 国 的 一 座 桥 梁 也 采 用 了 这 种 张 弦 气 肋 梁 结 构 。 我 国 目 前 尚 无 这<br />

类 工 程 实 例 , 有 关 高 校 和 科 研 单 位 , 对 其 结 构 理 论 、 分 析 方 法 、 构 造 措 施 和 应 用 前 景 在 进 一 步 研 究<br />

和 探 讨 。<br />

图 24 瑞 士 蒙 特 立 克 斯 车 站 汽 车 库 张 弦 气 肋 梁 及 其 与 气 承 结 构 的 连 接 构 造<br />

-14-


5.5. MRF5 预 应 力 网 架 ( 壳 )<br />

把 预 应 力 技 术 应 用 到 网 架 ( 壳 ) 中 , 构 成 杆 、 索 两 种 单 元 组 成 的 刚 柔 性 组 合 空 间 结 构 。 通 常 在<br />

网 架 下 弦 的 下 方 、 双 层 网 壳 的 周 边 设 置 裸 露 的 预 应 力 索 , 以 改 善 结 构 的 内 力 分 布 、 降 低 内 力 峰 值 、<br />

提 高 结 构 刚 度 、 节 省 材 料 耗 量 。 根 据 国 内 经 验 , 采 用 预 应 力 网 架 ( 壳 ) 比 非 预 应 力 网 架 ( 壳 ) 可 节<br />

省 钢 材 用 量 的 25%。<br />

1994 年 建 成 的 六 边 形 平 面 对 角 线 长 93.6m 广 东 清 远 体 育 馆 , 采 用 六 块 组 合 型 双 层 扭 网 壳 , 在 相<br />

邻 六 支 座 处 采 用 了 六 道 预 应 力 索 ( 图 25)。1995 年 建 成 的 间 等 边 八 边 形 74.8m×74.8m 攀 枝 花 体 育<br />

馆 , 采 用 双 层 球 面 网 壳 , 在 相 邻 八 支 座 处 设 置 八 榀 平 面 桁 架 , 其 下 弦 选 用 了 预 应 力 索 ( 图 26)。 上<br />

海 国 际 购 物 中 心 的 六 、 七 层 楼 层 为 缺 角 27m×27m 的 组 合 刚 架 , 在 网 架 下 弦 节 点 的 下 方 20cm 处 ,<br />

设 置 四 道 45° 方 向 预 应 力 索 , 围 成 一 斜 放 矩 形 环 向 索 加 强 。 广 东 高 安 露 天 球 场 新 加 屋 盖 , 采 用 平 面<br />

为 54.9m×69.3m 四 块 组 合 型 扭 网 壳 , 通 过 周 边 中 点 支 承 节 点 处 设 置 四 道 预 应 力 索 , 围 成 一 平 行 四<br />

边 形 环 向 索 加 强 。<br />

图 25 清 远 体 育 馆 模 型<br />

图 26 攀 枝 花 体 育 馆 模 型<br />

5.6. MRF6 斜 拉 网 架 ( 壳 )<br />

把 斜 拉 桥 的 斜 拉 体 系 应 用 到 网 架 ( 壳 ) 中 , 构 成 杆 、 索 两 种 单 元 组 成 的 刚 柔 性 组 合 空 间 结 构 。<br />

在 网 架 、 双 层 网 壳 的 顶 上 , 设 置 多 道 斜 拉 索 , 相 当 于 增 加 了 节 点 、 减 小 结 构 跨 度 、 提 高 刚 度 , 而 且<br />

斜 拉 索 尚 可 施 加 预 应 力 , 改 善 结 构 的 内 力 分 布 , 节 省 材 料 耗 量 。 斜 拉 索 宜 全 方 位 设 置 , 设 计 时 应 考<br />

虑 在 任 何 载 荷 作 用 下 不 出 现 松 弛 现 象 。<br />

代 表 性 的 工 程 有 :1993 年 建 成 的 新 加 坡 港 务 局 仓 库 采 用 4 幢 120m×96m 六 塔 柱 ,2 幢 96m×<br />

70m 四 塔 柱 斜 拉 网 架 ( 图 27),1995 年 建 成 的 山 西 太 旧 高 速 公 路 旧 关 收 费 站 采 用 14m×65m 独 塔 式<br />

斜 拉 双 层 柱 面 壳 ( 图 28)。2000 年 建 成 的 杭 州 黄 龙 体 育 中 心 体 育 馆 采 用 月 牙 形 平 面 50m×244m 双<br />

塔 柱 斜 拉 双 层 网 壳 , 屋 面 上 还 采 用 18 道 稳 定 索 加 强 ( 图 29)。2010 年 建 成 浙 江 大 学 紫 金 港 校 区 文<br />

体 中 心 采 用 斜 拉 索 网 悬 吊 单 层 网 壳 结 构 ( 图 30)。<br />

图 27 新 加 坡 港 务 局 仓 库<br />

图 28 山 西 太 旧 高 速 公 路 旧 关 收 费 站<br />

-15-


图 29 杭 州 黄 龙 体 育 中 心 体 育 场<br />

图 30 浙 江 大 学 紫 金 港 校 区 文 体 中 心<br />

5.7. MRF7 张 弦 ( 立 体 ) 桁 架<br />

以 ( 立 体 ) 桁 架 替 代 张 弦 梁 的 上 弦 梁 便 构 成 了 张 弦 ( 立 体 ) 桁 架 。 这 也 是 一 种 杆 、 索 单 元 组 成<br />

的 刚 柔 组 合 空 间 结 构 。 桁 架 可 设 计 成 平 面 桁 架 或 立 体 桁 架 , 但 都 要 设 置 屋 面 支 撑 体 系 以 保 证 张 弦 ( 立<br />

体 ) 桁 架 的 平 面 外 稳 定 性 。<br />

代 表 性 的 工 程 有 :2002 年 建 成 的 广 州 国 际 会 展 中 心 采 用 跨 度 为 126.6m 张 弦 立 体 桁 架 ( 图 31)。<br />

2007 年 建 成 北 京 北 站 雨 篷 , 采 用 最 大 跨 度 107m, 间 距 20m 的 张 弦 立 体 桁 架 , 覆 盖 建 筑 面 积 近 7 万<br />

m 2 ,2008 年 建 成 的 奥 运 会 国 家 体 育 馆 采 用 114m×144m 双 向 正 交 的 张 弦 桁 架 ( 平 面 桁 架 ) 结 构 ( 图<br />

32), 用 钢 指 标 90kg/m 2 , 这 也 是 当 前 世 界 上 跨 度 最 大 的 双 向 张 弦 桁 架 结 构 。<br />

图 31 广 州 国 际 会 展 中 心<br />

图 32 奥 运 会 国 家 体 育 馆<br />

5.8. MRF8 预 应 力 装 配 弓 式 结 构<br />

预 应 力 装 配 弓 式 结 构 是 我 国 科 技 工 作 者 提 出 一 项 专 利 技 术 , 它 是 通 过 逐 段 伸 展 预 应 力 装 配 法 来<br />

建 造 大 跨 度 弓 式 结 构 , 其 施 工 组 装 示 意 图 见 图 33。 图 中 串 式 拉 筋 一 方 面 是 连 接 装 配 的 手 段 , 另 一 方<br />

面 又 起 到 施 加 预 应 力 作 用 , 一 举 双 得 。 此 外 , 这 种 弓 式 结 构 施 工 安 装 时 无 需 大 型 的 极 具 设 备 , 只 要<br />

利 用 小 型 千 斤 顶 便 能 拼 装 大 跨 度 空 间 结 构 。 预 应 力 装 配 弓 式 结 构 也 是 仅 由 杆 、 索 二 种 单 元 组 成 的 刚<br />

柔 性 组 合 空 间 结 构 。<br />

早 年 这 种 预 应 力 装 配 弓 式 结 构 曾 用 于 小 型 机 库 。1994 年 建 成 了 45m 跨 度 北 京 钓 鱼 台 国 宾 馆 室<br />

内 网 球 场 弓 形 屋 盖 , 其 中 段 在 纵 向 可 开 启 ( 图 34)。2005 年 建 成 了 北 京 温 都 水 城 72m 跨 度 戏 水 乐 园 ,<br />

局 部 屋 盖 也 可 开 启 。 这 种 预 应 力 装 配 弓 式 结 构 特 别 适 宜 于 用 在 可 装 拆 的 临 时 性 仓 库 建 筑 和 舞 台 建<br />

筑 , 曾 建 成 跨 度 达 130m 构 筑 物 , 施 工 安 装 、 拆 除 都 很 方 便 。<br />

-16-


图 33 预 应 力 装 配 弓 式 结 构<br />

施 工 组 装 示 意 图<br />

图 34 钓 鱼 台 国 宾 馆 室 内 网 球 场<br />

5.9. MRF9 索 桁 结 构<br />

将 平 面 双 层 索 ( 向 下 凹 的 称 为 承 重 索 , 向 上 凸 的 称 为 稳 定 索 ) 布 置 成 图 35 所 示 的 各 种 形 式 ,<br />

索 间 设 置 受 压 撑 杆 ( 受 拉 时 一 般 设 置 成 拉 索 ), 当 承 重 索 ( 或 稳 定 索 ) 施 加 预 应 力 后 , 便 可 构 成 自<br />

平 衡 的 有 预 应 力 的 索 桁 结 构 。 为 改 善 索 桁 结 构 平 面 外 的 刚 度 , 承 重 索 和 稳 定 索 可 错 位 设 置 ( 其 水 平<br />

投 影 线 错 开 半 格 间 距 )。 索 桁 结 构 除 可 用 于 矩 形 平 面 建 筑 外 , 也 可 适 用 于 圆 形 平 面 和 环 形 平 面 的 大<br />

跨 度 体 育 场 馆 等 公 益 建 筑 。<br />

近 年 来 , 我 国 的 索 桁 结 构 已 有 较 多 的 应 用 , 其 代 表 性 的 工 程 有 :1986 年 建 成 59m 跨 度 的 吉 林<br />

冰 球 馆 , 采 用 了 上 、 下 索 错 位 布 置 的 索 桁 结 构 ( 图 36) [19] 。2006 年 建 成 的 佛 山 世 纪 莲 体 育 场 , 采<br />

用 了 环 形 平 面 折 板 形 索 桁 结 构 , 外 圆 直 径 310m, 内 孔 直 径 125m, 周 边 支 承 在 环 形 桁 架 上 ( 图 37)。<br />

2010 年 建 成 的 大 运 会 宝 安 体 育 场 , 采 用 230m×237m 椭 圆 环 平 面 的 索 桁 结 构 , 最 大 悬 挑 54m, 外 周<br />

边 设 箱 型 环 梁 , 内 孔 边 设 管 型 飞 柱 , 屋 面 为 支 承 在 小 拱 上 的 膜 结 构 ( 图 38)。2011 年 进 行 模 型 试 验<br />

和 施 工 的 浙 江 乐 清 体 育 中 心 体 育 场 , 采 用 229m×221m 月 牙 形 非 封 闭 环 形 平 面 索 桁 结 构 ( 图 39)。<br />

图 35 索 桁 结 构 的 形 式<br />

图 36 吉 林 冰 球 馆 结 构 模 型<br />

图 37 佛 山 世 纪 莲 体 育 场<br />

图 38 大 运 会 宝 安 体 育 场<br />

-17-


(a) 鸟 瞰 图<br />

(b) 计 算 模 型 图<br />

图 39 乐 清 体 育 中 心 体 育 场<br />

5.10. MRF10 索 穹 顶<br />

索 穹 顶 是 由 索 单 元 为 主 、 杆 单 元 和 膜 单 元 三 种 单 元 组 成 的 偏 柔 性 的 刚 柔 组 合 空 间 结 构 。 这 也 是<br />

一 种 周 边 支 承 在 受 压 环 梁 上 的 张 拉 整 体 结 构 , 完 全 体 现 了 富 勒 关 于 “ 压 力 的 孤 岛 存 在 于 拉 杆 的 海 洋 ”<br />

的 思 想 , 是 当 前 空 间 结 构 发 展 的 一 大 高 峰 , 具 有 极 高 的 结 构 效 率 。1986 年 由 美 国 工 程 师 盖 格 尔 首 创<br />

建 成 的 汉 城 奥 运 会 120m 跨 度 综 合 馆 索 穹 顶 ( 图 40) 是 世 界 首 例 肋 环 形 索 穹 顶 , 其 平 、 剖 面 的 基 本<br />

构 件 为 脊 索 、 谷 索 、 斜 拉 索 、 下 弦 环 索 、 立 柱 、 外 环 梁 、 中 心 环 、 膜 片 等 ( 图 41)。<br />

我 国 已 研 究 并 提 出 了 Kiewitt 型 、 混 合 Ⅰ 型 ( 肋 环 型 和 葵 花 型 重 叠 式 组 合 )、 混 合 Ⅱ 型 (Kiewitt<br />

型 和 葵 花 型 内 外 式 组 合 )、 鸟 巢 型 等 多 种 形 式 的 索 穹 顶 。 [21] 同 时 对 肋 环 型 、 葵 花 型 索 穹 顶 提 出 初 始<br />

预 应 力 模 态 的 快 速 计 算 法 , 对 一 般 索 穹 顶 提 出 求 解 整 体 自 应 力 模 态 的 二 次 奇 异 值 法 , 为 索 穹 顶 预 应<br />

力 设 计 提 供 了 创 新 的 分 析 方 法 。 [22~24] 索 穹 顶 在 我 国 的 工 程 应 用 刚 刚 开 始 ,2009 年 在 浙 江 金 华 晟 元 集<br />

团 标 准 厂 房 中 庭 采 用 平 面 尺 寸 19m×17m 的 索 穹 顶 ( 图 42), 这 是 我 国 首 例 索 穹 顶 。2010 年 建 成 了<br />

圆 形 平 面 跨 度 为 72m 鄂 尔 多 斯 市 伊 金 霍 洛 旗 体 育 馆 索 穹 顶 ( 图 43)。<br />

图 40 汉 城 奥 运 会 综 合 馆<br />

图 41 肋 环 型 索 穹 顶 平 、 剖 面 示 意 图<br />

图 42 浙 江 金 华 华 晟 元 集 团 标 准 厂 房 中 庭<br />

图 43 鄂 尔 多 斯 市 伊 金 霍 洛 旗 体 育 馆<br />

-18-


六 、 结 语<br />

综 上 所 述 , 可 得 到 如 下 一 些 结 论 :<br />

(1) 我 国 现 代 空 间 结 构 与 世 界 同 步 , 是 在 二 十 世 纪 七 、 八 十 年 代 左 右 发 展 起 来 的 。 发 展 历 史<br />

并 不 太 长 , 大 致 为 三 十 多 年 , 但 已 在 体 育 场 馆 、 会 展 中 心 、 影 剧 院 、 候 机 ( 车 ) 大 厅 、 铁 路 站 台 雨<br />

篷 、 工 业 厂 房 与 仓 库 中 获 得 大 量 应 用 。<br />

(2) 现 代 大 跨 空 间 结 构 可 认 为 是 由 二 部 分 组 成 , 一 是 由 采 用 轻 质 、 高 强 的 膜 材 和 钢 索 而 发 展<br />

起 来 的 诸 如 气 承 式 充 气 膜 结 构 、 索 膜 结 构 、 索 桁 结 构 、 张 弦 梁 结 构 、 弦 支 网 壳 结 构 、 索 穹 顶 结 构 等 ,<br />

二 是 由 薄 壳 结 构 、 网 格 结 构 、 一 般 悬 索 结 构 等 近 代 空 间 结 构 , 采 用 多 种 结 构 和 材 料 的 组 合 、 预 应 力<br />

技 术 与 形 体 的 创 新 而 发 展 起 来 的 诸 如 组 合 网 架 结 构 、 斜 拉 网 格 结 构 、 预 应 力 网 格 结 构 、 多 面 体 空 间<br />

框 架 结 构 等 。<br />

(3) 按 组 成 空 间 结 构 的 基 本 构 件 即 单 元 ( 包 括 板 壳 单 元 、 梁 单 元 、 杆 单 元 三 种 刚 性 单 元 及 索<br />

单 元 、 膜 单 元 两 种 柔 性 单 元 ) 分 类 , 现 代 空 间 结 构 又 可 分 为 现 代 刚 性 空 间 结 构 、 现 代 柔 性 空 间 结 构<br />

和 现 代 刚 柔 性 组 合 空 间 结 构 三 大 类 , 本 文 对 这 三 类 现 代 空 间 结 构 在 我 国 的 应 用 与 发 展 分 别 作 了 详 细<br />

的 说 明 和 简 述 。<br />

(4) 本 文 所 涉 及 的 现 代 大 跨 空 间 结 构 是 认 为 有 推 广 价 值 和 应 用 前 景 的 空 间 结 构 , 今 后 尚 应 总<br />

结 提 高 , 其 设 计 水 平 和 施 工 技 术 也 应 进 一 步 优 化 和 深 化 。17 种 现 代 大 跨 空 间 结 构 也 不 是 发 展 的 极 限<br />

数 , 今 后 也 必 然 会 按 组 成 空 间 结 构 的 分 类 方 法 发 展 为 更 多 数 量 的 、 形 体 各 异 的 现 代 大 跨 空 间 结 构 。<br />

(5) 由 本 文 所 述 可 知 , 从 大 跨 空 间 结 构 世 界 之 “ 最 ” 来 统 计 分 析 ( 包 括 新 颖 结 构 最 早 采 用 、<br />

结 构 跨 度 最 大 、 结 构 幢 数 最 多 、 建 筑 结 构 覆 盖 面 积 最 大 等 等 ), 中 国 不 愧 为 是 现 代 空 间 结 构 的 大 国 ,<br />

下 一 步 应 着 重 关 注 从 大 国 向 现 代 大 跨 空 间 结 构 强 国 发 展 。<br />

参 考 文 献<br />

[1] 董 石 麟 . 空 间 结 构 的 发 展 历 史 、 创 新 、 形 式 分 类 与 实 际 应 用 . 空 间 结 构 , 2009, 15(3): 22-43.<br />

[2] 董 石 麟 . 中 国 空 间 结 构 的 发 展 与 展 望 . 建 筑 结 构 学 报 , 2010, 31(6): 38-51.<br />

[3] 董 石 麟 , 罗 尧 治 , 赵 阳 . 新 型 空 间 结 构 分 析 、 设 计 与 施 工 , 北 京 : 人 民 交 通 出 版 社 , 2009.<br />

[4] 刘 锡 良 编 著 . 现 代 空 间 结 构 , 天 津 : 天 津 大 学 出 版 社 , 2003.<br />

[5] 斋 藤 公 男 著 , 季 小 莲 , 徐 华 泽 译 . 空 间 结 构 的 发 展 与 展 望 — 空 间 结 构 设 计 的 过 去 • 现 在 • 未 来 , 北 京 : 中 国 建 筑 工<br />

业 出 版 社 , 2006.<br />

[6] 梅 秀 魁 , 刘 德 明 , 姚 亚 雄 . 大 跨 建 筑 结 构 构 思 与 构 造 选 型 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 2002.<br />

[7] 董 石 麟 , 赵 阳 . 论 空 间 结 构 的 形 式 和 分 类 . 土 木 工 程 学 报 , 2004, 37(1): 7-12.<br />

[8] 董 石 麟 , 赵 阳 . 论 索 单 元 构 成 的 柔 性 空 间 结 构 与 刚 柔 组 合 空 间 结 构 . 第 十 三 届 空 间 结 构 学 术 会 议 论 文 集 , 深 圳 ,<br />

2010: 1-8.<br />

[9] 傅 学 怡 , 顾 磊 , 赵 阳 , 等 . 国 家 游 泳 中 心 水 立 方 结 构 设 计 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 2009.<br />

[10] 董 石 麟 , 马 克 俭 , 严 慧 , 等 . 组 合 网 架 结 构 与 空 腹 网 架 结 构 , 杭 州 : 浙 江 大 学 出 版 社 , 1992.<br />

[11] 汪 大 绥 , 张 伟 育 , 方 卫 , 等 . 世 博 轴 大 跨 度 索 膜 结 构 设 计 与 研 究 . 建 筑 结 构 学 报 , 2010, 31(5): 1-12.<br />

[12] 陈 志 华 . 弦 支 穹 顶 结 构 , 北 京 : 科 学 出 版 社 , 2010.<br />

[13] 董 石 麟 , 袁 行 飞 , 郭 佳 民 , 等 . 济 南 奥 体 中 心 体 育 馆 弦 支 穹 顶 结 构 分 析 与 试 验 研 究 . 工 业 建 筑 , 2009, 15( 增 刊 ):<br />

11-16.<br />

[14] 董 石 麟 , 王 振 华 , 袁 行 飞 . 一 种 由 索 穹 顶 与 单 层 网 壳 组 合 的 空 间 结 构 及 其 受 力 性 能 研 究 . 建 筑 结 构 学 报 , 2010,<br />

31(3): 1-8.<br />

[15] 王 振 华 . 索 穹 顶 与 单 层 网 壳 组 合 的 新 型 空 间 结 构 的 理 论 分 析 与 试 验 研 究 , 杭 州 : 浙 江 大 学 , 2009.<br />

[16] 宗 钟 凌 , 郭 正 兴 . 刚 性 网 格 索 穹 顶 结 构 初 始 形 态 的 确 定 方 法 . 空 间 结 构 , 2011, 17(1): 3-7.<br />

[17] Luchsinger, R. H., Pedretti, A. and Steingruber, P. et al. Light weight structures with tensigrity. Shell and Spatial<br />

Structures form Models to Realization, Montepllier, 2004.<br />

[18] Pedretti, M and Lvchsinger, R. H. Tensairity patent-a pneumatic tensile roof, Stablebau, 2007, 76(5): 314-319.<br />

[19] 沈 世 钊 , 徐 崇 宝 , 赵 臣 . 悬 索 结 构 设 计 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 1997.<br />

[20] 郭 彦 林 , 田 广 宇 , 王 昆 , 等 . 宝 安 体 育 场 车 辐 式 屋 盖 结 构 整 体 模 型 施 工 张 拉 试 验 . 建 筑 结 构 学 报 , 2011, 32(3):<br />

1-10.<br />

[21] 董 石 麟 , 袁 行 飞 . 索 穹 顶 结 构 体 系 若 干 研 究 新 发 展 . 浙 江 大 学 学 报 ( 工 学 版 ), 2008, 25(4): 134-139.<br />

[22] 董 石 麟 , 袁 行 飞 . 肋 环 型 索 穹 顶 初 始 预 应 力 分 布 的 快 速 计 算 法 . 空 间 结 构 , 2003, 9(2): 3-8.<br />

[23] 董 石 麟 , 袁 行 飞 . 葵 花 型 索 穹 顶 初 始 预 应 力 分 布 的 简 捷 计 算 方 法 . 建 筑 结 构 学 报 , 2004, 25(6): 9-14.<br />

[24] 袁 行 飞 , 董 石 麟 . 索 穹 顶 结 构 的 新 形 式 及 其 初 始 预 应 力 确 定 . 工 程 力 学 , 2005, 22(2): 22-26.<br />

-19-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

GROUND VIBRATIONS DUE TO UNDERGROUND TRAINS BY THE 2.5D<br />

FINITE/INFINITE ELEMENT APPROACH<br />

Y. B. Yang 1 and H. H. Hung 2<br />

1<br />

Department of Constructional Engineering, National Yunlin <strong>University</strong> of Science and Technology<br />

Douliu, Yunlin, Taiwan 64002; On leave from National Taiwan <strong>University</strong>.<br />

2<br />

National Center for Research on Earthquake Engineering<br />

Taipei 10617, Taiwan.<br />

ABSTRACT<br />

<strong>The</strong> 2.5D finite/infinite element approach for simulating the ground vibrations by underground trains is<br />

summarized in this paper. By assuming the soils to be uniform along the direction of the railway, only a profile<br />

of the soil perpendicular to the railway need be considered in establishing the mesh. Besides the two in-plane<br />

degrees of freedom (DOFs) per node conventionally used for plane strain elements, an extra DOF is introduced<br />

to account for the out-of-plane wave transmission. <strong>The</strong> profile of the half-space is divided into a near field and a<br />

semi-infinite far field, with the former simulated by finite elements and the latter by infinite elements to account<br />

for the effect of radiation. Enhanced by the automated mesh expansion procedure proposed previously (Yang et<br />

al. 1996), the far field impedance for each of the lower frequencies is generated recursively from the mesh<br />

created for the adjacent higher frequency considered. Besides, a load generation mechanism is proposed that<br />

takes the rail irregularity and dynamic properties of trains into account. Finally, a study was performed to<br />

evaluate the isolation efficiency of the elastic foundation inserted under the rails. Compared with the<br />

conventional 3D approach, the present approach appears to be simple, efficient and generally accurate.<br />

KEYWORDS<br />

Half space, infinite element, soil vibration, train, tunnel, 2.5D approach.<br />

INTRODUCTION<br />

Railway trains or mass rapid transit systems have become a major public transportation tool for most large cities<br />

in the past one and half centuries. Besides, following the successful operation of the bullet trains in Japan in<br />

1964, high speed trains have emerged as an effective tool for inter-city transportation in Europe and Asia.<br />

Recently, with the aim to reduce traffic congestion and oil dependence, while preserving the environment, the<br />

United States, where freeway and air transportations have been popular, is also considering the development of a<br />

national network for high-speed passenger rail lines.<br />

Ground-borne vibration can be generated by the passage of trains due to the inherent frequencies of the train,<br />

surface irregularities of wheels and rails, rise and fall of the axles over sleepers, etc. <strong>The</strong> vibrations then can be<br />

transmitted to adjacent structures through the track, underground tunnel and surrounding soil. This problem has<br />

aroused the attention of researchers since 1970s. Because of the limitation in computer capacities in the 1970s,<br />

this problem was dealt with mainly by the analytical solution with unavoidable simplifications or by in-situ<br />

measurements in the earlier year. However, in recent years, partly enhanced by the advance in computers,<br />

researchers and engineers began to work on more sophisticated numerical methods for simulating the problem in<br />

a more realistic way.<br />

Various numerical methods have been developed, including the two-dimensional model (Metrikine and<br />

Vrouwenvelder 2000; Paolucci, et al. 2003) and three-dimensional model (Mohammadi and Karabalis 1995;<br />

Celebi 2006). Both of these models have been developed along the lines of analytical, semi-analytical, finite<br />

element method (FEM), boundary element method (BEM), infinite element method (IFEM) or their<br />

combinations. <strong>The</strong> 2D analysis is attractive due to its computational efficiency. However, how to reasonably<br />

transfer a 3D force into an equivalent 2D force is a crucial concern for the results to be reasonable. In addition,<br />

2D models may underestimate the soil damping, while ignoring the wave propagation characteristics in the<br />

train-moving direction. In contrast, although 3D models can duly take into account the wave propagation<br />

behavior in the load-moving direction, the execution of a 3D analysis is extremely time-consuming in practice,<br />

-20-


egardless of the fact that better simulation results can be obtained.<br />

In order to simulate the wave propagation in the direction of the tunnel, while saving the computation time, the<br />

2.5D (two point five dimensional) analysis concept was proposed by Yang and Hung (2001, 2009). Based on the<br />

this approach, the surface irregularity of wheels and rails, which plays an influential role in ground-borne<br />

vibration, will be taken into account. <strong>The</strong> surface irregularities will be treated as a stationary ergodic Gaussian<br />

random process and simulated by trigonometry series (Lei and Noda 2002; Sheng et al. 2004). Besides, a load<br />

generation mechanism for the train-track system is presented. Finally, a case study is presented to investigate the<br />

efficiency of a floating slab track in reducing underground train-induced vibrations.<br />

PROBLEM FORMULATION AND BASIC ASSUMPTIONS<br />

Consider a series of vehicles moving with speed c along the z-axis on the ground surface or through an<br />

underground tunnel, as shown in Figure 1. <strong>The</strong> loading effect of the moving vehicles can be represented by:<br />

f ( xyzt , , , ) = ψ ( xy , ) φ( z− ctRt ) ( )<br />

(1)<br />

where R(t) represents the interaction forces between the wheels and rails, which include the vehicle weight and<br />

the dynamic components of the loading induced by the moving vehicles. <strong>The</strong> load distribution function φ ( z)<br />

accounts for the effect of wheel intervals along the z-axis. <strong>The</strong> function ψ (, x y)<br />

represents the location of the<br />

moving load on the xy plane. By performing double Fourier transformation to Eq. (1), one can express the<br />

external load in frequency and wave number domain as<br />

~<br />

~ ~<br />

f ( x,<br />

y,<br />

kz , ω ) = ψ ( x,<br />

y)<br />

φ ( kz<br />

) R(<br />

kzc<br />

+ ω)<br />

(2)<br />

~<br />

in which φ ( k z ) and R% ( ω)<br />

are the Fourier transforms of φ ( z)<br />

and R(t) with respective to z and t, respectively.<br />

By the inverse Fourier transformation, the external load in time and space domain can be recovered as<br />

∞ ∞<br />

f ( xyzt , , , ) = ψ ( xy , ) % φ( k) Rkc %( + ω)exp( ikz)exp( iω tdkd ) ω<br />

(3)<br />

∫−∞∫−∞<br />

z z z z<br />

<strong>The</strong> preceding equation shows that the external load can be expressed as the integration of a series of harmonic<br />

functions. For a linear system, which is the case considered herein, the total steady-state response of the<br />

half-space can be obtained by superposing the responses generated by all the harmonic functions of the external<br />

load. Let (<br />

z<br />

, )<br />

the total response of the half-space in time and space domain can be written as<br />

Hk ω denote the complex response function for each harmonic term ψ ( x, y)exp( ik z)exp( iω t)<br />

,<br />

∞ ∞<br />

∫−∞∫−∞<br />

d( x′ , y′ , z, t) = % φ ( k ) R% ( k c+<br />

ω) H( k , ω)exp( ik z)exp( iω t)<br />

dk dω<br />

(4)<br />

z z z z z<br />

In the following, we shall explain how to obtain the response function H( kz<br />

, ω ) by the 2.5D finite/infinite<br />

element approach in frequency and wave number domain, and the interaction force % φ ( kz) R % ( kzc+<br />

ω)<br />

by taking<br />

the distribution of wheels, configuration of trains and rail irregularity into account.<br />

z<br />

Figure 1 Typical structure of analysis<br />

DERIVATION OF 2.5D FINITE/INFINITE ELEMENTS<br />

Assume the material and geometry of the system in Figure 1 to be identical along the z-axis. In response to the<br />

external load ψ ( x, y)exp( ikz<br />

z)exp( iω t)<br />

, the displacements u ( x, yzt , , ) of the half-space can be expressed as<br />

u( x, yzt , , ) = u ) ( xy , )exp( ikz z )exp( iωt)<br />

(5)<br />

where u ˆ( x, y)<br />

denotes the 2D displacement field independent of the load-moving direction z. Consequently,<br />

the originally 3D continuous solid is represented by elements on the x-y plane via the function u ˆ( x, y)<br />

. Further,<br />

the half-space is divided into a near field and a semi-infinite far field, with the former simulated by finite<br />

elements, and latter by infinite elements. <strong>The</strong> displacement within each element can be interpolated in the same<br />

way as that for the plane elements. This is certainly an advantage of the 2.5D approach, since it enables us to<br />

compute the 3D response taking into account the load-moving effect using merely the 2D profile.<br />

-21-


As it is conventional, the displacements u ˆ( x, y)<br />

of each element of the profile can be interpolated as follows:<br />

n<br />

uˆ<br />

= ∑ Niu i<br />

(6)<br />

where N i is the shape function and n the number of nodes for each element. <strong>The</strong> coordinates x and y for each<br />

element can be expressed as<br />

n<br />

i=<br />

1<br />

x = Mx,<br />

y=<br />

My<br />

i i i i<br />

i= 1 i=<br />

1<br />

n<br />

∑ ∑ (7)<br />

where M i is the shape function for the coordinates. Substituting Eqs. (5) and (6), into the virtual work equation,<br />

the equation of motion in frequency and wave number domain can be assembled from the element equations as<br />

([ K ] − ω<br />

2<br />

[ M ]){ D } = { F }<br />

(8)<br />

in which {F} denotes the vector of external loads, {D} the vector of nodal displacements, and [K] and [M] the<br />

global stiffness and mass matrices, respectively. In particular, the load vector {F} contains the unit nodal forces<br />

corresponding to function ψ (, x y)<br />

, to represent the moving load locations on the xy plane. <strong>The</strong> displacements<br />

{D} solved from Eq. (8) should be interpreted as the response function Hk (<br />

z<br />

, ω ) for the unit loads.<br />

Accordingly, the displacement response in time domain can be computed from Eq. (4) using the fast Fourier<br />

transformation.<br />

<strong>The</strong> shape functions in Eqs. (6) and (7) are selected to be identical to those of the corresponding plane elements.<br />

Take the Q8 element as an example. By substituting the shape functions of the conventional Q8 element into the<br />

above equations, a Q8-based finite element of the 2.5D version can be established. In this case, the size of the<br />

element matrices is 24×24, instead of 16×16 as for the conventional ones. <strong>The</strong> mass matrices remain real and<br />

symmetric, but the stiffness matrices are complex and asymmetric because of the existence of the terms exp(ik z z)<br />

in Eq. (5). As for the infinite elements, both the mass matrices and stiffness matrices are complex. For a finite<br />

element, the displacement field is often approximated by simple polynomials because of their relative ease in<br />

computation, while the accuracy can always be improved via mesh refinement. However, for an infinite element,<br />

it does not make much sense to refine the mesh along the direction leading to infinity, as far as radiation<br />

damping is concerned. Thus, an accurate shape function based on the analytical solution for a uniform<br />

visco-elastic half-space subjected to a moving harmonic point load is adopted (Yang et al. 1996).<br />

(a)<br />

(b)<br />

Figure 2 Infinite element: (a) global coordinates, (b) local coordinates<br />

<strong>The</strong> mapping shape functions M i and the displacement shape functions N i of the infinite element (Figure 2)<br />

adopted for the present moving load case are given in Eqs. (9) and (10), respectively.<br />

⎧ ( ξ −1)( η− 1) η ξ( η+<br />

1)<br />

⎪M1 =− M4<br />

=<br />

⎪<br />

2 2<br />

⎪<br />

ξ( η−1)<br />

⎨M2 = ( ξ −1)( η− 1)( η+ 1 ) M5<br />

= −<br />

⎪<br />

2<br />

⎪ ( ξ − 1)( η + 1)<br />

η<br />

⎪M<br />

3<br />

=−<br />

⎪⎩<br />

2<br />

⎧ ηη ( −1)<br />

⎪N1<br />

= exp ( −αξ<br />

) exp ( ik′<br />

ξ )<br />

2<br />

⎪<br />

⎨N2<br />

=−( η − 1)( η+ 1) exp( −αξ) exp( ik′<br />

ξ)<br />

⎪<br />

⎪ ηη ( + 1)<br />

N3<br />

= exp ( −αξ<br />

) exp ( ik′<br />

ξ )<br />

⎪⎩ 2<br />

whereα denotes the displacement amplitude decay factor and the exponent term exp( ik ξ )<br />

′ was adjusted to<br />

account for the wave propagating outward with the wave number k ′ . <strong>The</strong> wave number k ′ is determined from<br />

the analytical solution as<br />

(9)<br />

(10)<br />

-22-


⎛ω<br />

⎞<br />

2<br />

k′ i<br />

= ⎜ ⎟ −k<br />

(11)<br />

z<br />

⎝ci<br />

⎠<br />

where the subscript i represents the Rayleigh (R-), compressional (P-) or shear (S-) wave. Based on the fact that<br />

Rayleigh waves are dominant near the free surface, and that the body waves are dominant in soils of greater<br />

depths, it is suggested that the R-, S- and P- waves numbers be used respectively for k′ in the three regions I,<br />

II and III in Figure 3. Besides, the following amplitude decay factors are suggested for the three types of waves,<br />

2<br />

⎧ 1 kz<br />

⎪α<br />

R<br />

=<br />

2 2<br />

2R<br />

k<br />

⎪<br />

z<br />

+ kR<br />

⎪<br />

2<br />

1 1 kz<br />

⎨α<br />

P<br />

= +<br />

2 2<br />

(12)<br />

⎪ 2R<br />

2R<br />

kz<br />

+ kP<br />

⎪<br />

2<br />

1 1 kz<br />

⎪α<br />

S<br />

= +<br />

2 2<br />

⎩ 2R<br />

2R<br />

kz<br />

+ kS<br />

In Eq. (12), R denotes the distance between the source and far field boundary. Other issues deserving special<br />

attention for implementation of the present approach, such as the selection of element and mesh sizes and the<br />

techniques for enhancing the computational efficiency, should be referred to Yang et al. (1996).<br />

2<br />

SIMULATION OF THE LOADING FUNCTION<br />

Figure 3 Selection of wave numbers<br />

<strong>The</strong> characteristics of the trains and rails affects significantly the soil vibration responses. How to simulate the<br />

wheel-rail interaction force is a crucial issue in the soil vibration analysis. In this study, a one-layer continuously<br />

supported beam and a moving suspension system as given in Figure 4 are used in simulating such an effect.<br />

Vehicle Model<br />

Figure 4 Train model: (a) car model, (b) wheelset model<br />

<strong>The</strong> moving train is assumed to contain a sequence of N carriages. <strong>The</strong> load distribution function φ (z)<br />

of the<br />

train in Eq. (1) is regarded as the superposition of the distribution function q 0(<br />

z)<br />

associated with each of the<br />

axle loads, determined from the deflection curve of an infinite elastically supported beam under a unit load,<br />

1 ⎛ − z ⎞⎡<br />

⎛ z ⎞ ⎛ z ⎞⎤<br />

q0 ( z)<br />

= exp⎜<br />

⎟⎢<br />

cos⎜<br />

⎟ + sin⎜<br />

⎟<br />

⎝<br />

α<br />

⎠⎣<br />

⎝<br />

α<br />

⎠ ⎝<br />

α<br />

2α<br />

⎠⎥<br />

(13)<br />

⎦<br />

where the characteristic length α is related to the bending stiffness EI of the beam and the spring constant of<br />

the foundation s (N/m 2 ) as<br />

4 EI<br />

α = 4 (14)<br />

s<br />

<strong>The</strong> above concept for a single wheel load can be extended to the case of a train consisting of N carriages of<br />

different lengths L i (Figure 4), with L i indicating the length of the i-th car. By assuming that the two bogies of<br />

the i-th car are separated by distance b i , each of which in turn comprises two wheelsets separated by distance a i ,<br />

the total distribution function of loading for the train is<br />

-23-


⎡<br />

n<br />

n<br />

⎤<br />

⎢q 1 0( z− Li) q0( z Li a 1)<br />

N − ∑ + −∑<br />

− n+<br />

⎥<br />

⎢ i= 0 i=<br />

0<br />

⎥<br />

φ()<br />

z = ∑ ⎢<br />

n<br />

n<br />

⎥<br />

(15)<br />

n=<br />

0 ⎢+ q0( z− Li −an+ 1− bn+ 1) + q0( z− Li −2 an+ 1−bn+<br />

1)<br />

⎥<br />

⎢<br />

∑<br />

∑<br />

⎣<br />

⎥<br />

i= 0 i=<br />

0<br />

⎦<br />

where L 0 is the distance between the observation point to the front wheelset of the first carriage.<br />

Correspondingly, the Fourier transform of the load distribution function in the wave number domain is<br />

%<br />

N−1⎧<br />

n<br />

⎪<br />

⎡1 + exp( − ikzan+ 1) + exp ( − ikz( an+ 1<br />

+ bn<br />

1)<br />

) ⎤⎫<br />

+ ⎪<br />

φ ( kz) = q % 0<br />

( kz) ∑⎨exp( −ikz∑Li)<br />

⎢ ⎥⎬<br />

n= 0⎪<br />

i= 0<br />

⎩<br />

⎢⎣<br />

+ exp ( − ikz(2 an+ 1<br />

+ bn+<br />

1)<br />

)<br />

⎥<br />

(16)<br />

⎦⎪⎭<br />

where q<br />

~<br />

0 ( k z ) is transformed from Eq. (13) as<br />

4<br />

q<br />

~<br />

0(<br />

kz<br />

) =<br />

4 4<br />

(17)<br />

4 + kzα<br />

which is a function of the characteristic length α .<br />

1<br />

W<br />

4 cb<br />

u b<br />

/2 m b<br />

k s<br />

d p<br />

u w<br />

m w<br />

R( t )<br />

α i<br />

L i<br />

Figure 5 Schematic of the wheelset model and wheel-rail interaction force<br />

Derivation of Wheel-Rail Interaction Forces<br />

Each carriage of the train is composed of four wheelsets (Figure 4). Further, each wheelset is simulated by the<br />

mass-spring-dashpot system shown in Figure 5, with the wheelset weight set equal to one-fourth the weight of<br />

the car body W cb , plus half of the bogie weight m b and the wheel weight m w . Only the primary suspension<br />

system is considered for the wheelset. By letting u b (t) and u w (t) denote the vertical displacement of the bogie and<br />

wheelset, k s and d p the spring and damping coefficient, respectively, of the suspension system of the vehicle load,<br />

and R(t) the wheel-rail interaction force, the equation of motion for each wheelset can be written as follows:<br />

⎡mb 2 0 ⎤⎧u&&<br />

b() t ⎫ ⎡ dp<br />

−dp⎤⎧u&<br />

b() t ⎫ ⎡ ks − ks⎤⎧ub() t ⎫ ⎧mbg 2+<br />

Wcb<br />

4⎫<br />

⎢<br />

0 m<br />

⎥⎨ ⎬+ ⎢<br />

w<br />

uw( t) dp<br />

d<br />

⎥⎨ ⎬+ p uw( t) ⎢<br />

ks k<br />

⎥⎨ ⎬=<br />

⎨ ⎬ (18)<br />

⎣ ⎦⎩&&<br />

⎭ ⎣<br />

−<br />

⎦⎩&<br />

⎭ ⎣− s ⎦⎩uw( t) ⎭ ⎩ mwg−R( t)<br />

⎭<br />

By the Fourier transformation, the preceding equation can be written in the frequency domain as<br />

2<br />

⎡mb 2 0 ⎤⎧u% b( ω) ⎫ ⎡ dp<br />

−dp⎤⎧u% b( ω) ⎫ ⎡ ks −ks⎤⎧u%<br />

b( ω)<br />

⎫<br />

− ω ⎢ ⎥⎨ ⎬+ iω⎢<br />

0 mw uw( ) dp<br />

d<br />

⎥⎨ ⎬+<br />

⎢ ⎥⎨ ⎬<br />

⎣ ⎦⎩% ω ⎭ ⎣<br />

−<br />

p ⎦⎩u %<br />

w( ω) ⎭ ⎣−ks ks ⎦⎩u%<br />

w( ω)<br />

⎭<br />

(19)<br />

⎧⎪<br />

( mg<br />

b<br />

2+<br />

Wcb<br />

4) δ ( ω)<br />

⎫⎪<br />

= ⎨<br />

mg<br />

w<br />

δ ( ω)<br />

− R% ⎬<br />

⎪⎩<br />

( ω)<br />

⎭⎪<br />

<strong>The</strong> interaction force will be included by letting a mass-spring-dashpot system travel along a random rail surface,<br />

as depicted in Figure 5, generated by the superposition of cosine functions given in Eq. (20) with random phase<br />

angles θ<br />

i<br />

in the interval [0,2π]:<br />

n<br />

uwr / () z = ∑ αicos( kiz−θi)<br />

(20)<br />

i=<br />

1<br />

where α i is the amplitude of the i-th cosine-shaped rail surface; k i = 2π/L i is the wavenumber of the i-th<br />

cosine-shaped rail surface and L i is the corresponding wavelength. By assuming the moving wheelset to be in<br />

full contact with the rails as shown in Figure 5, the vertical displacement u w (t) of the wheelset is equal to the<br />

irregularity profile u w/r (z) given in Eq. (20), with z replaced by ct, namely,<br />

n<br />

uw() t = ∑ αicos( kict−θi)<br />

(21)<br />

<strong>The</strong> Fourier transform of the vertical displacement of the wheelset in Eq. (21) is<br />

i=<br />

1<br />

-24-


n<br />

⎛1 1<br />

⎞<br />

u% w( ω) = ∑αi⎜<br />

δ ( ω−kc i ) exp( − iθi) + δ( ω+<br />

kc<br />

i ) exp( iθi)<br />

i=<br />

1 2 2<br />

⎟<br />

(22)<br />

⎝<br />

⎠<br />

~<br />

<strong>The</strong> interaction force R ( ω ) can be solved from Eq. (19) using Eq. (22) for u ~ ( ω ) ,<br />

where ω<br />

i<br />

= kc<br />

i<br />

,<br />

n<br />

∑<br />

( ) ( )<br />

R % ( ω) = Wδω ( ) + W δ ω− ω + W δ ω+<br />

ω<br />

(23)<br />

1 1<br />

1 = w +<br />

2 b +<br />

4 cb<br />

W m g m g W<br />

1 2, i i 3, i i<br />

i= 1 i=<br />

1<br />

,<br />

n<br />

∑<br />

( ks + iωi dp)<br />

⎛<br />

2<br />

⎞<br />

α<br />

W<br />

⎜<br />

k i d m<br />

⎟<br />

exp( i )<br />

i<br />

2<br />

2, i =<br />

1 2<br />

2<br />

s + ωi p − wωi − θi<br />

ks iωi dp 2<br />

mbω<br />

−<br />

⎜<br />

+ − i ⎟<br />

⎝<br />

⎠<br />

2<br />

( ks − iωi dp)<br />

⎛<br />

⎞<br />

α<br />

W<br />

⎜<br />

k i d m<br />

⎟<br />

exp( i )<br />

i<br />

2<br />

3, i =<br />

2<br />

s − ωi p − wωi −<br />

θ<br />

1 2 i<br />

ks iωi dp 2<br />

mbω<br />

⎜<br />

− − i ⎟<br />

⎝<br />

⎠<br />

~<br />

As indicated by Eqs. (23) and (24), the contact force R ( ω ) contains three terms with delta functions and each<br />

with the weight functions W 1 , W 2,i and W 3,i . <strong>The</strong> first term with delta function δ (ω)<br />

corresponds to the static<br />

weight of the wheelset, with W 1 equal to one quarter of the car weight. <strong>The</strong> second and third terms with delta<br />

functions δ ( ω− ω i<br />

) and δ ( ω+ ω i<br />

) represent the dynamic force with frequencies ω =± ωi( =± 2 πcLi)<br />

caused by rail irregularities, which depends not only on the wave length of rail irregularity L i , but on the train<br />

speed c. <strong>The</strong> weights W 2,i and W 3,i also depend on the frequency ω and the dynamic properties of the train.<br />

~<br />

By replacing ω in Eq. (23) with k z c + ω , the function R(<br />

k z c + ω)<br />

can be obtained. By substituting the two<br />

~ ~<br />

functions φ ( k z ) and R(<br />

k z c + ω)<br />

into Eq. (4), the inversion of the Fourier transform with respect to k z can be<br />

done analytically by use of Dirac’s delta functions in Eq. (23). Consequently, the original double integral<br />

reduces to a single integral with respect to frequency ω only, as given below:<br />

W ∞<br />

1<br />

d( x′ , y′ , z, t) = % φ( kz) H( kz, ω)exp( iω t)<br />

dω<br />

c ∫−∞<br />

ω<br />

n<br />

⎛<br />

⎞<br />

W2,<br />

∞<br />

i<br />

+<br />

⎜ % φ( kz) H( kz, ω)exp( iω t)<br />

dω<br />

⎟<br />

∑<br />

i 1 c ∫−∞<br />

ω−ω<br />

= ⎜<br />

i<br />

kz<br />

=− ⎟<br />

(25)<br />

⎝<br />

c ⎠<br />

n<br />

⎛<br />

⎞<br />

W3,<br />

∞<br />

i<br />

+<br />

⎜ % φ( kz) H( kz, ω)exp( iω t)<br />

dω<br />

⎟<br />

∑<br />

i 1 c ∫−∞ = ω+<br />

ω<br />

⎜<br />

i<br />

kz<br />

=− ⎟<br />

⎝<br />

c ⎠<br />

In this study, the transfer function H ( k z , ω)<br />

is calculated using the 2.5D finite/infinite element approach. <strong>The</strong><br />

total displacement response of the system in Eq. (25) is obtained by the fast Fourier transformation. <strong>The</strong> velocity<br />

and acceleration response can also be obtained by replacing the response function H ( k z , ω)<br />

with<br />

2<br />

i ω H ( k , ω)<br />

and − ω H ( k , ω)<br />

, respectively.<br />

z<br />

Simulation of Random Rail Irregularities<br />

z<br />

In order to take the randomness nature of the rail irregularity into account, the rail irregularities u w/r (y) in Eq. (20)<br />

is generated by a stochastic process characterized by a single-sided power spectral density (PSD) as given in Eq.<br />

(26), which is a function of the wavenumber k y (rad/m) (Hamid and Yang 1982).<br />

' 2 2 2<br />

A ky2( ky ky<br />

1)<br />

G % +<br />

wr /<br />

( ky)<br />

=<br />

4 2 2<br />

k ( k + k )<br />

(26)<br />

y y y2<br />

where k y1 = 0.1464 (rad/m) and k y2 = 0.8244 (rad/m) are the break frequencies that do not change significantly<br />

for different track classes. According to Federal Railroad Administration (FRA), the rail quality of the tracks can<br />

be divided into six classes, with class 6 track indicating the best quality and class 1 the poorest. <strong>The</strong><br />

'<br />

irregularities parameter A in Eq. (26) is strongly dependent on the track class as listed in Table 1.<br />

<strong>The</strong> artificial rail irregularities profile u w/r (y) adopted in this study is generated from the PSD function based on<br />

the superposition of simple random processes with corresponding statistical properties. <strong>The</strong> upper and lower<br />

bounds of the wavenumber [k yl , k yu ] are defined for the single sided PSD G% /<br />

( k ) considering the range of<br />

i<br />

w<br />

kz<br />

=−<br />

c<br />

wr<br />

y<br />

(24)<br />

-25-


frequencies of interest and the train speed. This range is divided into n intervals with width Δ k y and center<br />

wavenumber k yi , which can be obtained from the following:<br />

kyu<br />

− kyl<br />

1<br />

Δ ky<br />

= , kyi<br />

= ( i− ) Δ k , 1~<br />

2 y<br />

i = n<br />

(27)<br />

n<br />

in which k i represents the wavenumber of the i-th cosine-shaped rail surface. By substituting k yi and Δ k y into<br />

Eq. (28), the amplitudes of irregularity profile α<br />

i<br />

for the i-th cosine-shaped rail surface can be obtained.<br />

α = 2 G% ( k ) Δk<br />

(28)<br />

Case Study and Discussion<br />

i w/<br />

r yi y<br />

Table 1 Irregularities parameter for FRA track classes (Hamid and Yang 1982).<br />

Track class 6 5 4 3 2 1<br />

'<br />

A [10-7 m-cycle] 1.06 1.69 2.96 5.29 9.52 16.72<br />

As shown in Figure 6, an illustrative example is given on the vibration mitigation of the floating slab track for<br />

use in a tunnel embedded in a visco-elastic half-space subjected to a moving train. <strong>The</strong> 2.5D finite/infinite<br />

method described previously was employed to calculate the ground response caused by a train moving over<br />

uneven rails. Figure 7 shows the element mesh used, which has a width of 52 m and a depth of 22 m, with<br />

different colors representing different material properties. Only half of the system is adopted in analysis due to<br />

the symmetry consideration. To study the screening effect of the floating slab track on the vibrations caused by<br />

the moving train over irregular rails, two cases are studied. In the first case, an elastic foundation is inserted<br />

between the concrete slab and concrete tunnel lining to simulate the effect of the floating slab track. In the<br />

second case, a direct fastened track, i.e., with no consideration made for the elastic foundation, is considered.<br />

All the material properties, including soil properties, tunnel and track parameters, for the analytical model have<br />

been listed in Table 2, where the elastic foundation has a Young’s modulus much smaller than that of the<br />

concrete slab. Using the present data for soil, the shear and compressional wave velocities C s and C p computed<br />

are also shown in Figure 6. Besides, the centroid of the tunnel is located at a depth of h = 13.5 m beneath the<br />

ground, the inner diameter of the tunnel is 5.4 m, and the wall thickness of the tunnel is t = 30 cm.<br />

Figure 6 Soil-tunnel model adopted in analysis<br />

52 m<br />

13.5 m<br />

22 m<br />

Figure 7 Finite/infinite element mesh<br />

<strong>The</strong> train consists of 6 identical cars, with the following data: length = 19 m, a = 2.3 m, b = 12.6 m, weight of<br />

car body (including passengers) W cb = 411.6 MN, bogie mass m b = 3,600 kg, wheelset mass m w = 1,700 kg,<br />

-26-


vertical stiffness of primary suspension system k s = 1.4 MN/m, vertical damping of primary suspension system<br />

d p = 0.03MN-s/m, and characteristic length α = 0.745 m. FRA track class 1 is adopted in generating the<br />

artificial irregularities profile u w/r (y) for the rails, which corresponds to the poorest rail quality. <strong>The</strong> lower and<br />

upper bounds of the wavenumbers [k yl , k yu ] defined in the single sided PSD are chosen to be 0.1 and 15 rad/m,<br />

respectively, with n = 40 intervals. <strong>The</strong> corresponding wavelengths of rail unevenness considered range from L i<br />

= 0.42 to 62.8 m. By assuming the train speed to be c = 50 m/s, the major frequencies involved in the rail<br />

irregularities range from 7 to 119 Hz. Thus the analysis frequency range selected is from 0 to 150 Hz.<br />

Table 2. Material Properties.<br />

Young’s modulus<br />

E (MPa)<br />

Poisson’s ratio<br />

υ<br />

Mass density<br />

ρ (kg/m 3 )<br />

Damping ratio<br />

β<br />

Material<br />

Concrete tunnel lining 35,000 0.25 2,500 0.02<br />

Concrete slab 28,500 0.2 2,500 0.02<br />

Elastic Foundation 0.5 0.25 150 0.1<br />

Fill material 116.6 0.341 1,900 0.05<br />

Silty clay 289 0.313 2,023 0.04<br />

Gravel and pebble 704 0.223 1,963 0.03<br />

For train speed equal to c = 50 m/s, the responses computed for X = 0 m and X = 50 m on the ground with and<br />

without an elastic foundation supporting the concrete slab track were plotted in Figure 8, with parts (a) to (c)<br />

showing the time histories of displacement, velocity and acceleration, respectively. As indicated by Figure 8(a),<br />

there exists a localized quasi-static displacement combined with distinct fluctuating vibrations for the case<br />

without floating slab track. This is mainly caused by the moving tributary weight of the train, quasi-static in<br />

nature, while the fluctuating vibrations are induced by rail irregularities. For the case with floating slab track,<br />

although the fluctuating vibrations with high frequencies are reduced by the elastic foundation, the localized<br />

displacements and the fluctuating vibration with low frequencies increase.<br />

(a) Displacement<br />

(b) Velocity<br />

(c) Acceleration<br />

Figure 8 Effect of elastic foundation on the ground response caused by a moving train over uneven rails: (a)<br />

displacement, (b) velocity, (c) acceleration<br />

<strong>The</strong> isolation effect of the floating slab track is clearer from the velocity and acceleration responses in Figures<br />

8(b) and (c), as the high-frequency components have been suppressed. Such an observation is also confirmed by<br />

the spectra shown in Figures 9(a) and (b) for the responses at X = 0 m and 50 m, respectively. <strong>The</strong> isolation<br />

effect of elastic foundations is generally poor for vibrations at lower frequencies, but effective for higher<br />

frequencies. Because the acceleration has a frequency content covering mostly higher frequencies, the isolation<br />

effect of floating slab track is generally significant. In contrast, the frequency content of the displacement<br />

response is mostly localized on the low-frequency region, which makes the effect of isolation not so effective.<br />

-27-


Another phenomenon observed from Figure 8 is that the isolation effect of the floating slab at the location X =<br />

50 m (far field region) is not as good as that at location X = 0 m (near field region). This result can be attributed<br />

to the low attenuation rate of the amplified low-frequency vibrations and the high attenuation rate of the<br />

high-frequency vibrations. By comparing Figures 9 (a) and (b), one notes that the attenuation rate of vibration<br />

for high frequency components decay much faster than that of the low frequency components.<br />

Amplitude of displacement (m)<br />

Amplitude of velocity (m/s)<br />

10-4 Displacement<br />

without floating slab<br />

3<br />

with floating slab<br />

2<br />

1<br />

0<br />

0 50 100 150<br />

4 x Frequency (Hz)<br />

6 10-4 Velocity<br />

without floating slab<br />

4<br />

with floating slab<br />

2<br />

0<br />

0 50 100 150<br />

x<br />

Frequency (Hz)<br />

Amplitude of acceleration (m/s 2 )<br />

0.1<br />

0.05<br />

Amplitude of displacement (m)<br />

Amplitude of velocity (m/s)<br />

Acceleration<br />

without floating slab<br />

with floating slab<br />

0<br />

0 50 100 150<br />

Frequency (Hz)<br />

1<br />

x 10 -4<br />

Displacement<br />

(a) X = 0<br />

without floating slab<br />

with floating slab<br />

0<br />

0 50 100 150<br />

Frequency (Hz)<br />

10-4<br />

Velocity<br />

without floating slab<br />

3<br />

with floating slab<br />

2<br />

1<br />

0<br />

0 50 100 150<br />

4 x Frequency (Hz)<br />

Amplitude of acceleration (m/s 2 )<br />

0.015<br />

0.01<br />

0.005<br />

Acceleration<br />

without floating slab<br />

with floating slab<br />

0<br />

0 50 100 150<br />

Frequency (Hz)<br />

(b) X= 50 m<br />

Figure 9 Effect of elastic foundation on the ground response in frequency domain caused by a moving train over<br />

uneven rails: (a) at X = 0, (b) at X= 50 m<br />

<strong>The</strong> attenuation of ground vibration for the case with and without elastic foundation can also be evaluated by the<br />

method proposed below. <strong>The</strong> results for the displacement, velocity and acceleration responses in Figure 10 have<br />

been plotted on the following logarithmic scale:<br />

P1<br />

L ( dB)<br />

= 20log<br />

(29)<br />

P2<br />

where P 1 is the computed velocity or acceleration amplitude and P 2 a reference value, selected as P 2 =10 -11 m,<br />

10 -8 m/s and 10 -5 m/s 2 . As can be seen from Figure 10, the installation of elastic foundations can reduce both<br />

the velocity and acceleration levels by nearly 10 and 20 dB, respectively for the case considered. However, the<br />

displacement response is amplified after the floating slab is installed.<br />

CONCLUSIONS<br />

In this paper, the 2.5D finite/infinite element approach was briefly introduced. With the aid of the proposed<br />

train-load generating mechanism, both the moving load effect induced by the static train weights and the<br />

dynamic effect induced by the interaction between the moving trains and uneven rails with random irregularities<br />

are taken into account. <strong>The</strong> present approach is featured by the fact that the geometry considered is<br />

two-dimensional, i.e., only a profile perpendicular to the railway is considered, but the train loads are allowed to<br />

move along the third direction perpendicular to the 2D profile, and that the 3D response can be obtained in an easy<br />

way. As an illustration, a soil vibration problem caused by underground trains traveling on uneven rails is studied,<br />

with focus on the vibrations caused by uneven rails. By comparing the results for the case with an elastic<br />

foundation placed under the track slab with the other case with no elastic foundation, the screening effect of<br />

-28-


elastic foundations was also studied. Based on this case study, it is confirmed that the present method can be<br />

potentially applied to problems of practical nature.<br />

140<br />

)<br />

B<br />

138<br />

v<br />

e<br />

l<br />

(d<br />

le<br />

t<br />

136<br />

e<br />

n<br />

134<br />

c<br />

e<br />

m<br />

la<br />

132<br />

isp<br />

a<br />

x<br />

.<br />

d<br />

130<br />

M<br />

128<br />

without floating slab track<br />

with floating slab track<br />

0 5 10 15 20 25 30 35 40 45 50<br />

X (m)<br />

(a) Displacement level<br />

100<br />

)<br />

B90<br />

v<br />

e<br />

l<br />

(d<br />

80<br />

le<br />

n<br />

tio 70<br />

ra<br />

60<br />

50<br />

a<br />

x<br />

.<br />

a<br />

c<br />

e<br />

le<br />

M<br />

40<br />

110<br />

) 105<br />

B<br />

100<br />

v<br />

e<br />

l<br />

(d<br />

le<br />

95<br />

c<br />

ity<br />

90<br />

a<br />

x<br />

.<br />

v<br />

e<br />

lo<br />

M85<br />

80<br />

0 5 10 15 20 25 30 35 40 45 50<br />

X (m)<br />

without floating slab track<br />

with floating slab track<br />

0 5 10 15 20 25 30 35 40 45 50<br />

X (m)<br />

without floating slab track<br />

with floating slab track<br />

(b) Velocity level<br />

(c) Acceleration level<br />

Figure 10 Effect of elastic foundation on ground response attenuation for a train moving in underground tunnel: (a)<br />

displacement level, (b) velocity level, (c) acceleration level<br />

ACKNOWLEDGMENTS<br />

This study was sponsored partially by the National Science Council via Grant No. NSC 98-2221-E-002 -106<br />

-MY2. This paper was rewritten from the one presented at the Symposium on the Development of<br />

Computational Mechanics and Computational Methods in Engineering and Science in celebrating the 25th<br />

Anniversary of EPMESC Conferences and the 30th Anniversary of <strong>University</strong> of Macau, held on April 4, 2011<br />

in Macau.<br />

REFERENCES<br />

Celebi, E. (2006). “Three-dimensional modelling of train-track and sub-soil analysis for surface vibrations due to<br />

moving loads”, Applied Mathematics and Computation, 179(1), 209-230.<br />

Hamid, A. and Yang, T.-L. (1982). “Analytical Descriptions of Track-Geometry Variations”, Transportation<br />

Research Record, 838, 19-26.<br />

Lei, X. and Noda, N.A. (2002). “Analyses of dynamic response of vehicle and track coupling system with random<br />

irregularity of track vertical profile,” Journal of Sound and Vibration, 258(1), 147-165.<br />

Metrikine, A.V. and Vrouwenvelder, A.C.W.M. (2000). “Surface ground vibration due to a moving train in a<br />

tunnel: two-dimensional model”, Journal of Sound and Vibration, 234(1), 43-66.<br />

Mohammadi, M. and Karabalis, D.L. (1995). “Dynamic 3-D soil-railway track interaction by BEM-FEM”,<br />

Earthquake Engineering & Structural Dynamics, 24(9), 1177-1193.<br />

Paolucci, R., Maffeis, A., Scandella, L., Stupazzini, M., and Vanini, M. (2003). “Numerical prediction of<br />

low-frequency ground vibrations induced by high-speed trains at Ledsgaard, Sweden”, Soil Dynamics and<br />

Earthquake Engineering, 23(6), 425-433.<br />

Sheng, X., Jones, C.J.C., and Thompson, D.J. (2004). “A theoretical model for ground vibration from trains<br />

generated by vertical track irregularities”, Journal of Sound and Vibration, 272(3-5), 937-965.<br />

Yang, Y.B. and Hung, H.H. (2001). “A 2.5D finite/infinite element approach for modelling visco-elastic bodies<br />

subjected to moving loads”, International Journal for Numerical Methods in Engineering, 51(11),<br />

1317-1336.<br />

Yang, Y.B. and Hung, H.H. (2009). Wave Propagation for Train-induced Vibrations, Singapore, World<br />

Scientific.<br />

Yang, Y.B., Kuo, S.R., and Hung, H.H. (1996). “Frequency-Independent Infinite Elements for Analyzing<br />

Semi-Infinite Problems”, International Journal for Numerical Methods in Engineering, 39, 3553-3569.<br />

-29-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

THE ZHOUQU DEBRIS FLOW<br />

C. F. Lee<br />

Department of Civil Engineering,<br />

<strong>The</strong> <strong>University</strong> of <strong>Hong</strong> <strong>Kong</strong>, <strong>Hong</strong> <strong>Kong</strong>, China.<br />

ABSTRACT<br />

On August 8, 2010, a major debris flow involving 1.8Mm 3 of materials demolished part of the town of Zhouqu<br />

in southern Gansu Province, resulting in the loss of more than 1,760 lives. <strong>The</strong> town is located on an alluvial<br />

fan right on the flow path of a creek and hence on the runout path of debris flow in the event of heavy rainfall.<br />

This mountainous area, located at the northeastern edge of the Qinghai–Tibetan Plateau, has a long history of<br />

debris flow occurrence, particularly over the past two centuries. This talk will cover the causative mechanism of<br />

the Zhouqu debris flow, and the remedial measures adopted to mitigate future hazards of debris flow occurrence<br />

in the region. It will also present an overview of debris flow hazards and their mitigation in Sichuan, Yunnan,<br />

Gansu and other mountainous parts of China.<br />

-30-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

关 于 极 限 分 析 方 法 与 数 值 极 限 分 析 方 法<br />

郑 颖 人<br />

1,2 , 唐 晓 松<br />

1,2 1,2<br />

, 赵 尚 毅<br />

1. 后 勤 工 程 学 院 军 事 建 筑 工 程 系 , 重 庆 400041;<br />

2. 重 庆 市 地 质 灾 害 防 治 工 程 技 术 研 究 中 心 , 重 庆 400041;<br />

摘 要 : 传 统 极 限 分 析 方 法 已 有 百 年 以 上 的 历 史 , 尤 其 在 岩 土 工 程 中 广 为 应 用 , 并 作 为 设 计 的 依<br />

据 , 但 传 统 极 限 分 析 法 的 适 用 范 围 有 限 , 诸 多 工 程 问 题 无 法 求 解 , 影 响 了 它 的 发 展 。 近 年 来 , 数 值<br />

分 析 方 法 与 极 限 分 析 相 结 合 , 形 成 了 数 值 极 限 分 析 方 法 , 极 限 分 析 方 法 的 适 用 范 围 大 幅 扩 大 。 然 而<br />

人 们 对 传 统 极 限 分 析 法 与 数 值 极 限 法 的 内 涵 、 特 点 、 优 缺 点 和 今 后 需 要 研 究 的 问 题 尚 缺 乏 深 入 的 了<br />

解 。 本 文 对 两 者 的 内 涵 做 了 深 入 的 分 析 , 论 证 了 极 限 分 析 方 法 的 可 靠 性 , 比 较 了 两 种 方 法 的 不 同 点<br />

与 优 缺 点 , 也 指 出 了 数 值 极 限 分 析 法 存 在 的 问 题 , 以 及 当 前 急 需 解 决 的 难 点 。 并 望 数 值 极 限 分 析 法<br />

在 岩 土 工 程 的 各 个 领 域 发 扬 光 大 , 成 为 各 类 岩 土 工 程 设 计 计 算 的 有 力 工 具 。<br />

关 键 词 : 极 限 分 析 数 值 极 限 分 析 极 限 状 态 极 限 荷 载 稳 定 安 全 系 数 滑 移 线 场<br />

THE THEORY OF LIMIT ANALYSIS AND THE METHOD OF NUMERICAL LIMIT<br />

ANALYSIS<br />

Zheng Yingren 1,2 ,Tang Xiaosong 1,2 ,Zhao Shangyi<br />

(1. Department of Civil Engineering, Logistical Engineering <strong>University</strong>, Chongqing 400041, China;2.<br />

Chongqing Engineering and Technology Research Center of Geological Hazard Prevention and Treatment,<br />

Chongqing 400041, China)<br />

Abstract: <strong>The</strong> traditional limit analysis has been existed for over one hundred years, which is widely used in<br />

geotechnical engineering and taken as the design criterion. But due to its confined scope of application, this<br />

traditional method cannot solve some engineering problems so that its development is influenced. In recent<br />

years, the combination of numerical analysis method and limit analysis leads to numerical limit analysis, as a<br />

result, the scope of application has been enlarged greatly. However, researchers lack full understanding about<br />

the connotations, characteristics, advantages and disadvantages of traditional limit analysis and numerical<br />

analysis, as well as the problems which need to be solved in the future. This paper deeply analyzes the<br />

connotations of the these two methods, demonstrates the reliability of limit analysis, compares the differences,<br />

advantages and disadvantages of the two and points out the problems existing in the numerical limit analysis and<br />

some urgent difficulties at present. <strong>The</strong> writers hope that numerical limit analysis could be widely adopted in<br />

geotechnical engineering and become an effective instrument in the design and calculation of various<br />

geotechnical engineering.<br />

Keywords: limit analysis, numerical limit analysis, limit state, ultimate load, stability safety factor, slip line<br />

field<br />

一 、 前 言<br />

经 典 的 岩 土 稳 定 性 问 题 包 括 边 坡 稳 定 、 地 基 承 载 力 、 土 压 力 等 , 其 理 论 基 础 是 极 限 分 析 理 论 ,<br />

土 体 的 极 限 分 析 起 始 于 1773 年 的 库 仑 定 律 ,20 世 纪 20 年 代 Fellenius 等 人 建 立 了 极 限 平 衡 法 ,40<br />

年 代 , 又 相 继 出 现 了 Sokolovskii 等 人 的 滑 移 线 场 法 ( 特 征 线 法 ),50 年 代 又 提 出 了 极 限 分 析 的 上 、<br />

下 限 法 。 经 过 100 年 的 发 展 已 逐 趋 成 熟 。 从 工 程 实 践 上 看 , 极 限 分 析 法 具 有 很 好 的 效 果 , 解 决 了 岩<br />

土 工 程 的 一 些 设 计 问 题 , 尤 其 是 强 度 与 稳 定 问 题 。 但 对 复 杂 的 层 状 、 非 均 质 岩 土 材 料 , 各 类 工 程 的<br />

复 杂 情 况 , 这 一 方 法 往 往 无 能 为 力 。 随 着 岩 土 力 学 数 值 方 法 的 发 展 , 逐 渐 兴 起 了 数 值 极 限 分 析 方 法 ,<br />

它 既 有 很 广 的 适 用 性 , 又 有 很 好 的 实 用 性 。 但 工 程 人 员 对 极 限 分 析 法 的 基 本 理 论 、 方 法 的 内 在 联 系<br />

还 不 清 楚 , 本 文 对 其 实 质 与 方 法 之 间 的 联 系 作 了 简 要 的 叙 述 。 同 时 , 阐 述 了 数 值 极 限 分 析 法 的 含 义<br />

及 其 与 传 统 极 限 分 析 法 的 不 同 点 、 数 值 极 限 分 析 法 的 优 越 性 、 以 及 日 后 需 要 研 究 的 问 题 , 使 读 者 对<br />

这 两 种 极 限 分 析 方 法 有 较 好 的 理 解 。<br />

[1-6]<br />

二 、 关 于 极 限 分 析 理 论 的 含 义<br />

基 金 项 目 : 科 技 部 973 项 目 “ 重 大 工 程 灾 变 滑 坡 演 化 与 控 制 的 基 础 研 究 (2011CB710603)”; 重 庆 市 自 然 科 学 基 金 项 目 “ 地 震 作 用 下 边<br />

坡 稳 定 性 新 方 法 及 支 护 结 构 的 抗 震 性 能 研 究 (2010BC8002)”<br />

作 者 简 介 : 郑 颖 人 (1933-), 男 , 浙 江 宁 波 人 , 教 授 , 中 国 工 程 院 院 士 , 从 事 岩 土 力 学 与 岩 土 工 程 的 教 学 与 科 研 工 作 。<br />

-31-


刚 塑 性 平 面 应 变 极 限 分 析 传 统 理 论 采 用 理 想 塑 性 模 型 来 研 究 材 料 达 到 极 限 状 态 时 的 力 学 关 系 ,<br />

由 此 可 以 求 出 材 料 的 极 限 荷 载 或 稳 定 安 全 系 数 。 应 用 于 金 属 成 形 加 工 、 土 坡 稳 定 与 地 基 承 载 力 大 小<br />

等 领 域 。 极 限 荷 载 对 应 着 材 料 进 入 破 坏 状 态 , 此 时 荷 载 不 变 , 变 形 可 不 断 增 大 , 沿 滑 面 破 坏 材 料 达<br />

到 破 坏 状 态 , 对 应 的 荷 载 为 极 限 荷 载 ; 稳 定 安 全 系 数 对 应 着 滑 面 上 材 料 的 抗 滑 力 ( 与 材 料 强 度 有 关 )<br />

与 滑 动 力 ( 与 承 受 的 荷 载 有 关 ) 之 比 ; 也 可 写 成 极 限 荷 载 与 实 际 荷 载 之 比 , 当 安 全 系 数 为 1 时 , 材<br />

料 发 生 破 坏 。 应 当 注 意 滑 面 上 的 力 不 是 点 的 应 力 , 而 是 滑 面 上 的 总 力 。 它 是 当 前 判 别 材 料 整 体 剪 切<br />

失 稳 的 唯 一 判 据 , 从 上 述 含 义 看 极 限 分 析 理 论 不 仅 可 以 求 极 限 荷 载 与 稳 定 安 全 系 数 , 而 且 可 以 求 材<br />

料 中 的 破 坏 状 态 , 因 而 是 一 种 十 分 切 合 工 程 设 计 的 有 效 方 法 , 只 是 目 前 传 统 极 限 分 析 方 法 中 需 要 预<br />

先 知 道 滑 面 状 况 才 能 求 解 , 这 是 由 于 方 法 本 身 的 局 限 性 所 致 。<br />

然 而 材 料 是 否 都 会 达 到 极 限 状 态 是 有 条 件 限 制 的 , 一 般 要 求 材 料 内 产 生 一 定 的 位 移 , 此 时 强 度<br />

才 能 充 分 发 挥 , 当 位 移 受 到 制 约 时 , 有 时 材 料 达 不 到 极 限 状 态 , 此 时 极 限 分 析 失 效 。<br />

[1-6]<br />

三 、 关 于 极 限 分 析 的 力 学 方 法<br />

在 传 统 平 面 极 限 分 析 理 论 中 , 可 由 平 衡 方 程 屈 服 条 件 、 应 力 应 变 关 系 、 体 积 不 可 压 缩 条 件 等 五<br />

个 方 程 , 求 解 三 个 应 力 分 量 和 两 个 速 度 分 量 , 共 五 个 未 知 量 。 通 常 求 解 极 限 分 析 问 题 , 常 常 分 成 两<br />

步 走 , 先 应 用 平 衡 方 程 与 屈 服 条 件 求 出 三 个 应 力 分 量 , 也 可 由 此 求 出 极 限 荷 载 或 稳 定 安 全 系 数 。 然<br />

后 依 据 求 出 的 应 力 再 求 速 度 分 量 , 显 然 这 不 是 严 格 的 力 学 解 法 。 对 于 强 度 与 稳 定 问 题 , 目 的 是 求 极<br />

限 荷 载 或 稳 定 安 全 系 数 , 采 用 上 述 第 一 步 , 应 用 平 衡 方 程 与 屈 服 方 程 就 可 以 了 , 不 必 引 入 本 构 关 系 ,<br />

从 而 使 求 解 大 为 简 化 。 近 百 年 来 , 极 限 分 析 法 在 工 程 上 的 应 用 或 做 室 内 试 验 , 都 证 明 了 这 一 方 法 的<br />

可 行 性 , 尤 其 是 最 近 几 年 来 有 限 元 强 度 折 减 法 的 出 现 , 采 用 弹 塑 性 数 值 方 法 求 解 极 限 问 题 , 计 算 表<br />

明 传 统 的 极 限 分 析 方 法 与 数 值 极 限 分 析 方 法 可 以 得 到 同 样 的 结 果 , 进 一 步 论 证 了 极 限 分 析 法 的 可 靠<br />

性 , 也 证 明 了 上 述 求 解 方 法 在 工 程 上 的 可 行 性 。 这 也 表 明 在 求 极 限 荷 载 与 稳 定 安 全 系 数 时 连 续 性 条<br />

件 可 以 暂 时 不 满 足 。<br />

极 限 分 析 法 与 材 料 力 学 和 弹 性 力 学 计 算 的 不 同 。 材 料 力 学 和 弹 性 力 学 是 求 荷 载 作 用 下 , 材 料 所<br />

受 的 内 力 , 不 引 入 材 料 的 强 度 , 计 算 中 没 有 强 度 参 数 。 而 极 限 分 析 法 是 研 究 材 料 极 限 状 态 时 的 力 学<br />

关 系 , 计 算 中 引 入 了 屈 服 准 则 , 而 准 则 中 既 有 应 力 状 态 , 又 有 抗 力 状 态 , 它 与 强 度 有 关 。 当 求 得 真<br />

实 破 坏 面 上 的 滑 动 力 与 抗 滑 力 时 , 就 可 求 出 材 料 的 安 全 系 数 或 极 限 荷 载 。 因 而 多 年 来 被 作 为 岩 土 工<br />

程 设 计 的 依 据 , 它 对 二 维 、 三 维 的 岩 土 问 题 特 别 适 用 。<br />

[1-6]<br />

四 、 极 限 分 析 的 实 用 方 法<br />

经 典 塑 性 力 学 中 应 用 的 极 限 分 析 实 用 方 法 , 一 般 是 滑 移 线 场 法 与 极 限 分 析 中 的 上 、 下 限 法 ; 而<br />

在 岩 土 边 坡 稳 定 分 析 中 除 上 述 方 法 外 还 采 用 极 限 平 衡 法 。<br />

(1)、 滑 移 线 场 法<br />

滑 移 线 就 是 破 裂 面 的 迹 线 。 滑 移 线 场 法 就 是 按 照 滑 移 线 场 理 论 和 边 界 条 件 , 先 在 受 力 体 内 构 造<br />

相 应 的 滑 移 线 网 , 引 入 平 衡 方 程 发 展 成 为 滑 动 面 上 极 限 平 衡 方 程 , 然 后 利 用 滑 移 线 的 性 质 与 边 界 条<br />

件 求 出 塑 性 区 的 应 力 与 极 限 荷 载 。 可 以 证 明 滑 移 线 场 中 也 只 是 在 极 限 状 态 下 引 入 平 衡 方 程 与 屈 服 方<br />

程 , 而 与 本 构 关 系 无 关 , 它 不 受 变 形 参 数 、 弹 模 和 泊 松 比 的 影 响 。 虽 然 滑 移 线 场 与 采 用 的 塑 性 理 论<br />

或 本 构 有 关 , 但 求 解 中 并 不 引 入 本 构 关 系 , 因 而 计 算 结 果 仍 然 与 本 构 无 关 。 无 论 采 用 传 统 塑 性 的 滑<br />

移 线 场 还 是 广 义 塑 性 的 滑 移 线 场 , 求 解 过 程 不 同 , 但 其 最 终 求 得 的 极 限 荷 载 相 同 。<br />

(2)、 极 限 平 衡 法<br />

极 限 平 衡 法 是 岩 土 力 学 中 的 一 种 简 单 极 限 分 析 法 , 它 假 设 材 料 为 刚 性 体 或 刚 塑 性 体 , 采 用 隔 离<br />

体 方 法 , 并 假 定 隔 离 体 边 界 达 到 极 限 平 衡 状 态 , 然 后 利 用 平 衡 和 边 界 条 件 求 出 极 限 荷 载 , 或 求 出 抗<br />

滑 力 与 滑 动 力 之 比 , 从 而 达 到 稳 定 安 全 系 数 。 这 类 方 法 没 有 考 虑 本 构 关 系 与 机 动 条 件 , 得 不 出 应 力 ,<br />

应 变 与 位 移 速 度 , 只 能 给 出 极 限 荷 载 的 近 似 解 或 者 相 应 的 安 全 系 数 , 这 种 方 法 显 然 与 本 构 关 系 无 关 。<br />

(3)、 极 限 分 析 上 、 下 限 法<br />

极 限 分 析 法 是 将 岩 土 体 视 为 理 想 刚 塑 生 体 , 在 极 限 上 、 下 定 理 基 础 上 建 立 起 来 的 分 析 方 法 。 利<br />

用 连 续 介 质 中 的 虚 功 原 理 可 证 明 两 个 极 限 分 析 定 理 即 下 限 定 理 与 上 限 定 理 。 极 限 分 析 法 是 通 过 一 组<br />

极 限 定 理 即 上 限 定 理 或 下 限 定 理 , 推 求 极 限 荷 载 的 上 限 (pu+) 或 下 限 (pu-)。 上 限 解 满 足 机 动 条 件<br />

( 即 满 足 速 度 方 程 ) 与 屈 服 条 件 , 应 力 场 服 从 机 动 条 件 或 塑 性 功 率 不 为 负 的 条 件 , 这 样 就 可 有 虚 功<br />

方 程 , 求 出 极 限 荷 载 。 虚 功 方 程 是 外 功 率 与 能 量 耗 散 率 形 式 的 平 衡 方 程 , 可 见 , 上 限 法 求 解 不 只 是<br />

引 用 平 衡 方 程 与 屈 服 条 件 , 还 需 要 有 本 构 关 系 , 但 它 与 滑 移 线 场 一 样 , 求 解 中 只 要 求 满 足 本 构 关 系<br />

-32-


的 结 果 , 即 要 求 滑 面 与 位 移 矢 量 形 成 的 角 度 满 足 传 统 塑 性 或 广 义 塑 性 , 但 推 导 中 不 需 引 入 本 构 公 式 ,<br />

因 而 它 仍 属 于 极 限 分 析 方 法 。 下 限 解 只 要 求 满 足 平 衡 条 件 和 不 违 背 屈 服 条 件 , 也 与 本 构 关 系 无 关 。<br />

下 限 法 要 构 筑 一 个 合 适 的 静 力 许 可 的 应 力 分 布 来 求 得 下 限 解 , 由 于 很 难 找 到 合 适 的 静 力 许 可 应 力<br />

场 , 应 用 有 限 。<br />

由 上 述 可 见 传 统 的 几 种 极 限 分 析 方 法 求 解 应 力 场 都 不 需 要 引 入 本 构 关 系 , 从 而 使 求 解 简 化 。<br />

[7-10]<br />

五 、 传 统 极 限 分 析 法 的 优 缺 点 与 数 值 极 限 分 析 方 法 的 兴 起<br />

传 统 的 极 限 分 析 法 虽 然 解 法 简 便 , 但 却 求 解 不 易 , 适 用 的 范 围 十 分 有 限 , 一 般 只 能 用 于 均 匀 的<br />

土 体 中 , 获 得 的 经 典 解 答 很 少 。<br />

20 世 纪 下 半 期 , 随 着 数 值 分 析 法 兴 起 , 一 种 做 法 是 在 极 限 分 析 中 引 入 离 散 方 法 , 如 有 限 差 分 滑<br />

移 线 场 法 , 有 限 元 上 、 下 限 法 等 。 另 一 种 做 法 是 用 数 值 方 法 求 解 极 限 问 题 , 出 现 了 有 限 元 超 载 法 与<br />

强 度 折 减 法 , 直 接 采 用 有 限 元 求 解 极 限 荷 载 与 稳 定 安 全 系 数 。 对 于 后 者 作 者 将 其 统 称 为 数 值 ( 可 以<br />

是 有 限 元 法 、 有 限 差 分 法 、 离 散 元 法 等 ) 极 限 分 析 法 , 或 称 有 限 元 极 限 分 析 法 , 其 本 质 是 用 数 值 解<br />

方 法 进 行 极 限 分 析 , 求 解 极 限 荷 载 与 稳 定 安 全 系 数 。 这 种 方 法 不 必 事 先 知 道 滑 面 , 也 不 需 要 求 滑 面<br />

上 的 滑 动 力 与 抗 滑 力 , 可 直 接 获 得 极 限 荷 载 和 稳 定 安 全 系 数 , 还 可 确 定 滑 面 的 位 置 与 形 状 , 极 大 地<br />

扩 大 了 极 限 分 析 法 的 功 能 与 适 用 范 围 。 由 于 该 法 准 确 、 简 便 、 适 用 性 广 , 实 用 性 强 , 尽 管 目 前 还 主<br />

要 用 于 边 坡 稳 定 分 析 中 , 但 其 前 景 十 分 广 阔 。 它 给 予 了 极 限 分 析 第 二 个 春 天 。<br />

[7-13]<br />

六 、 数 值 极 限 分 析 方 法 的 优 势 及 其 存 在 的 问 题<br />

弹 塑 性 数 值 分 析 严 格 地 应 用 了 弹 塑 性 力 学 原 理 与 本 构 关 系 , 其 求 解 精 度 是 较 高 的 , 适 用 的 范 围<br />

是 很 广 的 , 但 数 值 分 析 不 能 获 得 岩 土 的 破 坏 状 态 与 破 坏 面 , 也 无 法 求 出 极 限 荷 载 与 稳 定 安 全 系 数 ,<br />

如 果 将 适 应 性 很 广 的 数 值 解 法 与 极 限 分 析 结 合 起 来 。 那 么 就 可 以 简 便 地 获 得 破 坏 状 态 , 也 可 以 求 出<br />

极 限 荷 载 与 稳 定 安 全 系 数 , 可 见 , 数 值 极 限 分 析 方 法 对 岩 土 工 程 的 设 计 有 很 大 的 优 越 性 , 这 种 方 法<br />

由 此 应 运 而 生 。<br />

数 值 极 限 方 法 与 传 统 极 限 分 析 方 法 相 比 , 其 优 势 首 先 是 减 少 了 求 解 的 条 件 , 增 加 了 求 解 的 功 能 。<br />

传 统 方 法 要 求 先 知 道 滑 动 面 , 而 数 值 极 限 方 法 不 必 事 先 知 道 滑 面 , 反 而 可 以 自 动 找 出 滑 动 面 , 获 得<br />

破 坏 状 态 。<br />

岩 土 工 程 中 往 往 是 场 破 坏 , 在 三 维 计 算 中 表 现 为 面 破 坏 , 二 维 计 算 中 表 现 为 线 破 坏 。 岩 土 体 内<br />

一 部 分 岩 土 体 处 于 弹 性 稳 定 状 态 , 另 一 部 分 岩 体 处 于 塑 性 屈 服 状 态 , 即 岩 土 体 内 同 时 存 在 一 系 列 稳<br />

定 面 与 屈 服 面 , 而 使 岩 土 发 生 整 体 破 坏 失 稳 的 破 坏 面 ( 滑 动 面 ) 只 是 其 中 某 一 个 屈 服 面 。 所 以 实 际<br />

工 程 中 , 岩 土 体 的 真 正 整 体 破 坏 面 并 不 是 指 所 有 的 屈 服 状 态 的 曲 面 , 而 是 指 其 中 一 个 最 危 险 的 屈 服<br />

曲 面 ( 如 安 全 系 数 最 小 的 屈 服 面 )。 但 要 寻 找 这 一 真 正 的 整 体 破 坏 面 并 不 容 易 , 如 果 找 出 来 了 , 求<br />

解 问 题 就 可 以 得 到 解 决 。 正 是 因 为 寻 找 真 实 破 坏 面 的 困 难 , 因 而 传 统 极 限 分 析 很 难 求 解 , 其 应 用 范<br />

围 严 重 受 制 。 而 数 值 极 限 分 析 方 法 , 求 解 过 程 与 传 统 方 法 不 同 , 通 过 不 断 地 降 低 材 料 强 度 或 增 大 荷<br />

载 , 使 其 在 数 值 计 算 中 最 终 达 到 极 限 破 坏 状 态 , 并 自 动 生 成 滑 动 面 , 而 达 到 破 坏 时 的 强 度 折 减 系 数<br />

即 为 稳 定 安 全 系 数 ; 达 到 破 坏 时 的 荷 载 就 是 极 限 荷 载 。<br />

数 值 极 限 方 法 与 传 统 极 限 分 析 方 法 相 比 , 第 二 个 优 势 是 大 幅 扩 大 了 极 限 分 析 法 的 适 用 性 。 例 如<br />

地 基 承 载 力 的 计 算 问 题 : 对 层 状 土 与 非 均 质 土 地 基 、 加 筋 土 地 基 的 承 载 力 等 情 况 , 传 统 方 法 无 法 计<br />

算 , 目 前 尚 没 有 工 程 界 满 意 的 计 算 方 法 。 而 采 用 数 值 极 限 分 析 法 , 就 可 顺 利 求 得 。 其 次 , 传 统 极 限<br />

分 析 法 无 法 求 出 岩 土 体 的 位 移 与 塑 性 区 , 以 及 真 正 的 滑 动 面 , 而 数 值 极 限 分 析 法 可 以 求 出 岩 土 体 内<br />

各 点 的 应 力 、 位 移 、 塑 性 区 与 滑 动 面 。 再 次 , 在 有 支 护 的 情 况 下 , 如 有 抗 滑 桩 情 况 下 , 传 统 方 法 无<br />

法 求 得 有 桩 情 况 下 的 边 坡 安 全 系 数 及 桩 上 的 推 力 , 而 数 值 极 限 方 法 不 受 这 种 条 件 的 限 制 。 应 当 注 意 ,<br />

有 桩 情 况 下 , 求 得 的 桩 上 推 力 不 一 定 全 是 极 限 状 态 下 的 推 力 , 即 主 动 土 压 力 , 而 是 桩 土 共 同 作 用 下<br />

的 推 力 , 这 一 推 力 更 符 合 实 际 受 力 情 况 。 这 是 因 为 达 到 极 限 状 态 需 要 有 一 定 的 位 移 , 如 果 位 移 不 足<br />

达 不 到 极 限 状 态 , 桩 上 推 力 就 会 增 大 。<br />

最 后 , 采 用 数 值 极 限 方 法 还 能 考 虑 岩 土 工 程 开 挖 、 支 护 的 施 工 过 程 , 以 及 岩 土 地 应 力 的 释 放 过<br />

程 等 , 而 传 统 极 限 分 析 方 法 很 难 做 到 这 点 。<br />

然 而 , 数 值 极 限 分 析 法 需 要 找 出 计 算 中 岩 土 体 发 生 破 坏 的 有 效 判 据 , 如 果 找 不 到 这 种 判 据 , 即<br />

使 岩 土 体 已 经 发 生 破 坏 , 而 求 解 者 并 不 知 道 。 或 者 由 于 种 种 原 因 不 能 顺 利 求 解 , 导 致 岩 土 体 不 能 达<br />

到 破 坏 状 态 。 例 如 网 格 剖 分 不 合 理 , 可 能 导 致 计 算 不 收 敛 , 尤 其 是 强 度 折 减 后 网 格 变 形 很 大 , 使 求<br />

解 更 为 困 难 , 类 似 求 解 中 的 各 种 问 题 还 需 要 通 过 计 算 实 践 加 以 解 决 。<br />

-33-


采 用 数 值 极 限 分 析 法 一 个 关 键 问 题 是 如 何 根 据 数 值 计 算 的 结 果 来 判 别 岩 土 体 是 否 达 到 极 限 破<br />

坏 状 态 。 目 前 静 力 状 态 下 采 用 如 下 三 个 判 据 :<br />

1 以 塑 性 应 变 从 边 坡 坡 脚 到 坡 顶 是 否 贯 通 作 为 判 据 , 即 以 塑 性 区 从 内 部 贯 通 至 地 面 或 临 空 面 作<br />

为 破 坏 的 判 据 。 但 塑 性 区 贯 通 只 意 味 着 达 到 屈 服 状 态 , 而 不 一 定 是 土 体 整 体 破 坏 状 态 , 可 见 塑 性 区<br />

贯 通 只 是 破 坏 的 必 要 条 件 , 而 不 是 充 分 条 件 ;<br />

2 在 数 值 计 算 过 程 中 , 边 坡 失 稳 与 数 值 计 算 不 收 敛 同 时 发 生 , 目 前 国 际 通 用 软 件 中 , 一 般 都 以<br />

数 值 计 算 , 即 位 移 或 力 计 算 不 收 敛 作 为 边 坡 失 稳 的 判 断 依 据 ;<br />

3 土 体 破 坏 标 志 着 土 体 滑 移 面 上 应 变 和 位 移 发 生 突 变 , 同 时 安 全 系 数 ( 强 度 折 减 系 数 或 超 载 系<br />

数 ) 与 位 移 的 关 系 曲 线 也 会 发 生 突 变 , 因 此 也 可 用 来 作 为 破 坏 的 判 据 。<br />

然 而 , 上 述 标 准 具 体 应 用 时 , 有 时 也 会 出 现 不 能 应 用 , 或 不 易 判 断 的 情 况 , 例 如 , 位 移 - 时 间 曲<br />

线 或 安 全 系 数 - 位 移 曲 线 , 有 些 突 变 很 明 显 , 有 些 突 变 不 明 显 。 对 于 动 力 荷 载 如 何 判 断 极 限 破 坏 状 态<br />

目 前 还 正 在 研 究 中 , 上 述 准 则 有 些 可 以 引 用 , 有 些 不 能 引 用 。 因 而 还 需 要 依 据 实 际 问 题 提 出 合 理 可<br />

靠 的 判 据 , 这 也 是 数 值 极 限 分 析 法 目 前 需 要 进 一 步 研 究 的 问 题 。<br />

[9-15]<br />

七 、 数 值 极 限 分 析 法 在 各 类 岩 土 工 程 中 的 应 用<br />

当 前 国 内 外 应 用 数 值 极 限 分 析 法 , 主 要 在 边 ( 滑 ) 坡 工 程 的 稳 定 分 析 , 地 基 工 程 中 也 有 少 量 的<br />

应 用 , 其 实 作 为 一 种 力 学 方 法 和 设 计 手 段 , 只 要 是 岩 土 稳 定 与 强 度 问 题 , 各 类 岩 土 工 程 中 都 可 应 用 。<br />

现 代 的 岩 土 工 程 力 学 状 态 更 为 复 杂 , 有 二 维 问 题 与 三 维 问 题 , 有 固 体 力 学 和 渗 流 力 学 问 题 ; 工 程 类<br />

型 十 分 复 杂 , 除 边 ( 滑 ) 坡 工 程 、 地 基 工 程 外 , 尚 有 隧 道 工 程 、 岩 土 环 境 工 程 等 ; 除 工 程 设 计 施 工<br />

外 , 尚 有 岩 土 监 测 、 检 测 、 现 场 试 验 、 工 程 施 工 与 自 然 滑 坡 的 预 警 预 报 等 。 这 些 项 目 都 可 运 用 数 值<br />

极 限 分 析 方 法 , 有 些 已 在 实 际 工 程 中 应 用 。 下 面 举 几 个 应 用 的 例 子 :<br />

[4、14]<br />

7.1. 在 边 坡 埋 入 式 抗 滑 桩 中 确 定 合 理 桩 长 的 应 用<br />

目 前 抗 滑 桩 的 设 计 只 计 算 桩 截 面 尺 寸 , 而 未 规 定 桩 的 长 度 设 计 , 因 而 设 计 中 一 直 沿 用 桩 顶 伸 到<br />

地 面 , 既 不 能 保 证 桩 不 出 现 “ 越 顶 ” 破 坏 , 又 会 使 桩 长 过 长 , 造 成 浪 费 。 应 用 有 限 元 强 度 折 减 法 ,<br />

可 以 确 定 合 理 的 桩 长 , 达 到 安 全 、 经 济 的 目 的 。 图 1 与 表 1 表 明 随 着 桩 长 的 增 加 , 由 于 桩 的 阻 挡 使<br />

滑 面 提 高 , 地 层 稳 定 安 全 系 数 也 随 之 增 加 。 桩 长 设 计 的 原 则 是 必 须 保 证 在 任 何 桩 长 情 况 下 都 要 使 地<br />

层 的 稳 定 系 数 大 于 设 计 安 全 系 数 , 如 果 达 不 到 安 全 系 数 , 桩 就 可 能 出 现 “ 越 顶 ” 破 坏 , 因 而 可 按 此<br />

原 则 确 定 合 理 桩 长 。 如 目 前 设 计 安 全 系 数 规 定 为 1.15, 则 22.5m 桩 长 可 减 为 9m 仍 然 满 足 设 计 要 求 ,<br />

而 且 桩 上 的 推 力 还 会 减 少 , 经 济 效 益 巨 大 。 已 在 重 庆 市 4 个 中 大 型 滑 坡 中 应 用 , 效 果 明 显 。<br />

表 1 桩 长 与 边 坡 安 全 系 数 之 间 的 关 系<br />

桩 0 7 9 1 1 1 1 1 2 2 2<br />

安 1 1 1 1. 1. 1. 1. 1. 1. 1. 1.<br />

图 1 桩 长 变 化 时 的 滑 动 面 位 置<br />

-34-


表 2 Nc、N q 、N γ 的 有 限 元 解 及 比 较<br />

系 数 ϕ (°) 0 5 10 15 20 25 30<br />

Nc<br />

N q<br />

Prandtl 5.14 6.49 8.34 10.98 14.84 20.72 30.14<br />

FEM( 光 滑 ) 5.22 6.60 8.50 11.19 15.18 21.21 31.00<br />

Reissner 1.00 1.56 2.47 3.93 6.38 10.62 18.32<br />

FEM( 光 滑 ) 1.01 1.60 2.51 4.01 6.63 11.03 18.92<br />

Terzaghi 0.00 0.09 0.46 1.41 3.52 8.07 17.99<br />

Meyerhof 0.00 0.07 0.37 1.12 2.86 6.73 15.58<br />

N γ<br />

Vesic 0.00 0.45 1.22 2.64 5.37 10.83 22.29<br />

W.F.Chen 0.00 0.46 1.31 2.93 6.17 12.90 27.51<br />

FEM( 光 滑 ) 0.00 0.21 0.83 1.79 3.87 10.55 18.35<br />

[4、9]<br />

7.2. 在 地 基 工 程 中 确 定 均 匀 地 基 承 载 力 的 应 用<br />

图 2、 图 3 中 , 对 Prandtl 解 与 有 限 元 解 的 滑 移 线 形 状 进 行 了 计 算 对 比 , 表 明 两 者 十 分 接 近 , 而 有<br />

限 元 法 是 自 动 求 出 滑 移 线 , 不 需 做 假 定 。<br />

q<br />

A<br />

B<br />

x<br />

图 2 半 平 面 无 限 体 的 Prandtl 解 的 滑 移 线<br />

图 3 有 限 元 计 算 滑 移 线<br />

依 据 传 统 极 限 平 衡 法 , 均 匀 地 基 极 限 承 载 力 可 近 似 的 表 示 为 :<br />

1<br />

Pu<br />

= cNc<br />

+ qN<br />

q<br />

+ BγN γ<br />

2<br />

下 面 应 用 有 限 元 增 量 加 载 法 对 极 限 承 载 力 系 数 进 行 求 解 , 并 与 传 统 极 限 分 析 法 进 行 比 较 , 列 于<br />

表 2。 N 、 N 两 种 方 法 ( 理 论 解 法 与 数 值 解 法 ) 计 算 结 果 十 分 相 近 , N γ<br />

比 其 它 经 验 解 更 为 准 确 。<br />

c<br />

q<br />

[15]<br />

7.3. 在 隧 洞 工 程 破 坏 机 理 中 的 应 用<br />

(1) 深 埋 隧 洞 破 坏 机 理 分 析<br />

为 了 观 察 隧 洞 破 坏 过 程 与 破 坏 机 理 , 首 先 采 用 混 合 材 料 进 行 模 型 试 验 , 隧 洞 的 跨 度 为 8cm, 侧<br />

墙 高 度 为 8cm, 拱 高 为 3cm, 矢 跨 比 为 1/2, 隧 洞 左 右 边 界 与 隧 洞 左 右 侧 墙 的 距 离 为 16cm, 上 侧 边 界<br />

距 离 隧 洞 拱 顶 为 24cm, 下 侧 边 界 距 离 隧 洞 底 部 为 16cm。 试 验 从 0 开 始 逐 级 加 载 直 至 隧 洞 发 生 破 坏 ,<br />

并 将 试 验 观 察 到 的 结 果 与 数 值 模 拟 的 结 果 进 行 对 比 。<br />

C<br />

(a) 模 型 试 验 破 裂 面 (b) 数 值 模 拟 破 裂 面 (a) 模 型 试 验 破 裂 面 (b) 数 值 模 拟 破 裂 面<br />

图 4 深 埋 隧 洞 破 坏 情 况<br />

图 5 浅 埋 隧 洞 破 坏 情 况<br />

图 4 中 , 列 出 了 模 型 试 验 与 数 值 模 拟 结 果 , 模 型 试 验 极 限 荷 载 为 59kN, 数 值 模 拟 极 限 荷 载 55kN。<br />

破 坏 区 大 小 也 很 接 近 , 数 值 模 拟 的 破 裂 面 位 置 是 依 据 破 裂 面 上 有 应 变 突 变 而 画 出 的 。<br />

(2) 浅 埋 隧 洞 破 坏 机 理 研 究<br />

-35-


为 了 推 测 浅 埋 隧 洞 破 坏 机 理 , 进 行 了 浅 埋 隧 洞 破 坏 模 型 试 验 , 并 将 其 与 数 值 模 拟 结 果 进 行 比 较 。<br />

浅 埋 隧 洞 洞 跨 8cm, 洞 高 12cm, 洞 深 15cm, 埋 深 4cm。 图 5 为 模 型 试 验 与 数 值 模 拟 结 果 图 。 模 型<br />

试 验 极 限 荷 载 为 28kN, 数 值 模 拟 极 限 荷 载 26kN, 两 者 破 裂 的 位 置 与 形 状 也 是 相 近 的 。<br />

由 于 传 统 极 限 分 析 方 法 至 今 没 有 求 出 隧 洞 的 破 裂 面 位 置 与 形 状 , 当 然 也 无 法 确 定 隧 洞 的 稳 定 安<br />

全 系 数 或 极 限 荷 载 。 而 采 用 有 限 元 强 度 折 减 法 能 自 动 求 出 破 坏 面 的 位 置 与 稳 定 安 全 系 数 或 极 限 荷<br />

载 , 首 次 给 隧 洞 围 岩 稳 定 性 提 供 了 定 量 判 据 , 从 而 有 可 能 进 行 合 理 的 隧 道 设 计 。<br />

(3) 不 同 埋 深 下 隧 洞 的 破 坏 形 态 与 深 浅 埋 的 分 界 标 准<br />

为 了 研 究 隧 洞 由 浅 埋 破 坏 逐 渐 转 向 深 埋 破 坏 的 过 程 , 我 们 采 用 有 限 元 强 度 折 减 法 , 对 一 个 洞 跨<br />

12m, 高 5m 的 矩 形 隧 洞 与 一 个 洞 跨 12m, 高 5m, 拱 高 3m 的 直 墙 拱 形 隧 洞 进 行 分 析 研 究 [21] , 图 6<br />

列 出 了 不 同 埋 深 下 矩 形 隧 洞 的 破 坏 情 况 及 其 安 全 系 数 。<br />

(a) 埋 深 3 米 , 安 全 系 数 0.52 (b) 埋 深 9 米 , 安 全 系 数 0.66 (c) 埋 深 10 米 安 全 系 数 0.69<br />

(d) 埋 深 15 米 安 全 系 数 0. 7 (e) 埋 深 18 米 安 全 系 数 0.7 (f) 埋 深 30 米 安 全 系 数 0.67<br />

图 6 矩 形 洞 室 的 等 效 塑 性 应 变 图<br />

由 图 6(a) 可 见 , 当 埋 深 3 米 时 , 最 大 的 塑 性 应 变 在 拱 肩 处 , 破 裂 面 自 拱 肩 处 起 呈 拱 形 直 至 地 表 ,<br />

但 拱 未 合 拢 , 安 全 系 数 为 0.52。 由 图 6(b) 可 见 , 当 埋 深 9m 时 , 形 成 了 明 显 的 浅 埋 压 力 拱 , 安 全 系<br />

数 为 0.66。 浅 埋 压 力 拱 的 形 成 与 埋 深 有 关 , 它 是 浅 埋 与 深 埋 的 分 界 线 。 由 图 6(c) 可 见 , 当 埋 深 10m<br />

时 , 拱 顶 上 方 浅 埋 压 力 拱 逐 渐 消 失 , 与 此 同 时 形 成 了 深 埋 压 力 拱 , 即 普 氏 压 力 拱 , 安 全 系 数 为 0.69。<br />

可 见 , 埋 深 10 米 时 出 现 了 突 变 , 由 浅 埋 转 为 深 埋 。 由 图 6(d、e) 可 见 , 当 埋 深 15m、18m 时 逐 渐 形<br />

成 两 条 破 裂 面 : 一 条 是 拱 顶 上 已 形 成 的 普 氏 压 力 拱 , 另 一 条 是 在 侧 面 逐 渐 形 成 的 破 裂 面 , 破 裂 面 自<br />

拱 角 至 墙 脚 , 安 全 系 数 均 为 0.7。 可 见 , 在 埋 深 10~18m 时 , 安 全 系 数 基 本 不 变 , 表 明 普 氏 压 力 拱 与<br />

埋 深 无 关 。 由 图 6 (f) 可 见 , 当 埋 深 30m 时 , 虽 然 普 氏 压 力 拱 仍 然 存 在 , 但 侧 壁 破 裂 面 明 显 先 破 坏 ,<br />

安 全 系 数 随 深 度 增 加 降 为 0.67。 由 上 反 映 了 随 深 度 增 加 隧 洞 破 坏 机 理 的 变 化 与 安 全 系 数 的 变 化 。 拱<br />

形 隧 洞 破 坏 规 律 与 矩 形 相 同 , 只 是 它 不 存 在 普 氏 压 力 拱 。<br />

参 考 文 献 :<br />

[1] 希 尔 R. 著 , 王 仁 等 译 . 塑 性 数 学 理 论 . 北 京 : 科 学 出 版 社 , 1966.<br />

[2] Chen W.F. Limit Analysis and Soil Plasticity. Elsevier Scientific Publishing Company, 1975.<br />

[3] 黄 传 志 . 土 体 极 限 分 析 理 论 与 应 用 . 北 京 : 人 民 交 通 出 版 社 , 2007.<br />

[4] 郑 颖 人 , 孔 亮 . 岩 土 塑 性 力 学 . 北 京 : 中 国 建 筑 工 业 出 版 社 , 2010.<br />

[5] 张 学 言 , 闫 澍 旺 . 岩 土 塑 性 力 学 基 础 ( 第 2 版 ). 天 津 : 天 津 大 学 出 版 社 , 2004.<br />

[6] 钱 家 欢 , 殷 宗 泽 . 土 工 原 理 与 计 算 ( 第 一 版 ). 北 京 : 水 利 电 力 出 版 社 , 1994.<br />

[7] 郑 颖 人 , 赵 尚 毅 . 岩 土 工 程 极 限 分 析 有 限 元 法 及 其 应 用 . 土 木 工 程 学 报 , 2005, 第 1 期 .<br />

[8] 郑 颖 人 , 赵 尚 毅 , 孔 位 学 , 邓 楚 键 . 岩 土 工 程 极 限 分 析 有 限 元 法 . 岩 土 力 学 , 2005, 26(1): 163-168.<br />

[9] 郑 颖 人 , 赵 尚 毅 , 邓 楚 建 . 有 限 元 极 限 分 析 法 及 其 在 岩 土 工 程 中 的 应 用 研 究 . 中 国 工 程 科 学 , 2006, 8(12): 39-61.<br />

[10] 郑 颖 人 , 赵 尚 毅 . 有 限 元 强 度 折 减 法 在 土 坡 与 岩 坡 中 的 应 用 . 岩 石 力 学 与 工 程 学 报 , 2004, 第 19 期 .<br />

[11] 赵 尚 毅 , 郑 颖 人 等 . 用 有 限 元 强 度 折 减 法 求 边 坡 稳 定 安 全 系 数 . 岩 土 工 程 学 报 , 2002, 第 3 期 .<br />

[12] 郑 颖 人 , 陈 祖 煜 , 王 恭 先 . 边 坡 与 滑 坡 工 程 治 理 . 北 京 : 人 民 交 通 出 版 社 , 2010.<br />

[13] 赵 尚 毅 , 郑 颖 人 , 张 玉 芳 . 有 限 元 强 度 折 减 法 中 边 坡 失 稳 的 判 据 探 讨 . 岩 土 力 学 , 2005, 26(2): 332-336.<br />

[14] 郑 颖 人 , 赵 尚 毅 , 雷 文 杰 等 , 基 于 有 限 元 强 度 折 减 法 的 抗 滑 桩 设 计 新 方 法 . 《 岩 土 工 程 学 报 》 创 刊 30 周 年 纪 念 文<br />

集 , 河 海 大 学 出 版 社 , 2009(1): 98-110.<br />

[15] 郑 颖 人 , 徐 浩 , 王 成 . 隧 洞 破 坏 机 理 及 深 浅 埋 分 界 标 准 . 浙 江 大 学 学 报 , 2010, 1: 1851-1856.<br />

-36-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

世 界 第 一 座 文 化 遗 址 类 水 下 博 物 馆 — 白 鹤 梁 古 水 文 题 刻 原 址 水<br />

下 保 护 工 程<br />

1,2,3<br />

葛 修 润<br />

( 2<br />

( 3<br />

( 1 上 海 交 通 大 学 上 海 200030)<br />

中 国 科 学 院 武 汉 岩 土 力 学 研 究 所 武 汉 430071)<br />

岩 土 力 学 与 工 程 国 家 重 点 实 验 室 武 汉 430071)<br />

摘 要 : 白 鹤 梁 乃 一 座 天 然 石 梁 , 位 于 长 江 涪 陵 段 靠 近 南 岸 的 深 水 航 道 旁 , 长 约 1600 米 。 每 年<br />

12 月 到 次 年 3 月 长 江 枯 水 季 节 时 梁 顶 才 能 露 出 江 面 。 三 峡 工 程 建 成 和 水 库 蓄 水 后 , 它 将 永 远 位 于 三<br />

峡 水 库 库 底 , 再 无 “ 出 头 ” 之 日 了 。<br />

古 人 认 为 , 冬 天 长 江 的 水 位 回 落 到 很 枯 位 置 时 , 第 二 年 一 定 是 个 风 调 雨 顺 的 丰 收 年 。 老 祖 宗 就<br />

用 刻 石 鱼 方 法 记 录 长 江 的 枯 水 水 位 , 这 一 做 法 从 唐 朝 延 续 至 今 。 每 隔 三 、 五 年 当 水 位 很 枯 , 石 鱼 露<br />

出 时 就 成 为 一 件 盛 事 。 人 们 在 长 江 上 聚 会 , 在 白 鹤 梁 上 刻 石 记 载 , 文 人 墨 客 也 赋 诗 填 词 。 岁 月 相 积 ,<br />

白 鹤 梁 上 刻 上 了 十 八 尾 石 鱼 , 有 题 刻 一 百 六 十 五 段 , 计 三 万 余 字 , 记 载 了 自 唐 朝 广 德 元 年 至 今 一 千<br />

二 百 余 年 间 的 长 江 七 十 二 个 极 枯 水 位 年 份 。 白 鹤 梁 题 刻 是 世 界 江 河 水 文 记 录 最 早 之 地 。 被 联 合 国 教<br />

科 文 组 织 誉 为 “ 保 存 完 好 的 世 界 唯 一 古 代 水 文 站 ”。1988 年 国 务 院 将 白 鹤 梁 列 为 国 家 级 文 物 保 护 单<br />

位 。<br />

白 鹤 梁 题 刻 在 科 学 、 历 史 和 艺 术 等 方 面 都 具 有 极 高 价 值 。 白 鹤 梁 一 定 要 保 护 , 这 样 的 共 识 不 难<br />

达 成 , 但 困 难 的 是 如 何 保 护 。 十 八 年 前 国 家 有 关 部 门 就 开 始 组 织 专 家 对 白 鹤 梁 文 化 遗 址 的 保 护 问 题<br />

进 行 研 究 。 历 经 “ 水 晶 宫 ” 方 案 等 等 直 到 上 世 纪 末 形 成 的 一 种 方 案 : 即 用 1:1 比 尺 复 制 白 鹤 梁 题 刻 陈<br />

列 于 库 岸 边 , 建 博 物 馆 而 原 文 物 将 随 三 峡 水 库 库 底 泥 沙 的 淤 积 而 埋 于 近 二 十 米 厚 的 淤 沙 之 中 。 二 ○○<br />

一 年 二 月 于 涪 陵 召 开 的 会 议 原 计 划 是 最 后 一 次 决 策 会 议 了 。 有 鉴 于 三 峡 水 库 不 久 即 将 蓄 水 , 时 间 已<br />

十 分 紧 迫 , 更 苦 于 无 更 好 的 保 护 方 案 , 只 能 无 可 奈 何 地 采 取 前 述 的 “ 异 地 陈 展 ”( 假 文 物 ),“ 就 地 於<br />

埋 ”( 真 国 宝 ) 的 方 案 了 。<br />

本 文 作 者 有 幸 参 加 了 这 次 会 议 。 在 仔 细 研 读 了 各 种 方 案 资 料 后 对 将 要 通 过 的 方 案 发 表 了 异 议 ,<br />

并 提 出 了 以 “ 无 压 容 器 ” 为 基 本 理 念 的 原 址 水 下 保 护 白 鹤 梁 的 创 新 方 案 。 后 经 近 一 年 的 努 力 和 论 证 ,<br />

此 方 案 最 终 被 国 家 主 管 部 门 所 采 纳 。<br />

此 新 方 案 为 保 证 水 下 文 化 遗 产 的 真 实 性 和 完 整 性 , 将 在 原 址 兴 建 水 下 保 护 工 程 保 护 白 鹤 梁 古 水<br />

文 题 刻 。 此 水 下 保 护 工 程 集 成 文 物 、 水 利 、 建 筑 、 市 政 、 航 道 、 潜 艇 、 特 种 设 备 等 多 种 专 业 、 多 学<br />

科 的 技 术 实 现 了 白 鹤 梁 题 刻 的 原 址 原 样 原 环 境 保 护 和 观 赏 。 保 护 工 程 由 水 下 保 护 体 、 交 通 及 参 观 廊<br />

道 、 地 面 陈 列 馆 三 部 分 组 成 。 总 建 筑 面 积 8433mm 2 , 工 程 总 投 资 1.9 亿 元 人 民 币 。 水 下 保 护 工 程 于<br />

2003 年 2 月 13 日 开 工 建 设 。2009 年 5 月 18 日 在 世 界 博 物 馆 日 建 成 开 馆 , 成 为 世 界 上 唯 一 在 水 深<br />

40 余 米 处 建 立 的 遗 址 类 水 下 博 物 馆 , 她 为 水 下 文 化 遗 产 的 原 址 保 护 提 供 了 成 功 的 工 程 范 例 , 也 为 我<br />

国 伟 大 的 长 江 三 峡 工 程 增 添 了 光 彩 。 三 峡 水 库 建 成 后 白 鹤 梁 仍 将 留 存 于 人 间 , 供 人 类 一 代 代 观 赏 下<br />

去 而 获 得 “ 新 生 ”。<br />

一 、 白 鹤 梁 古 水 文 题 刻 概 况 及 所 在 的 地 理 环 境<br />

“ 世 界 第 一 古 代 水 文 站 ”—— 白 鹤 梁 题 刻 位 于 正 在 兴 建 的 长 江 三 峡 水 利 枢 纽 水 库 区 涪 陵 城 北 长 江<br />

之 中 。 从 唐 朝 广 德 元 年 ( 公 元 七 六 三 年 ) 以 来 , 我 国 人 民 用 刻 石 鱼 的 方 式 将 历 年 来 的 枯 水 位 镌 刻 在 白<br />

-37-


鹤 梁 岩 壁 面 上 至 今 已 有 一 千 二 百 多 年 的 历 史 。“ 白 鹤 梁 ” 因 早 年 白 鹤 聚 集 梁 上 而 得 名 。<br />

白 鹤 梁 位 于 重 庆 市 所 属 涪 陵 城 北 江 心 , 距 乌 江 与 长 江 汇 合 处 约 800m, 是 一 道 天 然 石 梁 , 长 约<br />

1600m, 宽 约 25m, 东 西 向 延 伸 与 长 江 平 行 。 白 鹤 梁 梁 脊 标 高 为 138m, 比 长 江 最 高 洪 水 位 低 约 30m。<br />

白 鹤 梁 分 上 、 中 、 下 三 段 , 题 刻 集 中 在 长 约 220m 的 中 段 石 梁 上 , 特 别 是 约 65m 长 的 中 段 东 区 。 白<br />

鹤 梁 的 岩 面 是 较 平 整 的 浅 色 薄 层 砂 岩 , 以 14.5° 的 倾 角 北 向 长 江 主 航 道 , 为 题 刻 提 供 了 良 好 条 件 。<br />

涪 陵 市 地 处 乌 江 入 长 江 口 , 素 为 川 东 重 要 商 埠 , 乌 江 流 域 最 大 的 物 资 交 流 中 心 。 涪 陵 市 居 住 有<br />

汉 、 土 家 、 苗 、 回 、 蒙 古 等 民 族 , 历 史 悠 久 。 长 江 三 峡 库 区 文 物 古 迹 众 多 , 达 两 千 余 处 , 以 白 鹤 梁<br />

题 刻 最 为 有 名 , 也 是 长 江 三 峡 水 库 淹 没 区 内 最 早 的 一 处 全 国 重 点 文 物 保 护 单 位 。 长 江 涪 陵 河 段 河 势<br />

及 白 鹤 梁 位 置 , 请 参 见 图 1。 图 2 记 录 的 是 从 涪 陵 北 区 南 眺 白 鹤 梁 的 照 片 。 白 鹤 梁 题 刻 的 某 一 局 部<br />

情 景 照 片 见 图 3。 图 4 表 示 的 是 三 峡 水 利 枢 纽 — 涪 陵 — 白 鹤 梁 的 地 理 位 置 关 系 图 。<br />

据 不 完 全 统 计 : 白 鹤 梁 古 水 文 题 刻 计 有 文 字 题 刻 165 段 , 三 万 余 字 , 其 中 唐 代 1 段 , 宋 代 98<br />

段 , 元 代 5 段 , 明 代 16 段 , 清 代 24 段 , 民 国 14 段 , 年 代 不 详 者 7 段 。 石 鱼 雕 刻 18 尾 , 其 中 立 体<br />

浮 雕 1 尾 , 浅 浮 雕 2 尾 , 平 面 线 雕 15 尾 。 此 外 , 尚 有 线 雕 白 鹤 1 只 , 观 音 3 尊 。 这 些 题 刻 与 浮 雕<br />

分 布 于 不 同 位 置 、 没 于 冬 季 常 年 枯 水 位 线 以 下 , 只 有 在 水 位 很 枯 的 年 份 的 冬 季 , 江 水 枯 竭 时 才 显 露<br />

水 面 。 据 统 计 , 每 3、5 年 才 能 露 出 一 次 。 我 国 祖 先 刻 石 鱼 作 为 水 位 标 记 , 每 当 江 水 退 石 鱼 现 时 ,<br />

就 预 兆 丰 收 年 景 来 临 , 即 “ 石 鱼 出 水 兆 丰 年 ”。 历 代 的 人 们 将 石 鱼 出 水 的 时 间 , 石 鱼 距 水 位 线 之 间 的<br />

尺 度 , 观 察 者 的 姓 名 , 以 及 石 鱼 显 现 时 的 情 景 用 诗 词 、 题 文 等 形 式 刻 记 在 石 梁 上 。<br />

图 1 涪 陵 河 段 河 势 及 白 鹤 梁 位 置 示 意 图<br />

图 2 白 鹤 梁 紧 靠 涪 陵 河 段 的 长 江 深 水 航 道<br />

图 3 白 鹤 梁 题 刻 ( 局 部 )<br />

图 4 三 峡 水 利 枢 纽 — 涪 陵 — 白 鹤 梁 地 理 位 置 图<br />

-38-


二 、 白 鹤 梁 古 水 文 题 刻 的 科 学 价 值<br />

白 鹤 梁 上 所 刻 石 鱼 , 是 前 人 用 来 记 录 江 水 水 位 最 枯 的 标 志 , 又 称 石 鱼 水 标 , 为 研 究 长 江 水 文 、<br />

区 域 及 全 球 气 候 变 化 的 历 史 规 律 提 供 了 极 好 的 实 物 佐 证 , 具 有 很 好 的 科 学 价 值 。 唐 广 德 元 年 ( 公 元<br />

七 六 三 年 ) 以 前 , 白 鹤 梁 上 刻 有 石 鱼 两 尾 , 现 存 一 尾 , 长 六 十 厘 米 , 并 有 隶 书 “ 石 鱼 ” 二 字 。 该 鱼 刻<br />

究 竟 早 在 广 德 元 年 何 时 有 待 继 续 考 证 。 清 康 熙 二 十 四 年 , 涪 州 牧 肖 星 拱 命 重 刻 双 鲤 石 鱼 来 替 代 唐 鱼 ,<br />

其 下 题 有 “ 重 镌 双 鱼 记 ”( 图 5)。 据 考 证 , 双 鱼 鱼 眼 相 当 于 川 江 航 道 部 门 当 地 水 尺 零 点 , 而 唐 鱼 腹 相<br />

当 于 涪 陵 地 区 现 代 水 文 站 历 年 枯 水 位 的 平 均 值 。<br />

根 据 石 鱼 及 有 关 题 记 , 我 们 的 先 人 记 录 有 一 千 二 百 年 来 的 72 个 枯 水 年 份 的 水 位 , 留 下 极 其 珍<br />

贵 的 水 文 资 料 。 古 代 水 文 站 资 料 表 明 , 这 一 千 二 百 多 年 来 长 江 的 最 枯 水 位 发 生 在 宋 朝 绍 兴 十 年 ( 即<br />

公 元 一 一 四 ○ 年 ), 当 时 是 “ 水 去 鱼 下 十 尺 ”。<br />

上 述 古 水 文 资 料 对 研 究 长 江 流 域 的 综 合 开 发 、 内 河 航 运 、 农 田 灌 溉 、 桥 梁 建 设 、 城 市 供 水 等 有<br />

着 重 要 的 科 学 价 值 。 设 计 葛 洲 坝 电 站 和 三 峡 水 利 枢 纽 工 程 时 都 参 考 了 这 些 水 文 资 料 。1974 年 联 合 国<br />

教 科 文 组 织 在 巴 黎 召 开 的 国 际 水 文 会 议 上 , 我 国 代 表 介 绍 了 白 鹤 梁 题 刻 , 引 起 专 家 学 者 们 的 极 大 兴<br />

趣 。 白 鹤 梁 已 公 认 为 是 世 界 上 目 前 所 发 现 的 时 间 最 早 、 延 续 时 间 最 长 、 数 量 最 多 的 枯 水 位 水 文 题 刻 。<br />

埃 及 尼 罗 河 中 虽 有 类 似 的 水 文 石 刻 题 记 , 但 数 量 及 延 续 时 间 远 逊 于 白 鹤 梁 。<br />

图 5 唐 代 石 鱼 水 标 及 清 代 “ 重 镌 双 鱼 记 ”<br />

三 、 白 鹤 梁 古 水 文 题 刻 的 历 史 价 值 和 艺 术 价 值<br />

自 唐 迄 今 , 历 代 文 人 雅 士 、 官 吏 商 贾 , 过 往 涪 陵 , 值 石 鱼 出 水 , 治 舟 来 白 鹤 梁 上 , 驻 足 流 连 ,<br />

吟 诗 作 赋 , 题 铭 江 心 , 姓 名 可 考 者 300 余 人 , 史 有 传 者 如 黄 庭 坚 、 朱 昂 、 秦 九 韶 、 刘 甲 、 黄 寿 、 王<br />

士 桢 、 公 武 等 人 。 题 记 囊 括 了 各 派 书 法 , 文 字 有 汉 字 、 蒙 文 , 书 法 篆 、 隶 、 行 、 草 、 楷 皆 备 , 风<br />

格 颜 、 柳 、 欧 、 苏 具 全 。 尤 以 宋 代 大 文 学 家 黄 庭 坚 谪 居 涪 州 时 所 书 “ 元 符 庚 辰 涪 翁 来 ” 题 铭 , 寥 寥 数<br />

字 , 永 留 心 态 气 宇 ( 图 6)。 图 7 为 元 至 顺 四 年 ( 公 元 一 三 三 三 年 ) 模 刻 木 鱼 (46cm×18cm), 模 拟<br />

木 刻 技 法 , 奉 议 大 夫 涪 陵 守 张 八 歹 题 有 木 鱼 记 。 图 8 为 清 康 熙 四 十 五 年 董 维 祺 石 鱼 及 题 刻<br />

(140cm×47cm)。 图 9 为 清 嘉 庆 二 十 年 ( 公 元 一 八 一 五 年 ) 张 师 范 高 浮 雕 鱼 , 体 长 280 厘 米 。 图 10<br />

为 孙 海 题 刻 (97 cm×47cm), 是 清 光 绪 七 年 ( 公 元 一 八 八 一 年 ) 所 刻 , 镌 刻 点 划 有 神 、 结 构 端 庄 , 内 含<br />

奔 放 , 气 势 纵 横 。 图 11 是 在 白 鹤 梁 上 镌 刻 的 送 子 观 音 。 图 12 是 白 鹤 时 鸣 图 。<br />

白 鹤 梁 以 其 水 下 碑 文 之 多 、 历 史 之 悠 久 、 水 情 记 录 之 翔 实 、 题 记 内 容 之 丰 富 、 形 式 之 多 姿 多 彩 ,<br />

与 长 江 及 环 境 之 混 成 一 体 , 堪 称 一 大 水 下 奇 观 , 称 为 “ 水 下 碑 林 ” 也 不 为 其 过 。<br />

-39-


图 6 黄 庭 坚 题 铭 “ 元 符 庚 辰 涪 翁 来 ”<br />

图 7 元 至 顺 四 年 (1333 年 ) 模 刻 木 鱼<br />

图 8 董 维 祺 石 鱼 及 题 刻<br />

图 9 清 嘉 庆 二 十 年 (1875 年 ) 张 师 范 高 浮 雕 鱼<br />

图 10 清 光 绪 七 年 (1881 年 ) 孙 海 题 刻<br />

-40-


图 11 送 子 观 音 图<br />

图 12 白 鹤 时 鸣 图<br />

四 、 伟 大 的 母 亲 河 长 江 和 三 峡 水 利 枢 纽 及 白 鹤 梁 古 水 文 题 刻<br />

祖 国 的 母 亲 河 万 里 长 江 源 自 位 于 青 海 省 南 部 唐 古 拉 山 脉 的 主 峰 格 拉 丹 东 大 雪 山 ( 图 13 和 图<br />

14)。 涓 涓 细 流 汇 成 浩 瀚 的 大 江 切 穿 瞿 塘 、 巫 山 和 西 陵 三 大 峡 谷 ( 图 15 和 图 16), 经 过 荆 江 十 八 弯<br />

段 后 一 江 春 水 波 涛 滚 滚 向 东 流 , 直 奔 大 海 , 沿 途 物 产 丰 富 , 风 景 秀 美 ( 图 17)。<br />

图 13 长 江 源 头 — 格 拉 丹 东 大 雪 山<br />

图 14 万 里 长 江 源 头 — 格 拉 丹 东 大 雪 山<br />

图 15 浩 瀚 的 长 江<br />

-41-


图 16 巫 山 神 女<br />

图 17 长 江 美 景 — 小 三 峡<br />

一 九 九 二 年 4 月 3 日 全 国 人 大 七 届 五 次 会 议 通 过 兴 建 长 江 三 峡 水 利 枢 纽 工 程 的 决 议 ( 图 18),<br />

三 峡 大 坝 坝 址 选 择 在 中 堡 岛 ( 图 19), 这 是 一 块 以 花 岗 岩 为 基 底 的 稳 固 岩 基 。 历 时 17 载 终 于 在 二 ○○<br />

九 年 基 本 上 建 成 这 一 世 界 上 最 宏 伟 的 水 利 工 程 , 图 20 是 三 峡 水 利 枢 纽 的 鸟 瞰 图 , 它 包 括 宏 伟 的 溢<br />

洪 坝 段 ( 图 21), 左 右 两 个 坝 后 式 水 力 发 电 厂 房 坝 段 ( 图 22), 共 装 有 26 台 700MW 功 率 的 发 电 机<br />

组 。 三 峡 水 利 枢 纽 的 双 线 五 级 船 闸 能 保 5000 吨 船 队 从 宜 昌 直 达 重 庆 ( 图 23 和 图 24), 三 峡 水 库 可<br />

调 节 洪 水 , 其 蓄 洪 库 容 约 330 亿 立 方 米 , 当 遭 遇 百 年 一 遇 的 洪 水 时 能 保 证 下 游 各 大 城 市 的 安 全 。 三<br />

峡 水 库 长 度 达 600 公 里 淹 没 面 积 见 图 25。 三 峡 水 库 的 回 水 纵 剖 面 见 图 26。 根 据 科 学 实 验 得 知 , 三<br />

峡 工 程 完 工 后 三 十 年 左 右 白 鹤 梁 古 水 文 题 刻 将 葬 身 在 三 峡 水 库 的 淤 泥 之 中 。 由 于 国 家 电 力 供 应 的 急<br />

需 , 三 峡 水 利 枢 纽 在 右 岸 山 体 又 增 加 了 地 下 水 力 发 电 厂 (6×700MW 机 组 ), 目 前 土 建 已 完 工 , 正 在<br />

安 装 机 组 。 在 船 闸 旁 边 也 正 在 建 造 升 船 机 。<br />

图 18 全 国 人 大 七 届 五 次 会 议 通 过 兴 建 三 峡 议 案 (1992 年 4 月 3 日 )<br />

图 19 长 江 三 峡 大 坝 坝 址 — 中 堡 岛<br />

图 20 长 江 三 峡 工 程 鸟 瞰 图<br />

-42-


图 21 溢 流 坝 段 剖 面 图<br />

图 22 坝 后 式 厂 房 剖 面 图<br />

图 23 船 闸 和 升 船 机 剖 面 图<br />

图 24 船 闸 照 片<br />

图 25 三 峡 水 库 淹 没 区 示 意 图<br />

图 26 三 峡 水 库 回 水 示 意 图<br />

-43-


五 、 历 年 来 研 究 过 的 两 种 保 护 白 鹤 梁 古 水 文 题 刻 方 案 简 评 和 最 终 采 纳 方 案 的 提 出<br />

国 家 文 物 局 自 1994 年 以 来 曾 多 次 组 织 专 家 对 白 鹤 梁 保 护 方 案 进 行 深 入 研 究 。 在 许 多 方 案 里 本<br />

文 仅 对 两 个 具 有 代 表 性 方 案 作 简 要 介 绍 和 评 述 。<br />

5.1. 天 津 大 学 “ 水 晶 宫 ” 方 案<br />

天 津 大 学 提 出 的 “ 水 晶 宫 ” 方 案 属 于 最 早 提 出 的 方 案 之 一 。 此 方 案 建 议 建 造 如 图 27 所 示 的 在 水<br />

下 用 钢 筋 混 凝 土 浇 筑 成 的 双 层 拱 壳 来 保 护 白 鹤 梁 古 水 文 题 刻 , 长 度 为 120 米 , 宽 20 米 。 壳 体 内 无<br />

水 , 为 了 防 渗 、 阻 漏 、 保 护 基 岩 , 沿 基 础 进 行 帷 幕 灌 浆 。 参 观 者 经 过 过 江 通 道 进 入 壳 体 , 对 白 鹤 梁<br />

直 接 观 赏 。<br />

水 晶 宫 方 案 主 要 弊 病 可 归 纳 为 如 下 四 点 :<br />

(1) 内 外 压 差 大 , 由 于 损 伤 导 致 压 毁 的 可 能 性 很 大 。<br />

这 样 大 的 壳 体 外 部 将 长 期 承 受 40 余 米 水 头 的 高 压 , 内 部 压 力 为 大 气 压 力 , 实 际 上 是 一 种 “ 压 力<br />

容 器 ”。 壳 体 尺 度 大 , 作 用 荷 载 大 , 内 部 是 “ 大 通 仓 ”, 壳 体 在 建 造 时 在 某 处 若 有 损 伤 将 导 致 整 仓 溃 坏 ,<br />

无 可 挽 救 , 人 员 将 无 可 幸 免 。 在 运 行 期 间 如 遇 到 船 队 对 壳 体 碰 撞 或 有 重 物 坠 落 到 壳 体 上 可 导 致 壳 体<br />

破 裂 , 仓 内 人 员 也 无 法 逃 脱 。<br />

(2) 注 浆 帷 幕 施 工 过 程 和 建 成 后 的 渗 流 场 极 有 可 能 损 毁 白 鹤 梁 题 刻 。<br />

帷 幕 施 工 过 程 的 高 压 注 浆 必 将 危 及 白 鹤 梁 题 刻 安 全 。 白 鹤 梁 题 刻 是 镌 刻 在 薄 砂 岩 上 的 , 即 使 建<br />

成 帷 幕 , 由 于 巨 大 内 外 的 压 差 , 总 是 会 有 地 下 水 渗 流 场 , 地 下 水 会 从 层 状 岩 体 的 层 间 出 露 渗 漏 , 导<br />

致 白 鹤 梁 被 毁 的 可 能 性 极 大 。<br />

(3) 改 变 了 白 鹤 梁 古 水 文 题 刻 的 生 存 条 件 国 宝 必 将 毁 于 一 旦 。<br />

白 鹤 梁 古 水 文 题 刻 在 一 千 多 年 来 能 保 存 完 好 主 要 原 因 是 它 处 在 长 江 水 保 护 的 环 境 中 , 很 少 暴 露<br />

在 空 气 中 。 水 晶 宫 方 案 如 实 现 的 话 由 于 这 些 镌 刻 长 期 暴 露 在 空 气 中 必 将 风 化 而 且 很 快 都 会 损 毁 。<br />

(4)“ 水 晶 宫 ” 方 案 建 造 价 格 昂 贵 , 施 工 难 度 极 大 , 施 工 期 很 长 , 对 主 航 道 的 航 运 影 响 严 重 。<br />

由 于 上 述 诸 点 , 经 过 详 细 研 究 和 多 次 会 议 讨 论 ,“ 水 晶 宫 ” 方 案 在 1998 年 被 彻 底 否 定 , 但 随 之<br />

带 来 了 严 重 的 后 遗 症 : 似 乎 修 建 白 鹤 梁 水 下 保 护 工 程 之 路 是 一 条 “ 绝 路 ”。<br />

图 27 水 晶 宫 方 案 示 意 图<br />

5.2. “ 就 地 保 护 , 异 地 陈 展 ” 方 案<br />

这 类 方 案 的 所 谓 “ 就 地 保 护 ” 实 质 就 是 “ 就 地 淤 埋 ”。 这 种 方 案 认 为 , 在 目 前 的 施 工 技 术 水 平 与 经<br />

济 条 件 下 只 宜 采 取 水 下 泥 沙 淹 没 自 然 保 护 ( 预 先 对 水 文 题 刻 采 用 局 部 加 固 和 保 护 措 施 ), 以 便 将 来<br />

-44-


在 经 济 与 施 工 技 术 等 条 件 有 了 长 足 的 进 步 , 且 具 有 开 发 价 值 时 , 也 就 是 说 再 过 一 二 百 年 后 再 挖 掘 出<br />

来 使 它 的 原 始 风 貌 重 现 于 世 人 面 前 。 方 案 的 另 一 部 分 是 在 防 汛 堤 消 落 区 某 一 高 程 上 ( 图 28) 用 模 拟<br />

材 料 按 1∶1 的 尺 度 复 建 白 鹤 梁 ( 模 型 ) 陈 列 馆 , 和 在 岸 上 建 白 鹤 楼 以 期 再 现 白 鹤 梁 的 某 些 景 观 。<br />

这 方 案 的 实 质 是 “ 淤 埋 真 国 宝 , 复 建 假 文 物 ”。 这 种 保 护 方 案 必 将 在 国 际 和 国 内 对 伟 大 的 三 峡 工 程 和<br />

我 国 的 文 保 工 作 带 来 严 重 的 负 面 影 响 。 何 况 经 过 长 时 期 泥 沙 淤 埋 过 的 白 鹤 梁 古 水 文 题 刻 是 否 会 安 全<br />

无 损 也 无 科 学 定 论 。 此 方 案 也 不 符 合 世 界 上 公 认 的 文 物 保 护 方 面 的 主 要 原 则 。 由 于 时 间 紧 急 , 三 峡<br />

水 库 蓄 水 在 即 , 而 且 又 没 有 别 的 合 理 的 新 保 护 方 案 在 各 次 全 国 性 会 议 上 提 出 , 在 2001 年 2 月 25-26<br />

日 , 国 家 文 物 局 等 单 位 在 涪 陵 召 开 的 评 审 会 时 事 先 已 确 定 要 按 此 方 案 实 施 , 并 要 求 作 相 关 的 工 程 设<br />

计 了 。<br />

图 28 白 鹤 梁 复 建 平 台 示 意 图<br />

5.3. 具 有 创 新 性 的 原 址 水 下 保 护 白 鹤 梁 古 水 文 题 刻 方 案 的 提 出<br />

2001 年 2 月 初 , 笔 者 非 常 幸 运 地 收 到 了 参 加 2001 年 2 月 涪 陵 会 议 的 邀 请 信 。 同 年 2 月 中 旬 ,<br />

笔 者 因 有 事 出 差 去 北 京 , 顺 道 访 问 了 国 家 文 物 局 , 问 询 了 有 关 保 护 白 鹤 梁 的 情 况 , 这 是 本 人 第 一 次<br />

了 解 到 白 鹤 梁 概 况 及 其 保 护 方 面 的 历 史 沿 革 和 即 将 举 行 的 会 议 拟 采 取 的 方 案 大 致 内 容 , 听 完 了 情 况<br />

介 绍 后 我 总 感 到 会 议 将 采 取 的 方 案 并 不 理 想 。 我 直 率 地 向 文 物 局 文 保 司 的 领 导 和 工 作 人 员 表 达 了 我<br />

的 初 步 感 受 。 他 们 听 后 反 问 我 ,“ 那 么 您 有 什 么 好 的 方 案 吗 ?” 我 思 索 约 5 分 钟 初 步 分 析 了 以 前 一 些<br />

方 案 中 存 在 的 问 题 后 , 脑 子 中 形 成 了 以 “ 无 压 容 器 ” 为 基 本 理 念 的 原 址 水 下 保 护 国 宝 白 鹤 梁 的 粗 略 方<br />

案 。 他 们 听 了 我 的 简 略 表 述 后 , 感 到 很 突 然 , 也 很 高 兴 但 不 便 于 表 态 。 我 恳 请 他 们 借 给 我 一 些 资 料 ,<br />

仔 细 斟 酌 。 当 晚 我 在 软 卧 车 厢 彻 夜 未 眠 仔 细 阅 看 资 料 和 反 复 思 考 已 初 具 雏 形 的 “ 无 压 容 器 ” 方 案 。 到<br />

上 海 休 息 两 天 后 又 乘 火 车 于 2 月 24 日 到 达 重 庆 转 涪 陵 去 参 加 会 议 。 在 火 车 上 又 基 本 未 眠 , 思 考 和<br />

研 究 各 种 方 案 的 优 缺 点 。 这 次 涪 陵 会 议 是 由 国 家 文 物 局 、 重 庆 市 政 府 和 国 务 院 三 峡 工 程 建 设 委 员 会<br />

联 合 主 持 召 开 的 。 会 议 开 幕 前 邀 请 笔 者 担 任 会 议 专 家 评 审 组 组 长 , 我 私 下 向 他 们 表 示 , 我 对 即 将 通<br />

过 的 方 案 并 不 很 认 同 , 再 三 请 辞 组 长 一 职 。 组 织 会 议 的 领 导 同 志 向 我 诚 恳 表 示 , 由 于 三 峡 水 库 再 过<br />

三 四 年 就 要 蓄 水 了 , 时 间 已 经 没 有 了 , 现 在 只 能 采 用 这 次 会 议 将 要 通 过 的 方 案 。 他 们 坚 持 一 定 要 我<br />

担 任 评 审 组 长 , 使 原 定 方 案 能 顺 利 通 过 , 通 过 评 审 使 此 方 案 更 完 善 些 。 至 于 我 的 一 些 不 同 看 法 可 以<br />

-45-


在 会 议 结 束 前 给 我 半 小 时 做 个 发 言 备 个 案 。 我 只 好 勉 为 其 难 。 利 用 会 议 第 一 天 晚 上 的 时 间 要 了 数 张<br />

透 明 薄 膜 和 自 带 的 彩 色 笔 准 备 了 第 二 天 上 午 会 议 结 束 时 我 的 发 言 提 纲 和 方 案 简 图 。<br />

笔 者 在 会 议 通 过 了 “ 就 地 保 护 , 异 地 陈 展 ” 方 案 后 , 做 了 半 个 小 时 发 言 , 提 出 了 一 种 基 于 “ 无 压<br />

容 器 ” 概 念 的 新 的 原 址 水 下 保 护 工 程 方 案 。 我 很 庆 幸 “ 无 压 容 器 ” 方 案 立 即 获 得 了 全 体 评 委 的 一 致 赞<br />

同 。<br />

借 此 作 一 简 要 说 明 : 这 里 把 原 址 保 护 体 看 作 一 个 容 器 , 所 谓 “ 无 压 容 器 ” 不 是 指 什 么 压 力 都 没 有 ,<br />

而 是 指 作 用 在 水 下 保 护 体 外 面 的 水 压 力 压 强 与 内 壁 面 上 的 水 压 力 压 强 相 同 , 或 基 本 相 同 , 只 差 一 个<br />

很 小 的 量 。 这 样 有 关 损 伤 破 裂 、 渗 流 破 坏 、 帷 幕 灌 浆 等 工 程 难 题 全 可 排 除 了 , 这 就 是 说 在 保 护 体 内<br />

有 水 、 且 压 力 强 度 与 当 时 作 用 在 外 壁 面 上 的 长 江 水 压 力 压 强 同 步 变 化 。<br />

大 会 秘 书 组 起 草 的 专 家 评 审 会 意 见 书 在 原 有 六 段 文 字 表 达 已 评 审 方 案 已 经 获 得 通 过 并 要 求 作<br />

一 些 改 进 的 内 容 外 又 增 补 了 新 的 也 是 最 后 一 段 文 字 :“ 专 家 组 认 为 白 鹤 梁 古 水 文 题 刻 原 址 自 然 淤 埋 的<br />

保 护 方 案 仍 不 属 于 理 想 方 案 。 有 专 家 在 会 上 提 出 了 修 建 新 型 水 下 博 物 馆 进 行 原 址 保 护 的 设 想 。 评 审<br />

组 认 为 此 新 设 想 能 更 有 效 确 保 三 峡 水 库 运 行 后 的 题 刻 安 全 , 也 有 利 于 今 后 有 关 研 究 保 护 工 作 的 开<br />

展 , 建 议 国 家 主 管 部 门 在 加 以 实 施 现 有 规 划 的 同 时 , 抓 紧 开 展 对 此 新 设 想 的 研 究 和 论 证 工 作 。”<br />

当 时 在 场 的 各 部 门 负 责 同 志 虽 然 内 心 都 很 赞 同 笔 者 提 出 的 新 方 案 。 但 苦 于 三 峡 蓄 水 在 即 已 没 有<br />

时 间 再 做 新 方 案 研 究 而 不 便 在 会 上 表 态 。 很 多 领 导 同 志 向 我 表 白 说 :“ 如 果 您 参 加 了 前 几 次 会 议 , 说<br />

不 定 现 在 已 按 您 的 方 案 实 施 了 ”,“ 但 现 在 已 为 时 太 晚 了 , 不 好 办 了 ”。 因 此 , 各 项 施 工 方 案 仍 按 原 定<br />

方 案 进 行 。<br />

六 、“ 无 压 容 器 ” 原 址 水 下 保 护 白 鹤 梁 方 案 的 论 证 、 批 准 和 开 工<br />

6.1. 国 家 领 导 人 及 主 管 部 门 的 支 持<br />

在 已 有 确 定 方 案 的 情 况 下 要 推 翻 原 方 案 采 取 新 方 案 绝 非 一 件 易 事 , 要 争 取 得 到 国 家 高 层 领 导 人<br />

和 主 管 部 门 的 支 持 是 必 需 的 。<br />

从 2 月 26 日 笔 者 提 出 “ 无 压 容 器 ” 保 护 方 案 后 算 起 , 一 个 月 很 快 过 去 了 , 犹 如 石 沉 大 海 , 无 任<br />

何 信 息 反 馈 。 作 为 一 名 科 学 工 作 者 , 感 到 自 己 肩 负 的 责 任 。 为 了 国 家 的 利 益 和 为 了 保 护 好 祖 宗 留 下<br />

的 文 化 遗 存 的 神 圣 使 命 应 尽 一 切 力 量 去 争 取 。<br />

(1)2001 年 3 月 23 日 , 笔 者 用 人 民 来 信 的 方 式 , 给 当 时 的 朱 镕 基 总 理 写 信 , 希 望 国 家 能 更<br />

改 原 定 方 案 采 用 更 科 学 合 理 的 保 护 方 案 。<br />

图 29 是 国 家 信 访 办 4 月 9 日 给 我 的 回 信 。 图 30 是 笔 者 给 朱 总 理 建 议 信 中 所 附 的 方 案 示 意 图 。<br />

图 29 国 家 信 访 办 4 月 9 日 给 笔 者 的 回 信<br />

-46-


图 30 笔 者 给 朱 总 理 建 议 信 中 所 附 的 方 案 示 意 图<br />

(2) 在 中 国 工 程 院 的 大 力 支 持 下 2001 年 3 月 29 日 用 工 程 院 院 士 建 议 文 件 报 送 中 共 中 央 、 国<br />

务 院 有 关 领 导 , 中 共 中 央 办 公 厅 、 国 务 院 办 公 厅 、 全 国 人 大 、 全 国 政 协 等 , 图 31、 图 32 是 院 士 建<br />

议 文 件 的 首 页 和 末 页 。<br />

(3)2001 年 4 月 10 日 中 国 工 程 院 又 进 一 步 向 国 务 院 办 公 厅 秘 书 局 报 送 专 报 信 息 , 图 33 是 专<br />

报 信 息 的 首 页 图 片 。<br />

(4) 由 新 华 社 内 参 记 者 采 访 后 写 专 稿 给 中 央 和 国 务 院 领 导 同 志 参 阅 。<br />

(5) 由 时 任 中 国 工 程 院 院 长 宋 健 同 志 在 笔 者 给 他 的 信 上 批 示 后 转 给 三 峡 建 设 委 员 会 领 导 同 志 ,<br />

他 在 信 上 批 示 说 “ 请 听 听 院 士 的 呼 声 ”。<br />

-47-


图 31 院 士 建 议 文 件 的 首 页<br />

图 32 工 程 院 院 士 建 议 末 页 及 报 送 单 位<br />

-48-


图 33 专 报 信 息 的 首 页 图 片<br />

6.2. 基 于 “ 无 压 容 器 ” 概 念 的 保 护 新 方 案 开 始 启 动<br />

2001 年 9 月 经 国 务 院 三 峡 工 程 建 设 委 员 会 办 公 室 , 国 家 文 物 局 及 重 庆 市 政 府 同 意 , 由 笔 者 负<br />

责 , 由 长 江 水 利 委 员 会 长 江 规 划 勘 察 设 计 研 究 院 ( 以 下 简 称 长 江 设 计 院 ) 配 合 , 用 三 个 月 时 间 编 制<br />

相 关 的 可 行 性 研 究 报 告 向 上 呈 报 。2002 年 3 月 , 可 行 性 研 究 报 告 的 修 改 稿 得 到 有 关 领 导 部 门 的 正 式<br />

批 准 后 , 随 即 进 行 工 程 设 计 。 设 计 工 作 由 长 江 设 计 院 负 责 , 笔 者 任 设 计 院 该 项 目 的 顾 问 并 兼 任 投 资<br />

方 的 顾 问 。 由 于 此 项 工 程 的 复 杂 性 , 设 计 单 位 还 邀 请 了 中 科 院 武 汉 岩 土 力 学 研 究 所 、 上 海 交 大 岩 土<br />

力 学 与 工 程 研 究 所 、 铁 道 部 第 四 勘 测 设 计 院 、 武 昌 造 船 厂 、 华 中 科 技 大 学 、 武 汉 大 学 、 重 庆 西 南 水<br />

-49-


科 所 、 重 庆 交 通 学 院 等 单 位 分 工 开 展 了 九 项 专 题 研 究 , 这 九 项 专 题 名 称 如 下 :(1) 涪 陵 白 鹤 梁 题 刻<br />

原 址 水 下 保 护 工 程 ( 以 下 简 称 水 下 保 护 工 程 ) 对 流 态 、 流 势 影 响 的 实 验 研 究 ;(2) 水 下 保 护 工 程 三<br />

维 非 线 性 结 构 分 析 ;(3) 水 下 交 通 廊 道 ( 沉 管 方 案 ) 专 题 研 究 ;(4) 水 下 保 护 工 程 参 观 廊 道 设 计 的<br />

专 题 研 究 ;(5) 水 下 照 明 及 CCD 遥 控 观 测 系 统 ;(6) 水 下 保 护 工 程 内 外 压 平 衡 和 滤 清 的 循 环 水 系<br />

统 ;(7) 水 下 保 护 工 程 安 全 健 康 监 测 系 统 ;(8) 水 下 保 护 工 程 施 工 方 法 研 究 ;(9) 航 道 航 运 问 题 研<br />

究 。2002 年 10 月 总 体 设 计 完 成 , 同 年 12 月 工 程 设 计 和 概 算 得 到 国 家 有 关 部 门 批 准 。<br />

6.3. 白 鹤 梁 原 址 水 下 保 护 工 程 正 式 开 工<br />

2003 年 2 月 13 日 , 基 于 “ 无 压 容 器 ” 概 念 的 白 鹤 梁 古 水 文 题 刻 的 原 址 水 下 保 护 工 程 正 式 开 工 ,<br />

当 时 由 青 岛 海 军 某 部 队 负 责 施 工 , 主 要 是 炸 礁 修 筑 防 撞 墩 等 辅 助 工 作 等 , 图 34 是 开 工 典 礼 照 片 。<br />

图 34 白 鹤 梁 水 下 保 护 工 程 开 工<br />

七 、 白 鹤 梁 古 水 文 题 刻 原 址 水 下 保 护 工 程 的 基 本 内 容<br />

7.1. 方 案 的 基 本 要 点<br />

白 鹤 梁 古 水 文 题 刻 原 址 水 下 保 护 工 程 采 用 无 压 容 器 方 案 的 基 本 内 容 如 下 :<br />

(1) 水 库 水 位 与 水 下 保 护 壳 体 内 的 水 位 基 本 保 持 相 同 。<br />

(2) 长 江 水 水 质 好 , 是 保 护 白 鹤 梁 最 理 想 的 介 质 , 一 千 二 百 余 年 历 史 就 是 明 证 , 但 需 要 适 当<br />

滤 去 悬 移 质 , 以 防 淤 积 , 并 使 水 质 透 明 度 高 , 利 于 参 观 者 观 看 题 刻 。<br />

(3) 白 鹤 梁 中 段 东 部 65 余 米 区 段 上 集 中 有 大 部 分 主 要 题 刻 , 修 建 水 下 保 护 工 程 主 要 在 这 个 区<br />

段 。<br />

参 观 廊 道<br />

桩 基<br />

Piles<br />

图 35 白 鹤 梁 工 程 地 质 剖 面 图<br />

-50-


(4) 围 住 白 鹤 梁 题 刻 的 是 平 面 上 呈 椭 圆 型 、 厚 度 为 3.5 米 左 右 的 钢 筋 砼 墙 —— 称 之 为 导 墙 , 石<br />

刻 密 集 区 段 被 导 墙 及 穹 顶 保 护 着 。 工 程 地 质 及 保 护 体 示 意 图 见 图 35。<br />

(5) 导 墙 上 覆 盖 有 厚 度 为 1 米 左 右 高 配 筋 砼 穹 顶 壳 体 、 不 拆 卸 的 内 模 板 采 用 由 爆 炸 成 型 的 不<br />

锈 钢 复 合 板 。<br />

(6) 由 于 是 “ 无 压 容 器 ”, 机 理 上 保 证 不 会 出 现 严 重 事 故 , 这 种 主 体 保 护 工 程 具 有 可 修 复 性 , 造<br />

价 相 对 便 宜 、 施 工 期 短 等 特 点 。<br />

(7) 在 任 何 时 候 参 观 人 群 可 进 入 岸 上 陈 列 馆 通 过 耐 压 的 斜 坡 交 通 廊 道 和 耐 压 的 水 平 交 通 廊 道<br />

( 约 140 米 长 ) 进 入 位 于 保 护 壳 体 内 部 的 耐 压 的 钢 质 参 观 廊 道 , 外 径 约 3.8 米 的 钢 质 参 观 廊 道 按 耐<br />

60 米 水 头 的 潜 水 器 设 计 规 范 设 计 , 通 过 观 察 窗 直 接 观 赏 白 鹤 梁 古 水 文 题 刻 。 图 36 中 所 示 的 是 工 程<br />

方 案 示 意 图 。<br />

(8) 备 有 LED 大 功 率 水 下 灯 光 照 明 及 先 进 的 水 中 摄 像 装 置 , 参 观 者 可 在 参 观 廊 道 内 通 过 玻 璃<br />

窗 观 看 可 操 纵 设 备 对 白 鹤 梁 题 刻 进 行 观 赏 。 观 察 窗 直 径 800 毫 米 , 为 确 保 安 全 和 便 于 修 理 , 设 双 层<br />

玻 璃 窗 , 玻 璃 系 特 制 的 用 于 潜 艇 等 的 有 机 玻 璃 , 玻 璃 厚 度 80 毫 米 。<br />

图 36 涪 陵 白 鹤 梁 题 刻 原 址 水 下 保 护 工 程 方 案 示 意 图<br />

(9) 特 设 蛙 人 进 出 口 , 特 定 游 客 可 由 蛙 人 导 游 引 导 到 题 刻 前 参 观 。<br />

(10) 按 规 划 在 三 个 枯 水 季 节 将 水 下 关 键 工 程 部 分 完 成 , 符 合 国 家 做 出 的 三 峡 工 程 提 前 蓄 水 的<br />

进 度 。 施 工 期 碍 航 不 严 重 。<br />

(11) 与 “ 水 晶 宫 ” 方 案 相 比 , 费 用 相 对 要 低 很 多 , 施 工 难 度 小 , 最 主 要 的 是 没 有 损 毁 的 风 险 。<br />

7.2. 白 鹤 梁 原 址 水 下 保 护 工 程 遵 循 的 原 则<br />

从 前 述 的 基 本 要 点 可 以 看 出 白 鹤 梁 原 址 水 下 保 护 工 程 遵 循 了 如 下 原 则 :<br />

(1) 符 合 国 际 有 关 文 物 保 护 的 原 则 即 原 址 原 样 , 原 环 境 的 保 护 原 则 。<br />

(2) 由 于 古 水 文 镌 刻 分 布 范 围 很 广 , 但 主 要 和 重 要 部 分 集 中 在 白 鹤 梁 中 段 东 区 约 65 米 的 区 段<br />

内 , 所 以 根 据 重 点 保 护 原 则 , 我 们 保 护 体 的 保 护 范 围 集 中 在 东 区 约 65 米 地 段 。 图 37 所 示 的 为 中 段<br />

东 区 题 刻 分 布 图 。<br />

(3) 除 了 保 护 外 , 还 能 为 广 大 参 观 者 观 赏 原 则 。<br />

(4) 实 施 可 行 性 原 则 。<br />

(5) 工 程 完 整 性 原 则 。<br />

(6) 可 持 续 发 展 原 则 。<br />

-51-


图 37 白 鹤 梁 题 刻 中 段 东 区 题 刻 分 布 图<br />

八 、 白 鹤 梁 原 址 水 下 保 护 工 程 会 被 江 水 冲 毁 而 酿 成 灭 顶 之 灾 吗<br />

白 鹤 梁 水 下 保 护 工 程 开 工 不 久 , 就 有 一 位 据 说 专 门 研 究 河 流 动 力 学 的 某 名 校 教 授 , 在 一 本<br />

级 别 不 高 的 技 术 刊 物 上 发 表 了 一 篇 署 名 文 章 。 该 文 的 中 心 意 思 是 : 长 江 涪 陵 段 是 呈 弧 形 弯 段 ( 参 见<br />

图 28), 根 据 河 床 动 力 学 的 基 本 规 律 由 于 离 心 力 的 作 用 长 江 水 的 主 流 将 主 要 冲 刷 该 段 的 长 江 南 岸 ,<br />

因 此 建 在 长 江 南 岸 深 水 航 道 旁 边 的 白 鹤 梁 水 下 保 护 工 程 必 将 被 长 江 水 冲 毁 酿 成 灭 顶 之 灾 。 文 章 作 者<br />

认 为 白 鹤 梁 原 址 水 下 保 护 工 程 的 提 出 者 没 有 认 识 到 这 个 问 题 。 文 章 作 者 进 一 步 指 出 , 修 建 这 样 的 保<br />

护 工 程 违 反 了 自 然 的 规 律 , 是 注 定 要 失 败 的 。<br />

文 章 作 者 将 此 文 通 过 有 关 渠 道 在 极 短 的 时 间 就 已 送 达 到 国 家 主 管 三 峡 等 重 大 工 程 的 领 导<br />

同 志 前 。 这 位 领 导 同 志 要 求 主 管 白 鹤 梁 工 程 的 三 个 领 导 部 门 迅 速 向 中 央 写 一 份 报 告 把 问 题 讲 清 楚 。<br />

否 则 , 将 面 临 停 工 的 危 险 。<br />

根 据 三 个 领 导 部 门 指 示 , 我 必 需 在 不 长 时 间 内 提 出 我 个 人 的 意 见 以 供 他 们 考 虑 。 当 时 正 处<br />

于 SARS 的 高 峰 期 , 不 便 于 出 差 , 我 只 能 关 在 家 里 写 书 面 报 告 , 并 呈 送 给 国 家 文 物 局 和 三 建 委 等 领<br />

导 部 门 。<br />

我 的 主 要 意 见 是 如 下 两 点 :<br />

(1) 在 涪 陵 弯 段 的 长 江 水 不 会 因 离 心 力 而 发 生 冲 刷 其 南 岸 , 究 其 原 因 是 在 该 段 有 乌 江 入<br />

流 , 乌 江 是 一 大 河 , 常 年 平 均 流 量 很 大 , 其 入 江 口 距 我 们 的 水 下 保 护 工 程 又 不 足 一 公 里 。 由 于 乌 江<br />

的 顶 托 、 此 段 长 江 水 流 的 总 体 趋 势 是 逼 向 北 岸 偏 移 。 弯 段 河 流 的 冲 刷 原 理 虽 是 常 识 , 但 又 看 不 到 乌<br />

江 入 流 顶 托 影 响 确 是 缺 乏 最 基 本 的 常 识 和 地 理 知 识 。<br />

(2) 涪 陵 的 白 鹤 梁 历 经 千 余 年 还 保 存 十 分 完 好 。 就 是 由 于 上 述 原 因 , 历 史 也 证 明 上 述 理<br />

由 是 正 确 的 。 如 今 在 白 鹤 梁 上 加 了 一 保 护 壳 , 既 没 有 碍 航 , 又 没 有 阻 断 乌 江 入 流 , 基 本 情 况 未 变 ,<br />

怎 么 会 发 生 该 文 章 作 者 所 得 出 的 所 谓 科 学 论 断 ——“ 灭 顶 之 灾 ” 呢 ?<br />

根 据 我 回 答 的 理 由 , 有 关 部 门 拟 写 了 报 告 呈 送 中 央 后 , 一 场 处 于 停 工 边 缘 的 风 潮 终 于 平 息 了 。<br />

九 、 国 宝 白 鹤 梁 古 水 文 题 刻 原 址 水 下 保 护 工 程 的 施 工 、 地 面 陈 列 馆 和 八 大 技 术 系 统<br />

9.1. 水 下 主 体 工 程 的 施 工<br />

它 是 白 鹤 梁 水 下 保 护 工 程 成 败 的 关 键 。 由 于 椭 圆 形 导 墙 位 于 斜 坡 上 , 旁 边 就 是 深 水 航 道 , 流 速<br />

很 大 , 经 研 究 采 用 整 体 刚 性 模 板 ( 图 38), 浇 筑 水 下 砼 。 保 护 体 厚 3.5m 左 右 导 墙 施 工 浇 筑 的 情 况 见<br />

图 39。 预 埋 管 接 头 ( 直 径 约 4 米 ) 运 输 情 况 见 图 40。 经 过 艰 苦 努 力 , 保 护 体 导 墙 完 工 后 的 状 态 见<br />

图 41。 围 堰 施 工 时 的 情 况 见 图 42。 图 43 是 围 堰 快 接 近 完 成 时 的 情 况 。 胜 利 合 拢 围 堰 , 抽 干 水 后 为<br />

后 续 工 程 干 法 施 工 创 造 了 十 分 有 利 条 件 。 合 拢 后 的 施 工 现 场 见 图 44。 上 下 游 水 平 交 通 廊 道 的 施 工 情<br />

况 可 参 看 图 45,46, 和 47,48。 斜 坡 交 通 廊 道 施 工 情 况 见 图 49, 图 50, 图 51 和 图 52。 水 平 和 斜<br />

-52-


坡 交 通 廊 道 接 近 完 成 的 夜 景 见 图 53。 图 54 是 水 下 保 护 主 体 工 程 完 成 后 又 将 被 长 江 水 淹 没 时 的 情 景 。<br />

图 38 刚 性 模 板<br />

图 39 保 护 体 导 墙 施 工 的 情 况<br />

图 40 预 埋 管 接 头 运 输 情 况<br />

图 41 保 护 体 导 墙 完 工 后 的 状 态<br />

图 42 围 堰 施 工 时 的 情 况<br />

图 43 围 堰 快 接 近 完 成 时 的 情 况<br />

图 44 合 拢 后 的 施 工 现 场 图 45 上 下 游 水 平 交 通 廊 道 的 施 工 情 况 (1)<br />

-53-


图 46 上 下 游 水 平 交 通 廊 道 的 施 工 情 况 (2) 图 47 上 下 游 水 平 交 通 廊 道 的 施 工 情 况 (3)<br />

图 48 上 下 游 水 平 交 通 廊 道 的 施 工 情 况 (4) 图 49 斜 坡 交 通 廊 道 施 工 情 况 (1)<br />

图 50 斜 坡 交 通 廊 道 施 工 情 况 (2) 图 51 斜 坡 交 通 廊 道 施 工 情 况 (3)<br />

图 52 斜 坡 交 通 廊 道 施 工 情 况 (4)<br />

图 53 水 平 和 斜 坡 交 通 廊 道 接 近 完 成 时 的 夜 景<br />

-54-


图 54 水 下 保 护 工 程 主 体 部 分 完 成 后 又 将 被 长 江 水 淹 没 时 的 情 景 (2006 年 夏 )<br />

9.2. 参 观 廊 道 的 结 构 、 制 造 和 安 装<br />

参 观 廊 道 是 白 鹤 梁 原 址 水 下 保 护 工 程 中 极 重 要 的 金 属 结 构 。 直 径 为 3.2m 壁 厚 28mm 的 圆 形 钢<br />

结 构 管 要 承 受 40 多 米 水 头 的 压 力 , 完 全 按 照 潜 水 艇 的 标 准 来 设 计 和 制 造 。<br />

图 55 是 在 成 都 化 工 压 力 容 器 厂 内 制 造 的 一 节 参 观 廊 道 从 成 都 起 运 的 情 况 。 图 中 管 道 有 五 只 圆<br />

形 筒 是 观 察 窗 口 。 每 只 观 察 窗 都 安 装 有 双 层 玻 璃 ( 图 56)。 整 条 观 察 廊 道 共 有 七 节 管 道 组 成 , 装 有<br />

23 只 观 察 窗 。 一 只 救 生 球 仓 、 一 只 设 备 球 仓 ( 图 57)。 这 七 节 管 道 都 从 驳 船 上 吊 装 到 导 墙 腔 体 内 。<br />

最 大 一 节 管 道 重 量 达 45t。 各 节 管 道 要 准 确 定 位 和 无 水 焊 接 。 拼 缝 的 准 确 度 要 求 十 分 严 格 。 全 部 焊<br />

缝 要 经 过 多 种 手 段 严 格 检 查 并 必 需 100% 合 格 。 图 58、59、60 所 示 的 是 拼 装 焊 接 好 的 参 观 廊 道 安 装<br />

在 保 护 壳 腔 体 内 的 情 景 。<br />

图 55 参 观 廊 道 从 成 都 起 运<br />

图 56 观 察 窗<br />

图 57 救 生 球 窗 与 设 备 球 窗<br />

图 58 参 观 廊 道 安 装 在 保 护 壳 腔 体 内 (Ⅰ)<br />

-55-


图 59 参 观 廊 道 安 装 在 保 护 壳 腔 体 内 (Ⅱ)<br />

图 60 参 观 廊 道 安 装 在 保 护 壳 腔 体 内 (Ⅲ)<br />

9.3. 钢 骨 架 和 钢 筋 砼 穹 顶 的 浇 筑<br />

图 61、62 所 示 的 是 穹 顶 钢 骨 架 的 安 装 和 浇 筑 在 钢 筋 砼 穹 顶 内 的 钢 筋 网 。<br />

在 上 下 游 、 斜 坡 交 通 廊 道 内 各 装 有 一 台 垂 直 落 差 40 余 米 的 隧 道 式 自 动 扶 梯 ( 图 63)。 这 条 隧 道<br />

式 自 动 扶 梯 可 能 是 亚 洲 垂 直 落 差 最 大 的 一 条 且 隧 道 外 围 承 受 最 大 水 压 力 大 于 40 米 水 头 。<br />

为 了 施 工 方 便 , 穹 顶 内 模 板 采 用 爆 炸 成 型 的 厚 度 达 12mm 的 复 合 钢 板 制 成 , 其 内 模 面 系 2mm<br />

的 不 锈 钢 钢 板 。 穹 顶 浇 筑 完 成 后 不 需 拆 内 模 板 。<br />

图 61 穹 顶 钢 骨 架<br />

图 62 浇 筑 钢 筋 砼 穹 顶 的 钢 筋 网<br />

图 63 隧 道 式 自 动 扶 梯<br />

-56-


9.4. 地 面 陈 列 馆<br />

白 鹤 梁 古 水 文 题 刻 陈 列 馆 修 建 在 涪 陵 城 区 主 干 道 旁 边 的 防 汛 堤 上 以 节 省 用 地 , 地 面 陈 列 馆 的 效<br />

果 图 见 图 64。 地 面 陈 列 馆 鸟 瞰 图 见 图 65。 限 于 篇 幅 , 对 地 面 陈 列 馆 在 本 文 中 就 不 作 详 细 介 绍 了 。<br />

图 64 地 面 陈 列 馆 的 效 果 图<br />

图 65 地 面 陈 列 馆 鸟 瞰 图<br />

9.5. 安 装 在 水 下 保 护 体 内 八 大 技 术 系 统<br />

(1) 循 环 水 系 统 —— 保 证 保 护 体 内 外 水 压 差 很 小 , 符 合 设 计 要 求 , 并 滤 去 悬 浮 质 使 水 质 如 自<br />

来 水 一 样 洁 净 , 并 在 一 定 周 期 内 自 动 更 换 水 体 。<br />

(2) 水 下 照 明 系 统 —— 设 计 安 装 有 150 套 LED 大 功 率 灯 具 的 接 口 , 每 套 水 下 LED 白 色 光 灯 具<br />

功 率 达 63 瓦 。<br />

(3) 水 下 摄 像 系 统 —— 共 计 有 28 套 能 自 动 跟 踪 目 标 水 下 摄 像 装 置 , 以 供 参 观 者 使 用 。<br />

(4) 水 下 消 防 系 统 。<br />

(5) 救 生 和 供 高 压 气 系 统 。<br />

(6) 参 观 廊 道 及 交 通 廊 道 的 空 调 及 通 风 系 统 。<br />

(7) 保 护 体 内 低 压 供 电 照 明 系 统 。<br />

(8) 保 护 体 健 康 诊 断 系 统 。<br />

-57-


9.6. 白 鹤 梁 水 下 博 物 馆 的 建 成<br />

博 物 馆 由 水 下 保 护 体 、 交 通 廊 道 和 参 观 廊 道 以 及 地 面 陈 列 馆 三 部 分 组 成 。<br />

总 建 筑 面 积 8433m2。 工 程 总 投 资 1.9 亿 人 民 币 。<br />

白 鹤 梁 水 下 博 物 馆 自 2003 年 2 月 13 日 开 工 建 设 , 到 2009 年 5 月 18 日 在 世 界 博 物 馆 日 开<br />

馆 , 成 为 世 界 上 唯 一 的 文 化 遗 址 类 水 下 博 物 馆 ( 水 深 40 余 米 ), 她 为 水 下 文 化 遗 产 的 原 址 保 护 提 供<br />

了 成 功 的 工 程 范 例 , 也 为 我 国 伟 大 的 三 峡 工 程 增 添 了 光 彩 。<br />

三 峡 水 库 建 成 后 白 鹤 梁 仍 将 留 存 于 人 间 , 供 人 类 一 代 代 观 赏 下 去 而 获 得 “ 新 生 ”。<br />

2006 年 经 国 家 有 关 部 门 评 审 通 过 , 白 鹤 梁 古 水 文 题 刻 和 原 址 水 下 保 护 工 程 已 列 入 我 国 申 报 世 界<br />

文 化 遗 产 的 备 用 名 单 。<br />

十 、 白 鹤 梁 古 水 文 题 刻 原 址 水 下 保 护 工 程 的 社 会 影 响<br />

该 工 程 受 到 全 国 的 关 心 , 各 类 媒 体 纷 纷 做 了 大 量 报 道 。 自 方 案 获 得 批 准 之 日 至 博 物 馆 建 成 开 放<br />

的 七 年 左 右 时 间 里 , 据 不 完 全 统 计 , 平 面 媒 体 计 有 上 海 新 民 晚 报 , 北 京 青 年 报 , 重 庆 市 、 湖 北 省 武<br />

汉 市 和 许 多 省 市 地 方 报 纸 , 都 做 过 专 题 报 道 , 中 国 科 学 时 报 , 香 港 文 汇 报 , 人 民 日 报 也 有 大 量 报 道 。<br />

在 视 频 方 面 , 中 央 一 台 , 中 央 四 台 , 中 央 十 台 , 中 央 十 一 台 , 香 港 凤 凰 台 和 各 地 方 台 曾 做 过 多 次 报<br />

道 。2009 年 5 月 18 日 世 界 博 物 馆 日 中 央 台 有 关 台 (CCTV12) 对 白 鹤 梁 博 物 馆 开 馆 的 现 场 做 实 况 转<br />

播 。 笔 者 借 此 机 会 对 各 界 新 闻 媒 体 对 此 项 工 作 的 关 爱 深 表 感 谢 。 在 大 型 纪 录 片 “ 再 说 长 江 ” 第 14 集 中<br />

对 白 鹤 梁 及 其 保 护 工 程 做 了 相 关 介 绍 。<br />

至 于 白 鹤 梁 原 址 水 下 保 护 工 程 能 给 国 家 创 造 什 么 样 的 经 济 效 益 ? 我 作 为 一 名 科 技 工 作 者<br />

是 说 不 清 楚 的 。 白 鹤 梁 是 一 处 著 名 的 水 下 文 化 遗 址 , 虽 说 它 是 国 之 愧 宝 , 应 该 说 是 无 价 之 宝 , 但 保<br />

护 她 应 如 何 与 经 济 效 益 挂 起 钩 来 , 我 是 无 法 计 算 的 。 我 作 为 一 名 科 技 工 作 者 , 能 参 与 到 国 家 文 化 遗<br />

产 的 保 护 中 去 感 到 责 无 旁 贷 , 也 是 无 尚 光 荣 。 我 至 今 还 清 晰 回 忆 起 在 2002 年 9 月 在 北 京 市 重 庆 宾<br />

馆 召 开 的 关 于 白 鹤 梁 原 址 水 下 保 护 方 案 可 行 性 论 证 会 上 一 位 当 时 已 年 过 8 旬 至 今 还 健 在 的 国 家 文 物<br />

局 顾 问 、 著 名 的 文 物 鉴 定 和 保 护 专 家 谢 辰 生 先 生 的 一 段 话 。 他 当 时 非 常 激 动 地 从 座 位 上 站 起 来 说 :<br />

这 个 方 案 如 能 实 现 , 白 鹤 梁 就 不 会 再 遭 土 埋 的 命 运 ,“ 白 鹤 梁 正 是 枯 木 逢 春 , 从 此 有 了 生 命 ”。<br />

如 今 白 鹤 梁 的 建 成 也 已 引 起 国 际 同 行 的 深 切 关 注 。 联 合 国 教 科 文 组 织 驻 京 办 事 处 多 次 到 涪<br />

陵 考 察 , 由 联 合 国 教 科 文 组 织 (UNESCO), 中 国 文 化 遗 产 研 究 院 和 重 庆 市 文 物 局 共 同 商 定 , 已 在<br />

2010 年 11 月 24-26 日 在 我 国 重 庆 市 召 开 了 “ 水 下 文 化 遗 产 的 保 护 、 展 示 与 利 用 国 际 会 议 ”。 在 会 上 除<br />

了 中 外 专 家 各 自 介 绍 和 交 流 了 这 方 面 的 经 验 外 , 还 赴 涪 陵 参 观 了 白 鹤 梁 水 下 博 物 馆 。 笔 者 在 会 上 作<br />

了 近 50 分 钟 的 专 题 介 绍 。 出 席 会 议 的 除 中 方 代 表 外 , 有 外 国 专 家 14 名 , 他 们 分 别 来 自 德 国 、 法 国 、<br />

英 国 、 印 度 、 埃 及 与 瑞 典 等 国 。 会 上 专 家 们 一 致 认 为 ,“ 白 鹤 梁 水 下 博 物 馆 是 世 界 上 唯 一 的 遗 址 类 水<br />

下 博 物 馆 。”<br />

白 鹤 梁 水 下 保 护 工 程 在 国 内 社 会 影 响 也 是 比 较 大 的 , 我 仅 举 两 个 例 子 如 下 。<br />

(1) 近 年 来 全 国 义 务 教 育 标 准 实 验 教 科 书 小 学 六 年 级 下 学 期 语 文 课 本 中 列 入 一 篇 课 文 , 这 篇<br />

课 文 名 称 叫 作 “ 白 鹤 梁 的 沉 浮 ”( 见 附 录 1), 具 体 介 绍 白 鹤 梁 的 科 学 和 人 文 艺 术 价 值 和 “ 无 压 容 器 ” 方<br />

案 。 此 课 文 本 的 封 面 和 目 录 见 图 66、67。 据 有 关 部 门 告 知 , 至 今 已 有 近 1 亿 小 学 生 念 过 此 课 文 。<br />

(2)2004 年 全 国 高 考 语 文 试 卷 第 二 试 题 (12 分 , 每 小 题 3 分 ) 阅 读 下 面 一 段 文 字 , 完 成 7-10<br />

题 ( 见 附 录 2): 该 段 课 文 以 白 鹤 梁 为 主 题 。 据 说 当 时 有 大 约 400 万 高 考 生 用 过 此 试 卷 。<br />

-58-


图 66 小 学 六 年 级 下 学 期 语 文 课 本 封 面<br />

图 67 小 学 六 年 级 下 学 期 语 文 课 本 目 录<br />

白 鹤 梁 题 刻 原 址 的 水 下 保 护 工 程 研 究 与 实 践 由 长 江 水 利 委 员 会 设 计 院 作 为 申 报 单 位 已 获 得<br />

2009 年 度 国 家 文 物 局 文 物 保 护 科 学 和 技 术 创 新 奖 一 等 奖 的 第 一 名 ( 见 图 68)。<br />

图 68 2009 年 度 文 物 保 护 科 学 和 技 术 创 新 奖 一 等 奖 获 奖 证 书 照 片<br />

十 一 、 结 论<br />

(1) 白 鹤 梁 古 水 文 题 刻 是 我 国 古 代 科 学 文 明 成 就 的 优 秀 代 表 , 在 国 际 上 举 世 无 双 , 白 鹤 梁 水<br />

下 碑 林 也 是 文 化 瑰 宝 。 白 鹤 梁 古 水 文 题 刻 被 联 合 国 教 科 文 组 织 誉 为 “ 保 存 完 好 的 世 界 唯 一 古 代 水 文<br />

站 ”。 伟 大 的 长 江 三 峡 工 程 兴 建 将 使 其 位 于 三 峡 水 库 库 底 。 进 行 科 学 的 保 护 十 分 必 要 , 白 鹤 梁 古 水 文<br />

题 刻 原 址 水 下 保 护 工 程 的 建 成 将 为 三 峡 工 程 和 我 国 的 文 物 保 护 工 作 树 立 范 例 。<br />

(2) 由 于 这 一 古 代 水 文 站 是 以 石 鱼 水 标 为 指 示 器 , 如 果 脱 离 母 岩 、 进 行 搬 迁 的 方 法 是 不 可 取<br />

的 , 就 地 淤 埋 的 方 法 也 是 不 妥 的 。<br />

(3) 对 白 鹤 梁 古 水 文 题 刻 采 用 以 “ 无 压 容 器 ” 概 念 为 基 础 的 原 址 水 下 保 护 的 原 则 是 科 学 和 合 理<br />

的 。<br />

(4)“ 无 压 容 器 ” 概 念 克 服 了 修 建 原 址 水 下 保 护 工 程 在 力 学 、 结 构 和 岩 土 力 学 和 施 工 方 面 的 重 大<br />

技 术 难 题 , 技 术 上 是 可 行 和 合 理 的 。<br />

-59-


(5) “ 无 压 容 器 ” 水 下 保 护 工 程 方 案 之 所 以 得 以 成 立 是 各 级 领 导 和 各 方 人 士 大 力 支 持 的 结 果 。<br />

立 项 的 “ 过 程 也 说 明 党 和 政 府 对 文 物 保 护 和 科 学 建 议 是 重 视 和 采 纳 的 ”( 引 自 全 国 人 大 路 甬 祥 副 委 员<br />

长 的 批 语 )。<br />

(6) 白 鹤 梁 古 水 文 题 刻 水 下 博 物 馆 已 被 国 内 外 同 行 公 认 为 世 界 上 唯 一 遗 产 类 水 下 博 物 馆 。<br />

致 谢<br />

(1) 谨 以 此 文 向 所 有 为 白 鹤 梁 的 保 护 付 出 过 心 血 和 智 慧 的 人 表 示 感 谢 , 不 论 曾 经 提 出 的 方 案 、<br />

意 见 或 建 议 最 后 是 否 被 采 纳 。<br />

(2) 感 谢 国 务 院 三 峡 工 程 建 设 委 员 会 办 公 室 , 中 国 国 家 文 物 局 和 重 庆 市 人 民 政 府 对 白 鹤 梁 保<br />

护 工 程 的 关 心 、 支 持 和 指 导 !<br />

(3) 感 谢 中 国 工 程 院 , 中 国 科 学 院 对 本 项 工 作 的 关 心 和 支 持 !<br />

(4) 感 谢 中 国 长 江 三 峡 工 程 开 发 总 公 司 对 本 项 工 作 的 大 力 支 持 !<br />

(5) 感 谢 参 加 白 鹤 梁 古 水 文 题 刻 原 址 水 下 保 护 工 程 的 各 设 计 、 科 研 、 施 工 、 制 造 和 管 理 、 监<br />

理 等 各 单 位 广 大 员 工 的 辛 勤 劳 动 。<br />

(6) 感 谢 联 合 国 科 教 文 组 织 和 国 际 文 物 保 护 工 作 者 和 我 国 各 界 组 织 和 文 物 保 护 工 作 者 对 我 国<br />

此 项 原 址 水 下 文 物 保 护 工 程 的 关 心 。<br />

(7) 感 谢 第 五 届 海 峡 两 岸 结 构 与 岩 土 工 程 学 术 研 讨 会 组 织 委 员 会 和 香 港 理 工 大 学 的 盛 情 邀 请 ,<br />

使 我 有 机 会 能 在 香 港 向 海 峡 两 岸 同 行 同 仁 们 汇 报 白 鹤 梁 工 程 的 概 况 。<br />

(8) 感 谢 上 海 交 通 大 学 , 中 国 科 学 院 武 汉 岩 土 力 学 研 究 所 的 大 力 支 持 !<br />

参 考 书 目<br />

[1] 中 国 人 民 政 治 协 商 会 议 四 川 省 委 员 会 涪 陵 地 区 工 作 委 员 会 编 . 世 界 第 一 古 代 水 文 站 白 鹤 梁 . 中 国 三 峡 出 版<br />

社 , 1995.<br />

[2] 陈 曦 震 主 编 . 水 下 碑 林 白 鹤 梁 . 四 川 人 民 出 版 社 , 1995.<br />

[3] 葛 修 润 . 国 宝 “ 白 鹤 梁 ”, 中 国 三 峡 建 设 , 2006, (3): 73-79.<br />

-60-


附 录 1 白 鹤 梁 的 沉 浮 ( 全 国 义 务 教 育 标 准 实 验 教 科 书 小 学 六 年 级 下 学 期 语 文 课 文 )<br />

-61-


-62-


-63-


-64-


-65-


-66-


附 录 2 2004 年 普 通 高 等 学 校 招 生 全 国 统 一 考 试 语 文 试 题<br />

第 I 卷<br />

本 试 卷 共 14 小 题 , 每 小 题 3 分 , 共 42 分 。<br />

一 、(18 分 , 每 小 题 3 分 )( 题 略 )<br />

-67-


-68-


-69-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

低 层 冷 弯 薄 壁 型 钢 龙 骨 式 住 宅 结 构 足 尺 模 型 振 动 台 试 验 研 究<br />

沈 祖 炎 * 1,2<br />

1,2 2<br />

3<br />

4<br />

李 元 齐 刘 飞 秦 雅 菲 吴 曙 岽<br />

(1. 土 木 工 程 防 灾 国 家 重 点 实 验 室 , 上 海 200092;2. 同 济 大 学 建 筑 工 程 系 , 上 海 200092;3. 蓝 璀 建 筑 钢 结 构 ( 上<br />

海 ) 有 限 公 司 , 上 海 201613;4. 上 海 钢 之 杰 ( 集 团 ) 公 司 , 上 海 200949)<br />

摘 要 : 本 文 设 计 了 三 栋 两 层 冷 弯 薄 壁 型 钢 住 宅 房 屋 的 足 尺 模 型 , 龙 骨 体 系 模 型 I 采 用 屈 服 强 度<br />

550MPa 钢 材 , 模 型 II 和 III 采 用 S350 冷 轧 板 。 复 合 墙 体 外 墙 的 外 覆 面 板 分 别 采 用 OSB 板 和 带 肋 波<br />

纹 钢 板 , 内 覆 面 板 采 用 石 膏 板 ; 内 墙 内 外 覆 面 板 均 为 石 膏 板 。 进 行 了 三 栋 模 型 振 动 台 试 验 , 试 验 选<br />

取 3 条 实 测 地 震 波 记 录 和 1 条 人 工 合 成 波 , 加 载 工 况 考 虑 从 7 度 多 遇 到 9 度 罕 遇 水 平 。 结 果 表 明 ,<br />

结 构 破 坏 均 发 生 在 连 接 部 位 和 墙 板 局 部 区 域 , 主 要 破 坏 模 式 表 现 为 自 攻 螺 钉 的 脱 落 、 石 膏 板 局 部 破<br />

裂 和 带 肋 波 纹 钢 板 相 互 脱 离 。 模 型 底 层 墙 体 钢 龙 骨 斜 拉 条 可 能 屈 服 破 坏 , 墙 体 立 柱 基 本 无 破 坏 。 结<br />

构 设 计 时 要 关 注 门 、 窗 部 位 局 部 加 强 和 自 攻 螺 钉 连 接 可 靠 性 。 在 强 震 作 用 下 , 整 体 结 构 水 平 刚 度 虽<br />

然 会 有 大 削 弱 , 但 无 倒 塌 危 险 , 采 用 此 两 种 双 面 覆 板 构 造 复 合 墙 体 的 模 型 结 构 在 试 验 的 地 震 烈 度 水<br />

平 下 能 够 符 合 我 国 抗 震 设 防 地 区 大 震 不 倒 的 要 求 。 最 后 , 对 这 类 结 构 的 抗 震 设 计 方 法 提 出 了 建 议 。<br />

关 键 词 : 冷 弯 薄 壁 型 钢 振 动 台 试 验 动 力 特 性 抗 震 性 能<br />

SHAKING TABLE TESTS ON FULL-SCALE LOW-RISE COLD-FORMED<br />

THIN-WALLED STEEL RESIDENTIAL BUILDINGS USING LIGHT-GAUGE<br />

COMPOSITE WALLS<br />

Zuyan Shen* 1,2 , Yuan-Qi Li 1,2 , Fei Liu 2 , Yafei Qing 3 , Shudong Wu 4<br />

1 State Key Laboratory of Disaster Reduction in Civil Engineering, Shanghai 200092, China<br />

2 Department of Building Engineering, Tongji <strong>University</strong>, Shanghai 200092, China<br />

3 Bluescope Steel (Shanghai) Co., Ltd , Shanghai 201613, China<br />

4 Shanghai Beststeel Group, Shanghai 200949, China<br />

Abstract: In this paper, three full-scale models of cold-formed thin-walled steel residential structures were<br />

designed, one with G550 framing system, OSB board as external panel and gypsum board as internal panel, the<br />

other two with S350 framing system and OSB board and ribbed corrugated steel sheet as external panels,<br />

respectively, and gypsum board as internal panel. Shaking table tests were performed on these three models<br />

using three actual seismic records and one artificial wave with a loading scheme from intensity 7 0 of frequently<br />

occurred earthquake to intensity 9 0 of seldom occurred earthquake. It is shown that, the structural composite<br />

walls can satisfy seismic fortification level specified in the code, all the three models have excellent<br />

performance under the earthquake action. <strong>The</strong> main failures of the full-scale models were pulling out of<br />

self-drilling screws, rupture of local plasterboards and detachment of ribbed corrugated sheets from each other;<br />

Full attention should be paid to strengthen window, door opening regions and increase reliability of self-drilling<br />

screw connections. Finally, combining with domestic and foreign research achievements, seismic design method<br />

for this kind of residential structures with light-gauge composite walls were proposed.<br />

Keywords: light-gauge composite wall, monotonic shear test, cyclic test, full-scale shaking table test, seismic<br />

design method<br />

一 、 前 言<br />

冷 弯 薄 壁 型 钢 龙 骨 房 屋 体 系 近 年 来 在 国 际 上 发 展 十 分 迅 速 , 屈 服 强 度 345MPa 以 上 、 厚 度 2mm<br />

以 下 的 高 强 超 薄 钢 板 作 为 一 种 新 材 料 , 加 工 成 型 的 冷 弯 型 钢 在 北 美 、 日 本 、 澳 洲 等 发 达 国 家 和 地 区<br />

被 广 泛 应 用 于 低 多 层 钢 结 构 住 宅 等 建 筑 中 。 目 前 , 这 类 冷 弯 超 薄 壁 型 钢 在 中 国 大 陆 的 生 产 和 使 用 才<br />

刚 刚 起 步 , 但 市 场 应 用 前 景 广 阔 。 但 现 行 国 家 标 准 《 冷 弯 薄 壁 型 钢 设 计 规 范 》 [1] 仅 适 用 于 承 重 构 件<br />

板 材 厚 度 2mm~6mm 的 Q235 及 Q345 钢 材 , 对 强 度 更 高 且 厚 度 在 2mm 以 下 的 超 薄 壁 冷 弯 型 钢 承 重<br />

构 件 的 设 计 方 法 尚 无 明 确 条 文 可 依 。 同 时 , 中 国 大 陆 地 区 都 有 抗 震 设 防 的 要 求 , 考 虑 到 冷 弯 薄 壁 型<br />

钢 板 材 的 高 强 度 、 低 延 性 特 点 , 在 国 内 进 一 步 推 广 应 用 时 整 体 结 构 的 抗 震 性 能 是 一 个 关 键 性 问 题 。<br />

国 外 已 开 展 了 对 冷 弯 薄 壁 型 钢 结 构 体 系 抗 震 性 能 的 相 关 研 究 , 文 献 [2][3] 对 冷 弯 薄 壁 型 钢 抗 剪 墙<br />

基 金 项 目 : 国 家 自 然 科 学 基 金 资 助 项 目 (50878168)<br />

作 者 简 介 : 沈 祖 炎 (1935-), 男 , 浙 江 杭 州 人 , 同 济 大 学 教 授 , 中 国 工 程 院 院 士 , 从 事 钢 结 构 的 教 学 、 科 研 和 工 程 实 践 工 作 。<br />

-70-


体 在 单 调 和 低 周 反 复 荷 载 作 用 下 的 性 能 进 行 了 试 验 研 究 和 数 值 模 拟 理 论 分 析 , 文 献 [4] 进 行 了 一 个 两<br />

层 冷 弯 薄 壁 型 钢 墙 体 结 构 的 足 尺 模 型 振 动 台 试 验 , 模 型 的 抗 剪 墙 体 采 用 密 柱 带 斜 拉 条 的 形 式 , 但 没<br />

有 外 覆 面 板 , 钢 龙 骨 材 料 的 屈 服 强 度 为 300MPa。 随 着 冷 弯 薄 壁 型 钢 结 构 体 系 逐 步 引 进 国 内 , 文 献 [5-7]<br />

也 进 行 了 龙 骨 式 复 合 墙 体 抗 剪 性 能 的 试 验 研 究 , 文 献 [8]、[9] 进 行 了 开 洞 复 合 墙 体 抗 剪 性 能 的 试 验 研<br />

究 和 理 论 分 析 。 纵 观 国 内 外 研 究 现 状 , 对 于 结 构 体 系 抗 震 性 能 的 研 究 还 局 限 在 抗 剪 墙 体 的 层 面 , 关<br />

于 整 体 结 构 抗 震 性 能 和 分 析 方 法 的 研 究 较 少 , 足 尺 模 型 的 振 动 台 试 验 更 是 尚 未 见 报 道 。<br />

本 文 首 次 完 成 了 三 栋 两 层 冷 弯 薄 壁 型 钢 住 宅 结 构 的 足 尺 模 型 振 动 台 试 验 , 龙 骨 体 系 模 型 I 采 用<br />

屈 服 强 度 550MPa 钢 材 , 模 型 II 和 III 采 用 S350 冷 轧 板 。 复 合 墙 体 外 墙 的 外 覆 面 板 分 别 采 用 OSB 板<br />

和 带 肋 波 纹 钢 板 , 内 覆 面 板 采 用 石 膏 板 ; 内 墙 内 外 覆 面 板 均 为 石 膏 板 。 试 验 选 取 3 条 实 测 地 震 波 记<br />

录 和 1 条 人 工 合 成 波 , 加 载 工 况 考 虑 从 7 度 多 遇 到 9 度 罕 遇 水 平 。 基 于 振 动 台 试 验 , 研 究 了 结 构 的<br />

动 力 特 性 和 在 强 震 作 用 下 的 力 学 性 能 , 考 察 此 类 结 构 体 系 的 抗 震 能 力 , 以 检 验 其 在 不 同 抗 震 设 防 标<br />

准 地 区 的 适 用 性 。 最 后 基 于 试 验 结 果 , 对 这 类 结 构 的 抗 震 设 计 方 法 进 行 了 探 讨 。<br />

二 、 试 验 模 型<br />

2.1. 模 型 设 计 和 制 作<br />

为 了 尽 可 能 准 确 模 拟 结 构 在 地 震 作 用 下 的 真 实 响 应 , 试 验 采 用 了 足 尺 模 型 , 模 型 试 验 分 别 在 中<br />

国 建 筑 科 学 研 究 院 抗 震 试 验 室 和 同 济 大 学 土 木 工 程 防 灾 国 家 重 点 实 验 室 的 振 动 台 上 进 行 。 模 型 的 平<br />

面 尺 寸 为 4m×6m, 高 度 6.915m, 图 1 为 模 型 的 平 、 立 面 , 结 构 的 龙 骨 构 件 采 用 博 思 格 和 钢 之 杰 公 司<br />

的 产 品 。 龙 骨 式 复 合 墙 体 采 用 双 面 覆 板 形 式 , 外 墙 内 覆 12mm 厚 石 膏 板 , 外 覆 0.42mm 厚 带 肋 波 纹<br />

板 ; 内 隔 墙 双 面 均 覆 12mm 厚 石 膏 板 。 模 型 墙 体 龙 骨 、 楼 面 梁 、 覆 面 板 等 材 料 及 型 号 见 表 1。 模 型<br />

结 构 主 要 承 重 构 件 龙 骨 柱 、 梁 和 屋 架 杆 件 的 截 面 壁 厚 均 在 2mm 以 下 , 超 出 了 现 行 国 家 规 范 《 冷 弯<br />

薄 壁 型 钢 结 构 技 术 规 范 》 [1] 对 构 件 壁 厚 的 限 制 规 定 。<br />

C<br />

C<br />

B<br />

A<br />

A<br />

1 2 1 2<br />

(a) 平 面 图<br />

图 1 振 动 台 模 型 尺 寸<br />

(b) 立 面 图<br />

根 据 建 筑 结 构 荷 载 规 范 (GB50009-2001) [10] , 住 宅 楼 面 活 荷 载 取 2.0kN/m 2 ; 根 据 建 筑 抗 震 设 计<br />

规 范 (GB50011-2001) [11] , 在 抗 震 分 析 时 , 结 构 重 力 荷 载 代 表 值 计 算 中 楼 面 活 荷 载 的 组 合 值 系 数 取<br />

0.5, 所 以 本 试 验 施 加 楼 面 配 重 为 100kg/m 2 。 施 工 完 毕 后 的 模 型 在 振 动 台 面 上 如 图 2 所 示 。<br />

表 1 三 个 足 尺 模 型 材 料 情 况<br />

屋 架<br />

外 墙 覆 面 板<br />

模 型 编 号 龙 骨 材 料 墙 体 立 柱 楼 面 梁<br />

弦 杆 腹 杆 外 侧 内 侧<br />

模 型 I G550 C7575 C20019 C7510 C7575 波 纹 板<br />

模 型 II<br />

OSB 板 石 膏 板<br />

S350 U9008 C25019 U9008 C7008<br />

模 型 III<br />

波 纹 板<br />

内 墙 两 侧<br />

覆 面 板<br />

石 膏 板<br />

-71-


(a) 模 型 I (b) 模 型 II (c) 模 型 III (d) 楼 面 配 重<br />

图 2 振 动 台 模 型 照 片<br />

2.2. 传 感 器 布 置<br />

在 模 型 底 座 、 二 层 楼 面 和 二 层 屋 顶 布 置 加 速 度 传 感 器 , 以 分 别 测 试 实 际 输 入 的 地 震 波 激 励 、 楼<br />

面 以 及 屋 架 的 加 速 度 , 布 置 位 置 如 图 3 所 示 。 在 二 层 楼 面 布 置 两 个 Y 向 的 位 移 计 , 通 过 其 位 移 的 差<br />

值 来 测 试 整 体 结 构 的 扭 转 效 应 。<br />

模 型 底 座<br />

±0.000 地 面<br />

+3.000 混 凝 土 楼 面<br />

3.000<br />

truss003<br />

truss002<br />

结 构 底 导 槽<br />

truss001<br />

±0.000<br />

(a) 传 感 器 布 置 1 (b) 传 感 器 布 置 2<br />

图 3 加 速 度 传 感 器 布 置 ( 模 型 I, 其 他 同 )<br />

按 从 下 至 上 的 顺 序 , 依 次 对 结 构 底 层 、 二 层 、 楼 面 梁 和 屋 架 各 部 分 进 行 应 变 片 测 点 布 置 。 模 型<br />

整 体 在 水 平 两 个 方 向 是 基 本 对 称 的 , 布 置 时 偏 重 于 平 面 的 1/2 区 域 , 应 变 片 布 置 详 见 图 4。 图 4(a)<br />

为 应 变 片 测 点 分 布 的 平 面 示 意 , 图 4(b) 为 典 型 墙 体 龙 骨 立 面 的 应 变 片 布 置 , 立 柱 每 个 测 点 截 面 布<br />

置 3 个 应 变 片 ( 上 下 翼 缘 及 腹 板 中 部 ), 楼 面 梁 每 个 测 点 截 面 只 布 置 1 个 应 变 片 于 下 翼 缘 。 应 变 传<br />

感 器 结 构 底 层 布 置 了 10 个 测 点 , 共 24 片 ; 二 层 布 置 了 4 个 测 点 , 共 12 片 ; 楼 面 梁 布 置 了 2 个 测<br />

点 , 共 2 片 ; 屋 架 布 置 了 5 个 测 点 , 共 11 片 ; 合 计 21 个 测 点 ,49 个 应 变 片 。<br />

2.3. 加 载 方 案<br />

试 验 选 取 El.Centro 波 、 唐 山 迁 安 波 ( 图 5a)、 北 京 波 ( 唐 山 地 震 北 京 旅 馆 的 实 测 记 录 )3 条 实<br />

测 地 震 记 录 和 1 条 上 海 人 工 波 ( 图 5b), 考 虑 的 抗 震 设 防 等 级 从 7 度 多 遇 到 9 度 罕 遇 , 按 照 加 速 度<br />

峰 值 从 小 到 大 的 顺 序 加 载 。 在 每 个 荷 载 级 工 况 结 束 之 后 , 均 进 行 一 次 白 噪 声 扫 频 , 以 检 测 结 构 的 刚<br />

度 退 化 情 况 ; 在 烈 度 为 9 度 罕 遇 荷 载 级 的 第 一 个 工 况 (74) 结 束 后 , 也 进 行 了 一 次 白 噪 声 扫 频 ( 工<br />

况 74X), 目 的 是 判 断 模 型 结 构 的 安 全 状 况 以 及 试 验 工 况 继 续 进 行 的 可 能 性 。<br />

2<br />

F1S7-S9<br />

7<br />

1<br />

2<br />

6<br />

1<br />

2800<br />

7 2<br />

F1S4-S6<br />

F1S1-S3<br />

F2S10-S12<br />

F2S7-S9<br />

F2S4-S6<br />

F1S23<br />

F1S19-S21<br />

F2S1-S3<br />

F1S22<br />

F1S13-S15<br />

F1S1-S3<br />

F1S4-S6 F1S7-S9<br />

F1S16-S18<br />

F1S10-S12<br />

3<br />

8<br />

4 3<br />

5<br />

(a) 应 变 片 测 点 布 置 平 面<br />

(b) 典 型 墙 体 立 面 的 测 点 布 置<br />

图 4 应 变 片 布 置 ( 模 型 I, 其 他 两 个 模 型 类 似 )<br />

-72-


表 2 为 本 试 验 设 计 的 加 载 工 况 , 由 于 9 度 中 震 (0.40g) 和 8 度 罕 遇 地 震 (0.41g) 加 速 度 峰 值 接 近 ,<br />

试 验 过 程 中 取 消 9 度 中 震 相 应 的 加 载 工 况 (56-64)。 试 验 采 用 的 部 分 地 震 波 时 程 如 图 5 所 示 , 在 此 只<br />

列 出 了 典 型 的 唐 山 波 和 上 海 人 工 波 。<br />

表 2 振 动 台 试 验 加 载 工 况<br />

试 验 工<br />

地 震 输 入 值 (gal)<br />

工 况 名 称 烈 度 地 震 激 励<br />

况 序 号<br />

主 振 方 向 X 方 向 Y 方 向<br />

备 注<br />

1 W1 第 一 次 白 噪 声 30 30 双 向<br />

2 S7ELXY X 35 30<br />

El. Centro<br />

3 S7ELYX<br />

Y 30 35<br />

双 向<br />

4 S7TSXY X 35 30<br />

唐 山 波<br />

双 向<br />

5 S7TSYX<br />

Y 30 35<br />

7 度 多 遇<br />

6 S7BJXY X 35 30<br />

北 京 波<br />

双 向<br />

7 S7BJYX<br />

Y 30 35<br />

8 S7SHX X 35<br />

上 海 波<br />

9 S7SHY<br />

Y 35<br />

单 向<br />

10 W2 第 二 次 白 噪 声 30 30 双 向<br />

11 S8ELXY X 70 60<br />

El. Centro<br />

12 S8ELYX<br />

Y 60 70<br />

双 向<br />

13 S8TSXY X 70 60<br />

唐 山 波<br />

双 向<br />

14 S8TSYX<br />

Y 60 70<br />

8 度 多 遇<br />

15 S8BJXY X 70 60<br />

北 京 波<br />

双 向<br />

16 S8BJYX<br />

Y 60 70<br />

17 S8SHX X 70<br />

上 海 波<br />

18 S8SHY<br />

Y 70<br />

单 向<br />

19<br />

……<br />

W3<br />

……<br />

第 三 次 白 噪 声<br />

…… ……<br />

30<br />

……<br />

30<br />

……<br />

双 向<br />

……<br />

73 W9 第 九 次 白 噪 声 30 30 双 向<br />

74 G9ELXY 9 度 罕 遇 El.Centro X 620 527 双 向<br />

74X W10 第 十 次 白 噪 声 30 30 双 向<br />

75 G9ELYX El.Centro Y 527 620 双 向<br />

76 G9TSXY X 620 527<br />

唐 山 波<br />

77 G9TSYX<br />

Y 527 620<br />

双 向<br />

78 G9BJXY 9 度 罕 遇<br />

X 620 527<br />

北 京 波<br />

79 G9BJYX<br />

Y 527 620<br />

双 向<br />

80 G9SHX X 620<br />

81 G9SHY<br />

上 海 波 Y 527<br />

双 向<br />

82 W11 第 十 一 次 白 噪 声 30 30 双 向<br />

加 速 度 [m/s 2 ]<br />

1.2<br />

0.9<br />

0.6<br />

0.3<br />

0.0<br />

-0.3<br />

-0.6<br />

-0.9<br />

唐 山 波<br />

-1.2<br />

0 5 10 15 20 25<br />

时 间 t [s]<br />

(a) 唐 山 迁 安 波<br />

加 速 度 [m/s 2 ]<br />

0.4<br />

0.3<br />

0.2<br />

0.1<br />

0.0<br />

-0.1<br />

-0.2<br />

图 5 地 震 波 时 程 曲 线<br />

上 海 波<br />

-0.3<br />

0 5 10 15 20 25 30 35 40<br />

时 间 [s]<br />

(b) 上 海 波<br />

三 、 试 验 现 象 简 介<br />

3.1. 自 攻 螺 钉 连 接 破 坏<br />

从 整 个 试 验 过 程 中 观 察 到 的 现 象 看 , 结 构 的 破 坏 主 要 发 生 在 自 攻 螺 钉 连 接 部 位 以 及 墙 体 的 开 洞<br />

区 域 , 而 主 体 钢 龙 骨 基 本 没 有 破 坏 。 图 6 为 模 型 I 石 膏 板 拼 接 位 置 处 的 连 接 破 坏 。 试 验 中 发 现 螺 钉<br />

连 接 的 破 坏 主 要 在 内 墙 石 膏 板 的 拼 接 部 位 , 说 明 拼 接 缝 隙 处 是 结 构 这 类 连 接 的 薄 弱 区 。 图 7 模 型 I<br />

所 示 为 外 墙 波 纹 板 与 钢 龙 骨 的 连 接 破 坏 , 在 结 构 的 一 角 部 区 域 , 外 墙 波 纹 板 连 同 自 攻 螺 钉 一 起 脱 离<br />

龙 骨 柱 。 但 外 墙 与 钢 龙 骨 的 连 接 在 其 它 区 域 几 乎 完 好 无 损 。<br />

-73-


(a) 水 平 接 缝 处 螺 钉 松 动 (b) 竖 直 接 缝 处 螺 钉 松 动<br />

图 6 石 膏 板 拼 接 位 置 的 连 接 破 坏 ( 模 型 I) 图 7 外 墙 角 部 波 纹 板 与 龙 骨 脱 离 ( 模 型 I)<br />

3.2. 石 膏 板 局 部 破 坏<br />

石 膏 板 的 破 坏 主 要 出 现 在 门 框 、 窗 框 角 点 位 置 , 以 及 墙 板 和 楼 板 交 界 面 位 置 , 图 8 为 模 型 I 试<br />

验 中 观 测 到 的 门 框 角 点 破 坏 现 象 。 图 9 模 型 I 是 窗 框 角 部 石 膏 板 的 破 坏 。 图 10 是 模 型 I 墙 板 和 楼 板<br />

交 界 面 位 置 石 膏 板 的 挤 压 破 坏 情 况 。 从 石 膏 板 的 破 坏 现 象 可 发 现 , 墙 体 开 洞 角 点 和 石 膏 板 拼 接 区 域 ,<br />

板 相 互 发 生 挤 压 , 使 其 受 力 复 杂 而 产 生 局 部 破 坏 或 拼 接 错 位 , 在 结 构 设 计 时 应 该 引 起 重 视 。 如 果 在<br />

拼 接 石 膏 板 之 间 预 留 一 定 空 隙 , 则 其 相 互 挤 压 作 用 会 大 大 减 少 , 能 尽 量 避 免 挤 压 破 坏 的 发 生 。<br />

(a) 左 上 角 水 平 拼 接 脱 开<br />

(b) 右 上 角 局 部 破 坏<br />

图 8 门 框 角 点 石 膏 板 破 坏 ( 模 型 I)<br />

(a) 交 界 处 挤 压 破 坏 (b) 拼 接 处 错 位<br />

图 9 窗 框 角 点 石 膏 板 破 坏 ( 模 型 I) 图 10 石 膏 板 交 界 处 破 坏 ( 模 型 I)<br />

3.3. 龙 骨 的 情 况<br />

试 验 完 成 后 , 为 了 解 墙 体 内 部 龙 骨 的 可 能 破 坏 情 况 , 拆 卸 了 部 分 关 键 区 域 的 石 膏 板 , 如 图 11<br />

所 示 。 发 现 墙 体 立 柱 无 明 显 的 失 稳 破 坏 , 截 面 无 畸 变 发 生 。 观 测 到 的 现 象 和 试 验 应 变 测 量 数 据 反 映<br />

出 的 结 果 一 致 。<br />

(a) 模 型 I 底 层 柱 (b) 模 型 I 二 层 柱 (c) 模 型 II 底 层 柱 (d) 模 型 II 二 层 柱 (e) 模 型 II 层 间 连 接 件<br />

图 11 试 验 后 内 部 钢 龙 骨 情 况<br />

-74-


四 、 试 验 数 据 分 析<br />

4.1. 模 型 动 力 特 性 变 化<br />

在 每 一 个 荷 载 级 工 况 结 束 后 , 都 进 行 了 白 噪 声 扫 频 , 对 结 构 的 动 力 特 性 进 行 识 别 。 表 3 为 结 构<br />

在 各 烈 度 等 级 的 地 震 荷 载 工 况 作 用 结 束 后 , 白 噪 声 的 扫 频 的 分 析 结 果 。 可 以 看 出 : 结 构 两 方 向 刚 度<br />

衰 减 的 速 率 明 显 不 一 致 ,9 度 罕 遇 烈 度 的 地 震 持 时 作 用 使 结 构 产 生 损 伤 累 积 , 刚 度 连 续 大 幅 度 下 降 ,<br />

但 结 构 没 有 出 现 倒 塌 , 说 明 该 结 构 能 够 满 足 大 震 不 倒 的 抗 震 设 防 要 求 。 总 体 上 讲 , 结 构 在 7 度 和 8<br />

度 罕 遇 烈 度 地 震 作 用 下 , 刚 度 下 降 幅 度 不 大 ;9 度 罕 遇 烈 度 地 震 作 用 , 由 于 地 震 力 增 大 使 自 攻 螺 钉<br />

脱 落 , 造 成 覆 面 板 与 龙 骨 脱 离 , 从 而 导 致 结 构 刚 度 迅 速 退 化 。 另 外 , 结 构 阻 尼 比 的 取 值 , 也 是 抗 震<br />

分 析 计 算 值 得 关 注 的 问 题 , 从 地 震 响 应 和 白 噪 声 扫 频 的 结 果 来 看 , 这 类 结 构 的 整 体 阻 尼 比 在 弹 性 阶<br />

段 可 建 议 取 值 0.03, 非 线 性 阶 段 取 值 0.05。<br />

工<br />

况<br />

号<br />

表 3 白 噪 声 扫 频 结 果<br />

模 型 I 模 型 II 模 型 III<br />

X 向 Y 向 X 向 Y 向 X 向 Y 向<br />

频 率 阻 尼 比 频 率 阻 尼 比 频 率 阻 尼 比 频 率 阻 尼 比 频 率 阻 尼 比 频 率 阻 尼 比<br />

(Hz) (%) (Hz) (%) (Hz) (%) (Hz) (%) (Hz) (%) (Hz) (%)<br />

1 9.72 2.62 7.82 2.67 7.67 2.84 6.30 3.18 5.27 3.25 6.79 2.52<br />

10 9.63 1.85 7.22 2.23 7.62 3.21 6.30 3.47 5.27 3.85 6.79 2.49<br />

19 9.72 3.59 7.28 2.23 7.52 3.13 6.30 3.93 5.27 3.95 6.79 2.23<br />

28 9.56 3.28 7.09 1.76 7.23 3.42 6.20 3.94 5.18 3.79 6.40 2.75<br />

73 9.28 3.04 6.72 2.17 7.13 3.54 4.74 4.40 4.35 7.13 5.66 4.99<br />

74X 9.25 4.46 6.28 2.07 6.79 5.47 4.15 5.10 3.22 8.82 5.27 7.57<br />

82 8.59 6.64 5.00 7.83 6.59 6.94 1.86 9.50 - - - -<br />

4.2. 加 速 度 和 位 移<br />

由 典 型 加 载 工 况 下 模 型 最 大 加 速 度 值 和 位 移 值 可 以 看 出 :(1) 在 各 工 况 下 加 速 度 A2Y1 与 A2Y2<br />

所 测 得 的 峰 值 差 异 不 大 , 说 明 结 构 质 量 和 刚 度 分 布 对 称 合 理 , 无 明 显 扭 转 趋 势 ;(2) 四 种 地 震 波 的<br />

激 励 作 用 比 较 , 相 同 烈 度 条 件 下 结 构 对 唐 山 波 加 速 度 反 应 大 一 点 , 这 与 地 震 波 自 身 频 谱 特 性 有 关 系 。<br />

新 编 国 家 行 业 标 准 《 低 层 冷 弯 薄 壁 型 钢 房 屋 建 筑 技 术 规 程 》 [12] 规 定 , 结 构 在 多 遇 地 震 下 的 层 间<br />

位 移 角 限 值 为 1/300, 罕 遇 地 震 下 的 层 间 位 移 角 限 值 为 1/100。 本 模 型 结 构 的 层 高 为 3m, 要 求 多 遇<br />

地 震 下 层 间 位 移 小 于 10mm, 罕 遇 地 震 下 层 间 位 移 小 于 30mm。 表 4 列 举 了 模 型 I 典 型 试 验 工 况 结 构<br />

最 大 层 间 位 移 , 可 见 在 试 验 各 工 况 下 , 层 间 位 移 均 满 足 规 范 限 值 要 求 , 结 构 刚 度 设 计 合 理 。 同 样 ,<br />

模 型 II、 模 型 III 试 验 测 到 的 各 工 况 下 层 间 位 移 也 均 满 足 规 范 限 值 要 求 。 限 于 篇 幅 , 本 文 不 再 赘 述 。<br />

4.3. 构 件 应 变 和 应 力 分 布<br />

通 过 采 集 到 的 钢 龙 骨 各 测 点 应 变 及 应 力 数 据 , 可 用 来 评 估 龙 骨 构 件 的 破 坏 情 况 。 本 文 限 于 篇 幅 ,<br />

表 5 只 列 举 了 模 型 I 结 构 在 部 分 典 型 工 况 (9 度 罕 遇 烈 度 ) 下 的 最 大 应 变 和 应 力 值 。 表 格 中 仅 列 举<br />

了 部 分 应 变 片 测 点 的 数 据 。 未 列 举 一 般 应 变 均 很 小 , 少 数 几 个 测 点 在 试 验 中 有 损 坏 。<br />

表 4 模 型 I 典 型 工 况 结 构 的 最 大 层 间 位 移<br />

层 间 位 移<br />

9 度 多 遇 工 况 9 度 罕 遇 工 况<br />

29 30 31 32 74 75 76 77<br />

底 层 X 向 1.41 1.57 0.85 1.07 10.12 8.56 3.73 3.59<br />

底 层 Y 向 0.23 0.27 1.03 0.93 3.36 7.22 11.13 12.98<br />

二 层 X 向 0.76 0.47 0.58 0.39 3.79 1.54 1.85 1.59<br />

二 层 Y 向 0.08 0.08 1.04 1.13 2.38 6.67 4.11 5.57<br />

-75-


表 5 模 型 I 在 9 度 罕 遇 工 况 下 测 点 最 大 应 变 和 应 力 值<br />

工 况 74 (El.Centro 波 X 主 向 ) 75 (El. Centro 波 Y 主 向 )<br />

测 点 号 应 变 ( ´<br />

6<br />

10 - )<br />

2<br />

应 力 ( N/mm )<br />

6<br />

应 变 ( ´ 10 - )<br />

2<br />

应 力 ( N/mm )<br />

FIS1 76 -130 15.979 -27.237 148 -79 31.020 -16.651<br />

FIS2 16 -76 3.259 -15.932 38 -64 7.929 -13.342<br />

FIS3 78 -88 16.480 -18.521 128 -106 26.913 -22.359<br />

FIS4 67 -51 14.143 -10.764 45 -65 9.473 -13.627<br />

FIS5 4 -51 0.855 -10.733 10 -67 2.123 -14.121<br />

FIS6 11 -64 2.316 -13.479 22 -69 4.700 -14.492<br />

FIS7 82 -118 17.165 -24.790 89 -102 18.727 -21.358<br />

FIS8 43 -60 9.074 -12.655 58 -66 12.119 -13.829<br />

FIS9 29 -52 6.095 -11.018 31 -48 6.529 -10.171<br />

FIS12 44 -125 9.309 -26.238 37 -237 7.728 -49.874<br />

FIS13 324 -67 68.110 -14.009 407 -124 85.571 -26.091<br />

FIS15 78 -99 16.301 -20.764 76 -124 15.903 -26.058<br />

FIS18 40 -45 8.455 -9.475 29 -56 6.092 -11.694<br />

FIS19 59 -56 12.365 -11.811 67 -69 13.987 -14.516<br />

FIS20 69 -39 14.580 -8.190 81 -31 16.932 -6.475<br />

FIS22 142 -315 29.777 -66.201 118 -374 24.682 -78.640<br />

FIS23 4 -61 0.941 -12.803 0 -59 0.000 -12.370<br />

F2S3 79 -57 16.557 -11.869 62 -49 12.955 -10.316<br />

F2S5 6 -61 1.283 -12.779 12 -45 2.612 -9.472<br />

F2S6 31 -10 6.532 -2.061 24 -10 5.095 -2.194<br />

F2S9 74 -75 15.515 -15.685 84 -88 17.657 -18.417<br />

F2S11 10 -25 2.164 -5.216 10 -15 2.156 -3.250<br />

F2S12 40 -32 8.480 -6.782 25 -27 5.246 -5.644<br />

FBS2 96 -47 20.070 -9.836 72 -36 15.157 -7.506<br />

TRS8 13 -17 2.718 -3.673 12 -23 2.539 -4.890<br />

从 表 5 中 可 见 :(1) 模 型 的 整 体 质 量 轻 , 即 使 在 高 烈 度 的 地 震 作 用 下 , 地 震 荷 载 引 起 龙 骨 结 构<br />

柱 的 应 力 水 平 也 不 大 。9 度 罕 遇 地 震 作 用 时 构 件 的 最 大 应 力 响 应 基 本 都 小 于 100N/mm 2 , 说 明 钢 龙 骨<br />

是 均 处 于 弹 性 工 作 阶 段 。 (2) 底 层 柱 的 应 力 平 均 水 平 高 于 二 层 柱 , 因 底 层 复 合 墙 体 承 受 两 层 结 构<br />

地 震 作 用 产 生 的 水 平 力 之 合 剪 力 。 测 点 F1S1、F1S12、F1S23 的 应 力 最 大 , 其 分 别 对 应 墙 体 边 柱 、<br />

墙 体 洞 口 柱 和 斜 拉 条 位 置 , 说 明 边 柱 和 洞 口 柱 是 结 构 不 利 的 受 力 位 置 。(3) 拉 条 FIS22、FIS23 有<br />

明 显 的 应 变 反 应 , 说 明 当 墙 板 的 蒙 皮 效 应 减 弱 时 , 拉 条 对 提 高 龙 骨 式 复 合 墙 体 的 抗 侧 刚 度 将 有 明 显<br />

贡 献 , 其 中 斜 拉 条 的 作 用 更 加 明 显 。 (4) 对 比 白 噪 声 扫 频 结 构 刚 度 的 退 化 现 象 和 应 变 片 反 映 的 钢 龙<br />

骨 应 力 分 布 情 况 来 看 , 此 结 构 体 系 墙 板 的 蒙 皮 作 用 是 影 响 其 水 平 抗 力 的 关 键 , 虽 然 应 变 片 反 映 出 地<br />

震 荷 载 引 起 的 钢 龙 骨 构 件 应 力 小 于 材 料 的 屈 服 强 度 , 结 构 柱 处 于 弹 性 工 作 阶 段 , 但 整 体 水 平 刚 度 却<br />

发 生 了 明 显 退 化 , 主 要 原 因 在 于 结 构 局 部 外 覆 墙 板 与 龙 骨 立 柱 之 间 连 接 破 坏 , 蒙 皮 作 用 不 断 减 弱 。<br />

五 、 抗 震 设 计 方 法 建 议<br />

低 层 冷 弯 薄 壁 型 钢 房 屋 是 由 龙 骨 式 复 合 墙 板 组 成 的 “ 盒 子 ” 式 结 构 , 水 平 风 荷 载 或 水 平 地 震 作<br />

用 应 由 抗 剪 墙 体 承 担 。 本 文 基 于 振 动 台 试 验 , 通 过 对 结 构 在 多 遇 和 罕 遇 地 震 作 用 下 不 同 分 析 方 法 的<br />

对 比 研 究 , 结 合 《 低 层 冷 弯 薄 壁 型 钢 房 屋 建 筑 技 术 规 程 》 [12] ( 报 批 稿 ), 对 抗 震 设 计 方 法 建 议 如 下 :<br />

(1) 建 筑 结 构 系 统 宜 规 则 布 置 , 以 形 成 水 平 和 垂 直 抗 侧 力 系 统 。 抗 剪 墙 体 应 布 置 在 建 筑 结 构 的 两 个<br />

主 方 向 , 形 成 明 确 的 抗 风 和 抗 震 体 系 。 (2) 多 遇 地 震 下 水 平 地 震 对 结 构 的 作 用 , 可 采 用 底 部 剪 力 法 、<br />

振 型 分 解 反 应 谱 法 进 行 计 算 , 在 建 筑 结 构 的 两 个 主 轴 方 向 分 别 计 算 水 平 荷 载 的 作 用 。 每 个 主 轴 方 向<br />

的 水 平 荷 载 应 由 该 方 向 抗 剪 墙 体 承 担 , 可 根 据 其 抗 剪 刚 度 大 小 按 比 例 分 配 , 并 应 考 虑 门 窗 洞 口 对 墙<br />

体 抗 剪 刚 度 的 削 弱 作 用 。(3) 冷 弯 薄 壁 型 钢 复 合 墙 体 结 构 具 有 很 好 的 延 性 和 抗 震 性 能 , 墙 体 结 构 在<br />

动 力 试 验 中 表 现 出 的 抗 剪 承 载 能 力 和 静 力 加 载 试 验 吻 合 得 较 好 。 由 于 低 层 冷 弯 薄 壁 型 钢 房 屋 建 筑 的<br />

自 重 很 轻 , 地 震 作 用 对 其 影 响 不 明 显 。 振 动 台 试 验 的 结 果 表 明 , 对 于 体 型 规 则 的 房 屋 建 筑 , 在 结 构<br />

罕 遇 地 震 的 计 算 方 法 尚 不 完 善 时 , 结 构 在 多 遇 地 震 作 用 下 进 行 弹 性 设 计 后 , 采 取 合 理 的 墙 体 构 造 措<br />

施 , 能 够 满 足 罕 遇 地 震 作 用 下 的 抗 震 设 防 要 求 。(4) 特 殊 情 况 下 , 低 层 冷 弯 薄 壁 型 钢 结 构 体 系 , 对<br />

于 有 明 显 扭 转 、 体 型 复 杂 、 刚 度 和 质 量 分 布 不 均 匀 的 建 筑 , 必 须 建 立 实 际 结 构 的 空 间 杆 系 简 化 模 型 ,<br />

-76-


进 行 双 向 地 震 激 励 的 工 况 分 析 , 计 算 各 片 复 合 抗 剪 墙 体 实 际 承 受 的 水 平 剪 力 作 用 , 进 行 复 合 抗 剪 墙<br />

体 的 结 构 设 计 。<br />

六 、 结 论 和 建 议<br />

通 过 三 栋 两 层 冷 弯 薄 壁 型 钢 住 宅 结 构 的 足 尺 模 型 振 动 台 试 验 和 理 论 分 析 , 可 得 如 下 的 结 论 :<br />

(1) 总 体 而 言 , 采 用 此 类 双 面 覆 板 墙 体 构 造 形 式 的 结 构 能 够 符 合 7 度 、8 度 甚 至 更 高 抗 震 设 防<br />

地 区 (9 度 ) 小 震 不 坏 、 中 震 可 修 、 大 震 不 倒 抗 震 设 防 要 求 。<br />

(2) 墙 体 开 洞 部 位 ( 门 、 窗 口 ) 为 整 个 结 构 薄 弱 区 域 。 石 膏 板 由 于 其 脆 性 材 料 性 质 , 在 洞 口<br />

角 部 容 易 发 生 应 力 集 中 而 破 坏 。 虽 然 这 种 破 坏 只 是 局 部 性 的 , 但 设 计 时 要 充 分 关 注 门 、 窗 部 位 局 部<br />

加 强 以 及 自 攻 螺 钉 连 接 的 可 靠 性 , 必 要 时 自 攻 螺 钉 间 距 要 合 理 加 密 , 拼 接 石 膏 板 间 应 预 留 一 定 空 隙 。<br />

(3) 结 构 的 刚 度 满 足 抗 震 设 防 的 要 求 , 多 遇 地 震 和 罕 遇 地 震 作 用 下 的 层 间 位 移 角 均 小 于 规 范<br />

限 值 ; 结 构 的 质 量 和 刚 度 分 布 均 匀 , 没 有 出 现 明 显 的 扭 转 。<br />

(4) 抗 震 分 析 时 , 这 类 结 构 的 整 体 阻 尼 比 可 采 用 在 弹 性 阶 段 取 0.03、 非 线 性 阶 段 取 0.05。<br />

(5) 对 低 层 冷 弯 薄 壁 型 钢 房 屋 抗 震 设 计 , 多 遇 地 震 作 用 下 可 采 用 底 部 剪 力 法 进 行 地 震 力 计 算 ,<br />

并 按 各 主 方 向 上 有 效 抗 剪 墙 体 的 抗 剪 刚 度 大 小 按 比 例 分 配 该 层 的 地 震 力 。 在 罕 遇 地 震 作 用 下 , 由 于<br />

房 屋 一 般 自 重 较 轻 , 结 构 受 到 的 水 平 作 用 力 相 对 不 大 , 龙 骨 式 复 合 墙 体 进 入 非 线 性 的 程 度 并 不 严 重 ,<br />

可 以 实 现 房 屋 结 构 “ 大 震 不 倒 ” 的 抗 震 设 防 目 标 , 一 般 可 以 不 做 验 算 , 但 宜 根 据 抗 震 设 防 等 级 加<br />

强 墙 体 与 基 础 及 上 下 墙 体 之 间 的 相 应 连 接 构 造 。<br />

参 考 文 献<br />

[1] GB50018-2002, 冷 弯 薄 壁 型 钢 结 构 技 术 规 范 .<br />

[2] L.A.Fülöp, D.Dubina. Performance of wall-stud cold-formed shear panels under monotonic and cyclic loading. Part I:<br />

Experimental research. Thin-Walled Structures, 2004, 42: 321-338.<br />

[3] L.A.Fülöp, D.Dubina. Performance of wall-stud cold-formed shear panels under monotonic and cyclic loading. Part II:<br />

Numerical modelling and performance analysis.Thin-Walled Structures, 2004, 42: 339-349.<br />

[4] Tae-Wan Kim, James Wilcoski. Shake table tests of a cold-formed steel shear panel. Engineering Structures, 2006,<br />

28:1462-1470.<br />

[5] 周 天 华 , 石 宇 , 何 保 康 等 . 冷 弯 型 钢 复 合 墙 体 抗 剪 承 载 力 试 验 研 究 . 西 安 建 筑 科 技 大 学 学 报 , 2006, 38 (1): 83-88.<br />

[6] 聂 少 锋 . 冷 弯 型 钢 立 柱 复 合 墙 体 抗 剪 承 载 力 简 化 计 算 方 法 研 究 . 长 安 大 学 , 2006, 西 安 .<br />

[7] 郭 鹏 . 冷 弯 型 钢 骨 架 墙 体 抗 剪 性 能 试 验 与 理 论 研 究 . 西 安 建 筑 科 技 大 学 , 2008, 西 安 .<br />

[8] 熊 智 刚 . 冷 弯 薄 壁 型 钢 结 构 住 宅 开 洞 复 合 墙 体 抗 剪 性 能 研 究 . 长 安 大 学 , 2008, 西 安 .<br />

[9] 李 斌 . 开 门 窗 洞 口 的 冷 弯 薄 壁 型 钢 复 合 墙 体 抗 剪 性 能 . 苏 州 科 技 学 院 , 2008, 苏 州 .<br />

[10] GB50009-2001, 建 筑 结 构 荷 载 规 范 .<br />

[11] GB50011-2001, 建 筑 抗 震 设 计 规 范 .<br />

[12] JGJ 227-2011, 低 层 冷 弯 薄 壁 型 钢 房 屋 建 筑 技 术 规 程 ( 报 批 稿 ).<br />

-77-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

两 种 新 型 超 大 跨 度 (150m×150m) 部 分 预 应 力 空 间 钢 网 格 结 构<br />

1,2 1,2<br />

2<br />

1<br />

马 克 俭 申 波 徐 向 东 冯 更 帅<br />

(1 贵 州 大 学 空 间 结 构 研 究 中 心 , 贵 州 贵 阳 550003;2 湖 南 大 学 土 木 工 程 学 院 湖 南 长 沙 410000)<br />

摘 要 : 随 着 经 济 建 设 的 高 速 发 展 , 大 跨 度 或 超 大 跨 度 (l≥120m) 的 公 共 与 工 业 建 筑 , 如 大 型 体<br />

育 建 筑 , 大 型 展 厅 、 大 型 机 库 及 干 煤 棚 等 日 趋 增 多 , 如 何 使 这 些 超 大 跨 度 钢 结 构 屋 盖 达 到 “ 安 全 ,<br />

合 理 , 先 进 , 经 济 ” 的 结 构 设 计 原 则 , 我 们 针 对 将 兴 建 的 某 大 型 试 验 厅 , 其 平 面 为 方 形 l x =l y =150m,<br />

屋 盖 覆 盖 面 积 22500m 2 , 提 出 两 种 新 型 超 大 跨 度 预 应 力 空 间 钢 网 格 结 构 , 它 们 分 别 为 超 大 跨 度 预 应<br />

力 正 交 空 间 管 桁 架 钢 网 格 结 构 和 超 大 跨 度 周 边 简 支 承 十 字 形 三 层 与 双 层 组 合 预 应 力 扭 网 壳 结 构 , 通<br />

过 静 力 与 动 力 分 析 及 配 杆 设 计 , 达 到 了 上 述 结 构 设 计 的 要 求 。<br />

关 键 词 : 超 大 跨 度 (L≥120m) 部 分 预 应 力 空 间 钢 网 格 结 构<br />

TWO KINDS OF NEW PARTIAL PRE-STRESSED SPACE STEEL GRID<br />

STRUCTURE WITH SUPER LARGE SPAN(150M×150M)<br />

Ma Ke-jian 1,2 , Shen bo 1,2 , Feng Geng-shuai 1<br />

1. Space Structure Research Center, Guizhou <strong>University</strong>, Guiyang Guizhou 550003;<br />

2. School of Civil Engineering, Hunan <strong>University</strong>, Changsha Huan 410000<br />

Abstract: Many of the public and industrial buildings with large and super large span are constructed<br />

under the rapid economic development. <strong>The</strong> span of the buildings, such as large sport buildings, large<br />

exhibition halls, large hangars and dry coal sheds, is larger than 120 meters. It is designing principle that<br />

the steel grid roof structure of the buildings should be safe, reasonable, advanced and economical. A large<br />

experiment hall with 150 meters long and 150 meters wide will be built. <strong>The</strong> area of the hall’s roof is<br />

22500 square meters. Two kinds of new pre-stressed space steel grid structure with super large span are<br />

proposed. <strong>The</strong> first one is orthogonal space tube truss steel grid structure. <strong>The</strong> second one is composite<br />

twisted latticed shell structure of cross three-layer and two-layer, constrained by simply supported around.<br />

<strong>The</strong> two structures are analyzed by static and dynamic methods. <strong>The</strong> design of the two structures meets the<br />

designing principle.<br />

Keywords: Super large span(larger than 120 meters), partial pre-stressed, space steel grid structure.<br />

一 、 前 言<br />

1.1. 大 跨 度 与 超 大 跨 度 公 共 建 筑 应 使 建 筑 设 计 与 结 构 设 计 和 谐 统 一<br />

近 二 十 多 年 来 , 中 国 大 陆 大 型 公 共 建 筑 , 如 大 型 体 育 馆 、 大 型 高 耸 建 筑 , 在 全 国 各 地 如 雨 后 春<br />

笋 一 般 迅 速 出 现 , 这 些 大 型 或 超 大 型 公 共 建 筑 , 其 主 体 分 建 筑 设 计 和 结 构 设 计 两 大 环 节 , 此 两 部 分<br />

既 有 各 自 的 要 求 , 也 有 科 学 的 配 合 使 之 溶 为 一 体 , 从 而 符 合 “ 低 碳 经 济 ” 的 基 本 国 策 基 础 上 体 现 出<br />

现 代 建 筑 造 型 风 格 , 而 结 构 设 计 是 在 建 筑 设 计 的 基 础 上 ,“ 安 全 、 合 理 、 先 进 、 经 济 ” 地 选 择 建 筑<br />

物 的 结 构 体 系 , 作 到 两 者 和 谐 统 一 , 即 建 筑 工 程 界 提 出 的 “ 技 术 与 艺 术 的 完 美 结 合 ” 和 “ 力 学 与 美<br />

学 的 协 调 统 一 ”, 二 十 多 年 来 , 很 多 大 型 公 共 建 筑 都 作 到 两 者 的 和 谐 统 一 。 这 样 的 工 程 实 例 比 比 皆<br />

是 , 不 胜 枚 举 , 它 们 的 共 同 特 点 是 优 良 的 建 筑 功 能 , 合 理 、 先 进 的 结 构 体 系 , 并 在 较 好 的 技 术 经 济<br />

指 标 前 提 下 , 体 现 了 现 代 建 筑 的 风 格 和 造 型 , 使 人 们 产 生 美 的 感 受 。 但 现 代 建 筑 的 风 格 和 造 型 与 那<br />

些 违 背 基 本 力 学 规 律 , 只 强 调 “ 标 新 立 异 ”、“ 独 树 一 格 ” 的 建 筑 也 不 能 相 提 并 论 。<br />

作 者 简 介 : 马 克 俭 (1933-), 男 , 湖 南 岳 阳 人 , 教 授 , 博 导 , 中 国 工 程 院 院 士 , 从 事 空 间 结 构 的 教 学 、 科 研 和 工 程 实 践 工 作 。<br />

-78-


图 1 超 大 跨 度 站 房 单 层 柱 面 网 壳 建 筑 效 果 图<br />

500<br />

3500<br />

X 2<br />

X2<br />

3500<br />

500<br />

X 1<br />

X 1<br />

图 2a 网 壳 顶 部 箱 形 构 件 截 面 形 式<br />

图 2b 网 壳 柱 脚 箱 形 构 件 截 面 形 式<br />

-M M<br />

-M<br />

37000<br />

N<br />

N<br />

128000<br />

128000<br />

128000<br />

图 3a 两 铰 拱 简 图 图 3b 两 铰 拱 弯 矩 图 图 3c 两 铰 拱 轴 力 图<br />

随 着 经 济 建 设 的 发 展 , 中 国 大 陆 的 国 力 迅 速 增 强 , 长 期 以 来 中 国 人 民 固 有 的 “ 勤 俭 节 约 ” 美 德 ,<br />

在 某 些 人 的 脑 海 里 也 逐 渐 淡 薄 , 为 了 标 新 立 异 , 达 到 “ 世 界 第 一 ” 不 惜 花 费 巨 大 的 财 力 和 物 力 。 最<br />

典 型 的 工 程 如 “ 鸟 巢 ”, 其 投 影 面 积 用 钢 量 达 世 界 之 最 (650kg/m2), 北 京 人 称 之 为 “ 裤 衩 ” 的 新 建<br />

中 央 电 视 台 , 其 用 钢 量 比 “ 鸟 巢 ” 有 过 之 而 无 不 及 , 这 些 建 筑 单 纯 追 求 奇 异 的 建 筑 造 型 , 而 忽 视 结<br />

构 体 系 的 合 理 性 而 造 成 的 。 这 种 思 想 在 一 部 分 管 理 者 脑 海 里 仍 然 存 在 。2010 年 作 者 主 持 某 高 铁 站 房<br />

超 大 跨 度 钢 结 构 施 工 图 审 查 , 该 站 房 建 筑 设 计 方 为 德 国 GMP 建 筑 事 物 所 , 结 构 横 向 跨 度<br />

ly=128m(ly>120m) 纵 向 长 度 lx=367m 矢 高 f=37m, 矢 跨 比 f/l=1/3.46 投 影 面 积 46976m2, 采 用 由 钢 板<br />

焊 接 的 变 截 面 箱 形 拱 斜 交 组 成 超 大 跨 度 单 层 大 网 格 柱 面 网 壳 , 其 菱 形 大 网 格 沿 跨 度 方 向 (l y =128m) 分<br />

为 4 格 , 每 格 投 影 长 度 32m, 沿 纵 向 分 40 格 , 每 格 长 度 约 10m 左 右 , 矩 形 钢 管 檩 条 沿 纵 向 设 置 ,<br />

建 筑 设 计 者 单 纯 追 求 造 型 和 美 学 效 果 , 完 全 不 考 虑 如 此 巨 大 的 单 层 柱 面 网 壳 结 构 的 力 学 体 系 合 理<br />

性 , 图 1 为 该 建 筑 室 内 大 厅 效 果 图 , 从 图 1 可 知 其 顶 部 钢 箱 水 平 搁 置 , 如 图 2a 所 示 , 由 跨 中 向 两 侧<br />

至 网 壳 支 座 钢 箱 截 面 由 平 放 转 90 度 成 为 竖 立 , 如 图 2b 所 示 , 结 构 箱 形 构 件 由 跨 中 向 两 边 支 座 长 度<br />

64m 范 围 内 , 由 平 放 截 面 ( 图 2a) 转 变 为 竖 立 截 面 ( 图 2b), 不 考 虑 其 施 工 的 可 操 作 性 和 带 来 的 难 度 ,<br />

仅 从 其 结 构 体 系 的 合 理 性 和 结 构 力 学 基 本 规 律 来 讨 论 建 筑 设 计 与 结 构 设 计 不 和 谐 统 一 所 带 来 的 后<br />

果 。 图 3a 为 l=128m 两 铰 拱 , 它 在 均 布 荷 载 作 用 下 计 算 简 图 , 图 3b 为 其 弯 矩 分 布 布 置 图 , 图 3c 为<br />

其 轴 向 压 力 分 布 图 。 从 内 力 分 布 图 可 知 , 跨 中 正 弯 矩 最 大 , 轴 向 压 力 最 小 或 等 于 零 , 按 强 度 要 求 配<br />

杆 时 , 要 满 足 拱 顶 使 用 荷 载 下 钢 箱 下 部 最 大 拉 应 力 小 于 其 设 计 拉 应 力 , 即 f=M/w x1 ≤f c , 从 图 2a, 2b<br />

可 知 , 拱 顶 与 拱 脚 钢 箱 截 面 面 积 相 等 条 件 下 , 拱 脚 截 面 其 断 面 矩 w x2 =22w x1 , 仅 从 强 度 要 求 分 析 , 在<br />

截 面 形 式 不 变 条 件 下 , 唯 一 办 法 是 加 大 拱 顶 平 放 箱 形 截 面 上 、 下 板 厚 度 , 经 分 析 要 满 足 拱 强 度 要 求 ,<br />

拱 顶 箱 形 截 面 上 、 下 钢 板 厚 度 将 是 拱 脚 箱 形 截 面 上 、 下 板 厚 度 20 倍 以 上 , 其 用 钢 量 陡 增 , 该 站 房<br />

仅 钢 网 壳 用 钢 量 达 300kg/m 2 以 上 , 如 此 巨 大 的 用 钢 量 , 完 全 是 由 于 建 筑 设 计 单 纯 最 求 建 筑 造 型 的 “ 美<br />

感 ”, 而 违 背 基 本 的 力 学 规 律 而 造 成 的 , 此 种 建 筑 设 计 与 结 构 设 计 不 和 谐 的 现 象 是 时 有 出 现 的 , 应<br />

该 引 起 人 们 的 深 思 。<br />

-79-


1.2. 大 型 钢 结 构 设 计 应 遵 循 “ 安 全 、 合 理 、 先 进 、 经 济 ” 基 本 原 则<br />

“ 安 全 、 合 理 、 先 进 、 经 济 ” 这 四 句 话 八 个 字 是 矛 盾 的 统 一 体 , 对 结 构 体 系 而 言 , 安 全 是 第 一<br />

位 的 , 不 合 理 的 结 构 体 系 要 满 足 安 全 , 在 投 入 巨 大 的 财 力 和 物 力 条 件 下 是 可 以 达 到 安 全 要 求 的 , 不<br />

合 理 的 结 构 体 系 其 本 身 就 违 背 了 科 学 发 展 基 本 规 律 , 因 此 对 结 构 设 计 而 言 , 结 构 体 系 的 合 理 性 是 满<br />

足 基 本 原 则 的 根 本 。<br />

对 大 型 钢 结 构 设 计 而 言 , 要 达 到 优 良 的 技 术 经 济 指 标 , 必 然 要 有 客 观 的 评 价 标 准 , 对 超 大 跨 度<br />

钢 结 构 而 言 , 其 钢 结 构 自 重 (kg/m 2 ) 是 评 价 钢 结 构 设 计 “ 优 、 良 、 中 、 差 ” 的 标 准 , 文 献 [23] 指 出 “ 钢<br />

结 构 设 计 的 优 良 、 平 庸 与 拙 劣 设 计 可 以 钢 结 构 自 重 g 与 钢 结 构 屋 盖 总 重 量 p( 含 自 重 g 的 屋 盖 各 工 况<br />

总 重 量 ) 之 比 值 的 小 和 大 来 衡 量 , 即 : 优 秀 设 计 g/p=0.2~0.3, 平 庸 设 计 g/p=0.4~0.5, 拙 劣 设 计<br />

g/p=0.6~0.7”, 但 随 着 不 断 涌 现 的 超 大 跨 度 (L≥120m) 钢 结 构 , 将 此 评 价 标 准 适 当 放 宽 , 即<br />

优 秀 设 计 g/p=0.25~0.35<br />

(1a)<br />

平 庸 设 计 g/p=0.45~0.55<br />

(1b)<br />

拙 劣 设 计 g/p=0.65~0.75<br />

(1c)<br />

上 述 超 大 跨 度 除 钢 结 构 自 重 外 , 屋 盖 各 类 荷 载 的 总 重 为 200kg/m 2 时 , 其 g/P=300/500=0.6。 本<br />

文 提 出 的 两 类 超 大 跨 度 (150m×150m) 新 型 部 分 预 应 力 钢 网 格 结 构 , 它 们 的 用 钢 量<br />

g=80kg/m 2 ~100kg/m 2 之 间 , 其 g/p=(0.28~0.33)


称 之 为 “ 超 大 跨 度 ”, 相 关 规 定 指 出 : 当 跨 度 超 限 后 , 其 设 计 规 定 和 标 准 要 进 一 步 提 高 , 如 设 计 的<br />

地 震 烈 度 相 应 提 高 一 个 等 级 等 。 本 文 讨 论 的 超 大 跨 度 屋 盖 , 是 某 大 型 实 验 厅 的 屋 盖 , 使 用 方 要 求 钢<br />

网 格 屋 盖 结 构 自 重 q 0 ≤100kg/m 2 , 基 于 上 述 原 因 , 作 者 提 出 “ 下 撑 式 部 分 预 应 力 正 交 空 间 管 桁 架 钢<br />

网 格 结 构 ”(ZL201020170146.8)。<br />

2.2.1. 下 撑 式 部 分 预 应 力 正 交 空 间 管 桁 架 钢 网 格 结 构 基 本 组 成<br />

当 跨 度 达 150m×150m 时 , 若 采 用 常 规 的 各 类 平 板 型 钢 网 架 结 构 , 将 带 来 如 下 问 题 。 其 一 、 用<br />

钢 量 大 , 平 板 型 钢 网 架 上 、 下 弦 杆 的 轴 向 压 力 与 拉 力 是 由 网 架 弯 曲 变 形 得 出 , 网 架 弯 矩 (M) 与 结 构<br />

跨 度 平 方 呈 正 比 , 当 L 1 =150m 与 L 2 =60m 比 较 , 弯 矩 大 6.25 倍 , 当 L 2 =60m 时 , 上 弦 配 杆 为 Ф219<br />

× 7(g 2 =36.6kg/m) , 当 L 1 =150m 时 , 上 弦 配 杆 截 面 为 A 1 =291.4cm 2 , 配 杆 截 面 为 Ф630 ×<br />

15(g 1 =227.5kg/m), 上 、 下 弦 用 钢 量 陡 增 ; 其 二 、 当 跨 度 超 限 时 , 一 般 由 双 层 平 板 网 架 改 为 三 层 平 板<br />

网 架 , 目 的 是 减 少 由 构 造 控 制 的 腹 杆 用 钢 量 , 但 三 层 网 架 增 加 中 弦 层 弦 杆 和 节 点 , 在 超 大 跨 度 条 件<br />

下 , 采 用 三 层 平 板 网 架 , 并 不 能 很 好 地 改 善 屋 盖 结 构 的 技 术 经 济 指 标 ; 其 三 、 一 般 平 板 型 钢 网 架 其<br />

屋 面 排 水 是 在 上 弦 节 点 处 设 置 短 柱 , 按 坡 度 i=5% 考 虑 , 当 跨 度 为 30m 时 , 短 柱 高 750mm, 当 L 1 =150m<br />

时 , 其 短 柱 高 3750, 两 者 相 差 5 倍 , 屋 盖 整 体 性 受 到 影 响 的 同 时 , 其 赘 余 空 间 亦 增 大 。 如 何 克 服 常<br />

规 平 板 型 网 架 之 不 足 , 又 发 挥 空 间 管 桁 架 的 特 点 , 是 我 们 提 出 平 板 型 新 型 钢 网 格 结 构 体 系 的 原 因 。<br />

我 们 采 用 两 个 方 向 (x 与 y 方 向 ) 不 同 形 式 的 空 间 管 桁 架 彼 此 正 交 组 成 大 网 格 , 如 图 6 所 示 ,x 与<br />

y 方 向 空 间 管 桁 架 上 弦 宽 度 均 为 5000mm, 正 交 后 周 边 网 格 为 7500mm×15000mm, 四 角 隅 网 格<br />

7500mm×7500mm, 中 间 81 个 15000mm×15000mm 大 网 格 , 网 格 净 空 分 别 为 10000mm×10000mm<br />

及 5000mm×10000mm, 即 净 抽 空 率 达 44.4%。 周 边 及 四 角 隅 抽 空 部 位 设 置 上 弦 人 字 形 支 撑 , 提 高<br />

上 弦 层 整 体 性 。<br />

沿 x 方 向 10 榀 空 间 管 桁 架 考 虑 屋 面 排 水 坡 度 (i=7%) 两 上 弦 作 成 圆 弧 形 , 下 弦 为 直 线 , 其 10 榀<br />

管 桁 架 外 轮 廓 尺 寸 均 相 同 , 如 图 7 所 示 。y 方 向 空 间 管 桁 架 的 几 何 尺 寸 , 由 与 之 正 交 的 x 向 空 间 管<br />

桁 架 该 处 几 何 尺 寸 确 定 , 其 高 度 由 中 央 向 两 端 支 承 点 递 减 , 即 由 中 央 向 两 端 各 5 榀 的 几 何 尺 寸 各 不<br />

相 同 , 如 图 8 所 示 , 从 四 周 边 算 起 的 x,y 向 第 三 榀 空 间 管 桁 架 正 交 处 的 下 弦 节 点 下 部 设 置 了 “ 下 撑<br />

式 四 角 锥 钢 索 转 向 支 撑 架 ”, 两 根 折 线 钢 索 在 此 处 向 内 斜 向 至 此 处 转 为 水 平 线 , 从 而 形 成 折 线 钢 索<br />

的 夹 角 α x , α y 。<br />

Y<br />

10000<br />

37500<br />

10000<br />

5000<br />

5000<br />

75000<br />

150000<br />

37500<br />

X<br />

5000 5000 10000 5000 10000 5000 10000<br />

37500<br />

75000<br />

150000<br />

37500<br />

图 6 预 应 力 正 交 空 间 管 桁 架 钢 网 格 结 构 平 面 图<br />

-81-


5250<br />

5000<br />

i= 0.07<br />

1<br />

i= 0.07<br />

5000<br />

16000<br />

6250<br />

αix<br />

1 1-1<br />

αix<br />

11500<br />

4500<br />

37500 75000 37500<br />

150000<br />

图 7 沿 x 方 向 变 截 面 空 间 管 桁 架 简 图<br />

图 8 沿 y 方 向 等 截 面 空 间 管 桁 架 简 图<br />

2.2.2. 下 撑 式 部 分 预 应 力 正 交 空 间 管 桁 架 钢 网 格 结 构 主 要 构 造<br />

1) 上 弦 及 下 弦 节 点 构 造 : 此 处 不 能 沿 袭 常 规 的 拟 梁 式 节 点 构 造 , 其 原 因 有 两 方 面 。 其 一 、150m<br />

长 度 范 围 内 力 大 小 变 化 很 大 , 上 下 弦 杆 直 径 大 小 必 然 有 改 变 , 否 则 用 钢 量 增 大 显 著 ; 其 二 、 节 点 交<br />

汇 的 杆 件 增 多 , 与 之 正 交 管 桁 架 弦 杆 与 腹 杆 亦 在 该 节 点 交 汇 , 如 图 9a, b 为 上 弦 节 点 构 造 , 图 9c, d<br />

为 下 弦 节 点 构 造 。<br />

弯 管<br />

D2<br />

D 大<br />

锥 头 1<br />

D1<br />

D 小<br />

锥 头 2<br />

D2<br />

D1<br />

(a)<br />

(b)<br />

D 2<br />

D 1<br />

(c)<br />

图 9 a, b, c, d 上 、 下 弦 节 点 构 造 简 图<br />

(d)<br />

-82-


图 10 a, b 钢 索 转 向 支 撑 架 构 造 简 图<br />

“ 下 撑 式 四 角 锥 钢 索 结 构 支 撑 架 ” 是 设 置 在 x,y 两 榀 空 间 管 桁 架 两 下 弦 交 汇 节 点 ( 空 心 球 节 点 )<br />

沿 x,y 向 各 两 根 折 线 钢 索 在 此 处 转 向 , 设 x,y 向 钢 索 预 拉 力 分 别 为 P ax , P ay 两 个 方 向 折 线 钢 索 转 向<br />

夹 角 为 α x , α y , 此 处 四 根 折 线 钢 索 作 用 在 下 弦 节 点 向 上 的 作 用 力 为 ∑P b =2(P ax sinα x +P ay sinα y ), 图 10a, b<br />

为 下 撑 式 转 向 支 撑 架 构 造 简 图 。<br />

2.3. 150m×150m 预 应 力 正 交 空 间 管 桁 架 钢 网 格 结 构 屋 盖 设 计 简 介<br />

2.3.1. 150m×150m 非 预 应 力 正 交 空 间 管 桁 架 钢 网 格 结 构 设 计 简 述<br />

在 图 6 所 示 网 格 结 构 布 置 图 中 , 去 掉 井 字 形 折 线 钢 索 , 即 图 7、 图 8 下 弦 下 部 取 消 , 形 成 非 预<br />

应 力 正 交 空 间 管 桁 架 网 格 结 构 , 屋 盖 结 构 外 轮 廓 形 成 图 11 所 示 柱 面 型 , 屋 盖 中 央 高 度 好<br />

h 1 =11500mm(l/13), 支 座 处 高 度 h 2 =6250mm, 两 者 相 差 5250, 即 原 短 柱 高 度 , 屋 盖 按 常 规 保 温 屋 面<br />

作 法 , 考 虑 “ 正 常 使 用 极 限 状 态 工 况 组 合 ”, 正 常 使 用 极 限 状 态 的 抗 震 工 况 组 合 及 “ 承 载 力 极 限 状<br />

态 工 况 组 合 ”、“ 承 载 力 极 限 状 态 抗 震 工 况 组 合 ”, 进 行 了 静 力 与 动 力 的 线 性 与 非 线 性 分 析 , 其 最 大<br />

弹 性 位 移 (z) f=473.3, f/l=1/317


A<br />

Pb<br />

Pa<br />

Pb<br />

A<br />

Pb<br />

Pa<br />

Pb<br />

Pa<br />

Pa<br />

ax<br />

Pb<br />

Pa<br />

Pb<br />

ax<br />

图 13a 预 应 力 网 格 结 构 自 平 衡 体 系 简 图<br />

图 13b 左 图 为 A—A 剖 面 图<br />

2) 成 型 态 结 构 的 静 力 分 析 : 结 构 成 型 态 的 自 平 衡 体 系 为 结 构 受 自 重 及 预 应 力 共 同 作 用 后 的 变 形<br />

状 态 , 从 而 产 生 结 构 自 平 衡 内 力 。 成 型 态 静 力 分 析 必 须 考 虑 结 构 的 P-Δ 效 应 和 大 位 移 的 几 何 非 线 性<br />

影 响 。 分 析 中 进 行 了 结 构 自 重 作 用 和 结 构 自 重 与 钢 索 预 应 力 共 同 作 用 的 对 比 分 析 , 结 构 在 自 重 作 用<br />

下 z 向 位 移 为 -203.4mm,z 向 位 移 比 1/737, 结 构 在 成 型 态 ( 自 重 及 预 应 力 作 用 )z 位 移 为 -57.3mm,z<br />

向 位 移 比 1/2618, 两 者 相 差 3.55 倍 , 图 14a, b 分 别 为 结 构 在 自 重 作 用 下 竖 向 位 移 图 和 结 构 在 自 重 及<br />

预 应 力 共 同 作 用 下 的 竖 向 位 移 图 。 从 结 构 的 竖 向 变 形 分 析 , 预 应 力 有 改 善 结 构 内 力 分 布 和 减 小 部 分<br />

杆 件 内 力 作 用 。 具 体 分 析 内 容 详 见 文 献 [10]。<br />

A-A<br />

图 14a 自 重 下 z 方 向 位 移 图 14b 成 型 态 下 z 方 向 位 移<br />

3) 使 用 态 结 构 静 力 分 析 : 结 构 使 用 态 为 结 构 在 成 型 态 这 种 自 平 衡 体 系 状 态 下 , 受 到 屋 面 恒 载 ,<br />

下 弦 恒 载 , 屋 面 活 载 , 温 度 、 风 及 地 震 作 用 下 的 变 形 状 态 和 网 格 结 构 各 杆 件 内 力 分 布 情 况 , 经 分 析<br />

结 构 跨 中 竖 向 最 大 位 移 f z =-488.2mm, f/l-1/307, 图 15 为 结 构 使 用 态 竖 向 位 移 图 。<br />

图 15 结 构 使 用 态 z 向 位 移 图<br />

4) 使 用 态 结 构 动 力 分 析 : 利 用 成 型 态 结 构 刚 度 进 行 分 析 , 得 到 结 构 的 自 振 周 期 , 如 表 1 所 示 ,<br />

前 八 个 模 态 ( y , z,<br />

x ) 中 第 二 模 态 为 z 向 的 振 动 , 如 图 16 所 示 。 结 构 在 自 重 及 预 应 力 共 同 作 用 下 的 位<br />

移 动 力 响 应 采 用 振 型 分 解 反 应 谱 法 求 得 的 最 大 z 向 位 移 为 -393.2mm,z 向 位 移 比 1/381, 其 z 向 位 移<br />

云 图 如 图 17 所 示 。 表 1 为 结 构 的 自 振 周 期 。<br />

表 1 超 大 跨 度 预 应 力 正 交 空 间 管 桁 架 网 格 结 构 的 自 振 周 期<br />

阶 数 1 2 3 4 5 6 7 8 9 10<br />

周 期 (s) 1.697 1.512 1.255 1.202 1.030 0.903 0.810 0.761 0.742 0.701<br />

阶 数 11 12 13 14 15 16 17 18 19 20<br />

周 期 (s) 0.672 0.656 0.587 0.569 0.519 0.504 0.494 0.489 0.473 0.464<br />

-84-


图 16 第 二 模 态<br />

图 17 结 构 z 向 位 移 云 图<br />

5) 结 构 的 配 杆 设 计 : 依 据 结 构 最 不 利 工 况 下 网 格 结 构 各 杆 件 内 力 进 行 配 杆 设 计 , 上 弦 从 Ф95×4<br />

到 Ф550×12, 共 13 种 规 格 , 应 力 比 达 到 0.7 以 上 , 上 弦 杆 比 一 般 空 间 管 桁 架 上 弦 杆 利 用 较 充 分 ; 下<br />

弦 杆 从 Ф95×4 到 Ф650×20, 共 15 种 规 格 , 应 力 比 也 达 0.7 以 上 , 腹 杆 从 Ф159×6 到 Ф426×10, 共 8<br />

种 规 格 , 应 力 比 达 0.5 以 上 。150m×150m 预 应 力 正 交 空 间 管 桁 架 钢 网 格 结 构 用 钢 量 82.1kg/m 2 , 总<br />

用 钢 1847T。 按 钢 结 构 设 计 评 判 标 准 :g/P=82.1/(200+82.1)=0.291


(a) 预 应 力 索 布 置<br />

(b) 1/6 平 面<br />

(c) 预 应 力 组 合 扭 网 壳 对 角 线 剖 面<br />

(d) 单 块 扭 网 壳 钢 索 位 置 剖 面<br />

(e) 单 块 扭 网 壳 直 线 桁 架 简 图<br />

图 19a,b,c,d,e 广 东 清 远 市 预 应 力 双 层 组 合 扭 网 壳 简 图<br />

图 20a 为 广 东 省 高 要 市 体 育 馆 工 程 实 录 , 由 四 块 双 层 扭 网 壳 , 四 柱 支 承 形 成 62m×74m 矩 形 平<br />

面 预 应 力 双 层 组 合 扭 网 壳 , 图 20b 为 网 壳 平 面 图 , 结 构 用 钢 42kg/m 2 。<br />

图 20a 广 东 高 要 市 体 育 馆 工 程 实 录<br />

图 20b 双 层 预 应 力 组 合 扭 网 壳 平 面 图<br />

图 21 为 预 应 力 田 字 形 双 层 与 单 层 组 合 扭 网 壳 轴 测 图 , 它 由 四 块 周 边 双 层 中 央 单 层 扭 网 壳 组 成 ,<br />

网 格 双 层 与 单 层 布 置 时 根 据 内 力 分 布 大 与 小 来 确 定 。 图 22a 为 1997 年 建 成 的 广 东 省 新 兴 县 体 育 馆 ,<br />

平 面 尺 寸 60m×68m, 用 钢 量 34.8kg/m 2 , 图 22b 为 1998 年 建 成 的 广 西 省 桂 平 市 体 育 馆 平 面 尺 寸 为<br />

72m×80m, 用 钢 40.5kg/m 2 。<br />

-86-


图 21 预 应 力 田 字 形 双 层 与 单 层 组 合 扭 网 壳 轴 测 图<br />

图 22a 广 西 新 兴 县 体 育 馆 (60m×68m)<br />

图 22b 广 西 桂 平 市 体 育 馆 外 景 (72m×80m)<br />

上 述 预 应 力 组 合 扭 网 壳 结 构 , 其 结 构 跨 度 均 在 60m 到 78m 之 间 , 属 大 跨 度 钢 网 格 结 构 范 畴 ,<br />

其 结 构 自 重 g 与 屋 盖 结 构 总 重 P( 含 g) 之 比 , 均 在 优 秀 设 计 的 技 术 经 济 指 标 g/P=(0.25~0.30) 范 围 , 其<br />

中 清 远 市 体 育 馆 g/P=0.182, 广 东 高 要 市 体 育 馆 g/P=0.173, 广 东 新 兴 县 体 育 馆 g/P=0.144, 广 西 桂 平<br />

市 体 育 馆 g/P=0.168, 其 g/P 比 值 均 小 于 0.2, 这 是 合 理 的 网 格 结 构 布 置 和 采 用 现 代 预 应 力 所 带 来 的<br />

功 效 。<br />

3.2. 150m×150m 周 边 简 支 承 预 应 力 十 字 形 三 层 与 双 层 组 合 扭 网 壳 结 构<br />

上 述 点 支 承 预 应 力 组 合 扭 网 壳 结 构 , 应 用 于 跨 度 在 60m~80m 大 跨 度 范 围 其 平 面 角 部 网 壳 悬 挑<br />

跨 度 在 12.6m 至 25m 之 间 , 对 于 150m×150m 方 形 平 面 组 合 扭 网 壳 屋 盖 , 再 沿 用 方 形 平 面 四 点 支 承<br />

组 合 扭 网 壳 , 其 角 部 净 悬 挑 ( 不 算 挑 檐 ) 跨 度 达 53m, 当 网 壳 L=150m 时 , 其 结 构 跨 度 超 限 (L 1 >120m)<br />

的 同 时 , 其 角 部 悬 挑 跨 度 亦 超 限 (L 2 >40m) 形 成 “ 结 构 双 重 超 限 ”, 这 是 结 构 设 计 要 尽 量 避 免 的 , 本 文<br />

提 出 的 新 型 超 大 跨 度 预 应 力 组 合 扭 网 壳 结 构 , 它 消 除 了 角 部 悬 挑 超 限 而 导 致 的 “ 双 重 超 限 ”, 并 取<br />

得 良 好 的 技 术 经 济 指 标 。<br />

3.2.1. 结 构 布 置 及 相 关 几 何 尺 寸 的 确 定<br />

图 23 为 150m×150m 周 边 简 支 承 部 分 预 应 力 十 字 形 三 层 与 双 层 组 合 扭 网 壳 网 格 与 预 应 力 钢 索 布<br />

置 平 面 图 , 其 正 交 网 格 为 3750mm×3750mm, 在 此 方 向 , 沿 钢 索 方 向 布 置 斜 向 网 格 形 成 三 角 形 网 格 。<br />

周 边 柱 网 7500mm, 檐 口 悬 挑 两 个 网 格 , 即 7500mm, 沿 预 应 力 钢 索 合 力 方 向 , 即 中 央 受 力 最 大 处<br />

布 置 三 层 网 格 , 即 6×3750mm=22500mm 处 布 置 十 字 形 三 层 网 格 , 占 整 个 屋 盖 总 投 影 面 积 23%, 此<br />

屋 盖 若 为 四 点 支 撑 , 角 部 悬 挑 63.64m, 它 比 超 限 悬 挑 跨 度 40m 还 超 出 59%, 点 支 承 组 合 扭 网 壳 在<br />

跨 度 超 限 时 不 宜 采 用 。 图 24a,b,c,d 分 别 为 其 跨 度 中 央 剖 面 图 (I-I), 三 层 与 双 层 剖 面 (Ⅱ-Ⅱ) 预 应 力 钢<br />

索 位 置 剖 面 (Ⅲ-Ⅲ) 及 屋 盖 对 角 线 处 剖 面 (A-A), 图 25 为 此 屋 盖 网 格 结 构 的 计 算 模 型 。<br />

-87-


图 23 150m×150m 周 边 简 支 承 预 应 力 组 合 扭 网 壳 平 面 图<br />

图 24(a) Ⅰ-Ⅰ 剖 面 图<br />

图 24(b) Ⅱ-Ⅱ 剖 面 图<br />

图 24(c) Ⅲ-Ⅲ 剖 面 图<br />

10606 18562 68943 18562 18562<br />

68943 18562 10606<br />

212132<br />

图 24(d) 对 角 线 A-A 剖 面 图<br />

-88-


图 25 三 层 网 格 十 字 形 布 置 的 预 应 力 组 合 扭 网 壳 屋 盖 的 轴 测 图<br />

3.2.2. 周 边 简 支 承 非 预 应 力 十 字 形 三 层 与 双 层 组 合 扭 网 壳 结 构 设 计<br />

常 规 组 合 扭 网 壳 设 置 预 应 力 钢 索 部 位 为 钢 管 , 经 分 析 该 处 系 杆 为 Ф1420×30(518kg/m), 四 根 系<br />

杆 总 重 量 217T, 整 个 屋 盖 用 钢 量 109kg/m 2 , 其 g/P=0.35, 其 经 济 指 标 仍 满 足 要 求 , 但 如 此 巨 大 的 系<br />

杆 直 径 , 已 使 其 构 造 很 难 达 到 设 计 要 求 , 设 计 中 不 予 选 用 此 方 案 。<br />

3.3. 150m×150m 周 边 简 支 承 十 字 形 三 层 与 双 层 预 应 力 组 合 扭 网 壳 结 构 分 析<br />

3.3.1. 结 构 成 型 态 静 力 分 析<br />

1) 结 构 成 型 态 位 移 : 钢 索 调 直 张 紧 后 , 首 先 不 建 立 预 应 力 , 结 构 自 重 产 生 的 竖 向 位 移<br />

f z =-148.2mm,z 向 位 移 比 1/1012, 当 采 用 800Ф7mm 钢 丝 束 (Ф=200mm), 钢 索 安 全 系 数 K=2.5, 张<br />

拉 钢 索 建 立 预 应 力 后 , 结 构 最 大 竖 向 位 移 f z =101mm, 向 上 位 移 比 为 1/1485, 成 型 态 使 结 构 自 重 产 生<br />

的 竖 向 位 移 减 小 为 -47.2mm, 实 际 位 移 比 为 1/3177。<br />

预 应 力 作 用 下 结 构 刚 度 显 著 地 提 高 , 图 26a 为 结 构 在 自 重 作 用 下 z 向 位 移 图 , 图 26b 结 构 成 型<br />

态 ( 含 自 重 ) 下 z 向 位 移 云 图 。<br />

图 26a 结 构 自 重 作 用 下 z 向 位 移<br />

图 26b 结 构 成 型 态 作 用 下 z 向 位 移 图<br />

2) 结 构 成 型 态 内 力 简 述 : 成 型 态 承 载 力 极 限 状 态 的 工 况 组 合 分 两 项 : 第 一 项 为 :1.1×(1.2 结 构<br />

自 重 +1.0 预 应 力 ); 第 二 项 为 :1.1×(1.35 结 构 自 重 +1.0 预 应 力 ), 两 项 均 采 用 非 线 性 计 算 。 它 们 与 非<br />

预 应 力 计 算 结 果 对 比 分 析 , 由 于 预 应 力 作 用 , 结 构 杆 系 有 大 量 杆 件 内 力 减 少 , 即 “ 减 力 杆 ”, 而 在<br />

钢 索 支 座 附 近 出 现 压 力 增 大 , 即 “ 增 力 杆 ”。 图 27a 为 自 重 作 用 下 ( 非 预 应 力 ) 上 弦 层 弦 杆 压 力 , 图<br />

27b 为 成 型 态 ( 自 重 加 预 应 力 ) 上 弦 层 弦 杆 压 力 , 从 图 27 可 知 , 成 型 态 时 上 弦 层 中 各 杆 压 力 均 不 同 程<br />

度 的 减 小 。<br />

-89-


图 27(a) 自 重 作 用 下 结 构 上 弦 轴 力 图<br />

图 27(b) 成 型 态 ( 含 自 重 ) 结 构 上 弦 轴 力<br />

3.3.2 结 构 使 用 态 静 力 分 析<br />

1) 结 构 使 用 态 竖 向 位 移 : 结 构 成 型 态 下 考 虑 最 不 利 工 况 荷 载 态 作 用 下 ( 荷 载 态 ), 其 z 向 变 形 如<br />

图 28 所 示 , 其 网 壳 中 央 最 大 竖 向 挠 度 f z =263.1mm,z 向 位 移 比 为 1/570


图 29 网 壳 第 一 阶 自 振 模 态<br />

3) 结 构 位 移 的 动 响 应 : 表 4 为 结 构 在 地 震 作 用 下 各 向 位 移 , 分 3 种 正 常 使 用 极 限 状 态 的 使 用 态<br />

工 况 , 从 表 5 可 知 , 第 三 种 工 况 ( 水 平 地 震 作 用 和 竖 向 地 震 作 用 同 时 考 虑 ) 作 用 下 竖 向 挠 度 和 水 平 位<br />

移 均 为 最 大 值 , 但 比 静 力 作 用 下 最 大 挠 度 263.1mm 小 33%。 图 30 为 结 构 在 地 震 作 用 下 的 z 向 位 移<br />

图 。<br />

表 4 三 层 网 格 十 字 形 布 置 的 预 应 力 组 合 扭 网 壳 在 地 震 作 用 下 的 位 移<br />

工 况 x 向 位 移 (mm) y 向 位 移 (mm) z 向 位 移 (mm) z 向 位 移 比 值<br />

1 -44.5 -28.8 -162.9 1/921<br />

2 -31.2 -31.2 -174.4 1/860<br />

3 -47.4 -31.6 -175.6 1/854<br />

(a) x 向 位 移<br />

(b) y 向 位 移<br />

3.3.4. 结 构 的 配 杆 设 计<br />

(c) z 向 位 移<br />

图 30 结 构 在 地 震 作 用 下 x、y 与 z 方 向 地 震 工 况 的 位 移<br />

通 过 静 力 与 动 力 分 析 , 根 据 其 内 力 进 行 配 杆 设 计 , 扭 网 壳 上 层 杆 件 规 格 11 种 , 最 小 为 Ф76×3,<br />

最 大 为 Ф299×8, 上 层 腹 杆 9 种 规 格 , 最 小 Ф76×3, 最 大 Ф245×7; 中 弦 层 弦 杆 7 种 规 格 , 最 小 为 Ф76×3,<br />

最 大 为 Ф194×6, 十 字 形 三 层 中 下 层 腹 杆 有 13 种 规 格 , 最 小 为 Ф76×3, 最 大 为 Ф351×12, 下 弦 层 弦<br />

杆 11 种 规 格 最 小 Ф76×3, 最 大 Ф351×14。 含 钢 索 总 共 有 16 种 规 格 , 平 均 用 钢 量 77.3kg/m 2 , 总 用<br />

钢 1739.25T, 其 g/P=0.28


四 、 结 语<br />

150m×150m 空 间 钢 网 格 结 构 属 超 限 的 大 跨 度 结 构 , 如 何 使 此 类 超 大 跨 度 钢 网 格 达 到 满 意 的 技 术<br />

经 济 指 标 , 符 合 结 构 设 计 “ 安 全 、 合 理 、 先 进 、 经 济 ” 的 设 计 要 求 , 是 工 程 设 计 者 一 直 追 求 的 目 标 。<br />

“ 正 交 空 间 管 桁 架 预 应 力 钢 网 格 结 构 ”, 与 “ 周 边 简 支 承 十 字 型 三 层 与 双 层 预 应 力 组 合 扭 网 壳 结 构 ”,<br />

它 们 均 达 到 了 结 构 设 计 优 秀 标 准 , 起 到 了 “ 异 曲 同 功 ” 的 作 用 , 但 两 者 有 如 下 不 同 点 : 其 一 、 结 构<br />

力 学 模 型 不 同 , 前 者 为 “ 拟 夹 芯 板 ” 型 , 后 者 为 “ 拟 夹 芯 壳 ” 型 ; 其 二 、 结 构 网 格 布 置 不 同 , 前 者<br />

网 格 大 (5m×5m) 且 有 抽 空 网 格 10m×10m 和 5m×10m, 结 构 抽 空 率 大 。 后 者 网 格 小 (3.75m×3.75m) 且 为 三<br />

向 平 面 桁 架 交 叉 组 成 , 其 三 、 两 者 均 采 用 部 分 预 应 力 , 但 前 者 采 用 “ 转 向 支 撑 架 ”, 钢 索 建 立 预 应<br />

力 时 , 产 生 的 预 拉 力 P<br />

a<br />

, 在 转 向 位 置 产 生 在 下 弦 球 节 点 向 上 作 用 力 P b =P a sinα。 从 而 建 立 结 构 的 自<br />

平 衡 体 系 。 后 者 是 在 每 块 扭 网 壳 支 座 节 点 处 连 接 直 线 钢 索 , 并 设 若 干 吊 杆 后 建 立 预 应 力 , 钢 索 张 拉<br />

力 有 使 扭 网 壳 产 生 与 荷 载 作 用 相 反 的 变 形 。 从 而 形 成 结 构 自 平 衡 体 系 。 此 两 种 新 型 钢 网 格 结 构 , 符<br />

合 结 构 设 计 的 基 本 要 求 。<br />

科 学 发 展 的 客 观 规 律 告 诉 我 们 : 任 何 事 物 没 有 最 好 , 只 有 更 好 , 本 文 提 出 的 两 种 新 型 超 大 跨 度<br />

部 分 预 应 力 空 间 钢 网 网 格 结 构 , 旨 在 “ 抛 砖 引 玉 ”, 促 进 预 应 力 钢 结 构 进 一 步 发 展 。<br />

参 考 文 献<br />

[1] GB 50017-2003 钢 结 构 设 计 规 范 , 北 京 : 中 国 计 划 出 版 社 , 2003.<br />

[2]JGJ 61-2003 网 壳 结 构 技 术 规 程 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 2003.<br />

[3] JGJ 7-91 网 架 结 构 设 计 与 施 工 规 程 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 1992.<br />

[4] JGJ 99-98 高 层 民 用 建 筑 钢 结 构 技 术 规 程 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 1998.<br />

[5] GB 50011-2001 建 筑 抗 震 设 计 规 范 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 2001.<br />

[6] CECS 212 2006 预 应 力 钢 结 构 技 术 规 程 , 北 京 : 中 国 计 划 出 版 社 , 2006.<br />

[7] 马 克 俭 , 张 华 刚 , 等 . 新 型 建 筑 空 间 网 格 结 构 理 论 与 实 践 , 北 京 : 人 民 交 通 出 版 社 , 2006.<br />

[8] 马 克 俭 , 张 鑫 光 , 安 竹 石 , 等 . 大 跨 度 组 合 式 预 应 力 扭 网 壳 结 构 的 设 计 、 构 造 与 力 学 特 点 . 空 间 结 构 , 1994, 1(1):<br />

55-63.<br />

[9] 陆 赐 麟 , 尹 思 明 , 刘 锡 良 . 现 代 预 应 力 钢 结 构 ( 修 订 版 ), 北 京 : 人 民 交 通 出 版 社 , 2007.<br />

[10] 马 克 俭 , 申 波 . 新 型 超 大 跨 度 预 应 力 空 间 钢 网 格 结 构 , 北 京 : 人 民 交 通 出 版 社 , 2011.<br />

[11] 马 克 俭 , 张 华 刚 , 等 . 大 跨 度 大 柱 距 拟 肋 型 楼 盖 式 预 应 力 钢 网 格 屋 盖 结 构 : 中 国 专 利 , ZL200610200673.3,<br />

2008-10-01.<br />

[12] 马 克 俭 , 张 华 刚 , 等 . 大 跨 度 周 边 简 支 承 体 内 折 线 预 应 力 钢 网 架 结 构 : 中 国 专 利 , ZL200610200674.1, 2009-01-21.<br />

[13] 马 克 俭 , 郑 涛 , 等 . 大 跨 度 大 悬 挑 屋 面 斜 拉 式 双 层 预 应 力 钢 网 架 屋 盖 结 构 : 中 国 专 利 , ZL200610200672.2,<br />

2007-12-05.<br />

[14] 马 克 俭 , 郑 涛 , 等 . 一 种 短 撑 杆 式 张 弦 桁 架 及 其 制 作 方 法 : 中 国 专 利 , ZL200410155399.7, 2007-08-29.<br />

[15] 马 克 俭 , 张 华 刚 , 等 . 大 跨 度 张 弦 桁 架 的 制 作 方 法 : 中 国 专 利 , ZL200410022224.9, 2009-12-27.<br />

[16] 马 克 俭 , 张 华 刚 , 等 . 体 内 预 应 力 空 间 管 桁 架 四 点 支 承 大 跨 度 扁 网 壳 屋 盖 结 构 : 中 国 专 利 , ZL200610200676.0,<br />

2009-01-21.<br />

[17] 马 克 俭 , 郑 涛 , 等 . 大 跨 度 周 边 三 角 锥 简 支 承 体 内 折 线 预 应 力 钢 网 架 结 构 : 中 国 专 利 , ZL200610200675.6,<br />

2007-12-05.<br />

[18] 马 克 俭 , 申 波 , 等 . 超 大 跨 度 下 撑 式 正 交 空 间 管 桁 架 网 格 结 构 及 其 制 作 方 法 : 中 国 专 利 , 申 请 号 或 专 利 号<br />

201020170146.8, 2010-04-26 受 理 .<br />

[19] 马 克 俭 , 刘 建 军 , 等 . 超 大 跨 度 变 截 面 双 层 与 三 层 预 应 力 肋 梁 式 网 架 结 构 : 中 国 专 利 , 申 请 号 或 专 利 号<br />

201010155102.2. 2010-04-26 受 理 .<br />

[20] 马 克 俭 , 徐 向 东 , 等 . 超 大 跨 度 变 截 面 下 撑 式 折 线 预 应 力 抽 空 四 角 锥 网 架 结 构 : 中 国 专 利 , 申 请 号 或 专 利 号<br />

201010155102.2. 2010-04-26 受 理 .<br />

[21] 马 克 俭 , 周 绪 红 , 等 . 超 大 跨 度 周 边 简 支 承 预 应 力 双 层 与 三 层 组 合 扭 网 壳 结 构 : 中 国 专 利 , 申 请 号 或 专 利 号<br />

201010155393.5. 2010-04-26 受 理 .<br />

[22] 陈 荣 毅 , 董 石 麟 . 广 州 国 际 会 议 展 览 中 心 展 览 大 厅 钢 屋 盖 设 计 . 空 间 结 构 , 2002,8(3): 29-34.<br />

[23] 陆 赐 麟 . 用 科 学 标 准 促 进 钢 结 构 行 业 健 康 发 展 . 钢 结 构 , 2009, 24(12): 74-78.<br />

[24] 贺 拥 军 , 周 绪 红 , 黄 海 顺 . 超 大 跨 屋 面 结 构 风 速 时 程 的 数 值 模 拟 研 究 , 湖 南 大 学 学 报 ( 自 然 科 学 版 ), 2008, 35(9),<br />

1-5.<br />

[25] Dong, Shilin and Yuan, Xingfei. Pretension process analysis of prestressed space grid structures. Journal of<br />

Constructional Steel Research, 2007, 63(3): 406-411.<br />

-92-


Invited Lectures


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

中 国 高 铁 客 站 的 创 新 与 实 践<br />

郑 健<br />

( 中 华 人 民 共 和 国 铁 道 部 , 北 京 100844)<br />

摘 要 : 高 铁 客 站 既 是 高 速 铁 路 的 重 要 组 成 部 分 , 也 是 国 家 综 合 交 通 体 系 的 重 要 节 点 和 带 动 城<br />

市 发 展 不 可 或 缺 的 重 要 因 子 。 伴 随 着 我 国 高 速 铁 路 的 快 速 发 展 , 铁 路 客 站 的 建 设 迎 来 了 难 得 的 发 展<br />

机 遇 。 历 经 六 年 多 的 探 索 与 实 践 , 我 国 高 速 铁 路 客 站 在 建 设 理 念 、 关 键 技 术 、 工 程 管 理 等 方 面 取 得<br />

了 重 要 进 展 。 本 文 全 面 介 绍 了 中 国 铁 路 客 站 的 规 划 与 进 展 , 深 入 剖 析 了 中 国 高 铁 客 站 建 设 存 在 的 挑<br />

战 , 系 统 阐 述 了 中 国 高 铁 客 站 的 建 设 目 标 和 实 施 路 径 , 简 要 总 结 了 中 国 高 铁 客 站 取 得 的 显 著 成 效 。<br />

关 键 词 : 高 铁 客 站 规 划 进 展 实 现 路 径 成 效<br />

INNOVATION AND PRACTICE OF CHINA HIGH-SPEED RAIL STATIONS<br />

Zheng Jian<br />

Ministry of Railways of the People’s Republic of China, Beijing 100844,China<br />

Abstract: Railway stations, an important part of sophisticated railway network, also serve as important nodes<br />

for the national comprehensive traffic system and an indispensable drive force of urban development. With the<br />

rigorous development of high-speed rail in the country, railway station construction has also ushered in a rare<br />

opportunity of development. After six years’exploration and practice in building high speed rail stations, major<br />

progress has been made in terms of construction concept, key technologies, engineering management. <strong>The</strong> paper<br />

makes an intensive introduction on the planning and progress made in station construction and an in-depth<br />

analysis on the challenges ahead by elaborating on the goals of high speed rail stations and implementation<br />

approaches.<br />

Keywords: High-speed rail stations, planning and progress, implementation approaches, achievement.<br />

一 、 前 言<br />

伴 随 着 我 国 高 速 铁 路 大 规 模 快 速 推 进 , 作 为 高 速 铁 路 重 要 组 成 部 分 的 高 铁 客 站 , 迎 来 了 快 速 发<br />

展 的 难 得 机 遇 。 按 照 国 家 批 准 的 《 中 长 期 铁 路 网 规 划 》, 到 2015 年 , 我 国 将 建 成 4.5 万 公 里 的 快 速<br />

铁 路 客 运 网 , 其 中 , 新 建 时 速 250 公 里 及 以 上 的 高 速 铁 路 运 营 里 程 将 达 到 1.6 万 公 里 。 为 实 现 点 线<br />

能 力 匹 配 , 提 高 我 国 铁 路 客 运 网 的 综 合 效 率 和 服 务 质 量 , 需 建 成 高 铁 客 站 近 600 座 。<br />

截 至 2011 年 4 月 底 , 已 建 成 北 京 南 、 武 汉 、 广 州 南 、 上 海 虹 桥 等 高 铁 客 站 140 座 , 正 在 建 设 的<br />

139 座 。 这 些 客 站 无 论 在 功 能 布 局 、 交 通 流 线 、 建 筑 造 型 、 关 键 技 术 , 还 是 在 服 务 设 施 上 , 与 以 往<br />

客 站 相 比 都 有 重 大 创 新 和 突 破 , 普 遍 做 到 了 能 力 充 足 、 功 能 完 善 、 换 乘 便 捷 , 成 为 所 在 城 市 的 现 代<br />

化 综 合 交 通 枢 纽 。<br />

二 、 我 国 高 铁 客 站 建 设 目 标 与 挑 战<br />

高 速 铁 路 的 大 规 模 建 设 、 综 合 交 通 体 系 的 不 断 完 善 、 城 镇 化 的 快 速 推 进 , 对 我 国 高 铁 客 站 的 建<br />

设 提 出 了 新 的 需 求 。 在 当 下 的 时 代 背 景 和 国 情 条 件 下 , 高 铁 客 站 在 价 值 取 向 、 建 设 理 念 、 具 体 内 涵<br />

等 方 面 与 传 统 客 站 相 比 发 生 了 巨 大 变 化 。<br />

2.1 建 设 目 标<br />

围 绕 我 国 高 速 铁 路 和 综 合 交 通 体 系 的 发 展 目 标 , 研 究 确 立 了 “ 打 造 ‘ 百 年 不 朽 ’ 精 品 工 程 ” 的<br />

高 铁 客 站 建 设 总 目 标 。 其 内 涵 是 : 立 足 高 起 点 、 高 标 准 、 高 水 平 , 建 设 功 能 完 善 、 可 达 性 高 、 换 乘<br />

便 捷 、 设 施 先 进 、 文 化 特 色 突 出 的 高 铁 客 站 , 满 足 客 站 全 生 命 周 期 内 旅 客 候 车 方 式 和 客 流 量 变 化 的<br />

要 求 , 满 足 城 市 土 地 开 发 、 功 能 拓 展 和 空 间 结 构 优 化 的 需 求 , 着 力 实 现 价 值 工 程 。<br />

2.2 面 临 的 挑 战<br />

1. 建 设 规 模 大 。 按 照 铁 路 网 规 划 和 目 前 的 建 设 进 度 ,“ 十 二 五 ” 期 间 , 我 国 还 将 建 成 高 铁 客 站<br />

-93-


400 多 座 , 其 中 , 省 会 级 大 型 客 站 43 座 。 既 有 的 设 计 资 源 、 建 设 模 式 已 无 法 适 应 大 规 模 高 铁 客 站 的<br />

快 速 发 展 。<br />

2. 时 间 要 求 紧 。 高 铁 客 站 的 规 划 设 计 受 制 于 高 铁 线 路 、 车 场 以 及 城 市 轨 道 交 通 、 市 政 道 路 、 站<br />

区 规 划 等 多 重 因 素 , 开 工 时 间 一 般 滞 后 高 铁 线 路 两 年 左 右 , 但 又 需 与 高 铁 线 路 同 步 开 通 运 营 。 我 国<br />

高 铁 的 工 期 一 般 只 有 四 年 , 留 给 客 站 的 工 期 只 有 两 年 左 右 。<br />

3. 功 能 变 化 大 。 传 统 客 站 的 功 能 定 位 已 无 法 满 足 高 速 铁 路 、 综 合 交 通 体 系 以 及 城 镇 化 对 高 铁 客<br />

站 的 新 要 求 , 高 铁 客 站 的 功 能 定 位 需 要 实 现 从 单 一 铁 路 客 运 场 所 到 综 合 交 通 枢 纽 、 从 管 理 型 到 服 务<br />

型 两 个 转 变 。 这 给 高 铁 客 站 的 设 计 理 念 和 技 术 创 新 提 出 了 更 高 要 求 。<br />

4. 技 术 难 度 大 。 综 合 交 通 枢 纽 的 功 能 定 位 、 立 体 化 的 功 能 布 局 模 式 和 动 车 组 高 速 通 过 , 给 高 铁<br />

客 站 的 空 间 结 构 、 节 能 环 保 、 环 境 控 制 、 消 防 安 全 等 带 来 了 一 系 列 技 术 难 题 。 尤 其 是 承 受 动 车 组 高<br />

速 通 过 和 反 复 停 靠 的 大 跨 度 、 高 空 间 结 构 体 系 最 为 复 杂 , 是 实 现 综 合 交 通 枢 纽 功 能 定 位 的 基 本 前 提 。<br />

5. 专 业 接 口 多 。 高 铁 客 站 是 一 个 复 杂 、 庞 大 的 系 统 , 涉 及 33 个 专 业 , 同 时 还 与 地 铁 、 市 政 道<br />

路 、 城 市 规 划 等 行 业 密 不 可 分 , 专 业 接 口 管 理 和 系 统 集 成 管 理 的 难 度 大 。<br />

6. 协 调 难 度 大 。 高 铁 客 站 与 城 市 轨 道 交 通 、 市 政 设 施 等 工 程 关 系 紧 密 , 涉 及 市 政 、 规 划 、 地 铁 、<br />

甚 至 航 空 等 十 几 个 部 门 。 受 我 国 条 块 分 割 的 管 理 体 制 影 响 , 多 业 主 、 多 设 计 单 位 、 多 工 程 、 多 专 业 、<br />

多 工 种 、 多 操 作 面 之 间 的 交 叉 错 综 复 杂 。<br />

7. 施 工 组 织 难 。 特 大 型 高 铁 客 站 具 有 多 工 程 同 步 施 工 、 多 工 种 交 叉 施 工 的 特 点 , 具 有 场 地 局 促 、<br />

进 出 口 少 、 施 工 单 位 多 、 运 输 量 大 的 共 性 , 如 何 确 保 施 工 场 地 内 外 交 通 畅 通 和 多 层 次 立 体 交 叉 作 业<br />

状 态 下 各 工 序 的 有 序 转 换 , 对 施 工 组 织 提 出 了 巨 大 挑 战 。<br />

三 、 我 国 高 铁 客 站 建 设 路 径 与 成 效<br />

围 绕 我 国 高 铁 客 站 建 设 目 标 , 针 对 高 铁 客 站 建 设 资 源 制 约 的 现 状 , 以 系 统 工 程 和 项 目 群 管 理 等<br />

理 论 为 指 导 , 充 分 发 挥 我 国 铁 路 行 业 的 体 制 优 势 、 市 场 优 势 和 后 发 优 势 , 坚 持 “ 理 念 — 技 术 — 管 理 ”<br />

三 位 一 体 的 创 新 路 线 , 探 索 实 践 我 国 高 铁 客 站 的 发 展 路 径 , 以 实 现 高 铁 客 站 高 质 量 、 快 速 度 发 展 。<br />

3.1 理 念 与 理 论 创 新<br />

在 充 分 研 究 发 达 国 家 铁 路 客 站 的 演 变 过 程 和 发 展 趋 势 , 系 统 总 结 我 国 铁 路 客 站 建 设 几 十 年 发 展<br />

历 程 和 经 验 教 训 的 基 础 上 , 研 究 提 出 了 “ 适 应 时 代 需 求 , 服 务 交 通 功 能 , 体 现 地 域 文 化 , 构 建 以 铁<br />

路 为 主 的 综 合 交 通 枢 纽 ” 的 客 站 建 设 新 理 念 。 为 把 新 理 念 转 化 为 具 体 的 工 作 标 准 , 组 织 编 写 《 铁 路<br />

旅 客 车 站 设 计 指 南 》, 开 展 国 内 外 技 术 交 流 、 专 家 研 讨 , 全 过 程 引 导 和 促 进 参 建 人 员 的 观 念 更 新 ,<br />

为 新 理 念 在 工 程 实 践 中 的 贯 彻 奠 定 基 础 。 结 合 高 铁 客 站 工 程 实 践 , 出 版 了 《 中 国 当 代 铁 路 客 站 设 计<br />

理 论 探 索 》, 重 点 研 究 了 高 铁 规 划 与 总 体 布 局 、 功 能 布 局 与 流 线 组 织 、 空 间 形 态 与 文 化 表 现 等 重 大<br />

问 题 , 初 步 形 成 了 现 代 化 客 站 的 设 计 理 论 。<br />

3.2 关 键 技 术 创 新<br />

充 分 发 挥 铁 路 行 业 既 是 主 导 方 又 是 需 求 方 的 优 势 , 采 用 “ 产 学 研 用 紧 密 结 合 ” 的 技 术 创 新 路 径 ,<br />

组 织 建 立 高 铁 客 站 技 术 创 新 平 台 , 整 合 国 内 科 研 资 源 和 力 量 , 组 织 国 内 外 32 家 设 计 单 位 、22 家 高<br />

校 、5 家 科 研 机 构 、12 家 大 型 施 工 企 业 , 统 筹 实 施 重 大 科 研 和 技 术 攻 关 。<br />

-94-


通 过 上 述 技 术 创 新 平 台 , 我 们 在 较 短 的 时 间 内 取 得 了 “ 站 桥 合 一 ” 空 间 结 构 、 超 大 跨 度 空 间 钢<br />

结 构 、 列 车 震 动 控 制 等 27 项 技 术 创 新 成 果 , 解 决 了 制 约 我 国 高 铁 客 站 建 设 4 个 方 面 的 技 术 难 题 ,<br />

为 高 铁 客 站 的 建 设 提 供 技 术 支 撑 。<br />

3.3 管 理 创 新<br />

为 解 决 大 量 高 铁 客 站 同 期 建 设 导 致 的 组 织 协 调 、 资 源 共 享 、 工 期 控 制 和 质 量 保 证 等 工 程 管 理 难<br />

题 , 以 标 准 化 为 抓 手 , 以 信 息 化 为 支 撑 , 探 索 构 建 支 撑 多 客 站 同 期 建 设 的 管 理 模 式 , 推 动 高 铁 客 站<br />

的 集 约 化 发 展 。<br />

-95-


高 铁 客 站 项 目 群 管 理 模 型<br />

改 变 传 统 客 站 与 线 路 一 体 化 的 工 程 管 理 模 式 , 把 全 国 所 有 高 铁 客 站 从 高 铁 干 线 中 独 立 出 来 , 组<br />

成 客 站 工 程 项 目 群 。 构 建 “ 铁 道 部 —— 铁 路 局 —— 项 目 部 ” 三 级 管 理 组 织 体 系 , 统 一 制 定 战 略 目 标 、<br />

规 范 管 理 流 程 、 调 配 关 键 资 源 , 以 实 现 知 识 与 资 源 的 共 享 。 以 责 任 链 管 理 为 核 心 , 组 织 建 立 “ 结 构<br />

清 晰 、 职 责 分 明 、 权 责 对 等 ” 的 管 理 制 度 体 系 , 完 善 高 铁 客 站 建 设 的 规 则 框 架 , 以 实 现 高 铁 客 站 工<br />

程 项 目 群 管 理 的 制 度 化 和 规 范 化 。 客 站 建 设 专 业 接 口 多 , 交 叉 作 业 多 , 工 序 复 杂 , 工 作 面 小 , 建 立<br />

合 理 和 规 范 的 全 过 程 管 理 流 程 显 得 尤 为 重 要 。 通 过 流 程 的 再 造 , 实 现 “ 易 操 作 、 可 复 制 ” 的 标 准 化<br />

目 标 , 持 续 提 升 客 站 建 设 管 理 的 成 熟 度 。 针 对 高 铁 客 站 建 设 分 布 范 围 广 、 管 理 跨 度 大 、 专 业 性 强 等<br />

特 点 , 建 立 客 站 工 程 管 理 信 息 系 统 , 构 建 覆 盖 全 国 高 铁 客 站 的 信 息 支 撑 平 台 , 以 实 现 各 项 管 理 的 数<br />

据 化 、 规 范 化 , 提 高 决 策 指 挥 的 及 时 性 、 针 对 性 。<br />

3.4 取 得 的 成 效<br />

1. 确 立 了 一 套 全 新 的 建 设 理 念 ;<br />

2. 取 得 了 一 批 技 术 创 新 成 果 (27 项 );<br />

3. 构 建 了 一 种 高 效 的 项 目 群 管 理 模 式 ;<br />

4. 建 成 了 一 批 现 代 化 的 高 铁 客 站 (140 座 )。<br />

四 、 结 语<br />

展 望 中 国 铁 路 未 来 , 到 2015 年 , 随 着 一 大 批 现 代 化 铁 路 客 站 的 建 成 运 营 , 中 国 高 铁 客 站 将 为<br />

旅 客 提 供 更 加 便 捷 的 进 出 站 服 务 、 更 加 舒 适 的 候 车 环 境 、 更 加 人 性 化 的 服 务 设 施 , 与 城 市 的 关 系 更<br />

加 和 谐 , 对 城 市 综 合 交 通 乃 至 经 济 社 会 的 发 展 必 将 发 挥 更 加 重 要 的 作 用 。<br />

参 考 文 献<br />

[1] 何 华 武 , 郑 健 . 铁 路 旅 客 车 站 设 计 指 南 . 中 国 铁 道 出 版 社 , 2006, 北 京 .<br />

[2] 郑 健 , 沈 中 伟 等 . 中 国 当 代 铁 路 客 站 设 计 理 论 探 索 . 人 民 交 通 出 版 社 , 2009, 北 京 .<br />

[3] 唐 述 春 . 德 国 铁 路 客 运 营 销 特 点 及 思 考 . 中 国 铁 路 , 2005.<br />

[4] 朱 兆 慷 , 张 庄 . 铁 路 旅 客 车 站 流 线 设 计 和 建 筑 空 间 组 合 模 式 的 发 展 过 程 与 趋 势 . 建 筑 学 报 , 2005.<br />

[5] 顾 保 南 黄 志 华 等 . 上 海 南 站 的 综 合 交 通 换 乘 系 统 . 城 市 轨 道 交 通 研 究 , 2006.<br />

[6] 孙 伟 , 刘 德 明 . 铁 路 客 运 站 建 筑 空 间 开 放 性 解 析 . 华 中 建 筑 , 2008.<br />

-96-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

2<br />

軌 道 完 整 性 之 診 斷<br />

1<br />

1<br />

1, 2<br />

廖 慶 隆 陳 韋 帆 王 仲 宇<br />

1<br />

財 團 法 人 中 華 顧 問 工 程 司 , 台 北 , 台 灣<br />

國 立 中 央 大 學 土 木 工 程 學 系 , 桃 園 , 台 灣<br />

摘 要 : 軌 道 結 構 其 具 有 固 定 列 車 行 進 路 線 、 並 穩 定 地 將 列 車 載 重 傳 遞 至 下 部 土 建 結 構 之 功 能 ,<br />

是 軌 道 運 輸 系 統 中 十 分 重 要 的 組 成 部 分 。 而 軌 道 在 巨 大 的 移 動 載 重 作 用 之 下 可 能 迫 使 部 分 構 件 之 鬆<br />

動 現 象 , 或 是 基 礎 沉 陷 、 橋 面 大 梁 變 形 等 , 均 可 能 造 成 軌 道 結 構 之 幾 何 形 狀 與 原 先 設 計 不 同 而 產 生<br />

軌 道 不 整 , 而 一 旦 軌 道 不 整 之 現 象 產 生 , 輕 則 影 響 列 車 乘 客 的 舒 適 性 , 重 則 影 響 列 車 整 體 的 穩 定 性<br />

及 安 全 性 , 因 此 軌 道 不 整 的 判 斷 , 是 整 體 鐵 路 系 統 中 不 可 忽 視 的 一 環 。<br />

而 軌 道 列 車 於 空 間 三 維 曲 線 中 運 行 之 行 為 , 其 實 是 一 參 數 極 度 耦 合 的 力 學 問 題 , 因 此 若 以 此 角<br />

度 切 入 軌 道 不 整 問 題 對 軌 道 列 車 之 影 響 , 在 數 學 上 之 關 係 就 更 加 複 雜 了 。 本 研 究 利 用 軌 道 列 車 在 三<br />

維 空 間 中 之 運 動 , 於 動 座 標 系 中 之 加 速 度 反 應 結 果 , 經 由 數 學 上 的 反 算 流 程 得 到 其 所 包 含 的 各 方 向<br />

軌 道 面 曲 率 數 值 以 及 其 變 化 量 , 進 而 做 為 軌 道 線 型 及 軌 道 不 整 判 斷 之 依 據 。 此 方 法 的 優 點 為 ; 第 一 ,<br />

充 分 考 慮 列 車 三 維 運 動 中 各 方 向 曲 率 參 數 之 耦 合 現 象 使 運 動 反 應 完 整 呈 現 。 第 二 , 以 移 動 座 標 系 統<br />

為 本 位 , 針 對 軌 道 幾 何 變 化 造 成 之 動 態 反 應 進 行 計 算 , 使 模 擬 過 程 中 不 需 建 立 矩 陣 並 省 去 相 關 計 算<br />

疊 代 問 題 。<br />

本 文 中 首 先 推 導 空 間 軌 道 的 線 型 變 化 對 移 動 軌 道 點 的 加 速 度 方 程 式 , 接 著 進 一 步 利 用 數 學 反 算<br />

技 巧 推 解 方 程 式 , 可 得 到 代 表 軌 道 面 變 化 之 曲 率 值 , 最 後 利 用 擬 真 之 數 值 算 例 , 驗 證 此 方 法 之 可 行<br />

性 。<br />

關 鍵 詞 : 軌 道 完 整 性 檢 測 加 速 度 曲 率 反 算<br />

前 言<br />

為 降 低 軌 道 面 由 於 各 種 異 常 現 象 所 造 成 之 可 能 性 危 害 , 經 常 性 的 軌 道 檢 驗 及 維 護 工 作 是 軌 道 系<br />

統 中 不 能 忽 視 的 重 要 課 題 。 在 經 常 性 檢 測 中 由 於 人 工 檢 測 需 要 大 量 人 事 成 本 , 且 容 易 包 含 巡 查 人 員<br />

自 我 主 觀 意 識 , 使 得 在 軌 道 檢 測 上 難 免 有 所 遺 漏 。 因 此 發 展 更 加 精 確 、 快 速 且 低 成 本 的 軌 道 完 整 性<br />

檢 測 技 術 已 成 為 現 今 許 多 專 家 學 者 們 所 專 注 的 研 究 課 題 , 並 發 展 出 基 於 多 種 不 同 理 論 下 之 軌 道 檢 測<br />

模 式 。<br />

然 而 軌 道 列 車 在 軌 道 中 運 行 乃 是 一 種 空 間 曲 線 中 的 運 動 行 為 , 其 具 有 力 學 上 的 複 雜 性 , 且 加 速<br />

度 反 應 與 空 間 中 軌 道 線 型 之 關 聯 性 是 不 可 忽 略 的 。 過 去 研 究 空 間 曲 線 中 的 運 動 行 為 者 , 利 用 軌 道 列<br />

車 本 身 動 力 反 應 來 偵 測 軌 道 曲 率 變 化 ,Weston et al.’s (2007) 以 軌 道 列 車 垂 直 、 水 平 方 向 之 加 速 度 個<br />

別 換 算 得 到 軌 道 線 型 , 但 不 考 慮 各 方 向 曲 率 在 加 速 度 反 應 中 耦 合 之 現 象 。 為 控 制 軌 道 列 車 在 行 駛 過<br />

程 中 由 於 軌 道 線 型 所 帶 來 之 額 外 加 速 度 反 應 , 並 避 免 其 可 能 帶 來 的 行 車 安 全 顧 慮 , 確 實 地 掌 握 軌 道<br />

線 型 長 時 間 下 的 變 化 是 必 要 的 工 作 , 也 是 評 估 列 車 行 駛 舒 適 及 安 全 的 標 準 之 一 。 廖 慶 隆 (2004) 曾 以<br />

附 於 軌 道 列 車 之 動 座 標 系 探 討 軌 道 列 車 之 振 動 反 應 。 發 現 其 軌 道 面 上 各 方 向 的 曲 率 變 化 將 會 以 非 線<br />

性 偶 合 之 形 式 呈 現 於 空 間 中 動 座 標 系 方 位 之 動 力 反 應 行 為 中 。 而 本 研 究 基 於 此 理 論 基 礎 , 推 導 以 移<br />

動 車 軸 上 之 加 速 度 反 應 來 診 斷 軌 道 幾 何 特 徵 與 完 整 性 之 反 算 理 論 。 其 曲 率 反 應 則 代 表 軌 道 面 在 空 間<br />

中 之 旋 轉 、 扭 轉 之 情 形 , 足 以 代 表 整 個 軌 道 之 幾 何 線 型 變 化 。 此 理 論 特 色 在 於 : 其 一 , 考 慮 曲 率 參<br />

數 偶 合 之 關 係 , 反 應 真 實 之 空 間 動 力 反 應 。 其 二 , 有 別 於 一 般 有 限 元 素 模 擬 方 式 並 利 用 移 動 座 標 系<br />

為 主 之 計 算 概 念 , 減 少 矩 陣 運 算 後 置 轉 換 過 程 , 將 變 化 參 數 集 中 於 簡 單 之 空 間 集 合 曲 率 參 數 之 歷 時<br />

變 化 , 節 省 數 值 模 擬 計 算 時 間 。 其 三 , 未 來 可 結 合 實 際 量 測 數 據 與 反 算 理 論 進 行 實 際 之 軌 道 幾 何 不<br />

-97-


整 之 檢 測 工 作 , 利 用 安 裝 於 車 輪 軸 兩 側 的 加 速 度 計 (ABA, Axle-Boxes Acceleration) 進 行 實 地 量 測 , 提<br />

升 軌 道 安 全 檢 測 之 效 率 及 準 確 性 。 本 文 中 並 採 用 數 個 數 值 模 擬 案 例 , 驗 證 此 反 算 法 則 之 準 確 性 。<br />

軌 道 系 統 之 空 間 幾 何 特 徵 反 算<br />

由 廖 慶 隆 (2004) 所 建 立 之 軌 道 列 車 於 空 間 中 運 動 理 論 ( 詳 細 推 導 見 附 錄 A), 可 以 了 解 到 , 軌 道<br />

空 間 曲 線 的 幾 何 變 化 將 會 反 應 到 軌 道 車 之 加 速 度 反 應 。 依 此 理 論 , 若 需 診 斷 軌 道 空 間 曲 線 的 幾 何 變<br />

化 , 可 由 以 上 之 加 速 度 反 應 方 程 式 著 手 進 行 參 數 反 算 , 其 推 導 流 程 如 下 : 假 設 之 B 點 與 車 軌 接 觸 之<br />

R 點 間 無 相 對 位 移 , 兩 點 間 呈 完 全 剛 性 狀 態 , 但 若 實 際 之 B 點 位 於 列 車 輪 對 轉 向 架 以 上 之 列 車 結 構<br />

時 , 此 假 設 將 不 復 存 在 。 為 滿 足 此 假 設 並 利 於 未 來 實 際 檢 測 運 作 , 吾 人 假 設 兩 點 B1 及 B2 分 別 位 於<br />

軌 道 列 車 其 中 一 對 車 輪 軸 之 兩 側 , 如 圖 1 所 示 , 以 軌 道 面 左 右 兩 軌 中 線 做 為 基 準 線 , 兩 點 所 對 應 的<br />

r r<br />

動 座 標 系 下 之 位 置 向 量 分 別 為 au<br />

h h<br />

+ au 及<br />

v v<br />

− au<br />

r r<br />

h h<br />

+ au<br />

v v , 假 設 軌 道 車 輪 與 軌 道 之 間 為 完 全 剛 性 , 而<br />

B1、B2 向 量 和 u r 方 向 一 致 , 安 裝 於 B1 及 B2 點 上 之 加 速 度 計 正 向 振 動 方 向 也 與 u r h t<br />

、 u r h<br />

、 u r v<br />

方 向 一<br />

致 , 且 沿 軌 道 垂 直 面 上 具 有 相 同 的 幾 何 線 型 變 化 。 a v<br />

為 B 點 至 軌 道 之 鉛 直 距 離 , a h<br />

為 兩 輪 距 一 半<br />

之 長 度 。 依 此 由 式 (A4) 至 式 (A6) 可 得 到 六 組 參 數 極 度 耦 合 之 非 線 性 聯 立 方 程 式 :<br />

* * 2 * * * * * *<br />

B&& t1 = && s(1 + avθh − ahθv) + s&<br />

[( avθh ' − ahθv ') + ahθtθh + avθtθv]<br />

(1)<br />

* 2 * *2 *2 * * *<br />

B&& h1 =− && savθt + s&<br />

[ θv −ahθv − ahθt + avθhθv −avθt<br />

']<br />

(2)<br />

* 2 * * * * *2 *2<br />

B&& v1 = && sahθt + s&<br />

[ − θh + ahθvθh + ahθt ' −avθt −avθh<br />

]<br />

(3)<br />

* * 2 * * * * * *<br />

B&& t2 = && s(1 + avθh + ahθv) + s&<br />

[( avθh ' + ahθv ') − ahθtθh + avθtθv]<br />

(4)<br />

* 2 * *2 *2 * * *<br />

B&& h2 =− && savθt + s&<br />

[ θv + ahθv + ahθt + avθhθv −avθt<br />

']<br />

(5)<br />

* 2 * * * * *2 *2<br />

B&& v2 =− && sahθt + s&<br />

[ −θh −ahθvθh −ahθt ' −avθt −avθh<br />

]<br />

(6)<br />

依 此 理 論 之 結 果 可 得 知 ; 若 可 量 測 到 軌 道 列 車 在 軌 道 面 座 標 系 統 中 之 加 速 度 反 應 B & t<br />

、 B && h<br />

及 B && v<br />

,<br />

配 合 已 知 的 軌 道 列 車 瞬 時 速 度 ( s& ) 及 加 速 度 反 應 ( s&& ), 則 依 此 方 程 式 反 過 來 求 取 軌 道 路 線 的 空 間 幾 何<br />

線 型 變 化 , 由 於 輪 軸 是 車 輛 系 統 中 直 接 和 軌 道 接 觸 的 構 件 , 因 此 當 軌 道 之 幾 何 特 徵 及 局 部 瑕 疵 出 現<br />

時 , 將 會 在 輪 軸 之 振 動 行 為 中 反 應 出 來 。 本 研 究 由 佈 設 於 輪 軸 兩 側 軸 心 位 置 之 三 軸 加 速 度 計 量 測 值<br />

反 算 出 軌 道 之 幾 何 特 徵 , 以 便 於 透 過 移 動 車 軸 之 振 動 訊 號 來 檢 視 軌 道 幾 何 變 化 及 診 斷 軌 道 可 能 出 現<br />

的 缺 陷 或 不 完 整 性 。 在 此 由 式 (3) 與 式 (6) 作 相 加 再 除 二 , 可 得 到 下 式 (7):<br />

圖 1 軌 道 移 動 座 標 系 之 相 關 參 數 示 意 圖<br />

2<br />

2 2<br />

*2 * 1 s 2( Bv 1<br />

Bv2)<br />

av<br />

θ<br />

⎛ t<br />

θ<br />

⎞ & − && + &&<br />

+ ⎜ h<br />

+ ⎟ = = r<br />

2<br />

1<br />

⎝ 2av<br />

⎠ ( 2as&<br />

)<br />

同 理 將 式 (2) 與 式 (5) 作 相 減 再 除 二 , 可 得 到 下 式 (8):<br />

*2 *2 ( B&&<br />

h2 − B&&<br />

h1)<br />

θt<br />

+ θv<br />

= = r<br />

2<br />

2as&<br />

* *<br />

* *<br />

依 式 子 (7) 及 (8) 可 將 θ<br />

t<br />

與 θ 的 關 係 與 θ 與<br />

h<br />

t θ 的 關 係 分 別 以 極 座 標 (<br />

v<br />

r 1<br />

,θ) 及 ( r 2<br />

,θ) 表 示 , 其 關 係 如<br />

圖 2、 圖 3 所 示 :<br />

h<br />

v<br />

2<br />

2<br />

(7)<br />

(8)<br />

-98-


*<br />

θt r1cosθ r2cosθ2<br />

= = (9)<br />

* 1<br />

θh<br />

=− + r1<br />

sinθ<br />

(10)<br />

2a<br />

⎧<br />

2 2<br />

* 1 r<br />

⎫<br />

⎪<br />

⎡⎛<br />

⎞ ⎤<br />

− 1 ⎪<br />

2<br />

θv<br />

= r2sin θ2 = r2sin ⎨cos ⎢<br />

cosθ⎥⎬=± r2 −r1<br />

cos θ<br />

⎪ ⎢⎜r<br />

⎟ ⎥<br />

⎩ ⎣⎝<br />

2 ⎠ ⎦⎪⎭<br />

*<br />

θ h<br />

v<br />

*<br />

θ v<br />

(11)<br />

r 1<br />

2.89<br />

θ(s)<br />

* *<br />

( θt<br />

( s), θh( s))<br />

1<br />

2a<br />

v<br />

*<br />

θ t<br />

r 2<br />

2.89<br />

θ 2<br />

(s)<br />

* *<br />

( θ ( s), θ ( s))<br />

t v<br />

*<br />

θ t<br />

*<br />

圖 2 參 數 θ 與 參 數 θ 關 係 圖<br />

*<br />

t<br />

h<br />

圖 3<br />

*<br />

將 參 數 θ 做 一 次 微 分 可 得 :<br />

t<br />

* d<br />

dr1<br />

dθ<br />

θt<br />

' = ( r1cosθ)<br />

= ⎛ ⎞ cosθ −r1sinθ<br />

⎛ ⎞<br />

ds ⎜ ⎜ ⎟<br />

ds ⎟<br />

⎝ ⎠<br />

⎝ ds ⎠<br />

參 數 θ 與 參 數 θ 關 係 圖<br />

在 此 選 擇 將 式 (9)、(10)、(11) 及 (12) 代 入 式 (2) 並 做 簡 化 , 可 得 到 一 個 單 一 變 數 之 一 階 微 分 方 程 式 :<br />

⎛<br />

* 2 * * * *2 *2 2 dr ⎞<br />

1<br />

&& savθt + s& ( −θv − avθvθh + ahθt + ahθv ) + s&<br />

av cosθ<br />

+ B&&<br />

h1<br />

dθ<br />

⎜ ds ⎟<br />

⎛ ⎞ ⎝ ⎠<br />

(13)<br />

⎜ ⎟=<br />

ds<br />

2<br />

⎝ ⎠<br />

sar & sinθ<br />

在 此 假 設 初 始 三 方 向 曲 率 為 零 , 且 初 始 水 平 向 加 速 度 B && 也 為 零 , 此 時 之 θ 值 為 π/2, 接 著 並 可<br />

h<br />

透 過 數 值 方 式 處 理 微 分 部 份 , 直 接 計 算 出 下 一 個 時 間 點 的 θ 值 :<br />

( s+Δs) ⎛dθ<br />

⎞ ( s)<br />

θ =Δ s ⎜ ⎟+<br />

θ<br />

(14)<br />

⎝ ds ⎠<br />

*<br />

由 上 式 作 為 計 算 依 據 , 得 到 每 個 軌 道 弧 長 座 標 點 所 相 對 之 θ 值 , 可 帶 回 分 別 求 得 曲 率 參 數 θ<br />

t<br />

、<br />

* *<br />

θ 及<br />

h<br />

θ , 並 對 曲 率 參 數 對 位 移 做 數 值 微 分 , 可 求 軌 道 曲 率 之 變 化 量 θ t* ' 、 * v<br />

θ<br />

h<br />

' 及 θ<br />

* v<br />

' 。 由 以 上 推 導<br />

可 發 現 , 整 個 反 算 過 程 當 中 只 需 利 用 到 四 條 關 係 式 , 而 切 線 方 向 方 程 式 (1)(4) 均 無 使 用 , 代 表 在 此 演<br />

算 法 則 中 , 無 須 得 悉 軌 道 車 切 線 方 向 之 加 速 度 值 B&&<br />

() t1<br />

s 及 B&&<br />

()<br />

*<br />

t 2<br />

s , 即 可 達 到 反 算 θ 、 * t θ 、 * *<br />

θ 、<br />

h v<br />

θ<br />

t<br />

' 、<br />

θ<br />

* h<br />

' 及 θ * v<br />

' 之 效 果 。<br />

*<br />

而 在 設 θ 、 * t θ 、 θ * 為 變 數 之 情 況 , 並 針 對 B1 點 可 列 出 其 對 s 微 分 下 之 一 階 非 線 性 常 微 分 聯 立<br />

h v<br />

方 程 式 , 可 利 用 常 微 分 方 程 之 數 值 解 法 求 解 , 可 整 理 得 到 以 下 列 恆 等 式 :<br />

* * * 2 * * * * 2 *2 *2 2 *2 *2<br />

F( θ , θ , θ ) = s& [ aθ − aθ + 2 a aθ θ − a ( θ + θ ) − a ( θ + θ )] −a B&&<br />

− a B&&<br />

= 0 (15)<br />

t h v h v v h h v h v h t v v t h h h v v<br />

v 1<br />

由 上 式 可 發 現 , 其 不 含 任 何 曲 率 微 分 參 數 , 因 此 將 與 時 間 變 化 無 關 , 也 就 是 說 此 方 程 式 在 任 何<br />

時 間 點 均 可 單 獨 成 立 , 而 由 此 方 程 式 可 證 實 此 反 算 問 題 的 常 微 分 聯 立 方 程 組 是 具 有 相 依 性 的 。 並 可<br />

做 為 反 算 準 確 與 否 之 判 斷 依 據 : 若 判 斷 式 不 為 零 , 則 代 表 反 算 出 現 誤 差 , 而 誤 差 發 生 之 原 因 乃 由 於<br />

極 座 標 角 度 θ 在 正 負 號 判 斷 上 出 現 問 題 , 所 以 必 須 修 正 此 反 算 判 斷 , 具 體 之 修 正 方 式 將 於 範 例 中 加<br />

2<br />

以 介 紹 。<br />

數 值 案 例 驗 證<br />

數 值 案 例 乃 為 驗 證 利 用 軌 道 加 速 度 數 值 , 反 向 計 算 所 求 得 之 軌 道 幾 何 變 化 之 可 靠 性 及 準 確 性 。<br />

*<br />

t<br />

*<br />

v<br />

(12)<br />

-99-


在 此 假 設 之 數 值 算 例 驗 證 方 式 , 使 用 已 知 之 軌 道 幾 何 變 化 θ ( ) 、 θ * () s 及 θ * () s , 經 由 式 (4) 至 式 (6)<br />

之 關 係 式 做 計 算 , 可 推 算 得 局 部 座 標 系 中 B1 及 B2 兩 點 在 三 個 方 向 上 之 加 速 度 歷 時 , 在 此 B1 點 及<br />

B2 點 依 據 台 灣 高 速 鐵 路 之 軌 道 寬 度 及 輪 軌 高 度 之 可 能 量 測 位 置 , 令 軌 道 鉛 直 方 向 距 離 a 為 0.575<br />

v<br />

公 尺 , 軌 道 中 央 線 到 兩 側 B1 點 、B2 點 之 水 平 方 向 距 離 a 為 0.7275 公 尺 , 此 兩 參 數 為 恆 定 值 。 為<br />

h<br />

驗 證 本 反 算 法 則 之 可 靠 性 及 與 相 關 參 數 之 間 之 影 響 , 將 利 用 數 值 算 例 進 行 反 算 可 靠 度 分 析 。 以 下 依<br />

數 種 不 同 之 假 設 狀 況 進 行 案 例 處 理 及 探 討 。<br />

t* s<br />

h<br />

v<br />

圖 4 案 例 A 之 空 間 中 軌 道 線 型<br />

圖 5 案 例 B 之 空 間 中 有 大 曲 率 變 化 之 軌 道 線 型<br />

數 值 案 例 A<br />

模 擬 不 同 波 長 正 弦 波 疊 合 之 軌 道 曲 線 反 算 , 其 空 間 中 軌 道 線 型 如 圖 4 所 示 。 此 算 例 假 設 直 線 型<br />

*<br />

軌 道 表 面 變 化 為 不 同 長 度 波 長 之 正 弦 波 疊 合 做 為 假 設 , 其 幾 何 曲 率 變 化 令 θ 為 波 長 1.6、3.4、5.7<br />

t<br />

及 8.1 公 尺 且 振 幅 為 0.0001 之 正 弦 波 疊 合 , θ * 為 波 長 1.1、2.3、8.6 及 10.9 公 尺 且 振 幅 為 0.0001 之<br />

h<br />

正 弦 波 疊 合 , θ * 為 波 長 0.7、3.1、7.3 及 7.4 公 尺 且 振 幅 為 0.0001 之 正 弦 波 疊 合 , 並 假 設 列 車 以 時 速<br />

v<br />

120 公 里 等 速 度 行 進 , 其 由 式 (1-6) 所 得 到 之 輪 軸 兩 端 B1、B2 兩 點 之 加 速 度 反 應 如 圖 6-8 所 示 。 經 過<br />

以 上 流 程 反 算 之 結 果 , 可 以 得 到 在 移 動 座 標 系 中 三 個 方 向 的 曲 率 歷 程 變 化 量 (Output), 並 與 原 始 輸 入<br />

之 曲 率 變 化 量 做 比 較 的 結 果 , 如 圖 9-11 所 示 。<br />

圖 6 切 線 方 向 加 速 度 歷 時 反 應<br />

圖 7 平 行 軌 枕 方 向 加 速 度 歷 時 反 應<br />

圖 8 垂 直 軌 道 面 方 向 加 速 度 歷 時 反 應<br />

圖 9 切 線 方 向 之 幾 何 參 數 θ 反 算 比 較<br />

*<br />

t<br />

-100-


圖 10 水 平 軌 枕 方 向 之 幾 何 參 數 θ 反 算 比 較<br />

*<br />

h<br />

圖 11 垂 直 軌 道 面 方 向 之 幾 何 參 數 θ 反 算 比 較<br />

*<br />

v<br />

除 了 圖 11 在 水 平 軌 枕 方 向 沒 有 完 全 疊 合 之 外 , 其 餘 方 向 的 反 算 解 果 都 十 分 準 確 , 誤 差 都 非 常<br />

小 ( 小 於 0.01%)。 鉛 直 方 向 參 數 θ 之 反 算 結 果 , 發 現 計 算 結 果 會 出 現 恆 正 之 現 象 , 必 須 經 由 額 外 之<br />

*<br />

v<br />

* * *<br />

判 斷 式 F( θ , θ , θ ) 做 為 修 正 依 據 , 其 所 求 出 之 判 斷 式 歷 程 結 果 如 下 所 示 :<br />

t h v<br />

* * *<br />

圖 12 判 斷 式 F( θ , θ , θ ) 之 歷 程 變 化<br />

t h v<br />

由 結 果 可 以 發 現 , 由 於 判 別 式 發 生 不 趨 近 為 零 之 狀 況 , 即 是 θ 在 正 負 值 判 斷 出 現 異 常 的 位 置 點 ,<br />

依 此 方 程 式 可 做 為 θ 正 負 值 判 斷 之 依 據 。 修 正 後 所 得 到 之 反 算 結 果 如 下 圖 所 示 。 在 此 判 別 式 修 正 是<br />

*<br />

v<br />

否 大 於 十 的 負 六 次 方 (m 2 /s 2 ) 作 為 判 斷 依 據 , 經 過 修 正 之 結 果 如 下 圖 所 示 :<br />

*<br />

h<br />

圖 13 經 修 正 後 之 垂 直 軌 道 面 方 向 之 軌 道 幾 何 參 數 θ 比 較<br />

*<br />

v<br />

經 過 修 正 之 結 果 在 數 值 上 具 有 絕 佳 之 準 確 性 , 可 正 確 判 讀 並 反 算 軌 道 幾 何 變 化 。 依 此 範 例 之 結<br />

果 , 可 驗 證 當 軌 道 面 出 現 表 面 瑕 疵 或 是 不 平 整 度 時 , 本 演 算 方 式 可 以 準 確 地 由 其 曲 率 之 變 化 中 判 別<br />

出 來 軌 道 面 不 平 整 的 程 度 。<br />

數 值 案 例 B<br />

在 有 噪 訊 來 源 狀 況 下 , 模 擬 列 車 過 彎 並 考 慮 列 車 變 速 度 運 動 下 反 算 之 影 響 , 其 空 間 中 軌 道 線 型<br />

-101-


如 圖 5 所 示 。 此 數 值 案 例 以 疊 合 正 弦 波 做 噪 訊 來 源 , 並 給 予 波 長 為 100 公 尺 , 振 幅 為 0.001 之 正 弦<br />

波 來 模 擬 簡 單 之 列 車 過 彎 狀 況 , 其 狀 況 過 彎 最 小 曲 率 半 徑 為 1000 公 尺 , 並 考 慮 列 車 變 速 度 運 動 下<br />

反 算 之 影 響 , 如 下 圖 14 所 示 。 在 此 假 設 之 列 車 之 加 速 度 歷 程 如 下 圖 15-17 所 示 :<br />

圖 14<br />

列 車 速 度 歷 程 反 應<br />

其 中 , 軌 道 在 45 公 尺 處 加 有 垂 直 方 向 突 波 反 應 , 模 擬 軌 道 面 在 行 進 過 程 當 中 有 出 現 軌 道 瑕 疵<br />

或 損 壞 之 現 象 。 由 以 上 之 速 度 配 合 過 彎 模 擬 , 其 加 速 度 反 應 如 圖 15-17 所 示 。 其 中 可 見 切 線 方 向 加<br />

速 度 在 1.1 秒 左 右 的 位 置 有 突 然 異 常 的 加 速 度 產 生 , 並 經 過 以 上 流 程 反 算 之 結 果 , 可 以 得 到 在 移 動<br />

座 標 系 中 三 個 方 向 的 曲 率 歷 程 變 化 量 (Output), 並 與 原 始 輸 入 之 曲 率 變 化 量 (Input) 做 比 較 的 結 果 ( 圖<br />

18-20), 可 分 辨 出 垂 直 方 向 軌 道 不 整 的 突 波 線 型 , 如 圖 19 所 示 。<br />

圖 15 切 線 方 向 加 速 度 歷 時 反 應<br />

圖 16 平 行 軌 枕 方 向 加 速 度 歷 時 反 應<br />

圖 17 鉛 直 方 向 加 速 度 歷 時 反 應<br />

圖 18 切 線 方 向 之 幾 何 參 數 θ 反 算 比 較<br />

*<br />

t<br />

-102-


圖 19 水 平 軌 枕 方 向 之 幾 何 參 數 θ 反 算 比 較<br />

*<br />

h<br />

圖 20 垂 直 軌 道 面 方 向 之 幾 何 參 數 θ 反 算 比 較<br />

所 有 反 算 比 較 之 結 果 ( 圖 19,20) 都 具 有 高 準 確 度 , 而 針 對 θ 正 負 判 斷 必 須 經 過 判 斷 式<br />

F θ θ θ 做 修 正 , 利 用 修 正 之 結 果 如 下 :<br />

* * *<br />

(<br />

t<br />

,<br />

h, v)<br />

*<br />

v<br />

*<br />

v<br />

圖 21<br />

經 修 正 後 之 垂 直 軌 道 面 方 向 之 軌 道 幾 何 參 數 θ 比 較<br />

由 反 算 結 果 可 驗 證 , 此 反 算 法 在 變 速 度 之 情 況 下 亦 可 準 確 的 得 到 軌 道 之 幾 何 變 化 。<br />

與 過 去 常 見 做 法 之 比 較<br />

過 去 常 見 的 曲 率 取 得 方 式 , 乃 分 別 利 用 水 平 或 是 鉛 直 方 向 的 加 速 度 反 應 , 配 合 切 線 加 速 度 所 計<br />

算 出 ( 式 16,17)。 在 此 吾 人 可 比 較 本 反 算 法 則 的 計 算 與 基 本 曲 率 計 算 方 式 , 針 對 以 上 範 例 B 但 忽 略 突<br />

波 之 結 果 做 比 較 , 如 下 圖 所 示 。<br />

θ = B && / s& (16)<br />

* 2<br />

v h<br />

θ = B && / s& (17)<br />

* 2<br />

h v<br />

*<br />

v<br />

圖 22 水 平 軌 枕 方 向 之 幾 何 參 數<br />

*<br />

θ 反 算 比 較<br />

h<br />

圖 23 垂 直 軌 道 面 方 向 之 幾 何 參 數 θ 反 算 比 較<br />

由 於 使 用 之 加 速 度 方 程 式 中 考 慮 到 各 方 向 曲 率 及 曲 率 變 化 間 的 耦 合 現 象 , 可 發 現 其 間 計 算 結 果<br />

具 有 差 異 ; 使 用 基 本 曲 率 計 算 方 式 之 結 果 普 遍 振 動 幅 度 比 本 文 演 算 結 果 之 幅 度 來 的 大 , 可 見 一 般 之<br />

算 法 因 無 考 慮 曲 率 耦 合 效 應 之 結 果 , 尤 其 在 垂 直 軌 道 面 方 向 , 判 斷 之 幅 值 較 為 保 守 , 且 無 法 精 確 診<br />

斷 軌 跡 線 之 細 部 線 型 變 化 , 這 對 軌 道 完 整 性 之 判 斷 會 造 成 誤 差 。<br />

*<br />

v<br />

-103-


結 論<br />

本 文 提 出 一 套 由 移 動 中 輪 軸 加 速 度 信 號 去 反 算 運 動 軌 跡 線 沿 弧 長 之 曲 率 及 曲 率 變 化 量 的 方<br />

法 , 此 法 可 用 於 偵 測 軌 道 系 統 之 缺 陷 即 不 健 全 處 。 由 以 上 計 算 過 程 以 及 算 例 中 可 以 得 到 以 下 結 論 :<br />

(1) 基 於 廖 慶 隆 (2004) 所 推 導 之 軌 道 運 動 理 論 , 吾 人 可 以 利 用 適 當 的 測 量 點 位 置 假 設 配 合 數 學 上<br />

之 反 算 技 巧 , 可 由 參 數 互 相 耦 合 的 加 速 度 反 應 中 , 獲 得 軌 道 在 局 部 座 標 中 三 個 方 向 之 曲 率 歷 時 變<br />

化 。(2) 由 軌 道 運 動 理 論 推 到 出 之 三 個 方 向 之 加 速 度 反 應 , 在 運 算 中 可 發 現 其 中 之 相 依 性 , 導 致 在 反<br />

算 方 程 式 的 過 程 當 中 , 沿 切 線 方 向 之 加 速 度 反 應 可 以 不 需 使 用 , 並 且 可 推 導 出 一 判 斷 方 程 式 做 為 反<br />

算 過 程 當 中 修 正 結 果 之 用 。(3) 經 由 數 值 算 例 之 反 算 結 果 , 證 實 其 在 數 學 上 之 可 靠 性 並 擁 有 高 準 確<br />

度 , 在 高 頻 振 動 及 速 度 不 定 之 情 況 下 , 亦 可 獲 得 良 好 結 果 。(4) 在 反 算 過 程 中 , 鉛 直 方 向 參 數 則 會<br />

出 現 訊 號 恆 正 之 情 形 , 必 須 依 靠 判 斷 式 以 做 為 判 斷 。 經 過 判 斷 式 修 正 後 之 結 果 可 精 確 地 求 取 鉛 直 方<br />

向 參 數 , 達 到 準 確 的 反 算 結 果 。(5) 反 算 數 據 與 傳 統 計 算 方 式 做 比 較 之 結 果 , 可 見 傳 統 計 算 方 式 所 得<br />

到 之 曲 率 變 化 較 為 保 守 。<br />

參 考 文 獻<br />

Weston, P. F., Ling, C. S., Roberts, C., Goodman, C. J., Li, P. and Goodall, R. M. Monitoring vertical track irregularity from<br />

in-service railway vehicles, Proc. IMechE, 2007, 221:75-88.<br />

廖 慶 隆 . 軌 道 運 動 理 論 與 軌 道 不 整 之 應 用 . 中 國 土 木 水 利 工 程 會 刊 , 2004,31(2): 49-53.<br />

附 錄 A<br />

本 文 使 用 之 軌 道 列 車 在 空 間 中 之 運 動 理 論 由 廖 慶 隆 (2004) 提 出 , 經 由 軌 道 所 含 有 之 幾 何 曲 線 變<br />

化 , 推 得 對 應 之 軌 道 車 輛 車 軸 之 加 速 度 反 應 。 首 先 假 設 軌 道 曲 線 在 空 間 座 標 系 統 中 之 軌 道 列 車 與 軌<br />

uv<br />

道 接 觸 位 置 在 點 R()<br />

s , 其 中 s = s()<br />

t 為 時 間 點 t 之 接 觸 點 位 置 。 而 軌 道 車 輛 上 任 一 點 B 點 相 對 於 軌<br />

v<br />

道 曲 線 在 空 間 座 標 系 中 之 參 考 點 可 得 以 軌 道 面 之 動 座 標 系 統 為 基 準 之 位 置 向 量 rs ()。 其 下 標 u r 、<br />

t<br />

u r<br />

h<br />

及 u r 分 別 代 表 軌 道 面 之 局 部 動 座 標 系 統 中 沿 軌 道 切 線 方 向 、 垂 直 且 平 行 軌 枕 方 向 及 鉛 直 軌 道 面 方 向<br />

v<br />

向 量 , 其 中 鉛 直 方 向 向 量 u r 為<br />

v<br />

u r 及<br />

t<br />

u r 之 外 積 向 量 , 如 圖 A-1 所 示 :<br />

h<br />

圖 A-1 絕 對 空 間 座 標 與 軌 道 移 動 座 標 之 關 係 示 意 圖<br />

uv<br />

以 B()<br />

s 表 示 軌 道 車 輛 上 的 任 一 點 B 點 座 標 之 位 置 向 量 , 做 一 次 微 分 之 結 果 可 得 B 點 之 速 度 反 應 :<br />

uv uv v<br />

B() s = R() s + r()<br />

s<br />

(A1)<br />

&r &r r r &r r<br />

B = R + r&r<br />

= VRO<br />

+ ( VBR<br />

+ θ ×<br />

)<br />

(A2)<br />

其 中 V r<br />

為 R 點 相 對 於 整 體 座 標 系 之 線 速 度 , 即 為 列 車 切 線 方 向 速 度 。B 點 之 動 座 標 系 中 位 置<br />

RO<br />

-104-


向 量 r r &r r<br />

的 變 化 率 為 V BR<br />

+θ ×<br />

, 其 中 並 包 含 兩 部 分 : 其 一 為 V r<br />

BR<br />

, 其 乃 B 點 相 對 於 動 座 標 系 統 原 點<br />

R 之 相 對 速 度 , 假 設 吾 人 選 擇 之 B 點 與 軌 道 曲 線 參 考 點 R 點 之 間 無 相 對 變 位 , 兩 點 間 之 力 學 行 為 是<br />

完 全 剛 性 的 , 此 時 V r<br />

r<br />

BR<br />

為 零 。 其 二 為 & r θ × , 代 表 軌 道 面 運 動 帶 來 之 轉 角 速 度 &r θ 對 B 點 之 影 響 。 其 中<br />

&r θ 為 剛 體 轉 動 角 速 度 , 經 過 推 導 可 改 寫 為 r 。<br />

r r r θ′s &<br />

由 於 動 座 標 系 統 ( ut, uh,<br />

uv<br />

) 會 隨 軌 道 車 輛 之 運 動 一 直 在 做 轉 動 變 化 , 如 圖 A-2 所 示 , 也 就 是 說<br />

r r r<br />

向 量 ( u', t<br />

u' h, u'<br />

v ) 並 非 固 定 向 量 , 因 此 r 所 代 表 的 轉 角 量 對 距 離 s 作 微 分 之 結 果 , 可 視 為 軌 道 面 之<br />

r r r θ ′<br />

曲 率 向 量 , 可 用 基 底 向 量 ( u , u , u )<br />

* * *<br />

進 行 分 解 為 , 並 以 θ , θ , θ 做 為 動 座 標 系 統 之 轉 角 變 化 量 之 分<br />

量 。<br />

t h v<br />

t h v<br />

r<br />

u<br />

ht ( = t2)<br />

r<br />

u<br />

vt ( = t2)<br />

r<br />

r<br />

u<br />

ht ( = t3)<br />

u<br />

t( t = t2)<br />

r<br />

u<br />

vt ( = t3)<br />

r<br />

u<br />

t( t = t3)<br />

r<br />

u<br />

vt ( = t1)<br />

圖 A-2 移 動 座 標 系 隨 列 車 位 置 改 變 而 旋 轉 示 意 圖<br />

若 對 剛 體 而 言 ,B 點 及 R 點 相 對 位 置 為 定 值 , 即 r v 為 固 定 向 量 。 經 過 整 理 的 式 (2) 所 得 之 B 點 速<br />

度 反 應 再 經 由 一 次 微 分 , 可 得 到 B 點 之 加 速 度 反 應 :<br />

B<br />

& r v v v v v v<br />

' v 2 ' v " v ' ' v<br />

= && s( u + θ × r ) + s&<br />

θ × u + θ × r + θ × θ × r<br />

(A3)<br />

t<br />

[( ) ( ( )]<br />

t<br />

v "<br />

其 中 θ 為 轉 角 變 化 量 對 位 移 s 做 微 分 , 可 視 為 軌 道 面 曲 率 沿 弧 長 s 之 變 化 量 向 量 。 接 著 將 各 向<br />

量 以 分 量 形 式 帶 入 式 (3) 中 , 可 得 到 B 點 在 移 動 座 標 系 統 中 的 加 速 度 值 變 化 結 果 如 下 式 所 示 。<br />

* * 2 * * * * * * * *<br />

B && t<br />

= && s(1 + avθh − ahθv) + s & [( avθh ' − ahθv ') + ( ahθt −atθh) θh −( atθv −ahθt ) θv]<br />

(A4)<br />

* * 2 * * * * * * * * *<br />

B && h<br />

= && s( atθv − avθt ) + s & [ θv + ( avθh −ahθv) θv −( ahθt − atθh) θt + ( atθv ' −avθt<br />

')] (A5)<br />

* * 2 * * * * * * * * *<br />

B && v<br />

= && s( ahθt − atθh) + s & [ − θh + ( atθv −avθt ) θt −( avθh − ahθv) θh + ( ahθt ' −atθh<br />

')] (A6)<br />

以 上 方 程 式 經 過 轉 換 之 結 果 , 其 反 應 方 向 皆 在 軌 道 面 座 標 系 統 上 , 其 中 B & t<br />

、 B && h<br />

及 B && v<br />

分 別 代 表<br />

移 動 座 標 系 中 三 方 向 之 加 速 度 反 應 。s& 及 s&& 為 分 別 為 軌 道 列 車 之 瞬 時 速 度 及 加 速 度 。 a t<br />

、 a h<br />

及 a<br />

v<br />

分<br />

別 B 點 與 R 點 之 相 對 位 置 , θ *<br />

t<br />

、 θ * 、<br />

* *<br />

θ<br />

h v<br />

、 θ<br />

t<br />

' 、 θ<br />

* h<br />

' 及 θ<br />

* v<br />

' 分 別 代 表 各 方 向 上 之 瞬 時 扭 曲 率 及 其<br />

沿 弧 長 之 變 化 。<br />

由 上 式 可 以 發 現 , 影 響 軌 道 列 車 上 舒 適 度 的 實 際 加 速 度 反 應 不 指 單 是 由 軌 道 列 車 之 瞬 時 加 速 度<br />

s&& 所 控 制 , 並 且 與 軌 道 列 車 的 瞬 時 速 度 s& 、 軌 道 路 線 的 空 間 幾 何 線 型 變 化 以 及 實 際 軌 道 列 車 上 B 點<br />

位 置 有 關 。<br />

r<br />

u<br />

ht ( = t1)<br />

r<br />

u<br />

t( t = t1)<br />

-105-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

RECENT SEISMIC DESIGN AND RETROFIT STUDIES OF BRIDGES AT NCREE<br />

ABSTRACT<br />

K. C. Chang 1,2 , H. H. Hung 1 and K. Y. Liu 1<br />

1<br />

National Center for Research on Earthquake Engineering, Taipei, Taiwan<br />

2 Department of Civil Engineering, National Taiwan <strong>University</strong><br />

Email: kcchang@ncree.narl.org.tw<br />

Taiwan is located in a highly seismic active region of the world. <strong>The</strong>refore, the structures in Taiwan suffer<br />

frequently from earthquakes. Over the years, several earthquake events not only damaged the structures and<br />

caused life and property losses, but also impeded the economic development of the country. As such, the seismic<br />

design and retrofit of structures have become an essential issue for the sustainable development of Taiwan. As<br />

part of this, the National Center for Research on Earthquake Engineering (NCREE) was officially established in<br />

Taiwan in 1990. Thsis paper will introduce the researches at NCREE related to seismic design and retrofit of<br />

bridges in recent years. <strong>The</strong>se studies to be presented include the seismic retrofitting program, the rocking<br />

behavior of bridge column with spread footing, and an on-site experiment of three bridge columns performed in<br />

last year.<br />

KEYWORDS<br />

FRP, RC beams, strengthening, interfacial stresses, analytical solution.<br />

INTRODUCTION<br />

Taiwan is located along the western ridge of the Pacific Ocean earthquake belt and thus suffers frequently from<br />

earthquakes. Frequent earthquakes have inevitably caused significant damages to the existing bridges in Taiwan.<br />

<strong>The</strong> damages of the bridges due to earthquake not only lead to the loss of many innocent lives, but also interfere<br />

with the rescue efforts, generating countless amount of cost to the society and economy. <strong>The</strong>se circumstances<br />

lead special attention to be given to seismic design and retrofit in Taiwan, especially after the devastating<br />

Chi-Chi earthquake occurred in the central part of Taiwan on September 21, 1999. <strong>The</strong> National Center for<br />

Research on Earthquake Engineering (NCREE) in Taiwan, which was officially established in 1990, has<br />

continued to bring together academic resources and researchers to carry out joint projects to upgrade seismic<br />

design and retrofit technologies to reduce life and property losses resulting from earthquake in the recently two<br />

decades. This paper will introduce some latest researches at NCREE related to seismic design and retrofit of<br />

bridges. <strong>The</strong>se studies to be presented include a coordinated research program related to the seismic retrofitting<br />

of RC bridge columns, two series of experiments focused on the rocking behavior of bridge column with spread<br />

footing, and an on-site experiment of three bridge columns performed in last year.<br />

SEISMIC RETROFIT STUDY OF RC BRIDGE COLUMNS<br />

Background and Objectives<br />

Significant amount of retrofit research and actual implementations to enhance the seismic performance of<br />

existing bridges have been made in the United States, Japan and New Zealand after some major earthquakes<br />

occurred in these countries. In order to make sure that if the retrofit methods used in these countries were also<br />

effective for the existing RC bridge columns in Taiwan, a 4-years coordinated research effort on seismic retrofit<br />

of existing RC bridge columns were established at the NCREE from 1998 to 2002. Major objectives of this<br />

program were to develop effective seismic retrofit methods of existing bridges in Taiwan due to (1) inadequate<br />

design strength, (2) inadequate confinement at potential plastic hinge region, (3) inadequate shear strength due<br />

to large lateral steel spacing, and (4) lap-splicing in the plastic hinge zone. This coordinated research program<br />

includes a master plan administrated by NCREE and seven coordinated projects handled by the investigators<br />

from six universities and research institutions. <strong>The</strong> joint research effort has applied several retrofit techniques in<br />

the tests, including the steel jacketing, reinforced concrete jacketing, and the advanced composite material<br />

wrapping using the FRP. Results of this research program provided a domestic test database for seismic bridge<br />

engineering applications and seismic retrofit guidelines for highway officials in Taiwan.<br />

-106-


Experimental Program<br />

More than 60 large scale specimens were tested, including 24 benchmark specimens that were designed to<br />

represent typical pre- and after 1987 bridge columns in Taiwan. Cross sectional dimensions of the rectangular<br />

columns and circular columns were 60 cm by 75 cm and 76 cm diameter, respectively, roughly 2/5 scale of the<br />

prototype columns. Retrofit techniques used in the specimens include steel jacketing, FRP wrapping, and RC<br />

jacketing. Details of the test specimens are listed in Tables 1-3.<br />

Failure<br />

type<br />

Flexural<br />

Shear<br />

Lap-spli<br />

ces<br />

Table 1 Detail of benchmark specimens<br />

Specimen Cross<br />

section<br />

Height Axial<br />

load<br />

Longitudinal<br />

Reinforcement<br />

Transverse<br />

Reinforcement<br />

Arrange Cut off Arrangement<br />

-ment height PHZ Non-PHZ<br />

(mm) (mm) (f'c Ag) (mm) (mm) (mm) (mm)<br />

BMR1 750*600 3250 0.1 32-19Φ --- 9Φ@100 9Φ@100<br />

BMR2 750*600 3250 0.1 32-16Φ 1800 9Φ@130 9Φ@240<br />

BMR3 750*600 3250 0.15 32-16Φ 1800 9Φ@130 9Φ@240<br />

BMR4 750*600 3250 0.15 32-16Φ 1800 9Φ@230 9Φ@230<br />

BMR1-R 750*600 3250 0.15 34-19Φ 1800 9Φ@100 9Φ@100<br />

BMC1 D=760 3250 0.15 34-19Φ --- 9Φ@70 9Φ@100<br />

BMC2 D=760 3250 0.15 30-16Φ 1800 9Φ@130 9Φ@220<br />

BMC3 D=760 3250 0.15 30-16Φ 1800 9Φ@230 9Φ@230<br />

BMC4 D=760 3250 0.15 30-17Φ 1800 9Φ@130 9Φ@220<br />

SC1 D=760 3250 0.15 26-16Φ 1250 9Φ@140 9Φ@240<br />

SC1-R D=760 3250 0.15 26-16Φ 1250 9Φ@140 9Φ@240<br />

FC1 D=750 3250 0.15 32-16Φ --- 9Φ@100 9Φ@100<br />

FC4 D=750 3250 0.15 18-16Φ --- 9Φ@300 9Φ@300<br />

BMRS 750*600 1750 0.15 30-19Φ --- 9Φ@300 9Φ@300<br />

BMCS D=760 1750 0.15 30-19Φ --- 9Φ@300 9Φ@300<br />

BMRL100 750*600 3250 0.15 30-19Φ 760 9Φ@130 9Φ@220<br />

BMRL50 750*600 3250 0.15 30-19Φ 760 9Φ@130 9Φ@220<br />

BMCL100 D=760 3250 0.15 30-19Φ 760 9Φ@130 9Φ@220<br />

BMCL50 750*600 3250 0.15 30-19Φ 760 9Φ@130 9Φ@220<br />

Table 2 Retrofit methods of rectangular specimens<br />

Failure<br />

Type<br />

Flexural<br />

Shear<br />

Lap splices<br />

Retrofit/Repair Benchmark Note<br />

Spec men<br />

FR1 BMR2 FRP(8 layers)<br />

FR2 B R3 FRP(4 layers)<br />

SR1<br />

Large octagon<br />

SR2<br />

Ellipse<br />

SR3<br />

Small octagon<br />

S 4<br />

Ellipse<br />

FRS BMRS FRP(4 layers)<br />

SRS1<br />

Small octagon<br />

SRS2<br />

Ellipse<br />

BMRS-RC<br />

RC(9cm)<br />

FRL 100 BMRL 100 FRP(8 layers)<br />

SFRL 100<br />

Steel+FRP<br />

SRL1<br />

Small octagon<br />

SRL2<br />

Ellipse<br />

BMRL 100-RC<br />

RC(9cm)<br />

BMRL 50-RC BMRL 50 RC(9cm)<br />

Table 3 Retrofit methods of circular specimens<br />

Failure type<br />

Flexural<br />

Shear<br />

Lap splices<br />

Retrofit/repair Benchmark Note<br />

specimen<br />

SC2 SC1 Steel(3mm)<br />

SC3 BNC2 Steel(3mm)<br />

FC2 FC2 FRP(4 layers)<br />

FC3<br />

FRP(4 layers)<br />

RCC2 BMC2 RC(9 cm)<br />

BMC4-RC BMC4 RC(9 cm)<br />

SCS BMCS Steel(3mm)<br />

FCS<br />

FRP(4 Layers)<br />

FCS-1<br />

FRP(3 Layers)<br />

FCS-2<br />

FRP(2 Layers)<br />

FCS-3<br />

High pressure<br />

Epoxy injected<br />

SCL 100 BMCL 100 Steel(3mm)<br />

FCL 100<br />

FRP (6-2layers)<br />

FCL 100-1<br />

FRP (4-2layers)<br />

FCL 100-2<br />

FRP (6-2layers)<br />

FCL 100-3<br />

FRP (6-2layers)<br />

RCCL1<br />

RC(9cm)<br />

RCCL2<br />

RC(9cm)<br />

BMCL50-RC BMCL 50 RC(9cm)<br />

Test Results and Conclusions<br />

For seismic retrofit of rectangular RC bridge columns using steel jacketing and FRP wrapping, the hysteresis<br />

loop of the columns are shown in Fig. 1 (Tsai and Lin 2000) and Fig. 2 (Chang and Chung 2000; Chang and<br />

Chang 2000), respectively. Test results for Fig. 1 confirmed that the seismic performance of the rectangular RC<br />

bridge columns can be significantly and equally enhanced by properly constructed elliptical or octagonal steel<br />

jacket. Octagonal steel jacketing is cost-effectively and can provide lateral confinement and the shear strength to<br />

-107-


mitigate seismic failure of rectangular RC bridge columns due to a lack of lateral confinement, improper<br />

lap-splice or inadequate shear capacity. In addition, rectangular steel jacketing can effectively prevent a<br />

shear-deficient column from shear failure; however, it is not effective in improving the flexural ductility.<br />

500<br />

250<br />

BMR-3<br />

Vmax=290kN<br />

500<br />

250<br />

SR-1<br />

Vmax=456kN<br />

Force(kN)<br />

0<br />

-250<br />

Force(kN)<br />

0<br />

-250<br />

Force(kN)<br />

Force(kN)<br />

Force(kN)<br />

-500<br />

500<br />

250<br />

0<br />

-250<br />

-500<br />

800<br />

400<br />

0<br />

-400<br />

-800<br />

800<br />

400<br />

0<br />

-400<br />

-800<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(% radian)<br />

SR-2<br />

Vmax=400kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(% radian)<br />

BMRL100<br />

P=0.15f'cAg<br />

=1400kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(%)<br />

SRL2<br />

P=0.15f'cAg<br />

=1400kN<br />

Vmax=- 368kN<br />

Vmax=602kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(%)<br />

Force(kN)<br />

Force(kN)<br />

Force(kN)<br />

-500<br />

500<br />

250<br />

0<br />

-250<br />

-500<br />

800<br />

400<br />

0<br />

-400<br />

-800<br />

800<br />

400<br />

0<br />

-400<br />

-800<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(% radian)<br />

SR-3<br />

Vmax=418kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(% radian)<br />

SRL1<br />

P=0.15f'cAg<br />

=1400kN<br />

Vmax= - 622kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(%)<br />

BMRL100<br />

SRL1<br />

SRL2<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(%)<br />

Force(kN)<br />

Force(kN)<br />

Force(kN)<br />

500<br />

250<br />

0<br />

-250<br />

SR-4<br />

Vmax=423kN<br />

-500<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(% radian)<br />

1200<br />

BMRS<br />

600<br />

P=0.15f'cAg<br />

=1400kN<br />

0<br />

-600<br />

-1200<br />

1200<br />

600<br />

0<br />

-600<br />

-1200<br />

SRS2<br />

P=0.15f'cAg<br />

=1400kN<br />

Vmax=- 722kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(%)<br />

Vmax= 982kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(%)<br />

Force(kN)<br />

Force(kN)<br />

Force(kN)<br />

500<br />

250<br />

0<br />

-250<br />

-500<br />

1200<br />

600<br />

0<br />

-600<br />

-1200<br />

1200<br />

600<br />

0<br />

-600<br />

-1200<br />

P=0.15f'cAg<br />

=1400kN<br />

BMR-3<br />

SR-1<br />

SR-2<br />

SR-3<br />

SR-4<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(% radian)<br />

SRS1<br />

P=0.15f'cAg<br />

=1400kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(%)<br />

BMRS<br />

SRS1<br />

SRS2<br />

FRS<br />

Figure 1 Hysteresis loop of the retrofitted Rectangular RC columns using steel Jacket<br />

Vmax= 1086kN<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Drift Ratio(%)<br />

800<br />

400<br />

BMR2<br />

400<br />

BMR3<br />

600<br />

BMRS<br />

400<br />

BMRL100<br />

400<br />

200<br />

200<br />

200<br />

200<br />

0<br />

0<br />

0<br />

0<br />

-200<br />

-200<br />

-200<br />

-200<br />

-400<br />

-400<br />

-400<br />

-600<br />

-400<br />

-800<br />

-200 -100 0 100 200<br />

Displacement(mm)<br />

-200 -100 0 100 200<br />

Displacement(mm)<br />

1000<br />

-200 -100 0 100 200<br />

Displacement(mm)<br />

-200 -100 0 100 200<br />

Displacement(mm)<br />

400<br />

FR1<br />

400<br />

FR2<br />

750<br />

FRS<br />

400<br />

FRL100<br />

200<br />

200<br />

500<br />

200<br />

250<br />

0<br />

0<br />

0<br />

0<br />

-250<br />

-200<br />

-200<br />

-200<br />

-500<br />

-400<br />

-400<br />

-750<br />

-400<br />

-1000<br />

-200 -100 0 100 200<br />

Displacement(mm)<br />

-200 -100 0 100 200<br />

Displacement(mm)<br />

-200 -100 0 100 200<br />

Displacement(mm)<br />

-200 -100 0 100 200<br />

Displacement(mm)<br />

Figure 2 Hysteresis loop of the retrofitted Rectangular RC columns using FRP Jacket<br />

Test results for Fig. 2 show that failure of the flexural type specimen under larger axial load will result in<br />

speeding up the degradation of strength and energy dissipation capacity. In addition, standard hoop<br />

arrangements can gain better confinement than the double-U shaped alternation arrangement used in many<br />

existing bridges. For shear failure mode specimens, specimen retrofitted by wrapping FRP shows great<br />

performance in improving shear strength, and transfers the failure mode to flexural-shear type. For lap spliced<br />

-108-


failure mode specimens, applying CFRP directly can’t provide enough confinement stress to increase frictional<br />

force between the lap-spliced longitudinal reinforcements<br />

From the other test results for seismic retrofit of circular RC bridge column using steel jacketing and CFRP<br />

(Hwang and Hseih 1999; Hwang and Kuo 2000) concluded that the retrofit using steel jacketing is effective in<br />

enhancing the seismic resistance of existing circular RC bridge column. With steel jacketing and CFRP<br />

jacketing, the failure mode changes from flexural failure to the breaking of longitudinal bar in the bottom of the<br />

columns, and the ductility and maximum lateral force have increased. For lap splice failure mode specimens,<br />

using steel and CFRP jacketing can tremendously increase the confinement strength and ductility of bridges<br />

columns. Also, the more layers of CFRP can obtain higher ductility. On the other hand, for shear failure mode<br />

specimen, even though using steel and CFRP jacketing can also tremendously increase the confinement strength<br />

and ductility of bridge columns, more layers of CFRP may not have higher ductility.<br />

ROCKING EXPERIMENTS FOR BRIDGE PIERS WITH SPREAD FOOTINGS<br />

Background and Objectives<br />

After chi-chi earthquake, the design earthquake intensities of some area in Taiwan were shifted to a higher value;<br />

thus a great number of bridges need to be retrofitted. As the retrofitting work leads to a higher plastic moment<br />

capacity of the columns, the design force for the foundation needs to be increased based on capacity design. In<br />

order to satisfy the stability check of a spread footing under the application of this plastic moment transferred from<br />

the column base, some of the retrofitting works resulted in uneconomically large spread footings. Some newly<br />

designed engineering practices also met the similar situation. According to previous design code in Taiwan, the<br />

footing uplift involving separation of footing from subsoil is permitted only up to one-half of the foundation base<br />

area as the applied moment reaches the value of plastic moment capacity of the column. <strong>The</strong> reason for this<br />

provision is that rocking of spread footings is still not a favorable mechanism. However, recent researches have<br />

indicated that rocking itself may not be detrimental to seismic performance and in fact can act as a form of seismic<br />

isolation mechanism. In order to gain a better understanding of the problem of rocking and then to get more<br />

confidence to update the seismic design code and seismic evaluation guidelines, two series of rocking<br />

experiments were performed at NCREE.<br />

Experimental Program<br />

For the first series of experiments (Hung et al. 2008; 2010a), a total of three circular RC columns with spread<br />

footings were tested. Using pseudo-dynamic tests and a cyclic loading test, these columns were subjected to<br />

different levels of earthquake accelerations, including a near field ground motion. <strong>The</strong> focus of this experiment<br />

was to investigate the rocking behavior of both lightly transverse reinforced columns and retrofitted columns in<br />

order to clarify that if the widening and strengthening of the foundations to limit the rocking mechanism of spread<br />

footing is necessary for the retrofit work. <strong>The</strong>se three columns were named specimens A, B and C, respectively.<br />

<strong>The</strong>se columns measured 60 cm in diameter with a clear height of 3.4 m, were poorly confined and were<br />

lap-spliced above the top of the foundation. <strong>The</strong> columns were reinforced with 26-D19 longitudinal reinforcing<br />

bars, and were transversely reinforced with D10 perimeter hoops spaced 12.7 cm apart, corresponding to an<br />

insufficient volumetric confinement ratio of ρ s = 0.0039. Other material properties for these test columns were as<br />

follows: concrete compressive strength f c ’ = 278 kg/cm 2 ; yield strength of longitudinal reinforcements F y = 3840<br />

kg/cm 2 . In order to investigate the rocking behavior of a retrofitted column with a ductility capacity that meets<br />

the requirement specified by the design code, one of the test columns, specimen C, was wrapped with 6 mm thick<br />

A36 steel plate jacketing with a length of 150 cm. During the tests, one as-built test column (specimen B) and the<br />

retrofitted column (specimen C) were rested on a neoprene pad to allow the rocking to take place. Another as-built<br />

column (specimen A) was constrained to the strong floor during testing to represent a benchmark test with fixed<br />

base condition. <strong>The</strong> summary of the test sequence is shown in Table 4. In this table, TH1 and TH2 represent a<br />

code compatible medium earthquake and a code compatible design earthquake, respectively. TH3 and TH4 are<br />

the near-field ground motions recorded during Chi-Chi earthquake, but were scaled to have the same PGA of the<br />

code compatible ones TH1 and TH2, respectively.<br />

Fig. 3 illustrates the test setup. In the case where the rocking mode of the foundation was restrained (Fig. 3a), four<br />

tie-down rods were placed through the foundation and anchored into the strong floor of the laboratory. In the case<br />

where the rocking mechanism was considered (Fig. 3b), and the square footings were rested on a neoprene pad,<br />

simulating a spread footing foundation in a stiff soil. By comparing the experimental response of the retrofitted<br />

column with that of the as-built one, the interaction effect of the rocking on the ductility demand and the strength<br />

demand of the columns was identified. A critical side effect of increasing the displacement response at the deck<br />

-109-


level as a result of the rocking during an earthquake was also recognized. This effect was especially critical when<br />

the earthquake was induced by a near-fault ground motion.<br />

600<br />

180 cm<br />

neoprene pad<br />

5 cm<br />

(a) fixed base<br />

(b) rocking base<br />

Figure 3 Test setup for experiments of bridge piers with spread footings<br />

Table 4 Design details and experimental test schedule for the first series of rocking experiments<br />

Test Specimens Design details Base<br />

Tests<br />

condition<br />

A<br />

(lap-spliced specimen)<br />

Footing: 168cm×168cm<br />

26-D19 with stirrup: D10@ 12.7cm<br />

Fixed base cyclic loading test<br />

B<br />

(lap-spliced specimen)<br />

C<br />

(retrofitted specimen)<br />

Footing: 168cm×168cm<br />

26-D19 with stirrup: D10@ 12.7cm<br />

Footing: 168cm×168cm<br />

26-D19 with stirrup: D10@ 12.7cm<br />

6 mm thick A36 steel jacketing<br />

Rocking<br />

base<br />

Rocking<br />

base<br />

pseudo-dynamic test (TH3)<br />

pseudo-dynamic test (TH2)<br />

cyclic loading test<br />

pseudo-dynamic test (TH1)<br />

pseudo-dynamic test (TH2)<br />

pseudo-dynamic test (TH4)<br />

cyclic loading test<br />

Table 5 Design details and experimental test schedule for the second series of rocking experiments<br />

Test Specimens Design details Base condition Tests<br />

CD40FS-R<br />

Footing: 140cm×140cm<br />

Rocking base pseudo-dynamic test (TH1,TH2)<br />

18-D19 with stirrup: D13@ 9cm<br />

Fixed base<br />

cyclic loading test<br />

cyclic loading test<br />

CD30FS-R<br />

Footing: 140cm×140cm<br />

Rocking base pseudo-dynamic test (TH1,TH2)<br />

CD40FB-R<br />

12-D19 with stirrup: D13@ 9cm<br />

Footing: 170cm×170cm<br />

Rocking base<br />

cyclic loading test<br />

pseudo-dynamic test (TH1,TH2)<br />

CD30FB-R<br />

18-D19 with stirrup: D13@ 9cm<br />

Footing: 170cm×170cm<br />

Rocking base<br />

cyclic loading test<br />

pseudo-dynamic test (TH1,TH2)<br />

12-D19 with stirrup: D13@ 9cm<br />

cyclic loading test<br />

CB40FS-R<br />

Footing: 140cm×140cm<br />

Rocking base<br />

cyclic loading test<br />

18-D19 with stirrup: D13@ 18cm Fixed base cyclic loading test<br />

CD30FB-F<br />

Footing: 170cm×170cm<br />

12-D19 with stirrup: D13@ 9cm<br />

Fixed base<br />

cyclic loading test<br />

For the second series of experiments (Hung et al. 2010b; 2010c), a total of six circular RC columns were<br />

constructed and subjected to both quasi-static and pseudo dynamic loadings. <strong>The</strong> focus of this second experiment<br />

was to investigate the interaction relationship between the strength capacities of the column and the foundation<br />

as well as its effect on the rocking behavior. <strong>The</strong>refore, experimental variables included dimension of footings,<br />

strength and ductility capacity of columns, and level of the earthquake intensity applied. Results of each cyclic<br />

loading test under rocking base condition were also compared with the benchmark test with fixed base condition.<br />

<strong>The</strong> test setup was similar to that illustrated in Fig. 3. In this experiment, six reinforced concrete columns with<br />

two types of foundation size and three types of design details in column base were designed and constructed.<br />

<strong>The</strong>se circular RC columns were all 50 cm in diameter with a clear height of 2.5 m and a height of footing 0.5 m.<br />

<strong>The</strong>ir footing sizes were either B = 140 cm or B = 170 cm. In order to compare the rocking performance of<br />

specimens with different ratios of the moment capacity of the column to the capacity of the footing, these test<br />

-110-


columns were reinforced with three types of design details. One with 12-D19 main reinforcements was<br />

transversely reinforced with D13 perimeter hoops spaced 9 cm (volumetric confinement ratio ρ s = 0.012),<br />

corresponding to a case with sufficient transverse reinforcements. <strong>The</strong> other two with 18-D19 main<br />

reinforcements were transversely reinforced with D13 perimeter hoops spaced 9 cm and 18 cm, respectively. <strong>The</strong><br />

one with the transverse reinforcements spaced 18 cm was designed to represents a column with an insufficient<br />

volumetric confinement ratio (ρ s = 0.006) in order to investigate the ductility demand for a column allow to rock.<br />

<strong>The</strong> nominated material properties for these specimens are as follows: concrete compressive strength f c ’ =280<br />

kg/cm 2 ; yield strength of main reinforcements F y = 4200 kg/cm 2 ; yield strength of transverse reinforcements<br />

F yh =2800 kg/cm 2 . <strong>The</strong> test schedule is listed in Table 5, which also includes pseudo-dynamic loading test and<br />

quasi-static cyclic loading test.<br />

Test Results and Conclusions<br />

As mentioned previously, both series of rocking experiments include pseudo-dynamic test and cyclic loading<br />

test. However, only part of the cyclic loading test results will be presented in this paper due to the page limit of<br />

the paper. <strong>The</strong> test results of the first series of experiment are given Fig. 4. In this figure, (a) shows the lateral<br />

load versus the total lateral displacement curves and (b) shows the moment versus rotation curves at the column<br />

base. From these figures, it is evident that the hysteretic response for specimen B show a pattern of response<br />

behavior that is similar to that of specimen A, including exhibiting a sudden and significant loss of lateral<br />

resistance with low ductility under reversed cyclic deformation. However, the pinching effect in specimen B is<br />

not as serious as that in specimen A. If we further compare specimen A with specimen B at the same drift ratio<br />

of 5%, we can find that the rocking mechanism of specimen B resulted in an increase of lateral load resistance<br />

and a decrease of plastic deformation in the plastic hinge zone. This observation confirms the isolation effect of<br />

a rocking foundation. For the retrofitted specimen C, it demonstrates a nonlinear rocking behavior in Fig. (a)<br />

and that the seismic force that the pier sustained was limited to an almost constant value of 25 tonf. This means<br />

that specimen C was able to sustain large lateral displacement without significant strength degradation. <strong>The</strong><br />

plastic deformation of specimen C shown in the moment-rotation curve is also smaller than that of specimen B.<br />

30<br />

20<br />

Drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Specimen A<br />

(cyclic loading test)<br />

Drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

30<br />

Specimen B<br />

20 (cyclic loading test)<br />

30<br />

20<br />

Drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Specimen C<br />

(cyclic loading test)<br />

Lateral force (tonf)<br />

10<br />

0<br />

-10<br />

Pull<br />

Push<br />

Lateral force (tonf)<br />

10<br />

0<br />

-10<br />

Pull<br />

Push<br />

Lateral force (tonf)<br />

10<br />

0<br />

-10<br />

Pull<br />

Push<br />

-20<br />

-20<br />

-20<br />

Moment (tonf-m)<br />

-30<br />

-30 -20 -10 0 10 20 30<br />

Lateral displacement (cm)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

-20<br />

-40<br />

-60<br />

-80<br />

Specimen A<br />

(cyclic loading test)<br />

Pull<br />

Push<br />

-100<br />

-0.08 -0.04 0 0.04 0.08<br />

Rotation (radians)<br />

Moment (tonf-m)<br />

-30<br />

-30 -20 -10 0 10 20 30<br />

Lateraldisplacement (cm)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

-20<br />

-40<br />

-60<br />

-80<br />

(a) Lateral force-displacement curves<br />

Specimen B<br />

(cyclic loading test)<br />

Pull<br />

Push<br />

-100<br />

-0.08 -0.04 0 0.04 0.08<br />

Rotation (radians)<br />

Moment (tonf-m)<br />

-30<br />

-30 -20 -10 0 10 20 30<br />

Lateral displacement (cm)<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

-20<br />

-40<br />

-60<br />

-80<br />

Specimen C<br />

(cyclic loading test)<br />

Pull<br />

Push<br />

-100<br />

-0.08 -0.04 0 0.04 0.08<br />

Rotation (radians)<br />

(b) Moment-rotation curves<br />

Figure 4 Experimental results for the cyclic loading test of specimens A, B and C<br />

In theory, if the foundation of a pier is allowed to rock with the uplift, the foundation lifts off the ground once its<br />

moment of resistance provided by gravity is overcome. Thus, the base moment can be limited to the value<br />

required to produce uplift against the restraining forces due to gravity. <strong>The</strong> base moment limitation will then<br />

possibly reduce the inelastic deformation of the pier at the plastic hinge zone. <strong>The</strong>se experiments showed that<br />

there was a decrease of plastic deformation at the plastic hinge of a column as a result of the energy dissipation<br />

of the inelastic rocking mechanism of the footing. However, this effect is not very significant for the<br />

un-retrofitted case of specimen B. This is because before the base moment of the foundation could reach its limit<br />

value, the column yielded and the strength degraded earlier due to its inadequate lap-splicing of the main<br />

reinforcements and poor transverse confinement. Because the column yielded before the isolation effect of<br />

-111-


ocking can be fully developed, the moment capacity of the column governed the overall response behavior and<br />

the difference observed between specimen A with a fixed base and specimen B with a rocking base became<br />

insignificant. On the other hand, for the retrofitted case of specimen C, because the purpose of its retrofit was to<br />

enhance its ductility but not its moment strength, the increased moment strength of the column was not great<br />

enough to let the overturning moment at the foundation base completely govern the response. Hence, the<br />

hysteresis curves in Fig. 4 show that some plastic deformation still occurred in the column. From this<br />

observation, it is recognized that even thought there is a good potential for a spread footing foundation to<br />

dissipate a large amount of energy through the rocking mechanism, thereby reducing the ductility demand on the<br />

column, the column still has to possess a certain level of strength and ductility in order to allow the rocking<br />

mechanism to become effective.<br />

For the second series of experiment, the test results for cyclic loading of CD40xx-x (specimens with 18-D19 main<br />

reinforcements) and CD30xx-x (specimens with 12-D19 main reinforcements) are given Fig. 5 and 6,<br />

respectively. In the above figures, (a) show the lateral force-displacement curves and (b) show the<br />

moment-rotation curves. Among these specimens, CD40FS-F and CD30FS-F are the benchmark test,<br />

representing the fixed base case or case with a footing of very large size. they can also signify the capacity of<br />

columns with 18 main steels and 12 main steel, respectively. By comparing the results of CD40FS-F, CD40FB-R<br />

and CD40FS-R, it is noted that the rocking behavior becomes more pronounced as the dimension of footing<br />

decreases. Also, the maximum value of lateral forces that the column sustained decreases with the decrease in<br />

footing size. Besides, the maximum value of the bending moment that the column sustained decreases with the<br />

decrease in footing size, too. Because the maximum values of moment sustained by these two rocking case are less<br />

than the moment capacity of columns indicated by CD40FS-F, the moment-rotation curves for both cases are<br />

almost linear, implying that not much plastic deformation occurred in columns.<br />

drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Lateral force (kN)<br />

200<br />

100<br />

0<br />

-100<br />

CD40FS-F<br />

Lateral force (kN)<br />

200<br />

100<br />

0<br />

-100<br />

CD40FB-R<br />

Lateral force (kN)<br />

200<br />

100<br />

0<br />

-100<br />

CD40FS-R<br />

-200<br />

-200<br />

-200<br />

Moment (kN-m)<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

-100<br />

-200<br />

-300<br />

-400<br />

-500<br />

-600<br />

-200 -100 0 100 200<br />

lateral displacement (mm)<br />

CD40FS-F<br />

-0.08-0.06-0.04-0.02 0 0.02 0.04 0.06 0.08<br />

Rotation (radian)<br />

Moment (kN-m)<br />

-200 -100 0 100 200<br />

lateral displacement (mm)<br />

(a) Lateral force-displacement curves<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

-100<br />

-200<br />

-300<br />

-400<br />

-500<br />

-600<br />

CD40FB-R<br />

-0.08-0.06-0.04-0.02 0 0.02 0.04 0.06 0.08<br />

Rotation (radian)<br />

Moment (kN-m)<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

-100<br />

-200<br />

-300<br />

-400<br />

-500<br />

-600<br />

-200 -100 0 100 200<br />

lateral displacement (mm)<br />

CD40FS-R<br />

-0.08-0.06-0.04-0.02 0 0.02 0.04 0.06 0.08<br />

Rotation (radian)<br />

(b) Moment-rotation curves<br />

Figure 5 Experimental results for the cyclic loading test of specimens CD40xx-x<br />

Next, by comparing the results of CD30FS-F, CD30FB-R and CD30FS-R, it is also observed that with the<br />

decrease in footing size, the nonlinear rocking behavior becomes more pronounced. And consequently the<br />

plastic deformation occurred in column base became minor. For instance, the moment-rotation curve for the case<br />

with a smaller footing (CD30FS-R) is almost linear, while some plastic deformation was formed in the case with<br />

a larger footing (CD30FB-R). For CD30FB-R, the upper limit value of moment corresponding to the base<br />

moment limitation sustained by the foundation was higher than the moment capacity of column indicated by<br />

CD30FB-F. <strong>The</strong>refore, before the base moment of the footing could reach its limit value, the column already<br />

yielded. On the other hand, for specimen CD30FS-R, the upper limit value of the bending moment sustained by<br />

column was lower than the moment capacity indicated in CD30FB-F. Thus, the column can still remain in<br />

elastic state. <strong>The</strong>se experiments showed that if the footing of the column is allowed to rock, the moment that the<br />

column has to sustain can be limited to a certain value. This upper limit value can be calculated based on a<br />

simple equation only related to the footing size and the total vertical force of gravity (Hung et al. 2010a). If this<br />

limit value for moment is lower than bending moment strength of the column, the plastic deformation will not<br />

-112-


e formed at the column base and the ductility demand of the column can be reduced. In addition, results also<br />

shows that if the footing uplift took place, there was a decrease in plastic deformation at the plastic hinge of a<br />

column as a result of the energy dissipation of the inelastic rocking mechanism. <strong>The</strong> extent of decrease in plastic<br />

deformation depends on the ratio of the moment capacity of column to the limit value of moment that<br />

corresponds to the base moment limitation sustained by the foundation.<br />

200<br />

drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

200<br />

drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

200<br />

drift (%)<br />

-8 -6 -4 -2 0 2 4 6 8<br />

Lateral force (kN)<br />

100<br />

0<br />

-100<br />

CD30FB-F<br />

Lateral force (kN)<br />

100<br />

0<br />

-100<br />

CD30FB-R<br />

Lateral force (kN)<br />

100<br />

0<br />

-100<br />

CD30FS-R<br />

Moment (kN-m)<br />

-200<br />

-200 -100 0 100 200<br />

lateral displacement (mm)<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

-100<br />

-200<br />

-300<br />

-400<br />

-500<br />

CD30FB-F<br />

-0.08-0.06-0.04-0.02 0 0.02 0.04 0.06 0.08<br />

Rotation (radian)<br />

Moment (kN-m)<br />

-200<br />

-200 -100 0 100 200<br />

lateral displacement (mm)<br />

-200<br />

(a) Lateral force-displacement curves<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

-100<br />

-200<br />

-300<br />

-400<br />

-500<br />

CD30FB-R<br />

-0.08-0.06-0.04-0.02 0 0.02 0.04 0.06 0.08<br />

Rotation (radian)<br />

Moment (kN-m)<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

-100<br />

-200<br />

-300<br />

-400<br />

-500<br />

-200 -100 0 100 200<br />

lateral displacement (mm)<br />

CD30FS-R<br />

-0.08-0.06-0.04-0.02 0 0.02 0.04 0.06 0.08<br />

Rotation (radian)<br />

(b) Moment-rotation curves<br />

Figure 6 Experimental results for the cyclic loading test of specimens CD30xx-x<br />

ON-SITE EXPIRIMENT AT NIUDOU BRIDGE<br />

Background and Objectives<br />

Earthquakes and floods are the most dangerous threats to bridges in Taiwan. Earthquakes exert force on bridges<br />

and floods undermine their foundations; both can severely damage bridges and greatly shorten bridges' lives.<br />

Due to laboratory space restrictions, past engineering research almost never conducted experiments<br />

simultaneously involving bridge structures and bridge foundations, and compared the results with modern<br />

design theory. Since the Old Niudou Bridge was scheduled for demolition, the old bridge provided a very<br />

suitable experimental subject for investigation of current design theory. <strong>The</strong>refore, an on-site experiment at<br />

Niudou bridge was performed last year by NCREE. <strong>The</strong> research team felt that the experiment would certainly<br />

attract the interest of persons worldwide engaged in bridge research, and expected that the experiment could<br />

provide valuable experimental data and models concerning bridges' interaction with soil structure, as well as the<br />

mechanisms of earthquake damage and foundation erosion, enabling a better understanding of bridge earthquake<br />

resistance.<br />

LVDT<br />

Tilt meter<br />

Figure 7 Locations of jacks; use of a wall-type bridge pier as a reaction wall<br />

-113-


Experimental Program<br />

This experiment includes both structural experiment and geotechnical experiment. This paper will focus on the<br />

structural experiment that related to the cyclic loading test on the bridge column P2 to P4, and pseudo-dynamic<br />

test performed on the bridge columns P2, as shown in Fig. 8. P3 column represents the benchmark pier standing<br />

on a caisson foundation, while P4 column is similar but dealing with scouring issue on the caisson foundation<br />

with 4m exposed length from the bottom of the column base. <strong>The</strong> bridge column is 180cm in diameter and<br />

1030cm in height, reinforced with 30-D32 longitudinal reinforcing bars and were transversely reinforced with<br />

D16 perimeter hoops spaced 20 cm apart, shown in Fig.9. Similar to P3 and P4, P2 column is expected to<br />

observe the cyclic behaviour but designated to conduct pseudo-dynamic test in a pre-determined ground<br />

acceleration first, and then pushed in a single cycle to compare the hysteresis curve obtained in P3 and P4. This<br />

input ground motion for P2 column is a code-compatible artificial acceleration to simulate an earthquake event<br />

of 475 year return period, and the peak ground motion given is 0.32g, based on the Specification of Seismic<br />

Design for Bridge Structure(2009 version). Considering the biggest seismic hazard to the Niudou bridge in<br />

decades, records at strong motion station ILA025 in March 31, 2002 is selected. <strong>The</strong> peak ground motion for<br />

N-S and E-W directions are 118.62gal and 90.8 gal, respectively. Fortunately, this station is the closest station<br />

can be found. As for the loading system, this experiment use wall-type piers and two supplement A-shape<br />

reaction frames as a reaction wall, and employ two oil jacks (Fig. 7) to apply lateral force to the circular piers to<br />

the target displacement as shown in Fig. 10 and table 5. This will enable seismic resistance performance curves<br />

to be obtained for bridge piers with circular cross-sections.<br />

(a) P2 column (b) P3 column (c) P4 column (d) test setup<br />

Figure 8 Niudou bridge column specimen and test setup<br />

(c) longitudinal reinforcement<br />

(d) lateral reinforcement<br />

(a) column<br />

(b) caisson foundation<br />

Figure 9 Dimension and arrangement of reinforcement of the column and caisson<br />

60<br />

Displacement(cm)<br />

40<br />

20<br />

0<br />

-20<br />

-40<br />

-60<br />

0 2 4 6 8 10 12 14 16<br />

Cycle No.<br />

Figure 10 Displacement-control<br />

loading protocol<br />

Drift<br />

ratio<br />

(%)<br />

Target<br />

Disp.<br />

(cm)<br />

Table 5 Target drift ratio and displacement for cyclic loading test<br />

0.25 0.375 0.5 0.75 1.0 1.5 2.0 3.0 5.0<br />

2.575 3.86 5.15 7.73 10.3 15.45 20.6 30.9 51.5<br />

-114-


Acceleration (g)<br />

1<br />

0.5<br />

0<br />

-0.5<br />

Simulated<br />

Initial<br />

Sa (g)<br />

4<br />

3<br />

2<br />

1<br />

Response<br />

Initial<br />

Code<br />

-1<br />

0 10 20 30 40 50 60 70 80 90<br />

Time (sec)<br />

Figure 11 Seismic wave and response spectrum of the artificial ground acceleration for pseudo-dynamic test<br />

Test Results and Conclusions<br />

0<br />

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5<br />

Period (sec)<br />

Figures 12 to 15 present the failure modes and hysteresis curves of column P3, P4, and P2, respectively. Though<br />

given different boundary conditions and different detected material strengths, three columns demonstrate similar<br />

flexural failure mode, especially the backbone curve can be simplified as a bi-linear curve. <strong>The</strong> maximum lateral<br />

force is around 1000 kN, and the displacement ductility can reach 5.0. Besides, for each column, it is due to<br />

large spacing of 20cm of lateral reinforcement, the longitudinal reinforcements were buckled and large cover<br />

concrete blocks were spall off during the tests. Moreover, the lap splice length of lateral reinforcement is<br />

insufficient and lacks of seismic hook, leading to separation of lateral reinforcement observed at some particular<br />

heights in the plastic hinge zone. However, the axial load obtained from the design drawing and calibrated by<br />

the system identification is 291tonf, a relative small value about 0.04-0.05 f’cAg, as a result to explain why the<br />

bridge column performed ductile behaviour, compared to the conventional axial load of 0.15f’cAg on the<br />

reduced-size columns in the lab. In addition, the test results are not sensitive to the scouring of caisson<br />

foundation, since the geological condition is gavel. Figures 12 to 14 also show the analytical results, both from<br />

pushover analyses and cyclic loading analyses, by using the program SERCB for Bridge, developed by NCREE<br />

in 2010. This program has been used as a standard tool in current seismic evaluation and retrofitting program in<br />

provincial bridges in Taiwan since 2009. <strong>The</strong> analytical results are agreed with the experimental result very well.<br />

Based on the results, it is suggested to utilize Mander model for confined concrete model and TAKEDA<br />

hysteresis model with an adequate plastic hinge length proposed by Priestley, so that to simulate a<br />

flexural-failure dominated column if strength degradation and stiffness reduction are not obvious.<br />

Drift ratio (%)<br />

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6<br />

Drift ratio (%)<br />

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6<br />

1250<br />

125<br />

1500<br />

150<br />

1000<br />

P3 COLUMN<br />

100<br />

1200<br />

P3 COLUMN<br />

Experimental result<br />

120<br />

750<br />

75<br />

900<br />

Pushover curve<br />

90<br />

500<br />

50<br />

600<br />

Hyeteresis loop<br />

60<br />

Lateral Force (kN)<br />

250<br />

0<br />

-250<br />

PULL<br />

PUSH<br />

25<br />

0<br />

-25<br />

Laterl force (tonf)<br />

Lateral Force (kN)<br />

300<br />

0<br />

-300<br />

PULL<br />

PUSH<br />

30<br />

0<br />

-30<br />

Laterl force (tonf)<br />

-500<br />

-50<br />

-600<br />

-60<br />

-750<br />

-75<br />

-900<br />

-90<br />

-1000<br />

-100<br />

-1200<br />

-120<br />

-1250<br />

-125<br />

-1500<br />

-150<br />

-600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600<br />

Lateral Displacement (mm)<br />

-600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600<br />

Lateral Displacement (mm)<br />

(a) failure mode (b) hysteresis loop (c) analytical results<br />

Figure 12 Experimental and analysis results of P3 column<br />

Drift ratio (%)<br />

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6<br />

Drift ratio (%)<br />

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6<br />

1250<br />

125<br />

1500<br />

150<br />

1000<br />

P4 COLUMN<br />

100<br />

1200<br />

P4 COLUMN<br />

Experimental result<br />

120<br />

750<br />

75<br />

900<br />

Pushover curve<br />

90<br />

500<br />

50<br />

600<br />

Hyeteresis loop<br />

60<br />

Lateral Force (kN)<br />

250<br />

0<br />

-250<br />

PULL<br />

PUSH<br />

25<br />

0<br />

-25<br />

Laterl force (tonf)<br />

Lateral Force (kN)<br />

300<br />

0<br />

-300<br />

PULL<br />

PUSH<br />

30<br />

0<br />

-30<br />

Laterl force (tonf)<br />

-500<br />

-50<br />

-600<br />

-60<br />

-750<br />

-75<br />

-900<br />

-90<br />

-1000<br />

-100<br />

-1200<br />

-120<br />

-1250<br />

-125<br />

-1500<br />

-150<br />

-600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600<br />

Lateral Displacement (mm)<br />

-600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600<br />

Lateral Displacement (mm)<br />

(a) failure mode (b)hysteresis loop (c) analytical results<br />

Figure 13 Experimental and analysis results of P4 column<br />

-115-


Drift ratio (%)<br />

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6<br />

Drift ratio (%)<br />

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6<br />

1250<br />

125<br />

1500<br />

150<br />

1000<br />

P2 COLUMN<br />

100<br />

1200<br />

P2 COLUMN<br />

Experimental result<br />

120<br />

750<br />

75<br />

900<br />

Pushover curve<br />

90<br />

500<br />

50<br />

600<br />

60<br />

Lateral Force (kN)<br />

250<br />

0<br />

-250<br />

PULL<br />

PUSH<br />

25<br />

0<br />

-25<br />

Laterl force (tonf)<br />

Lateral Force (kN)<br />

300<br />

0<br />

-300<br />

PULL<br />

PUSH<br />

30<br />

0<br />

-30<br />

Laterl force (tonf)<br />

-500<br />

-50<br />

-600<br />

-60<br />

-750<br />

-75<br />

-900<br />

-90<br />

-1000<br />

-100<br />

-1200<br />

-120<br />

-1250<br />

-125<br />

-1500<br />

-150<br />

-600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600<br />

Lateral Displacement (mm)<br />

-600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600<br />

Lateral Displacement (mm)<br />

(a) failure mode (b) hysteresis loop (c) analytical results<br />

Figure 14 Experimental and analysis results of P2 column (cyclic loading test)<br />

Drift ratio (%)<br />

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6<br />

1250<br />

125<br />

1000<br />

P2 COLUMN<br />

100<br />

750<br />

75<br />

500<br />

50<br />

Lateral Force (kN)<br />

250<br />

0<br />

-250<br />

PULL<br />

PUSH<br />

25<br />

0<br />

-25<br />

Laterl force (tonf)<br />

-500<br />

-50<br />

-750<br />

-75<br />

-1000<br />

-100<br />

-1250<br />

-125<br />

-600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600<br />

Lateral Displacement (mm)<br />

(a) failure mode<br />

(b) hysteresis loop<br />

Figure 15 Experimental results of P2 column (pseudo-dynamic test)<br />

CONCLUSIONS<br />

This paper has briefly introduced three latest experimental programs performed at NCREE related to seismic<br />

design and retrofit of bridges. Some major conclusions drawn from these experiments were also presented.<br />

<strong>The</strong>se research results can be helpful for government to take proper steps to decrease seismic risks for bridges<br />

and to serve as a reference for future update of the design code.<br />

ACKNOWLEDGMENTS<br />

<strong>The</strong> authors gratefully acknowledge the financial support provided by the National Science Council through<br />

National Center for Research on Earthquake Engineering and National Applied Research Laboratories of<br />

Taiwan.<br />

REFERENCES<br />

Chang, K.C. and Chung F.S. (2000). Seismic shear and lap-spliced retrofit of rectangular bridge Columns using<br />

FRP, National Center for Research on Earthquake Engineering, Technical Report.<br />

Chang, K.C. and Chang H.F. (2000). Seismic flexural retrofit of rectangular bridge columns using FRP, National<br />

Center for Research on Earthquake Engineering, Technical Report.<br />

Tsai, K.C. and Lin, M.L. (2000), Steel jacketing for seismic retrofit of RC rectangular columns, National Center<br />

for Research on Earthquake Engineering, Technical Report.<br />

Hwang, J.S. and Hseih, Y.M. (1999) Seismic retrofit of RC bridge columns using steel jacket, National Center<br />

for Research on Earthquake Engineering, Technical Report.<br />

Hwang, J.S. and Kuo, M.Y. (2000) Seismic retrofit of existing RC bridge columns – shear strength and lap<br />

splice retrofit, National Center for Research on Earthquake Engineering, Technical Report.<br />

Hung, H.H, Chang, K.C, Liu, K.Y and Ho, T.H. (2008), A study on the rocking response of spread foundation.<br />

Report Number: NCREE-08-040. (in Chinese)<br />

Hung, H.H., Liu, K.Y. Ho, T.H. and Chang, K.C. (2010) “An Experimental Study on the Rocking Response of<br />

Bridge Piers with Spread Footing Foundations” Earthquake Engineering and Structural Dynamics,<br />

published online in Wiley Online Library. DOI: 10.1002/eqe.1057.<br />

Hung, H. H., Chang, K. C., Liu, K. Y., Wang, S. C. (2010), “Quasi-static and Pseudo-dynamic Testing of<br />

Rocking Spread Footings for Bridges”, 9th US National and 10th Canadian Conference on Earthquake<br />

Engineering, Toronto, Canada, July 25-29.<br />

Hung, H.H, Guntorojati, I., Liu, K.Y and Chang, K.C (2010), Seismic performance of bridge with rocking<br />

spread footing. Report Number: NCREE-10-025.<br />

-116-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

PSEUDO-DUCTILE PERMANENT FORMWORK FOR THE CONSTRUCTION OF<br />

DURABLE CONCRETE STRUCTURES<br />

Christopher K.Y. Leung and Changli Yu<br />

Department of Civil and Environmental Engineering,<br />

<strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> of Science and Technology, <strong>Hong</strong> <strong>Kong</strong>, China<br />

ABSTRACT<br />

To enhance the durability of reinforced concrete structures, high performance concrete with low water/binder<br />

ratio is often employed. However, once cracking occurs in the concrete cover due to mechanical loading or<br />

shrinkage, water and chloride can penetrate easily to induce steel corrosion. <strong>The</strong> present study focuses on an<br />

alternative approach for the construction of durable concrete members, with the use of permanent formwork.<br />

To make the formwork, pseudo-ductile cementitious composites (PDCC) is employed. With 2% of<br />

incorporated fibers, the PDCC exhibits multiple cracking behavior in tension, with crack openings controlled to<br />

below 60 micron at strain level up to several percent. With such small crack openings, the transport properties<br />

are similar to those of the un-cracked material. In addition to PDCC, glass fiber reinforced polymer (GFRP)<br />

rods can be incorporated into the formwork to provide flexural resistance. In some applications, the member<br />

can be made by casting concrete directly on the formwork. In others, a reduced amount of steel reinforcement<br />

can be added. With steel protected by the GFRP/PDCC formwork (which acts as part of the cover), high<br />

durability can be ensured. In this paper, we will describe the material employed for making the permanent<br />

formwork and the method to fabricate U-shape formwork for slabs and beams. Test results on components<br />

made with the permanent formwork will be presented. Failure behavior will be discussed and failure load<br />

compared to analytical values. An example will also be provided to illustrate the application of GFRP/PDCC<br />

formwork in practice.<br />

KEYWORDS<br />

Cementitious composites, permanent formwork, durability, construction technology.<br />

INTRODUCTION<br />

PDCC, also referred to as Engineered Cementitious Composites (ECC), are fiber reinforced composites<br />

designed based on fracture mechanics concepts and with the help of micromechanical models (Li & Leung 1992,<br />

Li 1993, Leung 1996, Kanda &Li 1999). Through the proper ‘engineering’ of micro-parameters, the composite<br />

exhibits strain hardening behavior in tension with failure strain up to several percents. Tensile failure of the<br />

composite is accompanied by the formation of closely-spaced multiple cracks that are controlled to very small<br />

openings (normally below 60 microns). According to the experimental results in Wang et al. (1997) and Lepech<br />

& Li (2005), cracks with such small openings have little effect on the transport properties of the material. In<br />

other words, for PDCC made with a matrix with sufficiently low water/binder ratio, high durability can be<br />

maintained even if the member surface is subjected to high tensile strain during service.<br />

As the durability of a concrete member is governed by the quality of the cover, PDCC (which is more costly<br />

than normal concrete) can be used strategically to prepare a permanent formwork first. <strong>The</strong> member is then<br />

constructed by the casting of normal concrete. With a surface layer exhibiting low permeability and high<br />

cracking resistance, penetration of corrosive agents can be resisted. Moreover, by replacing wooden formwork,<br />

construction efficiency is improved and site wastes are reduced.<br />

To further facilitate construction and improve member durability, the concept can be extended to include glass<br />

fiber reinforced plastics (GFRP) rods inside the formwork. <strong>The</strong> use of GFRP rods with PDCC creates a synergy<br />

between the two materials. If normal concrete is employed, the low modulus of GFRP, low bond strength and<br />

mismatch in thermal expansion coefficient between concrete and GFRP in the transverse direction will require a<br />

minimum cover of 35 mm to prevent the formation of unsightly cracks. <strong>The</strong> formwork will then be relatively<br />

thick and heavy. With the use of PDCC, a small cover to the GFRP is sufficient as crack openings are well<br />

controlled. Also, as shown in Fischer & Li (2002), due to the multiple cracking of PDCC which produces fine<br />

-117-


and closely spaced cracks, the interfacial shear deformation and stresses induced by cracking is significantly<br />

reduced. <strong>The</strong> required bond strength is hence also a lot lower. For components under light or moderate loading,<br />

the GFRP rods inside the PDCC formwork will be sufficient to carry the required loading. When the applied<br />

loading is higher, steel reinforcements can be added to increase the load-carrying capacity. In this case, the steel<br />

is well protected from corrosion as it is far away from the member surface. As a result, high durability of the<br />

member can be assured.<br />

When permanent formwork is employed, the interfacial bonding between the cast concrete and the formwork is<br />

always a concern. In an earlier paper (Leung & Cao 2010), we have focused on the development of flat plate<br />

PDCC formwork, to be placed at the bottom of slabs/decks. When GFRP is not incorporated into the formwork,<br />

the introduction of transverse grooves on the inner surface is found to be effective in preventing debonding.<br />

However, for GFRP/PDCC formwork, interfacial debonding is found to be the dominant failure mode of the<br />

final component. To improve the interface bonding, a feasible solution is to introduce a U-shape Permanent<br />

formwork to increase the total area of the interface. Furthermore, as the formwork will have a higher moment of<br />

inertia with the contribution from the bent-up legs, damage during construction handling and transportation can<br />

be minimized.<br />

In this paper, the effectiveness of U-shaped formwork relative to flat plate formwork will be studied. In the<br />

following, we will first present the material design of PDCC and the preparation of U-shaped formwork. <strong>The</strong>n,<br />

bending test results on beams made with GFRP/PDCC formwork (and with or without additional steel rebars)<br />

will be reported, with special attention paid on the load capacity and failure mode. Finally, a design example<br />

will be given to illustrate the feasibility of GFRP/PDCC formwork in the construction of lateral spanning deck<br />

for footbridges.<br />

SPECIMEN PREPARATION<br />

Materials<br />

From the literature, PDCC can be made with polyvinyl alcohol (PVA) fibers at a dosage of 2% in volume. <strong>The</strong><br />

properties of PVA fiber is shown in Table 1. To ensure uniform fiber distribution and to control the toughness of<br />

the matrix, fine silica sand is used in the matrix and no coarse aggregates are incorporated. In this study, 80%<br />

by weight of the cement was replaced by fly ash. As fly ash is a waste material, the use of a large amount of fly<br />

ash in the PDCC can be considered a ‘green’ approach (Yang et al. 2007). Some of the fly ash will undergo<br />

pozzolanic reaction to improve the long-term transport properties. However, a significant part of the fly ash will<br />

not hydrate and can be considered as inert fillers. According to the results in Song & Van Zijl (2004), increased<br />

deformability and toughness can be obtained with fly ash addition beyond 40%. Such a trend can be ascribed to<br />

the spherical shape of the unhydrated fly ash particles, which can reduce friction along the matrix-fiber interface<br />

and facilitate fiber pull-out (rather than rupture). High fly ash content will decrease the compressive strength of<br />

PDCC and increase its porosity. In our mix, silica fume is also added, as mixes with silica fume are found to<br />

exhibit less ductility reduction in the long term. <strong>The</strong> mix portion chosen for our test is shown in Table 1. With<br />

high content of unreacted fly ash, we expect a large number of distributed pores in the matrix which can<br />

facilitate the formation of multiple cracks and thus improve the ductility (Wang & Li 2007). Permeability<br />

measurement has been performed on our PDCC mix and a value of 5x10 -12 m/s was obtained. In Lepech & Li<br />

(2005), a value of 1x10 -11 m/s was obtained for a PDCC mix with a much lower fly ash to cement ratio. <strong>The</strong><br />

results indicate that the porosity in our mix is not highly connected, so it does not have detrimental effect on the<br />

permeability.<br />

<strong>The</strong> stress vs. strain curves for five PDCC specimens tested at 28 days are shown in Figure 1. From the figure,<br />

one can define the first cracking strength as the point where the curve exhibits a sharp decrease in slope from the<br />

initial linear behavior. <strong>The</strong> ultimate strength is the maximum stress carried by the material and the ultimate<br />

strain is the strain corresponding to the ultimate strength. After the peak load, cracking starts to localize and<br />

the rapid opening of a single crack is observed. <strong>The</strong> ultimate strain is hence a good indicator of material ductility,<br />

as the PDCC can be considered as a damaged homogenous material before this strain is reached. As shown in<br />

Figure 1, the PDCC mix can reach a ductility of 4.4% on average.<br />

Diameter<br />

μm<br />

Table 1 Properties of the PVA Fiber<br />

Young’s<br />

Length Elongation<br />

Modulus<br />

mm %<br />

GPa<br />

Tensile<br />

Strength<br />

MPa<br />

38 12 6.5 33 1530<br />

-118-


Table 2 Mix Proportion of the PDCC Mix<br />

Material Cement Fly Ash<br />

Silica<br />

Fume<br />

Water Sand Superplaticizer Fibre<br />

Proportion 0.18 0.8 0.02 0.22 0.2 0.0051 2% in Volume<br />

Fabrication of the Permanent Formwork<br />

Figure 1 Stress vs. Strain Curves of PDCC Mix<br />

Permanent formwork was prepared with the use of wooden moulds. Before casting of PDCC, Aslan GFRP rods<br />

(with properties shown in Table 3) were inserted through holes in the end plates of the mould and supported<br />

intermittently with spacers. Since the selected PDCC exhibited excellent workability, the formwork was<br />

fabricated without internal vibration or tampering. When the material was still fresh, lateral grooves were<br />

introduced on the surface. Flat plate formwork of 30mm thickness was prepared by direct casting into the mould.<br />

To make U-shape Formworks, a wood mould as shown in Figure 2 was employed. <strong>The</strong> mould is assembled of 3<br />

planks connected by 2 hinges. <strong>The</strong> inner surface of the mould is lined with a rubber sheet so the joints (at the<br />

hinged locations) are properly sealed to prevent water and fine particles from leaking out. <strong>The</strong>re is a wooden<br />

strip placed along each of the two side planks (See Fig. 2) to maintain a certain thickness of PDCC during<br />

casting. At the middle plank, an additional pair of strips (called the thickness adjuster) was placed to allow the<br />

casting of a thicker layer of PDCC at the bottom relative to the sides. At a suitable time after initial setting<br />

(determined by trial and error), the thickness adjuster is removed. Since the concrete is stiff enough, the middle<br />

part will remain higher than the sides. <strong>The</strong> two side planks are folded up to form the U-shape formwork.<br />

Transverse grooves (Fig. 3a) can be introduced on the formwork surface before or after the sides are folded up.<br />

As shown in Leung & Cao (2010), such a simple surface treatment can effectively improve the bonding between<br />

the PDCC formwork and the concrete to be cast. For a beam made with plain concrete and PDCC formwork<br />

with no GFRP reinforcement, discrete cracks in the concrete are arrested at the concrete/PDCC interface and<br />

turned into multiple fine cracks within the PDCC layer (Fig. 3b).<br />

Folding up<br />

Folding up<br />

Thickness of the side<br />

Thickness of the bottom<br />

Hole for GFRP rods<br />

Figure 2 Wooden Mould for U-shape Formwork<br />

In this study, the thickness at the bottom part of the U-shape formwork is 30mm to provide a proper cover to the<br />

GFRP rods inside. <strong>The</strong> two legs (formed by folding up of PDCC) are 20mm in thickness.<br />

With a reduced thickness, transportation and handling is facilitated by a reduced weight. <strong>The</strong> formwork cost is<br />

also reduced.<br />

-119-


Concrete<br />

PDCC<br />

(a)<br />

(b)<br />

Figure 3 (a) Transverse grooves on the PDCC, and (b) Crack control ability of the PDCC layer<br />

Specimen Design & Test Setups<br />

<strong>The</strong> Slab Specimens (SF and SU) were prepared using both flat plate and U-shaped formwork with 3 GFRP bars<br />

inside. <strong>The</strong> Beam Specimen BF1 and BU1 were designed to have the same Steel/GFRP ratio, but prepared with<br />

flat plate and U-shaped formwork respectively. An additional set of specimens BU2 with U-shape formwork and<br />

a different reinforcement ratio was also prepared. <strong>The</strong> design details are listed in Table 3. For each specimen<br />

type, 2 beams were prepared and tested.<br />

Table 3 Specimen Details<br />

Type Reinforcement Shape<br />

SF GFRP:3R6 Flat<br />

BF1 GFRP:4R6 Steel:2R10 High Yield Steel Flat<br />

SU GFRP:3R6 U<br />

BU1 GFRP:4R6 Steel:2R10 High Yield Steel U<br />

BU2 GFRP:5R6 Steel:4R8 Mild Steel U<br />

Test setup and reinforcement arrangement details are shown in Figure 4. Four point testing was performed on all<br />

specimens. For the beam specimens, the total span was 1800mm while the mid-span between loading points<br />

was 300mm. For the slab specimens, the total span was 2800mm and the loading points were 200mm apart. In<br />

all tests, the loading rate was set to be 0.5mm/min.<br />

750 300<br />

750<br />

1300 200<br />

1300<br />

@100spacing<br />

@100spacing<br />

GFRP Reinforcement<br />

PDCC Formwork<br />

100<br />

20<br />

150<br />

30 30<br />

250<br />

70<br />

20<br />

80<br />

30 30<br />

150<br />

150 150<br />

100 100<br />

BU1<br />

BF1<br />

SU<br />

Figure 4 Test Setup<br />

SF<br />

RESULTS<br />

Figure 5(a) and 5(b) show the typical load vs displacement behaviours for the various slabs and beams made<br />

with permanent formwork. Both slabs were able to reach the load capacity calculated from conventional<br />

reinforced concrete design. However, for the slab made with flat plate formwork (SF), delamination was<br />

-120-


observed between the formwork and cast concrete (Figure 6(a)), while the member made with the U-shaped<br />

formwork (SU) failed with the formation of a flexural crack in the middle. For the beam members BF1 and<br />

BU1, made with flat plate and U-shaped formwork respectively, BF1 failed by delamination (Figure 6(c)) at a<br />

loading much lower than that of BU1, which failed in flexure (Figure 6(d). For BU1, the failure load was<br />

higher than that predicted from conventional reinforced concrete theory. When the amount of GFRP in the<br />

formwork was further increased, as in BU2, delamination failure occurs despite the use of U-shaped formwork.<br />

However, delamination also led to a significantly higher deformation at failure.<br />

16<br />

SU<br />

14<br />

SF<br />

Load(kN)<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

0 10 20 30 40 50 60 70 80 90<br />

Displacement(mm)<br />

(a) Slabs SF & SU<br />

120<br />

BU1<br />

100<br />

BF1<br />

BU2<br />

80<br />

Load(kN)<br />

60<br />

40<br />

20<br />

0<br />

0 5 10 15 20 25 30 35 40 45<br />

Displacement(mm)<br />

(b) Beams BF1, BU1&BU2<br />

Figure 5 Test Results for the Various Beams<br />

-121-


(a)<br />

(b)<br />

(c)<br />

Figure 6 Failure Modes<br />

(d)<br />

AN EXAMPLE APPLICATION OF GFRP/PDCC FORMWORK<br />

A potential application of GFRP/PDCC formwork is in the construction of the deck of a footbridge under<br />

aggressive environment (such as marine environment or cold region where salt is used for deicing). To see if the<br />

PDCC formwork investigated above is suitable for such an application, a footbridge of 4m wide and 30m long is<br />

taken as example. We assume the use of two edge beams along the longitudinal direction, and so the concrete<br />

deck is spanning in the lateral direction. <strong>The</strong> ultimate design moment is calculated according to BS 5400<br />

(British Standard: Steel, concrete and composite bridges), which specifies a design load of 5kPa for loaded<br />

lengths of 36 m and under. For pedestrian traffic on bridges supporting footways and cycle tracks only, the live<br />

load shall be treated as uniformly distributed.<br />

For ultimate limit state design, the loads to be considered are the permanent loads, together with the appropriate<br />

primary live loads. Under this combination, the partial load factor is 1.15 for dead load and 1.5 for live load.<br />

For the sake of calculation, one can consider the deck to be composed of laterally-spanning members 0.1m in<br />

width and 0.15m in height, which is identical to the beam members tested in our study. <strong>The</strong> required moment<br />

capacity can then be calculated in the following way:<br />

G<br />

k<br />

= 24 × 0 .1 × 0 .15 = 0 .36 kN m<br />

Q<br />

k<br />

= 5 × 0 .1 = 0 .5 kN m<br />

DesignLoad = γ<br />

k<br />

G<br />

k<br />

+ γ<br />

q<br />

Q<br />

k<br />

= 1 .15 × 0 .36<br />

= 1 .164 kN m<br />

M =<br />

2<br />

PL<br />

8<br />

= 2 .33 kN ⋅ m<br />

+ 1 .5 × 0 .5<br />

For the member made with permanent formwork containing GFRP, both SF and SU failed at a maximum<br />

moment above 9.5kNm, which is over 4 times the required moment capacity calculated above. Even with an<br />

additional safety factor of 2.0, the design is more than adequate for the footbridge. <strong>The</strong> above simple calculation<br />

therefore illustrates the feasibility of using the permanent formwork in practical applications.<br />

In comparison with the flat plate formwork, the use of U-shape formwork in the above application has two<br />

advantages. Firstly, with U-shape formwork, flexural failure can be assured and the failure load is less variable<br />

than the case with debonding failure. Secondly, since the U-shape formwork has a significantly higher stiffness<br />

than the flat-plate formwork, the amount of falsework required to support the formwork during construction can<br />

be reduced.<br />

-122-


In the above example, the GFRP within the PDCC is already sufficient to provide the required load capacity. As<br />

no additional steel reinforcements (for both flexure and shear) are required, the site construction only involves<br />

the casting of concrete and this can be highly efficient. Also, since there is no steel, the corrosion problem which<br />

is a major cause of structural degradation is eliminated.<br />

CONCLUSIONS<br />

In this paper, the concept of using GFRP/PDCC permanent formwork to make durable concrete structures is<br />

first introduced. Test specimens are prepared with both flat-plate and U-shape formworks. With the flat-plate<br />

formwork containing a relatively high content of GFRP rods, interfacial delamination occurs along the<br />

PDCC/concrete interface, which leads to a reduction in failure load. With members made with U-shape<br />

formwork, the flexural failure with concrete crushing is favored, and the design failure load can be reached. In<br />

one set of tests with flat plate and U-shape formworks containing similar GFRP content, the failure load of the<br />

beam member increases by over 40% when the failure mode changes from delamination to concrete crushing<br />

under flexure. When the GFRP content becomes very high, delamination failure can occur even for the U-shape<br />

formwork. One interesting observation is that delamination can result in a higher deformation ability of the<br />

member. A simple design example is presented to show the feasibility of using permanent formwork for making<br />

the lateral spanning deck of a footbridge. <strong>The</strong> potential of the GFRP/PDCC permanent formwork for practical<br />

applications is hence demonstrated.<br />

REFERENCES<br />

Fischer, G., and Li, V.C. (2002). “Influence of matrix ductility on tension-stiffening behaviour of steel reinforced<br />

Engineered Cementitious Composites (ECC)”, ACI Struct. J., 99, 104-111.<br />

Kanda, T., and Li, V.C. (1999). “A new micromechanics design theory for pseudo strain hardening cementitious<br />

composite”, ASCE J. Eng. Mech. 125, 373-381.<br />

Lepech, M., and Li, V.C. (2005). “Water permeability of cracked cementitious composites”, Proc. 11 th Int. Conf.<br />

on Fracture, edited by A. Carpintari. (CD ROM)<br />

Leung, C.K.Y. (1996). “Design criteria for pseudo-ductile fiber composites”, ASCE J. Eng. Mech.,122, 10-18.<br />

Leung, C.K.Y., and Cao, Q. (2010). “Development of pseudo-ductile permanent formwork for durable concrete<br />

structures”, RILEM Mat. & Struct., 43, 993-1007<br />

Li, V.C., and Leung, C.K.Y. (1992). “Steady state and multiple cracking of short random fiber composites”,<br />

ASCE J. Eng. Mech. 118, 2246-2264.<br />

Li, V.C. (1993). “From micromechanics to structural engineering--the design of cementitous composites for civil<br />

engineering applications”, JSCE J. Struc. Mech. Earth. Eng., 10, 37-48.<br />

Song, G., and van Zijl, G.P.A.G. (2004). “Tailoring ECC for commercial application”, Proceedings of the 6th<br />

RILEM Symposium on Fiber-Reinforced Concretes (FRC)-BEFIB, Varenna, Italy.<br />

Wang, K., Jansen, D., Shah, S., and Karr, A. (1997). “Permeability study of cracked concrete”, Cem. & Concr.<br />

Res., 27, 381-393.<br />

Wang, S., and Li, V.C. (2007). “high-early-strength engineered cementitious composites”, ACI Mat. J., 103,<br />

97-105.<br />

Yang, E.H., Yang, Y.Z., and Li, V.C. (2007). “Use of high volumes of fly ash to improve ecc mechanical<br />

properties and material greenness”, ACI Materials J., 104, 620-628.<br />

-123-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

结 构 加 固 及 增 强 中 的 FRP 关 键 应 用 技 术 研 究<br />

1,2 1 1<br />

吴 智 深 汪 昕 吴 刚<br />

1 东 南 大 学 城 市 工 程 科 学 技 术 研 究 院 南 京 210096 中 国<br />

2 日 本 茨 城 大 学 都 市 系 统 工 学 科 日 立 316-8511 日 本<br />

摘 要 : 本 文 总 结 了 纤 维 增 强 复 合 材 料 (FRP) 在 结 构 加 固 及 增 强 领 域 的 研 究 和 应 用 成 果 , 分 析<br />

归 纳 了 目 前 急 需 克 服 的 不 足 和 需 要 高 效 开 发 FRP 关 键 技 术 , 包 括 以 下 四 个 方 面 :(1) 单 一 纤 维 , 如<br />

碳 纤 维 FRP、 玻 璃 纤 维 FRP、 芳 纶 纤 维 FRP 和 玄 武 岩 纤 维 FRP 等 作 为 结 构 体 内 或 体 外 增 强 材 料 不<br />

能 满 足 结 构 综 合 性 能 提 升 的 要 求 ;(2) 通 常 FRP 加 固 技 术 对 结 构 使 用 性 能 提 升 效 果 不 明 显 ;(3)FRP<br />

高 度 提 升 结 构 抗 震 性 能 技 术 , 如 损 伤 可 控 可 恢 复 结 构 , 还 有 待 于 开 发 ;(4) 提 升 结 构 的 可 持 续 性 发<br />

展 是 重 大 议 题 。 基 于 上 述 考 虑 和 现 状 , 结 合 作 者 研 究 团 队 的 研 究 成 果 , 重 点 介 绍 了 若 干 提 升 FRP 材<br />

料 及 增 强 结 构 的 关 键 技 术 , 以 解 决 或 改 善 上 述 问 题 , 主 要 包 括 混 杂 FRP 的 研 究 应 用 , 预 应 力 FRP<br />

技 术 , 损 伤 可 控 的 抗 震 设 计 理 论 及 利 用 混 杂 FRP 材 料 实 现 可 持 续 性 结 构 。 此 外 ,FRP 加 固 及 增 强 结<br />

构 将 来 面 临 的 挑 战 和 研 究 趋 势 在 文 中 也 进 行 了 讨 论 。<br />

关 键 词 :FRP 加 固 增 强 混 杂 预 应 力 抗 震 可 持 续 性<br />

KEY FRP TECHNOLOGIES IN STRUCTURAL RETROFITTING AND<br />

STRENGTHENING<br />

Wu Zhishen 1,2 Wang Xin 1 Wu Gang 1<br />

1 International Institute for Urban Systems Engineering, Southeast <strong>University</strong>, Nanjing, 210096<br />

2 Department of Urban and Civil Engineering, Ibaraki <strong>University</strong>, Hitachi, 316-8511, Japan<br />

Abstract: <strong>The</strong> paper reviewed the current R&D and applications of fiber reinforced polymer (FRP) composites<br />

in retrofitting and strengthening civil infrastructure and clarified the limitations which should be overcome<br />

directly and the challenges on developing effective FRP technologies, which comprised the following four<br />

aspects: 1) using of single type of FRP such as Carbon FRP(CFRP), Glass FRP(GFRP), Aramid FRP(AFRP)<br />

and Basalt FRP(BFRP) etc. as both external and internal reinforcements cannot receive integrated performance<br />

of strengthened structures; 2) limited improvements and enhancements of structural behavior under<br />

serviceability can be realized by regular FRP bond retrofitting techniques; 3) innovation of structures with<br />

highly integrated seismic behavior including damage controllability and high recoverability through adopting<br />

advanced FRP composites is greatly expected; 4) enhancement of structural sustainability has become a big<br />

issue. Based on these kinds of considerations and situations, several new attempts of improving and enhancing<br />

integrated performances of both FRP composites and FRP-strengthened structures are made recently by author’s<br />

research team. Some of major solutions/research results which consist of developments of hybrid FRP<br />

composites, prestressing FRP technology, damage controllable design methods and sustainable structures with<br />

hybrid FRP composites and technologies are summarized here. Moreover, the future challenges and R&D trends<br />

are also discussed through this paper.<br />

Keywords: FRP, retrofitting and strengthening, hybrid, prestressing, seismic behavior, sustainability.<br />

一 、 研 究 背 景<br />

以 碳 纤 维 为 代 表 的 纤 维 增 强 复 合 材 料 (Fiber Reinforced Polymer, FRP) 的 商 业 化 应 用 始 于 20 世<br />

纪 70 年 代 , 起 初 应 用 于 体 育 用 品 领 域 [1] ,80 年 代 后 FRP 材 料 逐 渐 开 始 在 土 建 交 通 领 域 得 到 了 推 广<br />

应 用 , 到 90 年 代 中 期 因 日 本 阪 神 地 震 对 土 建 交 通 基 础 设 施 造 成 了 极 大 的 灾 害 ,FRP 迅 速 发 展 成 为<br />

结 构 抗 震 加 固 的 重 要 手 段 [2] 。 此 后 ,FRP 在 结 构 加 固 及 增 强 领 域 得 到 了 广 泛 的 关 注 和 深 入 研 究 ,FRP<br />

也 从 初 期 的 结 构 加 固 改 造 , 到 作 为 建 筑 结 构 增 强 材 料 应 用 于 新 建 结 构 ; 从 开 始 的 FRP 纤 维 布 / 板 ,<br />

到 多 种 形 式 的 纤 维 复 合 索 、 网 格 、 型 材 , 应 用 范 围 和 应 用 形 式 越 来 越 广 泛 。FRP 材 料 的 发 展 也 为 研<br />

究 开 发 综 合 高 性 能 土 木 工 程 结 构 设 施 和 既 有 工 程 设 施 的 加 固 增 强 及 功 能 提 升 提 供 良 好 的 选 择 途 径<br />

[3-8] 。<br />

FRP 在 土 建 交 通 结 构 中 的 广 泛 应 用 , 根 据 应 用 的 目 的 不 同 , 主 要 可 分 为 两 类 : 一 是 用 于 既 有 结<br />

-124-


构 的 加 固 修 复 , 二 是 作 为 增 强 材 料 代 替 混 凝 土 结 构 中 的 普 通 钢 筋 或 用 于 开 发 新 型 FRP- 混 凝 土 复 合 结<br />

构 。FRP 加 固 技 术 开 始 于 上 世 纪 80 年 代 , 主 要 特 点 在 于 不 但 能 够 避 免 传 统 加 固 方 法 施 工 周 期 长 、<br />

难 度 大 、 费 用 高 、 复 杂 结 构 不 易 加 固 等 缺 点 , 而 且 不 增 加 结 构 的 原 有 尺 寸 、 不 额 外 增 加 结 构 的 重 量<br />

负 担 、 也 不 需 要 大 型 的 施 工 设 备 和 宽 敞 的 施 工 空 间 。 因 此 ,FRP 修 复 加 固 技 术 在 恢 复 和 提 升 钢 筋 混<br />

凝 土 (RC) 结 构 功 能 方 面 有 着 相 当 大 的 优 势 , 多 年 来 引 起 世 界 各 国 有 关 科 研 院 校 及 企 业 的 高 度 重 视 ,<br />

目 前 已 经 比 较 成 熟 的 运 用 于 各 种 需 要 加 固 或 改 造 的 多 种 结 构 形 式 。 除 用 作 结 构 加 固 改 造 外 ,FRP 材<br />

料 在 近 20 年 来 正 在 不 断 尝 试 用 于 增 强 结 构 新 建 , 如 国 内 外 ( 特 别 是 北 欧 、 美 国 等 冬 季 需 要 使 用 大<br />

量 融 雪 盐 的 国 家 和 地 区 ) 正 在 积 极 开 展 用 FRP 复 合 材 料 代 替 传 统 钢 材 ( 如 制 作 成 FRP 筋 埋 入 混 凝<br />

土 ) 制 作 混 凝 土 结 构 的 研 究 , 以 避 免 混 凝 土 中 钢 材 腐 蚀 所 带 来 的 危 害 [9] ; 日 美 欧 洲 和 中 国 都 开 展 了<br />

各 种 FRP 材 料 与 传 统 建 材 相 结 合 的 复 合 结 构 的 研 究 , 如 FRP- 混 凝 土 组 合 梁 板 结 构 、FRP 管 混 凝 土<br />

柱 、FRP- 混 凝 土 - 钢 组 合 结 构 等 [9] , 这 些 组 合 结 构 充 分 发 挥 了 各 种 材 料 的 优 势 , 取 得 综 合 高 性 能 ; 此<br />

外 ,FRP 材 料 作 为 独 立 的 结 构 单 元 , 整 体 或 部 分 作 为 结 构 也 已 开 展 了 初 步 的 研 究 , 如 FRP 管 网 架 结<br />

构 ,FRP 编 织 结 构 ,FRP 拉 索 大 跨 斜 拉 桥 结 构 ,FRP 型 材 作 为 桥 面 板 或 箱 梁 , 这 些 结 构 形 式 能 够 大<br />

幅 的 减 轻 原 先 结 构 的 重 量 , 提 升 结 构 的 耐 久 性 能 和 长 期 安 全 性 , 并 且 能 够 实 现 传 统 材 料 所 无 法 达 到<br />

的 跨 度 要 求 , 如 大 跨 斜 拉 桥 , 大 跨 编 织 结 构 等 [10-12] 。<br />

FRP 在 结 构 加 固 及 增 强 方 面 的 研 究 和 应 用 经 过 近 30 年 的 发 展 , 虽 然 很 多 理 论 , 方 法 和 技 术 已<br />

相 当 成 熟 , 但 有 由 于 FRP 材 料 本 身 物 理 、 力 学 特 性 的 限 制 和 结 构 性 能 要 求 的 不 断 提 高 , 如 FRP 的<br />

相 对 低 弹 性 模 量 , 破 断 延 伸 率 低 , 强 度 离 散 大 等 固 有 特 性 , 及 结 构 综 合 高 强 度 、 刚 度 和 延 性 要 求 等 。<br />

因 此 , 在 FRP 实 现 结 构 加 固 增 强 和 高 性 能 化 上 , 不 可 避 免 的 出 现 一 些 限 制 因 素 和 瓶 颈 问 题 , 分 析 概<br />

括 有 以 下 几 个 方 面 :(1) 单 一 纤 维 不 能 满 足 结 构 综 合 性 能 提 升 的 要 求 。 结 构 的 综 合 性 能 通 常 表 现 为<br />

对 强 度 、 刚 度 和 延 性 的 综 合 要 求 , 单 一 的 提 升 强 度 , 没 有 足 够 的 刚 度 保 证 , 结 构 将 发 生 不 能 容 许 的<br />

过 大 变 形 ; 同 样 , 如 果 没 有 足 够 的 延 性 , 结 构 将 可 能 发 生 脆 性 破 坏 , 这 对 于 保 证 生 命 安 全 是 不 能 允<br />

许 的 。 由 于 结 构 的 这 种 综 合 性 能 要 求 , 结 构 材 料 就 必 须 满 足 相 应 的 强 度 、 刚 度 和 延 性 的 综 合 指 标 ,<br />

而 单 一 FRP 材 料 , 通 常 无 法 同 时 满 足 这 种 综 合 性 能 要 求 , 如 碳 纤 维 具 有 很 高 的 强 度 , 刚 度 , 但 破 坏<br />

时 的 延 伸 率 很 低 , 脆 性 特 征 明 显 ;E 玻 璃 纤 维 , 虽 然 延 伸 率 较 高 , 但 强 度 和 刚 度 较 低 , 尤 其 是 刚 度 ,<br />

仅 为 钢 材 的 三 分 之 一 左 右 。 因 此 , 在 结 构 的 加 固 及 增 强 中 , 要 发 挥 FRP 材 料 的 优 势 , 保 证 结 构 的 综<br />

合 性 能 , 势 必 要 改 善 单 一 FRP 材 料 的 不 足 。(2) 既 有 FRP 加 固 技 术 对 结 构 使 用 性 能 提 升 效 果 不 明<br />

显 。 利 用 FRP 对 既 有 结 构 加 固 , 由 于 FRP 材 料 的 弹 性 模 量 总 体 低 于 钢 材 ( 除 碳 纤 维 与 钢 材 相 当 ),<br />

而 且 既 有 结 构 本 身 已 在 持 荷 作 用 下 , 加 固 后 FRP 材 料 和 原 结 构 共 同 工 作 ,FRP 材 料 力 学 性 能 远 未 充<br />

分 发 挥 , 结 构 既 已 进 入 屈 服 阶 段 。 结 构 的 开 裂 荷 载 和 屈 服 荷 载 很 难 得 到 明 显 改 善 , 虽 然 极 限 承 载 力<br />

能 够 得 到 明 显 提 升 , 但 对 于 结 构 处 于 使 用 状 态 下 的 性 能 却 无 法 得 到 明 显 提 升 。(3)FRP 增 强 抗 震 新<br />

结 构 需 要 突 破 。 由 于 目 前 抗 震 设 计 思 想 的 局 限 ,“ 中 震 可 修 ” 未 具 体 量 化 , 导 致 大 量 结 构 在 中 大 震 作<br />

用 下 发 生 过 大 的 不 可 恢 复 变 形 , 结 构 虽 然 不 倒 , 但 也 无 法 继 续 使 用 及 修 复 。 对 既 有 结 构 , 采 用 FRP<br />

布 约 束 或 嵌 入 式 加 固 的 方 法 , 可 以 明 显 改 善 传 统 RC 结 构 的 抗 震 性 能 不 足 , 使 结 构 在 屈 服 后 具 有 明<br />

显 的 二 次 刚 度 和 较 小 的 残 余 变 形 。 但 对 于 新 建 结 构 , 如 何 采 用 FRP 筋 或 FRP 和 钢 材 复 合 筋 实 现 结<br />

构 在 地 震 作 用 下 损 伤 可 控 可 修 复 , 仍 需 要 大 力 研 究 , 其 中 包 括 新 抗 震 设 计 思 想 、 高 性 能 材 料 和 设 计<br />

方 法 的 建 立 ;(4) 结 构 综 合 高 性 能 和 可 持 续 性 能 需 要 高 度 提 升 。 仅 采 用 传 统 结 构 材 料 往 往 受 到 力 学<br />

性 能 限 制 , 无 法 实 现 大 跨 高 性 能 结 构 ; 而 一 味 采 用 先 进 的 纤 维 复 合 材 料 造 价 往 往 成 为 瓶 颈 因 素 , 因<br />

此 如 何 合 理 使 用 先 进 的 纤 维 复 合 材 料 实 现 传 统 材 料 无 法 达 到 的 高 性 能 结 构 将 是 结 构 工 程 发 展 的 一<br />

个 重 要 方 向 。 另 外 , 除 考 虑 结 构 的 基 本 力 学 性 能 外 , 结 构 在 寿 命 周 期 内 会 受 到 各 种 恶 劣 环 境 的 影 响<br />

及 长 期 荷 载 的 作 用 , 结 构 的 可 持 续 性 不 容 忽 视 。 目 前 的 大 多 数 研 究 , 往 往 只 局 限 于 短 期 力 学 性 能 ,<br />

结 构 的 长 期 性 能 无 法 得 到 定 量 的 评 价 , 这 也 很 大 的 限 制 了 新 兴 FRP 材 料 的 广 泛 使 用 。 因 此 , 如 何 采<br />

用 FRP 材 料 作 为 结 构 构 件 , 或 FRP 材 料 与 传 统 结 构 材 料 的 有 机 结 合 , 以 实 现 结 构 体 系 的 高 度 可 持<br />

续 化 发 展 是 目 前 急 需 发 展 的 一 个 重 要 方 向 。<br />

由 此 可 见 ,FRP 作 为 高 性 能 的 结 构 加 固 和 增 强 材 料 经 过 多 年 的 研 究 和 应 用 , 在 积 累 经 验 的 同 时<br />

也 充 分 暴 露 了 自 身 的 不 足 , 如 何 改 善 及 克 服 这 些 问 题 , 将 决 定 FRP 能 否 向 更 高 更 远 的 方 向 发 展 。 本<br />

文 将 在 作 者 研 究 团 队 多 年 研 究 成 果 的 基 础 上 , 针 对 上 述 FRP 应 用 瓶 颈 问 题 , 重 点 介 绍 利 用 FRP 实<br />

现 结 构 高 安 全 性 、 高 性 能 化 和 高 度 可 持 续 性 的 一 些 关 键 应 用 技 术 , 并 通 过 这 些 关 键 技 术 特 征 的 分 析<br />

为 解 决 FRP 当 前 应 用 的 瓶 颈 提 供 参 考 和 借 鉴 。<br />

二 、FRP 基 本 力 学 性 能 及 制 品<br />

-125-


4000<br />

3500<br />

3000<br />

©<br />

a<br />

£<br />

P<br />

¨M<br />

È£<br />

¿<br />

­Ç<br />

¹À<br />

¿<br />

2500<br />

2000<br />

1500<br />

1000<br />

碳 纤 维 布<br />

( 高 弹 模 型 )<br />

PBO 纤 维<br />

芳 纶<br />

纤 维<br />

碳 纤 维<br />

( 高 强 型 )<br />

玄 武 岩 纤 维<br />

E 玻 璃 纤 维<br />

Dyneema 纤 维<br />

500<br />

0<br />

钢 筋<br />

0 10000 20000 30000 40000<br />

应 变 ( 礶 )<br />

图 1 FRP 材 料 应 力 - 应 变 关 系<br />

首 先 对 常 用 的 FRP 材 料 进 行 回 顾 和 比 较 。 常 用 的 FRP 材 料 由 碳 纤 维 (Carbon fibers)、PBO 纤<br />

维 (Poly-p-phenylenebenzobisthiazole)、 玻 璃 纤 维 (Glass fibers)、 玄 武 岩 纤 维 (Basalt fibers) 或 芳<br />

纶 纤 维 (Aramid fibers) 分 别 与 基 体 材 料 ( 如 环 氧 树 脂 ) 含 浸 硬 化 后 复 合 形 成 。 相 对 于 传 统 建 筑 材<br />

料 ( 钢 材 , 混 凝 土 , 木 材 )FRP 材 料 具 有 优 越 的 力 学 ( 高 比 强 度 ) 和 物 理 化 学 性 能 ( 轻 质 、 耐 腐 蚀<br />

等 ), 及 其 他 优 秀 特 性 , 如 绝 缘 ( 除 碳 纤 维 弱 导 电 )、 耐 高 低 温 等 , 其 密 度 只 有 钢 材 的 三 分 之 一 到 四<br />

分 之 一 , 但 拉 伸 强 度 为 普 通 低 碳 钢 的 5 倍 以 上 , 与 预 应 力 钢 丝 / 索 相 当 甚 至 更 高 。 碳 纤 维 FRP(CFRP)<br />

在 各 种 纤 维 复 合 材 料 中 , 拥 有 最 为 突 出 的 力 学 性 能 和 化 学 稳 定 性 , 典 型 的 CFRP 抗 拉 强 度 为 3400MPa,<br />

弹 性 模 量 达 到 230GPa, 密 度 只 有 1.8 kg/cm2, 并 且 能 够 抵 抗 酸 碱 盐 紫 外 线 等 各 种 环 境 腐 蚀 , 耐 高 温<br />

达 600 度 。PBO 纤 维 不 仅 具 有 和 碳 纤 维 相 似 甚 至 更 高 的 力 学 性 能 , 而 且 具 有 良 好 的 冲 击 能 力 吸 收 性<br />

能 , 这 也 意 味 着 PBO 纤 维 作 为 干 丝 具 有 更 高 的 可 张 拉 性 能 。 玻 璃 纤 维 FRP(GFRP) 力 学 和 化 学 性<br />

能 在 常 用 各 种 纤 维 复 合 材 料 中 相 对 较 低 , 但 因 其 相 对 低 廉 的 价 格 得 到 广 泛 的 应 用 , 常 用 的 E 玻 璃 纤<br />

维 FRP 的 抗 拉 强 度 为 1500MPa, 弹 性 模 量 为 80GPa, 密 度 为 2.6 kg/cm2,GFRP 虽 然 能 够 抵 抗 酸 ,<br />

盐 和 紫 外 线 等 腐 蚀 作 用 , 但 对 碱 作 用 敏 感 , 并 且 适 用 温 度 范 围 相 对 较 小 -60~+250℃, 在 长 期 荷 载 下<br />

会 出 现 应 力 破 断 现 象 , 应 用 面 有 一 定 的 限 制 。 玄 武 岩 纤 维 FRP(BFRP) 是 一 种 以 纯 天 然 火 山 岩 ( 玄<br />

武 岩 为 主 ) 为 原 料 , 在 1450~1500℃ 高 温 熔 融 后 拉 丝 而 成 的 连 续 纤 维 。 连 续 玄 武 岩 纤 维 在 我 国 已 经<br />

能 产 业 化 工 业 化 生 产 。 该 纤 维 具 有 高 性 价 比 、 耐 高 低 温 (-269~650℃)、 抗 水 损 害 性 能 好 、 耐 紫 外<br />

线 、 电 绝 缘 、 纤 维 表 面 呈 极 性 、 纯 天 然 环 保 、 防 火 阻 燃 等 特 点 [8,13]。 从 力 学 性 能 角 度 , 玄 武 岩 纤 维<br />

比 E 玻 璃 纤 维 具 有 更 高 的 抗 拉 强 度 (2100MPa) 和 弹 性 模 量 (91GPa) 以 及 更 广 的 适 用 温 度 范 围 。<br />

芳 纶 纤 维 FRP(AFRP) 的 力 学 性 能 介 于 CFRP 和 BFRP 之 间 , 但 长 期 荷 载 作 用 下 会 出 现 应 力 破 断 现<br />

象 , 而 且 化 学 性 能 中 对 紫 外 线 很 敏 感 , 加 上 价 格 和 CFRP 接 近 , 因 此 实 际 应 用 不 如 CFRP 和 GFRP<br />

广 泛 。 上 述 几 种 纤 维 材 料 和 其 他 几 种 土 木 领 域 应 用 的 纤 维 材 料 典 型 的 应 力 - 应 变 关 系 见 图 1 所 示 , 可<br />

见 相 对 于 传 统 钢 材 , 纤 维 复 合 材 料 具 有 明 显 的 强 度 优 势 , 但 部 分 纤 维 材 料 的 刚 度 不 足 或 延 伸 率 不 足<br />

也 是 需 要 避 免 和 改 善 的 重 要 问 题 。<br />

上 述 纤 维 材 料 通 过 与 树 脂 等 基 体 材 料 复 合 , 在 手 糊 或 拉 挤 或 真 空 辅 助 等 工 艺 下 , 形 成 结 构 加 固<br />

及 增 强 用 的 各 种 FRP 制 品 , 如 图 2 所 示 。<br />

-126-


FRP 网 格<br />

FRP 筋 和 索<br />

FRP 布<br />

FRP 型 材<br />

FRP 板<br />

图 2 各 种 FRP 制 品<br />

三 、 混 杂 FRP 技 术<br />

如 上 文 所 说 , 单 一 FRP 材 料 很 难 满 足 结 构 的 综 合 性 能 要 求 , 如 强 度 、 刚 度 、 延 性 、 长 期 性 能 等 ,<br />

因 此 要 实 现 结 构 的 综 合 高 性 能 , 改 善 单 一 FRP 材 料 的 性 能 局 限 势 在 必 行 。 根 据 纤 维 复 合 材 料 的 本 身<br />

特 征 , 采 用 混 杂 方 式 实 现 材 料 的 性 能 设 计 和 性 能 提 升 已 被 证 明 是 一 条 有 效 和 稳 定 的 途 径 。<br />

3.1. 混 杂 FRP 设 计 理 念<br />

纤 维 按 力 学 性 能 分 为 3 大 类 : 高 弹 性 模 量 纤 维 , 高 强 度 纤 维 和 高 延 性 纤 维 。 高 弹 性 模 量 纤 维 被<br />

活 用 来 提 高 材 料 在 正 常 使 用 情 况 下 的 刚 度 , 如 在 混 凝 土 工 程 中 高 弹 性 模 量 纤 维 的 使 用 可 以 通 过 提 高<br />

结 构 的 刚 度 来 控 制 结 构 在 正 常 情 况 下 的 变 形 ; 高 强 度 纤 维 被 活 用 来 提 高 材 料 / 结 构 的 承 载 能 力 ; 高 延<br />

性 纤 维 可 以 活 用 来 提 高 材 料 / 结 构 的 变 形 能 力 , 特 别 是 极 限 条 件 下 如 地 震 等 情 况 下 结 构 的 高 延 性 要 求<br />

等 。 通 过 适 当 混 杂 配 比 , 三 类 纤 维 的 共 同 作 用 可 以 提 高 材 料 的 综 合 力 学 性 能 , 如 高 刚 度 、 高 承 载 能<br />

力 以 及 高 延 性 等 。 从 而 可 以 解 决 单 种 纤 维 增 强 FRP 复 合 材 料 脆 性 和 延 性 不 足 所 带 来 的 一 些 问 题 。 根<br />

据 实 际 情 况 可 以 对 混 杂 FRP 材 料 中 的 纤 维 种 类 及 混 杂 比 例 进 行 适 当 的 调 整 。<br />

在 此 基 础 上 , 为 进 一 步 发 挥 混 杂 设 计 优 势 , 采 用 传 统 钢 材 ( 包 括 钢 丝 和 钢 筋 ) 与 纤 维 材 料 复 合 ,<br />

结 合 两 者 优 势 , 进 一 步 发 挥 混 杂 材 料 的 特 点 。 如 采 用 玄 武 岩 纤 维 和 高 强 钢 丝 混 杂 , 可 以 提 升 整 体 初<br />

始 弹 性 模 量 , 满 足 结 构 刚 度 要 求 , 并 且 由 于 混 杂 少 量 钢 丝 , 混 杂 纤 维 复 合 材 料 整 体 重 量 不 会 明 显 增<br />

加 , 但 造 价 能 够 得 到 进 一 步 降 低 。 采 用 玄 武 岩 纤 维 和 钢 筋 复 合 , 利 用 玄 武 岩 纤 维 材 料 的 先 弹 性 力 学<br />

特 征 和 钢 筋 的 屈 服 特 性 , 混 杂 形 成 具 有 较 高 初 始 刚 度 和 显 著 二 次 刚 度 的 复 合 筋 材 , 为 提 高 结 构 抗 震<br />

性 能 、 使 用 性 能 和 耐 久 性 能 具 有 显 著 作 用 。<br />

3.2. 混 杂 FRP 设 计 理 论 及 方 法<br />

混 杂 设 计 总 的 原 则 是 在 一 种 纤 维 出 现 断 裂 破 坏 后 所 产 生 的 冲 击 以 及 其 所 承 担 载 荷 能 够 平 稳 的<br />

被 其 他 纤 维 来 承 担 , 或 钢 材 屈 服 或 进 入 非 弹 性 阶 段 后 纤 维 承 担 荷 载 提 高 整 体 弹 性 模 量 , 如 混 杂 FRP<br />

复 合 材 料 中 的 高 弹 性 摸 量 纤 维 的 断 裂 所 产 生 的 载 荷 转 移 和 冲 击 荷 载 可 以 有 效 地 被 高 强 度 和 高 延 性<br />

纤 维 吸 收 , 钢 筋 屈 服 后 由 纤 维 承 担 荷 载 并 维 持 一 定 的 弹 性 模 量 。 在 这 个 原 则 的 基 础 上 , 作 者 研 究 团<br />

队 研 究 了 各 种 纤 维 及 钢 材 在 混 杂 FRP 复 合 材 料 中 的 混 杂 比 例 关 系 并 建 立 了 混 杂 FRP 纤 维 复 合 材 料<br />

的 设 计 理 论 。<br />

(1) 多 种 纤 维 混 杂 设 计<br />

碳 纤 维 由 于 其 脆 性 特 征 , 往 往 导 致 结 构 延 性 不 足 , 使 结 构 出 现 脆 性 破 坏 或 在 地 震 作 用 下 过 早 的<br />

丧 失 承 载 力 。 通 过 混 杂 高 延 性 纤 维 或 钢 筋 , 可 以 利 用 这 些 材 料 的 延 性 特 征 , 设 计 混 杂 材 料 整 体 取 得<br />

高 延 性 特 征 , 从 而 保 证 结 构 的 延 性 破 坏 需 要 。 混 杂 设 计 见 下 图 所 示 , 图 中 采 用 三 种 纤 维 : 高 弹 摸 、<br />

高 强 度 和 高 延 性 纤 维 , 通 过 一 定 的 混 杂 比 例 , 实 现 初 期 高 刚 度 , 中 期 高 强 度 和 后 期 高 延 性 [14] 。<br />

-127-


图 3 混 杂 复 合 材 料 性 能 设 计 图<br />

由 于 混 杂 FRP 复 合 材 料 中 因 延 伸 率 较 低 的 纤 维 断 裂 造 成 的 载 荷 降 低 会 引 起 应 力 波 动 现 象 , 这 种<br />

应 力 波 动 会 造 成 对 剩 余 纤 维 材 料 和 结 构 的 冲 击 , 因 而 在 设 计 中 尽 量 降 低 这 种 波 动 的 幅 度 , 在 混 杂 设<br />

计 中 提 出 了 主 动 控 制 应 力 波 动 的 设 计 方 法 , 即 通 过 合 理 的 混 杂 设 计 和 配 比 来 对 应 力 波 动 进 行 有 效 的<br />

控 制 并 降 低 其 给 其 他 纤 维 和 结 构 所 带 来 的 冲 击 , 另 外 还 可 以 选 用 能 量 吸 收 能 力 ( 高 耐 冲 击 性 ) 好 的<br />

纤 维 材 料 来 吸 收 较 低 延 性 纤 维 断 裂 所 产 生 的 冲 击 能 来 控 制 应 力 波 动 [15] , 如 图 4 所 示 。<br />

Evaluation index (Exp.)<br />

1<br />

0.9<br />

0.8<br />

0.7<br />

0.6<br />

0.5<br />

0.4<br />

0.3<br />

0.2<br />

0.1<br />

0<br />

Experimental values<br />

Average line<br />

0.00 0.20 0.40 0.60 0.80 1.00<br />

Control index (Cal.)<br />

Evaluation index (Exp.)<br />

1<br />

0.9<br />

0.8<br />

0.7<br />

0.6<br />

0.5<br />

0.4<br />

0.3<br />

0.2<br />

0.1<br />

0<br />

Experimental values<br />

Average line<br />

0.00 0.20 0.40 0.60 0.80 1.00<br />

Control index (Cal.)<br />

Evaluation index (Exp.)<br />

1<br />

0.9<br />

0.8<br />

0.7<br />

0.6<br />

0.5<br />

0.4<br />

0.3<br />

0.2<br />

0.1<br />

0<br />

Experimental values<br />

Average line<br />

0.00 0.20 0.40 0.60 0.80 1.00<br />

Control index (Cal.)<br />

(a) 高 弹 摸 / 高 强 度 碳 纤 维 (b) 高 弹 摸 碳 纤 维 /PBO 纤 维 (c) 高 弹 摸 碳 纤 维 /Dyneema 纤 维<br />

图 4 应 力 波 动 控 制 研 究<br />

通 过 大 量 不 同 比 例 混 杂 的 纤 维 材 料 试 验 研 究 , 包 括 高 弹 摸 碳 纤 维 、 高 强 度 碳 纤 维 、PBO 纤 维 和<br />

Dyneema 纤 维 的 混 杂 , 得 到 了 保 证 较 低 应 力 波 动 以 实 现 混 杂 设 计 的 纤 维 比 例 关 系 。 由 图 3 可 以 看 出 ,<br />

采 用 高 弹 模 碳 纤 维 和 PBO 纤 维 混 杂 , 当 两 者 的 比 例 控 制 在 0.7 以 下 时 , 应 力 波 动 范 围 可 以 稳 定 在 0.3<br />

以 下 , 相 比 较 而 言 高 强 度 碳 纤 维 和 Dyneema 纤 维 的 混 杂 比 例 要 控 制 在 0.5 以 下 , 才 能 保 证 稳 定 的 0.3<br />

应 力 波 动 。PBO 纤 维 的 能 量 吸 收 能 力 起 到 重 要 作 用 。 上 述 研 究 用 于 定 量 的 指 导 混 杂 纤 维 的 比 例 设 计 ,<br />

确 保 混 杂 效 果 的 发 挥 。 通 过 对 应 力 波 动 的 研 究 , 能 够 合 理 设 计 混 杂 纤 维 复 合 材 料 的 力 学 特 征 [16] , 如<br />

图 3 所 示 , 为 高 弹 摸 、 高 强 度 碳 纤 维 和 高 延 性 纤 维 ( 玻 璃 纤 维 或 Dyneema 纤 维 ) 的 混 杂 复 合 材 料 ,<br />

实 现 了 和 混 杂 设 计 图 一 致 的 混 杂 纤 维 复 合 材 料 , 参 见 图 5。<br />

玻 璃 纤 维 : 高 弹 摸 碳 纤 维 : 高 强 度 碳 纤 维<br />

Dyneema 纤 维 : 高 弹 摸 碳 纤 维 : 高 强 度 碳 纤 维<br />

图 5 三 种 纤 维 混 杂 设 计 试 验 结 果 图<br />

-128-


(2) 钢 筋 - 连 续 纤 维 复 合 筋 混 杂 设 计<br />

FRP 具 有 强 度 高 、 弹 模 低 、 延 性 差 、 耐 久 性 好 、 重 量 轻 等 特 点 , 而 钢 材 具 有 强 度 低 、 弹 模 高 、<br />

延 性 好 、 耐 久 性 差 、 重 量 重 等 特 点 , 两 者 互 补 性 极 强 , 复 合 后 可 以 扬 长 避 短 , 得 到 综 合 性 能 更 高 的<br />

钢 - 连 续 纤 维 复 合 材 料 。 另 外 , 线 弹 性 的 FRP 与 弹 塑 性 的 钢 材 复 合 还 可 以 带 来 力 学 性 能 上 的 变 化 ,<br />

如 得 到 的 钢 - 连 续 纤 维 复 合 筋 (SFCB) 具 有 稳 定 的 二 次 刚 度 ( 图 6)。 其 特 点 是 能 动 地 控 制 结 构 或 构<br />

件 的 屈 服 后 刚 度 ( 二 次 刚 度 )、 震 后 残 余 变 形 、 极 限 状 态 的 破 坏 模 式 以 及 结 构 系 统 耗 能 机 理 , 为 实<br />

现 “ 大 震 ” 不 倒 乃 至 “ 大 震 ” 可 修 的 定 量 化 设 计 提 供 了 有 效 的 途 径 。<br />

70<br />

60<br />

B20<br />

纤 维 断 裂<br />

Steel<br />

rebar<br />

FRP<br />

SFCB<br />

荷 载 (kN)<br />

50<br />

40<br />

屈 服<br />

30<br />

试 件 M2<br />

20<br />

试 件 M4<br />

试 件 M8<br />

10<br />

试 件 M10<br />

理 论 计 算<br />

0<br />

0 5000 10000 15000 20000 25000 30000 35000<br />

应 变 (με)<br />

图 6 钢 筋 和 FRP 复 合 实 现 稳 定 二 次 刚 度<br />

图 7 典 型 的 SFCB 应 力 - 应 变 关 系<br />

根 据 材 料 的 复 合 法 则 推 导 出 SFCB 单 向 拉 伸 的 理 论 计 算 模 型 , 并 与 试 验 数 据 进 行 比 较 , 吻 合 良 好 ,<br />

证 明 了 可 以 根 据 复 合 法 则 对 SFCB 的 性 能 进 行 预 测 和 设 计 , 典 型 的 SFCB 应 力 - 应 变 关 系 如 图 7 所 示<br />

[17] 。<br />

70<br />

60<br />

SFCB (S10-B30)<br />

钢 筋<br />

荷 载 (kN)<br />

50<br />

40<br />

30<br />

20<br />

10<br />

峰 值 应 变<br />

卸 载 残 余 应 变<br />

0<br />

0 5000 10000 15000 20000 25000<br />

应 变 (με)<br />

图 8 SFCB 和 普 通 钢 筋 往 复 拉 伸 力 学 性 能 比 较<br />

往 复 拉 伸 典 型 SFCB 的 荷 载 - 应 变 关 系 如 图 8 所 示 , 在 同 样 的 卸 载 应 变 下 ,SFCB 具 有 较 小 的 残<br />

余 应 变 , 从 而 使 SFCB 增 强 混 凝 土 结 构 在 震 后 具 有 较 小 的 残 余 变 形 。<br />

(3) 钢 丝 - 连 续 纤 维 复 合 板 的 混 杂 设 计<br />

混 杂 设 计 采 用 高 强 钢 丝 和 FRP 复 合 , 要 求 高 强 钢 丝 均 匀 分 布 在 FRP 片 材 之 间 。 采 用 玄 武 岩 纤<br />

维 和 高 强 钢 丝 混 杂 后 的 复 合 板 力 学 性 能 如 图 9 所 示 。 复 合 板 相 对 玄 武 岩 FRP 具 有 较 高 的 初 期 刚 度 ,<br />

同 时 保 留 了 玄 武 岩 纤 维 的 高 强 度 和 一 定 的 延 性 , 材 料 综 合 性 能 得 到 改 善 [18] 。<br />

2400<br />

2000<br />

1600<br />

a<br />

)<br />

P<br />

1200 (M<br />

s<br />

S<br />

tre<br />

800<br />

400<br />

0<br />

HS Steel<br />

BFRP<br />

Hybrid B/SFRP 20%<br />

Hybrid B/SFRP 30%<br />

0 0.005 0.01 0.015 0.02 0.025<br />

Strain (e)<br />

-129-


图 9 钢 丝 - 纤 维 复 合 板 应 力 - 应 变 关 系<br />

图 10 钢 丝 直 径 对 混 杂 效 果 影 响<br />

为 了 达 到 最 优 化 的 FRP 与 高 强 钢 丝 混 杂 设 计 的 效 果 , 针 对 钢 丝 接 触 面 积 和 混 杂 效 果 进 行 了 研<br />

究 , 如 图 10 所 示 。 试 验 研 究 了 0.25mm,0.3mm 和 0.5mm 直 径 钢 丝 和 玄 武 岩 纤 维 布 混 杂 效 果 , 结 果<br />

表 明 采 用 较 细 直 径 的 钢 丝 能 够 保 证 纤 维 和 钢 丝 之 间 的 粘 结 性 能 , 而 直 径 相 对 较 大 的 钢 丝 在 高 应 力 状<br />

态 下 则 会 发 生 滑 移 , 影 响 材 料 整 体 性 能 [18] 。<br />

3.3. FRP 混 杂 效 应 研 究<br />

通 过 纤 维 之 间 混 杂 的 方 法 得 到 混 杂 FRP 材 料 , 除 了 能 够 提 升 材 料 的 基 本 力 学 性 能 外 , 混 杂 设 计<br />

对 材 料 的 其 他 性 能 , 如 疲 劳 性 能 、 耐 久 性 能 和 耐 高 温 性 能 也 会 发 生 影 响 , 基 于 此 作 者 研 究 团 队 开 展<br />

了 以 下 研 究 , 旨 在 揭 示 及 理 解 混 杂 设 计 对 提 升 材 料 综 合 性 能 作 用 。<br />

[19]<br />

(1) 纤 维 混 杂 提 升 疲 劳 性 能<br />

由 于 相 当 一 部 分 的 土 建 交 通 基 础 设 施 处 于 循 环 荷 载 或 者 动 荷 载 作 用 下 , 如 桥 面 板 , 大 跨 桥 主 梁 ,<br />

路 面 , 拉 索 等 频 繁 的 受 到 交 通 荷 载 和 风 荷 载 的 作 用 , 因 而 利 用 纤 维 复 合 材 料 对 这 些 结 构 设 施 进 行 增<br />

强 或 者 建 造 , 其 耐 疲 劳 强 度 必 须 得 到 保 证 。 作 者 研 究 团 队 多 种 FRP 片 材 的 耐 疲 劳 性 能 进 行 了 长 期 研<br />

究 , 其 中 包 括 CFRP,PBO,GFRP 和 BFRP, 依 据 JSCE-E541 [20] 规 程 对 其 疲 劳 寿 命 进 行 了 统 计 和 分<br />

析 预 测 , 详 见 图 11 所 示 。 各 种 FRP 片 材 都 具 有 相 对 传 统 建 筑 材 料 优 秀 的 耐 疲 劳 性 能 (55% 最 大 应<br />

力 比 、 幅 值 R=0.1、 能 够 循 环 加 载 两 百 万 次 而 不 断 裂 ), 能 够 满 足 多 数 土 建 设 施 的 疲 劳 性 能 要 求 ; 其<br />

中 以 CFRP 和 PBO 的 耐 疲 劳 性 能 最 好 , 但 性 价 比 更 为 突 出 的 GFRP 和 BFRP 则 相 对 较 低 , 对 于 循 环<br />

荷 载 幅 值 较 大 的 结 构 , 在 一 定 程 度 上 也 限 制 了 两 者 的 应 用 面 。 为 此 , 结 合 纤 维 混 杂 设 计 方 法 , 研 究<br />

了 混 杂 纤 维 耐 疲 劳 性 能 , 包 括 混 杂 玄 武 岩 纤 维 - 碳 纤 维 和 混 杂 玻 璃 纤 维 - 碳 纤 维 。 试 验 结 果 显 示 ( 图<br />

12):(1) 通 过 玄 武 岩 - 碳 纤 维 混 杂 ( 比 例 1:1), 混 杂 纤 维 布 在 循 环 荷 载 下 的 稳 定 性 能 和 耐 疲 劳 性<br />

能 都 能 得 到 显 著 的 提 高 。 混 杂 后 的 玄 武 岩 / 碳 纤 维 复 合 材 料 在 70% 的 最 大 应 力 下 ( 幅 值 R=0.1), 能<br />

够 维 持 两 百 万 次 循 环 而 不 发 生 断 裂 , 相 对 于 原 先 玄 武 岩 纤 维 的 55% 疲 劳 强 度 得 到 了 大 幅 提 升 。 这 一<br />

特 征 使 得 这 种 混 杂 纤 维 布 能 够 在 循 环 荷 载 作 用 下 的 结 构 中 发 挥 更 高 的 效 率 ;(4) 在 循 环 荷 载 作 用 下 ,<br />

玄 武 岩 纤 维 和 碳 纤 维 能 够 粘 结 良 好 , 使 得 混 杂 效 应 突 出 。 相 比 较 而 言 , 玻 璃 纤 维 和 碳 纤 维 的 混 杂 效<br />

应 较 不 明 显 , 混 杂 纤 维 的 疲 劳 强 度 也 未 能 得 到 提 高 , 如 图 12 所 示 , 其 中 影 响 因 素 还 有 待 进 一 步 研<br />

究 。<br />

110%<br />

CFRP PBO GFRP BFRP<br />

110%<br />

CFRP GFRP BFRP C1G2 C1B1<br />

100%<br />

100%<br />

ρ=σmax/σavg<br />

90%<br />

80%<br />

70%<br />

ρ=Pmax/Pavg<br />

90%<br />

80%<br />

70%<br />

60%<br />

60%<br />

50%<br />

50%<br />

40%<br />

-0.5 0.5 1.5 2.5 3.5 4.5 5.5 6.5<br />

LogN<br />

图 11 S-N 疲 劳 寿 命 (FRP)<br />

40%<br />

-0.5 0.5 1.5 2.5 3.5 4.5 5.5 6.5<br />

LogN<br />

图 12 S-N 疲 劳 寿 命 ( 混 杂 FRP)<br />

(2) 纤 维 混 杂 提 升 冻 融 循 环 作 用 下 的 材 料 性 能<br />

[21]<br />

为 了 满 足 在 严 酷 环 境 下 ( 如 中 国 北 方 , 西 北 地 区 以 及 临 海 地 区 ), 结 构 的 使 用 性 能 和 安 全 性 能 ,<br />

对 FRP 材 料 在 冻 融 循 环 作 用 下 的 性 能 进 行 了 研 究 。 试 验 依 据 《 普 通 混 凝 土 长 期 性 能 和 耐 久 性 能 试 验<br />

方 法 标 准 (GB-T50082-2009)》 和 《 碳 纤 维 片 材 单 向 拉 伸 试 验 标 准 (GB/T 3354-1999)》 对 1B1C、<br />

1C2B 四 种 FRP 片 材 进 行 了 冻 融 循 环 试 验 和 单 调 拉 伸 试 验 。 每 组 试 件 包 含 7 个 试 件 , 试 验 结 果 取 其<br />

中 的 5 个 有 效 数 据 。 图 13 示 出 了 经 历 不 同 冻 融 循 环 次 数 作 用 后 的 FRP 片 材 力 学 性 能 的 平 均 值 。 总<br />

体 而 言 , 单 一 纤 维 组 成 的 FRP 片 材 经 历 了 冻 融 循 环 作 用 后 性 能 略 有 降 低 ,BFRP 片 材 抗 冻 性 能 好 于<br />

CFRP, 经 过 混 杂 配 置 ,FRP 片 材 的 抗 冻 性 能 基 本 未 有 降 低 , 整 体 上 好 于 单 一 纤 维 片 材 。1C1B 在 冻<br />

融 循 环 作 用 后 三 项 拉 伸 性 能 指 标 均 高 于 未 冻 融 的 对 比 试 件 ,1C2B 试 件 在 拉 伸 强 度 和 极 限 应 变 与 1B<br />

片 材 相 类 似 , 弹 性 模 量 指 标 较 两 种 单 一 纤 维 片 材 的 增 长 幅 度 略 低 , 但 相 对 值 仍 保 持 在 1 以 上 , 未 发<br />

-130-


生 退 化 。<br />

拉 伸 强 度 相 对 值<br />

1.15<br />

1.1<br />

1.05<br />

1<br />

0.95<br />

0.9<br />

0.85<br />

0.8<br />

0.75<br />

1B<br />

1C1B<br />

0 50 100 150 200 250<br />

冻 融 循 环 次 数<br />

极 限 应 变 相 对 值<br />

1.1<br />

1.05<br />

1<br />

0.95<br />

0.9<br />

0.85<br />

0.8<br />

0.75<br />

0.7<br />

1C<br />

1C2B<br />

1B<br />

1C1B<br />

弹 性 模 量 相 对 值<br />

1.15<br />

1.1<br />

1.05<br />

1<br />

0.95<br />

0.9<br />

1B<br />

1C1B<br />

0 50 100 150 200 250<br />

冻 融 循 环 次 数<br />

0 50 100 150 200 250<br />

1C<br />

1C2B<br />

冻 融 循 环 次 数<br />

图 13 冻 融 循 环 作 用 下 各 FRP 片 材 力 学 性 能 指 标 平 均 值 比 较<br />

另 外 , 以 变 异 系 数 (CV) 为 指 标 讨 论 了 各 种 冻 融 循 环 次 数 作 用 后 同 组 FRP 片 材 试 件 力 学 性 能<br />

的 离 散 程 度 。 总 体 而 言 , 冻 融 循 环 作 用 后 各 组 试 件 的 离 散 程 度 均 有 变 大 的 趋 势 , 这 和 材 料 性 能 的 差<br />

异 和 冻 融 循 环 作 用 的 不 均 匀 性 有 一 定 的 关 系 。 单 一 纤 维 片 材 在 三 种 拉 伸 性 能 指 标 方 面 均 表 现 出 较 大<br />

的 离 散 程 度 , 混 杂 纤 维 大 部 分 数 据 点 的 变 异 系 数 数 值 均 小 于 单 一 纤 维 片 材 , 混 杂 纤 维 试 验 数 据 的 稳<br />

定 性 要 好 于 单 一 纤 维 片 材 。<br />

[22-23]<br />

(3) 纤 维 混 杂 提 升 耐 高 温 性 能<br />

突 发 灾 难 下 结 构 的 安 全 性 能 一 直 是 土 建 交 通 领 域 关 注 的 重 要 课 题 , 其 中 火 灾 就 是 一 种 较 常 见 并<br />

且 危 害 很 大 的 突 发 灾 难 。 因 而 , 对 于 火 灾 等 高 温 环 境 下 , 纤 维 材 料 的 力 学 性 能 评 价 尤 为 重 要 。 另 外 ,<br />

在 火 灾 作 用 下 , 纤 维 复 合 材 料 是 否 释 放 有 毒 气 体 , 也 是 我 们 选 择 FRP 材 料 所 需 关 注 的 。 对 于 FRP<br />

材 料 , 由 于 纤 维 本 身 具 有 较 好 的 耐 高 温 性 能 ( 约 300~700℃), 因 而 其 耐 高 温 性 能 很 大 程 度 上 决 定<br />

于 基 体 材 料 在 高 温 下 的 性 能 。 对 此 , 首 先 对 多 种 纤 维 布 和 普 通 环 氧 树 脂 基 体 复 合 形 成 的 FRP 片 材 进<br />

行 了 高 温 下 试 验 , 结 果 表 明 纤 维 布 在 200 度 的 高 温 下 , 虽 然 基 体 材 料 已 经 完 全 玻 璃 化 ( 玻 璃 化 温 度<br />

38 度 ), 但 FRP 仍 然 维 持 65% 左 右 的 抗 拉 强 度 ( 图 14), 明 显 高 于 干 丝 纤 维 布 的 强 度 (40-50% 左 右 ),<br />

这 一 残 余 强 度 对 于 保 证 FRP 增 强 结 构 的 安 全 性 具 有 一 定 的 作 用 。<br />

在 此 基 础 上 , 为 进 一 步 提 高 结 构 安 全 性 , 研 究 了 混 杂 玄 武 岩 纤 维 对 碳 纤 维 FRP 片 材 耐 高 温 性 能<br />

的 影 响 , 试 验 结 果 表 明 : 混 杂 玄 武 岩 纤 维 和 碳 纤 维 布 , 虽 然 不 能 提 高 碳 纤 维 布 在 高 温 下 的 抗 拉 强 度 ,<br />

但 能 够 降 低 复 合 纤 维 布 在 高 温 下 的 强 度 离 散 程 度 , 强 度 值 更 加 稳 定 。 因 此 , 使 得 FRP 增 强 结 构 在 高<br />

温 下 的 强 度 具 有 更 高 的 利 用 效 率 和 可 设 计 性 。<br />

根 据 上 述 研 究 成 果 , 为 了 保 证 在 复 合 材 料 中 发 挥 玄 武 岩 纤 维 的 耐 高 温 性 能 , 必 须 保 证 FRP 中 的<br />

基 体 材 料 也 具 备 理 想 的 耐 高 温 性 能 , 同 时 增 加 防 火 处 理 , 使 FRP 材 料 具 有 一 定 的 防 火 耐 高 温 性 能 ,<br />

保 证 与 结 构 的 共 同 工 作 。 目 前 正 在 进 行 多 种 耐 高 温 树 脂 材 料 基 本 性 能 的 研 究 , 并 且 正 在 建 立 FRP 在<br />

高 温 下 力 学 性 能 预 测 模 型 , 对 指 导 FRP 增 强 结 构 设 计 提 供 依 据 。 同 时 , 对 于 在 建 筑 室 内 结 构 中 采 用<br />

具 有 高 温 下 无 毒 无 烟 的 树 脂 研 究 正 在 开 展 , 研 究 成 果 为 推 动 FRP 防 火 耐 高 温 结 构 具 有 指 导 作 用 。<br />

1C<br />

1C2B<br />

-131-


图 14 混 杂 效 应 对 高 温 下 性 能 影 响<br />

3.4. 混 杂 FRP 增 强 结 构 性 能<br />

通 过 高 弹 性 模 量 、 高 强 度 和 高 延 性 纤 维 按 一 定 的 比 例 混 杂 , 从 而 实 现 混 杂 纤 维 复 合 材 料 整 体 高<br />

初 始 刚 度 , 高 强 度 和 充 足 的 延 性 。 利 用 这 种 混 杂 纤 维 复 合 材 料 增 强 混 凝 土 结 构 , 可 以 实 现 高 开 裂 后<br />

刚 度 、 高 屈 服 荷 载 、 高 承 载 力 和 充 足 的 延 性 。 如 图 15 显 示 了 混 杂 纤 维 加 固 混 凝 土 梁 的 设 计 概 念 ,<br />

通 过 高 碳 模 纤 维 作 用 , 提 高 开 裂 荷 载 和 屈 服 荷 载 , 提 升 结 构 的 使 用 性 能 ; 在 结 构 屈 服 后 , 伴 随 高 弹<br />

模 纤 维 的 逐 渐 断 裂 , 应 力 逐 渐 释 放 , 避 免 界 面 剥 离 的 发 生 , 并 在 高 强 度 纤 维 的 配 合 作 用 下 , 结 合 强<br />

度 得 到 维 持 ; 之 后 , 随 着 高 强 度 纤 维 的 逐 渐 断 裂 , 高 延 性 纤 维 发 挥 作 用 , 结 构 延 性 得 到 保 证 [14] 。<br />

高 度 的 承 载 能 力<br />

高 度 的 承 载 能 力<br />

图 15 混 杂 纤 维 复 合 材 料 增 强 混 凝 土 结 构 荷 载 - 变 形 图<br />

根 据 上 述 设 计 理 念 , 首 先 试 验 研 究 了 高 强 度 和 高 弹 模 的 混 杂 碳 纤 维 的 力 学 性 能 和 试 验 加 固 效 果<br />

[24-25] 。 相 对 于 普 通 高 强 碳 纤 维 , 该 混 杂 纤 维 具 有 更 高 的 屈 服 强 度 和 弹 性 模 量 , 并 且 能 够 保 持 良 好 的<br />

屈 服 后 刚 度 。 利 用 该 混 杂 纤 维 对 预 加 载 试 件 进 行 了 受 弯 加 固 的 试 验 , 分 析 了 不 同 的 纤 维 层 数 和 混 杂<br />

比 例 的 影 响 。 结 果 表 明 预 加 载 的 试 验 梁 在 混 杂 纤 维 增 强 后 能 够 恢 复 到 原 先 的 承 载 能 力 , 同 时 通 过 纤<br />

维 混 杂 比 例 的 优 化 分 析 可 以 更 加 有 效 的 提 高 构 件 的 承 载 能 力 以 及 控 制 荷 载 的 下 降 过 程 。 如 图 16 所<br />

示 ,C1 表 示 高 强 纤 维 ,C7 表 示 高 弹 模 纤 维 ,C 前 数 字 代 表 纤 维 层 数 。 由 图 可 见 对 预 加 载 混 凝 土 梁<br />

同 样 能 够 取 得 更 高 的 刚 度 和 屈 服 荷 载 。 图 16 中 增 强 梁 屈 服 后 齿 状 曲 线 表 明 高 弹 模 纤 维 的 连 续 断 裂 ,<br />

界 面 应 力 不 断 进 行 重 新 分 布 , 缓 解 了 局 部 应 力 集 中 问 题 , 与 单 纯 采 用 高 强 度 纤 维 相 比 , 具 有 更 好 的<br />

延 性 和 更 高 的 刚 度 。<br />

-132-


图 16 两 种 纤 维 混 杂 增 强 梁 荷 载 - 挠 度 曲 线 图<br />

图 17 三 种 纤 维 混 杂 增 强 梁 荷 载 - 挠 度 曲 线<br />

在 两 种 纤 维 复 合 的 基 础 上 进 一 步 研 究 了 三 种 纤 维 的 混 杂 [26] : 高 强 度 , 高 弹 模 和 高 延 性 纤 维 。 这<br />

种 混 杂 纤 维 除 了 能 得 到 更 高 的 弹 性 模 量 和 屈 服 强 度 外 , 还 能 够 取 得 了 更 好 的 延 性 。 试 验 结 果 如 图 17<br />

所 示 , 图 中 高 弹 模 (HM), 高 强 (HS) 和 高 延 性 (HD) 纤 维 按 不 同 比 例 进 行 混 杂 , 其 增 强 效 果 和<br />

控 制 梁 相 比 刚 度 , 强 度 得 到 了 很 大 的 提 高 , 并 且 取 得 很 好 的 延 性 。 与 单 纯 采 用 高 强 碳 纤 维 3C1 增 强<br />

的 混 凝 土 梁 , 如 果 不 采 用 U 形 锚 固 措 施 , 延 性 提 高 不 大 , 当 采 用 U 形 锚 固 后 , 延 性 提 高 了 3 倍 以 上 。<br />

可 见 混 杂 纤 维 具 备 良 好 的 延 性 储 备 , 在 保 证 端 部 锚 固 的 条 件 下 高 延 性 纤 维 的 优 势 得 到 了 充 分 发 挥 。<br />

对 于 纤 维 和 钢 筋 复 合 的 SFCB 筋 主 要 用 于 结 构 柱 的 抗 震 设 计 , 试 验 增 强 效 果 将 在 下 文 FRP 提 升<br />

结 构 抗 震 性 能 部 分 详 述 。<br />

利 用 玄 武 岩 纤 维 - 钢 丝 混 杂 纤 维 布 结 合 预 应 力 张 拉 方 法 增 强 混 凝 土 梁 , 试 验 结 果 表 明 梁 的 开 裂 荷<br />

载 , 屈 服 荷 载 都 有 明 显 提 高 , 纤 维 和 钢 丝 的 材 料 性 能 得 到 更 充 分 的 发 挥 , 结 构 的 综 合 性 能 得 到 很 大<br />

的 改 善 , 具 体 如 图 18 所 示 [18] 。<br />

图 18 玄 武 岩 纤 维 - 钢 丝 混 杂 布 预 应 力 增 强 梁<br />

四 、 预 应 力 FRP 技 术<br />

如 上 文 所 述 , 由 于 FRP 材 料 的 弹 性 模 量 与 钢 材 相 当 或 偏 低 , 在 和 既 有 结 构 共 同 工 作 后 , 结 构 的<br />

使 用 性 能 提 升 不 明 显 ,FRP 材 料 仅 在 结 构 钢 筋 屈 服 后 才 能 发 挥 作 用 。 因 此 , 要 提 高 FRP 材 料 的 使 用<br />

效 率 , 提 升 FRP 材 料 对 结 构 使 用 性 能 的 增 强 效 果 , 一 方 面 我 们 可 以 采 用 高 弹 性 模 量 的 纤 维 和 普 通 纤<br />

维 混 杂 实 现 对 结 构 使 用 性 能 的 明 显 改 善 , 另 一 方 面 可 以 采 用 预 应 力 技 术 和 FRP 材 料 相 结 合 , 从 工 法<br />

技 术 上 提 升 单 一 FRP 加 固 结 构 的 使 用 性 能 。 针 对 结 构 加 固 常 用 的 片 材 和 筋 材 两 种 形 式 , 下 文 介 绍 外<br />

粘 结 预 应 力 FRP 片 材 加 固 和 嵌 入 式 FRP 索 材 加 固 两 种 方 法 。<br />

4.1. 外 粘 结 预 应 力 FRP 片 材 加 固 技 术<br />

(1) 原 理 及 设 计 理 念<br />

对 FRP 材 料 施 加 预 应 力 以 发 挥 其 高 强 度 , 即 对 FRP 预 先 张 拉 , 然 后 利 用 粘 结 材 料 粘 贴 于 混 凝<br />

土 表 面 共 同 受 力 , 这 样 在 提 高 结 构 强 度 的 同 时 还 能 有 效 改 善 结 构 的 使 用 性 能 , 充 分 发 挥 FRP 纤 维 材<br />

料 的 高 强 特 性 。 然 而 在 实 际 操 作 中 , 由 于 普 通 纤 维 材 料 抗 冲 击 性 能 差 , 必 须 先 用 树 脂 含 浸 硬 化 成 板<br />

-133-


材 后 才 能 作 为 预 应 力 材 料 张 拉 , 这 样 不 仅 费 时 费 力 而 且 往 往 不 能 和 混 凝 土 表 面 取 得 良 好 的 粘 结 , 极<br />

大 了 影 响 了 预 应 力 FRP 加 固 结 构 的 效 率 和 效 果 [27] 。 为 此 , 作 者 研 究 团 队 根 据 纤 维 材 料 的 耐 冲 击 性 能<br />

研 究 , 提 出 了 两 种 可 用 于 结 构 预 应 力 加 固 的 方 法 : 一 是 直 接 张 拉 具 有 高 能 量 吸 收 能 力 的 PBO 纤 维 布 ;<br />

二 是 采 用 混 杂 纤 维 直 接 张 拉 。<br />

4<br />

Load(kN)<br />

3<br />

2<br />

1<br />

PBO fiber<br />

Carbon fiber<br />

Aramid fiber<br />

0<br />

0 5 10 15<br />

Displacement(mm)<br />

图 19 多 种 纤 维 的 抗 冲 击 试 验<br />

图 20 混 杂 提 升 张 拉 率<br />

PBO 纤 维 具 有 远 高 于 碳 纤 维 和 芳 纶 纤 维 的 能 量 吸 收 能 力 , 如 图 19 所 示 。 在 一 个 纤 维 发 生 断 裂<br />

时 , 周 围 的 纤 维 有 足 够 的 抗 冲 击 能 力 吸 收 断 裂 纤 维 产 生 的 冲 击 , 避 免 纤 维 发 生 连 续 性 断 裂 。 因 此 ,<br />

利 用 PBO 纤 维 干 丝 张 拉 对 结 构 进 行 加 固 是 可 行 的 [28] 。 另 外 , 研 究 表 明 为 了 提 高 纤 维 干 丝 的 可 张 拉<br />

性 能 , 通 过 混 杂 多 种 纤 维 也 是 一 个 很 好 的 方 法 。 图 20 显 示 通 过 混 杂 玄 武 岩 纤 维 或 玻 璃 纤 维 , 碳 纤<br />

维 布 的 张 拉 率 能 从 干 丝 的 30% 左 右 提 升 到 54%~65% 以 上 , 张 拉 性 能 大 幅 提 升 [16] 。<br />

利 用 预 应 力 纤 维 片 材 增 强 结 构 的 设 计 理 念 如 图 21 所 示 , 通 过 预 应 力 纤 维 片 材 , 混 凝 土 梁 的 开<br />

裂 荷 载 , 屈 服 荷 载 和 极 限 荷 载 都 得 到 明 显 提 高 , 同 时 初 期 刚 度 也 能 显 著 提 升 。 因 此 , 达 到 了 通 过 施<br />

加 预 应 力 提 升 结 构 使 用 性 能 的 目 标 , 纤 维 材 料 本 身 的 利 用 效 率 也 得 到 了 极 大 的 提 升 [27] 。<br />

RC beam strengthened with<br />

prestressed FRP sheets<br />

FRP rupture<br />

P<br />

Limited reinforcement effect<br />

(Load)<br />

Initiation of Debonding FRP debonding failure<br />

Great enhancement of<br />

of FRP Sheets<br />

steel yielding load<br />

P y ’’<br />

P y ’<br />

P y<br />

P cr ’’<br />

P cr<br />

RC beam strengthened with FRP sheets<br />

Normal RC beam<br />

Great enhancement of stiffness<br />

Limited reinforcement effect<br />

σ cr<br />

(Cracking)<br />

σ y<br />

(Steel yielding)<br />

σ<br />

(Displacement)<br />

图 21 预 应 力 纤 维 片 材 增 强 效 果 图<br />

(2) 工 艺 技 术 难 点<br />

张 拉 原 理 见 图 22 所 示 , 纤 维 张 拉 后 通 过 粘 结 材 料 与 结 构 结 合 , 保 证 了 界 面 粘 结 的 可 靠 。 同 时<br />

为 解 决 端 部 锚 固 区 剪 应 力 集 中 问 题 , 采 用 逐 步 减 少 纤 维 层 数 或 者 分 部 释 放 预 应 力 的 方 法 减 缓 了 应 力<br />

集 中 ( 图 23), 并 增 加 U 形 FRP 布 以 取 得 更 好 的 锚 固 效 果 。 实 际 操 作 中 在 端 部 PBO 纤 维 层 间 插 入<br />

不 粘 结 薄 片 以 达 到 逐 层 减 少 层 数 的 目 的 。 在 此 基 础 上 开 发 了 一 套 完 整 的 操 作 系 统 (P-PUT), 包 括 一<br />

整 套 张 拉 , 养 护 , 辅 助 设 备 , 保 证 了 纤 维 材 料 和 混 凝 土 的 粘 结 质 量 和 “ 一 天 内 完 成 快 速 施 工 ” 的 高 效<br />

目 标 [29] 。<br />

-134-


σ tf<br />

混 凝 土 构 件<br />

粘 结 材 料 ( 环 氧 )<br />

σ tf : 预 应 力<br />

τf<br />

剪 应 力<br />

(1) FRP 布 端 部 剪 应 力 集 中 分 布<br />

1. 预 张 拉 不 含 浸 纤 维 布<br />

Pre-bonded FRP sheets<br />

σ tf<br />

σ tf<br />

2. 粘 结 和 养 护<br />

τf1 τf2 τf3<br />

(2) 逐 层 减 少 以 释 放 应 力 集 中<br />

1) 粘 结 1st 部 分 后 释 放<br />

3. 剪 断 释 放 端 部 纤 维 布<br />

-σ c<br />

截 面 应 力 分 布<br />

σf1<br />

3rd 2nd<br />

σf2<br />

σf3<br />

1st<br />

σf3<br />

2) 粘 结 2nd 部 分 后 释 放<br />

σf2<br />

3) 粘 结 3rd 部 分 后 释 放<br />

(3) 分 部 释 放 减 少 应 力 集 中<br />

σf1<br />

图 22 预 应 力 纤 维 外 贴 加 固 原 理 图<br />

图 23 预 应 力 张 拉 端 部 纤 维 锚 固 方 法<br />

(3) 加 固 效 果<br />

利 用 预 应 力 张 拉 技 术 对 PBO 和 混 杂 纤 维 张 拉 后 进 行 结 构 加 固 , 增 强 效 果 如 图 24,25 所 示 , 预<br />

应 力 效 果 得 到 充 分 的 发 挥 , 开 裂 荷 载 , 屈 服 荷 载 都 得 到 显 著 提 高 , 保 证 了 结 构 在 使 用 状 态 下 的 力 学<br />

性 能 。 同 时 , 由 于 锚 固 措 施 有 效 , 结 构 破 坏 时 具 有 一 定 的 延 性 [30] 。<br />

Load(kN)<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

Rupture of FRP sheets near the mid-span<br />

Debonding of fiber sheets near the mid-span<br />

Crack initiation<br />

0 50 100<br />

Displacement (mm)<br />

图 24 10 m T 形 梁 加 固 效 果<br />

PC girder with 3 lay ers of fiber<br />

sheets (45% prestress)<br />

PC girder with 3 lay ers of fiber<br />

sheets (33% prestress)<br />

PC girder with 3 lay ers of fiber<br />

sheets (no prestress)<br />

PC girder (no fiber reinforce)<br />

Load(kN)<br />

With 33%-prestressed 1layer Basalt and 3layers carbon fiber sheets<br />

100<br />

With non-prestressed<br />

80<br />

1layer Basalt and 3layers<br />

60<br />

carbon fiber sheets<br />

40<br />

20<br />

0<br />

Without fiber sheets<br />

0 5 10 15 20 25<br />

Displacement (mm)<br />

图 25 混 杂 纤 维 预 应 力 张 拉 加 固<br />

4.2 嵌 入 式 预 应 力 FRP 索 加 固 技 术<br />

在 FRP 预 应 力 粘 结 加 固 技 术 中 , 考 虑 到 提 高 FRP 与 混 凝 土 界 面 的 粘 结 性 能 以 及 避 免 外 界 对 FRP<br />

的 机 械 损 伤 , 作 者 研 究 团 队 采 用 了 预 应 力 CFRP 绞 线 嵌 入 式 增 强 混 凝 土 梁 受 弯 承 载 力 [31] , 如 图 26<br />

所 示 。 该 方 法 能 够 全 面 提 升 了 构 件 的 开 裂 荷 载 , 抗 裂 性 能 , 刚 度 , 屈 服 强 度 和 极 限 强 度 。 在 预 应 力<br />

嵌 入 式 增 强 方 法 中 为 解 决 实 际 工 程 中 嵌 入 深 度 不 够 导 致 粘 结 不 充 分 的 问 题 , 还 研 究 了 混 凝 土 槽 外 表<br />

面 增 加 覆 盖 层 , 以 提 高 FRP 筋 和 混 凝 土 粘 结 效 果 的 方 法 。 对 粘 结 材 料 和 覆 盖 层 材 料 , 如 环 氧 树 脂 和<br />

水 泥 砂 浆 , 进 行 了 比 较 分 析 , 试 验 结 果 表 明 采 用 环 氧 树 脂 和 水 泥 砂 浆 分 别 作 为 粘 结 材 料 和 覆 盖 层 材<br />

料 能 够 获 得 最 优 化 的 构 件 强 度 , 刚 度 和 延 性 , 如 图 27 所 示 。<br />

槽 深 小 于 保 护 层 厚 度<br />

端 部 锚 固 端 附 近<br />

应 力 集 中 消 去 措 施<br />

混 凝 土 梁<br />

中 部 注 入 高 刚 度 树 脂<br />

张 拉<br />

连 续 纤 维 布 或 柔 性 索<br />

CFRP 筋<br />

张 拉 装 置 与 CFRP 索 的 连 接 装 置<br />

连 续 纤 维 布 或 柔 性 索<br />

图 26 嵌 入 式 预 应 力 CFRP 筋 加 固 结 构 图<br />

张 拉<br />

-135-


图 27 嵌 入 式 加 固 外 覆 盖 层 示 意 图 和 张 拉 装 置 示 意<br />

试 验 结 果 表 明 采 用 环 氧 树 脂 和 水 泥 砂 浆 分 别 作 为 粘 结 材 料 和 覆 盖 层 材 料 能 够 获 得 最 优 化 的 构<br />

件 强 度 , 刚 度 和 延 性 , 如 图 28 所 示 。<br />

五 、 损 伤 可 控 FRP 抗 震 设 计 理 论 及 方 法<br />

5.1. 新 型 抗 震 设 计 思 想 的 提 出<br />

图 28 嵌 入 式 FRP 索 预 应 力 加 固 效 果<br />

由 于 目 前 抗 震 设 计 思 想 的 局 限 ,“ 中 震 可 修 ” 未 具 体 量 化 , 导 致 大 量 结 构 在 中 大 震 作 用 下 发 生 过<br />

大 的 不 可 恢 复 变 形 , 结 构 虽 然 不 倒 , 但 也 无 法 继 续 使 用 及 修 复 。 以 往 研 究 表 明 , 对 既 有 结 构 , 采 用<br />

FRP 布 约 束 或 嵌 入 式 加 固 的 方 法 , 可 以 明 显 改 善 传 统 RC 结 构 在 地 震 作 用 下 的 性 能 , 使 结 构 在 屈 服<br />

后 具 有 明 显 的 二 次 刚 度 和 较 小 的 残 余 变 形 。 但 对 于 新 建 结 构 , 如 何 发 挥 FRP 筋 或 SFCB 筋 实 现 结 构<br />

在 地 震 作 用 下 损 伤 可 控 可 修 复 , 仍 是 目 前 急 需 深 入 研 究 和 解 决 的 问 题 。 作 者 研 究 团 队 从 二 次 刚 度 和<br />

[32]<br />

残 余 位 移 两 个 指 标 入 手 , 在 日 本 抗 震 规 范 的 荷 载 - 位 移 曲 线 基 础 上 , 提 出 理 想 抗 震 结 构 的 设 计 理 念 ,<br />

如 下 图 29 所 示 。<br />

荷 使 用 阶 段 可 修 复 阶 段 极 限 阶 段<br />

载 中 、 小 震<br />

无 需 修 复<br />

中 、 大 、 特 大 震 可 修 不 倒 塌<br />

损 伤 可 控 阶 段<br />

E<br />

V y<br />

P<br />

Q<br />

损<br />

伤<br />

可<br />

控<br />

结<br />

构<br />

极 限 点<br />

( 不 倒 塌 )<br />

O<br />

传 统 钢 筋 混 凝 土 结 构<br />

开 裂<br />

残 余 位 移 < h/100 ( 可 修 复 限 值 ) 位 移<br />

(a) 不 同 二 次 刚 度 的 混 凝 土 结 构 残 余 位 移 示 意 (b) 理 想 抗 震 结 构 的 荷 载 - 位 移 关 系<br />

图 29 新 型 抗 震 结 构 的 荷 载 - 位 移 关 系<br />

-136-


新 型 抗 震 结 构 的 荷 载 - 位 移 曲 线 可 以 分 为 四 个 主 要 阶 段 。 阶 段 1, 为 OB 段 ; 阶 段 2, 为 BC 段 ;<br />

阶 段 3, 为 CD 段 ; 阶 段 4, 为 DE 段 。 阶 段 1 代 表 了 结 构 整 体 屈 服 之 前 , 对 于 应 结 构 的 弹 性 阶 段 和<br />

开 裂 后 的 状 态 。 小 震 作 用 下 的 结 构 响 应 在 阶 段 1 以 内 , 保 持 体 系 的 弹 性 阶 段 , 结 构 或 构 件 在 相 应 的<br />

地 震 烈 度 作 用 后 无 需 修 复 。 图 29b 的 阶 段 2 代 表 了 结 构 在 中 等 地 震 作 用 下 , 增 强 纵 筋 进 入 屈 服 阶 段 ,<br />

但 具 有 明 显 的 硬 化 特 征 , 即 稳 定 的 二 次 刚 度 , 结 构 变 形 可 以 得 到 有 效 控 制 , 并 且 震 后 损 伤 ( 中 震 、<br />

大 震 ) 可 以 得 到 快 速 修 复 。 阶 段 3 相 应 于 结 构 体 系 在 二 次 刚 度 段 之 后 的 变 形 能 力 , 从 而 使 结 构 在 大<br />

震 作 用 下 具 有 足 够 的 延 性 , 避 免 结 构 由 于 过 高 的 二 次 刚 度 而 产 生 太 大 的 地 震 响 应 , 结 构 震 后 可 以 通<br />

过 替 换 部 分 单 元 进 行 修 复 。 而 在 罕 遇 的 特 大 震 作 用 下 , 结 构 可 能 进 入 第 4 阶 段 , 这 一 阶 段 的 结 构 应<br />

能 够 避 免 倒 塌 , 当 荷 载 下 降 至 极 限 荷 载 的 20% 时 , 定 义 为 极 限 阶 段 , 并 能 够 保 持 结 构 不 倒 , 这 和 传<br />

统 结 构 的 定 义 相 同 。<br />

值 得 注 意 的 是 , 新 型 抗 震 体 系 由 于 可 以 具 有 更 小 的 卸 载 刚 度 , 进 一 步 减 小 残 余 位 移 , 实 现 较 高<br />

的 可 修 复 性 。<br />

5.2. 利 用 钢 - 连 续 纤 维 复 合 筋 实 现 损 伤 可 控 型 抗 震 混 凝 土 结 构<br />

在 提 出 理 想 抗 震 结 构 设 计 理 念 的 基 础 上 , 通 过 FRP 材 料 如 何 实 现 设 计 理 念 是 研 究 的 重 点 。 作 者<br />

研 究 团 队 已 有 的 研 究 表 明 , 通 过 FRP 约 束 抗 震 加 固 的 RC 柱 , 可 以 实 现 一 定 程 度 的 二 次 刚 度 [33-37] ,<br />

然 而 由 于 通 过 FRP 约 束 实 现 的 RC 柱 的 二 次 刚 度 提 高 并 不 明 显 , 而 且 由 于 尺 寸 效 应 、 方 柱 和 圆 柱 的<br />

不 同 处 理 方 式 等 因 素 使 得 该 途 径 实 现 的 结 构 二 次 刚 度 具 有 不 可 设 计 性 。 为 此 , 作 者 研 究 团 队 提 出 了<br />

对 于 既 有 结 构 采 用 柱 脚 嵌 入 式 FRP 加 固 混 凝 土 桥 墩 的 技 术 和 针 对 新 建 结 构 利 用 SFCB 筋 实 现 高 抗 震<br />

性 能 的 方 法 。<br />

(1) 柱 脚 嵌 入 FRP 筋 加 固 混 凝 土 桥 墩<br />

通 过 FRP 约 束 抗 震 加 固 可 以 成 功 的 保 证 既 有 混 凝 土 桥 墩 柱 在 强 震 作 用 下 的 安 全 性 , 但 震 后 残 余<br />

位 移 损 伤 的 有 效 控 制 依 然 需 要 进 一 步 的 提 高 。 因 此 , 作 者 研 究 团 队 提 出 了 新 型 抗 震 加 固 方 法 , 即 在<br />

柱 脚 塑 性 铰 区 域 嵌 入 BFRP 筋 并 根 据 具 体 情 况 决 定 是 否 采 用 外 包 FRP 进 行 组 合 加 固 ( 图 30) [38] ,<br />

可 以 有 效 的 控 制 混 凝 土 柱 的 损 伤 水 平 、 残 余 变 形 和 二 次 刚 度 。 值 得 注 意 的 是 , 在 适 用 该 加 固 方 法 时 ,<br />

应 避 免 发 生 类 似 于 在 阪 神 地 震 中 部 分 桥 墩 由 于 纵 筋 切 断 而 引 起 的 剪 切 破 坏 , 这 要 求 柱 脚 嵌 入 BFRP<br />

筋 的 加 固 长 度 应 限 制 在 合 理 长 度 内 以 避 免 改 变 柱 子 的 破 坏 模 式 。<br />

图 30 柱 脚 嵌 入 FRP 筋 加 固 设 计<br />

图 31 嵌 入 BFRP 筋 加 固 混 凝 土 柱 滞 回 曲 线<br />

在 进 行 BFRP 筋 加 固 混 凝 土 柱 之 前 , 进 行 了 模 拟 加 固 方 式 的 BFRP 筋 与 混 凝 土 的 粘 结 性 能 试 验 ,<br />

结 果 表 明 随 着 粘 结 长 度 的 增 加 , 试 验 破 坏 时 的 BFRP 筋 应 力 呈 增 加 趋 势 , 最 终 破 坏 模 式 从 拔 出 破 坏<br />

(20 倍 直 径 粘 结 长 度 ) 转 变 为 BFRP 筋 拉 断 破 坏 (40 倍 直 径 ), 因 此 文 献 [38] 推 荐 60 倍 直 径 的 锚 固<br />

长 度 可 以 确 保 嵌 入 的 BFRP 筋 的 强 度 得 到 充 分 发 挥 。<br />

柱 脚 嵌 入 BFRP 加 固 前 后 的 RC 柱 荷 载 - 位 移 曲 线 如 图 31 所 示 , 其 中 包 括 了 普 通 钢 筋 混 凝 土 的<br />

滞 回 曲 线 。 可 以 看 出 , 由 于 柱 脚 嵌 入 BFRP 筋 , 并 没 有 改 变 混 凝 土 柱 的 弹 性 刚 度 , 而 可 以 实 现 和 SFCB<br />

增 强 混 凝 土 柱 类 似 的 稳 定 二 次 刚 度 , 加 固 柱 由 于 BFRP 筋 的 拔 出 而 发 生 破 坏 , 相 应 的 柱 顶 侧 移 为<br />

30mm。 加 固 柱 相 对 于 普 通 RC 柱 , 可 以 显 著 减 小 卸 载 残 余 位 移 , 具 有 较 高 的 可 修 复 性 的 效 果 。<br />

-137-


(2) 利 用 钢 - 连 续 纤 维 复 合 筋 (SFCB) 实 现 损 伤 可 控 型 新 建 混 凝 土 结 构<br />

[39]<br />

对 于 新 建 结 构 , 作 者 研 究 团 队 提 出 以 钢 - 连 续 纤 维 复 合 筋 (SFCB) 增 强 混 凝 土 抗 震 结 构 ( 图<br />

32), 利 用 钢 筋 和 FRP 材 料 的 优 势 互 补 , 在 实 现 复 合 筋 材 性 稳 定 二 次 刚 度 的 基 础 上 , 实 现 SFCB 增<br />

强 结 构 的 稳 定 二 次 刚 度 , 并 给 出 了 SFCB 关 键 构 造 工 艺 , 实 现 了 工 业 化 批 量 生 产 。SFCB 增 强 结 构<br />

其 特 点 包 括 :1 在 正 常 使 用 荷 载 或 中 小 地 震 作 用 下 , 具 有 与 普 通 钢 筋 混 凝 土 结 构 相 同 的 强 度 抵 抗 能<br />

力 , 可 以 充 分 发 挥 SFCB 内 芯 钢 筋 带 来 的 高 弹 性 模 量 作 用 ;2 利 用 外 包 FRP 的 高 强 度 特 性 使 SFCB<br />

增 强 的 结 构 具 有 截 面 层 次 上 稳 定 的 二 次 刚 度 , 这 一 特 征 可 以 预 防 塑 性 铰 在 柱 脚 小 范 围 内 集 中 转 动 形<br />

成 的 过 大 的 塑 性 变 形 , 实 现 在 一 个 更 长 的 区 域 内 实 现 曲 率 的 较 均 匀 分 布 , 减 小 截 面 的 需 求 曲 率 , 因<br />

而 相 应 的 减 小 SFCB 中 内 芯 钢 筋 塑 性 的 应 变 ;3 用 SFCB 代 替 普 通 钢 筋 , 不 改 变 截 面 的 初 始 屈 服 强<br />

度 , 有 利 于 结 构 抗 震 设 计 时 控 制 损 伤 截 面 的 位 置 , 强 地 震 荷 载 作 用 下 , 可 以 控 制 损 伤 截 面 的 位 置 对<br />

结 构 的 变 形 能 力 具 有 重 要 的 影 响 。<br />

图 32 利 用 SFCB 或 普 通 钢 筋 和 FRP 筋 混 杂 配 筋 实 现 混 凝 土 结 构 的 稳 定 二 次 刚 度<br />

利 用 SFCB 所 增 强 的 混 凝 土 结 构 具 有 稳 定 的 二 次 刚 度 , 这 同 样 可 以 通 过 钢 筋 和 FRP 筋 的 混 杂 配<br />

筋 来 实 现 , 然 而 SFCB 由 于 外 侧 全 部 为 FRP, 具 有 高 度 的 耐 久 性 特 征 , 可 以 适 用 于 各 种 恶 劣 环 境 下 。<br />

作 者 研 究 团 队 进 行 了 SFCB 单 向 和 往 复 拉 伸 试 验 [40] [41]<br />

、SFCB 与 混 凝 土 粘 结 性 能 试 验 和 SFCB<br />

增 强 混 凝 土 柱 抗 震 性 能 的 试 验 研 究 [42] 。 粘 结 试 验 表 明 ,SFCB 与 混 凝 土 的 粘 结 强 度 约 为 相 应 带 肋 钢<br />

筋 的 94%; 特 别 的 , 拉 拔 试 件 的 破 坏 形 态 可 分 为 钢 筋 内 芯 屈 服 后 SFCB 拔 出 、 钢 筋 内 芯 屈 服 前 SFCB<br />

拔 出 和 SFCB 拉 断 3 种 。 在 SFCB 粘 结 试 件 自 由 端 发 生 轻 微 滑 移 后 , 复 合 筋 的 钢 / 纤 维 种 类 和 比 例 、<br />

试 件 有 效 粘 结 长 度 和 混 凝 土 强 度 等 级 等 因 素 决 定 了 在 发 生 SFCB 拔 出 或 拉 断 破 坏 , 以 及 SFCB 拔 出<br />

之 前 钢 筋 内 芯 是 否 屈 服 。<br />

对 于 SFCB 增 强 混 凝 土 抗 震 结 构 , 在 大 震 作 用 下 应 能 够 避 免 FRP 的 断 裂 , 保 证 结 构 震 后 残 余 位 移<br />

在 可 修 复 范 围 ( 图 29b 中 OC 段 )。 在 特 大 震 作 用 下 , 为 了 避 免 由 于 FRP 断 裂 引 起 的 倒 塌 ,SFCB 与 混<br />

凝 土 之 间 应 该 能 够 进 行 滑 移 并 保 持 承 载 力 水 平 , 如 图 29b 中 SFCB 荷 载 - 滑 移 曲 线 峰 值 点 后 的 虚 线 所<br />

示 。 具 体 如 何 实 现 SFCB 与 混 凝 土 之 间 的 这 种 具 有 高 度 韧 性 的 粘 结 - 滑 移 关 系 , 有 待 于 进 一 步 的 研 究 。<br />

为 了 研 究 钢 - 连 续 纤 维 复 合 筋 (SFCB) 增 强 混 凝 土 柱 的 抗 震 性 能 , 共 制 作 了 截 面 为 300×300mm<br />

的 4 个 SFCB 增 强 混 凝 土 柱 和 1 个 RC 对 比 柱 , 试 件 设 计 参 数 可 见 文 献 [42] , 试 验 结 果 表 明 , 所 有 柱 试 件<br />

均 发 生 弯 曲 破 坏 , 表 现 为 先 是 柱 脚 附 近 混 凝 土 开 裂 , 纵 筋 或 SFCB 内 芯 钢 筋 屈 服 ,SFCB 外 包 纤 维 部<br />

分 ( 或 全 部 ) 断 裂 , 继 续 加 载 , 柱 脚 混 凝 土 压 碎 并 剥 落 , 纵 筋 压 弯 屈 曲 。<br />

SFCB 柱 和 RC 对 比 柱 荷 载 - 位 移 滞 回 曲 线 (V-δ 曲 线 ) 如 图 33 所 示 [42] 。 从 按 力 加 载 阶 段 SFCB 增 强<br />

混 凝 土 柱 和 RC 柱 荷 载 - 位 移 曲 线 ( 图 33a) 比 较 可 以 看 出 :SFCB 增 强 混 凝 土 柱 由 于 所 用 复 合 筋 轴 向<br />

刚 度 小 于 柱 C-S14 所 用 钢 筋 , 在 混 凝 土 开 裂 后 ,SFCB 增 强 混 凝 土 柱 刚 度 比 柱 C-S14 柱 小 。 普 通 RC 柱<br />

滞 回 曲 线 呈 梭 形 , 较 丰 满 , 表 现 出 良 好 的 延 性 和 耗 能 能 力 ( 图 33b); 由 于 钢 筋 的 屈 服 后 刚 度 基 本 为<br />

“0”, 柱 屈 服 后 荷 载 - 位 移 曲 线 基 本 呈 平 台 状 。SFCB 增 强 柱 的 开 裂 荷 载 和 RC 柱 相 差 不 多 , 都 为 51kN<br />

左 右 , 复 合 筋 内 芯 钢 筋 屈 服 后 , 由 于 外 包 FRP 的 高 强 度 特 征 , 滞 回 曲 线 在 正 向 卸 载 及 反 向 加 载 阶 段<br />

表 现 出 稳 定 的 二 次 刚 度 ( 屈 服 后 刚 度 ), 而 卸 载 阶 段 有 一 定 的 捏 拢 效 应 。 柱 C-S10B20 和 柱 C-S10C24<br />

-138-


的 破 坏 过 程 也 与 柱 C-S10B30 和 柱 C-S10C40 类 似 , 滞 回 曲 线 表 现 出 稳 定 的 二 次 刚 度 , 卸 载 阶 段 有 一 定<br />

的 捏 拢 效 应 。<br />

90<br />

60<br />

C-S14<br />

C-S10B30<br />

C-S10C40<br />

150<br />

100<br />

RC 柱 屈 服 平 台<br />

( 二 次 刚 度 接 近 为 “0”)<br />

荷 载 (kN)<br />

30<br />

0<br />

-30<br />

SFCB 柱 开 裂 后 ,<br />

初 始 刚 度 小 于 RC 柱<br />

荷 载 (kN)<br />

50<br />

0<br />

-50<br />

-60<br />

-90<br />

-6 -4 -2 0 2 4 6<br />

柱 顶 侧 移 (mm)<br />

(a) 按 力 加 载 阶 段<br />

150<br />

试 验 SBFCB 柱 二 次 刚 度<br />

复 合 筋<br />

( 有 效 段 较 长 )<br />

S10-B20<br />

100<br />

外 包 维 断 裂<br />

-100<br />

-150<br />

150<br />

100<br />

C-S14<br />

-40 -20 0 20 40<br />

柱 顶 侧 移 (mm)<br />

(b)C-S14<br />

试 验 SBFCB 柱 二 次 刚 度<br />

( 有 效 段 较 长 )<br />

复 合 筋 S10-B30<br />

外 包 玄 武 岩<br />

纤 维 断 裂<br />

荷 载 (kN)<br />

50<br />

0<br />

-50<br />

荷 载 (kN)<br />

50<br />

0<br />

-50<br />

-100<br />

C-S10B20<br />

-100<br />

C-S10B30<br />

-150<br />

150<br />

100<br />

-40 -20 0 20 40<br />

柱 顶 侧 移 (mm)<br />

(c)C-S10B20<br />

复 合 筋 S10-C24<br />

外 包 碳 纤 维 断 裂<br />

-150<br />

150<br />

100<br />

-40 -20 0 20 40<br />

柱 顶 侧 移 (mm)<br />

(d)C-S10B30<br />

试 验 SCFCB 柱 二 次 刚 度<br />

( 有 效 段 较 短 )<br />

复 合 筋 S10-C40<br />

外 包 碳 纤 维 断 裂<br />

50<br />

50<br />

荷 载 (kN)<br />

0<br />

荷 载 (kN)<br />

0<br />

-50<br />

-50<br />

-100<br />

-150<br />

-100<br />

C-S10C24<br />

C-S10C40<br />

-150<br />

-40 -20 0 20 40<br />

-40 -20 0 20 40<br />

柱 顶 侧 移 (mm)<br />

柱 顶 侧 移 (mm)<br />

(e)C--S10C24<br />

(f)C--S10C40<br />

图 33 各 柱 滞 回 曲 线<br />

卸 载 刚 度 决 定 了 结 构 在 震 后 的 残 余 位 移 , 是 提 高 结 构 可 修 复 性 的 重 要 参 数 。 在 同 样 卸 载 点 , 卸<br />

载 刚 度 越 小 , 相 应 残 余 位 移 越 小 。 对 五 个 混 凝 土 柱 按 屈 服 荷 载 、 屈 服 位 移 无 量 纲 化 骨 架 曲 线 如 图 34<br />

所 示 。 可 以 看 出 各 柱 在 屈 服 位 移 以 前 的 荷 载 - 位 移 曲 线 基 本 重 合 。SFCB 增 强 混 凝 土 柱 由 于 FRP 高 强<br />

度 特 征 , 表 现 出 稳 定 的 二 次 刚 度 , 不 同 卸 载 刚 度 将 直 接 带 来 不 同 的 残 余 位 移 。<br />

SFCB 增 强 混 凝 土 柱 试 件 的 残 余 侧 移 率 见 图 34 SFCB 柱 无 量 纲 化 试 验 柱 骨 架 曲 线 图 , 普<br />

通 RC 柱 由 于 钢 筋 的 塑 性 发 展 , 残 余 位 移 随 着 加 载 位 移 的 增 加 而 增 加 ( 图 35), 且 最 先 达 到 日 本 规 范<br />

的 可 修 复 性 限 值 ( 柱 顶 加 载 位 移 24.37mm)。<br />

SFCB 增 强 混 凝 土 柱 C-S10C40 和 C-S10B30 可 修 复 性 限 值 的 柱 顶 侧 移 分 别 达 到 29.76mm 和<br />

32.80mm, 是 RC 柱 的 122.11% 和 134.59%; 随 着 加 载 位 移 增 加 ,SFCB 外 包 FRP 逐 渐 断 裂 ,SFCB 增 强<br />

混 凝 土 柱 残 余 侧 移 率 加 速 变 大 ( 曲 线 增 加 ), 相 应 普 通 RC 柱 残 余 侧 移 率 与 加 载 位 移 基 本 呈 线 性 关 系 ,<br />

这 个 前 面 统 计 的 普 通 RC 柱 的 趋 势 相 同 。<br />

-139-


P/P y<br />

1.4<br />

1.2<br />

1.0<br />

0.8<br />

柱 屈 服 点<br />

柱 C-S10B30 二 次 刚 度 效 果 最 好<br />

0.6<br />

柱 卸 载 刚 度 K u<br />

S14<br />

0.4 混 凝 土 开 裂<br />

B20<br />

B30<br />

0.2<br />

C24<br />

卸 载 残 余 位 移<br />

C40<br />

0.0<br />

0 1 2 3 4 5<br />

δ/δ y<br />

残 余 侧 移 率 (mm)<br />

图 34 SFCB 柱 无 量 纲 化 试 验 柱 骨 架 曲 线<br />

30<br />

20<br />

10<br />

0<br />

-10<br />

-20<br />

-30<br />

日 本 规 范 规 定<br />

可 修 复 残 余 位 移 限 值<br />

(1% 柱 高 ,11mm)<br />

-40 -20 0 20 40<br />

卸 载 点 位 移 (mm)<br />

图 35 残 余 侧 移 率 比 较<br />

S14<br />

B30<br />

C40<br />

1% 柱 高<br />

六 、 结 构 综 合 高 性 能 及 可 持 续 性 FRP 提 升 技 术<br />

如 上 文 所 述 , 只 采 用 传 统 结 构 材 料 往 往 受 到 力 学 性 能 限 制 无 法 实 现 大 跨 高 性 能 结 构 , 而 一 味 采<br />

用 先 进 的 纤 维 复 合 材 料 造 价 往 往 成 为 瓶 颈 因 素 , 因 此 如 何 合 理 使 用 先 进 的 纤 维 复 合 材 料 实 现 传 统 材<br />

料 无 法 达 到 的 高 性 能 结 构 将 是 结 构 工 程 需 要 解 决 的 重 要 问 题 。 同 时 , 除 考 虑 结 构 的 基 本 力 学 性 能 外 ,<br />

结 构 在 寿 命 周 期 内 会 受 到 各 种 恶 劣 环 境 的 影 响 及 长 期 荷 载 的 作 用 , 结 构 的 可 持 续 性 设 计 和 评 价 不 容<br />

忽 视 。 目 前 的 FRP 结 构 加 固 及 增 强 的 大 多 数 研 究 , 只 局 限 于 短 期 力 学 性 能 , 结 构 的 长 期 性 能 无 法 得<br />

到 定 量 的 评 价 , 这 也 很 大 程 度 上 限 制 了 FRP 材 料 的 推 广 使 用 。 因 此 , 如 何 采 用 FRP 材 料 作 为 结 构<br />

构 件 , 或 FRP 材 料 与 传 统 结 构 材 料 的 有 机 结 合 , 以 实 现 结 构 体 系 的 高 度 可 持 续 化 发 展 是 目 前 欠 缺 的<br />

一 个 重 要 方 向 。 为 此 作 者 研 究 团 队 研 发 了 若 干 综 合 性 能 优 异 、 可 持 续 性 的 新 结 构 , 主 要 包 括 : 综 合<br />

高 性 能 超 大 跨 混 杂 纤 维 复 合 拉 索 斜 拉 桥 结 构 和 具 有 高 耐 久 性 和 良 好 综 合 性 能 的 湿 法 外 包 式 混 杂<br />

FRP- 混 凝 土 组 合 结 构 。<br />

6.1. 综 合 高 性 能 混 杂 纤 维 复 合 拉 索 大 跨 斜 拉 桥 结 构<br />

为 了 满 足 世 界 各 国 对 于 跨 越 海 峡 海 湾 或 江 河 大 跨 度 桥 梁 的 不 断 需 求 , 如 2008 年 开 通 的 1088 米<br />

主 跨 的 苏 通 大 桥 ,2009 年 建 设 完 成 的 1018 米 主 跨 的 香 港 昂 船 洲 大 桥 , 正 在 筹 建 的 意 大 利 墨 西 拿 海<br />

峡 大 桥 (3300 米 主 跨 ) 等 , 大 跨 桥 梁 正 在 不 断 进 行 技 术 和 材 料 的 革 新 。 采 用 传 统 钢 材 或 混 凝 土 的 大<br />

跨 桥 梁 , 由 于 材 料 自 身 的 物 理 、 化 学 和 力 学 性 能 限 制 , 桥 梁 的 力 学 性 能 和 长 期 使 用 性 能 正 在 受 到 不<br />

断 的 挑 战 , 如 大 跨 桥 梁 用 拉 索 的 耐 久 性 , 随 着 跨 度 增 加 拉 索 承 载 效 率 的 下 降 , 结 构 在 使 用 周 期 内 的<br />

维 护 或 更 换 成 本 等 。 要 从 本 质 上 解 决 这 些 问 题 , 实 现 具 有 可 持 续 发 展 的 结 构 , 就 必 须 从 革 新 现 有 材<br />

料 , 充 分 发 挥 材 料 的 优 势 , 以 满 足 结 构 可 持 续 性 发 展 要 求 。<br />

基 于 上 述 背 景 , 作 者 研 究 团 队 从 大 跨 斜 拉 桥 出 发 , 研 究 了 纤 维 复 合 拉 索 千 米 级 大 跨 斜 拉 桥 的 静<br />

力 和 动 力 性 能 , 阐 明 了 混 杂 纤 维 复 合 拉 索 的 综 合 高 性 能 和 优 势 , 在 此 基 础 上 进 一 步 拓 展 纤 维 复 合 拉<br />

索 的 应 用 跨 度 , 分 析 推 导 了 多 种 纤 维 复 合 拉 索 在 千 米 至 万 米 级 大 跨 斜 拉 桥 中 的 设 计 方 法 、 拉 索 材 料<br />

利 用 效 率 、 造 价 和 各 种 拉 索 的 最 优 适 用 跨 径 , 为 纤 维 复 合 材 料 适 材 所 用 提 供 依 据 。 并 通 过 上 述 研 究 ,<br />

发 现 纤 维 复 合 拉 索 随 着 桥 梁 跨 度 增 加 , 振 动 响 应 将 会 越 发 难 以 控 制 。 针 对 该 问 题 , 设 计 提 出 了 基 于<br />

混 杂 纤 维 的 自 减 振 拉 索 , 并 理 论 证 明 了 自 减 振 拉 索 的 阻 尼 性 能 和 在 外 激 励 下 的 振 动 响 应 , 具 体 如 下 。<br />

[43-45]<br />

(1) 纤 维 复 合 拉 索 千 米 级 大 跨 斜 拉 桥 静 动 力 性 能 分 析<br />

为 保 证 千 米 级 大 跨 斜 拉 桥 在 使 用 荷 载 下 的 性 能 要 求 , 对 多 种 纤 维 复 合 拉 索 , 包 括 碳 纤 维 、 玄 武<br />

岩 纤 维 、 玄 武 岩 - 碳 混 杂 纤 维 拉 索 采 用 等 刚 度 设 计 方 法 , 等 效 替 代 钢 拉 索 进 行 结 构 静 动 力 分 析 , 研 究<br />

结 果 表 明 :(a) 根 据 等 效 弹 性 模 量 分 析 , 混 杂 玄 武 岩 - 碳 纤 维 拉 索 比 钢 拉 索 和 单 纯 玄 武 岩 纤 维 拉 索 具<br />

有 更 高 的 等 效 弹 性 模 量 , 虽 然 没 有 碳 纤 维 高 , 但 在 2000m 跨 度 下 , 混 杂 纤 维 拉 索 依 然 能 够 保 持 接 近<br />

90% 的 高 等 效 弹 性 模 量 , 材 料 的 使 用 效 率 高 。(b) 由 于 混 杂 纤 维 拉 索 具 有 接 近 纯 碳 纤 维 拉 索 的 等 效<br />

弹 性 模 量 , 在 2000m 跨 度 下 具 有 相 对 于 钢 拉 索 斜 拉 桥 更 小 的 跨 中 挠 度 , 这 也 说 明 高 等 效 弹 性 模 量 可<br />

以 提 高 纤 维 复 合 拉 索 的 刚 度 , 提 高 材 料 使 用 效 率 。(c) 在 2000m 跨 度 下 , 纤 维 复 合 拉 索 的 刚 度 优 势<br />

得 到 体 现 并 且 具 有 更 高 的 承 载 能 力 , 尤 其 是 混 杂 纤 维 拉 索 , 表 现 出 与 碳 纤 维 拉 索 相 似 的 线 弹 性 荷 载<br />

- 挠 度 关 系 ( 图 36)。(d) 纤 维 复 合 拉 索 具 有 明 显 高 于 钢 拉 索 的 自 振 频 率 , 因 此 有 利 于 避 免 拉 索 和 桥<br />

面 发 生 共 振 : 一 方 面 可 能 发 生 共 振 的 拉 索 数 量 比 钢 拉 索 斜 拉 桥 大 为 减 少 ; 另 一 方 面 可 能 发 生 共 振 的<br />

-140-


拉 索 频 率 阶 数 也 低 于 钢 拉 索 ( 图 37)。<br />

Distributed load (kN/m)<br />

1400<br />

1200<br />

1000<br />

800<br />

600<br />

400<br />

200<br />

0<br />

with steel cable<br />

with CFRP cable<br />

with BFRP cable<br />

with hybrid B/CFRP cable<br />

0 20 40 60 80 100 120 140<br />

Vertical displacement of mid-span (m)<br />

图 36 各 种 纤 维 复 合 拉 索 斜 拉 桥 L-D 关 系<br />

First natural frequency (Hz<br />

3<br />

2.5<br />

2<br />

1.5<br />

1<br />

0.5<br />

Steel cable<br />

CFRP cable<br />

BFRP cable<br />

Hybrid B/CFRP cable<br />

0<br />

150 200 250 300 350 400 450 500 550 600<br />

Length of stay cables (m)<br />

图 37 拉 索 和 全 桥 自 振 频 率 比 较<br />

[46-47]<br />

(2) 千 米 到 万 米 级 大 跨 斜 拉 桥 纤 维 复 合 拉 索 设 计 方 法 研 究<br />

研 究 考 虑 对 于 不 同 跨 度 斜 拉 桥 , 不 同 纤 维 复 合 拉 索 可 以 具 有 不 同 的 设 计 安 全 系 数 , 从 而 更 加 充<br />

分 的 利 用 各 自 材 料 性 能 。 从 拉 索 的 设 计 安 全 系 数 和 等 效 弹 性 模 量 和 等 效 刚 度 的 关 系 出 发 , 得 到 安 全<br />

系 数 设 计 上 限 和 下 限 ( 疲 劳 强 度 控 制 ), 并 提 出 三 段 式 系 数 模 型 描 述 这 一 规 律 及 指 导 拉 索 设 计 , 如<br />

图 38 所 示 。<br />

同 时 , 从 材 料 刚 度 利 用 效 率 和 强 度 利 用 效 率 两 方 面 ( 即 等 效 弹 性 模 量 和 等 效 刚 度 ) 出 发 , 得 到<br />

最 优 化 材 料 强 度 和 刚 度 使 用 效 率 的 拉 索 设 计 系 数 λ 2 ( 图 39), 据 此 推 导 提 出 了 各 种 纤 维 复 合 拉 索 的<br />

优 化 适 用 跨 径 , 阐 明 了 各 种 纤 维 复 合 拉 索 的 不 同 优 势 跨 度 ( 图 40), 最 后 结 合 造 价 比 较 , 定 量 分 析<br />

了 1000-3000m 跨 度 , 各 种 拉 索 的 造 价 比 , 结 果 表 明 随 着 跨 度 的 增 加 , 碳 纤 维 及 其 和 玄 武 岩 纤 维 的<br />

混 杂 纤 维 拉 索 与 钢 拉 索 的 相 对 造 价 不 断 减 小 , 但 3000m 跨 度 内 它 们 仍 维 持 在 2.7~3.8 倍 的 造 价 之 间 ;<br />

在 此 跨 度 内 , 玄 武 岩 纤 维 - 钢 丝 混 杂 拉 索 具 有 与 钢 拉 索 非 常 接 近 的 造 价 比 ( 低 于 1.5 倍 ), 具 有 很 高<br />

的 性 价 比 ( 图 41)。<br />

2<br />

λ<br />

6.00<br />

5.00<br />

4.00<br />

3.00<br />

2.00<br />

1.00<br />

0.00<br />

图 38 三 段 式 拉 索 安 全 系 数 模 型<br />

HS steel cable CFRP cable B/CFRP 25% cable<br />

B/SFRP 20% cable B/SFRP 30% cable B/CFRP 50% cable<br />

BFRP cable<br />

2<br />

λ =1.78<br />

500 1000 1500 2000 2500 3000 3500 4000 4500 5000<br />

Horizantal length of cable (m)<br />

图 40 各 种 拉 索 适 用 跨 径<br />

Esec/E0 and EsecA/(EsecA)max<br />

Esec/E0 EsecA/(EsecA)max<br />

1.0<br />

0.9<br />

0.8<br />

0.7<br />

0.6<br />

0.5<br />

0.4<br />

0.3<br />

0.2<br />

0.1<br />

0.0<br />

2<br />

λ =1.78<br />

0.01 0.1 1 10 100 1000<br />

2 λ<br />

图 39 拉 索 等 效 弹 性 模 量 和 等 效 刚 度 与 λ2 关 系<br />

Relative cost<br />

5.00<br />

4.50<br />

4.00<br />

3.50<br />

3.00<br />

2.50<br />

2.00<br />

1.50<br />

1.00<br />

0.50<br />

0.00<br />

1000 2000 3000<br />

Main span (m)<br />

HS steel<br />

CFRP<br />

BFRP B/CFRP 25%<br />

B/CFRP 50% B/SFRP 20%<br />

B/SFRP 30%<br />

图 41 多 种 拉 索 造 价 比 较<br />

(3) 混 杂 纤 维 复 合 拉 索 自 减 振 设 计 方 法 研 究<br />

[48-49]<br />

由 于 大 跨 斜 拉 桥 拉 索 振 动 问 题 突 出 , 并 考 虑 充 分 利 用 混 杂 纤 维 复 合 材 料 的 特 性 , 在 普 通 混 杂 拉<br />

索 的 基 础 上 设 计 提 出 了 自 减 振 混 杂 纤 维 拉 索 。 具 体 如 图 42 所 示 , 自 减 振 拉 索 将 两 种 纤 维 材 料 相 对<br />

分 开 , 把 拉 索 从 截 面 上 分 为 内 拉 索 和 外 拉 索 , 两 种 拉 索 间 填 充 粘 弹 性 材 料 。 因 为 内 外 拉 索 具 有 不 同<br />

-141-


的 自 振 频 率 , 在 拉 索 发 生 振 动 时 , 内 外 拉 索 发 生 相 互 作 用 , 挤 压 粘 弹 性 材 料 从 而 消 耗 振 动 能 量 , 起<br />

到 自 减 振 的 作 用 。 同 时 , 粘 弹 性 材 料 沿 拉 索 纵 向 不 连 续 布 置 , 可 起 到 减 轻 拉 索 自 重 同 时 又 获 得 等 效<br />

减 振 效 果 的 作 用 。<br />

Hybrid B/CFRP tendons(strands)<br />

Viscoelastic material<br />

BFRP tendons(strands)<br />

Viscoelastic<br />

material<br />

Outer cable<br />

Inner cable<br />

Inner sleeve<br />

l<br />

L/(N+1)<br />

l<br />

图 42 自 减 振 混 杂 纤 维 拉 索 设 计<br />

研 究 从 理 论 上 证 明 了 该 拉 索 设 计 方 法 的 减 振 效 果 。 基 于 Hamilton 原 理 , 从 动 能 , 势 能 和 非 保 守<br />

力 做 功 平 衡 角 度 出 发 , 推 导 建 立 了 包 含 粘 弹 性 材 料 的 拉 索 平 面 内 和 平 面 外 振 动 方 程 , 并 根 据 动 力 平<br />

衡 方 程 , 引 入 内 外 拉 索 的 相 互 作 用 力 , 从 而 得 到 拉 索 整 体 阻 尼 比 方 程 。<br />

为 验 证 方 程 的 适 用 性 , 根 据 设 计 拉 索 参 数 , 设 计 了 自 减 振 混 杂 拉 索 , 并 选 取 常 用 材 料 系 数 , 得<br />

到 自 减 振 拉 索 在 不 同 振 幅 下 的 阻 尼 比 , 证 明 了 该 设 计 的 有 效 性 。 同 时 , 也 证 明 对 平 面 外 振 动 的 抑 制<br />

效 果 要 更 为 明 显 。<br />

[50-51]<br />

(4) 纤 维 复 合 拉 索 在 间 接 激 励 作 用 下 的 响 应 分 析<br />

为 分 析 评 价 多 种 纤 维 复 合 拉 索 的 振 动 响 应 性 能 , 同 时 比 较 自 减 振 拉 索 减 振 效 果 , 研 究 根 据 拉 索<br />

在 参 数 激 励 和 外 激 励 下 的 振 动 响 应 方 程 , 结 合 各 种 拉 索 设 计 参 数 , 数 值 分 析 了 各 种 拉 索 随 激 励 大 小<br />

的 振 幅 响 应 和 索 力 响 应 。 具 体 如 图 43 所 示 。<br />

Amplitude (mm)<br />

6000<br />

5000<br />

4000<br />

3000<br />

2000<br />

1000<br />

0<br />

Steel<br />

CFRP<br />

BFRP B/CFRP 25%<br />

B/CFRP 50% B/SFRP 20%<br />

B/SFRP 30%<br />

B/CFRP 25% SD<br />

0 50 100 150 200<br />

Amplitude of external excitation (mm)<br />

Increment of cable tension (%)<br />

20%<br />

18%<br />

16%<br />

14%<br />

12%<br />

10%<br />

8%<br />

6%<br />

4%<br />

2%<br />

0%<br />

Steel<br />

CFRP<br />

BFRP B/CFRP 25%<br />

B/CFRP 50% B/SFRP 20%<br />

B/SFRP 30%<br />

B/CFRP 25% SD<br />

0 50 100 150 200<br />

Amplitude of external excitation (mm)<br />

(a) 振 幅<br />

(b) 应 力 增 加<br />

图 43 外 激 励 下 拉 索 振 动 响 应<br />

纤 维 拉 索 在 外 激 励 作 用 下 , 表 现 出 相 对 于 钢 拉 索 不 同 程 度 的 较 小 响 应 , 但 响 应 的 整 体 趋 势 相 近 ;<br />

对 于 自 减 振 混 杂 拉 索 , 在 较 小 激 励 下 , 表 现 出 较 小 响 应 , 随 着 激 励 的 增 大 , 响 应 幅 值 也 逐 渐 与 其 他<br />

纤 维 拉 索 接 近 。 在 外 激 励 下 , 各 纤 维 拉 索 的 应 力 增 加 幅 值 要 远 低 于 钢 拉 索 , 主 要 因 为 各 纤 维 拉 索 相<br />

对 较 低 的 弹 性 模 量 , 降 低 了 应 力 增 加 的 幅 值 。<br />

-142-


A m plitude (m m )<br />

6000<br />

5000<br />

4000<br />

3000<br />

2000<br />

1000<br />

0<br />

Steel<br />

CFRP<br />

BFRP B/CFRP 25%<br />

B/CFRP 50% B/SFRP 20%<br />

B/SFRP 30%<br />

B/CFRP 25% SD<br />

0 50 100 150 200<br />

Amplitude of parametric excitation (mm)<br />

(a) 振 幅<br />

Increm ent of cable tension (% )<br />

25%<br />

20%<br />

15%<br />

10%<br />

5%<br />

0%<br />

Steel<br />

CFRP<br />

BFRP B/CFRP 25%<br />

B/CFRP 50% B/SFRP 20%<br />

B/SFRP 30%<br />

B/CFRP 25% SD<br />

0 50 100 150 200<br />

Amplitude of parametric excitation (mm)<br />

(b) 应 力 增 加<br />

图 44 参 数 激 励 下 拉 索 振 动 响 应<br />

在 参 数 激 励 作 用 下 , 除 自 减 振 拉 索 , 其 他 纤 维 拉 索 表 现 出 和 钢 拉 索 一 致 的 振 幅 响 应 。 自 减 振 拉<br />

索 由 于 具 有 较 高 的 阻 尼 比 , 表 现 出 振 动 响 应 滞 后 , 因 此 相 对 其 他 拉 索 具 有 更 大 的 优 势 。 玄 武 岩 纤 维<br />

- 钢 丝 复 合 拉 索 表 现 出 和 钢 拉 索 类 似 的 应 力 幅 值 响 应 , 而 其 他 纤 维 拉 索 应 力 增 加 则 较 低 ( 图 44)。<br />

此 外 , 对 于 纤 维 复 合 拉 索 , 随 着 跨 度 的 增 加 , 参 数 激 励 对 拉 索 的 控 制 因 素 将 逐 渐 变 为 外 激 励 和<br />

参 数 激 励 响 应 相 当 , 因 此 , 在 分 析 设 计 时 需 要 同 时 考 虑 两 者 响 应 。<br />

6.2. 湿 法 外 包 式 混 杂 FRP- 混 凝 土 组 合 结 构 研 究<br />

钢 筋 混 凝 土 结 构 的 锈 蚀 问 题 在 世 界 范 围 内 都 很 严 重 , 并 造 成 了 巨 大 的 经 济 损 失 , 发 展 和 建 设 具<br />

有 高 耐 久 性 的 结 构 已 迫 在 眉 睫 ; 过 去 的 几 十 年 中 , 关 于 如 何 解 决 钢 筋 锈 蚀 问 题 已 经 进 行 了 很 多 研 究 。<br />

镀 锌 、 不 锈 钢 、 阴 极 保 护 法 、 环 氧 涂 层 、 混 凝 土 外 加 剂 等 等 之 类 的 方 法 都 已 被 试 验 过 。 然 而 , 上 述<br />

方 法 中 没 有 一 种 方 法 能 够 彻 底 解 决 问 题 。 具 有 轻 质 、 高 强 、 耐 腐 蚀 等 优 异 性 能 的 FRP 材 料 的 出 现 ,<br />

为 从 根 本 上 解 决 钢 筋 锈 蚀 问 题 带 来 了 希 望 。<br />

目 前 , 具 有 较 高 耐 久 性 的 FRP- 混 凝 土 组 合 结 构 正 在 成 为 国 际 上 研 究 的 热 点 。 初 步 研 究 表 明 :FRP<br />

型 材 、FRP 夹 芯 板 以 及 FRP- 混 凝 土 组 合 板 等 等 具 有 轻 质 高 强 、 施 工 方 便 、 耐 久 、 耐 疲 劳 的 特 性 , 因<br />

此 非 常 适 合 作 为 桥 面 板 以 及 轻 质 人 行 桥 等 结 构 。 目 前 , 学 者 们 已 经 进 行 了 多 种 形 式 的 FRP- 混 凝 土 组<br />

合 梁 试 验 研 究 , 然 而 对 于 组 合 梁 合 理 的 截 面 形 式 尚 未 取 得 一 致 的 观 点 。 轻 型 化 设 计 的 组 合 梁 往 往 面<br />

临 着 刚 度 小 、 挠 度 大 、 腹 板 失 稳 、 抗 剪 性 能 差 、 脆 性 破 坏 、 连 接 构 造 难 、 造 价 高 等 方 面 的 问 题 。 开<br />

发 研 究 能 够 在 公 路 桥 梁 中 应 用 的 中 、 大 跨 度 的 组 合 梁 成 为 了 一 种 历 史 使 命 。<br />

为 了 充 分 利 用 纤 维 材 料 和 降 低 成 本 , 作 者 所 在 团 队 提 出 了 一 种 新 型 的 FRP 外 包 混 凝 土 组 合 梁<br />

[52] 。 该 组 合 梁 采 用 纤 维 为 主 要 增 强 材 料 , 并 采 用 最 小 的 钢 筋 配 筋 率 , 通 过 改 变 配 纤 率 和 侧 面 纤 维 来<br />

研 究 组 合 梁 受 力 。 组 合 梁 的 破 坏 模 式 表 现 为 纤 维 拉 断 和 纤 维 剥 离 。 组 合 梁 具 有 较 高 的 极 限 承 载 力 ,<br />

CFRP 配 纤 率 为 0.27%, 钢 筋 配 筋 率 也 为 0.27% 的 组 合 梁 比 配 筋 率 为 2.5% 的 组 合 梁 承 载 力 高 出 60%。<br />

组 合 梁 的 开 裂 后 刚 度 和 配 筋 率 为 1.0% 的 普 通 混 凝 土 梁 相 当 。 同 年 , 作 者 所 在 团 队 又 提 出 了 一 种 采 用<br />

湿 粘 结 技 术 的 组 合 梁 。 这 种 组 合 梁 先 在 预 先 制 作 好 的 混 杂 FRP 模 壳 内 壁 上 涂 刷 粘 结 树 脂 , 后 浇 筑 混<br />

凝 土 内 芯 ( 湿 粘 结 ),FRP 模 壳 作 为 钢 筋 混 凝 土 内 芯 的 保 护 材 料 和 组 合 梁 的 部 分 抗 弯 增 强 材 料 。 他<br />

们 首 先 对 低 配 筋 率 ( 钢 筋 配 筋 率 为 0.2%, 纤 维 参 与 受 力 比 例 在 83% 左 右 ) 的 矩 形 混 杂 FRP-RC 组 合<br />

梁 进 行 了 试 验 研 究 , 混 杂 纤 维 为 高 弹 模 碳 纤 维 与 玻 璃 纤 维 的 混 杂 ,FRP 与 混 凝 土 之 间 的 界 面 采 用 湿<br />

粘 结 和 粘 粗 砂 两 种 , 试 件 及 尺 寸 见 图 45。 初 步 研 究 表 明 : 湿 粘 结 施 工 方 便 快 速 , 特 别 适 合 于 新 建 结<br />

构 。 低 配 筋 率 (ρ=0.2%) 的 该 种 组 合 梁 与 适 筋 率 (ρ=1.5%) 的 普 通 混 凝 土 梁 在 初 期 刚 度 、 开 裂 荷 载 、<br />

屈 服 强 度 以 及 极 限 强 度 相 当 。 然 而 这 种 低 配 筋 率 组 合 梁 的 使 用 刚 度 较 小 , 造 价 较 高 , 因 而 后 续 的 研<br />

究 中 提 高 了 钢 筋 受 力 比 例 , 研 究 了 钢 筋 和 纤 维 受 力 比 例 不 同 的 组 合 梁 。<br />

-143-


a) 截 面 尺 寸 b) 纤 维 受 力 为 主 的 组 合 梁 c) 纤 维 受 力 为 主 的 组 合 梁 荷 载 位 移 曲 线<br />

图 45 纤 维 受 力 为 主 的 矩 形 组 合 梁<br />

作 者 研 究 团 队 对 纤 维 受 力 为 主 的 FRP- 混 凝 土 T 形 梁 进 行 了 受 弯 性 能 试 验 研 究 [53] 。 该 种 T 形 组<br />

合 梁 由 纵 向 布 置 在 梁 底 的 混 杂 FRP(2 层 CFRP 和 5 层 GFRP 的 混 杂 布 )、 横 向 布 置 的 GFRP 和 钢 筋<br />

混 凝 土 内 芯 组 成 ,FRP 与 混 凝 土 之 间 采 用 湿 粘 结 界 面 方 式 , 通 过 改 变 配 筋 率 来 研 究 其 受 力 性 能 。 所<br />

有 试 件 的 荷 载 位 移 曲 线 见 图 46。 试 验 结 果 表 明 , 湿 粘 结 是 一 种 有 效 的 界 面 处 理 技 术 , 组 合 梁 具 有 较<br />

高 的 刚 度 和 极 限 承 载 力 。 通 过 合 理 的 设 计 可 以 使 FRP- 混 凝 土 组 合 梁 达 到 损 伤 可 控 的 要 求 , 并 提 出 了<br />

理 想 的 损 伤 可 控 荷 载 位 移 曲 线 , 如 图 46c 所 示 。<br />

125 150 125<br />

150<br />

5D6<br />

4 层 高 延 性 GFRP<br />

箍 筋<br />

受 拉 钢 筋<br />

40<br />

20<br />

160<br />

2 层 高 强 度 CFRP<br />

+5 层 高 延 性 GFRP<br />

40<br />

Load (kN)<br />

350<br />

300<br />

250<br />

200<br />

150<br />

100<br />

50<br />

0<br />

FRP-RC-0.29%<br />

FRP-RC-0.43%<br />

FRP-RC-0.77%<br />

FRP-RC-1.20%<br />

FRP-RC-1.73%<br />

RC-2.38%<br />

0 5 10 15 20 25 30 35<br />

Displacement (mm)<br />

荷 载<br />

PHD<br />

PHS<br />

Py<br />

Pc<br />

正 常 使<br />

用 阶 段 损 伤 可 控 阶 段 极 限 阶 段<br />

B( 钢 筋 屈 服 )<br />

A( 混 凝 土 开 裂 )<br />

钢 筋 混 凝 土 结 构<br />

C(FRP 开 始 逐 步<br />

断 裂 或 剥 离 )<br />

D<br />

(FRP 拉 断 或 剥 离 )<br />

0δc δy δHS δHD<br />

位 移<br />

a) 截 面 形 式 和 尺 寸 b) 荷 载 位 移 曲 线 c) 损 伤 可 控 结 构 的 理 想 荷 载 位 移 曲 线<br />

图 46 纤 维 受 力 为 主 T 形 组 合 梁<br />

在 上 述 研 究 基 础 上 , 作 者 研 究 团 队 设 计 并 研 究 了 纤 维 参 与 受 力 比 例 在 50% 左 右 的 混 杂 纤 维 - 混 凝<br />

土 ( 简 称 HFRP-RC) 组 合 梁 [54] 。 通 过 12 根 四 点 受 弯 T 形 梁 的 试 验 研 究 , 采 用 3 种 界 面 方 式 和 3 种 不<br />

同 混 杂 比 例 , 进 一 步 研 究 了 湿 粘 结 等 界 面 方 式 的 可 靠 性 以 及 HFRP-RC 组 合 梁 的 性 能 。 试 验 结 果 表<br />

明 ,HFRP-RC 组 合 梁 在 钢 筋 屈 服 后 具 有 明 显 的 二 次 刚 度 , 具 有 较 高 的 极 限 承 载 力 和 较 好 的 位 移 延 性 。<br />

荷 载 位 移 曲 线 见 图 47。 试 验 中 所 有 使 用 C50 混 凝 土 组 合 梁 的 破 坏 模 式 均 为 梁 底 部 纤 维 拉 断 , 但 是 组<br />

合 梁 的 侧 面 纤 维 发 生 空 鼓 , 因 此 界 面 方 式 需 要 进 一 步 改 进 。 将 碳 纤 维 (CFRP) 与 玄 武 岩 纤 维 (BFRP)<br />

进 行 混 杂 , 可 以 提 高 CFRP 的 极 限 应 变 值 , 而 且 BFRP 比 例 越 高 , 提 高 越 明 显 , 因 此 计 算 临 界 混 杂<br />

比 和 HFRP-RC 组 合 梁 极 限 承 载 力 时 , 应 考 虑 CFRP 极 限 应 变 提 高 系 数 。<br />

荷 载 /kN<br />

250<br />

200<br />

150<br />

100<br />

1C3B-W-CH<br />

1C3B-W-FCS<br />

1C3B-BG-CH<br />

RC<br />

1C3B-W-CH-SCC<br />

荷 载 /kN<br />

250<br />

200<br />

150<br />

100<br />

1C4B-W-CH<br />

1C4B-W-FCS<br />

1C4B-EB-CH<br />

RC<br />

荷 载 /kN<br />

300<br />

250<br />

200<br />

150<br />

100<br />

1C5B-W-CH<br />

1C5B-W-FCS<br />

1C5B-BG-CH<br />

RC<br />

50<br />

50<br />

50<br />

0<br />

0 10 20 30 40 50 60 70<br />

跨 中 位 移 /mm<br />

0<br />

0 10 20 30 40 50 60<br />

跨 中 位 移 /mm<br />

0<br />

0 10 20 30 40 50 60 70 80<br />

跨 中 位 移 /mm<br />

a)1C3B<br />

b) 1C4B<br />

图 47 纤 维 受 力 比 例 50% T 形 组 合 梁 荷 载 位 移 曲 线<br />

c) 1C5B<br />

上 述 的 试 验 结 果 表 明 , 纤 维 材 料 参 与 受 力 对 组 合 梁 刚 度 的 提 高 作 用 不 明 显 , 而 且 造 价 较 高 , 因<br />

此 作 者 研 究 团 队 进 一 步 降 低 了 纤 维 参 与 受 力 的 比 例 , 开 发 了 钢 筋 受 力 为 主 , 纤 维 主 要 起 耐 腐 蚀 作 用<br />

的 FRP- 混 凝 土 组 合 梁 。 首 先 , 在 前 期 手 糊 制 作 的 混 杂 FRP 模 壳 的 基 础 上 , 对 FRP 模 壳 的 工 业 化 生<br />

-144-


产 工 艺 进 行 探 索 性 研 究 , 制 备 出 了 适 应 工 业 化 生 产 的 性 能 稳 定 的 FRP 模 壳 。 接 着 , 对 湿 粘 结 界 面 进<br />

行 了 改 进 , 提 出 了 FRP 剪 力 键 & 湿 粘 结 组 合 界 面 , 这 种 界 面 方 式 具 有 更 高 的 可 靠 性 。 作 者 研 究 团 队<br />

完 成 了 2 根 钢 筋 混 凝 土 对 比 梁 、6 根 BFRP-RC 组 合 梁 以 及 2 根 钢 丝 复 合 BFRP-RC 组 合 梁 的 受 弯 性<br />

能 试 验 研 究 , 对 其 荷 载 位 移 关 系 ( 图 48)、 破 坏 模 式 、 承 载 力 、 刚 度 、 荷 载 应 变 关 系 等 进 行 了 较 为<br />

深 入 的 讨 论 和 分 析 , 证 明 了 以 钢 筋 受 力 为 主 的 FRP-RC 组 合 梁 不 仅 延 性 好 , 而 且 具 有 良 好 、 稳 定 并<br />

可 设 计 的 二 次 刚 度 , 因 而 也 具 有 较 好 的 可 恢 复 性 ; 梁 底 纵 向 FRP、 钢 筋 和 混 凝 土 能 良 好 地 共 同 变 形 、<br />

协 同 工 作 , 所 有 梁 中 FRP 剪 力 键 & 湿 粘 结 界 面 均 是 安 全 可 靠 的 , 能 有 效 保 证 FRP 与 钢 筋 混 凝 土 内 芯<br />

共 同 工 作 , 表 明 是 一 种 可 靠 的 界 面 方 式 ; 将 FRP-RC 组 合 梁 的 上 述 理 论 分 析 结 果 与 试 验 值 进 行 比 较 ,<br />

计 算 公 式 能 够 较 准 确 地 预 测 组 合 梁 的 二 次 刚 度 和 抗 弯 承 载 力 , 表 明 理 论 分 析 是 正 确 可 行 的 。<br />

Load /kN<br />

250<br />

200<br />

150<br />

100<br />

50<br />

0<br />

0 20 40 60 80 100<br />

deformation /mm<br />

a) B1/B2/B5/B6/B7<br />

B1<br />

B1(FA)<br />

B2<br />

B2(FA)<br />

B5<br />

B5(FA)<br />

B6<br />

B7<br />

Load /kN<br />

400<br />

350<br />

300<br />

250<br />

200<br />

150<br />

100<br />

50<br />

0<br />

0 20 40 60 80 100<br />

Deformation /mm<br />

B2<br />

B2(FA)<br />

B3<br />

B3(FA)<br />

B4<br />

B4(FA)<br />

B8<br />

B8(FA)<br />

b) B2/B3/B4/B8<br />

图 48 钢 筋 受 力 为 主 组 合 梁 荷 载 位 移 曲 线<br />

Load /kN<br />

350<br />

300<br />

250<br />

200<br />

150<br />

100<br />

50<br />

0<br />

0 20 40 60 80 100<br />

Deformation /mm<br />

c) B2/B7/B9/B10<br />

以 上 研 究 结 果 表 明 : 随 着 纤 维 种 类 以 及 钢 筋 和 纤 维 比 例 的 不 同 , 组 合 梁 的 力 学 性 能 明 显 不 同 。<br />

以 钢 筋 受 力 为 主 , 纤 维 主 要 起 耐 腐 蚀 作 用 的 FRP- 混 凝 土 组 合 梁 具 有 优 异 的 综 合 性 能 。 在 这 种 FRP<br />

组 合 结 构 中 ,FRP 高 强 度 可 得 到 充 分 发 挥 ,FRP 的 弹 性 模 量 较 低 可 以 得 到 有 效 弥 补 , 并 且 FRP 的 高<br />

耐 久 性 和 良 好 抗 疲 劳 性 能 也 可 以 极 大 提 升 和 改 善 钢 材 、 钢 筋 和 混 凝 土 等 传 统 结 构 材 料 不 足 , 并 能 使<br />

结 构 实 现 良 好 的 可 修 复 性 。 不 难 预 见 , 经 深 入 研 究 与 优 化 设 计 的 FRP 组 合 结 构 能 够 定 能 发 挥 出 显 著<br />

性 能 优 势 , 在 桥 梁 结 构 及 其 他 工 程 结 构 中 都 具 有 良 好 应 用 前 景 。<br />

七 、 结 论<br />

针 对 FRP 加 固 及 增 强 结 构 存 在 的 问 题 和 不 足 , 结 合 作 者 研 究 团 队 多 年 研 究 成 果 , 总 结 并 提 出 了<br />

一 些 提 升 FRP 高 效 使 用 和 高 效 增 强 结 构 的 方 法 和 技 术 , 结 论 总 结 如 下 :<br />

1) 为 解 决 单 一 纤 维 无 法 满 足 结 构 综 合 性 能 提 升 的 需 要 , 提 出 并 发 展 了 纤 维 混 杂 设 计 理 论 , 研<br />

发 了 多 种 混 杂 纤 维 及 混 杂 纤 维 - 钢 材 复 合 材 料 , 克 服 了 单 一 纤 维 力 学 性 能 的 不 足 。 并 深 入 研 究 了 纤 维<br />

混 杂 效 应 对 于 提 升 材 料 多 种 综 合 性 能 的 作 用 。 在 此 基 础 上 , 应 用 混 杂 纤 维 于 结 构 加 固 及 增 强 , 使 结<br />

构 能 满 足 良 好 的 使 用 刚 度 , 极 限 强 度 及 延 性 要 求 , 显 著 改 善 了 单 一 纤 维 增 强 结 构 脆 性 破 坏 的 不 足 。<br />

研 究 成 果 为 更 加 合 理 、 更 加 充 分 的 利 用 纤 维 复 合 材 料 提 供 参 考 和 借 鉴 。<br />

2) 针 对 FRP 加 固 结 构 使 用 性 能 效 果 不 明 显 的 问 题 , 研 发 了 预 应 力 片 材 和 索 材 张 拉 加 固 结 构 技<br />

术 。 其 中 包 括 PBO 纤 维 布 , 混 杂 纤 维 布 等 预 应 力 片 材 , 碳 纤 维 拉 索 等 预 应 力 索 材 , 解 决 了 一 般 纤 维<br />

布 干 丝 张 拉 预 应 力 度 不 足 和 锚 固 端 部 应 力 集 中 的 问 题 , 开 发 了 高 效 实 用 的 成 套 工 法 , 提 升 了 结 构 在<br />

使 用 状 态 下 的 服 役 性 能 。<br />

3) 在 现 有 FRP 约 束 加 固 混 凝 土 结 构 抗 震 性 能 研 究 的 基 础 上 , 为 充 分 发 挥 FRP 材 料 优 势 , 从 本<br />

质 上 提 高 结 构 抗 震 性 能 , 提 出 了 损 伤 可 控 的 新 抗 震 设 计 理 论 , 并 通 过 嵌 入 式 FRP 筋 加 固 和 钢 - 连 续<br />

纤 维 筋 (SFCB) 的 应 用 , 实 现 结 构 在 中 大 震 作 用 下 保 持 稳 定 的 二 次 刚 度 和 较 小 的 残 余 变 形 , 推 动 新<br />

抗 震 设 计 理 论 及 方 法 的 完 善 。<br />

4) 针 对 当 前 结 构 综 合 高 性 能 及 可 持 续 性 发 展 不 足 的 现 状 , 提 出 了 采 用 FRP 及 混 杂 FRP 作 为 结<br />

构 构 件 实 现 可 持 续 性 结 构 的 理 念 和 方 法 , 并 通 过 对 千 米 至 万 米 级 大 跨 斜 拉 桥 混 杂 FRP 拉 索 和 外 包 式<br />

混 杂 FRP- 混 凝 土 组 合 结 构 的 研 究 , 发 现 并 理 论 和 试 验 证 明 了 FRP 及 混 杂 FRP 在 高 度 提 升 结 构 综 合<br />

性 能 及 可 持 续 性 发 展 方 面 的 突 出 作 用 , 为 结 构 向 更 大 跨 度 , 更 轻 质 量 , 更 高 性 能 和 更 可 持 续 性 方 向<br />

发 展 提 供 了 途 径 。<br />

参 考 文 献 (References)<br />

[1] http://www.toray.jp/products/carbon/index.html.<br />

[2] Ueda, T. and Dai, J. G. FRP Composites for Infrastructures in Japan: Review and Outlook, Innovation & Sustainability<br />

B2<br />

B2(FA)<br />

B7<br />

B7(FA)<br />

B9<br />

B9(FA)<br />

B10<br />

B10(FA)<br />

-145-


of Structures [C]. Proceedings of the International Symposium on Innovation & Sustainability of Structures in Civil<br />

Engineering, Nanjing, Vol.1, 20-22 November, 2005: 305-320.<br />

[3] Ueda, T. Future of FRP in Far East [C]. Third International Conference on FRP composites in Civil Engineering,<br />

December 13-15 2006, Miami, Florida, USA, pp11-16.<br />

[4] Keller, T. Recent all-Composite and Hybrid Fiber-Reinforced Polymer Bridges and Buildings [J]. Progress in<br />

Structural Engineering and Materials, 2001, 3: 132-140.<br />

[5] Bakis, C. E., Bank, L. C., Brown, V. L., Cosenza, E., Davalos, J. F., Lesko, J. J., Machida, A., Rizkalla, S. H., and<br />

Triantafillou, T. C. Fibre-Reinforced Polymer Composites for Construction-State-of-the-Art Review [J]. Journal of<br />

Composite for Construction, 2002, 7(2): 73-87.<br />

[6] Wu, Z. S., Wang, X. and Iwashita, K. State-of-the-Art of Advanced FRP Applications in Civil Infrastructure in Japan<br />

(Invited paper) [C/OL]. COMPOSITES & POLYCON 2007, American Composites Manufacturers Association,<br />

October 17-19, 2007, Tampa, FL USA. http://www.acmanet.org/resources/07papers/Wu179.pdf.<br />

[7] 吴 智 深 , FRP 粘 贴 结 构 加 固 中 的 几 个 关 键 问 题 和 技 术 [J], 建 筑 结 构 , 2007, 37(s1):8-14.<br />

[8] 吴 智 深 , 吴 刚 , 汪 昕 等 , 玄 武 岩 纤 维 在 土 建 交 通 基 础 设 施 领 域 研 究 与 应 用 若 干 新 进 展 [J], 工 业 建 筑 , 2009, 39<br />

(suppl):1-14.<br />

[9] Bank, L. C. Composites for Construction: Structural Design with FRP Materials. John Wiley and Sons, Inc., New<br />

Jersey (2006).<br />

[10] Feng, P., Ye, L. P. and Teng, J.G. Large-span Woven Web Structure Made of Fiber Reinforced Polymer [J]. Journal of<br />

Composites for Construction, ASCE, 2007, 11(2):110-119<br />

[11] Meier, U. Proposal for a Carbon Fibre Reinforced Composite Bridge Across the Strait of Gibraltar at its Narrowest Site<br />

[C]. Proceedings of the Institution of Mechanical Engineers, 1987, 201(B2): 73-78.<br />

[12] Lu, Z. and Mei, K. First Application of CFRP Cables for a Cable-Stayed Bridge in China [J]. China Civil Engineering<br />

Journal, 2007; 40(1): 50-59<br />

[13] http://www.chinagbf.com.<br />

[14] Wu, Z. S. Structural Strengthening and Integrity with Hybrid FRP Composites [C]. In: Proceedings of the Second<br />

International Conference on FRP Composites in Civil Engineering (CICE-2), Adelaide, Australia, 8-10 December 2004:<br />

905-912.<br />

[15] Iwashita, K., Yoshikiyo, K., Wu, Z. S. and Hamaguchi, Y. Experimental Study on Control Index of Stress Drop for<br />

Designing Hybrid FRP Sheets [C]. 16th International Conference on Composite Materials, Kyoto, Japan, 2007.<br />

[16] Wu, Z. S., Wang, X. and Iwashita, K. Development of High Performance Structures with Hybrid FRP Composites<br />

(Keynote paper) [C]. <strong>The</strong> 10th International Symposium on Structural Engineering for Young Experts (ISSEYE10),<br />

Changsha, China, Oct. 19-21, 2008:11-20.<br />

[17] 吴 刚 , 罗 云 标 , 吴 智 深 等 . 钢 - 连 续 纤 维 复 合 筋 力 学 性 能 试 验 研 究 与 理 论 分 析 . 土 木 工 程 学 报 . 2010, 43(3): 53-61.<br />

[18] Hashimoto, T. Study on Hybrid FRP-steel Wire Sheets for Prestressing Strengthening Structures [D]. Master <strong>The</strong>sis,<br />

Department of Urban and Civil Engineering, Ibaraki <strong>University</strong>, 2009. (in Japanese)<br />

[19] Wu Z. S., Wang X., and Iwashita K, Sasaki T., Hamaguchi Y., Tensile fatigue behavior of FRP and hybrid FRP sheets,<br />

Composites Part B: Engineering, 2010,41(5):396-402.<br />

[20] JSCE. Research Committee on Upgrading of Concrete Structures with Use of Continuous Fiber Sheet [S].<br />

Recommendations for Upgrading of Concrete Structures with Use of Continuous Fiber Sheets, Concrete Engineering<br />

Series, 2001, No.41, Tokyo, Japan.<br />

[21] 施 嘉 伟 , 朱 虹 , 吴 智 深 , 吴 刚 . BFRP/HFRP 布 - 混 凝 土 界 面 黏 结 性 能 试 验 研 究 [J]. 东 南 大 学 学 报 ( 自 然 科 学 版 ),<br />

2010, 40(3): 554-558.<br />

[22] Cao, S. H., Wu, Z. S. and Wang, X. Tensile Properties of CFRP and Hybrid FRP Composites at Elevated Temperatures,<br />

Journal of Composite Materials, 2009, Vol. 43, No.4, 315-330.<br />

[23] Cao, S. H. and Wu, Z. S. Tensile Properties of FRP Composites at Elevated and High Temperatures [J]. Journal of<br />

Applied Mechanics, JSCE, 2008, 11: 963-970.<br />

[24] Wu, Z. S., Shao, Y. X., Iwashita, K. and Sakamoto, K. Strengthening of Preloaded RC Beams Using Hybrid Carbon<br />

Sheets [J]. Journal of Composites for Construction, ASCE, 2007, 11(3): 299-307.<br />

[25] Wu, Z. S., Sakamoto, K., Iwashita, K. and Kobayashi, A. Enhancement of Flexural Performances through FRP<br />

Hybridization with High Modulus Type of Carbon Fibers [J]. Journal of Applied Mechanics, JSCE, 2005, 8: 987-997.<br />

[26] Wu, Z. S., Iwashita, K., Yang, C. Q., Sakamoto, K. and Ahmed, E. Layered Hybrid FRP Sheets for Strengthening<br />

Structures with Integrated Performances [J]. Journal of Reinforced Plastics and Composites, 2010. (In press)<br />

[27] Wu, Z. S., Iwashita, K. and Wang, X. Integrated Strengthening of Structures with Bonded Prestressed FRP<br />

Reinforcements [C]. 5th International Conference on Advanced Composite Materials in Bridges and Structures<br />

(ACMBS-V),Winnipeg, Manitoba, Canada, September 22-24, 2008. (CD-ROM)<br />

[28] Wu, Z. S., Iwashita, K. and Niu, H. D. Study on Strengthening Technique with Prestressed PBO Fiber Sheets [J].<br />

Engineering Science, Vol. 9, No. 7, pp. 18-24, 2005.9. (In Chinese)<br />

[29] Wu, Z. S., Iwashita, K., Hayashi, K., Higuchi, T., Murakami, S. and Koseki, Y. Strengthening Prestressed-Concrete<br />

-146-


Girders with Externally Prestressed PBO Fiber Reinforced Polymer Sheets [J]. Journal of Reinforced Plastics and<br />

Composites, 2003, 22(14):1269-1286.<br />

[30] Wu, Z. S., Hemdan, S. and Iwashita, K., Performance Evaluation and Design of Strengthened Structures with<br />

Prestressed PBO Fiber Reinforced Polymer Sheets [C]. Proceedings of IABSE Symposium on Responding to<br />

Tomorrow’s Challenges in Structural Engineering, Budapest, Hungary, (Lead speaker lecture), 2006.9. (CD-ROM)<br />

[31] Wu, Z. S., Iwashita, K. and Sun, X. H. Structural Performance of RC Beams Strengthened with Prestressed Near<br />

Surface Mounted CFRP Tendons [J]. ACI Special Publication on “Case Histories and Use of FRP for Prestressing<br />

Applications”, 2007, 245(suppl):165-178.<br />

[32] Japan Society of Civil Engineering (JSCE). Earthquake Resistant Design Codes in Japan [S]. JSCE Earthquake<br />

Engineering Committee, 2000, Tokyo, Japan.<br />

[33] 吴 刚 , 吴 智 深 , 吕 志 涛 . FRP 约 束 混 凝 土 圆 柱 有 软 化 段 时 的 应 力 - 应 变 关 系 研 究 . 土 木 工 程 学 报 , 2006, 39(11):<br />

7-14.<br />

[34] Wu, G., Lu, Z. T. and Wu, Z. S. Strength and Ductility of Concrete Cylinders Confined with FRP Composites [J].<br />

Journal of Construction and Building Materials, 2006, 20(3): 134-148.<br />

[35] Wu, G., Wu, Z. S. and Lu, Z. T. Design-Oriented Stress-Strain Model for Concrete Prisms Confined with FRP<br />

Composites [J]. Journal of Construction and Building Materials, 2007, 21(5): 1107-1121.<br />

[36] Fahmy, M. F. M., Wu, Z. S. and Wu, G. Seismic Performance Assessment of Damage-controlled FRP-retrofitted RC<br />

Bridge Columns Using Residual Deformations [J]. Journal of Composites for Construction, ASCE. 2009, 13(6):<br />

498-513.<br />

[37] Fahmy, M. F. M., Wu, Z. and Wu, G. Post-earthquake Recoverability of Existing RC Bridge Piers Retrofitted with FRP<br />

Composites [J]. Construction and Building Materials, 2010, 24(6): 980-998.<br />

[38] Fahmy, M. F. M. Enhancing Recoverability and Controllability of Reinforced Concrete Bridge Frame Columns Using<br />

FRP Composites [D]. Doctor of Philosophy, Ibaraki <strong>University</strong>, Hitachi, Japan, 2010.<br />

[39] 吴 智 深 , 吴 刚 , 吕 志 涛 . 钢 - 连 续 纤 维 复 合 筋 增 强 混 凝 土 抗 震 结 构 [P]. 中 国 : CN1936206A, 2006.<br />

[40] Wu, G., Wu, Z. S., Luo, Y. B., Sun, Z. Y. and Hu, X. Q. Mechanical Properties of Steel-FRP Composite Bar Under<br />

Uniaxial and Cyclic Tensile Loads [J]. Journal of Materials in Civil Engineering, ASCE, 2010, 22(10): 1056-1066.<br />

[41] 孙 泽 阳 , 吴 刚 , 吴 智 深 等 . 钢 - 连 续 纤 维 复 合 筋 (SFCB) 与 混 凝 土 粘 结 性 能 试 验 研 究 [J]. 工 程 抗 震 与 加 固 改 造 ,<br />

2009, 31(1): 21-27.<br />

[42] 孙 泽 阳 , 吴 刚 , 吴 智 深 , 张 敏 . 钢 - 连 续 纤 维 复 合 筋 增 强 混 凝 土 柱 抗 震 性 能 试 验 研 究 [J]. 土 木 工 程 学 报 . 正 式 录<br />

用 , 稿 号 (0020179).<br />

[43] Wu, Z. S. and Wang, X. Investigation on a Thousand-meter Scale Cable-stayed Bridge with Fiber Composite Cables<br />

[C], the Fourth International Conference on FRP Composites in Civil Engineering (CICE2008), July, 22-24, 2008,<br />

Zurich, Switzerland. (CD-ROM)<br />

[44] Wang, X. and Wu, Z. S. Dynamic Behavior of Thousand-meter Scale Cable-Stayed Bridge with Hybrid FRP Cables [J],<br />

Journal of Applied Mechanics (JSCE), 2009, 12: 935-943.<br />

[45] Wang, X. and Wu, Z. S. Integrated High-Performance Thousand-meter Scale Cable-Stayed Bridge with Hybrid FRP<br />

Cables [J]. Composites Part B: Engineering, 2010, 41(2): 166-175.<br />

[46] Wang, X. and Wu, Z. S. Design Considerations of FRP Stay Cable for Extra Long Span Cable-stayed Bridge [J]. World<br />

Journal of Engineering, 2009, 6(suppl.): 1103-1104.<br />

[47] Wang, X. and Wu, Z. S. Evaluation of FRP and Hybrid FRP Cables for Super Long-span Cable-stayed Bridges [J].<br />

Composite Structures, 2010, 92(10): 2582-2590.<br />

[48] Wang, X. and Wu, Z. S. Effectiveness of Smart Dampers for Hybrid FRP Cable in Long-span Cable-stayed Bridge [C].<br />

Proceedings of the Fourth International Conference on FRP Composites in Civil Engineering (CICE2010), 2010, 1:<br />

243-247.<br />

[49] Wang, X. and Wu, Z. S. Modal Damping Evaluation of Hybrid FRP Cable with Smart Dampers for Long-span<br />

Cable-stayed Bridges. Composite Structures (Elsevier), 2011, 93(4): 1231-1238.<br />

[50] Wang, X. and Wu, Z. S. Vibration Behavior of Hybrid FRP Cable under Indirect Excitation [C]. <strong>The</strong> International<br />

Symposium on Innovation & Sustainability of Structures in Civil Engineering (ISISS2009), Guangzhou, China, Nov.<br />

28-30, 2009: 1669-1675.<br />

[51] Wang, X. and Wu, Z. S. Vibration Control of Different FRP Cables in Long-span Cable-stayed Bridge under Indirect<br />

Excitations [J]. Journal of Earthquake & Tsunami, 2011, 5(2): 167-188.<br />

[52] Wu, Z. S., Li, W. X. and Sakuma, N. Innovative Externally Bonded FRP/concrete Hybrid Flexural Members [J].<br />

Composite Structures, 72, 3, pp.289-300, 2006<br />

[53] Wu, Z. S., Iwashita, K. and Li, W. Development of Externally Bonded FRP/concrete Hybrid Members with Hybrid<br />

FRP Sheets [J]. Journal of Applied Mechanics, 2007, 10: 995-1002. (In Japanese)<br />

[54] 张 普 , 朱 虹 , 孟 少 平 等 . T 形 混 杂 FRP- 混 凝 土 组 合 梁 受 弯 性 能 试 验 研 究 . 东 南 大 学 学 报 ( 自 然 科 学 版 ), 2010, 40(3):<br />

548-553.<br />

-147-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

THE EVOLUTION OF TRANSVERSELY CONFINED STRUCTURAL COLUMNS<br />

Y. Xiao 1, 2<br />

1<br />

MOE Key Laboratory of Building Safety and Energy Efficiency<br />

College of Civil Engineering, Hunan <strong>University</strong>, Changsha, China<br />

2<br />

Department of Civil and Environmental Engineering<br />

<strong>University</strong> of Southern California, Los Angeles, CA, USA<br />

Email: yanxiao@usc.edu, cipres@hnu.cn<br />

ABSTRACT<br />

This paper summarizes the author’s more than twenty five years of research efforts and innovation on so-called<br />

tubed concrete columns or transversely confined concrete columns. Systematic research has been conducted by<br />

the author and his colleagues, making the tubed reinforced concrete column originally proposed by Tomii et al.<br />

evolve into transversely confined structural columns, which is a much broader concept. <strong>The</strong> transversely<br />

confined structural columns can include different confinement materials such as steel and fiber reinforced<br />

polymers, transversely confined concrete columns and transversely confined steel columns.<br />

KEYWORDS<br />

Columns, transverse confinement, ductility, strength, seismic behaviour, tubed columns, FRP, innovation.<br />

INTRODUCTION<br />

<strong>The</strong> concept of using steel tube as primarily transverse reinforcement for reinforced concrete (RC) columns was<br />

first studied by a research group at the Kyushu <strong>University</strong>, Japan lead by Tomii. While many initial research<br />

papers and reports were published in Japanese, the English papers were first published in the Proceedings of the<br />

International Speciality Conference on Concrete Filled Steel Tubular Structures, Harbin, China in 1985 [Tomii<br />

et al. 1985a, b.], and then in the Transaction of Japan Concrete Institute [Xiao et al. 1986] and a more widely<br />

publicized international conference in New Zealand [Tomii, Sakino and Xiao 1987]. <strong>The</strong> terminology of “tubed<br />

column” first adopted by Tomii et al. [1985] refers to the function of the steel tube as that of the steel hoops in<br />

a hooped RC column. Thus, a tubed column differs from the conventional concrete filled tubes (CFT) from the<br />

fact that the composite action between the steel tube and concrete is primarily expected in the transverse<br />

direction.<br />

(a)<br />

(b)<br />

Figure 1 Comparison of conventional CFT and tubed columns<br />

Figure 1 compares the conventional concrete filled steel tubular (CFT) column and the tube column. As shown<br />

in Fig.1(b), the tube in the tubed column stops short leaving a small gap at the ends of the column, in order to<br />

intentionally avoid the interaction in the longitudinal direction between the steel tube and the concrete infill, and<br />

in other words, to maximize the transverse interaction between the two. Such details were originally proposed<br />

by Tomii, Sakino and Xiao (1985a and b). Other original development related to steel tube transversely confined<br />

-148-


concrete columns were documented by the author in 2004 in two separate papers [Xiao 2004a, b]. As stated in<br />

Tomii et al.’s original work, a tubed column not only reserves the merits of conventional CFT, but also can still<br />

be considered as RC structural system, thus, connection detailing and other design as well as construction<br />

methods used in conventional RC structures are applicable without significant modification.<br />

TUBED COLUMN CONCEPT FOR RETROFITTING EXISTING DEFICIENT RC COLUMNS<br />

Steel Jacketing<br />

In the original pioneer work of Tomii et al. (1985), the tubed column concept was targeted to new structural<br />

column design, however, it is applicable to retrofit of existing deficient RC columns for improving the ductile<br />

behavior. Such work was systematically carried out for bridge columns by a research team at the <strong>University</strong> of<br />

California at San Diego lead by Priestley and Seible, and the author was among the team and contributed<br />

particularly to the development of steel jacketing for shear retrofit of bridge columns [Priestley et al. 1994a, b].<br />

For retrofitting a bridge column with circular section, two half cylindrical steel shells are welded together to<br />

enclose the existing column with the gap filled by cement paste, however, elliptical steel jacket is proposed.<br />

Figure 2 shows the construction of steel jacketing retrofit of an actual bridge column.<br />

Partially Stiffened Rectilinear Jacketing<br />

Due to the geometrical shape, transversely confining square or rectangular sectioned concrete columns poses<br />

special problems, as the transverse confinement efficiency is reduced near the middle of section sides. Although,<br />

retrofitting bridge columns with rectangular sections with elliptical jacketing may be permitted, it is not<br />

desirable for retrofitting building columns, due to the limitation in floor area. In 2003, Xiao and Wu proposed<br />

and experimentally validated a more efficient steel transverse confinement design using stiffened elements, such<br />

as angles, tubes or pipes [Xiao and Wu 2003].<br />

Figure 2 Steel jacketing retrofit of bridge<br />

column<br />

Figure 3 Prefabricated GFRP jacketing retrofit of more than<br />

3400 bridge columns in California<br />

FRP Jacketing<br />

<strong>The</strong> tubed column or transversely confined concrete column concept can also be extended to the use of other<br />

materials rather than steel for transverse confinement. Fiber reinforced polymer or plastic (FRP) confinement is<br />

particularly promising since the FRP can be engineered to have all the fibers oriented in the transverse direction<br />

to enhance the confinement efficiency. <strong>The</strong> author and his colleagues were among the earlier explorers in<br />

developing prefabricated FRP composite jacketing [Xiao and Ma 1997; Xiao et al. 1999; Ma and Xiao 1999],<br />

and studying in-situ fabricated FRP composite jacketing for transverse confinement of reinforced concrete<br />

columns [Ma, Xiao and Li 2000]. <strong>The</strong> author’s research led to the application of prefabricated GFRP composite<br />

jacketing retrofit of more than 3400 columns for a several miles long bridge viaduct in California, as shown in<br />

Fig.3.<br />

CONFINED CONCRETE FILLED STEEL TUBES (CCFT)<br />

In order to improve the seismic behavior and load carrying capacity of concrete filled tubular (CFT) columns,<br />

the author invented and experimentally validated transversely confined CFT columns or CCFT columns, in<br />

which additional transverse reinforcement is designed for the potential plastic hinge regions [Xiao et al. 2005;<br />

Mao and Xiao 2006]. Figure 4 shows the concept of CCFT. Based on fundamental principles of mechanics, the<br />

design concept is aimed at controlling the local buckling of the steel tube, more efficiently confining concrete<br />

-149-


and delaying or avoiding low-cycle fatigue rupture of the steel tube. Exponential tests were conducted for<br />

columns under cyclic lateral force with constant axial load, and the superior performance of CCFT was<br />

validated. Figure 5 compares the column end moment and lateral displacement skeleton curves of CFT and<br />

CCFT model columns along with the demonstration of final failure patterns.<br />

Figure 4 Concept of CCFT<br />

Column End Moment (kNm)<br />

C2-CCFT3<br />

C1-CFT3<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

-10 -8 -6 -4 -2-100<br />

0 2 4 6 8 10<br />

-200<br />

-300<br />

-400<br />

-500<br />

Drift Ratio (%)<br />

CCFT<br />

CFT<br />

Figure 5 Comparisons of CFT and CCFT behaviors<br />

(b)<br />

(c)<br />

(a)<br />

Figure 6 Transversely confined steel column: (a) critical locations with confinement, (b) H-shape section<br />

confined by steel strips, and (c) H-shape section confined by reinforced concrete<br />

TRANSVERSELY CONFINED STEEL COLUMNS<br />

<strong>The</strong> transverse confinement concept was recently extended to the design of stronger and more ductile steel<br />

columns by the author. Under cyclic loading, steel columns shape such as H-steel (or W steel) may develop<br />

post-yielding local buckling in the potential plastic hinge region. <strong>The</strong> buckled elements are straightened during<br />

the reverse loading, and such repetitive buckling and straightening process significantly reduces the low-cycle<br />

-150-


fatigue life of steel columns. Xiao invented methods to transversely confine the plastic hinge region of a steel<br />

column, and test results validated the concept to have superior seismic behavior [Xiao et al. 2009]. Figure 6<br />

shows the concept of transversely confined steel columns. Cyclic lateral loading test results of full-scale<br />

columns without and with transverse confinement are compared in Fig.7, demonstrating the significantly<br />

improved seismic behaviour of confined steel column.<br />

1.5<br />

1.5<br />

1.0<br />

0.5<br />

HWW<br />

N/N y =0.3<br />

1.0<br />

0.5<br />

HWW-RCC<br />

N/N y =0.3<br />

V/Vp<br />

0.0<br />

V/Vp<br />

0.0<br />

-0.5<br />

-0.5<br />

-1.0<br />

-1.0<br />

-1.5<br />

-10 -8 -6 -4 -2 0 2 4 6 8 10<br />

Drift Ratio(%)<br />

-1.5<br />

-10 -8 -6 -4 -2 0 2 4 6 8 10<br />

Drift Ratio(%)<br />

(a)<br />

(b)<br />

Figure 7 Lateral force and drift ratio hysteretic responses of (a) steel H column, and (b) confined steel H column<br />

MODELLING OF CONFINED CONCRETE<br />

<strong>The</strong> author also devoted significant efforts on fundamental studies aimed to develop rational mechanical models<br />

of confined concrete. In his Doctor of Engineering thesis defended in 1989, the author conducted carefully<br />

designed axial loading tests on more than thirty concrete stub columns confined by steel tubes with different<br />

thickness and concrete strength. Through the instrumentation of strains and analysis of stresses, an octahedral<br />

stress and strain relationship, as shown in Fig.8, along with failure criterion and plastic flow potential were<br />

proposed [Xiao 1989].<br />

For<br />

σ − 3 ε<br />

c<br />

oct c oct<br />

relationship<br />

⎧ ⎛ f ⎞⎫<br />

a = ⎨ − ⎜ ⎟⎬×<br />

⎩ ⎝35.3<br />

⎠⎭<br />

c B<br />

−4<br />

14.60 3.65 10<br />

⎧ ⎛ f ⎞⎫<br />

b = ⎨ + ⎜ ⎟⎬×<br />

⎩ ⎝35.3<br />

⎠⎭<br />

c B<br />

−4<br />

8.27 11.60 10<br />

⎧<br />

⎛ f ⎞⎫<br />

c = ⎨− + ⎜ ⎟⎬×<br />

⎩<br />

⎝35.3<br />

⎠⎭<br />

c B<br />

−4<br />

1.57 5.04 10<br />

For<br />

c<br />

τ<br />

−<br />

γ<br />

oct c oct<br />

relationship<br />

⎧ ⎛ f ⎞⎫<br />

a = ⎨ + ⎜ ⎟⎬×<br />

⎩ ⎝35.3<br />

⎠⎭<br />

c B<br />

−4<br />

8.19 2.03 10<br />

⎧ ⎛ f ⎞⎫<br />

b = ⎨ + ⎜ ⎟⎬×<br />

⎩ ⎝35.3<br />

⎠⎭<br />

c B<br />

−4<br />

7.25 5.85 10<br />

⎧<br />

⎛ f ⎞⎫<br />

c = ⎨− + ⎜ ⎟⎬×<br />

⎩<br />

⎝35.3<br />

⎠⎭<br />

c B<br />

−4<br />

2.82 11.13 10<br />

Figure 8 Octahedral stress-strain relationship for confined concrete proposed by Xiao [1989]<br />

This model was recently used satisfactorily in analyses of CFT and CCFT columns under axial loading [Choi<br />

and Xiao 2010a, b], as shown in Fig.9.<br />

<strong>The</strong> author also carried out axial loading tests on FRP confined concrete stub columns and established a data set<br />

widely used by other researchers [Xiao and Wu 2000]. Based on the test results, stress strain models for FRP<br />

confined concrete were developed [Xiao and Wu 2000; Xiao and Wu 2003].<br />

-151-


Figure 9 Prediction of axial compressive behaviours of CFT columns using Xiao’s confinement model (Choi<br />

and Xiao 2010b)<br />

RECENT RESEARCH PROGRESS ON IMPACT BEHAVIORS<br />

<strong>The</strong> author and his research team are recently conducting experimental studies on the mechanical behaviour of<br />

confined concrete under impact loading and have made some progress. Using a gas gun facility, conventional<br />

CFT and the proposed CCFT specimens under high-speed impact were tested and the results validated the<br />

improved impact behavior of confined CFT [Shan et al. 2007]. Using the split Hopkinson pressure bar (SHPB)<br />

equipment, concrete filled tubes under high strain rate axial compression were studied [Xiao et al. 2009].<br />

Through the analysis of the testing data, the author defined the dynamic coefficient of the transverse<br />

confinement and found it was within the range of conventional values obtained from static tests. Most recently,<br />

CFT and CCFT stub columns were tested using a large-capacity drop-weight testing facility developed by the<br />

author and his team. Figure 8 shows the differences between the failure patterns of CFT and CCFT specimens<br />

under drop-weight impact. As exhibited in Fig.10(a), the CFT specimen had apparent local buckling in the tubes,<br />

and the core concrete infill crushing was severe. However, there was little sign of local buckling in the tube of<br />

the counterpart CCFT specimen, and the core concrete had only a few countable thin cracks identified after the<br />

removal of the steel tube.<br />

(a)<br />

(b)<br />

Figure 10 Comparisons of failure patterns: (a) CFT specimen, (b) CCFT specimen<br />

<strong>The</strong> improved impact behaviour of using transverse confinement to the CFT is also clearly demonstrated in<br />

Fig.11 by comparing the residual strain and impact energy relationships between CFT and CCFT specimens. In<br />

Fig.11, the vertical axis shows a performance index as the ratio of the residual strain and the yield strain,<br />

whereas the horizontal axis shows the impact intensity index defined as the impact energy divided by the static<br />

yield energy. As shown in Fig.9, the difference between residual strains of CFT and CCFT specimens with two<br />

layers of CFRP at relatively lower impact energy is small, however, becomes significant when subjecting to<br />

larger impact energy. It is also shown that the increase of transverse confinement by using more CFRP layers,<br />

the residual strains can be drastically reduced.<br />

-152-


Figure 11 Relationships between residual strain and impact energy for CFT specimens and CCFT specimens<br />

with different layers of CFRP.<br />

<strong>The</strong> transverse confinement is also expected to be very effective to improve the performance of concrete and<br />

composite columns subjecting to lateral impacts, which are more often occurred in bridges and other<br />

transportation facilities. In a collaboration project with <strong>University</strong> of Nebraska, concrete filled tubes were tested<br />

using the large-capacity drop-weight testing equipment developed by the author’s team [Deng et al. 2011].<br />

Several recent incidents around the worlds alerted the society of the threats posed by car bombs by terrorists.<br />

One of the most effective ways to reduce the hazards by car bombs is to set bollards around the important<br />

buildings to enlarge the distance between the point of potential explosion and the crucial buildings. Studies for<br />

the bollards generally only stay in the software analysis phase. <strong>The</strong> author’s team recently established a field<br />

laboratory of vehicular impact on structures and conducted truck collision tests of concrete filled tube (CFT)<br />

bollards, as shown in Fig.12. <strong>The</strong> total weight of truck was 6.8 tons, with the speed of about 40 kph collision.<br />

<strong>The</strong> truck was completely stopped by bollards, and the penetration distance was less than 1 m, in full<br />

compliance with the U.S. standard SD-STD-02.01A under the K4 level. <strong>The</strong> research shows that concrete filled<br />

tubular bollards system can play a good role in anti-car bomb project measures. Further studies are under way<br />

on testing the effects of confinement on impact behaviour of CFT and RC columns.<br />

CONCLUSIONS<br />

Figure 12 Real field testing of truck collision on CFT bollards<br />

Through the summary review of the development of the transversely confined structural columns, it is clearly<br />

evident that the concept has evolved to a generic design concept for achieving superior performance in structural<br />

columns. <strong>The</strong> transverse confinement can be applied using steel, reinforced concrete and fiber reinforced<br />

plastics, etc. With transverse confinement, the mechanical behaviours including impact behaviours of structural<br />

columns, such as reinforced concrete, steel or concrete filled tubular columns, etc. are shown to be greatly<br />

improved.<br />

-153-


ACKNOWLEDGMENTS<br />

<strong>The</strong> author sincerely thanks the kind advice, cooperation and supports from his previous and current mentors,<br />

colleagues and students.<br />

REFERENCES<br />

1. Tomii, M., Sakino, K., Watanabe, K., and Xiao, Y. (1985a). “Lateral Load Capacity of Reinforced Concrete<br />

Short Columns Confined by Steel Tube”, Proceedings of the International Speciality Conference on<br />

Concrete Filled Steel Tubular Structures, Harbin, China, August, 19-26.<br />

2. Tomii, M., Sakino, K., Xiao, Y., and Watanabe, K. (1985b). “Earthquake Resisting Hysteretic Behavior of<br />

Reinforced Concrete Short Columns Confined by Steel Tube”, Proceedings of the International Speciality<br />

Conference on Concrete Filled Steel Tubular Structures, Harbin, China, August, 119-125.<br />

3. Xiao, Y., Tomii, M., and Sakino, K. (1986). “Experimental Study on Design Method to Prevent Shear<br />

Failure of Reinforced Concrete Short Circular Columns by Confining in Steel Tube”, Transactions of Japan<br />

Concrete Institute, 8, 535-542.<br />

4. Tomii, M., Sakino, K., and Xiao, Y. (1987). “Ultimate moment of reinforced concrete short columns<br />

confined in steel tube”, Proceedings of the Pacific Conference on Earthquake Engineering, New Zealand, 2,<br />

11-22.<br />

5. Xiao, Y. (2004a). “Development of Tubed Concrete Column Structures”, China Civil Engineering Journal,<br />

37(4), 8-12. (in Chinese)<br />

6. Xiao, Y. (2004b). “Applications of FRP Composites in Concrete Columns”, International Journal of<br />

Advances in Structural Engineering, A Multi-Science Publication, UK, 7(4), 335-344.<br />

7. Priestley, M.J.N., Seible, F., Xiao, Y., and Verma, R. (1994a). “Steel Jacket Retrofit of Squat RC Bridge<br />

Columns for Enhanced Shear Strength - Part 1 - <strong>The</strong>oretical Considerations and Test Design”, American<br />

Concrete Institute, ACI Structural Journal, 91(4), 394-405.<br />

8. Priestley, M.J.N., Seible, F., Xiao, Y., and Verma, R. (1994b). “Steel Jacket Retrofit of Squat RC Bridge<br />

Columns for Enhanced Shear Strength - Part 2 - Experimental Results”, American Concrete Institute, ACI<br />

Structural Journal, 91(5), 537-551.<br />

9. Xiao, Y., and Wu, H. (2003). “Retrofit of Reinforced Concrete Columns Using Partially Stiffened Steel<br />

Jackets”, American Society of Civil Engineers, ASCE Journal of Structural Engineering, 129(6), 725-732.<br />

10. Xiao, Y. and Ma, R. (1997). “Seismic Retrofit of RC Circular Columns Using Prefabricated Composite<br />

Jacketing”, American Society of Civil Engineers, ASCE Journal of Structural Engineering, 123(10),<br />

1357-1364.<br />

11. Xiao, Y., H. Wu, and G.R. Martin (1999) “Prefabricated Composite Jacketing of Circular Columns for<br />

Enhanced Shear Resistance”, American Society of Civil Engineers, ASCE Journal of Structural Engineering,<br />

125(3), 255-264.<br />

12. Ma, R., and Xiao, Y. (1999). “Seismic Retrofitting and Repairing of Circular Bridge Columns with<br />

Advanced Composite Materials”, Earthquake Engineering Research Institute, EERI Spectra, 15(4),<br />

747-764.<br />

13. Ma, R., Xiao, Y., and Li, K.N. (2000). “Full-Scale Testing of A Parking Structure Column Retrofitted with<br />

Carbon Fiber Reinforced Composites”, Journal of Construction and Building Materials, Elsevier, 14(2),<br />

63-71.<br />

14. Xiao, Y., He, W.H., and Choi, K.G. (2005). “Confined Concrete Filled Tubular Columns”, American Society<br />

of Civil Engineers, ASCE Journal of Structural Engineering, 131(3), 488-497.<br />

15. Mao, X.Y. and Xiao, Y. (2006). “Seismic Behavior of Confined Square CFT Columns”, Journal of<br />

Engineering Structures, 28, 1378-1386.<br />

16. Xiao Y., H. Li, and T. Zhou (2009). “Seismic Behavior of Wide-Flange Steel Column with Confined<br />

Potential Plastic Hinge”, Journal of Constructional Steel Research, 65(4), 808–817.<br />

17. Xiao, Y. (1989). “Experimental Study and Analytical Modeling of Triaxial Compressive Behavior of<br />

Confined Concrete in Steel Tube”, Doctor of Engineering <strong>The</strong>sis, Structural Division, Department of<br />

Architecture, Faculty of Engineering, Kyushu <strong>University</strong>, Fukuoka, Japan, 200p. (in Japanese)<br />

18. Choi, K.G. and Xiao, Y. (2010a). “Analytical Model of Circular CFRP Confined Concrete-Filled Steel<br />

Tubular Columns under Axial Compression”, American Society of Civil Engineers, ASCE Journal of<br />

Composite for Construction, 14(1), 125-133.<br />

19. Choi, K.G. and Xiao, Y. (2010b). “Analytical Studies of Concrete-Filled Circular Steel Tubes under Axial<br />

Compression”, American Society of Civil Engineers, ASCE Journal of Structural Engineering, 136(5),<br />

565-573.<br />

-154-


20. Xiao, Y. and Wu, H. (2000). “Compressive Behavior of Concrete Cylinders Confined by Carbon Fiber<br />

Composite Jackets”, American Society of Civil Engineers, ASCE Journal of Materials in Civil Engineering,<br />

12(2), 139-146.<br />

21. Xiao, Y., Wu, H. (2003). “Compressive Behavior of Concrete Confined by Various Types of FRP Composite<br />

Jackets”, Journal of Reinforced Plastics and Composites, 22(13), 1187-1202.<br />

22. Shan, J.H., Chen, R., Zhang, W.X., Xiao, Y., Lu, F.Y. (2007). “Behavior of concrete filled tubes and<br />

confined concrete filled tubes under high speed impact”, International Journal of Advances in Structural<br />

Engineering, UK, 10(2), 209-218.<br />

23. Xiao, Y., Shan, J.H., Zheng, Q., Chen, B.S. and Shen, Y.L. (2009). “Experimental studies on concrete filled<br />

steel tubes under high-strain rate loading”, American Society of Civil Engineers, ASCE Journal of Materials<br />

in Civil Engineering, 21(10), 569-577.<br />

24. Deng, Y., Tuan, C.Y., Zhou, Q. and Xiao, Y. (2011). “Flexural Strength Analysis of Non-post-tensioned<br />

versus Post-tensioned Concrete-Filled Circular Steel Tubes”, Journal of Construction Steel Research, 67(2),<br />

192-202.<br />

-155-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

SPATIAL GROUND MOTION MODELLING AND ITS EFFECT ON BRIDGE<br />

RESPONSES<br />

ABSTRACT<br />

<strong>Hong</strong> Hao 1 and Kaiming Bi 2<br />

1 School of Civil and Resource Engineering, the <strong>University</strong> of Western Australia,<br />

35 Stirling Highway, Crawley WA 6009, Australia. Email: hao@civil.uwa.edu.au<br />

2 School of Civil and Resource Engineering, the <strong>University</strong> of Western Australia,<br />

35 Stirling Highway, Crawley WA 6009, Australia.<br />

Earthquake ground motions at multiple supports of large dimensional structures inevitably vary owing to the<br />

seismic wave propagation effect. It has been realized that this ground motion spatial variation affects structural<br />

responses. Many researchers have studied ground motion spatial variation and its effect on structures. For<br />

simplicity, most of previous studies assumed that the site is flat and homogeneous although it is known that<br />

varying local site conditions will further enhance ground motion spatial variations. In this paper, an approximate<br />

method is proposed to model and simulate spatially varying ground motions on surface of a non-uniform canyon<br />

site. This method takes into consideration the combined wave passage effect, coherency loss effect and local site<br />

effect, therefore leads to a more realistic modelling of spatial ground motions on non-uniform sites as compared<br />

to the some simplified approaches in previous studies. <strong>The</strong> simulated multi-component spatially varying time<br />

histories are then used as inputs to study the seismic pounding responses of a two-span simply-supported bridge<br />

structure. <strong>The</strong> influence of torsional response induced eccentric poundings is highlighted in this study based on a<br />

detailed 3D FE model. Numerical results demonstrate that the influence of local site effect on the seismic<br />

ground motion spatial variations and torsional response induced eccentric pounding between adjacent bridge<br />

spans, which are usually neglected in previous studies, are significant and should be considered in bridge<br />

response analysis.<br />

KEYWORDS<br />

Ground motion simulation, local site effect, wave passage effect, coherency loss effect, torsional response,<br />

eccentric pounding, 3D FEM.<br />

INTRODUCTION<br />

For large dimensional structures, such as long span bridges, pipelines, communication transmission systems, the<br />

ground motions at different stations during an earthquake are inevitably different, which is known as the ground<br />

motion spatial variation effect. <strong>The</strong>re are many reasons that may result in the spatial variability in seismic<br />

ground motions, e.g., the wave passage effect owing to the different arrival times of waves at different locations;<br />

the loss of coherency due to seismic waves scattering in the heterogeneous medium of the ground; the site<br />

amplification effect owing to different local soil properties. It has been proved that ground motion spatial<br />

variations have great influence on the structural responses and in some cases might even govern the structural<br />

responses (Saxena et al. 2000).<strong>The</strong> ground motion spatial variations are usually modelled by a<br />

theoretical/semi-empirical power spectral density function and a coherency loss function. Many ground motion<br />

spatial variation models have been proposed especially after the installation of the SMART-1 array in Lotung,<br />

Taiwan. Zerva and Zervas (2002) overviewed these models. It should be noted that most of these models were<br />

proposed based on the seismic data recorded from the relatively flat-lying sites. Taking different soil conditions<br />

into consideration, Der Kiureghian (1996) proposed a theoretical coherency loss function, in which the ground<br />

motion power spectral density function was represented by a site-dependent transfer function and a white noise<br />

spectrum. Typical site-dependent parameters, i.e., the central frequency and damping ratio for three generic site<br />

conditions, namely, firm, medium and soft site were proposed. <strong>The</strong> advantage of the model is that it can consider<br />

different soil properties at different support locations and it is straightforward to use. <strong>The</strong> drawback is that it can<br />

only implicitly represent the local site effects on ground motions. For example, it is well known that seismic<br />

wave will be amplified and filtered when propagating through a layered soil site. <strong>The</strong> amplifications occur at<br />

various vibration modes of the site. <strong>The</strong>refore, the energy of surface motions will concentrate at a few<br />

frequencies. <strong>The</strong> power spectral density function of the surface motion then may have multiple peaks. This<br />

-156-


phenomenon, however, cannot be considered in Der Kiureghian’s model. It does not explicitly model the<br />

influences of the sites effects on ground motion spatial variations.<br />

<strong>The</strong> ground motion power spectral density functions and spatial variation models can be used directly as inputs<br />

at multiple supports of structures in spectral analysis of structural responses. This approach, however, is usually<br />

applied to relatively simple structural models and for linear response of the structures owing to its complexity.<br />

For complex structural systems and for nonlinear seismic response analysis, only the deterministic solution can<br />

be evaluated with sufficient accuracy. In this case, the generation of artificial seismic ground motions is required.<br />

Many methods are available to generate artificial spatially correlated time histories at different structural<br />

supports. Hao et al. (1989) presented a method of generating spatially varying time histories at different<br />

locations on ground surface based on the assumption that all the spatially varying ground motions have the same<br />

intensity, i.e., the same power spectral density or response spectrum. <strong>The</strong> variation of the spatial ground motions<br />

is modelled by an empirical coherency loss function and a phase delay depending on a constant apparent wave<br />

propagation velocity. If the considered site is flat with uniform soil properties, the uniform ground motion<br />

intensity assumption for spatial ground motions in the site is reasonable. However, for a canyon site or a site<br />

with varying soil properties, because local site conditions affect the wave propagation hence the ground motion<br />

intensity and frequency contents as discussed above, the uniform ground motion power spectral density<br />

assumption is no longer valid. Deodatis (1996) developed a method to simulate spatial ground motions with<br />

different power spectral densities at different locations. <strong>The</strong> method is based on a spectral representation<br />

algorithm (Shinozuka 1972; Shinozuka and Jan 1972) to generate sample functions of a non-stationary,<br />

multivariate stochastic process with evolutionary power spectrum. Similar to the Der Kiureghian (1996) model,<br />

the considered varying spectral densities are filtered white noise functions with different central frequency and<br />

damping ratio. This method thus can only approximately represent local site effects on ground motions.<br />

Moreover, trying to establish an analytical expression for a realistic ground motion evolutionary power spectrum<br />

related to the local site conditions is quite difficult since very limited information is available on the spectral<br />

characteristics of propagating seismic waves (Shinozuka and Deodatis 1988).<br />

<strong>The</strong> first part of this paper extends the work by Der Kiureghian (1996) by using the 1D wave propagation theory<br />

(Wolf 1985) to more realistically model the influence of local site conditions on seismic waves. <strong>The</strong> spectral<br />

representation method (Shinozuka 1972; Shinozuka and Jan 1972) and 1D wave propagation theory are<br />

combined to derive the power spectral density functions of the spatially varying ground motions on surface of a<br />

canyon site with multiple soil layers. <strong>The</strong> ground motion spatial variations are modelled in two steps: firstly, the<br />

spatially varying base rock ground motions are assumed to consist of out-of-plane SH wave or in-plane<br />

combined P and SV waves and propagate into the layered soil site with an assumed incident angle. <strong>The</strong> spatial<br />

base rock motions are assumed to have the same intensity and frequency contents and are modelled by the<br />

filtered Tajimi-Kanai power spectral density function (Tajimi 1960). <strong>The</strong> spatial variation effect is modelled by a<br />

theoretical coherency loss function (Sobczky 1991). <strong>The</strong> surface motions of a canyon site with multiple soil<br />

layers are derived based on the deterministic 1D wave propagation theory. <strong>The</strong> auto power spectral density<br />

functions of ground motions at various points on ground surface and the cross power spectral density functions<br />

between ground motions at any two points are derived by neglecting the wave scattering on the uneven canyon<br />

surface. <strong>The</strong> spectral representation method is then used to generate spatially varying ground motion time<br />

histories. Compared to the work by Deodatis (1996), in this study the power spectral density functions at<br />

different locations of a canyon site are derived based on the 1D wave propagation theory, which directly relates<br />

the local soil conditions and base rock motion characteristics with the surface ground motions, thus local site<br />

effect can be realistically considered. <strong>The</strong> current approach also allows for a consideration of different incoming<br />

wave types and incident angles to the soil site, which have great influence on the surface motions. <strong>The</strong> proposed<br />

approach can be used to simulate ground motion time histories at an uneven site with known non-uniform site<br />

conditions.<br />

For bridge structures with conventional expansion joints, a complete avoidance of pounding between bridge<br />

decks during strong earthquakes is often impossible since the separation gap of an expansion joint is usually a<br />

few centimetres to ensure a smooth traffic flow. <strong>The</strong>refore, pounding damages of adjacent bridge structures have<br />

always been observed in almost all the previous major earthquakes. Pounding is an extremely complex<br />

phenomenon involving damage due to plastic deformation, local cracking or crushing, fracturing due to impact,<br />

and friction when two adjacent bridge decks are in contact with each other. To simplify the analysis, many<br />

researchers modelled a bridge girder as a lumped mass (e.g. Malhotra 1998; Jankowski et al. 1998;<br />

Ruangrassamee and Kawashima 2001; DesRoches and Muthukumar 2002; Chouw and Hao 2005, 2008) or<br />

beam-column element frame (e.g. Jankowski et al. 2000; Chouw et al. 2006). Based on theses simplified<br />

lumped mass model or beam-column element model, only 1D point to point pounding, usually in the axial<br />

direction of the structures, can be considered. In a real bridge structure under seismic loading, pounding could<br />

-157-


take place along the entire surfaces of the adjacent structures. Moreover, it was observed from previous<br />

earthquakes that most poundings actually occurred at corners of adjacent bridge decks, this is because torsional<br />

responses of the adjacent decks induced by spatially varying transverse ground motions at multiple bridge<br />

supports resulted in eccentric poundings. To more realistically model the pounding phenomenon between<br />

adjacent bridge structures, a detailed 3D finite element analysis is necessary. However, previous studies of<br />

seismic pounding responses of bridge structures based on the detailed 3D FEM are limited. To the best<br />

knowledge of the authors, the only two studies in which bridge girders were modelled as 3D finite elements<br />

were reported by Zanardo et al. (2002) and Zhu et al. (2002). <strong>The</strong>se two studies, however, either neglected the<br />

surface to surface and eccentric poundings (Zanardo et al. 2002) or the algorithm for searching for the pounding<br />

locations is quite complicated (Zhu et al. 2002).<br />

Pounding between adjacent bridge decks occurs because of large relative displacement responses. Ground<br />

motion spatial variation, besides differences in vibration properties of adjacent bridge structures, is a source of<br />

relative displacement responses. Owing to the difficulty in modelling ground motion spatial variation, many<br />

studies assumed uniform excitations (Malhotra 1998; Ruangrassamee and Kawashima 2001; DesRoches and<br />

Muthukumar 2002; Zhu et al. 2002) or assumed variation was caused by wave passage effect only (Jankowski et<br />

al. 1998, 2000). Only a few studies considered the combined wave passage effect and coherency loss effect in<br />

analyzing relative displacement responses of adjacent bridge structures (Chouw and Hao 2005, 2008; Chouw et<br />

al. 2006; Zanardo et al. 2002). It should be noted that all these studies mentioned above assumed that the<br />

analyzed structures locate on a flat-lying site, the influence of local site effect, which further intensifies ground<br />

motion spatial variation at multiple structural supports, are neglected. Studies revealed that local site effect not<br />

only causes further phase difference (Der Kiureghian 1996; Bi et al. 2010), but also affects the coherency loss<br />

between spatial ground motions (Bi and Hao 2010). <strong>The</strong>se differences will significantly affect the structural<br />

responses. Consequently, neglecting local soil effect on the spatial ground motion variations at multiple supports<br />

of a bridge structure crossing a canyon site may lead to inaccurate estimation of bridge responses.<br />

In the last part of this study, seismic pounding responses between the abutment and the adjacent bridge deck and<br />

between two adjacent bridge decks of a two-span simply-supported bridge located on a canyon site are<br />

investigated. <strong>The</strong> three-dimensional spatially varying ground motions at different supports of the bridge<br />

structure are stochastically simulated as inputs based on the methodology proposed in the first section of the<br />

present paper. A detailed 3D finite element model of the bridge is constructed in ANSYS, and then LS-DYNA is<br />

employed to calculate the structural responses. <strong>The</strong> influences of pounding effect and local soil conditions on<br />

ground motion spatial variation and on structural responses are investigated in detail.<br />

SPATIALLY VARYING GROUND MOTION SIMULATION<br />

Ground Motion Generation<br />

Consider a canyon site with horizontally extended multiple soil layers resting on an elastic half-space as shown<br />

in Figure 1, in which h m , G m , ρ m , ξ m and υ m are the depth, shear modulus, mass density, damping ratio<br />

and Poisson’s ratio of layer m. <strong>The</strong> spatially varying base rock motions are assumed to consist of out-of-plane<br />

SH wave or in-plane combined P and SV waves and propagating into the layered soil site with an assumed<br />

incident angle. <strong>The</strong> incident motions at different locations on the base rock are assumed to have the same power<br />

spectral density, and are modelled by a filtered Tajimi-Kanai power spectral density function. <strong>The</strong> spatial<br />

variation of ground motions at base rock is modelled by an empirical coherency function for spatial ground<br />

motions on a flat site. <strong>The</strong> cross power spectral density functions of surface motions at n locations of the layered<br />

site can be written as:<br />

⎡ S11(<br />

ω)<br />

S12<br />

( iω)<br />

⋅⋅⋅ S1n<br />

( iω)<br />

⎤<br />

⎢<br />

⎥<br />

⎢<br />

S21(<br />

iω)<br />

S22<br />

( ω)<br />

⋅⋅⋅ S2n<br />

( iω)<br />

S ( iω)<br />

=<br />

⎥<br />

(1)<br />

⎢ ⋅⋅⋅ ⋅⋅⋅ ⋅⋅⋅ ⋅⋅⋅ ⎥<br />

⎢<br />

⎥<br />

⎣Sn1(<br />

iω)<br />

Sn2<br />

( iω)<br />

⋅⋅⋅ Snn<br />

( ω)<br />

⎦<br />

where<br />

S<br />

S<br />

jj<br />

jk<br />

( ω)<br />

= H ( iω)<br />

S<br />

*<br />

k<br />

( ω)<br />

= H ( iω)<br />

H<br />

j<br />

j<br />

2<br />

g<br />

( ω)<br />

j = 1,2, L,<br />

n<br />

( iω) S ( ω)<br />

γ ( d , iω)<br />

j = 1,2, L,<br />

n<br />

g<br />

are the auto and cross power spectral density function respectively. In which S (ω)<br />

is the ground motion power<br />

'<br />

spectral density on the base rock; γ j k d j k , iω)<br />

is the coherency function between location j and k ' on the<br />

' '( ' '<br />

j'<br />

k'<br />

j'<br />

k'<br />

g<br />

(2)<br />

-158-


ase rock; H j ( iω)<br />

, H k ( iω)<br />

are the site transfer function at locations j and k on the ground surface, which can<br />

be formulated based on one-dimensional wave propagation theory (Wolf 1985). Because the transfer function is<br />

directly related to the site conditions, the current approach yields more realistic representation of the influences<br />

of non-uniform site conditions on surface ground motions as compared to that proposed by Der Kiureghian<br />

(1996) by using different central frequencies and damping ratios in Tajimi-Kanai power spectral density<br />

function to model the influences of non-uniform site conditions. However, it should be noted that the scattering<br />

of seismic waves by the uneven canyon, which is a 3D wave propagation problem, is not considered when<br />

formulating Eq.2. Modelling 3D wave propagation in a site is beyond the scope of the present study.<br />

Decomposing the Hermitian, positive definite matrix S ( iω)<br />

into the multiplication of a complex lower<br />

triangular matrix L ( iω)<br />

and its Hermitian L H ( iω)<br />

H<br />

S( iω)<br />

= L(<br />

iω)<br />

L ( iω)<br />

(3)<br />

the stationary time series u j ( t),<br />

j = 1,2, L,<br />

n , can be simulated in the time domain directly (Hao et al. 1989)<br />

where<br />

j<br />

j<br />

N<br />

u ( t)<br />

= ∑∑A<br />

A<br />

jm<br />

m= 1n=<br />

1<br />

jm<br />

( ω)<br />

=<br />

jm<br />

β ( ω)<br />

= tan<br />

( ω )cos[ ω t + β ( ω ) + ϕ<br />

n<br />

4Δω<br />

L<br />

n<br />

( iω),<br />

jm<br />

n<br />

mn<br />

0 ≤ ω ≤ ω<br />

( ω )]<br />

n<br />

jm<br />

N<br />

Im[ L<br />

jm(<br />

iω)]<br />

−1 (5)<br />

(<br />

), 0 ≤ ω ≤ ωN<br />

Re[ L<br />

jm(<br />

iω)]<br />

are the amplitudes and phase angles of the simulated time histories which ensure the spectra of the simulated<br />

time histories compatible with those given in Eq.2; ϕ<br />

mn( ωn)<br />

is the random phase angles uniformly distributed<br />

over the range of [ 0,2π ] , ϕ<br />

mn and ϕ rs should be statistically independent unless m = r and n = s ; ω<br />

N<br />

represents an upper cut-off frequency beyond which the elements of the cross power spectral density matrix<br />

given in Eq.1 is assumed to be zero.<br />

<strong>The</strong> generated time series by Eq.4 are stationary processes. In order to obtain the non-stationary time histories,<br />

an envelope function ζ (t)<br />

is applied to u j (t)<br />

. <strong>The</strong> non-stationary time histories at different locations are then<br />

obtained by<br />

f j ( t)<br />

= ζ ( t)<br />

u j ( t),<br />

j = 1,2,...,<br />

n<br />

(6)<br />

Numerical Example<br />

An alluvium canyon site with multiple soil layers shown in Figure 1 is selected as an example, the mean value<br />

of the corresponding soil properties of each soil layer and the base rock are also given in the Figure.<br />

(4)<br />

Figure 1 A canyon site with multiple soil layers (not to scale)<br />

<strong>The</strong> motions on the base rock are assumed to be the same and have the following form:<br />

4<br />

2 2 2<br />

ω<br />

1+<br />

4ξ<br />

gωgω<br />

S g ( ω)<br />

= H P ( ω)<br />

S0<br />

( ω)<br />

=<br />

Γ<br />

2 2 2<br />

2 2 2 2 2 2 2<br />

(7)<br />

( ω f −ω<br />

) + (2ω<br />

f ωξ f ) ( ωg<br />

−ω<br />

) + 4ξ<br />

gωgω<br />

In which H P (ω ) is a high pass filter, S 0 ( ω)<br />

is the Tajimi-Kanai power spectral density function, ω g and<br />

-159-


ξ g are the central frequency and damping ratio of the Tajimi-Kanai power spectral density function,<br />

ω f and<br />

ξ f are the central frequency and damping ratio of the high pass filter. In the analysis, the horizontal<br />

out-of-plane motion is assumed to consist of SH wave only, while the in-plane horizontal and vertical motion<br />

are assumed to be combined P and SV wave. <strong>The</strong> parameters of the horizontal motion are assumed as<br />

ω g = 10πrad<br />

/ s , ξ g = 0. 6 , ω f = 0. 5π<br />

, ξ f = 0. 6 and 2 / 3<br />

Γ = 0.0034m<br />

s . <strong>The</strong>se parameters correspond to a<br />

ground motion time history with duration T=20s and peak ground acceleration (PGA) 0 .2g<br />

and peak ground<br />

displacement (PGD) 0.082m based on the standard random vibration method (Der Kiureghian 1980). <strong>The</strong><br />

vertical motion on the base rock is also modelled with the same filtered Tajimi-Kanai power spectral density<br />

function, but the amplitude is assumed to be 2/3 of the horizontal component.<br />

<strong>The</strong> Sobczyk model (1991) is selected to describe the coherency loss between the ground motions at points<br />

and k ' ( j ≠ k ) at the base rock:<br />

γ<br />

'<br />

j k<br />

iω)<br />

= γ ' ' ( iω)<br />

exp( −iωd<br />

' ' cosα<br />

/ v ) = exp( −βωd<br />

' ' / v ) ⋅exp(<br />

−iωd<br />

' cosα<br />

/ v ) (8)<br />

' ( '<br />

j k<br />

j k<br />

app<br />

in which, γ ' '<br />

j k<br />

( iω)<br />

is the lagged coherency loss, β is a coefficient which reflects the level of coherency loss,<br />

β = 0.0005 is used in the present paper, which represents highly correlated motions; d '<br />

is the distance<br />

between the points j ' and k ' , and d ' ' = d ' ' = 100m<br />

is assumed; α is the incident angle of the incoming<br />

12 23<br />

wave to the site, and is assumed to be 60°; v app is the apparent wave velocity at the base rock, which is<br />

1768m/s according to the base rock property and the specified incident angle.<br />

In the simulation, the sampling frequency and the upper cut-off frequency are set to be 100Hz and ω N = 25Hz<br />

,<br />

and the time duration is assumed to be T=20s. To improve the computational efficiency, the ground motions are<br />

generated in the frequency domain by using the FFT technique, and 2048 sampling points are used.<br />

Assuming the incoming motions at the base rock consist of SH wave with an incident angle of<br />

o<br />

α = 60 , the<br />

horizontal out-of-plane acceleration and displacement time histories on the ground surface are simulated and<br />

shown in Figure 2. It is obvious that the site amplification effects alter the frequency contents and increase the<br />

amplitudes of the incoming wave. Different wave paths result in different site amplification effect. For the given<br />

example, the PGAs and PGDs on the ground surface reach 4.32, 6.53, 3.31m/s2 and 0.0565, 0.0540, 0.0712m at<br />

the three different locations as shown in Figure 2. As compared with the motions at the base rock (the PGAs on<br />

the base rock are 2.27, 2.21 and 2.23 m/s2 respectively for the three sites, which are not shown herein owing to<br />

the page limit), site 2 significantly amplifies the horizontal out-of-plane ground acceleration. This is because the<br />

fundamental vibration frequency of site 2 is 4.85Hz, which is very close to the central frequency of the filtered<br />

Tajimi-Kanai power spectral density function of ground motions at the base rock, so resonance occurs. Sites 1<br />

and 3 also amplify the base rock motion, but with a less extent. It is interesting to find that, though the site<br />

amplifies the PGA of surface motions, it is not necessarily result in larger PGD (the PGD on the base rock are<br />

0.0861, 0.0857 and 0.0839m respectively). For the given example, the PGDs even decrease 34%, 37% and 15%<br />

respectively. This can be attributed to the fact that the local soil layers also filter the frequency content of the<br />

incoming waves. Although PGA of site 2 is the largest, the PGD is the smallest because site 2 is the stiffest<br />

among the three sites. On contrary, PGA of site 3 is the smallest, but PGD is the largest because site 3 is the<br />

softest. In general, the softer is the site, the larger is the PGD. Figure 3 shows the comparisons of the simulated<br />

power spectral densities with the theoretical values, good agreements are observed.<br />

2<br />

j k<br />

app<br />

j k<br />

j ' k<br />

app<br />

j '<br />

Figure 2 Generated horizontal out-of-plane motions on ground surface: (a) acceleration, and (b) displacement<br />

-160-


For the coherency loss function between surface motions at a canyon site, the analysis based on the recorded<br />

seismic data showed larger variability than that on the flat-lying sites (Somerville et al. 1991; Liao and Zerva<br />

2007). Further studies revealed that local site effect not only causes phase difference of the coherency function<br />

(Der Kiureghian 1996), but also affects its modulus (Lou and Zerva 2005; Bi and Hao 2010). Figure 4 shows the<br />

comparison of the lagged coherency loss functions of the base rock motions (solid line) with those of the<br />

simulated surface motions (dashed line) in Figure 2. As shown, the lagged coherency loss on the base rock is the<br />

upper bound of that on the ground surface. <strong>The</strong>se results are consistent with those obtained from recorded<br />

surface ground motions (Somerville et al. 1991; Liao and Zerva 2007). <strong>The</strong>y indicate wave propagation through<br />

non-uniform paths cause further coherency loss between spatial ground motions. More detailed discussions of<br />

the influences of local site conditions on surface motion coherency loss can be found in Bi and Hao (2010).<br />

Figure 3 Comparison of power spectral density of the generated horizontal out-of-plane acceleration on ground<br />

surface with the respective theoretical power spectral density<br />

Figure 4 Comparison of the coherency loss functions between base rock motions (solid line) and those between<br />

surface motions (dashed line)<br />

<strong>The</strong> in-plane surface motions can be simulated similarly by assuming incident wave consists of combined P and<br />

SV waves. Owing to the page limit, they are not shown here, good agreements can also be observed between the<br />

simulated ground motions and the specific theoretical models.<br />

3D FEM ANALYSIS OF POUNDING RESPONSE OF BRIDGE STRUCTURES<br />

Bridge Model<br />

<strong>The</strong> seismic pounding responses of a two-span simply-supported bridge structure on a canyon site as shown in<br />

Figure 5 are investigated in this Section. Figure 5(a) shows the elevation view of the bridge. <strong>The</strong> box-section<br />

bridge girders with the cross section shown in Figure 5(b) have the same span length of 50m. <strong>The</strong> Young’s<br />

modulus and density of the bridge girders are 3.45×10 10 Pa and 2500kg/m 3 , respectively. <strong>The</strong> L-type abutment is<br />

8.1m long in the transverse direction and its cross section is shown in Figure 5(c). <strong>The</strong> height of the rectangular<br />

central pier is 20m, with the cross section shown in Figure 5(d). <strong>The</strong> materials for the two abutments and the<br />

pier are the same, with Young’s modulus and density of 3.0×10 10 Pa and 2400 kg/m 3 , respectively. <strong>The</strong> two<br />

bridge girders are supported by 8 high-damping rubber bearings. <strong>The</strong> horizontal effective stiffness and<br />

equivalent damping ratio of a bearing are 2.33×10 7 N/m and 0.14 respectively (Jankowski et al. 1998; Zanardo<br />

et al. 2002). <strong>The</strong> stiffness of the bearing in the vertical direction is much larger than those in the horizontal<br />

directions, and is assumed to be 1.87×10 10 N/m (Zanardo et al. 2002). To allow for contraction and expansion of<br />

the bridge decks from creep, shrinkage, temperature fluctuations and traffic without generating constraint forces<br />

in the structure, a 5cm expansion joint is introduced between the abutments and the bridge decks and between<br />

-161-


the adjacent bridge decks.<br />

<strong>The</strong> bridge locates on a canyon site, consisting of horizontally extended soil layers on a half-space (base rock).<br />

<strong>The</strong> foundations of the bridge are assumed rigidly fixed to the ground surface and soil-structure interaction (SSI)<br />

is not involved. Points A, B and C are the three bridge support locations on the ground surface, the<br />

corresponding points on the base rock are A’, B’ and C’.<br />

Figure 5 (a) Elevation view of the bridge, (b) Cross-section of the bridge girder, (c) Cross-section of the<br />

abutment, and (d) Cross-section of the pier (unit: cm)<br />

<strong>The</strong> 3D finite element model of the bridge is constructed by using the finite element code ANSYS. <strong>The</strong> bridge<br />

girders, abutments and pier are modelled by eight-node solid elements. <strong>The</strong> bearings are modelled by the<br />

spring-dashpot elements. <strong>The</strong> detailed geometric characteristics in Figure 5 and the material properties have<br />

been implemented in the model. 85428 nodes and 59824 elements are included in the finite element model. <strong>The</strong><br />

first four vibration frequencies of the bridge are 1.081, 1.138, 1.254 and 1.313Hz for the in-phase longitudinal (x<br />

direction), in-phase transverse (z direction), out-of-phase transverse and out-of-phase longitudinal vibrations,<br />

respectively. Rayleigh damping of 5% is assumed in the analysis.<br />

Poundings may occur between the abutment and the bridge girder and/or between two adjacent bridge girders.<br />

To realistically consider the poundings between entire surfaces of adjacent structures, the contact type of<br />

*CONTACT_AUTOMATIC_SURFACE_TO_SURFACE in LS-DYNA is employed in the simulations. <strong>The</strong><br />

Coulomb friction coefficient of 0.5 is assumed in the analysis.<br />

In the present study, only one soil layer is considered and soil conditions for the three sites are assumed to be the<br />

same with the corresponding parameters for the base rock and soil shown in Table 1. Soil depths for the three<br />

sites are 48.6, 30 and 48.6m respectively. <strong>The</strong> horizontal in-plane, horizontal out-of-plane and vertical in-plane<br />

spatially varying ground motions at different supports of the bridge are stochastically simulated based on<br />

method proposed in the first part of present paper, and are applied simultaneously to the longitudinal, transverse<br />

and vertical directions of the bridge. Owing to the page limit, the simulated multi-component spatially varying<br />

ground motions are not plotted.<br />

Table 1 Parameters for local site conditions<br />

Type Density (kg/m 3 ) Shear modulus(MPa) Damping ratio Poisson’s ratio<br />

Base rock 2500 1800 0.05 0.33<br />

Soil 2000 320 0.05 0.4<br />

-162-


Numerical Results<br />

<strong>The</strong> earthquake-induced pounding responses of the bridge as shown in Figure 5 are investigated. For<br />

comparison purpose, the case without pounding is also considered, which is achieved by assuming the<br />

separation gaps between the abutments and the girders and between two adjacent girders are large enough so<br />

that pounding phenomenon can be completely precluded and the structure vibrates freely. To obtain a less biased<br />

estimation, 3 sets of spatially varying ground motions are used. <strong>The</strong> results presented here are the<br />

assembled-mean responses.<br />

Poundings may occur between the abutments and bridge girders and between two bridge girders as mentioned<br />

above. Though the bridge considered in the present paper is a symmetrical structure, the responses of different<br />

parts will be different owing to ground motion spatial variations and pounding effects. To obtain a general idea<br />

of the earthquake-induced structural responses, the following 12 nodes as shown in Figure 6 are selected to<br />

examine the results. <strong>The</strong> peak responses, which are mostly concerned in engineering practice, are listed in Table<br />

2, where the torsional responses is illustrated by the relative longitudinal displacements between the inside node<br />

and the corresponding outside node at the same section.<br />

1<br />

3<br />

Left abutment<br />

4<br />

Left girder<br />

2<br />

Left girder<br />

5<br />

7<br />

6<br />

8<br />

Right girder<br />

11<br />

12<br />

Right girder<br />

Right abutment<br />

9<br />

10<br />

Pier<br />

Figure 6 Different nodes examined in the present study.<br />

Table 2 Influence of pounding effect on the mean peak displacements (m)<br />

Node<br />

Longitudinal Transverse Vertical Torsional<br />

with without with without with without with without<br />

1 0.181 0.182 0.135 0.135 0.072 0.072<br />

3 0.182 0.182 0.135 0.135 0.072 0.072<br />

0.0045 0.0002<br />

2 0.210 0.274 0.184 0.207 0.098 0.186<br />

4 0.217 0.287 0.184 0.206 0.102 0.161<br />

0.0418 0.0373<br />

5 0.208 0.266 0.270 0.263 0.100 0.201<br />

7 0.209 0.275 0.269 0.262 0.104 0.129<br />

0.0354 0.0304<br />

6 0.233 0.272 0.241 0.293 0.136 0.133<br />

8 0.237 0.261 0.240 0.292 0.138 0.143<br />

0.0371 0.0321<br />

9 0.226 0.267 0.179 0.200 0.096 0.109<br />

11 0.220 0.255 0.178 0.197 0.112 0.107<br />

0.0454 0.0371<br />

10 0.179 0.179 0.119 0.119 0.067 0.067<br />

12 0.179 0.179 0.119 0.119 0.067 0.067<br />

0.0024 0.0001<br />

It can be noted in Table 2 that the responses of the abutments are almost unaffected by pounding. This is because<br />

the abutment is quite rigid compared to the adjacent girder. <strong>The</strong> influence of collisions on the girder response is,<br />

however, significant. Poundings usually result in smaller displacements in the longitudinal direction. This is<br />

because the rigid abutment acts as a constraint to the flexible girder. For the displacements in the transverse and<br />

vertical directions, smaller values are also observed when poundings are involved. This may be because of the<br />

friction forces between the adjacent surfaces during poundings, which reduce the displacement responses of the<br />

bridge structures in the transverse and vertical directions. Pounding effects, however, result in larger torsional<br />

responses as shown in the last two columns of Table 2. This is because eccentric poundings induced by spatially<br />

varying ground motions generate large eccentric impact forces that enhance the torsional responses.<br />

-163-


To examine the occurrence of poundings, the longitudinal displacements of nodes 1 and 2 and nodes 3 and 4 are<br />

plotted in the same figure with the displacements of nodes 1 and 3 shifted by the initial gap of 5cm. Thus, in the<br />

figure, the instants when the displacements of the two adjacent points coinciding with each other indicate the<br />

occurrence of poundings. As shown in Figure 7(a), node 1 and node 2 come into contacts 15 times, at the time<br />

instants 3.26, 5.29, 6.29, 6.68, 7.30, 7.72, 8.20, 8.63, 9.13, 9.66, 11.13, 11.89, 12.44, 13.70 and 14.26s. Whereas<br />

between nodes 3 and 4 as shown in Figure 7(b), the poundings at 6.29, 11.89 and 12.44s do not occur, but two<br />

more collisions can be observed at 3.76 and 13.20s. Since these points locate at the opposite corners of the<br />

bridge deck cross section, pounding at these points occurring simultaneously implies the entire cross sections<br />

are in contact, i.e. surface to surface pounding occurs. Otherwise, they are torsional response induced eccentric<br />

poundings. In this example, pounding occurring at 6.29, 11.89 and 12.44s are eccentric poundings between<br />

nodes 1 and 2, and those at 3.76 and 13.20s are eccentric poundings between nodes 3 and 4. Torsional response<br />

induced eccentric poundings between other corner points shown in Figure 6 are also observed. Owing to page<br />

limit, they are not shown here. <strong>The</strong>se observations indicate that if a 3D model with tri-axial ground motion<br />

inputs is considered, more number of poundings will be observed than the lumped mass and 2D beam-column<br />

element model because the two letter models cannot capture the possible eccentric poundings induced by<br />

torsional responses.<br />

Initial gap=5cm<br />

(a) with pounding<br />

(b) without pounding<br />

Initial gap=5cm<br />

Figure 7 Longitudinal displacements of different nodes with<br />

Pounding effects<br />

Figure 8 Stresses in the longitudinal direction<br />

at left abutment when t=6.27s<br />

By using the traditional lumped-mass model or beam-column element model, the stress on the entire contact<br />

surface will be the same. However, the use of 3D finite element model allows a more detailed prediction of the<br />

largest stresses and its locations, where earthquake-induced damage may occur. Figure 8 shows the stress<br />

distributions in the longitudinal direction at the left abutment with and without pounding effect at t=6.27s. This<br />

time instant is selected because the resultant pounding force reaches the maximum value. As shown in Figure<br />

8(a), the maximum longitudinal stress at the bottom outside corner of the bridge girder reaches around 90MPa.<br />

This value is much larger than the compressive strength of concrete, which is usually 30-65MPa for impact<br />

loading (Bischoff and Perry 2005), thus concrete damage are expected. <strong>The</strong>se results are consistent with the<br />

observations in the past major earthquakes, in which the damage around the corners of the structure were<br />

usually the most serious. Compared with Figure 8(b), it is obvious that pounding effect significantly increases<br />

the intensity of longitudinal stresses.<br />

Conclusions<br />

A method to model and simulate spatially varying earthquake ground motion time histories at sites with<br />

non-uniform conditions is proposed. This method takes into consideration the local site effect on ground motion<br />

amplification and spatial variations. It is believed leading to a more realistic modelling of spatial ground<br />

motions on non-uniform sites as compared to the common assumption of uniform ground motion intensity in<br />

most previous studies. <strong>The</strong> simulated time histories can be used as inputs to multiple supports of long-span<br />

structures on non-uniform sites in engineering practice.<br />

Earthquake-induced pounding responses between adjacent components of a two-span simply-supported bridge<br />

structure located at a canyon site are studied based on a detailed 3D FE model. It is found that a detailed 3D FE<br />

model gives more accurate predictions of the earthquake-induced pounding responses of bridge structures since<br />

torsional vibrations of the structure, which play an important role in the overall structure response, can be<br />

modelled. With a 3D model, the potential damage locations in the structure can be identified. Pounding effects<br />

usually results in smaller longitudinal, transverse and vertical displacements while lead to larger torsional<br />

responses.<br />

-164-


REFERENCES<br />

Bi, K., Hao, H. and Ren, W. (2010). “Response of a frame structure on a canyon site to spatially varying ground<br />

motions”, Structural Engineering and Mechanics, 36(1), 111-127.<br />

Bi, K., and Hao, H. (2010). “Influence of irregular topography and random soil properties on coherency loss of<br />

spatial seismic ground motions”, Earthquake Engineering and Structural Dynamics (published online).<br />

Bischoff, P.H. and Perry, S.H. (1995). “Impact behaviour of plain concrete loaded in uniaxial compression”,<br />

Journal of Engineering Mechanics, 121(6), 685-693.<br />

Chouw, N. and Hao, H. (2005). “Study of SSI and non-uniform ground motion effects on pounding between<br />

bridge girders”, Soil Dynamics and Earthquake Engineering, 23, 717-728.<br />

Chouw, N. and Hao, H. (2008). “Significance of SSI and non-uniform near-fault ground motions in bridge<br />

response I: Effect on response with conventional expansion joint”, Engineering Structures, 30(1), 141-153.<br />

Chouw, N., Hao, H. and Su, H. (2006). “Multi-sided pounding response of bridge structures with non-linear<br />

bearings to spatially varying ground excitation”, Advances in Structural Engineering, 9(1), 55-66.<br />

Deodatis, G. (1996). “Non-stationary stochastic vector processes: seismic ground motion applications”,<br />

Probabilistic Engineering Mechanics, 11(3), 149-167.<br />

Der Kiureghian, A. (1980). “Structural response to stationary excitation”, Journal of Engineering Mechanics,<br />

106, 1195-1213.<br />

Der Kiureghian, A. (1996). “A coherency model for spatially varying ground motions”, Earthquake Engineering<br />

and Structural Dynamics, 25(1), 99-111.<br />

DesRoches, R. and Muthukumar, S. (2002). “Effect of pounding and restrainers on seismic response of<br />

multi-frame bridges”, Journal of Structural Engineering (ASCE), 128(7), 860-869.<br />

Hao, H., Oliveira, C.S. and Penzien, J. (1989). “Multiple-station ground motion processing and simulation based<br />

on SMART-1 Array data”, Nuclear Engineering and Design, 111, 293-310.<br />

Jankowski, R., Wilde, K. and Fujino, Y. (1998). “Pounding of superstructure segments in isolated elevated<br />

bridge during earthquakes”, Earthquake Engineering and Structural Dynamics, 27, 487-502.<br />

Jankowski, R., Wilde, K. and Fujino, Y. (2000). “Reduction of pounding effects in elevated bridges during<br />

earthquakes”, Earthquake Engineering and Structural Dynamics, 29, 195-212.;<br />

Liao, S., Zerva, A. and Stephenson, W.R. (2007). “Seismic spatial coherency at a site with irregular subsurface<br />

topography”, Proceedings of Sessions of Geo-Denver, Geotechnical Special Publication, 1-10.<br />

Lou, L. and Zerva, A. (2005). “Effects of spatially variable ground motions on the seismic response of a skewed,<br />

multi-span, RC-highway bridge”, Soil Dynamics and Earthquake Engineering, 25, 729-740.<br />

Malhotra, P.K. (1998). “Dynamics of seismic pounding at expansion joints of concrete bridges”, Journal of<br />

Engineering Mechanic, 124(7), 794-802.<br />

Ruangrassamee, A. and Kawashima, K. (2001). “Relative displacement response spectra with pounding effect”,<br />

Earthquake Engineering and Structural Dynamics, 30(10), 1511-1538.<br />

Saxena, V., Deodatis, G. and Shinozuka, M. (2000). “Effect of spatial variation of earthquake ground motion on<br />

the nonlinear dynamic response of highway bridges”, Proceeding of 12th World Conference on Earthquake<br />

Engineering, Auckland, New Zealand.<br />

Shinozuka, M.(1972). “Monte Carlo solution of structural dynamics”, Computers and Structures, 2, 855-874.<br />

Shinozuka, M. and Jan, C.M. (1972). “Digital simulation of random processes and its applications”, Journal of<br />

Sound and Vibration, 25(1), 111-128.<br />

Shinozuka, M. and Deodatis, G. (1988). “Stochastic process models for earthquake ground motion”,<br />

Probabilistic Engineering Mechanics, 3(3), 114-123.<br />

Sobczky, K. (1991). Stochastic Wave Propagation, Kluwer Academic Publishers, Netherlands.<br />

Somerville, P.G., McLaren, J.P., Sen, M.K. and Helmberger, D.V. (1991). “<strong>The</strong> influence of site conditions on<br />

the spatial incoherence of ground motions”, Structural Safety, 10(1), 1-13.<br />

Tajimi, H. (1960). “A statistical method of determining the maximum response of a building structure during an<br />

earthquake”, Proceeding of 2nd World Conference on Earthquake Engineering, Tokyo, Japan, 781-796.<br />

Wolf, J.P. (1985). Dynamic Soil-structure Interaction, Englewood Cliffs, NJ, USA.<br />

Zanardo, G., Hao, H. and Modena, C. (2002). “Seismic response of multi-span simply supported bridges to<br />

spatially varying earthquake ground motion”, Earthquake Engineering and Structural Dynamics, 31(6),<br />

1325-1345.<br />

Zerva, A., and Zervas, V. (2002). “Spatial variation of seismic ground motions: An overview”, Applied<br />

Mechanics Reviews, 56(3), 271-297.<br />

Zhu, P., Abe, M. and Fujino, Y. (2002). “Modelling three-dimensional non-linear seismic performance of<br />

elevated bridges with emphasis on pounding of girders”, Earthquake Engineering and Structural Dynamics,<br />

31, 1891-1913.<br />

-165-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

DYNAMIC MECHANICAL ANALYSIS OF MAGNETORHEOLOGICAL SMART<br />

NANOCOMPOSITES<br />

L. Z. Sun<br />

Departments of Civil & Environmental Engineering and Chemical Engineering & Materials Science<br />

<strong>University</strong> of California, Irvine, USA. Email: lsun@uci.edu<br />

ABSTRACT<br />

<strong>The</strong>re has been an emphasis on the development of smart materials and composites such as magnetorheological<br />

(MR) elastomer composites whose mechanical behavior can be controlled rapidly and reversibly when subjected<br />

to an external magnetic field. While promising, MR elastomers have comparatively low storage moduli which<br />

impair their potential in industrial applications. Meanwhile, carbon nanotube (CNT) reinforced elastomers have<br />

proved to provide the increase in stiffness, strain to failure and extraordinary damping performances, giving rise<br />

to a possible approach to enhance the original mechanical properties of conventional MR elastomers. We<br />

develop a new type of CNT-reinforced MR elastomer nanocomposites and investigate their dynamic mechanical<br />

behavior. This research combines the advantages of MR elastomers and CNT nanocomposites in the fabrication<br />

and analysis of novel smart composite materials with enhanced stiffness and field-dependent storage moduli.<br />

<strong>The</strong> results are helpful for investigation and optimization of the nanocomposites performance for such<br />

engineering applications as smart structures.<br />

-166-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

A THEORETICAL PLATE END FLEXURAL DEBONDING MODEL FOR PLATED<br />

BEAMS<br />

J.F. Chen 1, * , V. Narayanamurthy 1,2 , J. Cairns 2 and D.J. Oehlers 3<br />

1<br />

Institute for Infrastructure and Environment,<br />

School of Engineering, <strong>The</strong> <strong>University</strong> of Edinburgh, Edinburgh, U.K. Email: J.F.Chen@ed.ac.uk<br />

2<br />

School of the Built Environment, Heriot Watt <strong>University</strong>, Edinburgh, U.K.<br />

3<br />

School of Civil, Environmental and Mining Engineering, <strong>The</strong> <strong>University</strong> of Adelaide, SA 5005, Australia.<br />

ABSTRACT<br />

Reinforced concrete (RC) beams strengthened in flexure by adhesive bonding of a fibre reinforced polymer<br />

(FRP) or steel plate on their tension face are susceptible to premature plate end debonding failures. Safe design<br />

of such a strengthened RC beam demands a reliable and predictive debonding strength model. <strong>The</strong>re are two<br />

special cases of plate end debonding failures: flexural debonding for the case when the plate terminates within a<br />

constant bending moment region (CMR) and shear debonding for the case when the plate terminates where the<br />

shear force is large but the bending moment is minimal. An interaction equation between these two extreme<br />

cases is usually used to model a general plate end debonding case. This paper presents the development of a first<br />

ever theoretical model for flexural debonding.<br />

KEYWORDS<br />

FRP, steel, RC beams, interfacial stresses, fracture energy, debonding, debonding models, test database.<br />

INTRODUCTION<br />

Reinforced concrete (RC) beams strengthened in flexure by adhesive bonding of a fibre reinforced polymer<br />

(FRP) or steel plate to their tension face are vulnerable to various modes of debonding failure. <strong>The</strong>re are many<br />

different classifications of these debonding failures, but they may be broadly classified into two types: plate end<br />

debonding (PED) and intermediate crack induced debonding (ICD) (Teng and Chen, 2009). Several ICD<br />

strength models are now available, either based on extensive finite element analysis (FEA) (Lu et al., 2005a, b)<br />

or rigorous mechanics analysis (Teng et al., 2003; Chen et al., 2006; Teng et al., 2006; Chen et al., 2007). By<br />

contrast, all existing PED strength models (e.g. Oehlers, 1992; Teng and Yao, 2007) are either empirical or<br />

semi-empirical.<br />

PED failure initiates at or near the plate ends. Several different modes may occur (Teng and Chen 2009): (1)<br />

concrete cover separation; (2) interfacial debonding; (3) mixed mode with a combination of (1) and (2); and (4)<br />

critical diagonal crack (CDC) initiated concrete cover separation or interfacial debonding. <strong>The</strong>se debonding<br />

failures are brittle and occur well before the full flexural capacity of the plated section is achieved. Many<br />

empirical or semi-empirical models have been developed for prediction of the PED strength. A thorough review<br />

and classification of all debonding strength models prior to 2002 can be found in Smith and Teng (2002a, b).<br />

Oehlers (1992) introduced a methodology to develop a PED strength model in three stages based on the location<br />

where the plate terminates: (1) a flexural debonding model that caters for the case where the plate terminates in<br />

a CMR; (2) a shear debonding model that caters for the case where the plate terminates in a high shear region<br />

with zero or low bending moment; and (3) an interaction model based on these two special models that caters<br />

for any region in between the above two cases (shear-moment region). Thus although PED in a CMR represents<br />

a special case of PED which is not itself of practical application, the importance of the corresponding flexural<br />

debonding strength model forms the basis of a general model for predicting PED.<br />

This paper is concerned with the development of the first theoretical flexural debonding strength model for the<br />

special case of RC beams with the plate terminating in a CMR (Figure 1). This is followed by an assessment of<br />

accuracy of the model with a database of 67 test results.<br />

-167-


Figure 1 Flexural-strengthened RC beam with plate terminated in CMR<br />

ASSESSMENT OF PEELING AND SHEAR EFFECT FROM ADHESION ANALYSIS<br />

Higher interfacial shear and normal (peeling) stresses are introduced between the beam and the plate at the plate<br />

end {τ(0) and σ(0) respectively} when the beam is loaded. <strong>The</strong> PED failure is usually due to the combination of<br />

these two stresses, but their relative significance depends on the geometry, material and loading of the plated<br />

beam. A simplistic assessment of these stresses based on an elastic adhesion analysis using Narayanamurthy et<br />

al.’s (2010) model for a plate terminated within the CMR reveals that their magnitude is primarily influenced by<br />

the shear stiffness of the adhesive and the axial stiffness of the plate, Figure 2.<br />

Shear to normal interfacial<br />

stress ratio τ (0)/σ (0)<br />

5<br />

4<br />

3<br />

2<br />

Ea/ta=0.5 GPa/mm<br />

Ea/ta=1 GPa/mm<br />

Ea/ta=2 GPa/mm<br />

Ea/ta=5 GPa/mm<br />

Ea/ta=10 GPa/mm<br />

1<br />

0.0 0.6 1.2 1.8 2.4 3.0<br />

Axial stiffness of plate E 2 t 2 (GPa-m)<br />

Figure 2 Interfacial shear-to-normal stress ratio at the plate end in a beam with plates terminating in the CMR<br />

Figure 2 shows that the peak interfacial shear stress is several times greater than the peak interfacial normal<br />

stress over a wide range of these two parameters. For practical values of adhesive and plate stiffnesses, the peak<br />

shear stress exceeds the peak normal stress by between 50% and 500%. <strong>The</strong>refore, it may be reasonable to<br />

assume that the Mode-2 shear fracture dominates in flexural debonding, and Mode-1 tension fracture is<br />

neglected in the following analysis for simplicity. <strong>The</strong> accuracy of the model will be assessed with a test<br />

database.<br />

INTERFACIAL SHEAR BEHAVIOUR<br />

Bonded Joint Model<br />

Consider a simply supported RC beam with a span of l under four point bending (Figure 1). It is strengthened in<br />

flexure by adhesive bonding of an FRP or steel plate of length L within the CMR. <strong>The</strong> PED in this beam may be<br />

analysed by investigating the plated segment between the major flexural cracks at each end of the plate. Cracks<br />

between the two ends of the plate are neglected but their effects on the axial and flexural stiffnesses can be<br />

easily taken into consideration by using cracked section values in the analysis. <strong>The</strong> segment is idealised as a<br />

bonded joint and is subjected to identical axial forces N 1 (0) and bending moments M 1 (0) on both sides as shown<br />

in Figure 3. End faces of the plates are unstressed. <strong>The</strong> adhesive is assumed to have a constant thickness and the<br />

bond stress is assumed to be constant through its thickness. <strong>The</strong> adherends are assumed to behave linearly. <strong>The</strong><br />

left (x = 0) and right (x = L) crack locations are referred as left end and right end respectively. <strong>The</strong> moment at<br />

the left is equal to that at the right in the case considered in this study (Figure 3).<br />

-168-


Figure 3 Idealised bonded joint model<br />

Figure 4 Softening local bond-slip model<br />

Local Deformation (bond-slip) Model<br />

<strong>The</strong> actual local FRP-to-concrete bond-slip behaviour can be closely approximated by a bilinear bond-slip<br />

model (Yuan et al., 2004; Lu et al., 2005). However, Chen et al. (2007) have demonstrated that using a<br />

rigid-linear softening bond-slip model as shown in Figure 4 and retaining the fracture energy G f of a bilinear<br />

model leads to almost the same final predictions while significantly simplifying the calculation process. This<br />

linear softening bond-slip model is adopted in the present formulation where δ is the longitudinal slip at the<br />

interface, τ f is the peak bond stress (bond strength) and δ f is the ultimate slip when shear stress reduces to zero<br />

and the plate is considered to be debonded. Friction and aggregate interlock in the debonded area is ignored,<br />

leading to the absence of any residual shear strength after debonding. <strong>The</strong> bond-slip model in Figure 4<br />

represents a constitutive relation that can be described mathematically by the following equation:<br />

⎧ 0 when δ ( x)<br />

= 0<br />

⎪<br />

τ ( x)<br />

= f ( δ ) = ⎨<br />

τ f<br />

[ δ f − δ ( x)]<br />

when 0 < δ ( x)<br />

≤ δ<br />

(1)<br />

f<br />

⎪<br />

⎩δ<br />

f<br />

Failure Process and States of Interface<br />

Figure 5 shows the sequence of debonding propagation for a typical failure process and the corresponding<br />

interfacial shear stress distribution for the bonded joint shown in Figure 3. A point on the interface can be in a<br />

rigid, softening, or debonded state. Letters R (rigid), S (softening) and D (debonding) are used to describe the<br />

states of the interface from the plate end to mid-length. <strong>The</strong> entire interface is initially in a rigid state as the<br />

adopted local bond–slip model neglects elastic deformations. Softening initiates at the plate ends as soon as any<br />

loading is applied, resulting in micro-cracking and interfacial slip. <strong>The</strong> softening length of the interface a<br />

increases with increases in loading and reaches a d when debonding (macro-cracking) initiates at a slip of δ f . <strong>The</strong><br />

interface ahead of the softening front remains rigid and has no interfacial stresses in the present analysis (Figure<br />

5a-c), which makes the softening front abrupt. When the more accurate bilinear bond-slip model is used the<br />

actual stress distribution would be smooth ahead of the softening front throughout the whole loading process<br />

(Teng et al., 2006).<br />

Figure 5 Failure progression and interfacial shear stress distribution at different states of the bonded joint<br />

interface<br />

-169-


In the CMR, the interface during the failure process is symmetrical about the middle of the bonded joint so only<br />

half of the interface (from x = 0 to 0.5L) needs to be considered in this analysis. Assuming that the bond length<br />

is sufficiently large so that L > 2a d , (if L≤2a d no debonding state exists), the interface progressively experiences<br />

the following states during the loading process:<br />

1) softening–rigid (S–R) state (Figure 5a);<br />

2) debonding–softening–rigid (D–S–R) state (Figure 5b,c); and<br />

3) debonding–softening (D–S) state (Figure 5d,e).<br />

Governing Equations and Solutions for Different States of Interfaces<br />

<strong>The</strong> equilibrium, constitutive relations and the interface compatibility requirements for a differential segment of<br />

the plated beam between two flexural cracks are used to derive the governing differential equation to analyse the<br />

initiation and propagation of plate end debonding in the plated beam. Considering the local deformation model<br />

shown in Figure 4 for the case of 0 < δ(x) ≤ δ f , the governing equation for the interfacial slip for the softening<br />

interface is obtained as:<br />

2<br />

d δ ( x)<br />

2<br />

+ λ δ ( x)<br />

= λ<br />

2 δ<br />

2<br />

f<br />

(2)<br />

dx<br />

and the axial force in the plate is given by<br />

2<br />

τ f b2<br />

⎛ dδ<br />

( x)<br />

2 ⎞<br />

N2(<br />

x)<br />

= ⎜ + mλ<br />

M ( x)<br />

⎟<br />

(3a)<br />

2<br />

T<br />

2G<br />

⎝ dx<br />

f λ<br />

⎠<br />

where<br />

2<br />

2<br />

τ f b2<br />

⎡( y + )( + + )<br />

⎤<br />

=<br />

1 y2<br />

y1<br />

ta<br />

y2<br />

1 1<br />

1 ⎡ y + ⎤<br />

λ ⎢<br />

+ + ⎥ ; =<br />

1 y<br />

m<br />

2<br />

⎢<br />

⎥<br />

2G<br />

f ⎣ E1I1<br />

+ E2I2<br />

E1<br />

A1<br />

E2A2<br />

⎦<br />

2 ⎣(<br />

E1I1<br />

+ E2<br />

I (3b,c)<br />

λ 2)<br />

⎦<br />

in which y 1 and y 2 represent the distances from the bottom of adherend-1 (the original beam) and the top of<br />

adherend-2 (the plate) to their respective centroids; N 2 (x) and M T (x) refer respectively to the axial force in the<br />

plate and total bending moment due to applied external loading; t, b, E, A and I represent the thickness,<br />

breadth, elastic modulus, cross-sectional area and second moment of area of the adherends and adhesive<br />

respectively; and subscripts 1, 2 and a respectively refer to beam, plate and adhesive.<br />

<strong>The</strong> interfacial slip, the interfacial shear stress and the axial force in the plate for the softening region in<br />

different states of interface can be found by solving Eq. 2 for the appropriate boundary and continuity<br />

conditions.<br />

Softening–rigid (S–R) interface<br />

<strong>The</strong> interface remains in the S–R state as shown in Figure 5a until the debonding bending moment M Td is<br />

reached at the left plate end (x = 0) as the loads are increased gradually from zero. <strong>The</strong> general solution for the<br />

softening region of the interface (0 ≤ x ≤ a) with 0 < δ(x) ≤ δ f is given by<br />

δ ( x)<br />

= A1 sin[ λ(<br />

x − a)]<br />

+ B1<br />

cos[ λ(<br />

x − a)]<br />

+ δ f<br />

(4a)<br />

τ f<br />

τ ( x)<br />

= − ( A1 sin[ λ(<br />

x − a)]<br />

+ B1<br />

cos[ λ(<br />

x − a)]<br />

)<br />

(4b)<br />

δ<br />

where<br />

f<br />

2<br />

( A λ cos[ λ(<br />

x − a)]<br />

− B λ sin[ λ(<br />

x − a)]<br />

+ mλ<br />

M ( ))<br />

N2(<br />

x)<br />

= m1<br />

1<br />

1<br />

x . (4c)<br />

2<br />

f b2<br />

2<br />

f<br />

τ<br />

m 1 = (4d)<br />

2G<br />

λ<br />

T<br />

<strong>The</strong> boundary and continuity conditions at the softening region of the interface are:<br />

N 2 (x) = 0 at x = 0<br />

N 2 (x) is continuous at x = a<br />

δ(x) = 0 at x = a<br />

(5a)<br />

(5b)<br />

(5c)<br />

<strong>The</strong> constants of integration A 1 and B 1 and the softening length a are obtained by applying Eqs 5a-c to Eqs 4a-c<br />

as given below.<br />

-170-


A = 0 1 (6a)<br />

B1 = −δ f<br />

(6b)<br />

f sin( λa)<br />

mλM T<br />

(6c)<br />

δ =<br />

<strong>The</strong> softening length a increases gradually with applied loading. <strong>The</strong> relative displacement Δ 0 between the<br />

adherends at the left end can be obtained from Eq. 4a as:<br />

Δ = δ [1 − cos( λ )]<br />

(7)<br />

Debonding–softening–rigid (D–S–R) interface<br />

0 f a<br />

<strong>The</strong> S–R interface in Figure 5b is on the verge of becoming a D–S–R interface (Figure 5b when Δ 0 reaches δ f at<br />

x = 0 and τ reduces to zero). From Eq. 7 this gives<br />

π<br />

a d =<br />

(8)<br />

2λ<br />

<strong>The</strong> applied bending moment at the initiation of debonding M Td can be obtained from Eq. 6c as:<br />

δ f<br />

MTd = (9)<br />

mλ<br />

<strong>The</strong> shape and length of the softening region and the applied loading remains constant until the rigid region of<br />

the interface ahead of the softening front vanishes as debonding propagates. This is achieved by the movement<br />

of the softening front steadily towards the mid-length of the bonded joint (x = 0.5L) as shown in Figure 5c-d. In<br />

this process, the debonded length d at the left end of the interface increases from 0 to (0.5L–a d ).<br />

Solutions given in Eqs 4–6 are valid if x is replaced by [x – (0.5L–d)] and a by a d . <strong>The</strong> axial plate force within<br />

the debonded zone is zero. Considering Eq. 3, the slip or the relative displacement at x = 0 during D–S–R state<br />

of interface can be obtained as<br />

2<br />

Δ 0 = δ f + mλ<br />

MTdd<br />

(10)<br />

Debonding–softening (D–S) interface<br />

<strong>The</strong> D–S–R interface in Figure 5c becomes a D–S interface when the length of the rigid region ahead of the<br />

softening front reduces to zero and d = (0.5L – a d ) as shown in Figure 5d. <strong>The</strong> general solutions in Eqs 4a–c are<br />

valid if (x–a) is replaced by [x – (0.5L – a)] and the constants of integration A 1 and B 1 are replaced by A 2 and B 2 ,<br />

which can be determined from the following boundary and symmetric conditions for the softening region of the<br />

interface:<br />

N 2 (x) = 0 ; δ(x) = δ f at x = (0.5L–a) (11a,b)<br />

δ(x) = 0 at x = 0.5L<br />

(11c)<br />

Application of Eqs 11a-c to the modified general solutions from Eqs 4a–c as described above provide the<br />

following expressions for the softening length a and the integration constants A 2 and B 2 :<br />

δ f<br />

A2 = −mλ MT<br />

; B2<br />

= 0 ; sin( λ a)<br />

=<br />

(12a,b,c)<br />

mλM<br />

Any further increase in applied loading is accompanied by a decrease in the softening length a as per Eq. 12c.<br />

<strong>The</strong> relative displacement at x = 0 during the debonding process can be obtained from Eq. 10. <strong>The</strong> distribution<br />

of interfacial stress is shown in Figure 5e for this state which notionally ends when a = 0. Mathematically<br />

complete debonding occurs only when a →0 and M T →∞ (Eq. 12c) but this is not practical and is a result of the<br />

assumed static behaviour. If dynamic behaviour during the debonding process is considered, any increase of the<br />

loading during this stage is unlikely to be experienced in practice. What is significant from this analysis is the<br />

debonding initiation moment (Eq. 9).<br />

Load–displacement response<br />

Figure 6 shows the load-displacement (plate end slip ∆ 0 ) response predicted using the above formulations (Eqs<br />

6c-10 and 12c) for the test steel plated RC beam specimen S13/20 reported in Oehlers and Moran (1990). <strong>The</strong><br />

T<br />

-171-


ond-slip parameters calculated using Lu et al.’s (2005b) bi-linear bond-slip model are: δ f = 0.16 mm; τ f = 3.5<br />

MPa; and G f = 0.28 N/mm. <strong>The</strong>se parameters may also be obtained using other models such as Seracino et al.<br />

(2007). Key points in the load-displacement response are marked as O, A and B in Figure 6. <strong>The</strong> curve OA<br />

represents the S–R state of the interface, with point A representing the initiation of debonding. <strong>The</strong> plateau AB<br />

represents the D–S–R state, with point B representing the end of the D-S-R state when the softening front<br />

reaches mid-span. <strong>The</strong> curve starting from B onwards (dotted line in Figure 6) refers to the D–S state which<br />

has little practical significance as discussed above.<br />

20<br />

Bending moment (kNm)<br />

16<br />

12<br />

8<br />

4<br />

O<br />

•<br />

A<br />

B<br />

• •<br />

0<br />

0.0 0.3 0.5 0.8 1.0 1.3 1.5<br />

Relative displacement at the left plate end (mm)<br />

Figure 6 Load-displacement response of the bonded joint in CMR<br />

FLEXURAL DEBONDING TEST DATABASE<br />

A large database containing 67 RC beams with the flexural strengthening plate terminated in a CMR and which<br />

failed due to flexural debonding has been assembled from an extensive literature survey.. <strong>The</strong> database includes<br />

54 steel plated beams, 9 CFRP plated beams, 2 GFRP plated beams and 2 C-GFRP plated beams. All the beams<br />

failed due to flexural PED failure either by concrete cover separation or interfacial failure or a combination of<br />

these two modes. Geometric and material parameters for the beams can be found in Oehlers and Moran (1990),<br />

Oehlers (1992), Mohammed Ali et al. (2001), Smith and Teng (2003) and Yao and Teng (2007). Measured<br />

debonding moments and debonding moments predicted by the expressions derived earlier are presented in Table<br />

1. It is seen that these tests cover a wide range of important parameters: (1) elastic modulus of plate E 2 = 8.8-257<br />

GPa; (2) nominal thickness of plate t 2 = 0.165-32 mm; (3) cube strength of the concrete f cu = 25-59 MPa; (4)<br />

split tensile strength of the concrete f t = 2.55-4.9 MPa; (5) elastic modulus of concrete E 1 = 20-32 GPa; (6) axial<br />

stiffness of plate per unit width E 2 t 2 = 37.6-3150 (x10 3 ) N/mm; (7) effective axial stiffness of beam per unit<br />

width E 1 d e = 2.4-7.1 (x10 6 ) N/mm; (8) flexural rigidity of the cracked plated RC beam section (EI) c,p = 0.83-6.7<br />

(x10 12 ) Nmm 2 ; (9) flexural rigidity of the cracked un-plated RC beam section (EI) c,0 = 0.38-4.7 (x10 12 ) Nmm 2 ;<br />

(10) width ratio of beam to plate α w = 1.0-4.8; (11) internal tensile reinforcement steel ratio ρ s = 0.44-5.40 %;<br />

(12) interfacial fracture energy G f = 0.25-0.88 Nmm/mm 2 ; and (13) interfacial bond strength τ f = 2.7-6.8<br />

N/mm 2 .<br />

<strong>The</strong> formulation of the present theoretical model includes the adhesive thickness t a , but the results are<br />

insensitive to the value selected. <strong>The</strong> predictions from the theoretical model presented in the rest of the paper do<br />

not include the effect of adhesive parameters. Not all concrete properties such as f t , f cu and cylinder strength f c<br />

were available for many test beams. In such cases, the relationships provided in BS 8110 (1997) and ACI 318-08<br />

(2008) have been used to calculate the missing properties from those available.. Compression reinforcement is<br />

ignored in all calculations.<br />

-172-


Table 1 Test and predicted flexural debonding moment<br />

Reference Specimen M e ε e M p* /M e M p** /M e M Oeh /M e M T-Y /M e<br />

(kNm) (µε) (kNm) (kNm) (kNm) (kNm)<br />

Oehlers & S1/1 9.2 618 1.0 1.1 0.8 1.0<br />

Moran (1990) S1/2 9.9 639 0.9 1.0 0.7 0.9<br />

S2/1 12.2 502 0.7 0.7 0.3 0.7<br />

S2/2 11.5 491 0.7 0.8 0.4 0.7<br />

S3/1 11.3 606 0.8 0.9 0.5 0.7<br />

S3/2 10.8 546 0.9 0.9 0.5 0.8<br />

S3/3 12.6 588 0.7 0.8 0.5 0.6<br />

S3/4 13.5 748 0.7 0.7 0.4 0.6<br />

S4/1 12.6 512 0.7 0.9 0.5 0.7<br />

S4/2 14.0 562 0.7 0.8 0.4 0.7<br />

S4/3 13.5 543 0.7 0.8 0.5 0.7<br />

S4/4 14.4 588 0.6 0.8 0.4 0.6<br />

S5/1 11.5 603 0.8 0.9 0.5 0.8<br />

S5/2 15.1 717 0.6 0.7 0.4 0.6<br />

S5/3 11.5 615 0.8 0.8 0.4 0.8<br />

S5/4 9.9 639 0.9 1.0 0.5 0.9<br />

S6/1 16.0 617 0.5 0.6 0.3 0.6<br />

S6/2 12.6 551 0.7 0.7 0.4 0.7<br />

S6/3 12.6 536 0.7 0.8 0.4 0.7<br />

S6/4 14.2 558 0.6 0.7 0.3 0.6<br />

S9/1 16.7 1103 0.6 0.6 0.4 0.6<br />

S9/2 14.0 999 0.7 0.7 0.4 0.7<br />

S9/3 13.5 825 0.7 0.8 0.4 0.6<br />

S9/4 15.3 1109 0.6 0.7 0.4 0.5<br />

S10/1 12.3 244 0.8 1.0 0.5 0.8<br />

S10/2 10.8 231 1.0 1.1 0.5 0.9<br />

S10/3 11.2 348 0.9 1.1 0.6 1.0<br />

S10/4 11.7 305 0.9 1.0 0.5 1.0<br />

S11/1 12.3 573 1.0 1.1 0.8 1.1<br />

S11/2 13.1 498 0.9 1.0 0.8 1.0<br />

S11/3 7.9 128 1.3 1.6 0.7 1.2<br />

S11/4 9.2 174 1.1 1.4 0.6 1.1<br />

S12/1 10.4 261 1.0 1.1 0.4 0.9<br />

S12/2 12.7 403 0.9 0.8 0.3 0.8<br />

S12/3 16.4 1004 0.7 0.7 0.2 0.6<br />

S12/4 10.6 268 1.1 1.1 0.5 1.0<br />

S13/6 24.4 682 0.6 0.6 0.4 0.6<br />

S13/7 23.9 682 0.7 0.8 0.7 0.8<br />

S13/9 17.6 971 0.8 1.0 0.8 0.8<br />

S13/10 16.1 938 0.9 1.1 0.9 0.9<br />

S13/11 13.1 461 0.9 1.1 0.7 0.8<br />

S13/13 12.6 765 0.8 0.9 0.6 0.8<br />

S13/14 9.0 425 0.9 1.0 0.5 0.9<br />

S13/15 15.3 870 0.9 1.0 0.5 0.7<br />

S13/16 11.4 472 0.9 1.0 0.4 0.8<br />

S13/17 6.3 214 1.1 1.2 0.7 1.0<br />

S13/18 7.9 629 1.0 1.1 0.9 1.0<br />

S13/19 10.6 618 0.7 0.7 0.7 0.7<br />

S13/20 10.2 395 0.6 0.7 0.5 0.6<br />

Note: * with Lu et al’s (2005b) bond model; ** with Seracino et al’s (2007) bond model<br />

-173-


Table 1 (Contd.) Test and predicted flexural debonding moment<br />

Reference Specimen M e ε e M p* /M e M p** /M e M Oeh /M e M T-Y /M e<br />

(kNm) (µε) (kNm) (kNm) (kNm) (kNm)<br />

Oehlers (1992) 7/2/* 13.5 - 0.9 1.1 0.8 1.0<br />

8/2/* 12.9 - 0.9 1.0 0.7 0.9<br />

Yao & Teng (2007) CS-A 12.5 1158 1.3 1.4 4.9 1.2<br />

CS-L1-A 17.0 2425 1.3 1.3 6.6 1.0<br />

CS-L3-A 13.5 919 1.1 1.2 4.0 1.0<br />

CS-W50-A 13.8 1689 1.7 1.5 3.7 1.1<br />

CS-W100-A 13.3 1038 1.4 1.3 4.3 1.1<br />

CP-A 11.4 404 1.2 1.4 4.1 1.1<br />

SP-A 11.5 215 1.1 1.1 2.9 0.9<br />

GS-A 19.1 2924 1.2 1.4 8.2 0.9<br />

CS-C10-A 18.9 1540 1.1 1.2 4.9 0.9<br />

CS-C50-A 11.9 1211 1.1 1.1 4.0 1.1<br />

Smith & Teng (2003) 6-A 17.0 - 1.0 1.0 3.4 0.8<br />

Mohomed Ali FP-S-5 26.3 325 1.3 1.2 0.7 0.8<br />

et. al. (2001) FP-C-8.5 28.1 342 1.2 1.1 0.7 0.8<br />

FP-CG2-16 28.0 455 1.3 1.2 0.9 0.9<br />

FP-G-32 40.4 809 1.2 1.1 1.1 0.7<br />

FP-CG-16 62.1 1775 1.1 1.0 1.4 0.7<br />

Note: * with Lu et al’s (2005b) bond model; ** with Seracino et al’s (2007) bond model<br />

2.0<br />

2.0<br />

1.5<br />

1.5<br />

Mp* / Me<br />

1.0<br />

Mp** / Me<br />

1.0<br />

0.5<br />

0.0<br />

Average<br />

95% lower bound<br />

0 500 1000 1500 2000 2500 3000<br />

ε e (x10 -6 )<br />

(a) With Lu et al.’s (2005b) bond parameters<br />

Figure 7 Comparison of model predictions with test data<br />

NEW FLEXURAL DEBONDING STRENGTH MODEL<br />

0.5<br />

0.0<br />

Average<br />

95% lower bound<br />

0 500 1000 1500 2000 2500 3000<br />

ε e (x10 -6 )<br />

(b) With Seracino et al.’s (2007) bond parameters<br />

A new theoretical model developed in this section directly employs the theoretical debonding moment M p (Eq. 9)<br />

from the preceding interfacial shear analysis, i.e.<br />

δ f<br />

M p = φ1<br />

(13)<br />

m λ<br />

where λ and m are given in Eqs 3b-c in which the parameters A 1 and I 1 should be replaced respectively by A c,0<br />

and I c,0 that are the elastic sectional properties of the cracked unplated beam section. Coefficient φ 1 is introduced<br />

in Eq. 13 so the model can be turned into a lower bound design model.<br />

Calculation of M p (Eq. 13) needs the bond properties δ f ; τ f and G f for which any bond models such as Lu et al.’s<br />

(2005b) or Seracino et al.’s (2007) can be used. This is referred as M p* and M p** respectively when the above<br />

two bond models are used. <strong>The</strong> statistical results of prediction-to-test using both bond models give similar<br />

accuracy (Table 2). For predicting the debonding strength, φ 1 = 1.0, and Eq. 13 is the same as Eq. 9. Based on<br />

the data presented in Table 1, φ 1 = 0.737 or 0.701 for the 95 percentile (1.645 × standard deviation) lower bound<br />

respectively when Lu et al. (2005b) or Seracino et al. (2007) bond parameters are used.<br />

-174-


COMPARISON OF FLEXURAL DEBONDING MODELS WITH TEST DATABASE<br />

<strong>The</strong> test flexural debonding moment and the predicted values from the present theoretical model (Eq. 13) are<br />

listed in Table 1. <strong>The</strong> prediction-to-test moment ratios are plotted against measured plate strains at mid-span ε e at<br />

debonding failure in Figures 7a-b. It may be noted that present theoretical model does not rely on test results and<br />

is fully based on the interfacial fracture mechanics. Its prediction (Figure 7a) is conservative for 49 out of the 67<br />

test data. <strong>The</strong> statistical results of the natural logarithm of predicted-to-test debonding moment ratios are also<br />

provided in Table 2. <strong>The</strong> average, maximum and minimum should be close to zero, and the standard deviation<br />

should be small for a good model. <strong>The</strong> positive and negative sign in average, maximum and minimum can be<br />

considered as a representation of the unconservative and conservative predictions respectively.<br />

Table 2 Statistics of prediction-to-test debonding moment ratio<br />

Prediction-to-test ratio Log of prediction-to-test ratio<br />

Model Ave CoV Max Min Ave SD Max Min<br />

Present model with Lu et al’s (2005b)<br />

bond model 0.912 0.268 1.681 0.540 -0.126 0.262 0.519 -0.616<br />

Present model with Seracino et al’s<br />

(2007) bond model 0.989 0.236 1.588 0.585 -0.039 0.240 0.462 -0.536<br />

CONCLUSIONS<br />

This paper has presented an investigation on the plate end debonding behaviour of flexurally strengthened RC<br />

beams with a bonded plate terminated in the constant bending moment region (CMR), the so called flexural<br />

debonding behaviour. A simple assessment of the interfacial stresses near the plate end suggests that the<br />

interfacial shear behaviour, rather than the peeling behaviour dominates in this case which may appear<br />

counterintuitive. Based on an interfacial fracture mechanics analysis between the bonded plate and the RC beam,<br />

a theoretical model has been developed for prediction of flexural debonding moment. This model is assessed<br />

against a database containing 67 test results and is shown to be in close agreement with test data.<br />

ACKNOWLEDGEMENTS<br />

<strong>The</strong> authors would like to thank the support from the Scottish Funding Council for the Joint Research Institute<br />

between Edinburgh and Heriot-Watt Universities which is a part of the Edinburgh Research Partnership in<br />

Engineering and Mathematics (ERPem), and from the UKIERI (UK India Education and Research Initiative)<br />

project (IND/CONT/07-08/E/133).<br />

REFERENCES<br />

ACI (2008). Building code requirements for reinforced concrete (318-08), American Concrete Institute,<br />

Farmington Hills Michigan.<br />

Ali, M.S., Oehlers, D.J. and Park, S.M. (2001). “Comparison between FRP and steel plating of reinforced<br />

concrete beams”, Composites: Part A, 32(9), 1319-1328.<br />

BS 8110 (1997). Structural use of concrete-part 1: Code of practice for design and construction, British<br />

Standards Institute, London.<br />

Chen, J.F., Teng, J.G. and Yao, J. (2006). “Strength model for intermediate crack debonding in FRP-strengthened<br />

concrete members considering adjacent crack interaction”, Proceedings of the 3 rd Intl Conference on FRP<br />

Composites in Civil Engg (CICE2006), Miami, Florida, USA, 67-70.<br />

Chen, J.F., Yuan, H. and Teng, J.G. (2007). “Debonding failure along a softening FRP-to-concrete interface<br />

between two adjacent cracks in concrete members”, Engineering Structures, 29, 259-270.<br />

Lu, X.Z., Ye L.P., Teng, J.G. and Jiang, J.J. (2005a). “Meso-scale finite element model for FRP sheets/plates<br />

bonded to concrete”, Engineering Structures, 27(4), 564-575.<br />

Lu, X.Z., Teng, J.G., Ye, L.P. and Jiang, J.J. (2005b). “Bond-slip models for FRP sheets/plates bonded to<br />

concrete”, Engineering Structures, 27(6), 920-937.<br />

Narayanamurthy, V., Chen, J. F. and Cairns, J. (2010). “A general analytical method for the analysis of<br />

interfacial stresses in plated beams under arbitrary loading”, Advances in Struct Engg, 13 (5), 975-988.<br />

Oehlers, D.J. (1992). “Reinforced concrete beams with plates glued to their soffits”, Journal of Structural<br />

Engineering, ASCE, 118(8), 2023-2038.<br />

Oehlers, D.J. and Moran, J.P. (1990). “Premature failure of externally plated reinforced concrete beams”,<br />

Journal of Structural Engineering ASCE, 116(4), 979-995.<br />

-175-


Seracino, R., Saifulnaz, M.R. and Oehlers, D.J. (2007). “Generic debonding resistance of EB and NSM<br />

plate-to-concrete joints”, Journal of Composites for Construction, ASCE, 11 (1), 62-70.<br />

Smith, S.T. and Teng, J.G. (2002a). “FRP-strengthened RC beams - I: Review of debonding strength models”,<br />

Engineering Structures, 24 (4), 385-395.<br />

Smith, S.T. and Teng, J.G. (2002b). “FRP-strengthened RC beams - II: Assessment of debonding strength<br />

models”, Engineering Structures, 24 (4), 397-417.<br />

Smith, S.T. and Teng, J.G. (2003). “Shear-bending interaction in debonding failures of FRP-plated RC beams”,<br />

Advances in Structural Engineering, 6(3), 183-199.<br />

Teng, J.G. and Chen, J.F. (2009). “Mechanics of debonding in FRP-Plated RC beams”, Proceedings of ICE,<br />

Structures and Buildings, 162, SB5, 335-45.<br />

Teng, J.G., Smith, S.T. and Chen, J.F. (2003). “Intermediate crack-induced debonding in RC beams and slabs”,<br />

Construction and Building Materials, 17, 447-462.<br />

Teng, J.G. and Yao, J. (2007). “Plate end debonding in FRP-plated RC beams - II: Strength model”, Engineering<br />

Structures, 29(10), 2472-2486.<br />

Teng, J.G., Yuan, H. and Chen, J.F. (2006). “FRP-to-concrete interface between two adjacent cracks: <strong>The</strong>oretical<br />

model for debonding failure”, International Journal of Solids and Structures, 43, 5750-5778.<br />

Yao, J. and Teng, J.G. (2007). “Plate end debonding in FRP-plated RC beams - I: Experiments”, Engineering<br />

Structures, 29(10), 2457-2471.<br />

Yuan, H., Teng, J.G., Seracino, R., Wu, Z.S. and Yao, J. (2004). “Full-range behavior of FRP-to-concrete bonded<br />

joints”, Engineering Structures, 26 (5), 553-565.<br />

-176-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

SMART AGGREGATE-BASED DAMAGE DETECTION OF<br />

FRP-STRENGTHENED COLUMNS UNDER REVERSED CYCLIC<br />

LOADING<br />

Haichang Gu 1 , Rachel Howser 2 , Yashar Moslehy 2 , , Hemant Dhonde 2 , Gangbing Song 1 , Y.L. Mo 2 ,<br />

and Ashraf Ayoub 2<br />

2 Department of Mechanical Engineering<br />

1 Department of Civil and Environmental Engineering<br />

<strong>University</strong> of Houston<br />

4800 Calhoun, Houston, TX, 77004<br />

ABSTRACT<br />

Structural health monitoring is an important aspect of maintenance of large civil infrastructures,<br />

especially for bridge columns in areas of high seismic activity. In this project, recently developed<br />

innovative multi-functional piezoceramic-based transducers, called smart aggregates, were utilized<br />

to perform health monitoring of a shear-critical reinforced concrete (RC) bridge column before<br />

and after reinforced polymer (FRP) sheets were wrapped around the column for more strength.<br />

During the experiment, the concrete column without FRP was first subjected to reversed-cyclic<br />

loading protocol until failure. After the column failed, it was wrapped with fiber reinforced<br />

polymer sheets, commonly used to retrofit seismically damaged structures, and the FRP<br />

strengthened column was retested under the same reversed cyclic loading pattern. During the<br />

reversed loading tests, based on developed smart aggregates, an active-sensing and an impact<br />

response-based approach were used to evaluate the health status of the RC column and the FRP<br />

strengthened column. Wave transmission energy is attenuated by the existence of cracks during the<br />

loading procedure, and this attenuation phenomenon alters the transfer function curve between the<br />

actuator and sensor. To detect the damage occurrence and evaluate the damage severity, transfer<br />

function curves were compared with those obtained during the healthy status. A transfer<br />

function-based damage index matrix was developed to demonstrate the damage severity at<br />

different locations. Furthermore, a weighted damage index is developed to quantitatively evaluate<br />

the overall damage status. Experimental results verified the effectiveness of the smart aggregates<br />

in health monitoring of the FRP-strengthened column as well as the un-strengthened column. <strong>The</strong><br />

experimental results show that the proposed smart aggregate-based approach can successfully<br />

detect the damage occurrence and evaluate its severity.<br />

KEY WORDS<br />

FRP, RC columns, piezoceramic, smart aggregate, health monitoring, weighted damage index.<br />

INTRODUCTION<br />

Natural hazards, such as earthquakes, cause severe damage in RC structural systems. RC columns<br />

are typically the primary elements of energy dissipation in bridges subject to seismic loads. Failure<br />

of bridge columns can result in the collapse of bridge girders, which was the scenario for the<br />

majority of the bridges damaged in past earthquakes. It is imperative to quickly assess the severity<br />

of the damage and the health status of an RC structure in real time or near real time after a disaster<br />

event to provide vital structural safety information for decision makers. This is especially true for<br />

large-scale infrastructures where it is desirable to design an automated and distributed system to<br />

perform structural health monitoring.<br />

In this project, innovative piezoceramic-based devices, called smart aggregates, were used as<br />

transducers for structural health monitoring of RC columns under a reversed-cyclic loading<br />

procedure. <strong>The</strong> proposed smart aggregate is a multi-functional device. In additional to the<br />

-177-


structural health monitoring function, the smart aggregate can be used to perform early-age<br />

concrete strength monitoring (Gu et al. 2006), impact detection (Song et al. 2007a) and dynamic<br />

seismic detection (Gu et al. 2010). In previous studies, smart aggregates have been successfully<br />

used in health monitoring of various concrete structures under different loading cases. <strong>The</strong><br />

proposed smart aggregate-based approach has been used to perform health monitoring for the<br />

following cases: a concrete bridge bent-cap under static loading (Song et al. 2007b), a concrete<br />

frame under static loading (Laskar et al. 2009, Zhao et al. 2006), a shear wall under reversed<br />

loading (Yan et al. 2009), circular columns under seismic loading (Liao et al. 2008), and circular<br />

columns under reversed cyclic loading (Moslehy et al. 2010). <strong>The</strong> purpose of this project was to<br />

verify the effectiveness of the smart aggregate in health monitoring of an FRP-strengthened RC<br />

column under reversed cyclic loading.<br />

In the experiments, an active-sensing approach and an impact response-based approach were used<br />

to perform the health monitoring. In the active-sensing approach, one smart aggregate was used to<br />

generate the sweep sine signals to propagate through the concrete structure; the other distributed<br />

smart aggregates were used as sensors to detect the responses. In the impact response-based<br />

approach, an impact hammer is used to strike on the concrete column on marked pre-determined<br />

location to generate stress waves to propagate through the concrete structure; the other distributed<br />

smart aggregates were used as sensors to detect the impact responses. In both approaches, transfer<br />

functions of the sensor signals were obtained during the testing process. <strong>The</strong> existence of cracks<br />

attenuated the wave propagation and affected the frequency response curves. A damage index<br />

matrix was formed based on the change of frequency response curves to demonstrate the health<br />

status at different locations. Furthermore, a weighted damage index was developed to<br />

comprehensively evaluate the overall damage status.<br />

To verify the effectiveness of the proposed piezoceramic-based health monitoring approach,<br />

structural health monitoring tests were performed on a shear-critical RC column in the Thomas<br />

T.C. Hsu Structural Research Laboratory at the <strong>University</strong> of Houston (UH). A reversed cyclic<br />

loading protocol was used to gradually load the column to failure, while the smart aggregates were<br />

used as transducers to perform the structural health monitoring during the loading procedure. After<br />

the column failed, it was repaired using an FRP fabric, which was wrapped around the plastic<br />

hinge region of the damaged column. Wrapping a column with FRP confines the concrete in the<br />

column, allowing the confined concrete to carry significant stresses at higher strains. Confined<br />

concrete demonstrates an increase in both compressive strength and ultimate strain. In an<br />

FRP-wrapped column, the transverse strains greatly increase at high load values because of<br />

internal cracking. Concrete bares out against the confining FRP, which eventually ruptures,<br />

causing failure of the column. In this research, the FRP-strengthened column was again loaded<br />

until failure by a reversed-cyclic loading protocol, and the smart aggregates were once again used<br />

to evaluate the health status of the column during the loading procedure. It was observed that<br />

although the failure mode was different from a typical RC column, the smart-aggregate based<br />

health monitoring system could successfully detect both the failure occurrence and its severity.<br />

EXPERIMENTAL SETUP<br />

<strong>The</strong> smart aggregates, as shown in Figure 1, were formed by embedding a water-proof coated<br />

piezoceramic patch with lead wires into a small concrete block before casting. A RC column was<br />

fabricated at UH, and six smart aggregates were installed at pre-determined distributive locations<br />

in the shear-critical column before casting. <strong>The</strong>se locations are shown in Figures 2 and 3.<br />

Water-proof coating<br />

Electric wires<br />

Piezoceramic patch<br />

Figure 1 Illustration of a smart aggregate<br />

-178-


Figure 2 Location of smart aggregates (view from east<br />

side of column )<br />

Figure 3 Location of smart aggregates<br />

SMART AGGREGATE-BASED STRUCTURAL HEALTH MONITORING SYSTEMS<br />

In recent years, the demand for structural health monitoring of large-scale structures has increased<br />

greatly in order to reduce maintenance costs and enhance safety. Traditional health monitoring<br />

methods, such as using x-ray or ultrasonic C-scan technologies, are expensive and sometimes<br />

ineffective for large-scale structures with limited or no accessibility. Fiber optical sensors,<br />

including the Fiber Bragg Grating sensors, are now used for the health monitoring of various RC<br />

structures (Mehrani et al. 2009, Zhang et al. 2006, Ren et al. 2006, Lu and Xie 2006, Chen and<br />

Ansari, 2000). However, fiber optical sensors offer only local measurements, which greatly limit<br />

their applications. Piezoelectric transducers have emerged as a new tool for health monitoring of<br />

large-scale structures due to having advantages including solid-state actuation, low cost, quick<br />

response and availability in different shapes. In general, there are two major categories of the<br />

piezoelectric-based health monitoring approaches for concrete structures: the impedance-based<br />

health monitoring approach (Hey et al. 2006, Naidu and Bhalla 2003, Tseng and Wang, 2004, Soh<br />

et al. 2000, Bhalla and Soh 2004, Park et al. 2006, Zhao and Li, 2006, Raju et al. 1999) and the<br />

vibration-characteristic approach (Okafor et al. 1996, Song et al. 2007 b , Saafi et al. 2001, Miller et<br />

al. 2002, Na et al. 2003).<br />

In this study, a piezoceramic-based vibration-characteristic approach was applied to perform the<br />

structural health monitoring of a RC column under reversed-cyclic loading. <strong>The</strong> wave response<br />

detected by the proposed piezoceramic-based smart aggregates was used to evaluate the health<br />

status of the concrete structure during the reversed cyclic loading. Two smart aggregate-based<br />

active sensing systems, as shown in Figure 4, were used for the health monitoring of the concrete<br />

column. In System 1, as shown in Figure 4(a), the piezoelectric transducer in one smart aggregate<br />

was used as an actuator to excite the desired guided waves. In System 2, as shown in Figure 4(b),<br />

an impact hammer was used to strike the column on marked locations to generate the stress waves<br />

to propagate through the column. <strong>The</strong> piezoelectric transducers in the distributed smart aggregates<br />

were used as sensors to detect wave responses. Cracks inside the RC column acted as stress relief<br />

in the wave propagation path. <strong>The</strong> amplitude of the wave and the transmission energy will<br />

decrease due to the existence of cracks. <strong>The</strong> value of the transmission energy drop will correlate<br />

with the degree of the damage inside. For System 1, the transfer function is obtained by using the<br />

smart aggregate actuator signals as input and the smart aggregate sensor signals as output. For<br />

System 2, the signal from the embedded force sensor inside the impact hammer is used as the input<br />

signal for the transfer function; the sensor signal detected from the smart aggregate is used as the<br />

output signal for the transfer function.<br />

-179-


Smart aggregate<br />

(sensor)<br />

Smart aggregate<br />

(sensor)<br />

Smart aggregate<br />

(actuator)<br />

Impact hammer<br />

(a) (b<br />

Figure 4 (a) Diagram of active-sensing system, (b) Diagram of impact response-based health<br />

monitoring system<br />

<strong>The</strong> energy deviation of the j th sensor at the k th test is defined as:<br />

(1)<br />

<strong>The</strong> maximum energy deviation value for the j th sensor for all tests is defined as:<br />

Max (j) = , (k=1, 2, …, m ) (2)<br />

An element in a damage index matrix is the normalized energy deviation, which is defined as:<br />

(3)<br />

where k is the test number, j is the sensor number and m is the total number of tests. Based on<br />

experimental data, a three-dimensional damage index matrix plot can be constructed to reveal the<br />

damage evolution.<br />

For a complex structure, it may be confusing and time-consuming to obtain a comprehensive<br />

overall health monitoirng result based on multiple sensors. In this paper, a weighted damage index<br />

is developed to provide a comprehensive quantitative result based on multiple sensors. <strong>The</strong><br />

weighted damage index is defined as:<br />

(4)<br />

where w(j) is weight for a damage index value, n is the total number of sensor data sets used,<br />

is the distance from actuator to sensor, k is the test number, and j is the sensor number. As shown<br />

in equation (4), the weight value w(j) is proportional to the reciprocal of a distance between a<br />

sensor- actuator pair, which means that the smaller the distance between an actuator- sensor pair,<br />

the greater the weight value .<br />

TEST PROCEDURE<br />

<strong>The</strong> tested RC column was subjected to reversed-cyclic loading using the experimental setup<br />

shown in Figure 5. An axial load equal to 10% of the column’s axial capacity was applied using<br />

the jack shown in Figure 5. A reversed cyclic loading protocol, described in Figure 6, was then<br />

applied on the concrete column to load the structure until failure. <strong>The</strong> failed column can be<br />

observed in Figure 7. During the loading procedure, the proposed piezoceramic-based smart<br />

aggregates were used as transducers to perform health monitoring to evaluate the health status of<br />

the RC column.<br />

After the failure of the column, all the cracked concrete pieces were removed from the structure,<br />

and it was repaired by replacing grouts in spalling locations. Subsequently, the column was<br />

wrapped with FRP sheets as show in Figure 8. Based on material tests, the FRP sheets used in this<br />

study can be specified with a modulus of elasticity equal to 15,000 ksi and an ultimate strength<br />

equal to 13.2 ksi. <strong>The</strong> FRP fabric thickness was 0.011 in. <strong>The</strong> column was retested under<br />

reversed-cyclic loading, as seen in Figure 9. In Figures 10 and 11, rupture of FRP in the repaired<br />

column at failure can be observed. <strong>The</strong> load versus displacement curves for both tests are shown in<br />

Figure 12. It can be seen from Figure 12 that the FRP effectively confined the concrete. <strong>The</strong><br />

strengthening provided by the FRP proved to be successful in recovering the strength loss of the<br />

damaged column, although a reduction in stiffness was observed. During the loading procedure,<br />

the piezoceramic-based smart aggregates were used as transducers to perform health monitoring in<br />

order to evaluate the damage status of the column.<br />

-180-


Figure 5 Experimental setup<br />

Figure 6 Loading protocol for column<br />

Figure 7 Column damage<br />

Figure 8 Column repaired with FRP<br />

Figure 9 Loading protocol for FRP-wrapped column<br />

Figure 10 FRP rupture<br />

Figure 11 Close up of FRP rupture<br />

-181-


Figure 12 Force vs. displacement curves for RC column and FRP-wrapped RC column<br />

EXPERIMENTAL RESULTS<br />

Experimental Results of RC Column<br />

<strong>The</strong> reversed cyclic loading protocol shown in Figure 6 was implemented to gradually test the RC<br />

column until failure. During the loading procedure, smart aggregate-based health monitoring<br />

system 1 (active-sensing approach) was implemented at the positive loading peaks, the zero<br />

loading points and the negative loading peaks of the cyclic loading protocol. Figure 13 shows the<br />

damage index matrix for the column, defined in Equation (3). Figure 14 shows the weighted<br />

damage index, defined in Equation (4).<br />

During the health monitoring test using an active-sensing based approach, PZT2 was used as an<br />

actuator to excite a sweep sine wave and the other smart aggregates were used as sensors to detect<br />

the wave responses. From the damage index matrix results shown in Figure 13, the damage index<br />

increases significantly at Test Number 6, one cycle before the first major cracks were seen in the<br />

RC column. This likely indicates internal damage in the column before it could be seen externally.<br />

<strong>The</strong> weighted damage index matrix reveals the overall damage status of the column. From the<br />

weighted damage index matrix results shown in Figure 14, it can be seen that the index<br />

dramatically increases after test number 6 which means the overall damage status became severe<br />

after test 6.<br />

Damage index<br />

1<br />

0.5<br />

0<br />

1 2 3 4 5 6<br />

Sensor No.<br />

5<br />

Test No.<br />

Figure 13 Sensor History Damage Index<br />

Matrix plot for RC column (active-sensing<br />

approach)<br />

10<br />

15<br />

Weighted damage index<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0 5 10 15 20<br />

Test No.<br />

Figure 14 Weighted damage index for RC column<br />

(active-sensing approach)<br />

Experimental Results of FRP-Wrapped RC Column<br />

<strong>The</strong> reversed cyclic loading protocol shown in Figure 9 was implemented to retest the<br />

FRP-wrapped RC column. During the loading procedure, smart aggregate-based health monitoring<br />

(impact response-based approach) was implemented at the positive loading peaks, the zero loading<br />

points as well as the negative loading peaks of the cyclic loading protocol.<br />

-182-


Figure 15 Impact location on FRP wrapped RC column<br />

During the health monitoring test, an impact hammer was used to generate stress waves, and smart<br />

aggregates were used as sensors to detect the wave responses. <strong>The</strong> impact hammer was used to<br />

repeatedly strike the column on marked location I and location II, as shown in Figure 15. <strong>The</strong><br />

health monitoring results for impact tests on location I are shown in Figures 16, 17 and 18. From<br />

the damage index matrix based on the frequency response magnitude ranging from 2k-10k Hz, as<br />

shown in Figure 16, the damage index increased slightly at Test Number 3 and continues to<br />

progressively increase during the course of the test. This indicates slight increased internal damage<br />

during the test. <strong>The</strong> damage index increased greatly in Test Number 17, when the FRP ruptured,<br />

indicating major structural damage in the column. This shows that the column loses its<br />

confinement as a result of FRP rupturing, and that the existing cracks inside the column open up<br />

substantially at this stage.<br />

Figure 17 show the damage index matrix based on the frequency response magnitude ranging from<br />

10k-20k Hz, while Figure 18 show the damage index matrix based on the frequency response<br />

magnitude ranging from 18k-20k Hz. While comparing the damage index matrices based on<br />

frequency response among different frequency ranges, as shown in Figures 16, 17 and 18, it can<br />

clearly be observed that damage index matrix based on a higher frequency range is more sensitive<br />

in detecting the damage. <strong>The</strong> development trend of the damage index matrix based on the higher<br />

frequency range (10 k-20 kHz) in Figure 17 is more sensitive and smoother than that in Figure 16.<br />

This smooth development trend is more apparent in Figure 18, where the frequency range used to<br />

calculate the damage index matrix ranges from 18 to 20 kHz. Figure 19 shows the damage index<br />

matrix for FRP wrapped RC column when the column was impacted on location II. <strong>The</strong> frequency<br />

range used to calculate the damage index matrix was from 18 to 20 KHz. From the damage index<br />

matrix, it can be seen that at test 17, there is a tremendous increase in damage index values, which<br />

indicates major structural damage in the column. <strong>The</strong> health monitoring results from impact<br />

location II is consistent with the results from impact location I. Figure 20 and Figure 21 show the<br />

weighted damage index for the FRP wrapped RC column when the column was impacted on<br />

location I and II, respectively. From the weighted damage index for both cases, it can be seen that<br />

the damage index increases tremendously at test 17, which reveals that the overall damage status<br />

of the FRP wrapped RC column failed at test 17.<br />

Damage index<br />

1<br />

0.5<br />

0<br />

1 2 3 4 5 6<br />

Sensor No.<br />

5<br />

Test No.<br />

Figure 16 Damage index matrix based on<br />

frequency response magnitude among<br />

2k-10kHz (struck on location I)<br />

10<br />

15<br />

Damage index<br />

1<br />

0.5<br />

0<br />

1 2 3 4 5 6<br />

Sensor No.<br />

5<br />

Test No.<br />

Figure 17 Damage index matrix based on<br />

frequency response magnitude among 10k-20kHz<br />

(struck on location I)<br />

10<br />

15<br />

-183-


Damage index<br />

1<br />

0.5<br />

0<br />

1 2 3 4 5<br />

Sensor No. 6<br />

5<br />

10<br />

Test No.<br />

Figure 18 Damage index matrix based on<br />

frequency response magnitude among 18k-20kHz<br />

(struck on location I)<br />

15<br />

Damage index<br />

1<br />

0.5<br />

0<br />

1 2 3 4 5 6<br />

Sensor No.<br />

5<br />

Test No.<br />

Figure 19 Damage index matrix for FRP<br />

wrapped RC column (struck on location II)<br />

10<br />

15<br />

W eighted dam age index<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0 5 10 15 20<br />

Test No.<br />

Figure 20 Weighted damage index (struck on<br />

location I)<br />

Weighted damage index<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0 5 10 15 2<br />

Test No.<br />

Figure 21 Weighted damage index (struck on<br />

location II)<br />

CONCLUSIONS<br />

In this study, innovative piezoceramic-based smart aggregates were used as transducers for<br />

structural health monitoring of a RC column before and after being retrofitted with FRP composite<br />

sheets under reversed-cyclic loading protocols. <strong>The</strong> RC column was first tested until failure. <strong>The</strong><br />

column was then wrapped with an FRP sheet and retested until failure. <strong>The</strong> active sensing-based<br />

approach and the impact response-based approach were used for health monitoring purposes. To<br />

quantitatively evaluate the severity degree of damage, a damage index matrix and a weighted<br />

damage index were developed based on the attenuation of the transmitted wave. <strong>The</strong> experimental<br />

results demonstrated that the damage inside attenuated the transmitted energy during the loading<br />

procedure and the proposed damage index matrix extracted useful damage development<br />

information, such as a prediction point of failure. From experimental results, it can be seen that the<br />

transmitted wave attenuated in wide frequency ranges and the attenuation level can be correlated<br />

with damage severity. It was also observed that high frequency signals proved to be more sensitive<br />

in detecting the damage existence. <strong>The</strong> proposed smart aggregate–based approach could<br />

successfully detect not only failure occurrence but also evaluate its severity. <strong>The</strong> study confirms<br />

that the proposed piezoceramic-based smart aggregates have the great potential to be applied to<br />

perform structural health monitoring for large-scale infrastructures.<br />

ACKNOWLEDGEMENTS<br />

<strong>The</strong> research study described herein was sponsored by the National Science Foundation under<br />

Award No. CMMI-0724190, EEC-0649163 and DGE-0840889. <strong>The</strong> opinions expressed in this<br />

study are those of the authors and do not necessarily reflect the views of the sponsor. <strong>The</strong> authors<br />

would also like to thank Fyfe Co. LLC. who generously donated fabrics and epoxies, and Master<br />

Builders Inc. and CMC Construction Services for donating chemical admixtures and formwork for<br />

this study.<br />

-184-


REFERENCES<br />

Bhalla, S. and Soh, C.K. (2004). “High frequency piezoelectric signatures for diagnosis of<br />

seismic/blast induced structural damage”, NDT and E International, 37(1), 23-33.<br />

Chen, X. and Ansari, F. (2000). “Fiber optic stress wave sensor for detection of internal flaws in<br />

concrete structures”, Journal of Intelligent Material Systems and Structures, 10(4),<br />

274-279.<br />

Gu, H. , Song G., Dhonde, H., Mo Y.L. and Yan, S. (2006). “Concrete early-age strength<br />

monitoring using embedded piezoelectric transducers”, Smart Materials and Structures,<br />

15(6), 1837-1845.<br />

Gu, H., Moslehy, Y., Sanders, D., Song, G. and Mo, Y.L. (2010). “Multi-functional Smart<br />

Aggregate-based Structural Health Monitoring of Circular Reinforced Concrete Columns<br />

subjected to Seismic Excitations”, Smart Materials and Structures, 19, 065026 (7pp).<br />

Hey, F., Bhalla, S. and Soh, C.-K,. (2006). “Optimized parallel interrogation and protection of<br />

piezo-transducers in electromechanical impedance technique”, Journal of Intelligent<br />

Material Systems and Structures, 17(6), 457-468.<br />

Laskar, A ., Gu, H., Mo, Y.-L. and Song, G. (2009). “Progressive collapse of a 2-story reinforced<br />

concrete frame with embedded smart aggregate”, Smart Materials and Structures, 18(7),<br />

075001(10pp).<br />

Liao, W.I., Gu, H., Olmi, C., Song, G., Mo, Y.-L. and Loh, C.H. (2008). “Structural health<br />

monitoring of a concrete column subjected to shake table excitations using smart<br />

aggregates”, Intelligent Sensor and Actuator Symposium at ASCE Earth and Space<br />

conference, Long Beach, California.<br />

Lu, S. and Xie, H. (2006). “Real-time monitoring and simulating the load effects of smart CFRP -<br />

Strengthened RC beams”, Key Engineering Materials, 324-325 I, 129-132.<br />

Mehrani, E., Ayoub, A.S. and Ayoub, A. (2009). “Condition assessment of reinforced concrete<br />

bridges using structural health monitoring techniques - a case study”, Journal of Smart<br />

Structures and Systems, 5(4), 381-396.<br />

Miller, T., Hauser, C.J. and Kundu, T. (2002). “Nondestructive inspection of corrosion and<br />

delamination at the concrete-steel reinforcement interface”, American Society of<br />

Mechanical Engineers (Publication) NDE, v 23, Proceedings of the ASME<br />

Nondestructive Evaluation Engineering Division, 121-128.<br />

Moslehy, Y., Gu, H., Belarbi, A., Mo, Y. L.and Song, G. (2010). “Smart Aggregate-Based Damage<br />

Detection of Circular RC columns under Cyclic Combined Loading”, Smart Materials<br />

and Structures, 19, 065021 (12pp).<br />

Na, W.-B. and Kund, T. (2003). “Inspection of interfaces between corroded steel bars and concrete<br />

using the combination of a piezoelectric zirconate-titanate transducer and an<br />

electromagnetic acoustic transducer”, Experimental Mechanics, 43(1), 24-31.<br />

Naidu, A.S.K. and Bhalla, S. (2003). “Damage detection in concrete structures with smart<br />

piezoceramic transducers”, Proceedings of the SPIE - <strong>The</strong> International Society for<br />

Optical Engineering, 5062(1), 684-690.<br />

Okafor, A.C. (1996). “Chandrashekhara and Jiang Y.P. Delamination prediction in composite<br />

beams with built-in piezoelectric devices using modal analysis and neural network”,<br />

Smart Materials and Structures, 5(3), 338-347.<br />

Park, S., Ahmad, S., Yun, C.-B. and Roh, Y. (2006). “Multiple crack detection of concrete<br />

structures using impedance-based structural health monitoring techniques”, Experimental<br />

Mechanics, 46(5), 609-618.<br />

Raju, V., Gyuhae, Park and Cudney, H.H. (1999). “Impedance-based health monitoring of<br />

composite reinforced structures”, Ninth International Conference on Adaptive Structures<br />

and Technologies, 448-457.<br />

Ren, L., Li, H., Li, X., Zhou, J. and Xiang, L. (2006). “Application of FBG sensors in rolled<br />

concrete dam model”, Proceedings of SPIE - <strong>The</strong> International Society for Optical<br />

Engineering, Smart Structures and Materials - Sensors and Smart Structures<br />

Technologies for Civil, Mechanical, and Aerospace Systems, 6174.<br />

Saafi, M. and Sayyah, T. (2001). “Health monitoring of concrete structures strengthened with<br />

advanced composite materials using piezoelectric transducers”, Composites Part B<br />

(Engineering), 32(4), 333-342.<br />

Soh, C.H., Tseng, K.K., Bhalla, S. and Gupta, A. (2000). “Performance of smart piezoceramic<br />

patches in health monitoring of a RC bridge”, Smart Materials and Structures, 9(4),<br />

533-542.<br />

-185-


Song, G., Olmi, C. and Gu, H. (2007a). “An overheight vehicle - bridge collision monitoring<br />

system using piezoelectric transducer”, Smart Materials and Structures, 16(2), 462-468.<br />

Song, G, Gu, H., Mo, Y.L., Hsu, T.T.C. and Dhonde, H. (2007b). “Concrete structural health<br />

monitoring using embedded piezoceramic transducers”, Smart Materials and Structures,<br />

16, 959-968.<br />

Song, G., Gu, H. and Mo, Y.-L. (2008). “Smart aggregates, multi-functional sensors for concrete<br />

structures-A tutorial and a review”, Smart Materials and Structures, 17(3), 033001<br />

(17pp).<br />

Sun, F.P., Chaudhry, Z., Rogers, C.A. and Majmundar, M. (1995). “Automated real-time structure<br />

health monitoring via signature pattern recognition”, Proceedings of SPIE - <strong>The</strong><br />

International Society for Optical Engineering, 2443, 236-247.<br />

Tseng, K.K.-H. and Naidu, A.S.K. (2002). “Non-parametric damage detection and characterization<br />

using smart piezoceramic material”, Smart Materials and Structures, 11(3), 317-329.<br />

Tseng, K.K. and Wang, L. (2004). “Smart piezoelectric transducers for in situ health monitoring of<br />

concrete” Smart Materials and Structures, 13(5), 1017-1024.<br />

Yan, S., Sun, W., Song, G., Gu, H., Huo, L.-S., Liu, B. and Zhang, Y.-G. (2009). “Health<br />

monitoring of reinforced concrete shear walls using smart aggregates”, Smart Materials<br />

and Structures, 18(4), 047001 (6pp).<br />

Zhang, W., Gao, J. Shi, J., Cui, H.and Zhu, H. (2006). “Health monitoring of rehabilitated concrete<br />

bridges using distributed optical fiber sensing”, Computer-Aided Civil and Infrastructure<br />

Engineering, 21, 411-424.<br />

Zhao, X. and Li, H. (2006). “Health monitoring of reinforced concrete frame-shear wall using<br />

piezoceramic transducer”, Journal of Vibration and Shock, 25(4), 82-84.<br />

-186-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

ON THE MULTI-SCALE MODELING OF HETEROGENEOUS GEOMATERIALS<br />

J.F. Shao 1,2 , A. Guery 1 , T. Jiang 2 , Q.Z. Zhu 1 , D. Kondo 3<br />

1<br />

LML, UMR8107 CNRS, Email: jian-fu.shao@polytech-lille.fr<br />

<strong>University</strong> of Lille, Villeneuve d’Ascq, France<br />

2<br />

Faculty of Civil and Transportation Engineering,<br />

Hohai <strong>University</strong>, Nanjing, China<br />

3<br />

IJLRA, UMR7190 CNRS<br />

<strong>University</strong> of Paris 6, Paris, France<br />

ABSTRACT<br />

Most geomaterials are characterized by multi-scale heterogeneous structures such as pores, mineral grains,<br />

bedding planes, microcracks, interfaces etc. <strong>The</strong> macroscopic behaviour of such materials inherently depends on<br />

the mineral composition and the evolution of microstructure at various scales such as microcrack propagation,<br />

pore expansion or collapse. <strong>The</strong> macroscopic properties of such heterogeneous materials are also governed by<br />

physical mechanisms at relevant scales. Constitutive modelling of geomaterials should take into account such<br />

physical mechanisms. In this paper, we present a general framework for multi-scale modelling of geomaterials.<br />

Two specific situations will be in particular discussed. Firstly, based on linear homogenization technique, we<br />

present a micromechanics-based formulation for damage modelling in brittle rocks, by taking into account<br />

coupling between crack growth and frictional sliding. Secondly, the micromechanical modelling is extended to<br />

heterogeneous geomaterials containing non linear mineral phases. A non-linear homogenization method is<br />

presented to describe coupled plastic damage behaviour in semi-brittle rocks.<br />

KEYWORDS<br />

Micromechanics, damage, homogenization, multi-scale modelling, heterogeneous rocks<br />

INTRODUCTION<br />

In many engineering applications such as stability and failure analysis of underground structures, the knowledge<br />

of mechanical behaviours of geomaterials is required, together with other properties such permeability and heat<br />

conductivity. Extensive experimental investigations have shown that the mechanical behaviours of geomaterials<br />

are general complex and inherently related to their microstructure and mineral compositions. For instance, the<br />

anisotropic properties of sedimentary rocks are due to existence of bedding planes and other weakness planes.<br />

Damage due to nucleation and propagation of microcracks is an essential mechanism of inelastic deformation<br />

and failure in many brittle rocks. On the other hand, for a given class of rocks, the macroscopic behaviour is<br />

strongly affected by the variation of mineral compositions. For instance, in clayey rocks, the macroscopic<br />

swelling capacity is directly depending on the clay content. In Figure 1, we show the microscopic image of<br />

mineral compositions of a clayed rock. This rock called Callovo-Oxfordian argillite has been extensively studied<br />

in France in the context of feasibility studies for geological disposal of nuclear waste. Due to its low<br />

permeability and high mechanical strength, the Callovo-Oxfordian argillite is considered as potential host<br />

formation for underground storage of high level radioactive wastes. It is seen that the mechanical behaviour of<br />

the argillite varies with layer depth due to variation of mineral composition. At the macroscopic scale, the<br />

mechanical response of the argillite can be characterized by significant plastic deformation coupled with<br />

induced damage, strong sensitivity to confining pressure and water content, transition from volumetric<br />

compressibility and dilatancy as well as slight anisotropy (Chiarelli et al. 2003; Shao et al. 2006; Jia et al. 2010).<br />

A series of microscopic investigations have also been conducted in order to identify physical mechanisms at<br />

relevant scales involved in macroscopic mechanical responses (ANDAR 2005). It is found that the argillite is<br />

composed of three principal phases: clay matrix, quartz and calcite grains. <strong>The</strong> macroscopic deformation seems<br />

to mainly be related to distortion of clay matrix, similar to dislocation of crystal structures. Microcracks are<br />

observed both at interfaces between mineral grains and clay matrix and inside clay matrix. Such microcracks are<br />

responsible to mechanical damage of argillite such as deterioration of elastic modulus. More importantly, the<br />

macroscopic permeability can significantly change due to nucleation and propagation of microcracks. <strong>The</strong><br />

porosity of argillite is mainly due to voids between clay particles which is an assembly of parallel clay platelets.<br />

-187-


<strong>The</strong> distance between clay platelets constitutes nanoscopic porosity and significant controls water sensitivity and<br />

time dependent behaviour of argillite. Further, the poromechanical properties are also influenced by interaction<br />

between nanoscopic and microscopic pores.<br />

Figure 1 Micrograph of a typical argillite structure:<br />

calcite grains (C), quartz grains (tectosilicates) (T)<br />

and clay matrix (MA)<br />

Figure 2 Nucleation and propagation of microcrack<br />

around mineral grains under applied stress<br />

All these experimental data and microscopic investigations clearly show that the macroscopic properties of<br />

argillite are inherently related to the evolution of microstructure at different scales. A number of constitutive<br />

models have been proposed for geomaterials, elastoplastic models, damage models, coupled plastic damage<br />

models. Most models are based on phenomenological approaches using the framework of thermodynamics of<br />

irreversible process. In practice, these models can capture main features of mechanical behaviours of<br />

geomaterials and used for engineering analysis and design. However, the phenomenological models fail in<br />

linking macroscopic responses to microscopic mechanisms. For example, they are not able to properly describe<br />

influences of mineral compositions. Some micromechanical models have been proposed for modelling induced<br />

damage in brittle geomaterials. <strong>The</strong>y are so far generally based on linear fracture mechanics without using<br />

rigorous up-scaling methods. With the rapid use of new composite materials, significant advances have been<br />

realised in homogenization methods for determination of effective properties of linear and heterogeneous<br />

materials. <strong>The</strong> application of such techniques to geomaterials represents an interesting challenge in constitutive<br />

modelling of such complex materials. In this short review paper, we intend to present the general framework of<br />

multi-scale modelling of geomaterials using proper homogenization methods. Two typical cases will be<br />

considered. In the first case, a micromechanical model for anisotropic damage in brittle rocks will be presented.<br />

In the second one, we will show the application of a nonlinear homogenization method to plastic damage<br />

modelling of the Callovo-Oxfordian argillite.<br />

LINEAR HOMOGENIZATION<br />

Consider a component of structure in geomaterials (Figure 3). Each material point constituting the structure<br />

component is seen as homogeneous medium at the macroscopic scale ( L ). However, at smaller scales, the<br />

material point contains various heterogeneities as above mentioned. <strong>The</strong> macroscopic properties of the material<br />

point depend on the mineral composition and microstructure at this position, and can be estimated by different<br />

homogenization schemes. <strong>The</strong> homogenization methods are based on the proper definition of a representative<br />

volume element (RVE). <strong>The</strong> definition of RVE must verify the principle of scale separation. <strong>The</strong> seize of the<br />

RVE ( l ) must be large enough compared with the characteristic length of heterogeneities (noted as a ) to be<br />

studied such as average grains size, and at the same time smaller enough compared with the characteristic length<br />

of structure component ( L ). Thus the scale separation implies a ≤ l ≤ L.<br />

x i<br />

L<br />

l<br />

x j<br />

Figure 3 From macroscopic to microscopic scale: definition of representative volume element (RVE)<br />

-188-


<strong>The</strong> RVE, occupying a geometrical domain V , is composed of N material phases with r = 0, N − 1 . Note V r<br />

and f r be respectively the volume and volume fraction of the phase r . Introduce the notations w and w r to<br />

denote the average of the local field function wz ( ) respectively on the entire volume of the RVE and on the<br />

partial volume of the<br />

th<br />

r phase, such as:<br />

N −1<br />

w 1<br />

= w = w ( z ) dz = f V<br />

∑ r w r<br />

V<br />

∫ ;<br />

r=<br />

0<br />

1<br />

wr = w = w( z)<br />

dz<br />

r<br />

V<br />

∫ (1)<br />

V r<br />

Consider first the estimation of effective properties of heterogeneous materials. Assume that the RVE is<br />

subjected to a uniform macroscopic strain field at its boundary, E . <strong>The</strong> local strain field inside each phase is<br />

related to the macroscopic strain by an appropriate strain concentration tensor A( z)<br />

as:<br />

r<br />

ε ( z) = A ( z):<br />

E<br />

(2)<br />

In elastic materials, each phase is described by the linear elastic law with local elastic stiffness tensor L ( ):<br />

e<br />

e<br />

z<br />

σ ( z) = L ( z): ε( z)<br />

(3)<br />

<strong>The</strong> macroscopic stress tensor is obtained by the averaging local stress field over the volume of RVE, one<br />

obtains:<br />

e<br />

hom<br />

Σ= σ = L ( z): A( z) : E = L : E<br />

(4)<br />

<strong>The</strong> macroscopic elastic stiffness tensor is then given by:<br />

e−hom<br />

L = L( z): A ( z)<br />

(5)<br />

One can note that the estimation of macroscopic elastic properties of heterogeneous materials is entirely related<br />

to the strain concentration tensor A ( z)<br />

. Its specific forms are directly related to the homogenization schemes<br />

used. For most geomaterials, three different schemes are generally considered (Zaoui 2002): dilute scheme,<br />

Mori-Tanaka scheme (MT) and self-consistent scheme.<br />

MESOMECHANICAL MODELLING OF ANISOTROPIC DAMAGE IN BRITTLE ROCKS<br />

Effective elastic properties of damaged material<br />

Damage due to nucleation and growth of microcrack is an essential mechanism of dissipation and failure in<br />

brittle rocks. <strong>The</strong> mechanical and other (hydraulic for instance) properties are affected by displacement<br />

discontinuities induced by the microcracks. For the sake of clarity, consider here the RVE composed of an<br />

elastic solid weakened a family of penny-shaped parallel microcracks (Figure 4). <strong>The</strong> family of microcracks is<br />

characterized by the unit normal ( n ) and average radius ( a ). <strong>The</strong> aspect ratio of penny-shaped microcracks is<br />

defined by ò = c/<br />

aand generally very small.<br />

s<br />

n<br />

Ω<br />

c<br />

E<br />

n<br />

a<br />

∂Ω<br />

2c<br />

Figure 4 RVE of cracked solid and schematic representation of a penny-shaped crack (Zhu et al. 2008)<br />

s<br />

<strong>The</strong> mechanical property of the elastic solid is defined by the elastic stiffness tensor C . An elastic stiffness<br />

c<br />

tensor ( C ) is also defined for the families of microcracks in order to account for unilateral effects. In the case<br />

of isotropic elastic solid matrix, the stiffness tensor reads C s = 3k<br />

J + 2μ<br />

s<br />

K, with k and μ s being the bulk<br />

and shear modulus of undamaged material respectively. According to (5), the effective elastic stiffness of<br />

cracked material depends on the strain concentration tensor of each phase. Assume now that the strain<br />

concentration tensor is constant in each phase and using the identity relation A( z ) = I, the effective stiffness<br />

of cracked material can be written as:<br />

hom s c<br />

C = C +ϕ<br />

c s<br />

C −C : A (6)<br />

( )<br />

-189-


where<br />

c<br />

ϕ is the volume fraction of cracks. This relation clearly show the crucial role of the strain concentration<br />

tensor A whose determination depends on the homogenization scheme. <strong>The</strong> simplest scheme corresponds to the<br />

case of dilute distribution of cracks. This scheme is simply based on a direct use of the basic Eshelby’s solution<br />

to the matrix-inclusion problem (Eshelby 1957) by neglecting interaction between cracks. In the case of<br />

matrix-inclusion system, the Mori-Tanaka (MT) scheme (Mori-Tanaka, 1973) is largely used and in a simple<br />

way allows the consideration of crack interaction. This scheme was applied to cracked material by Benveniste<br />

(1987) and other researchers. <strong>The</strong> (MT) scheme is requires only the shape of inclusions and does not account for<br />

the spatial distribution of inclusions. As a consequence, it can not properly predict the effects of cracks<br />

interaction. To overcome this limitation, Ponte Castaneda and Willis (1995) proposed a new scheme by taking<br />

into account both the inclusions shape and their spatial distribution by introducing two Hill-type tensors ( P ò<br />

associated with the shape and Pd<br />

corresponding to the spatial distribution). Accordingly the strain concentration<br />

tensor takes the following general form (Zhu et al. 2008):<br />

−1 1<br />

−1<br />

c s ⎧<br />

−<br />

c c s c s ⎫<br />

A = ⎡I+ P :( − )<br />

⎤ : ⎨ +ϕ ⎡ + ( − d ):( − )<br />

⎤⎡ + :( − )<br />

⎤<br />

⎣<br />

蝌 C C I I P P C C I P<br />

⎦ ⎣<br />

⎦⎣<br />

C C<br />

⎦<br />

⎬ (7)<br />

⎩<br />

⎭<br />

s<br />

s<br />

<strong>The</strong> Hill tensors are related to the Eshelby ones by S蝌 = P : C and Sd<br />

= Pd<br />

: C . Thus, for the case of open<br />

cracks ( C<br />

c = 0 ), one gets:<br />

−1 1<br />

c −1<br />

−<br />

A = ( I− S ) : ⎡ +ϕ d : ( − ) ⎤<br />

蝌 ⎢<br />

I S I S<br />

⎣<br />

⎥<br />

(8)<br />

⎦<br />

<strong>The</strong> effective elastic stiffness of damaged material with open cracks is given by:<br />

1<br />

hom s ⎧ c −1 c<br />

−1<br />

− ⎫<br />

C = C : ⎨I−ϕ ( I− S ) : ⎡ +ϕ d :( ) ⎤<br />

蝌 ⎢<br />

I S I−<br />

S<br />

⎣<br />

⎥⎦<br />

⎬<br />

(9)<br />

⎩<br />

⎭<br />

Assuming now a spherical spatial distribution of microcracks, the tensor S is isortopic in nature and given by:<br />

11+ v<br />

2 4−5v<br />

Sd =α 1J+α2K; ; with α 1 = and α 2 =<br />

(10)<br />

3<br />

s<br />

1 15<br />

s<br />

−v<br />

1−v<br />

Note that the above general expression can be reduced to that of the dilute scheme by setting α 1 =α 2 = 0 and<br />

that of the Mori-Tanaka estimate by α 1 =α 2 = 1 . <strong>The</strong> expression of the Eshelby tensor ( S ò ) for penny-shaped<br />

cracks has been discussed in previous works (Horii and Nemat Nasser 1983; Mura, 1987; Zhu et al. 2008). In<br />

the particular case of cracks with ò → 0 , the tensor ( I−<br />

S ) −1<br />

ò become singular whereas the tensor<br />

ò→0<br />

−<br />

( ) 1<br />

T= lim ò I−S ò remains with limited value. <strong>The</strong>refore, the effective stiffness tensor of damaged material<br />

with open cracks is finally expressed as (Zhu et al. 2008):<br />

⎡<br />

−1<br />

hom s 4 ⎛ 4 ⎞ ⎤<br />

C = C : ⎢I− πω T:<br />

⎜I+ πωS d : T⎟<br />

⎥<br />

(11)<br />

⎢⎣<br />

3 ⎝ 3 ⎠ ⎥⎦<br />

3<br />

<strong>The</strong> variable ω=N a is identical to the crack density parameter initially introduced by Budiansky and<br />

O'Connell (1976).<br />

Coupled damage-frictional model<br />

In most cases, geomaterials are subjected to complex loading paths. Microcracks can be in either open and<br />

closed states. In closed microcracks, due to roughness of crack surfaces, the damage evolution related to crack<br />

propagation is inherently coupled with frictional sliding along crack surfaces. <strong>The</strong> frictional sliding is the main<br />

mechanism of irreversible strains in brittle rocks. A coupled damage-frictional model should be formulated. Due<br />

to the limited page number, the details of the formulation are not given in this paper and can be found some<br />

previous papers by the authors (Zhu et al. 2008, 2009). Only the main lines are summarized here. <strong>The</strong> condition<br />

of unilateral contact on the crack surface can be written as:<br />

[ un] ≥0; σnn ≤0; [ un] σ nn = 0<br />

(12)<br />

[ un<br />

] denotes the normal displacement jump across cracks and σ nn being the local normal stress applied to crack<br />

c<br />

surface. <strong>The</strong> macroscopic inelastic strain E due to microcracks can be expressed in the following form:<br />

pl<br />

1<br />

E =β( n⊗ n)<br />

+ ( γ⊗ n+ n⊗γ ) , β=<br />

+<br />

[ un<br />

] dS, γ<br />

+<br />

[ ut<br />

] dS<br />

2<br />

N ∫ = S ∫<br />

(13)<br />

S<br />

s<br />

d<br />

s<br />

-190-


[ ut<br />

] is the tangential vector if displacement jump. By making used of Eshelby’s basic solution and based on<br />

appropriate problem decompositions (Zhu et al. 2008), the macroscopic free energy function of cracked material<br />

can be determined:<br />

1 pl s pl 1<br />

2<br />

W = ( E−E ): C :( E− E ) + ⎡H0( 1−χnω) β + H1( 1 −χtω)<br />

γγ . ⎤<br />

2 2ω (14)<br />

s<br />

3 E<br />

s<br />

H0 = , H<br />

2 1 = H0( 1 −v<br />

/2)<br />

16<br />

s<br />

1 − ( v )<br />

(15)<br />

⎧<br />

s<br />

2<br />

s<br />

⎪<br />

16( 1−v<br />

) 16( 1−v<br />

)<br />

⎪ χ n = , χ t =<br />

for the dilute scheme<br />

s<br />

s<br />

⎪ 31 ( −2v<br />

) 32 ( −v<br />

)<br />

⎪<br />

⎨ χ n = 0, χ t = 0 for the scheme MT<br />

⎪<br />

s<br />

⎪ 128 16( 7 − 5v<br />

,<br />

)<br />

⎪ χ n = χ t =<br />

for the scheme PCW<br />

45 s<br />

⎪<br />

45( 2 − v )<br />

⎩<br />

(16)<br />

By making standard derivative, macroscopic state law (constitutive relations) as well as the thermodynamic<br />

conjugated force to inelastic strains can be deduced and expressed as:<br />

pl<br />

pl<br />

s pl β ∂W ∂W ∂E γ ∂W ∂W ∂E<br />

∑= C :( E− E ) ; F =− =− ; F =− =−<br />

∂β<br />

pl<br />

pl<br />

∂E<br />

∂β ∂γ ∂E<br />

∂γ<br />

(17)<br />

<strong>The</strong> frictional sliding is then described by the definition of a friction criterion. In geomaterials, Coulomb type<br />

criteria are generally used. <strong>The</strong> friction criterion is a scalar-valued function of the thermodynamic conjugate<br />

forces such as:<br />

g<br />

c<br />

σ<br />

γ β<br />

= F +μ F ≤ (18)<br />

μ c is the local frictional coefficient.<br />

( ) c 0<br />

On the other hand, the thermodynamic force conjugated with damage variable van also deduced (Zhu et al.<br />

2008):<br />

d ∂W<br />

1 2<br />

F = − = ( H0β + H1γγ<br />

. )<br />

(19)<br />

∂ω 2ω<br />

<strong>The</strong> damage evolution due to crack propagation is governed by the damage criterion which can be written in the<br />

following general form (Zhu et al. 2008):<br />

d d<br />

f F , ω = F −R ω<br />

(20)<br />

( ) ( )<br />

<strong>The</strong> scalar-valued function R ( ω)<br />

defines the current materials toughness against damage evolution. Various<br />

forms can be adopted based on experimental identification. We can clearly see that the damage evolution is<br />

inherently coupled with the frictional sliding due to the inter-dependency between the thermodynamic forces.<br />

Note that the general framework above formulated for a single family of microcracks can be extended to the<br />

case with multi families of microcracks. An appropriate numerical integration algorithm over all the crack<br />

families should be adopted. Uncoupled and coupled schemes can be considered regarding interactions between<br />

crack families, for detail see Zhu et al. (2009).<br />

Example of application<br />

<strong>The</strong> proposed micromechanical damage model has been applied to simulating mechanical behaviors of different<br />

brittle rocks (Zhu et al. 2009, 2008b). In Figure 5, we show the simulation of uniaxial tension test on concrete<br />

with the comparison between three homogenization schemes. One can see that macroscopic responses depend<br />

on the scheme used and the PCW scheme provides the best simulation by taking into account crack interaction<br />

and spatial distribution. Figure 6 shows the rosette surface of damage density distribution during uniaxial test. It<br />

is seen that the maximum damage density is obtained in the direction of applied tensile stress. An example of<br />

experimental verification is shown in figure 7. Two triaxial compression tests performed on typical granite with<br />

different confining pressures are simulated using the MT scheme. We can see that the main features of<br />

mechanical responses are correctly reproduced. In particular, the micromechanical model well describes the<br />

volumetric dilatancy induced by frictional sliding along microcracks.<br />

-191-


10<br />

[ MPa<br />

11 ]<br />

∑ [ ]<br />

8<br />

6<br />

Expérience<br />

4<br />

2<br />

E % 11<br />

0<br />

0.00 0.02 0.04 0.06 0.08 0.10 0.12<br />

Figure 5 Simulation of uniaxial tension test and<br />

influence of homogenization scheme (data from<br />

Bazant and Pijaudier-Cabot 1989)<br />

Figure 6 Rosette surface of damage density<br />

distribution during uniaxial tension<br />

E33<br />

Ev<br />

400<br />

300<br />

∑ −∑ [ ]<br />

11 MPa<br />

33<br />

E11<br />

E33<br />

Ev<br />

500<br />

400<br />

300<br />

−∑ [ ]<br />

∑ 11 MPa<br />

33<br />

E11<br />

200<br />

Pc=10MPa<br />

200<br />

Pc=20<br />

100<br />

0<br />

Experiments<br />

With dilatation<br />

Without dilatation<br />

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0<br />

100<br />

0<br />

Experiments<br />

Model<br />

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0<br />

Figure 7 Comparisons between experimental data and simulations for triaxial compression tests with different<br />

confining pressures on Lac du Bonnet granite (Zhu et al. 2008b)<br />

MESOMECHANICAL MODELLING OF QUASI-BRITTLE ROCKS<br />

Hill’s incremental method<br />

In brittle geomaterials, the local behavior of material phase is described by linear elastic model. <strong>The</strong><br />

nonlinear and inelastic deformation is due to propagation of microcracks and frictional sliding. In quasi-brittle<br />

geomaterials, such as hardened clay and cement-based materials, plastic deformation can develop inside<br />

different mineral phases. As mentioned above, the macroscopic deformation of clayey rocks (argillite, shale) is<br />

controlled by plastic sliding of clay sheets and propagation of microcracks around grain-matrix interfaces as<br />

well as inside clay matrix. <strong>The</strong> general linear homogenization procedure mentioned above cannot be directly<br />

applied to heterogeneous materials containing inelastic phases. <strong>The</strong> determination of effective properties of<br />

nonlinear heterogeneous materials has been being the most attractive topic in solid mechanics during the last<br />

decades. Various analytical and numerical techniques have been proposed. Due to the limited page number here,<br />

we don’t attend to give an exhaustive review of existing methods and the readers can refer to the reference paper<br />

(Ponte-Castaneda and Suquet 1998). All these methods are based on some appropriate linearization procedure<br />

such as linear comparison composite. Among these methods, the incremental model initially proposed by Hill<br />

(1965) for polycrystal composites has been used and extended to modeling different kinds of heterogeneous<br />

materials. <strong>The</strong> main advantage of this method is its incremental formulation allowing consideration of complex<br />

loading paths. Further the numerical implementation in standard computer codes is quite easy. In the framework<br />

of feasibility study for geological disposal of nuclear waste storage, the incremental method has been applied to<br />

describe macroscopic plastic damage behavior of argillite by the authors (Abou-Charka et al. 2008, Jiang et al.<br />

2009a, 2009b). <strong>The</strong> incremental method is based on the extension of linear homogenization procedure to<br />

nonlinear materials described by incremental constitutive relations. <strong>The</strong> principle of the incremental method<br />

consists in the determination of the overall tangent stiffness tensor for each incremental loading step, based on<br />

the knowledge of local constitutive behaviors. <strong>The</strong> local constitutive relations for each phase can be written as:<br />

σ & ( z) = L ( z): ε& ( z)<br />

(21)<br />

σ& and ε& denotes respectively local stress and strain rate fields while L ( z)<br />

is the local tangent stiffness tensor<br />

which depends on the position point ( z ) inside the local scale frame.<br />

Due to the fact that the local constitutive relations are expressed in an incrementally linear form with nonlinear<br />

tangent operator, the classical Eshelby-based linear homogenization procedure as that mentioned above can be<br />

-192-


extended to the resolution of the nonlinear problem. For this purpose, a tangent concentration tensor A ( z)<br />

is<br />

introduced to relate the local strain rate to that of macroscopic strain:<br />

ε &( z) = A ( z):<br />

E&<br />

(22)<br />

Substituting (22) for (21) and defining the macroscopic stress tensor rate by Σ & = σ& , the macroscopic<br />

incremental constitutive relations read:<br />

Σ= & L hom : E&<br />

(23)<br />

<strong>The</strong> macroscopic tangent stiffness tensor is given by:<br />

hom<br />

L = L( z): A ( z)<br />

(24)<br />

One can see that the determination of the macroscopic tangent operator is based on the volume averaging of<br />

local tangent operator and concentration tensor which depend on at point position at the local scale. Since the<br />

local strain and stress fields are strongly heterogeneous at the local scale, it is nearly impossible to obtain the<br />

analytical form of such volume average. <strong>The</strong>refore, for the effective implementation of the incremental method,<br />

an approximation of the tangent operator in each phase is needed. For this purpose, the following hypothesis is<br />

adopted: at any point z inside the phase r , the stress rate can be related to strain rate by a uniform tangent<br />

operator:<br />

∀z∈Vr<br />

, σ & ( z) = L r : ε& ( z)<br />

(25)<br />

Such an assumption appears very strong regarding non-uniform distribution of local fields and then represents<br />

indeed a critical aspect of the incremental method. However, this assumption makes it possible to deduce in an<br />

analytical way the macroscopic tangent operator. In practice, the local stiffness tensor L r is evaluated for a<br />

suitable reference strain state, which is classically taken as the average value of local strain field in the phase<br />

( r ), noted by ε . r<br />

E &<br />

Figure 8 Schematic presentation of RVE of argillite<br />

In the following, we present the application of incremental method to elastoplastic damage modeling of argillite.<br />

<strong>The</strong> argillite is seen as a two phase composite with clay matrix (index m ) and mineral inclusions (quartz and<br />

calcite, index I ) as shown in Figure 8. Using the linearization method defined above, the tangent concentration<br />

law (22) for each material phase becomes:<br />

⎧ε &<br />

⎪<br />

= Am<br />

: E&<br />

⎨<br />

m<br />

(26)<br />

ε & = AI<br />

: E&<br />

⎪⎩ I<br />

A m and A I are the tangent concentration tensors respectively for the clay matrix and for mineral inclusions.<br />

Using now the Mori-Tanaka scheme (1973), the tangent concentration tensors are given by:<br />

⎧ 1<br />

−1<br />

⎡ 0<br />

− ⎤<br />

⎪Am = fm f ⎡<br />

I I :( I m)<br />

⎤<br />

⎢<br />

I + I + P L −L<br />

⎪ ⎣ ⎣<br />

⎦ ⎥<br />

⎦<br />

⎨<br />

(27)<br />

1 1<br />

−1<br />

⎪ 0<br />

− ⎡ 0<br />

− ⎤<br />

⎪AI = ⎡I + PI :( LI − Lm) ⎤ : fm + f ⎡<br />

I + I :( I − m)<br />

⎤<br />

⎣ ⎦ ⎢<br />

I I P L L<br />

⎣ ⎣ ⎦ ⎥<br />

⎩<br />

⎦<br />

0<br />

E<br />

<strong>The</strong> Hill tensor ( P I ) is related to Eshelby tensor ( S ) by PI<br />

0 = S E : L −<br />

m 1 . Using the incremental strain<br />

concentration tensors, the macroscopic tangent operator of the argillite can be determined:<br />

hom<br />

L = f L : A + f L : A (28)<br />

m m m I I I<br />

For a given loading history, the tangent stiffness tensors ( Lm<br />

for clay matrix and L I for mineral inclusions) are<br />

updated in each incremental loading step according the constitutive models used for each material phase.<br />

-193-


As the main difference with the linear problem, the Eshelby tensor depends now on the tangent operator of the<br />

clay matrix L m . Depending on the local constitutive model used for clay matrix, the tangent operator can<br />

become anisotropic in nature as it is the case for non-associated plastic model. As a consequence, the Eshelby<br />

(or Hill) tensor can not be evaluated in analytical way and a suitable numerical integration method should be<br />

adopted. Further, as shown in some previous works (Doghri, and Ouaar, 2003; Chaboche et al. 2005;<br />

Abou-Charka et al., 2008), the use of anisotropic local tangent operator in the determination of incremental<br />

strain concentration tensors general leads to too stiff responses of composites. <strong>The</strong> physical reason of such a<br />

result is mainly the fact that the heterogeneity of local fields is neglected in the proposed method. As a<br />

pragmatic corrective method, it is proposed to decompose the local tangent operator into an isotropic part and<br />

anisotropic one and then to use the isotropic part in the evaluation of the macroscopic tangent operator. As a<br />

example, we propose here to use the isotropic part of local tangent operator of the clay matrix only for the<br />

evaluation of Eshelby tensor. <strong>The</strong> isotropic part is determined by (Bornert et al. 2001):<br />

δij<br />

are components of the second order unit tensor,<br />

iso<br />

1<br />

L m = ( J :: L m ) J + ( K:: L m ) K= = 3k<br />

t J + 2μ<br />

t K<br />

5<br />

(29)<br />

⎧ 1<br />

Iijkl = ( δikδ jl +δilδjk<br />

)<br />

⎪ 2<br />

⎨<br />

⎪ 1<br />

Jijkl = δijδ kl ; Kijkl = Iijkl −Jijkl<br />

⎪⎩ 3<br />

(30)<br />

k t and μ t denote respectively the tangent bulk and shear<br />

modulus related to the isotropic part of the tangent operator of the clay matrix. Considering now spherical<br />

mineral inclusions, the Eshelby tensor is given by:<br />

0 iso 3kt 6 kt + 2μt<br />

SI( Lm<br />

) = J + K (31)<br />

3k<br />

+ 4μ 53k<br />

+ 4μ<br />

t t t t<br />

And the Hill tensor for the inclusion phase becomes:<br />

0 0 iso 1<br />

P = S ( L ): L (32)<br />

Example of application<br />

I I m m −<br />

For the argillite studied here, the clay matrix is described by an elastoplastic model coupled with damage due to<br />

microcracks. <strong>The</strong> mineral grains (quartz and calcite) are described by the linear elastic model. An important<br />

challenge in micromechanical modeling of heterogeneous materials is the determination of local behavior of<br />

different mineral phases. Direct identification remains a serious technical challenge even if significant advances<br />

have been made such as nano indentation tests. In the present studies, an indirect method is adopted. <strong>The</strong><br />

model’s parameters are determined for a given mineral composition by numerical fitting. And then the obtained<br />

set of parameters is checked for other mineral compositions. Note that the main advantage of micromechanical<br />

modeling is that the mineral composition becomes direct input data.<br />

In Figure 9, we show the simulation of a triaxial compression test with 10MPa confining pressure. <strong>The</strong><br />

macroscopic responses of argillite are compared with experimental data and also with those of clay matrix alone.<br />

We can see that there is a good agreement between numerical results and test data. <strong>The</strong> main features of argillite<br />

mechanical behavior are well reproduced by the micromechanical model. Further, the effects of mineral<br />

inclusions are clearly illustrated. In Figure 10, a series of simulations are presented. <strong>The</strong>se triaxial compression<br />

tests are performed under different confining pressures and on the argillites samples with different mineral<br />

compositions. Again, the numerical results are in good agreement with experimental data. In most important<br />

feature of the micromechanical model is that the influences of mineral composition are inherently taken into<br />

account in the formulation of the model.<br />

-194-


469.1m,f m =55%,f I =45%<br />

-50<br />

Σ 3 -Σ 1 (MPa)<br />

-40<br />

-30<br />

Argilite<br />

Matrix<br />

1.0<br />

E 1 (×10 -2 )<br />

0.5<br />

0.0<br />

-20<br />

-10<br />

0<br />

-0.5<br />

-1.0<br />

E 3 (×10 -2 )<br />

Figure 9 Comparison of macroscopic responses of argillite and those of clay matrix alone during a triaxial<br />

compression test with 10MPa confining pressure – influence of mineral inclusions<br />

-1.5<br />

-2.0<br />

466.8m,f m =51%,f I =49%<br />

451.5m,f m =49%,f I =51%<br />

-40<br />

Σ 3 -Σ 1 (MPa)<br />

-50<br />

Σ 3 -Σ 1 (MPa)<br />

-30<br />

-40<br />

-30<br />

-20<br />

-20<br />

1.0<br />

E 1 (×10 -2 )<br />

0.5<br />

-10<br />

0.0 -0.5<br />

0<br />

(a: Pc=0)<br />

E 3 (×10 -2 )<br />

-1.0<br />

-1.5<br />

1.0<br />

E 1 (×10 -2 )<br />

0.5<br />

-10<br />

0.0 -0.5 -1.0<br />

0<br />

(b: Pc=5MPa)<br />

E 3 (×10 -2 )<br />

-1.5<br />

-2.0<br />

451.4m,f m =47%,f I =53%<br />

482.2m,f m =60%,f I =40%<br />

-60<br />

Σ 3 -Σ 1 (MPa)<br />

-50 Σ 3 -Σ 1 (MPa)<br />

-50<br />

-40<br />

-40<br />

-30<br />

-30<br />

-20<br />

-20<br />

E 1 (×10 -2 )<br />

1.0 0.5<br />

-10<br />

0.0<br />

0<br />

-0.5<br />

-1.0<br />

E 3 (×10 -2 )<br />

-1.5<br />

-2.0<br />

E 1 (×10 -2 )<br />

0.5<br />

0.0<br />

-10<br />

0<br />

-0.5<br />

-1.0<br />

E 3 (×10 -2 )<br />

-1.5<br />

-2.0<br />

(c: Pc=10MPa)<br />

(d: Pc=20MPa )<br />

Figure 10 Simulation of triaxial compression tests under different confining pressures and on argillite<br />

samples with different mineral compositions<br />

CONCLUSIONS<br />

In this short review paper, we have presented the general framework for multi-scale modelling of heterogeneous<br />

geomaterials using rigorous homogenization procedures. Physical mechanisms involved at different material<br />

scales can be directly involved in the formulation of the models. Influences of mineral compositions and<br />

microstructure evolution can be properly taken into account. Such models appear very promising for the<br />

physics-based modelling of geomaterials. It remains a number of challenges for future works such as influences<br />

of microstructure morphology, intra phase heterogeneities, multi-physical and chemical coupling, linking<br />

different materials scales from nano to macro scales.<br />

-195-


ACKNOWLEDGMENTS<br />

<strong>The</strong> present work is partially supported by French National Agency for radioactive waste management<br />

(ANDRA), which is gratefully acknowledged.<br />

REFERENCES<br />

Abou-Charka Guery, A., Cormery, F., Shao, J.F. and Kondo, D. (2008). “A micromechanical model of<br />

elastoplastic and damage behaviour of a cohesive geomaterials”, International journal of solids and<br />

structures, 45(5), 1406–1429<br />

ANDRA (2005), Geological referential of argillite of the site Meuse/Haute Marne<br />

Benveniste, Y. (1987). “A new approach to the application of Mori-Tanaka’s theory in composite materials”,<br />

Mechanics of materials, 6, 147-157.<br />

Bazant, Z.P., Pijaudier-Cabot, G. (1989), “Measurement of characteristic length of nonlocal continuum”, Journal<br />

of Engineering Mechanics, 115(4), 755-767.<br />

Budiansky, B., O’Connell, R.-J. (1976). “Elastic moduli of a cracked solid”, International Journal of Solids and<br />

Structures, 12, 81-97.<br />

Chaboche, J.L., Kanouté, P., Roos, A. (2005). “On the capabilities of mean-field approaches for the description<br />

of plasticity in metal matrix composites”, International Journal of Plasticity, 21, 1409-1434<br />

Chiarelli A.S., Shao J.F. and Hoteit N. (2003), “Modelling of elastoplastic damage behaviour of a claystone”,<br />

International Journal of Plasticity , 19, 23-45<br />

Doghri, I., Ouaar, A. (2003). “Homogenization of two-phase elasto-plastic composite materials and structures.<br />

Study of tangent operators, cyclic plasticity and numerical algorithms”, International Journal of Solids and<br />

Structures, 40, 1681–1712<br />

Eshelby, J.D. (1957). “<strong>The</strong> determination of the elastic field of an ellipsoidal inclusion and related problems”,<br />

Proc. R. Soc. Lond., A241, 375-396.<br />

Hill, R. (1965), “Continuum micro-mechanics of elastoplastic polycrystals”, Journal of Mechanics and Physics<br />

of Solids, 13, 89-101<br />

Horii, H., Nemat-Nasser, S. (1983). “Overall moduli of solids with microcracks: load-induced anisotropy”, J.<br />

Mech. Phys. Solids, 31(2), 155-171.<br />

Jia Y., Bian H. B., Su K., Kondo D. and J. F. Shao (2010). “Elastoplastic damage modelling of desaturation<br />

and resaturation in argillites”, Int. J. for numerical and analytical methods in Geomechanics, 34, 187-220<br />

Jiang, T. Abou-Chakra Guéry, A., Kondo, D. and Shao, J.F. (2009a). “Multi-scale modeling for inelastic<br />

behavior of a cohesive geomaterials”, Mechanics Research Communications, 36(6), 673-681<br />

Jiang, T. and Shao, J.F. (2009b). “On the incremental approach for nonlinear homogenization of composite<br />

and influence of isotropisation”, Computational Material Science, 46(2), 447-451<br />

Mori, T., Tanaka, K. (1973). “Average stress in matrix and average elastic energy of materials with misfitting<br />

inclusions”, Acta Metallurgica, 21, 571-574.<br />

Mura, T. (1987). Micromechanics of defects in solids, Martinus Nijhoff Publ., <strong>The</strong> Hague, Boston.<br />

Ponte-Castaneda, P., Willis, J.R. (1995). “<strong>The</strong> effect of spatial distribution on the behavior of composite<br />

materials and cracked media”, J. Mech. Phys. Solids, 43, 1919-1951.<br />

Ponte-Castaneda, P., Suquet P. (1998). “Nonlinear composites”, in Advances in Applied Mechanics, 34, 171-302.<br />

Shao J. F., Jia Y., Kondo D. and Chiarelli A.S. (2006). “A coupled elastoplastic damage model for semi-brittle<br />

materials and extension to unsaturated conditions”, Mechanics of Mechanics, 38, 218-232<br />

Zaoui, A. (2002). “Continuum micromechanics: survey”, Journal of Engineering Mechanics, ASCE, 128(8),<br />

808-816.<br />

Zhu Q. Z., Kondo, D. and Shao, J. F. (2008a). “Micromechanical analysis of coupling between anisotropic<br />

damage and friction in quasi brittle materials: role of the homogenization scheme”, International Journal of<br />

Solids and Structures, 45(5), 1385–1405.<br />

Zhu Q.Z., Kondo D., Shao J.F. and Pensee V. (2008b). “Micromechanical modelling of anisotropic damage in<br />

brittle rocks and application”, Int. Journal of Rock Mechanics and Mining Sciences, 45(4), 467-477<br />

Zhu, Q.Z., Kondo, D. and Shao, J.F. (2009). “Homogenization-based analysis of anisotropic damage in brittle<br />

materials with unilateral effect and interactions between microcracks”, Int. J. for numerical and analytical<br />

methods in Geomechanics, 33(6), 749-772<br />

-196-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

台 灣 地 錨 和 地 錨 邊 坡 破 壞 態 樣<br />

廖 洪 鈞 鄭 世 豪<br />

( 台 灣 科 技 大 學 營 建 工 程 系 , 台 灣 台 北 10607)<br />

摘 要 : 台 灣 地 區 因 山 坡 地 質 破 碎 且 風 化 情 形 嚴 重 , 地 錨 邊 坡 的 破 壞 常 伴 隨 多 種 因 素 而 發 生 ,<br />

較 常 見 情 況 為 邊 坡 表 層 覆 土 受 雨 水 滲 透 和 風 化 後 發 生 淺 層 滑 動 , 造 成 坡 面 擋 土 結 構 滑 動 , 進 而 導 致<br />

整 個 背 拉 地 錨 系 統 的 破 壞 。 此 外 , 由 本 文 收 集 之 地 錨 邊 坡 破 壞 案 例 顯 示 , 若 地 錨 之 防 蝕 保 護 不 佳 ,<br />

即 使 初 期 狀 況 良 好 , 長 期 使 用 也 會 成 為 地 錨 邊 坡 破 壞 之 潛 在 因 素 。 目 前 台 灣 地 區 所 使 用 之 地 錨 完 工<br />

至 今 大 多 已 服 務 數 十 年 之 久 , 其 現 況 和 功 能 性 須 作 有 系 統 的 調 查 , 才 可 確 保 地 錨 邊 坡 的 長 期 安 全<br />

性 。 本 文 將 從 台 灣 地 區 常 見 之 地 錨 和 地 錨 邊 坡 破 壞 態 樣 中 , 歸 納 出 其 破 壞 原 因 , 並 建 議 適 當 的 維 護<br />

檢 測 和 劣 化 處 理 對 策 。<br />

關 鍵 詞 : 地 錨 地 錨 邊 坡 破 壞 態 樣 淺 層 滑 動 防 蝕 保 護<br />

FAILURE TYPES OF ANCHORS AND ANCHORED SLOPES IN TAIWAN<br />

H. J. Liao 1 and S. H. Cheng 1<br />

1 Department of Construction Engineering, Taiwan <strong>University</strong> of Science and Technology, Taipei 10607, Taiwan<br />

Abstract: Due to the complicated geological condition of nature slopes in Taiwan, failure of anchored slopes is<br />

usually resulted by a combination of causes. <strong>The</strong> most commonly encountered one is a shallow slip behind or<br />

above the anchored slope retaining structure followed by a total failure of the entire anchor system. Rainfall<br />

infiltration and weathering of slope surface material are the main causes to trigger shallow landslip. But, totally<br />

or partially loss of tieback anchor load due to poor corrosion protection is often the reason for the failure of<br />

anchored slopes under the adverse geological and weather conditions of Taiwan. So it is necessary to carry out a<br />

systematic investigation on current status of anchors which had been in service for decades to understand the<br />

corrosion condition of anchors and to keep the safety of anchored slope. This paper summarizes the common<br />

cases and failure types of anchors and anchored slopes failures in Taiwan. Some counter-measures against<br />

anchor corrosion are proposed including the routine anchor investigation program and the improved details of<br />

corrosion protection measure for the entire anchor (e.g., anchor head, steel strands in fixed and free ends of the<br />

anchor).<br />

Keywords: anchor, anchored slope, failure type, shallow slip, corrosion protection.<br />

一 、 前 言<br />

長 期 以 來 , 地 錨 一 直 被 廣 泛 地 應 用 於 坡 地 工 程 上 。 台 灣 地 錨 的 使 用 , 從 首 次 用 於 達 見 水 壩 之 興<br />

建 迄 今 , 也 有 五 十 年 以 上 的 歷 史 。 但 因 台 灣 地 區 的 地 錨 絕 大 多 數 是 使 用 在 地 下 水 以 下 或 是 地 下 水 豐<br />

沛 區 域 , 錨 碇 地 層 多 位 於 土 壤 與 岩 層 交 界 的 風 化 岩 層 或 節 理 發 達 的 岩 層 、 甚 至 錨 碇 在 土 壤 中 。 在 如<br />

此 不 利 之 外 在 環 境 下 , 如 何 維 持 地 錨 構 件 ( 包 括 錨 頭 、 自 由 段 和 錨 碇 段 鋼 絞 線 ) 的 防 蝕 性 與 地 錨 本 身<br />

的 功 能 性 , 一 直 都 是 工 程 界 所 關 心 的 課 題 。 雖 然 地 錨 在 台 灣 邊 坡 的 使 用 有 少 數 破 壞 案 例 , 但 大 多 數<br />

地 錨 邊 坡 大 致 尚 能 維 持 穩 定 , 只 是 因 缺 乏 長 期 之 追 蹤 檢 測 資 料 , 且 有 些 地 錨 使 用 年 代 久 遠 , 其 現 況<br />

基 金 項 目 : 台 灣 國 道 高 速 公 路 局 資 助 項 目 (097A22P002)<br />

作 者 簡 介 : 廖 洪 鈞 (1957-), 男 , 台 灣 台 北 人 , 教 授 , 副 校 長 , 從 事 校 務 行 政 、 教 學 、 地 盤 改 良 和 坡 地 工 程 等 研 究 工 作 。<br />

-197-


如 何 不 得 而 知 , 在 安 全 上 , 存 在 著 相 當 大 的 不 確 定 性 。<br />

二 、 常 見 地 錨 損 壞 型 態<br />

通 常 以 目 視 可 見 的 地 錨 損 壞 型 態 大 致 可 分 為 錨 頭 / 鋼 絞 線 銹 蝕 、 錨 頭 脫 落 和 鋼 絞 線 斷 裂 等 三<br />

種 , 至 於 自 由 段 或 是 錨 碇 段 鋼 絞 線 的 銹 蝕 , 以 及 地 錨 預 力 喪 失 的 部 份 , 則 無 法 以 目 視 方 式 檢 測 , 而<br />

須 輔 以 儀 器 檢 測 方 式 進 行 確 認 。 以 下 內 容 就 可 目 視 檢 測 之 地 錨 損 壞 型 態 進 行 說 明 :<br />

2.1. 錨 頭 / 鋼 絞 線 銹 蝕<br />

2.1.1. 發 生 原 因<br />

台 灣 地 區 地 錨 之 錨 頭 絕 大 多 數 都 以 混 凝 土 護 蓋 保 護 , 但 因 錨 頭 保 護 之 時 間 點 都 是 在 整 個 地 錨 護<br />

坡 工 程 接 近 尾 聲 時 進 行 , 所 以 這 些 混 凝 土 可 能 在 現 場 以 少 量 拌 合 方 式 製 作 , 混 凝 土 品 質 較 難 掌 控 ,<br />

而 影 響 到 混 凝 土 的 水 密 性 , 且 在 風 吹 日 曬 雨 淋 的 情 況 下 , 易 使 包 覆 在 混 凝 土 護 蓋 中 之 錨 頭 銹 蝕 ; 再<br />

者 這 些 混 凝 土 都 是 以 二 次 施 工 方 式 灌 置 , 與 地 錨 承 座 之 混 凝 土 間 易 有 冷 縫 存 在 , 形 成 水 路 , 使 得 錨<br />

頭 銹 蝕 。 錨 頭 銹 蝕 後 , 體 積 會 膨 脹 , 造 成 混 凝 土 護 蓋 的 龜 裂 , 更 進 一 步 加 速 錨 頭 的 銹 蝕 。 至 於 , 錨<br />

頭 護 蓋 之 混 凝 土 品 質 可 以 榔 頭 敲 擊 做 初 步 判 斷 , 通 常 混 凝 土 品 質 越 差 , 錨 頭 之 銹 蝕 越 嚴 重 。<br />

自 由 段 鋼 絞 線 銹 蝕 多 半 是 因 自 由 段 防 蝕 保 護 不 足 所 致 , 通 常 在 自 由 段 與 錨 頭 交 界 處 , 為 方 便 鋼<br />

絞 線 與 夾 片 的 咬 合 , 施 工 時 都 會 將 鋼 絞 線 之 護 管 切 斷 , 而 使 得 該 處 鋼 絞 線 出 露 , 形 成 防 蝕 保 護 死 角 ,<br />

而 該 鋼 絞 線 出 露 處 又 幾 乎 在 地 表 處 , 鋼 絞 線 極 易 受 到 大 氣 水 分 和 降 雨 之 影 響 而 銹 蝕 ( 圖 1)。 此 外 ,<br />

地 錨 孔 有 時 也 會 因 施 工 過 程 中 , 未 確 實 將 地 錨 孔 以 水 泥 漿 灌 滿 , 使 得 地 錨 孔 變 成 地 下 水 之 排 水 路<br />

徑 , 從 錨 頭 處 流 出 ( 圖 2), 易 造 成 錨 頭 及 其 下 方 鋼 絞 線 的 銹 蝕 。<br />

圖 1 錨 頭 下 方 鋼 絞 線 出 露 和 銹 蝕 示<br />

意 圖 ( 重 繪 自 [1] )<br />

圖 2 地 下 水 由 錨 頭 處 流 出 和 錨 頭 下 方 鋼 絞 線 銹 蝕 情 況<br />

2.1.2. 改 善 方 式<br />

由 於 場 鑄 混 凝 土 護 蓋 常 有 水 密 性 不 佳 和 二 次 施 工 之 冷 縫 問 題 , 不 宜 單 獨 使 用 為 錨 頭 保 護 措 施 ,<br />

以 免 日 後 銹 蝕 。 通 常 錨 頭 發 生 銹 蝕 後 , 即 很 難 再 更 換 夾 片 和 鋼 絞 線 , 換 句 話 說 , 此 時 地 錨 已 經 很 難<br />

再 回 復 到 新 品 時 的 性 能 , 頂 多 只 能 避 免 其 繼 續 銹 蝕 , 而 銹 蝕 嚴 重 者 , 甚 至 需 廢 除 重 作 。<br />

錨 頭 之 檢 測 可 依 圖 3 之 流 程 圖 進 行 , 通 常 目 視 和 錨 頭 敲 擊 方 式 是 最 基 本 之 檢 測 方 式 , 但 卻 很 有<br />

效 。 若 檢 測 結 果 無 明 顯 之 異 狀 , 則 該 地 錨 只 需 列 入 例 行 檢 查 對 象 即 可 ; 若 有 明 顯 的 異 狀 , 則 可 將 錨<br />

-198-


頭 混 凝 土 護 蓋 敲 除 , 檢 視 錨 頭 之 銹 蝕 狀 況 。 若 情 況 尚 可 , 則 將 鋼 絞 線 和 夾 片 外 表 混 凝 土 和 銹 屑 清 除<br />

乾 淨 , 塗 上 防 銹 漆 , 並 蓋 上 填 有 防 銹 油 脂 之 鍍 鋅 錨 頭 護 蓋 , 再 以 油 槍 將 護 蓋 內 部 灌 滿 防 銹 油 脂 ( 圖<br />

4), 以 完 全 斷 絕 水 氣 之 入 侵 。 然 而 , 為 確 保 錨 頭 之 水 密 性 , 可 於 鍍 鋅 錨 頭 護 蓋 外 側 , 再 澆 灌 混 凝 土<br />

護 蓋 保 護 之 。<br />

圖 3 錨 頭 檢 測 和 維 護 流 程 示 意 圖<br />

圖 4 本 文 建 議 之 錨 頭 保 護 方 式<br />

錨 頭 經 防 銹 處 理 後 , 仍 須 評 估 地 錨 因 錨 頭 構 件 銹 蝕 所 喪 失 的 拉 拔 力 多 寡 。 雖 然 因 銹 蝕 所 減 少 的<br />

鋼 絞 線 斷 面 積 , 可 用 金 相 檢 查 技 術 檢 測 之 , 但 錨 頭 構 件 其 實 是 握 線 器 ( 含 夾 片 、 錨 座 ) 和 鋼 絞 線 等 之<br />

-199-


共 同 組 合 體 , 當 各 構 件 發 生 銹 蝕 後 , 要 評 估 錨 頭 整 體 功 能 性 的 減 少 量 , 目 前 尚 未 有 成 熟 技 術 可 進 行<br />

偵 測 。 實 務 上 , 可 選 擇 少 數 幾 支 錨 頭 銹 蝕 之 地 錨 進 行 揚 起 試 驗 , 估 計 其 剩 餘 之 抗 拔 能 力 , 作 為 設 計<br />

新 擋 土 措 施 之 參 考 。 若 選 擇 新 作 地 錨 時 , 其 鋼 絞 線 組 立 和 錨 碇 段 、 自 由 段 、 以 及 錨 頭 之 防 蝕 保 護 方<br />

式 可 參 考 圖 5 施 作 , 其 中 須 特 別 注 意 的 是 , 在 錨 頭 下 方 、 自 由 段 鋼 絞 線 和 封 漿 器 等 位 置 須 有 相 關 保<br />

護 措 施 , 以 提 供 錨 頭 下 方 鋼 絞 線 的 防 蝕 保 護 。<br />

圖 5 新 作 地 錨 之 鋼 絞 線 組 立 和 全 長 之 防 蝕 保 護 措 施<br />

2.2. 錨 頭 脫 落<br />

2.2.1. 發 生 原 因<br />

台 灣 地 區 大 部 份 地 錨 都 使 用 握 線 器 ( 含 夾 片 及 錨 座 ) 作 為 預 力 鎖 定 的 構 件 , 握 線 器 能 否 有 效 將 鋼<br />

絞 線 鎖 定 , 直 接 影 響 地 錨 的 使 用 功 能 。 當 錨 頭 產 生 銹 蝕 時 , 夾 片 與 鋼 絞 線 間 的 咬 合 能 力 將 會 大 受 影<br />

響 , 由 圖 6 林 肯 大 郡 邊 坡 地 錨 的 銹 蝕 磨 損 夾 片 可 見 , 夾 片 的 銹 蝕 以 及 鋼 絞 線 與 夾 片 間 之 磨 痕 , 兩 者<br />

皆 會 造 成 夾 片 與 鋼 絞 線 間 的 咬 合 失 效 , 進 而 引 致 錨 頭 脫 落 ( 圖 7)。 錨 頭 脫 落 後 , 若 鋼 絞 線 的 露 頭 呈<br />

現 平 整 排 列 , 則 可 研 判 錨 頭 脫 落 之 當 時 , 鋼 絞 線 幾 乎 未 受 預 力 ; 反 之 , 若 錨 頭 脫 落 當 時 , 鋼 絞 線 有<br />

受 到 預 力 , 則 鋼 絞 線 將 會 往 地 錨 孔 內 縮 , 在 外 頭 看 不 到 鋼 絞 線 的 露 頭 ( 圖 7)。<br />

圖 6 夾 片 銹 蝕 與 磨 痕<br />

圖 7 錨 頭 脫 落 ( 左 ) 無 預 力 鋼 絞 線 未 內 縮 ( 右 ) 有 預 力 鋼 絞 線 內 縮<br />

此 外 , 另 一 錨 頭 脫 落 的 原 因 則 是 錨 頭 構 件 不 良 , 當 鋼 絞 線 承 受 拉 力 後 , 需 藉 由 錨 頭 握 線 器 , 將<br />

拉 力 由 鋼 絞 線 移 轉 到 擋 土 結 構 上 , 以 穩 定 邊 坡 。 一 般 工 作 地 錨 每 條 鋼 絞 線 之 受 力 都 在 10 公 噸 左 右 ,<br />

因 此 夾 片 與 鋼 絞 線 之 咬 合 點 會 受 到 很 大 的 力 量 , 一 旦 其 材 質 不 良 或 是 加 工 製 作 的 精 度 不 佳 , 很 容 易<br />

-200-


發 生 夾 片 咬 合 失 敗 , 錨 頭 握 線 器 無 法 固 定 住 鋼 絞 線 的 現 象 。 若 牆 後 土 水 壓 力 在 大 雨 後 增 加 、 或 是 邊<br />

坡 開 始 產 生 下 滑 而 位 移 時 , 即 會 發 生 錨 頭 脫 落 現 象 。<br />

另 外 , 還 有 一 種 造 成 夾 片 磨 損 及 鋼 絞 線 滑 脫 的 狀 況 , 即 是 握 線 器 未 能 與 地 錨 孔 之 軸 線 保 持 垂<br />

直 , 致 使 鋼 絞 線 在 錨 頭 處 形 成 異 常 的 折 角 ( 圖 8), 夾 片 無 法 有 效 地 與 鋼 絞 線 咬 合 , 造 成 鋼 絞 線 傳 力<br />

不 良 , 甚 至 發 生 錨 頭 脫 落 的 現 象 。 因 台 灣 許 多 地 錨 皆 有 此 問 題 , 為 改 善 其 缺 失 , 應 注 意 擋 土 構 造 物<br />

之 地 錨 承 座 與 地 錨 孔 軸 線 要 保 持 垂 直 , 以 免 發 生 上 述 地 錨 鋼 絞 線 傳 力 不 良 之 問 題 。<br />

2.2.2. 改 善 方 式<br />

因 錨 頭 夾 片 咬 合 失 敗 , 致 使 地 錨 之 鋼 絞 線 內 縮 至 地 錨 孔 內 時 , 此 地 錨 即 可 視 同 報 廢 , 須 以 補 作<br />

新 地 錨 的 方 式 處 理 。 若 錨 頭 夾 片 咬 合 不 良 並 非 起 因 於 施 工 精 準 度 , 而 是 夾 片 與 錨 座 間 的 問 題 時 , 則<br />

可 在 新 作 地 錨 時 , 採 取 圖 9 所 示 之 現 場 噴 漆 作 記 的 方 式 加 以 檢 測 , 以 確 定 鋼 絞 線 與 夾 片 之 咬 合 是 否<br />

正 常 。<br />

圖 8 握 線 器 與 地 錨 孔 之 軸 線 未 保 持 垂 直<br />

圖 9 現 場 噴 漆 作 記 檢 測 鋼 絞 線 與 夾 片 之 咬 合 效 果<br />

2.3. 鋼 絞 線 斷 裂<br />

2.3.1. 發 生 原 因<br />

鋼 絞 線 斷 裂 大 都 發 生 在 地 錨 之 自 由 段 與 錨 頭 、 或 是 自 由 段 與 錨 碇 段 交 界 處 , 其 原 因 不 外 乎 鋼 絞<br />

線 受 力 超 過 其 抗 拉 強 度 ; 或 是 鋼 絞 線 因 銹 蝕 減 少 斷 面 , 降 低 其 抗 拉 能 力 , 進 而 發 生 斷 裂 。 其 中 後 者<br />

已 於 2.1 節 中 說 明 , 前 者 將 於 本 節 中 說 明 。 鋼 絞 線 因 受 力 超 過 其 抗 拉 強 度 而 斷 裂 時 , 常 伴 隨 著 鋼 絞 線<br />

射 出 的 現 象 ( 圖 10)。 至 於 造 成 鋼 絞 線 受 力 超 過 其 抗 拉 強 度 的 原 因 , 除 鋼 絞 線 銹 蝕 外 , 多 半 是 因 地 錨<br />

之 設 計 量 不 足 所 致 , 當 邊 坡 產 生 下 滑 位 移 時 , 即 將 地 錨 之 鋼 絞 線 扯 斷 ; 另 外 , 擋 土 牆 背 後 地 下 水 位<br />

上 升 時 , 也 會 造 成 牆 後 土 水 推 力 的 增 加 , 而 使 得 地 錨 受 力 超 出 鋼 絞 線 之 拉 力 強 度 而 斷 裂 。<br />

圖 10 鋼 絞 線 斷 裂 射 出 情 形<br />

-201-


2.3.2. 改 善 方 式<br />

當 發 生 地 錨 鋼 絞 線 斷 裂 時 , 即 顯 示 該 地 錨 系 統 不 足 以 抵 抗 擋 土 牆 背 後 之 土 水 壓 , 此 時 可 增 設 新<br />

地 錨 、 或 是 加 強 牆 後 之 地 下 水 排 水 措 施 以 改 善 此 問 題 , 例 如 : 打 設 橫 向 排 水 孔 , 降 低 牆 後 和 邊 坡 內<br />

部 之 水 壓 力 , 或 是 兩 種 方 式 併 用 。<br />

三 、 地 錨 邊 坡 破 壞 原 因<br />

雖 然 大 多 數 之 地 錨 邊 坡 即 使 在 大 雨 和 地 震 狀 況 下 , 皆 能 維 持 穩 定 狀 態 , 但 仍 有 少 部 分 地 錨 邊 坡<br />

在 雨 後 發 生 坍 滑 , 本 節 將 藉 由 台 灣 之 地 錨 邊 坡 破 壞 案 例 和 收 集 的 破 壞 態 樣 , 歸 納 出 造 成 地 錨 邊 坡 破<br />

壞 的 原 因 如 下 :<br />

3.1. 地 質 問 題<br />

台 灣 因 地 質 環 境 關 係 , 大 多 數 邊 坡 經 常 覆 蓋 結 構 鬆 散 的 土 層 或 風 化 的 岩 層 , 但 這 類 地 質 條 件 在<br />

地 錨 系 統 的 設 計 過 程 中 , 通 常 只 被 視 為 需 以 地 錨 穩 定 之 可 能 滑 動 塊 體 , 其 內 部 材 料 性 質 ( 尤 其 是 材<br />

料 的 劣 化 問 題 ), 並 未 受 到 特 別 的 關 注 , 因 此 也 未 採 取 特 別 的 處 理 措 施 , 例 如 : 擋 土 結 構 的 基 礎 座<br />

落 於 鬆 弱 的 覆 蓋 層 上 , 或 採 用 擋 土 功 能 較 差 的 格 梁 構 造 物 , 而 非 版 狀 構 造 物 , 作 為 擋 土 措 施 等 。 在<br />

此 狀 況 下 , 當 施 加 地 錨 拉 力 後 , 向 下 之 分 力 便 極 易 使 地 錨 擋 土 牆 產 生 基 礎 承 載 力 破 壞 ; 或 是 風 化 表<br />

土 日 積 月 累 後 會 經 由 格 框 滑 出 , 淘 空 格 梁 下 方 之 土 層 等 現 象 , 進 而 導 致 整 個 地 錨 背 拉 系 統 的 破 壞 。<br />

此 外 , 造 成 邊 坡 表 面 淺 層 滑 動 的 可 能 性 , 也 會 隨 著 地 層 之 風 化 程 度 和 雨 水 的 滲 透 而 增 加 , 尤 其<br />

在 頁 岩 和 砂 岩 地 層 最 為 明 顯 。 當 擋 土 結 構 背 後 或 上 方 之 地 層 發 生 淺 層 滑 動 時 ( 圖 11), 地 錨 之 自 由 段<br />

鋼 絞 線 易 受 下 滑 土 體 和 擋 土 結 構 的 帶 動 影 響 , 而 受 剪 變 形 , 使 地 錨 失 去 拉 力 , 而 喪 失 其 穩 定 邊 坡 的<br />

功 能 。 造 成 此 一 問 題 之 主 因 , 是 因 擋 土 牆 之 背 拉 地 錨 屬 單 力 構 件 , 雖 可 對 邊 坡 提 供 正 向 壓 力 , 增 加<br />

邊 坡 抵 抗 滑 動 的 能 力 , 但 地 錨 本 身 並 無 法 如 剪 力 鎨 般 地 提 供 抵 抗 剪 力 的 能 力 , 因 此 地 錨 若 受 到 如 圖<br />

11 之 淺 層 滑 坡 影 響 時 , 將 因 無 法 承 受 剪 力 而 破 壞 , 並 連 帶 地 引 發 大 規 模 的 邊 坡 滑 動 。<br />

至 於 地 錨 設 計 過 程 中 , 最 受 注 意 的 錨 碇 段 地 質 情 況 , 其 實 並 非 最 常 出 現 問 題 的 地 方 , 因 地 錨 之<br />

錨 碇 效 果 是 否 確 實 , 由 現 場 地 錨 施 加 拉 力 的 過 程 中 便 可 瞭 解 , 所 以 地 錨 邊 坡 的 破 壞 絕 大 多 數 都 不 是<br />

因 錨 碇 段 的 問 題 而 發 生 , 反 而 是 發 生 在 錨 頭 和 自 由 段 位 置 。<br />

3.2. 設 計 問 題<br />

圖 11 地 錨 擋 土 結 構 上 方 或 後 方 之 淺 層 滑 動 示 意 圖<br />

地 錨 邊 坡 常 因 基 地 調 查 不 足 , 以 致 無 法 充 分 了 解 邊 坡 內 部 之 可 能 滑 動 位 置 , 因 此 部 分 邊 坡 地 錨<br />

-202-


會 因 設 計 長 度 不 足 , 而 使 得 地 錨 隨 著 滑 動 塊 體 掉 落 , 這 種 情 形 較 可 能 發 生 於 擴 座 型 地 錨 , 因 擴 座 地<br />

錨 之 地 錨 總 長 度 和 錨 碇 段 長 度 較 短 。 另 一 常 見 的 設 計 問 題 則 是 未 能 適 當 地 考 量 擋 土 結 構 後 方 之 地 下<br />

水 狀 況 , 通 常 設 計 時 假 設 邊 坡 內 部 之 地 下 水 , 可 藉 由 打 入 邊 坡 之 橫 向 排 水 孔 或 擋 土 結 構 上 之 排 水 孔<br />

順 利 排 出 。 但 事 實 上 , 邊 坡 內 部 的 地 下 水 不 一 定 能 藉 由 這 些 排 水 設 施 作 有 效 的 排 放 。 因 此 , 當 進 行<br />

邊 坡 擋 土 結 構 設 計 時 , 可 能 會 低 估 邊 坡 和 擋 土 結 構 後 方 之 地 下 水 壓 力 , 而 造 成 地 錨 設 計 數 量 的 不<br />

足 。 但 在 此 也 必 須 指 出 , 假 設 不 合 理 的 高 地 下 水 壓 力 也 是 沒 必 要 , 因 過 於 保 守 的 擋 土 結 構 設 計 , 將<br />

會 造 成 資 源 的 浪 費 。<br />

3.3. 施 工 問 題<br />

[2]<br />

預 力 地 錨 的 主 要 優 點 是 , 所 有 地 錨 之 抗 拔 能 力 均 可 於 施 工 期 間 , 藉 由 地 錨 施 拉 的 標 準 程 序 加<br />

以 驗 證 。 換 言 之 , 所 有 地 錨 的 抗 拔 能 力 皆 可 藉 由 地 錨 施 拉 的 過 程 加 以 驗 證 , 因 此 要 確 保 現 場 工 作 地<br />

錨 之 錨 碇 能 力 , 並 不 困 難 。 而 且 在 大 多 數 的 情 況 下 , 地 錨 之 錨 碇 能 力 在 施 工 階 段 獲 得 確 定 後 , 往 後<br />

的 錨 碇 能 力 將 不 會 有 太 大 的 變 化 。 可 是 在 錨 頭 和 自 由 段 的 防 蝕 保 護 部 份 就 不 一 樣 了 , 尤 其 是 台 灣 的<br />

天 候 、 地 質 和 地 下 水 狀 況 , 對 地 錨 之 錨 頭 和 自 由 段 的 防 蝕 保 護 都 相 當 不 利 , 因 此 地 錨 構 件 之 銹 蝕 便<br />

成 為 台 灣 地 錨 的 最 大 殺 手 。 然 而 過 去 因 較 缺 乏 正 確 的 防 蝕 保 護 觀 念 , 非 常 多 的 地 錨 都 只 採 用 單 層 防<br />

蝕 保 護 , 而 非 雙 重 防 蝕 保 護 。 隨 著 使 用 年 限 的 增 加 , 這 些 地 錨 之 錨 頭 和 自 由 段 鋼 絞 線 的 銹 蝕 問 題 將<br />

會 日 趨 嚴 重 , 若 未 採 取 適 當 的 補 強 和 改 善 措 施 , 那 麼 因 地 錨 銹 蝕 而 破 壞 之 地 錨 邊 坡 , 將 會 陸 續 發 生 。<br />

四 、 地 錨 邊 坡 破 壞 案 例<br />

針 對 上 述 地 錨 邊 坡 破 壞 的 原 因 , 本 文 將 以 台 灣 北 部 三 個 地 錨 邊 坡 破 壞 案 例 進 行 說 明 。 這 三 個 案<br />

例 的 地 質 情 況 , 主 要 為 砂 / 頁 岩 互 層 , 地 質 構 造 為 順 向 坡 山 體 , 三 者 在 整 地 過 程 中 都 有 削 坡 動 作 ,<br />

且 都 採 用 背 拉 地 錨 來 增 加 其 穩 定 性 。<br />

4.1. 林 肯 大 郡 地 錨 邊 坡 滑 動<br />

1997 年 8 月 18 日 早 上 8 點 30 分 溫 妮 颱 風 侵 襲 台 灣 , 林 肯 大 郡 發 生 了 悲 慘 的 邊 坡 滑 動 事 件 ( 圖<br />

12), 以 擋 土 牆 和 背 拉 地 錨 保 護 的 邊 坡 突 然 向 下 滑 動 , 滑 動 範 圍 約 140 m 長 ,60 m 寬 。 當 時 覆 蓋 於<br />

頁 岩 上 之 砂 岩 層 以 高 速 下 滑 , 衝 向 緊 鄰 邊 坡 下 方 的 5 層 樓 房 屋 , 造 成 居 民 多 人 死 傷 之 慘 劇 。 其 滑 動<br />

面 分 為 兩 部 分 , 一 為 上 層 滑 動 塊 體 , 厚 度 約 2.4~4 m; 另 一 為 下 層 滑 動 塊 體 , 厚 度 約 6~8 m, 如 圖<br />

13 所 示 。 此 一 事 件 之 主 要 發 生 原 因 , 在 於 地 下 水 滲 入 邊 坡 造 成 邊 坡 地 質 材 料 之 軟 化 和 劣 化 , 使 得 原<br />

先 打 設 在 邊 坡 上 的 背 拉 地 錨 , 無 法 有 效 地 抑 制 順 向 坡 岩 層 的 向 下 滑 動 所 致 。<br />

圖 12 林 肯 大 郡 坡 地 滑 動 案 例 示 意 圖 ( 摘 自 [3] )<br />

-203-


30 o 40 o 20 o<br />

29.3 o<br />

A<br />

EL. 56.76<br />

EL. 53.76<br />

(a) 地 質 剖 面 圖<br />

(b) 背 拉 地 錨 剖 面 圖<br />

圖 13 林 肯 大 郡 之 地 質 剖 面 和 背 拉 地 錨 剖 面 圖<br />

事 實 上 , 在 該 地 錨 邊 坡 崩 塌 發 生 前 , 即 已 有 許 多 跡 象 可 循 , 例 如 : 擋 土 牆 上 已 有 部 分 地 錨 錨 頭<br />

脫 落 , 距 牆 趾 高 約 6~7 公 尺 處 之 牆 面 有 地 下 水 滲 出 等 , 這 些 跡 象 均 顯 示 該 邊 坡 已 在 滑 動 , 且 地 下 水<br />

壓 力 亦 在 牆 後 累 積 增 加 中 , 邊 坡 之 穩 定 性 在 溫 妮 颱 風 侵 襲 前 即 已 接 近 臨 界 狀 況 。 該 基 地 之 地 質 主 要<br />

由 砂 岩 和 1~10 公 分 厚 的 薄 層 頁 岩 互 層 沿 28~30 度 坡 面 所 組 成 , 其 中 , 砂 岩 之 無 圍 壓 縮 強 度 約 為<br />

30~190 kg/cm 2 且 表 面 存 在 明 顯 的 垂 直 裂 縫 。 由 於 砂 岩 之 垂 直 裂 縫 明 顯 , 雨 水 可 輕 易 地 入 滲 至 邊 坡 內<br />

部 , 累 積 一 定 時 間 後 , 地 下 水 壓 力 上 升 , 砂 岩 下 方 之 頁 岩 浸 水 軟 化 , 剪 力 阻 抗 降 低 , 最 後 導 致 邊 坡<br />

下 滑 。 由 試 驗 室 之 直 接 剪 力 試 驗 結 果 得 知 , 該 處 頁 岩 浸 水 80 小 時 後 , 其 殘 餘 摩 擦 角 降 低 約 50%,<br />

但 邊 坡 擋 土 結 構 在 設 計 過 程 中 並 未 考 慮 地 下 水 壓 力 和 頁 岩 浸 水 軟 化 的 影 響 , 而 高 估 邊 坡 材 料 強 度 ,<br />

設 計 為 數 不 足 的 地 錨 。 由 此 案 例 可 見 , 地 下 水 對 邊 坡 穩 定 性 有 絕 對 性 的 影 響 , 即 使 是 32 小 時 內 累<br />

積 區 區 150 公 釐 的 雨 量 , 也 會 造 成 如 此 大 規 模 的 邊 坡 崩 塌 事 件 , 令 人 非 常 驚 訝 。<br />

除 了 在 邊 坡 穩 定 設 計 階 段 , 並 未 考 慮 滑 動 面 之 水 壓 力 外 , 地 錨 拉 拔 能 力 之 喪 失 , 也 是 造 成 邊 坡<br />

崩 塌 主 要 原 因 之 一 。 經 現 地 揚 起 試 驗 得 知 , 邊 坡 下 滑 後 , 遺 留 在 邊 坡 上 的 地 錨 , 其 極 限 拉 拔 力 較 設<br />

計 拉 力 (40 噸 ) 高 出 約 35~40 噸 左 右 , 因 此 在 現 地 並 未 發 現 有 地 錨 錨 碇 段 被 拉 出 之 跡 象 。 但 在 現 場 卻<br />

觀 察 到 有 鋼 絞 線 射 出 錨 頭 混 凝 土 護 蓋 , 以 及 錨 頭 掉 落 的 現 象 。 此 外 , 由 現 場 出 露 的 錨 頭 ( 含 夾 片 和<br />

鋼 絞 線 ) 外 觀 可 看 出 , 錨 頭 銹 蝕 情 形 相 當 嚴 重 , 經 測 試 銹 蝕 夾 片 與 鋼 絞 線 之 咬 合 能 力 後 得 知 , 其 咬<br />

合 能 力 僅 剩 原 地 錨 設 計 拉 力 的 1/10 左 右 , 因 此 當 擋 土 牆 因 牆 後 之 推 力 增 加 而 外 移 時 , 錨 頭 便 因 銹<br />

蝕 , 使 得 夾 片 與 鋼 絞 線 間 無 法 作 有 效 地 咬 合 而 滑 脫 掉 落 。<br />

另 外 , 經 比 對 地 錨 數 量 後 亦 發 現 , 實 際 打 設 之 地 錨 總 數 (=497) 足 足 比 原 設 計 數 量 短 少 103 支 ,<br />

所 以 在 地 錨 數 目 不 足 、 錨 頭 銹 蝕 、 和 未 考 慮 地 下 水 壓 等 因 素 之 綜 合 影 響 下 , 造 成 了 林 肯 大 郡 地 錨 邊<br />

坡 的 滑 動 破 壞 。<br />

4.2. 台 灣 科 技 大 學 基 隆 校 區 邊 坡<br />

台 灣 科 技 大 學 基 隆 校 區 之 邊 坡 曾 分 別 發 生 兩 次 崩 塌 ( 圖 14): 第 一 次 發 生 在 山 坡 整 地 期 間 (1995<br />

年 ); 第 二 次 發 生 在 完 工 12 年 後 (2007 年 )。 這 兩 次 事 件 皆 造 成 邊 坡 格 梁 式 擋 土 牆 和 背 拉 地 錨 的 嚴 重<br />

破 壞 。1995 年 事 件 發 生 在 連 續 2 天 的 豪 雨 後 , 格 梁 擋 土 結 構 先 發 生 滑 動 破 壞 , 進 而 引 致 整 個 山 坡 的<br />

向 下 滑 動 , 其 滑 動 範 圍 約 150 公 尺 長 、80 公 尺 寬 、40 公 尺 深 ; 2007 年 事 件 也 是 發 生 在 連 續 豪 雨 之<br />

後 , 其 破 壞 規 模 雖 較 前 例 為 小 , 但 兩 者 之 破 壞 模 式 卻 很 相 似 , 都 是 起 因 於 邊 坡 表 層 鬆 散 的 覆 土 層 之<br />

淺 層 滑 動 , 進 而 造 成 整 個 背 拉 地 錨 系 統 破 壞 的 案 例 。<br />

該 破 壞 邊 坡 之 地 質 情 況 可 區 分 為 兩 部 分 , 上 層 為 覆 土 層 , 厚 度 約 4~8 公 尺 ; 下 層 為 較 厚 的 砂 、<br />

頁 岩 互 層 , 其 中 砂 岩 風 化 嚴 重 , 其 岩 石 品 質 指 標 (RQD) 約 在 0~30 之 間 [4] 。 同 時 , 邊 坡 上 亦 可 觀 察 到<br />

垂 直 裂 縫 之 存 在 與 分 佈 , 地 下 水 相 當 豐 沛 , 但 在 鋼 筋 混 凝 土 格 梁 間 , 並 無 充 足 的 排 水 設 施 和 排 水 孔 ,<br />

可 供 邊 坡 內 部 之 過 量 地 下 水 排 出 , 因 此 該 邊 坡 之 地 下 水 幾 乎 以 漫 流 的 方 式 由 邊 坡 流 出 。<br />

由 埋 置 於 邊 坡 內 之 傾 斜 儀 讀 數 可 知 ,1995 年 崩 塌 案 例 之 滑 動 面 深 度 約 位 於 坡 面 下 方 11.5~22.5<br />

公 尺 , 滑 動 坡 角 約 為 16 度 , 如 圖 15 所 示 。 此 外 , 依 滑 動 土 塊 之 幾 何 尺 寸 ( 未 打 設 地 錨 ) 和 估 計 滑 動<br />

-204-


面 之 反 推 分 析 可 知 , 當 該 滑 動 面 之 凝 聚 力 等 於 零 , 摩 擦 角 約 為 16~18 度 時 , 邊 坡 產 生 滑 動 。 反 推 分<br />

析 所 得 之 參 數 , 相 當 接 近 室 內 直 接 剪 力 試 驗 結 果 : 飽 和 頁 岩 土 樣 之 凝 聚 力 為 1.8 t/m 2 、 摩 擦 角 為 16.5<br />

度 。 然 而 , 經 打 設 地 錨 系 統 後 ( 圖 15), 抵 抗 滑 動 之 安 全 係 數 將 高 於 1.25。 但 在 此 高 安 全 係 數 下 , 仍<br />

然 發 生 邊 坡 滑 動 破 壞 , 顯 示 未 打 設 地 錨 之 邊 坡 滑 動 面 與 打 設 地 錨 後 邊 坡 崩 塌 之 滑 動 面 並 不 一 致 , 後<br />

者 較 可 能 是 因 擋 土 牆 後 方 之 淺 層 滑 動 而 被 驅 動 的 後 續 破 壞 。<br />

圖 14 台 灣 科 技 大 學 基 隆 校 區 坡 地 崩 塌<br />

圖 15 台 灣 科 技 大 學 基 隆 校 區 坡 地 之 地 錨 邊 坡 和 傾 斜 管 剖 面<br />

當 進 一 步 觀 察 用 來 穩 定 邊 坡 之 地 錨 時 ( 圖 15), 可 看 出 擋 土 牆 之 水 平 夾 角 約 75 度 , 地 錨 之 打 設<br />

傾 角 高 達 50 度 。 在 如 此 大 的 傾 角 情 況 下 , 大 部 份 之 地 錨 分 力 , 會 轉 變 成 對 擋 土 牆 的 下 拉 作 用 力 。<br />

因 此 , 當 邊 坡 因 雨 水 滲 透 而 軟 化 後 , 擋 土 牆 之 基 礎 可 能 會 發 生 承 載 力 破 壞 , 而 且 若 採 型 框 擋 土 牆 時 ,<br />

土 壤 還 有 可 能 從 型 框 中 間 漏 出 , 並 引 發 邊 坡 之 淺 層 滑 動 , 進 而 導 致 整 體 擋 土 系 統 向 下 滑 落 破 壞 。 當<br />

失 去 背 拉 地 錨 拉 力 後 , 滑 動 面 亦 將 向 下 發 展 至 更 深 層 的 位 置 。 然 而 , 較 可 惜 的 是 此 邊 坡 並 未 設 置 監<br />

測 系 統 , 因 此 沒 有 邊 坡 位 移 情 形 之 相 關 紀 錄 。<br />

此 外 , 由 2007 年 邊 坡 滑 動 案 例 之 地 錨 格 梁 破 壞 狀 況 , 發 現 邊 坡 滑 動 掉 落 的 過 程 , 位 於 最 上 層<br />

之 地 錨 有 整 支 被 拔 出 的 情 形 ( 圖 16), 顯 示 地 錨 之 錨 碇 力 失 效 , 該 地 錨 可 能 是 因 長 度 不 足 , 以 致 無 法<br />

錨 碇 於 深 層 之 堅 硬 岩 層 上 , 因 此 當 邊 坡 表 層 發 生 淺 層 滑 動 時 , 地 錨 即 無 法 提 供 抗 滑 所 需 的 錨 碇 力 ,<br />

反 而 被 下 滑 的 土 石 拉 出 。 類 似 的 情 形 可 能 也 發 生 在 1995 年 之 滑 坡 案 例 中 。<br />

-205-


地 錨 被 拔 出<br />

地 錨 被 拔 出<br />

圖 16 最 上 層 之 地 錨 在 滑 坡 過 程 中 整 支 被 拔 出<br />

4.3. 公 路 地 錨 邊 坡<br />

2008 年 7 月 28 日 鳳 凰 颱 風 侵 台 , 導 致 高 速 公 路 發 生 地 錨 邊 坡 破 壞 案 例 ( 圖 17)。 該 邊 坡 下 方 之<br />

地 層 主 要 為 頁 岩 層 , 岩 塊 沿 順 向 坡 面 向 下 滑 動 , 災 變 範 圍 約 70 公 尺 寬 、30 公 尺 高 和 1.5 公 尺 厚 ,<br />

屬 淺 層 滑 動 。 該 處 地 錨 邊 坡 之 建 造 年 限 已 有 30 年 以 上 , 坡 面 之 角 度 約 45 度 , 滑 動 面 之 傾 角 約 為 50<br />

度 ~60 度 。 邊 坡 上 設 有 兩 層 地 錨 , 上 層 有 3 排 背 拉 地 錨 ( 間 距 = 寬 2 公 尺 × 高 3 公 尺 ), 每 支 地 錨 有 7<br />

條 直 徑 12.7 mm 之 鋼 絞 線 。 下 層 有 2 排 背 拉 地 錨 ( 間 距 = 寬 3 公 尺 × 高 3 公 尺 ), 每 支 地 錨 有 6~7 條 直<br />

徑 12.7 mm 之 鋼 絞 線 。 上 層 地 錨 之 設 計 拉 力 為 60 公 噸 , 打 設 傾 角 ( 水 平 之 夾 角 ) 約 30 度 ; 下 層 地 錨 之<br />

設 計 拉 力 為 60 公 噸 , 打 設 傾 角 ( 水 平 之 夾 角 ) 約 30 度 。<br />

地 錨 構 件 銹 蝕<br />

圖 17 公 路 旁 之 地 錨 邊 坡 破 壞<br />

鋼 絞 線 銹 蝕 整 束 被 拔 出<br />

地 錨 格 梁 邊 坡 上 層 , 噴 凝 土 護 坡 已 嚴 重 風 化 、 表 面 開 裂 、 雜 草 叢 生 , 地 表 水 可 輕 易 地 經 由 這 些<br />

裂 縫 和 風 化 的 頁 岩 入 滲 到 邊 坡 內 部 。 長 期 下 來 , 表 層 頁 岩 逐 漸 風 化 , 喪 失 大 部 分 的 強 度 。 依 現 場 岩<br />

石 試 體 之 弱 面 直 接 剪 力 試 驗 結 果 顯 示 , 頁 岩 之 殘 餘 強 度 參 數 約 為 C r = 2.3 t/m 2 , φ r =25 度 。 此 外 , 鳳<br />

凰 颱 風 挾 帶 之 雨 量 並 不 算 豐 沛 , 累 計 雨 量 只 有 220 公 釐 左 右 , 在 附 近 地 區 並 沒 有 出 現 其 他 坡 地 崩 塌<br />

災 情 , 因 此 , 此 邊 坡 之 破 壞 原 因 , 除 了 降 雨 外 , 可 能 另 有 其 因 。<br />

事 實 上 , 由 邊 坡 崩 塌 現 場 可 觀 察 到 一 些 與 地 錨 相 關 之 重 要 訊 息 , 例 如 : 鋼 絞 線 和 錨 頭 之 夾 片 已<br />

經 嚴 重 銹 蝕 ( 圖 17), 部 分 地 錨 之 鋼 絞 線 也 有 因 銹 蝕 而 斷 裂 , 或 整 束 被 拉 斷 的 情 形 ( 圖 17), 顯 然 地 錨<br />

所 能 提 供 之 背 拉 效 果 , 已 因 銹 蝕 而 大 為 減 少 。 此 外 , 早 在 該 邊 坡 滑 動 前 (2003), 上 下 排 地 錨 間 之 坡<br />

面 噴 凝 土 即 顯 示 出 異 狀 , 但 未 作 積 極 處 理 , 可 能 也 是 地 錨 邊 坡 破 壞 之 原 因 。<br />

此 外 , 整 座 地 錨 格 梁 擋 土 牆 在 滑 動 過 程 中 , 係 以 溜 滑 梯 的 方 式 下 滑 , 並 產 生 90 度 的 旋 轉 ( 如 圖<br />

18)。 顯 然 地 , 原 先 打 設 在 該 格 梁 擋 土 牆 上 的 24 支 地 錨 , 可 能 早 就 因 銹 蝕 而 失 去 提 供 錨 碇 力 的 功 用 ,<br />

-206-


因 此 連 擋 土 牆 本 身 之 自 重 也 無 法 支 撐 。 同 時 , 在 失 去 地 錨 錨 碇 力 的 情 況 下 , 也 使 得 該 邊 坡 之 頁 岩 地<br />

層 , 在 颱 風 帶 來 的 大 雨 驅 動 下 , 發 生 順 向 坡 滑 動 。<br />

滑 動 方 向<br />

格 梁 旋 轉 下 滑<br />

圖 18 格 梁 擋 土 牆 旋 轉 下 滑 情 形<br />

五 、 結 論<br />

一 般 而 言 , 本 文 說 明 之 地 錨 邊 坡 破 壞 案 例 常 伴 隨 著 多 種 因 素 發 生 , 並 非 單 一 因 素 所 造 成 , 根 據<br />

這 些 地 錨 邊 坡 破 壞 案 例 的 彙 整 研 究 , 可 得 結 論 如 下 :<br />

1. 本 文 所 舉 之 三 個 地 錨 邊 坡 破 壞 案 例 中 , 有 兩 個 與 地 錨 發 生 銹 蝕 有 關 ( 林 肯 大 郡 和 公 路 地 錨 邊<br />

坡 ), 顯 示 地 錨 之 防 蝕 保 護 與 地 錨 邊 坡 穩 定 性 息 息 相 關 。 尤 其 是 台 灣 地 區 多 雨 且 邊 坡 表 層 地 質 破<br />

碎 , 雨 水 入 滲 邊 坡 情 況 嚴 重 , 常 造 成 邊 坡 表 層 材 料 性 質 之 老 劣 化 和 地 錨 銹 蝕 等 問 題 , 因 此 須 特<br />

別 注 意 地 錨 之 防 蝕 保 護 措 施 。<br />

2. 台 灣 之 邊 坡 表 面 經 常 覆 蓋 著 鬆 軟 地 層 , 若 需 將 護 坡 結 構 直 接 設 置 在 此 類 地 層 時 , 應 留 意 覆 蓋 層<br />

承 載 力 不 佳 , 可 能 會 造 成 的 淺 層 邊 坡 滑 動 。 就 風 化 情 況 嚴 重 之 邊 坡 而 言 , 宜 採 用 擋 土 能 力 較 佳<br />

之 版 式 擋 土 牆 , 避 免 使 用 擋 土 能 力 較 差 之 格 式 擋 土 牆 或 使 用 傾 角 過 大 之 地 錨 , 以 免 因 牆 後 土 壤<br />

流 失 或 過 大 之 向 下 分 力 , 造 成 邊 坡 之 淺 層 滑 動 破 壞 。<br />

3. 台 灣 使 用 之 地 錨 , 大 多 數 已 有 相 當 久 的 歷 史 , 其 現 況 和 功 能 均 須 作 有 系 統 的 調 查 , 以 維 持 地 錨<br />

邊 坡 之 安 全 。 一 般 而 言 , 以 目 視 方 式 配 合 榔 頭 等 簡 易 工 具 進 行 檢 查 , 是 既 經 濟 又 有 效 之 地 錨 和<br />

地 錨 邊 坡 檢 測 方 式 , 若 地 錨 之 缺 陷 無 法 以 目 視 檢 測 時 , 再 輔 以 其 他 儀 器 設 備 , 如 電 阻 量 測 、 內<br />

視 鏡 攝 影 或 其 他 先 進 儀 器 設 備 進 行 詳 細 檢 查 。<br />

参 考 文 献<br />

[1] Weerasinghe, R. B. and Anson, R. W. W. Investigation of the Long Term Performance and Future Behaviour of Existing<br />

Ground Anchorages. Proceedings of Ground Anchorages and Anchored Structures, London, 1997, 353-362.<br />

[2] F.I.P. Recommendations for the Design and Construction of Prestressed Concrete Ground Anchors. Federation<br />

Internation De La Prectntrainte, 1982.<br />

[3] 陳 堯 中 廖 洪 鈞 林 宏 達 陳 志 南 . 汐 止 林 肯 大 郡 災 變 原 因 探 討 . 地 工 技 術 雜 誌 . 1998, 68: 29-40.<br />

[4] Liao, H. J., Chen, C. N. and Liao, J. T. Landslides of Cut and Tieback Slopes in Northern Taiwan. Proceedings of 8 th<br />

International Symposium on Landslides, Cardiff, 2000, 923-928.<br />

-207-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

砂 土 液 化 剪 切 波 速 判 别 法 研 究 进 展<br />

1, 2<br />

1, 2<br />

陈 云 敏 周 燕 国<br />

( 1 浙 江 大 学 软 弱 土 与 环 境 土 工 教 育 部 重 点 实 验 室 , 2 浙 江 大 学 岩 土 工 程 研 究 所 , 浙 江 杭 州 310058)<br />

摘 要 : 砂 土 地 震 液 化 判 别 是 岩 土 地 震 工 程 领 域 的 重 要 课 题 。 剪 切 波 速 法 是 砂 土 液 化 判 别 的 主 要<br />

方 法 之 一 , 其 核 心 在 于 建 立 砂 土 抗 液 化 强 度 与 剪 切 波 速 的 相 关 关 系 。 本 文 对 近 30 年 来 各 种 砂 土 液<br />

化 剪 切 波 速 判 别 方 法 作 了 回 顾 , 针 对 循 环 剪 应 变 方 法 和 循 环 剪 应 力 方 法 体 系 , 分 析 了 液 化 标 准 和 土<br />

体 结 构 性 对 建 立 各 种 相 关 性 经 验 公 式 的 重 要 影 响 。 液 化 标 准 不 同 , 使 得 具 有 相 同 剪 切 波 速 试 样 的 抗<br />

液 化 强 度 显 著 不 同 ; 而 不 同 类 型 砂 土 的 结 构 性 差 异 显 著 , 其 抗 液 化 强 度 随 剪 切 波 速 的 变 化 规 律 也 不<br />

同 , 因 此 现 有 各 种 剪 切 波 速 方 法 在 实 际 工 程 应 用 中 的 适 用 性 和 可 靠 性 存 在 差 异 。 建 立 经 济 、 可 靠 的<br />

剪 切 波 速 判 别 法 有 待 于 提 出 与 土 类 相 关 的 CRR-V s 相 关 关 系 。<br />

关 键 词 : 砂 土 液 化 剪 切 波 速 抗 液 化 强 度 液 化 标 准 结 构 性 地 震<br />

RESEARCH PROGRESS OF LIQUEFACTION EVALUATION OF SANDY SOILS<br />

BY SHEAR WAVE VELOCITY<br />

Yun-Min CHEN 1, 2 and Yan-Guo ZHOU 1, 2<br />

1 MOE Key Laboratory of Soft Soils and Geoenvironmental Engineering, Zhejiang <strong>University</strong>; 2 Institute of<br />

Geotechnical Engineering, Zhejiang <strong>University</strong>, Hangzhou 310058, P. R. China<br />

Abstract: Evaluation of the liquefaction potential of sandy soils is important content in many geotechnical<br />

earthquake engineering problems. Shear wave velocity-based evaluation is one of the main methods, and the key<br />

issue is to establish the correlation between liquefaction resistance ratio (CRR) and shear wave velocity (V s ) of<br />

sands. In this paper, the progress of V s -based evaluation methods in the last three decades were reviewed based<br />

on either cyclic stress or strain approaches, and the influences of liquefaction criteria and soil structure on the<br />

derivation of CRR-V s correlations are highlighted. Different liquefaction criteria leads to different cyclic<br />

strength evaluation for a given sand with the same shear wave velocity, while different types of soils have<br />

different CRR-V s correlations due to the variation of soil structures. <strong>The</strong>refore the currently existing V s -based<br />

methods differ from each other with respect to the suitability and reliability in engineering practices. <strong>The</strong><br />

economic and reliable Vs-based evaluation method requires soil type specific CRR-V s correlations.<br />

Keywords: Sand liquefaction, shear wave velocity, liquefaction resistance, liquefaction criterion, soil structure,<br />

earthquake.<br />

一 、 前 言<br />

土 体 地 震 液 化 是 最 常 见 也 是 最 具 破 坏 性 的 场 地 震 害 , 场 地 液 化 判 别 一 直 是 岩 土 地 震 工 程 研 究 和<br />

实 践 的 重 要 课 题 。 根 据 Seed-Idriss 简 化 方 法 [1] , 地 震 液 化 判 别 需 要 确 定 两 个 主 要 参 数 : 一 是 地 震 在<br />

场 地 引 起 的 循 环 剪 应 力 水 平 , 即 CSR (Cyclic Stress Ratio); 二 是 场 地 土 的 抗 液 化 强 度 , 即 CRR (Cyclic<br />

Resistance Ratio)。 当 CSR 大 于 CRR 时 , 土 体 将 发 生 液 化 。 由 于 地 震 剪 应 力 CSR 可 由 如 下 Seed 简<br />

化 公 式 确 定 :<br />

基 金 项 目 : 国 家 自 然 科 学 基 金 资 助 项 目 (No. 50908207);973 计 划 资 助 项 目 (No. 2007CB714203); 中 央 高 校 基 本 科 研 业 务 费 专 项 资 助<br />

( 项 目 批 准 号 :No.2010XZZX001,No.2010KYJD006), 浙 江 省 重 点 创 新 团 队 支 持 计 划 资 助 ( 项 目 批 准 号 :No.2009R50050)。<br />

作 者 简 介 : 陈 云 敏 (1962-), 男 , 浙 江 温 岭 人 , 教 授 , 博 导 , 主 要 从 事 岩 土 工 程 教 学 、 科 研 和 工 程 咨 询 工 作 ;<br />

-208-<br />

周 燕 国 (1978-), 男 , 浙 江 富 阳 人 , 副 教 授 , 主 要 从 事 岩 土 地 震 工 程 与 防 震 减 灾 方 向 的 教 学 与 科 研 工 作 。


⎛τ<br />

⎞<br />

av<br />

⎛amax ⎞⎛σ<br />

⎞<br />

v0<br />

CSR = ⎜ ⎟ = 0.65⎜<br />

⎟⎜ ⎟rd<br />

⎝σ′ ⎠ ⎝ g ⎠⎝σ′<br />

⎠<br />

v0 e<br />

v0<br />

(1)<br />

其 中 ,τ av = 等 效 水 平 地 震 剪 应 力 ;σ′ v0 = 有 效 上 覆 应 力 ;σ v0 = 为 总 的 上 覆 应 力 ;a max = 为 场 地 水 平 峰<br />

值 加 速 度 ;g = 为 重 力 加 速 度 ;r d = 剪 应 力 折 减 系 数 [2] 。<br />

可 见 , 液 化 判 别 的 关 键 问 题 在 于 抗 液 化 强 度 CRR 的 确 定 [3] 。 室 内 固 结 不 排 水 等 应 力 幅 循 环 三 轴<br />

( 或 扭 剪 、 单 剪 ) 试 验 能 较 准 确 模 拟 地 震 荷 载 在 土 单 元 上 产 生 的 动 应 力 和 液 化 破 坏 过 程 , 是 直 接 确<br />

定 饱 和 砂 土 抗 液 化 强 度 的 有 效 手 段 。 然 而 , 通 过 常 规 取 样 或 重 塑 样 进 行 液 化 试 验 确 定 CRR, 由 于 存<br />

在 试 样 扰 动 或 制 样 方 法 差 异 , 破 坏 了 原 位 土 体 结 构 性 , 无 法 准 确 评 价 原 位 土 体 的 抗 液 化 强 度 [4] 。 现<br />

场 以 SPT、CPT 等 贯 入 指 标 确 定 土 体 抗 液 化 强 度 的 经 验 方 法 , 虽 然 能 直 接 反 应 土 体 大 应 变 行 为 , 但<br />

物 理 意 义 不 明 确 , 无 法 与 场 地 地 震 液 化 过 程 关 联 , 且 存 在 场 地 适 用 性 较 差 ( 如 砾 石 场 地 无 法 测 量 )、<br />

不 能 连 续 测 试 等 缺 点 。 而 土 体 剪 切 波 速 (V s ) 物 理 意 义 明 确 , 综 合 反 映 了 砂 土 结 构 性 和 应 力 状 态 ,<br />

可 在 绝 大 多 数 场 地 开 展 波 速 测 试 , 加 上 可 在 室 内 单 元 体 尺 度 和 现 场 场 地 尺 度 进 行 无 损 测 量 , 将 室 内<br />

单 元 体 液 化 试 验 与 现 场 原 位 测 试 优 势 有 机 结 合 , 推 动 了 利 用 土 体 剪 切 波 速 评 价 砂 土 抗 液 化 强 度 这 一<br />

重 要 方 法 的 产 生 和 发 展 。 与 基 于 SPT、CPT 指 标 的 Seed-Idriss 简 化 方 法 类 似 , 这 种 方 法 的 主 要 任 务<br />

在 于 建 立 能 可 靠 反 映 砂 土 剪 切 波 速 与 抗 液 化 强 度 之 间 的 CRR-V s 定 量 相 关 关 系 。<br />

Stokoe et al. (1988) [5] 较 早 开 展 基 于 现 场 震 害 调 查 的 CRR-V s 相 关 关 系 研 究 。 该 方 法 综 合 地 震 场<br />

地 液 化 调 查 与 现 场 剪 切 波 速 测 试 , 将 地 震 场 地 的 剪 切 波 速 ( 经 上 覆 应 力 修 正 ) 与 地 震 循 环 应 力 比 ( 经 震<br />

级 修 正 ) 相 关 联 , 通 过 液 化 和 非 液 化 场 地 的 (CRR, V s ) 数 据 拟 合 获 得 CRR-V s 经 验 相 关 关 系 ( 即 液 化 -<br />

非 液 化 分 界 线 ) [6] ( 见 图 1)。 近 年 来 也 有 学 者 将 SPT 或 CPT 贯 入 指 标 与 剪 切 波 速 关 联 , 进 一 步 丰 富<br />

了 基 于 现 场 调 查 的 波 速 方 法 [7] 。 然 而 , 这 种 基 于 现 场 震 害 调 查 的 CRR-V s 相 关 关 系 仍 然 不 够 完 善 ,<br />

主 要 原 因 在 于 它 仅 是 对 地 震 后 状 态 进 行 被 动 评 价 , 不 能 反 映 地 震 液 化 灾 变 过 程 信 息 , 同 时 还 受 到 地<br />

震 液 化 数 据 稀 少 的 限 制 , 尤 其 是 具 有 强 地 震 条 件 (a max >0.4 g 或 CSR>0.3)、 较 深 场 地 土 层 (>10 m)、<br />

高 密 实 度 土 层 (V s >200 m/s) 等 特 征 的 数 据 较 稀 少 。 而 且 , 现 场 波 速 测 试 通 常 不 会 同 时 进 行 取 样 , 缺 少<br />

所 测 土 层 的 详 细 物 性 数 据 。 另 外 , 场 地 调 查 获 得 是 地 震 后 的 特 性 , 通 常 与 地 震 前 土 体 性 质 有 所 差 别 ,<br />

不 能 完 全 反 映 液 化 前 土 体 刚 度 和 强 度 特 性 。<br />

[6]<br />

图 1 基 于 现 场 调 查 的 液 化 判 别 CRR-V s1 曲 线<br />

-209-


鉴 于 此 , 更 多 学 者 将 注 意 力 集 中 到 建 立 将 室 内 试 验 与 现 场 波 速 测 试 相 结 合 的 波 速 判 别 方 法 上<br />

来 , 其 核 心 思 想 在 于 基 于 理 论 推 导 或 室 内 试 验 提 出 砂 土 CRR-V s 相 关 性 , 然 后 结 合 现 场 波 速 测 试 评<br />

价 现 场 土 体 的 抗 液 化 强 度 。 由 于 室 内 液 化 试 验 可 全 面 考 虑 诸 如 有 效 围 压 、 相 对 密 度 、 应 力 历 史 等 因<br />

素 , 测 得 的 试 样 剪 切 波 速 ( 或 小 应 变 剪 切 模 量 ) 直 接 与 该 试 样 的 抗 液 化 强 度 对 应 。 迄 今 主 要 分 为 两<br />

大 类 方 法 : 一 类 是 以 Dobry 等 人 提 出 的 基 于 循 环 剪 应 变 概 念 的 波 速 评 价 方 法 , 另 一 类 是 基 于 Seed<br />

等 人 提 出 的 循 环 剪 应 力 概 念 发 展 的 波 速 方 法 。 这 两 类 方 法 都 基 于 同 一 个 认 识 , 即 砂 土 的 剪 切 波 速 和<br />

抗 液 化 强 度 受 到 土 体 结 构 性 、 应 力 状 态 等 诸 多 因 素 的 共 同 影 响 , 存 在 良 好 的 相 关 关 系 。 本 文 对 这 些<br />

剪 切 波 速 判 别 法 作 了 回 顾 , 分 析 了 砂 土 液 化 标 准 和 土 体 结 构 性 在 建 立 各 种 CRR-V s 相 关 关 系 过 程 中<br />

的 重 要 作 用 , 并 对 剪 切 波 速 法 研 究 和 应 用 中 存 在 的 一 些 问 题 作 了 讨 论 。<br />

二 、 循 环 剪 应 变 方 法<br />

对 砂 土 CRR-V s 相 关 性 的 理 论 考 虑 最 早 开 始 于 Dobry et al. (1980) [8] 的 基 于 “ 门 槛 剪 应 变 ” 的 液 化<br />

判 别 工 作 。 他 们 提 出 采 用 应 变 控 制 的 循 环 三 轴 试 验 , 以 动 剪 应 变 幅 代 替 动 应 力 比 作 为 判 别 砂 土 液 化<br />

的 指 标 , 结 合 式 (1) 将 水 平 地 面 深 度 h 处 的 等 效 地 震 动 剪 应 变 γ e 表 示 为 :<br />

τ ⎛a<br />

⎞<br />

γ<br />

e<br />

= = 0.65⎜ ⎟<br />

⎝ ⎠<br />

av max v0<br />

d<br />

( G) g G ( G/<br />

G )<br />

γ e<br />

max<br />

σ r<br />

max<br />

γ e<br />

, (2)<br />

其 中 ,(G/G max ) γe 为 相 应 于 剪 应 变 γ e 时 的 动 剪 模 量 比 。 将 γ e 与 液 化 时 相 应 的 土 体 临 界 剪 应 变 γ cr 相 比 ,<br />

当 γ e < γ cr 时 , 土 体 不 会 液 化 , 而 当 γ e >γ cr 时 , 土 体 液 化 。 可 见 , 土 体 的 临 界 剪 应 变 γ cr 的 确 定 是 关 键 。<br />

Dobry 等 人 研 究 发 现 , 在 不 排 水 循 环 中 存 在 一 个 “ 门 槛 剪 应 变 ”(γ t ), 只 有 当 剪 应 变 幅 超 过 γ t , 试<br />

样 中 才 开 始 产 生 超 静 孔 压 , 而 且 该 值 受 相 对 密 度 、 固 结 压 力 等 因 素 影 响 较 小 。 对 砂 土 而 言 , 该 门 槛<br />

剪 应 变 值 大 致 在 10 -2 % 量 级 ( 见 图 2)。<br />

[8]<br />

图 2 循 环 剪 应 变 与 孔 压 增 长 规 律<br />

在 门 槛 剪 应 变 水 平 的 假 设 下 ,Dobry 根 据 地 震 剪 应 力 比 概 念 以 及 动 应 力 应 变 关 系 给 出 如 下 式 :<br />

τ<br />

⎛<br />

av<br />

ρ ⎞⎛ G ⎞<br />

CSR = = f ( γ ) V = γ ⎜ ⎟⎜ ⎟ V<br />

σ′ ′<br />

v<br />

2 2<br />

av s av s<br />

⎝σv<br />

⎠⎝Gmax<br />

⎠γ<br />

av<br />

(3)<br />

可 见 , 在 门 槛 剪 应 变 假 设 下 , 统 计 意 义 上 砂 土 抗 液 化 强 度 将 与 其 剪 切 波 速 的 2 次 方 成 正 比 。<br />

自 Dobry 应 变 法 原 理 提 出 用 预 测 砂 土 液 化 势 以 来 , 我 国 也 开 展 了 类 似 的 研 究 工 作 , 提 出 若 干 判<br />

别 方 法 和 判 别 式 。 汪 闻 韶 (1984) [9] 进 一 步 建 立 了 剪 切 波 速 与 抗 液 化 强 度 之 间 的 理 论 关 系 , 提 出 了 场<br />

-210-


地 液 化 与 否 初 步 判 别 的 剪 切 波 速 方 法 。 首 先 , 根 据 地 震 剪 应 力 比 CSR 定 义 表 提 出 地 基 波 速 V s0 与 地 震<br />

过 程 中 地 基 模 量 、 孔 压 及 剪 应 变 的 关 系 如 下 式 :<br />

V<br />

s0<br />

=<br />

⎛a<br />

0.65⎜<br />

⎝ g<br />

max<br />

av<br />

⎝ σ<br />

v0 ⎠ ⎝Gmax<br />

⎠γ<br />

av<br />

n<br />

⎞<br />

⎟σ<br />

vrd<br />

⎠<br />

⎛ Δu<br />

⎞ ⎛ G ⎞<br />

ργ ⎜1−<br />

′<br />

⎟ ⎜ ⎟<br />

(4)<br />

当 γ av ≤γ t 时 ,Δu = 0。 若 取 γ t = 0. 01 % , 则 (G/G max )γ t = 0. 01 % = 0. 75, 于 是 相 应 的 现 场 地 震 前 的 “ 门<br />

槛 剪 切 波 速 ”V s0t 可 如 下 计 算 :<br />

σ<br />

vrd<br />

Vs0t<br />

= 291 kh<br />

(5)<br />

ρ g<br />

式 中 k h = a hmax /g, 是 与 地 震 烈 度 有 关 的 水 平 地 震 加 速 度 系 数 ( 相 应 于 地 震 烈 度 7、8、9 度 分 别 取 0. 1、<br />

0. 2 和 0. 4)。 当 实 测 剪 切 波 速 V s0 等 于 或 大 于 V s0t 时 , 可 不 考 虑 地 震 液 化 可 能 性 , 从 而 节 省 有 关 勘 探<br />

和 试 验 工 作 量 。 该 方 法 被 《 水 工 建 筑 物 抗 震 设 计 规 范 》 采 纳 为 场 地 液 化 的 初 判 条 件 。<br />

注 意 到 “ 门 槛 剪 应 变 ” 对 应 试 样 中 孔 隙 水 压 力 刚 开 始 出 现 的 状 态 , 即 土 骨 架 结 构 处 于 从 弹 性 进<br />

入 塑 性 的 临 界 状 态 , 显 然 远 未 达 到 破 坏 。 而 Seed 等 在 循 环 应 力 方 法 基 础 上 提 出 的 “ 初 始 液 化 ” 标<br />

准 , 是 指 饱 和 砂 样 在 固 结 不 排 水 等 应 力 幅 循 环 三 轴 试 验 中 首 次 出 现 孔 隙 水 压 力 峰 值 Δu max 等 于 试 样<br />

初 始 周 围 有 效 压 力 σ' 0 时 的 瞬 间 情 况 , 此 时 土 颗 粒 间 有 效 应 力 丧 失 , 土 骨 架 完 全 破 坏 。 并 以 此 时 的<br />

循 环 次 数 N L 和 相 应 的 动 剪 应 力 比 σ d /2σ' 0 ( 在 试 样 45° 斜 面 上 的 动 剪 应 力 比 ) 两 个 数 值 来 描 述 动 强 度<br />

的 。 因 此 , 对 于 某 一 试 样 , 相 同 振 次 作 用 下 达 到 “ 门 槛 剪 应 变 ” 状 态 和 “ 初 始 液 化 ” 状 态 要 求 的 动<br />

剪 应 变 ( 或 动 剪 应 力 ) 水 平 显 著 不 同 , 取 门 槛 剪 应 变 作 为 液 化 标 准 得 到 的 动 强 度 过 于 保 守 ( 见 图 3)。<br />

门 槛 剪 应 变<br />

初 始 液 化<br />

抗 液 化 强 度 , (τ d<br />

/σ'c )<br />

0.6<br />

0.4<br />

0.2<br />

门 槛 剪 应 变 标 准<br />

CRR 7.5 = 0.2<br />

图 3 不 同 液 化 标 准 下 的 动 强 度 比 较 示 意 图<br />

初 始 液 化 标 准<br />

CRR 7.5<br />

= 0.13<br />

0.0<br />

1 10 100<br />

循 环 破 坏 振 次 , Nl<br />

针 对 这 一 问 题 , 石 兆 吉 (1986) [10] 根 据 唐 山 、 海 城 等 地 震 液 化 资 料 和 室 内 实 验 指 出 ,Dobry 取 临<br />

界 剪 应 变 为 10 -4 作 为 判 别 液 化 标 准 过 于 保 守 , 以 孔 隙 水 压 力 比 等 于 1 作 为 液 化 标 准 才 合 理 , 此 时 剪<br />

应 变 约 为 2%, 相 应 的 模 量 比 约 为 0.0125( 砂 土 ) 和 0.02808( 粉 土 )。 由 此 提 出 了 判 别 饱 和 土 地 震 液 化 与<br />

否 的 “ 临 界 动 剪 应 变 幅 ” 和 相 应 的 “ 临 界 剪 切 波 速 ”, 得 到 经 验 判 别 式 。 首 先 , 在 Dobry 提 出 的 基<br />

本 表 达 式 中 引 入 两 个 影 响 因 子 因 考 虑 地 下 水 位 和 粘 粒 含 量 影 响 , 得 下 述 表 达 式 :<br />

( amax<br />

)<br />

γ ( G/<br />

G )<br />

⎡ g gh r ⎤<br />

V = ⎢<br />

⎥<br />

⎢<br />

⎣<br />

⎥<br />

⎦<br />

0.5<br />

a a<br />

s d<br />

s w c<br />

e max γ e<br />

(6)<br />

-211-


结 合 我 国 现 场 资 料 进 一 步 提 出 判 别 式 , 用 剪 切 波 速 判 别 地 面 下 15 米 范 围 内 饱 和 砂 土 和 粉 土 的 地<br />

震 液 化 。 当 实 测 剪 切 波 速 Vs 大 于 按 下 式 计 算 的 临 界 剪 切 波 速 时 , 可 判 为 不 液 化 :<br />

0.5 ⎡<br />

2<br />

⎛d<br />

⎞⎤<br />

w<br />

⎛ 3 ⎞<br />

Vscr = Vs 0 ( ds −0.0133ds<br />

) ⎢1.0 −0.185⎜ ⎟⎥⎜ ⎟<br />

⎢⎣<br />

⎝ ds<br />

⎠⎥⎝ ⎦ ρ ⎠<br />

0.5<br />

(7)<br />

其 中 V s0 是 与 烈 度 、 土 类 有 关 的 经 验 系 数 。 这 一 成 果 被 《 岩 土 工 程 勘 察 规 范 》(GB50021-1994)<br />

采 用 , 并 被 《 天 津 市 建 筑 地 基 基 础 设 计 规 范 》(TBJ1-88) 结 合 当 地 情 况 引 用 。<br />

黄 博 等 (2000) [11] 对 该 问 题 作 了 进 一 步 研 究 , 由 应 变 控 制 液 化 试 验 分 析 了 剪 应 变 γ 和 达 到 初 始 液 化<br />

时 循 环 次 数 N L 之 间 的 关 系 , 并 对 “ 临 界 动 剪 应 变 幅 ” 的 取 值 提 出 了 建 议 。 研 究 表 明 ,γ 和 N L 之 间 是<br />

一 种 递 减 关 系 , 即 随 着 循 环 次 数 的 增 大 , 产 生 初 始 液 化 所 需 的 剪 应 变 就 减 小 ( 如 图 4)。 如 同 临 界 动<br />

应 力 比 是 液 化 循 环 次 数 N L 的 函 数 一 样 , 临 界 剪 应 变 也 是 液 化 循 环 次 数 的 函 数 , 而 不 是 一 个 定 值 。<br />

[11]<br />

图 4 初 始 液 化 时 γ 和 N L 的 关 系 曲 线<br />

可 见 , 在 循 环 应 变 方 法 框 架 内 发 展 剪 切 波 速 判 别 法 , 关 键 在 于 土 体 液 化 的 临 界 剪 应 变 γ cr 的 确 定 。<br />

对 于 “ 门 槛 剪 应 变 ”,Dobry 等 人 从 试 验 到 理 论 都 作 过 充 分 的 论 证 , 但 对 于 超 过 “ 门 槛 应 变 ” 后 的<br />

复 杂 灾 变 状 态 , 是 否 存 在 较 固 定 的 “ 临 界 动 剪 应 变 幅 ” 等 判 别 指 标 则 缺 乏 统 一 认 识 。 因 此 , 目 前 基<br />

于 循 环 剪 应 变 的 波 速 判 别 方 法 作 为 一 种 保 守 的 初 判 方 法 有 其 适 用 性 , 但 对 于 初 始 液 化 状 态 的 判 别 尚<br />

存 在 定 理 分 析 的 困 难 , 需 要 进 一 步 研 究 。<br />

三 、 循 环 剪 应 力 方 法<br />

这 类 方 法 通 过 等 应 力 幅 控 制 式 的 动 三 轴 液 化 试 验 , 并 采 用 初 始 液 化 标 准 ( 或 双 幅 轴 向 应 变 ε d =<br />

5%) 确 定 土 体 抗 液 化 强 度 。 在 研 究 方 法 上 还 可 进 一 步 分 为 两 种 : 第 一 种 , 通 过 对 非 扰 动 土 样 进 行 室<br />

内 试 验 评 价 土 体 抗 液 化 强 度 , 然 后 将 其 与 现 场 剪 切 波 速 对 应 , 直 接 判 别 现 场 土 体 CRR; 第 二 种 , 基<br />

于 重 塑 样 室 内 试 验 获 得 的 CRR-V s 相 关 关 系 , 通 过 原 位 剪 切 波 速 间 接 判 别 原 位 土 体 抗 液 化 强 度 。<br />

严 格 讲 , 第 一 类 是 较 理 想 的 抗 液 化 强 度 波 速 评 价 方 法 。 室 内 液 化 试 验 揭 示 液 化 灾 变 本 质 , 而 高<br />

质 量 原 状 样 避 免 了 取 样 扰 动 和 制 样 方 法 的 影 响 , 能 反 映 原 位 土 体 结 构 特 性 , 获 得 的 抗 液 化 强 度 代 表<br />

了 现 场 原 位 土 体 强 度 。 因 此 , 这 种 方 式 下 原 位 剪 切 波 速 V s 与 室 内 动 强 度 CRR 通 过 同 一 原 位 结 构 性<br />

高 度 对 应 , 是 一 种 综 合 了 室 内 试 验 和 原 位 波 速 测 试 优 势 的 土 体 抗 液 化 强 度 可 靠 评 价 方 法 。 但 这 种 方<br />

法 需 要 进 行 现 场 冻 土 法 等 复 杂 的 原 状 取 样 , 代 价 高 昂 且 工 作 量 巨 大 [12] 。 针 对 这 一 问 题 , 不 少 学 者 探<br />

索 出 通 过 预 剪 方 法 恢 复 土 体 结 构 性 的 方 法 , 在 循 环 三 轴 试 验 中 使 试 件 制 备 成 V s = V s0 , 然 后 进 行 循<br />

环 试 验 以 获 得 与 原 位 土 体 相 当 的 CRR。Tokimatsu & Uchida (1990) [13] , 刘 小 生 等 (1996) [14] 以 及 黄 博<br />

-212-


等 (2000) [15] 在 试 验 中 都 发 现 , 预 剪 可 以 显 著 提 高 砂 土 的 波 速 和 抗 液 化 强 度 。Tokimatsu 研 究 进 一 步 证<br />

实 , 若 试 样 预 剪 后 的 G max 与 相 对 密 度 都 达 到 现 场 测 试 值 , 则 土 样 抗 液 化 强 度 与 由 原 位 冻 结 法 得 到 的<br />

高 质 量 原 状 土 抗 液 化 强 度 一 致 。 因 此 , 通 过 小 应 变 幅 的 先 期 振 动 控 制 土 样 初 始 剪 切 波 速 , 恢 复 土 样<br />

的 结 构 性 , 为 根 据 现 场 剪 切 波 速 V s0 试 验 直 接 获 得 相 应 的 抗 液 化 强 度 提 供 了 有 效 手 段 。<br />

[16-18]<br />

另 一 方 面 , 大 量 基 于 重 塑 样 的 试 验 研 究 表 明 , 砂 土 剪 切 波 速 与 抗 液 化 强 度 有 良 好 的 相 关 性<br />

( 如 图 5 和 图 6)。 因 此 , 即 使 无 法 获 得 高 质 量 的 原 状 试 样 , 只 要 在 室 内 用 不 同 相 对 密 度 重 塑 样 进 行<br />

测 量 剪 切 波 速 的 液 化 试 验 , 建 立 一 种 砂 土 CRR-V s 相 关 性 曲 线 , 就 可 根 据 曲 线 确 定 对 应 于 现 场 原 位<br />

波 速 V s 的 抗 液 化 强 度 CRR。 尽 管 都 是 基 于 重 塑 样 的 试 验 , 与 通 过 预 剪 恢 复 室 内 试 样 原 位 结 构 性 方<br />

法 不 同 , 后 者 建 立 的 CRR-Vs 相 关 关 系 具 有 更 广 的 适 用 性 , 能 对 一 种 砂 土 在 不 同 结 构 状 态 和 应 力 条<br />

件 下 的 抗 液 化 强 度 做 出 准 确 评 价 , 是 一 种 更 一 般 性 的 方 法 。<br />

[17]<br />

图 5 不 同 砂 土 CRR-V s 相 关 性 曲 线<br />

[13]<br />

图 6 归 一 化 剪 切 模 量 与 CRR 相 关 关 系<br />

需 要 指 出 的 是 , 上 述 研 究 中 的 CRR-V s 相 关 性 都 是 基 于 大 量 试 验 统 计 分 析 得 到 的 , 缺 乏 理 论 指<br />

导 。 而 且 , 很 多 研 究 中 进 行 循 环 液 化 试 验 的 土 样 与 测 量 剪 切 波 速 的 土 样 是 分 离 的 , 加 上 极 小 应 变 下<br />

剪 切 波 速 ( 或 模 量 ) 测 试 技 术 手 段 的 限 制 , 影 响 了 CRR-V s 相 关 性 曲 线 的 定 量 准 确 性 。 近 年 来 发 展<br />

起 来 的 压 电 弯 曲 元 等 无 损 测 量 技 术 , 可 安 装 在 室 内 仪 器 上 对 同 一 试 样 进 行 波 速 测 试 然 后 施 加 循 环 荷<br />

载 直 至 液 化<br />

[19-21] , 为 这 一 问 题 的 研 究 提 供 了 有 效 手 段 。 基 于 这 一 认 识 ,Chen et al. (2005) [22] 首 先 在 理<br />

论 上 初 步 建 立 了 基 于 室 内 重 塑 砂 土 的 CRR-V s 理 论 相 关 关 系 。Zhou & Chen (2007) [23] 在 此 基 础 上 进 一<br />

步 完 善 了 该 公 式 表 达 和 确 定 了 主 要 参 数 , 并 开 展 室 内 液 化 试 验 和 离 心 机 振 动 台 试 验 验 证 这 一 关 系 的<br />

可 靠 性 。 其 相 关 性 公 式 在 室 内 试 验 条 件 下 的 表 达 为 :<br />

CRR<br />

tx<br />

min<br />

f / n<br />

2<br />

1 ⎡ ρV<br />

⎤<br />

s<br />

= ⎢k ⎥<br />

(8)<br />

σ ′ ⎣ Fe ( ) ⎦<br />

m<br />

其 中 f 为 与 土 体 密 实 度 有 关 的 参 数 , 对 于 不 同 密 度 砂 土 取 值 在 0.6 到 1.0 之 间 ;n 为 Hardin 公 式 的<br />

指 数 , 通 常 砂 土 取 n = 0.5。<br />

上 述 推 导 表 明 , 对 不 同 的 砂 土 其 CRR tx -V s 相 关 性 不 唯 一 , 收 到 指 数 f/n 的 显 著 影 响 。 以 往 研 究<br />

表 明 n 在 1/3 到 2/3 之 间 变 化 [24] , 假 设 f 从 0.6-1.0 变 化 , 则 f/n 从 0.9 变 化 至 3.0。 此 时 式 (8) 中 剪 切<br />

波 速 指 数 (i.e., 2f/n) 将 在 1.8 到 6.0 之 间 变 化 , 这 也 是 与 很 多 试 验 研 究 发 现 一 致 的 [25] 。<br />

将 式 (8) 经 过 适 当 转 换 到 现 场 条 件 后 , 并 采 用 n=0.5, 得 到 相 关 性 公 式 在 现 场 条 件 下 的 表 达 式 :<br />

1 ⎡ k ρ ⎤<br />

= ⎢ ⎥<br />

⎣ ⎦<br />

N<br />

4<br />

CRR rc<br />

Vs<br />

1<br />

Pa<br />

F( emin<br />

)<br />

2<br />

(9)<br />

其 中 Pa = 100 kPa, 为 参 考 应 力 。<br />

-213-


上 式 表 明 , 尽 管 不 同 砂 土 的 CRR-V s 相 关 性 曲 线 不 同 , 但 在 统 计 意 义 上 可 以 得 到 砂 土 抗 液 化 强<br />

度 与 其 剪 切 波 速 4 次 方 成 正 比 的 结 论 。 因 此 , 采 用 式 (9) 和 下 限 (lower-bound) 参 数 , 可 对 各 种 场 地<br />

[26]<br />

液 化 可 能 性 作 相 对 保 守 判 别 。 图 7 给 出 了 全 球 地 震 现 场 数 据 库 与 上 述 典 型 CRR-V s 相 关 性 曲 线 的<br />

比 较 。Zhou & Chen (2007) 提 出 的 下 界 线 能 很 好 的 区 分 现 场 震 害 液 化 和 非 液 化 数 据 , 但 Andrus &<br />

Stokoe (2000) 曲 线 却 在 强 震 区 (CSR>0.3) 和 高 波 速 段 (V s1 >200 m/s) 重 叠 区 域 内 误 判 了 大 部 分 液<br />

化 数 据 。 而 基 于 循 环 剪 应 变 的 方 法 , 尤 其 是 基 于 “ 门 槛 剪 应 变 ” 概 念 的 Dobry( 或 汪 闻 韶 ) 方 法 显<br />

[27]<br />

然 是 很 保 守 的 。 相 比 之 下 , 基 于 初 始 液 化 标 准 下 “ 临 界 剪 应 变 ” 概 念 的 石 兆 吉 方 法 在 V s1 =100-200<br />

m/s 的 区 段 能 较 好 判 别 现 场 数 据 , 但 在 高 波 速 阶 段 仍 然 是 相 当 保 守 的 。<br />

抗 液 化 强 度 CSR or CRR<br />

四 、 相 关 讨 论<br />

0.6<br />

0.4<br />

0.2<br />

全 球 地 震 数 据<br />

Kayen et al. (2004)<br />

液 化<br />

非 液 化<br />

M w<br />

= 7.5<br />

σ' v<br />

=100 kPa<br />

Andrus & Stokoe<br />

(2000)<br />

石 兆 吉 等<br />

(1993)<br />

汪 闻 韶 (1984)<br />

0.0<br />

Dobry et al. (1980)<br />

0 100 200 300<br />

修 正 剪 切 波 速 , V s1<br />

(m/s)<br />

Zhou & Chen<br />

(2007)<br />

图 7 典 型 CRR-V s1 曲 线 在 全 球 (CSR, V s1 ) 数 据 库 中 的 表 现<br />

从 物 理 意 义 上 看 , 剪 切 波 速 V s0 反 映 了 土 体 的 初 始 刚 度 或 初 始 剪 切 模 量 G max0 。 初 始 剪 切 模 量 越<br />

大 , 在 相 同 剪 应 力 幅 τ d 作 用 下 的 变 形 越 小 , 其 体 变 和 液 化 发 展 越 慢 , 因 此 砂 土 CRR-V s 相 关 性 曲 线<br />

随 剪 切 波 速 增 长 的 整 体 趋 势 都 是 一 致 的 。 但 砂 土 液 化 灾 变 过 程 还 受 到 其 他 因 素 的 制 约 , 例 如 颗 粒 级<br />

配 、 细 粒 / 粗 粒 含 量 、 颗 粒 间 胶 结 程 度 等 土 体 结 构 性 因 素 影 响 , 导 致 其 剪 切 波 速 和 CRR 定 量 变 化 规<br />

[28,<br />

律 存 在 差 异 , 因 砂 土 CRR-V s 曲 线 应 该 与 土 类 相 关 29] 。 具 体 可 从 以 下 几 方 面 来 看 :<br />

首 先 , 颗 粒 级 配 对 砂 土 剪 切 波 速 和 抗 液 化 强 度 都 存 在 显 著 影 响 [30] , 因 此 对 于 典 型 砂 土 仍 然 有 必<br />

要 进 一 步 研 究 诸 如 粒 径 和 级 配 参 数 对 CRR-V s 相 关 性 的 影 响 规 律 。 同 时 , 对 于 目 前 备 受 关 注 的 含 细<br />

粒 砂 土 液 化 问 题 , 目 前 基 本 上 以 洁 净 砂 土 液 化 判 别 式 为 基 础 , 进 一 步 考 虑 细 粒 含 量 影 响 得 到 经 验 判<br />

别 公 式 或 相 关 性 曲 线 ( 如 图 1)。 对 于 细 粒 含 量 较 少 的 砂 土 , 尤 其 是 含 粉 粒 砂 土 , 这 种 方 法 具 有 一 定<br />

的 可 靠 性 。 然 而 ,Chu et al. [31] 对 1999 年 台 湾 地 区 集 集 地 震 现 场 液 化 数 据 研 究 后 发 现 , 图 1 的 判 别<br />

曲 线 严 重 低 估 了 含 粘 粒 砂 土 的 抗 液 化 强 度 , 表 明 仅 用 细 粒 含 量 指 标 无 法 可 靠 评 价 其 抗 液 化 强 度 。 后<br />

续 研 究 表 明 , 含 细 粒 砂 土 抗 液 化 强 度 随 细 粒 含 量 的 增 加 不 是 单 调 变 化 的 , 存 在 使 抗 液 化 强 度 最 低 的<br />

临 界 细 粒 含 量 (20-40% 之 间 )。 这 对 提 出 合 理 的 CRR-V s 相 关 性 曲 线 提 出 了 新 的 挑 战 , 尤 其 对 含 粘 粒<br />

[<br />

砂 土 需 要 更 深 入 的 研 究 32, 33] 。 另 外 , 实 际 震 害 经 验 表 明 , 含 粗 粒 的 砂 土 ( 如 砂 砾 土 ) 也 存 在 液 化 可<br />

能 性 [34] , 在 这 一 方 面 已 有 一 定 的 研 究 [35] , 但 建 立 可 靠 的 砂 砾 土 剪 切 波 速 判 别 法 有 待 于 更 多 的 现 场 调<br />

查 数 据 和 进 一 步 的 大 尺 寸 三 轴 试 验 , 以 及 理 论 上 的 深 入 认 识 。 另 外 , 砂 土 的 胶 结 效 应 也 对 其 刚 度 和<br />

强 度 特 性 有 显 著 影 响 [36] , 也 值 得 深 入 研 究 。<br />

-214-


五 、 结 论<br />

剪 切 波 速 法 是 砂 土 地 震 液 化 判 别 的 重 要 方 法 之 一 。 本 文 阐 述 了 该 方 法 的 应 用 原 理 和 发 展 过 程 ,<br />

指 出 了 液 化 标 准 和 土 体 结 构 性 对 建 立 砂 土 抗 液 化 强 度 与 其 剪 切 波 速 相 关 性 的 重 要 影 响 。 研 究 表 明 ,<br />

液 化 标 准 不 同 , 使 得 具 有 相 同 剪 切 波 速 试 样 的 抗 液 化 强 度 显 著 不 同 ; 而 不 同 类 型 砂 土 的 结 构 性 差 异<br />

显 著 , 其 抗 液 化 强 度 随 剪 切 波 速 的 变 化 规 律 也 不 同 , 因 此 现 有 各 种 剪 切 波 速 方 法 在 实 际 工 程 应 用 中<br />

的 适 用 性 和 可 靠 性 存 在 差 异 。 在 实 际 工 程 中 开 展 可 靠 的 剪 切 波 速 液 化 判 别 , 有 赖 于 建 立 与 场 地 土 类<br />

相 关 的 CRR-V s 相 关 关 系 。<br />

参 考 文 献<br />

[1] Seed, H. B. and Idriss, I. M. Simplified procedure for evaluating soil liquefaction potential. Journal of the Soil Mechanics<br />

and Foundations Division, ASCE, 1971, 97(9): 1249-1273.<br />

[2] Liao, S. S. C. and Whitman, R. V. Overburden correction factors for SPT in sands. Journal of the Geotechnical<br />

Engineering Division, ASCE, 1986, 112(3): 373-377.<br />

[3] Youd, T. L., Idriss, I. M. and Andrus, R. D., et al. Liquefaction resistance of soils: summary report from the 1996<br />

NCEER and 1998 NCEER/NSF workshops on evaluation of liquefaction resistance of soils. Journal of the Geotechnical and<br />

Geoenvironmental Engineering, ASCE, 2001, 127(10): 817- 833.<br />

[4] Yoshimi, Y., Tokimatsu, K., Kaneko, O. and Makihara, Y. Undrained cyclic shear strength of a dense Niigata sand. Soils<br />

and Foundations, 1984, 24(4): 131-145.<br />

[5] Stokoe, K. H., II, Roesset, J. M., Bierschwale, J. G. and Aouad, M.. Liquefaction potential of sands from shear wave<br />

velocity. Proc., 9th World Conf. on Earthquake Engrg., 1988, 3, 213-218.<br />

[6] Andrus, R. D. and Stokoe, K. H., II. Liquefaction resistance of soils from shear-wave velocity. Journal of Geotechnical<br />

and Geoenvironmental Engineering, ASCE, 2000, 126(11): 1015-1025.<br />

[7] Andrus, R. D., Piratheepan, P., Ellis, B. S., Zhang, J. F. and Juang, C. H. Comparing liquefaction evaluation methods<br />

using penetration-Vs relationships. Soil Dynamics and Earthquake Engineering, 2004, 24(9-10): 713-721.<br />

[8] Dobry, R. and Powell, D. J., et al. Liquefaction potential of saturated sand-the stiffness method. Proceedings of seventh<br />

world conference on earthquake engineering, 1980, 3: 25-32.<br />

[9] 汪 闻 韶 . 关 于 饱 和 砂 土 液 化 机 理 和 判 别 方 法 的 某 些 探 讨 . 水 利 水 电 科 学 研 究 院 论 文 集 : 第 16 集 , 北 京 : 水 利 电 力 出<br />

版 社 , 1984. 1-8.<br />

[10] 石 兆 吉 . 判 别 水 平 土 层 液 化 势 的 剪 切 波 速 法 . 水 文 地 质 工 程 地 质 , 1986, (4): 9-13.<br />

[11] 黄 博 . 粉 土 和 粉 砂 的 动 力 特 性 及 地 震 液 化 研 究 , 杭 州 , 浙 江 大 学 , 2000.<br />

[12] Chen, Y. C. and You, P. S. Evaluation of liquefaction potential by the test results of in-situ frozen samples. Proceedings<br />

of <strong>The</strong> Fourteenth International Offshore and Polar Engineering Conference, Toulon, France, 2004, 1: 563-570.<br />

[13] Tokimatsu, K. and Uchida, A. Correlation between liquefaction resistance and shear wave velocity. Soils and<br />

Foundations, 1990, 30(2): 33–42.<br />

[14] 刘 小 生 , 汪 闻 韶 , 常 亚 屏 . 饱 和 砂 土 的 原 位 动 强 度 确 定 方 法 探 讨 . 土 木 工 程 学 报 , 1996, (2): 65-74.<br />

[15] 黄 博 , 陈 云 敏 , 殷 建 华 等 . 控 制 试 样 初 始 剪 切 摸 量 的 动 三 轴 液 化 试 验 . 岩 土 工 程 学 报 , 2000, 22(6): 682-685.<br />

[16] De, A. P., Baldwin, K. C., Janno, V., Roe, G. and Celikko, B. Elastic-wave velocities and liquefaction potential.<br />

Geotechnical Testing Journal, ASTM, 1984, 7(2): 77-87.<br />

[17] 刘 小 生 , 赵 冬 , 汪 闻 韶 . 原 状 结 构 性 对 饱 和 砂 土 动 力 变 形 特 性 影 响 的 试 验 研 究 . 水 利 学 报 , 1993, (2) :32-42.<br />

[18] Chen, Y. C. and Lee, C. G. Evaluation of liquefaction resistance of sand by maximum shear modulus. Journal of<br />

Chinese Institute of Engineer, 1994, 17(5): 689-699.<br />

[19] Huang, Y. T., Huang, A. B., Kuo, Y. C. and Tsai, M. D. A laboratory study on the undrained strength of a silty sand<br />

from Central Western Taiwan. Soil Dynamics and Earthquake Engineering, 2004, 4(9-10): 733-743.<br />

[20] Lin, P. S., Chang, C. W. and Chang, W. J. Characterization of liquefaction resistance in gravelly soil: large hammer<br />

-215-


penetration test and shear wave velocity approach. Soil Dynamics and Earthquake Engineering, 2004, 24(9-10): 675-687.<br />

[21] 程 国 勇 , 张 立 , 王 建 华 . 用 扭 剪 波 测 试 土 样 剪 切 波 速 的 新 技 术 . 岩 土 工 程 学 报 , 2005, 27(3): 358-359.<br />

[22] Chen, Y. M., Ke, H. and Chen, R. P. Correlation of shear wave velocity with liquefaction resistance based on laboratory<br />

tests. Soil Dynamics and Earthquake Engineering, 2005, 25(6): 461-469.<br />

[23] Zhou, Y. G. and Chen, Y. M. Laboratory investigation on assessing liquefaction resistance of sandy soils by shear wave<br />

velocity. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 2007, 133(8): 959-972.<br />

[24] Cascante, G. and Santamarina, J. C. Interparticle contact behavior and wave propagation. Journal of Geotechnical<br />

Engineering, 1996, 122(10): 831-839.<br />

[25] Ishibashi, I. and Capar, O. F. Anisotropy and its relation to liquefaction resistance of granular material. Soils and<br />

Foundations, 2003, 43(5): 149-159.<br />

[26] Kayen, R., Seed, R. B., Moss, R. E., Cetin, O., Tanaka, Y. and Tokimatsu, K.. Global shear wave velocity database for<br />

probabilistic assessment of the initiation of seismic-soil liquefaction. Proc., 11th Int. Conf. On Soil Dynamics and<br />

Earthquake Engrg., 2, Berkeley, CA, 2004, 506-512.<br />

[27] 石 兆 吉 , 郁 寿 松 , 丰 万 玲 . 土 壤 液 化 势 的 剪 切 波 速 判 别 法 . 岩 土 工 程 学 报 , 1993, 15(1): 74-80.<br />

[28] 汪 闻 韶 . 剪 切 波 速 在 评 估 地 基 饱 和 砂 层 地 震 液 化 可 能 性 中 的 应 用 . 岩 土 工 程 学 报 , 2001, 23(6): 655-658.<br />

[29] 王 建 华 , 程 国 勇 . 饱 和 砂 土 的 剪 切 波 速 与 抗 液 化 强 度 相 关 性 研 究 . 岩 土 工 程 学 报 , 2005, 27(4): 369-373.<br />

[30] Wichtmann, T. and Triantafyllidis, T. Influence of the grain-size distribution curve of quartz sand on the small strain<br />

shear modulus G max . Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 2009, 135(10): 1404-1418.<br />

[31] Chu, D. B., Stewart, J. P. and Lee, S., et al. Documentation of soil conditions at liquefaction and non-liquefaction sites<br />

from 1999 Chi-Chi (Taiwan) earthquake. Soil Dynamics and Earthquake Engineering, 2004, 24(9-10): 647-657.<br />

[32] 刘 雪 珠 , 陈 国 兴 . 粘 粒 含 量 对 南 京 粉 细 砂 液 化 影 响 的 试 验 研 究 . 地 震 工 程 与 工 程 振 动 , 2003, 23(3): 150-155.<br />

[33] Gratchev, I. B., Sassa, K. and Fukuoka H. How reliable is the plasticity index for estimating the liquefaction potential of<br />

clayey sands. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 2006, 132(1): 124-127.<br />

[34] 袁 晓 铭 , 曹 振 中 , 孙 锐 , 孙 锐 , 陈 龙 伟 , 孟 上 九 , 董 林 , 王 维 铭 , 孟 凡 超 , 陈 红 娟 , 张 建 毅 , 蔡 晓 光 . 汶 川 8.0 级<br />

地 震 液 化 特 征 初 步 研 究 . 岩 石 力 学 与 工 程 学 报 , 2009, 28(6): 1288-1296.<br />

[35] Hatanaka, M., Uchida, A. and Suzuki, Y. Correlation between undrain cyclic shear strength and shear wave velocity for<br />

gravelly soils. Soils and Foundations, 1997, 37(4): 85-92.<br />

[36] Andrus, R. D., Hayati, H. and Mohanan, N. P. Correcting liquefaction resistance for aged sands using measured to<br />

estimated velocity ratio. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 2009, 135(6): 735-744.<br />

-216-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

钢 管 约 束 混 凝 土 柱 的 设 计 理 论 与 工 程 应 用<br />

周 绪 红<br />

1 1,2<br />

, 刘 界 鹏<br />

(1. 兰 州 大 学 土 木 工 程 与 力 学 学 院 , 兰 州 730000;2. 哈 尔 滨 工 业 大 学 建 筑 设 计 研 究 院 , 哈 尔 滨 150090)<br />

摘 要 : 本 文 对 钢 管 约 束 混 凝 土 柱 的 轴 压 力 学 性 能 和 抗 震 性 能 进 行 了 研 究 。 钢 管 约 束 混 凝 土 短 柱<br />

的 轴 压 试 验 结 果 表 明 , 圆 钢 管 约 束 混 凝 土 短 柱 的 钢 管 在 峰 值 荷 载 点 处 屈 服 , 而 方 钢 管 约 束 混 凝 土 短<br />

柱 的 钢 管 在 峰 值 荷 载 后 屈 服 。 根 据 试 验 结 果 建 立 了 钢 管 约 束 混 凝 土 短 柱 的 轴 压 承 载 力 计 算 公 式 。 框<br />

架 短 柱 的 抗 震 试 验 结 果 表 明 , 钢 管 约 束 混 凝 土 框 架 短 柱 具 有 良 好 的 延 性 、 弹 塑 性 变 形 能 力 和 耗 能 性<br />

能 ; 根 据 框 架 短 柱 的 试 验 和 理 论 分 析 结 果 建 立 了 抗 剪 承 载 力 公 式 。 压 弯 构 件 的 抗 震 试 验 结 果 表 明 ,<br />

钢 管 约 束 混 凝 土 压 弯 构 件 的 抗 震 性 能 优 越 ; 根 据 试 验 结 果 建 议 了 钢 管 约 束 混 凝 土 的 截 面 抗 弯 承 载 力<br />

计 算 方 法 。 根 据 框 架 短 柱 和 压 弯 构 件 的 抗 震 试 验 结 果 , 建 议 了 钢 管 约 束 混 凝 土 柱 的 抗 震 构 造 措 施 。<br />

介 绍 了 钢 管 约 束 混 凝 土 在 体 育 馆 和 超 高 层 结 构 中 的 应 用 概 况 , 可 为 工 程 实 践 提 供 参 考 。<br />

关 键 词 : 钢 管 约 束 混 凝 土 短 柱 压 弯 构 件 轴 压 承 载 力 抗 剪 承 载 力 抗 弯 承 载 力<br />

DESIGN THEORY AND APPLICATION OF TUBED CONCRETE COLUMNS<br />

X.H. Zhou 1 and L. J. Peng 1, 2<br />

1 School of Civil Engineering and Mechanics, Lanzhou <strong>University</strong>, Lanzhou 730000, China<br />

2<br />

Architectural Design and Research Institute, Harbin Institute of Technology, Harbin 150090, China<br />

Abstract: <strong>The</strong> axial loading behavior and seismic behavior of tubed concrete columns were studied in this paper.<br />

<strong>The</strong> test results indicated that the steel tube yielded at the peak load point for circular stub columns, while the<br />

steel tube yielded after the peak load point for squre stub columns. Equations for the prediction of the ultimate<br />

axial load strength of tubed concrete columns were proposed based on the test results. <strong>The</strong> test results<br />

indicated that tubed concrete short columns exhibited excellent ductility, elasto-plastic deformation capacity and<br />

energy dissipation ability. A design equation to calculate the shear strength of tubed concrete columns was<br />

proposed. <strong>The</strong> tubed concrete beam-columns exhibited excellent seismic behavior. Design method was<br />

proposed to calculate the flexural strength of tubed concrete columns. <strong>The</strong> details for seismic design were<br />

proposed based on the test and analysis results. <strong>The</strong> application of tubed concrete columns in gymnasium and<br />

high building structures was presented in this paper, which can be a reference for structural engineer.<br />

Keywords: Tubed concrete columns, stub column, beam-column, axial load strength, shear strength, flexural<br />

strength.<br />

一 、 前 言<br />

钢 管 约 束 混 凝 土 是 在 钢 管 内 填 充 混 凝 土 , 但 钢 管 不 直 接 承 担 纵 向 荷 载 , 且 只 对 核 心 混 凝 土 起 约<br />

束 作 用 的 一 种 组 合 柱 。 在 实 际 工 程 中 框 架 柱 一 般 要 承 受 轴 力 、 弯 矩 和 剪 力 的 共 同 作 用 , 因 此 将 钢 管<br />

约 束 混 凝 土 应 用 于 框 架 柱 中 时 需 配 置 纵 筋 或 型 钢 , 形 成 钢 管 约 束 钢 筋 混 凝 土 柱 [1] 或 钢 管 约 束 型 钢 混<br />

凝 土 柱 [2]。 钢 管 约 束 混 凝 土 柱 与 传 统 的 钢 管 混 凝 土 柱 有 着 本 质 不 同 。 图 1 显 示 了 钢 管 混 凝 土 框 架 柱<br />

与 钢 管 约 束 钢 筋 混 凝 土 及 钢 管 约 束 型 钢 混 凝 土 框 架 柱 的 区 别 。 钢 管 约 束 钢 筋 混 凝 土 和 钢 管 约 束 型 钢<br />

基 金 项 目 : 国 家 自 然 科 学 基 金 资 助 项 目 (50708027)<br />

作 者 简 介 : 周 绪 红 (1956- ), 男 , 湖 南 南 县 人 , 博 士 , 教 授 , 从 事 钢 结 构 及 组 合 结 构 的 教 学 与 科 研 工 作<br />

刘 界 鹏 (1978- ), 男 , 山 东 青 岛 人 , 博 士 , 副 研 究 员 , 从 事 钢 结 构 及 组 合 结 构 的 科 研 与 工 程 实 践 工 作<br />

-217-


混 凝 土 框 架 柱 的 外 包 钢 管 在 梁 柱 节 点 处 断 开 , 不 通 过 节 点 核 心 区 , 因 此 其 钢 管 不 直 接 承 担 纵 向 荷 载 ,<br />

只 对 核 心 混 凝 土 起 约 束 作 用 ; 而 传 统 钢 管 混 凝 土 中 的 钢 管 在 整 个 框 架 柱 中 不 断 开 , 既 承 担 纵 向 荷 载 ,<br />

又 在 横 向 约 束 混 凝 土 的 变 形 。<br />

在 钢 管 约 束 混 凝 土 中 , 钢 管 对 核 心 混 凝 土 的 约 束 作 用 提 高 了 核 心 混 凝 土 的 强 度 和 延 性 , 同 时 由<br />

于 钢 管 和 混 凝 土 对 纵 筋 和 型 钢 的 有 效 约 束 可 以 防 止 纵 筋 和 型 钢 板 件 失 稳 , 因 此 钢 管 约 束 钢 筋 混 凝 土<br />

框 架 柱 中 只 需 设 少 量 箍 筋 以 架 立 纵 筋 即 可 。 钢 管 约 束 型 钢 混 凝 土 框 架 柱 中 也 可 不 设 钢 筋 笼 , 减 少 钢<br />

筋 施 工 工 序 , 便 于 混 凝 土 浇 筑 ; 同 时 由 于 柱 中 无 钢 筋 笼 , 可 使 型 钢 截 面 更 加 开 展 , 提 高 了 型 钢 部 分<br />

的 抗 弯 能 力 , 充 分 利 用 了 钢 材 的 优 点 [2]。<br />

混 凝 土 梁<br />

钢 梁<br />

钢 管 贯 通 节 点 区<br />

钢 管 不 贯 通 节 点 区<br />

钢 管 不 贯 通 节 点 区<br />

(a) 钢 管 混 凝 土 柱<br />

(b) 钢 管 约 束 钢 筋 混 凝 土 柱<br />

(c) 钢 管 约 束 型 钢 混 凝 土 柱<br />

图 1 钢 管 混 凝 土 柱 与 钢 管 约 束 混 凝 土 柱<br />

钢 管 约 束 混 凝 土 柱 中 , 钢 管 不 直 接 承 担 竖 向 荷 载 , 只 对 核 心 混 凝 土 起 横 向 的 约 束 作 用 , 对 核 心<br />

混 凝 土 的 约 束 效 应 很 强 。 因 此 钢 管 约 束 混 凝 土 继 承 了 钢 管 混 凝 土 柱 承 载 力 高 和 抗 震 性 能 优 越 的 特<br />

点 。 钢 管 约 束 混 凝 土 柱 中 , 外 包 薄 壁 钢 管 的 作 用 是 在 承 载 能 力 极 限 状 态 下 或 水 平 地 震 作 用 下 对 核 心<br />

混 凝 土 起 约 束 作 用 , 同 时 抵 抗 由 水 平 地 震 作 用 产 生 的 水 平 剪 力 ; 而 在 火 灾 作 用 下 , 结 构 内 力 为 准 永<br />

久 组 合 产 生 的 效 应 , 框 架 柱 并 非 处 于 承 载 能 力 极 限 状 态 , 钢 管 对 核 心 混 凝 土 的 约 束 应 力 很 小 , 且 钢<br />

管 不 直 接 承 担 纵 向 荷 载 , 钢 管 即 使 受 火 失 效 , 核 心 钢 筋 混 凝 土 或 型 钢 混 凝 土 柱 仍 具 有 良 好 的 耐 火 性<br />

能 , 钢 管 外 不 需 做 防 火 保 护 层 ; 因 此 钢 管 约 束 混 凝 土 又 继 承 了 型 钢 混 凝 土 柱 抗 火 性 能 优 越 的 特 点 ,<br />

经 济 效 果 好 。<br />

钢 管 约 束 混 凝 土 柱 在 继 承 了 传 统 钢 管 混 凝 土 或 型 钢 混 凝 土 柱 力 学 性 能 特 点 的 同 时 , 还 避 免 了 传<br />

统 的 组 合 框 架 柱 梁 柱 节 点 连 接 较 为 复 杂 的 问 题 。 在 工 程 实 践 中 , 钢 管 混 凝 土 和 型 钢 混 凝 土 柱 是 常 用<br />

的 框 架 柱 形 式 , 但 当 框 架 梁 采 用 钢 筋 混 凝 土 时 , 节 点 连 接 复 杂 , 施 工 难 度 大 , 用 钢 量 高 。 钢 管 约 束<br />

钢 筋 混 凝 土 柱 这 一 新 型 的 组 合 结 构 形 式 能 够 解 决 钢 管 混 凝 土 或 型 钢 混 凝 土 柱 与 钢 筋 混 凝 土 梁 连 接<br />

复 杂 的 问 题 。 钢 管 约 束 混 凝 土 柱 相 当 于 将 钢 管 混 凝 土 或 型 钢 混 凝 土 柱 中 的 部 分 钢 管 或 型 钢 以 钢 筋 的<br />

形 式 弥 散 于 混 凝 土 中 , 而 将 剩 余 部 分 以 薄 壁 钢 管 的 形 式 外 包 于 混 凝 土 柱 ; 外 包 钢 管 在 梁 柱 节 点 区 断<br />

开 , 不 直 接 承 担 纵 向 荷 载 , 只 对 核 心 混 凝 土 起 约 束 作 用 。 由 于 钢 管 不 通 过 节 点 区 , 因 此 钢 筋 混 凝 土<br />

梁 与 钢 管 约 束 钢 筋 混 凝 土 柱 的 连 接 节 点 施 工 工 艺 与 普 通 钢 筋 混 凝 土 框 架 节 点 相 同 , 施 工 方 便 快 捷 ,<br />

造 价 低 。 当 框 架 梁 采 用 钢 梁 时 , 采 用 钢 管 混 凝 土 柱 的 梁 柱 节 点 用 钢 量 大 , 经 济 效 益 不 明 显 ; 而 型 钢<br />

混 凝 土 柱 与 钢 梁 的 连 接 节 点 则 存 在 柱 纵 筋 和 箍 筋 需 贯 穿 型 钢 的 问 题 , 施 工 复 杂 , 混 凝 土 浇 筑 困 难 ,<br />

现 场 施 工 难 度 大 , 难 以 发 挥 钢 结 构 施 工 方 便 快 捷 的 优 点 。 当 框 架 梁 为 钢 梁 时 , 采 用 钢 管 约 束 型 钢 混<br />

凝 土 框 架 柱 可 有 效 避 免 梁 柱 节 点 连 接 的 复 杂 的 问 题 , 且 节 点 用 钢 量 低 , 经 济 效 益 明 显 。 钢 管 约 束 型<br />

-218-


钢 混 凝 土 相 当 于 将 型 钢 混 凝 土 中 的 钢 筋 笼 外 置 , 并 以 薄 壁 钢 管 的 形 式 外 包 于 混 凝 土 ; 外 包 钢 管 在 梁<br />

柱 节 点 区 断 开 , 不 直 接 承 担 纵 向 荷 载 , 只 对 核 心 混 凝 土 起 约 束 作 用 。 在 钢 管 约 束 型 钢 混 凝 土 柱 中 ,<br />

由 于 外 包 钢 管 能 够 对 核 心 柱 产 生 有 效 约 束 作 用 , 保 证 型 钢 与 混 凝 土 共 同 工 作 , 因 此 钢 管 约 束 型 钢 混<br />

凝 土 中 不 必 再 设 置 钢 筋 笼 , 避 免 了 传 统 的 型 钢 混 凝 土 柱 中 钢 筋 笼 与 钢 梁 交 接 复 杂 的 问 题 。 钢 管 约 束<br />

型 钢 混 凝 土 中 不 设 置 钢 筋 笼 , 则 混 凝 土 浇 筑 振 捣 方 便 迅 速 , 有 利 于 发 挥 钢 结 构 施 工 快 捷 的 优 势 。<br />

[3-4]<br />

钢 管 钢 筋 混 凝 土 柱 的 概 念 (tubed column) 是 Tomii 等 在 1985 年 首 次 提 出 , 就 是 在 钢 筋 混 凝<br />

土 柱 外 设 置 钢 管 且 钢 管 在 节 点 处 断 开 , 以 防 止 框 架 短 柱 或 边 柱 发 生 剪 切 破 坏 并 提 高 其 延 性 。 随 后<br />

Aboutaha [5] 进 行 了 3 个 矩 形 压 弯 构 件 的 试 验 研 究 , 试 验 结 果 表 明 , 矩 形 钢 管 的 约 束 改 善 了 高 强 混 凝<br />

[6-8]<br />

土 柱 的 延 性 ; 肖 岩 进 行 了 剪 跨 比 为 2.0 的 方 形 截 面 局 部 加 劲 钢 套 管 加 固 钢 筋 混 凝 土 短 柱 的 抗 剪 性<br />

能 试 验 研 究 , 研 究 结 果 表 明 , 采 用 方 钢 管 约 束 钢 筋 混 凝 土 短 柱 能 有 效 提 高 构 件 的 抗 剪 承 载 力 、 延 性<br />

和 耗 能 性 能 。 在 对 钢 管 约 束 钢 筋 混 凝 土 的 早 期 研 究 中 , 研 究 者 主 要 针 对 框 架 短 柱 的 抗 震 性 能 进 行 研<br />

究 , 其 研 究 目 的 主 要 是 用 于 对 已 有 结 构 的 抗 震 加 固 [8-9] , 因 此 缺 乏 对 钢 管 约 束 钢 筋 混 凝 土 结 构 的 深 入<br />

[2,<br />

系 统 研 究 ; 而 对 钢 管 约 束 型 钢 混 凝 土 结 构 的 研 究 , 除 本 课 题 组 外 10-12] , 还 未 见 其 它 报 道 。 本 文 在<br />

国 内 外 和 本 课 题 组 的 试 验 和 理 论 分 析 结 果 的 基 础 上 建 立 了 钢 管 约 束 混 凝 土 结 构 的 设 计 理 论 , 提 出 了<br />

设 计 建 议 , 并 介 绍 了 课 题 组 推 广 应 用 钢 管 约 束 混 凝 土 结 构 的 几 个 典 型 工 程 , 可 供 工 程 实 践 参 考 。<br />

2 短 柱 的 轴 压 承 载 力 公 式<br />

2.1. 圆 钢 管 约 束 混 凝 土 短 柱 的 轴 压 承 载 力 公 式<br />

试 验 结 果 表 明 , 圆 钢 管 约 束 混 凝 土 柱 的 最 大 承 载 力 点 对 应 于 钢 管 屈 服 点 ; 在 轴 压 承 载 力 计 算 中<br />

可 假 定 钢 管 沿 环 向 完 全 屈 服 , 根 据 钢 管 的 屈 服 点 计 算 钢 管 对 核 心 混 凝 土 的 约 束 作 用 , 从 而 最 终 得 到<br />

极 限 轴 压 承 载 力 。 圆 钢 管 约 束 混 凝 土 短 柱 的 轴 压 承 载 力 公 式 为 :<br />

式 中 ,A c 、A b 、A a — 分 别 为 混 凝 土 、 钢 筋 和 型 钢 的 面 积 ;<br />

f b 、f a — 分 别 为 钢 筋 和 型 钢 的 屈 服 强 度 ;<br />

f — 核 心 约 束 混 凝 土 的 轴 心 抗 压 强 度 ;<br />

cc<br />

其 中 , f co<br />

— 非 约 束 混 凝 土 的 轴 心 抗 压 强 度 ;<br />

'<br />

f<br />

rc ,<br />

— 圆 钢 管 约 束 混 凝 土 的 有 效 约 束 应 力 :<br />

2tf<br />

' y<br />

fr<br />

=<br />

D−<br />

2t<br />

f<br />

cc<br />

5000<br />

Nu = fccAc + fbAb + faAa<br />

(1)<br />

'<br />

'<br />

= frc<br />

,<br />

frc<br />

,<br />

f ( − co<br />

1.254 + 2.254 1 + 7.94 2 )<br />

f<br />

− f<br />

(2)<br />

co<br />

co<br />

(3)<br />

4000<br />

Nuc(kN)<br />

3000<br />

本 文<br />

2000<br />

文 献 [1] 14<br />

文 献 [2] 15<br />

1000<br />

文 献 [3] 16<br />

文 献 [4] 17<br />

文 献 [5] 18<br />

0<br />

0 1000 2000 3000 4000 5000<br />

N ue (kN)<br />

图 2 圆 钢 管 约 束 混 凝 土 轴 压 短 柱 计 算 结 果 与 试 验 结 果 的 对 比<br />

-219-


图 2 为 计 算 结 果 与 本 文 及 其 它 文 献 的 试 验 研 究 结 果 的 对 比 , 其 中 文 献 [14]~[18] 为 仅 核 心 混 凝 土<br />

受 压 的 圆 钢 管 混 凝 土 轴 压 短 柱 的 试 验 结 果 , 由 图 中 可 见 , 本 文 公 式 与 试 验 结 果 吻 合 较 好 ; 公 式 适 用<br />

于 钢 管 约 束 素 混 凝 土 、 钢 管 约 束 钢 筋 混 凝 土 和 钢 管 约 束 型 钢 混 凝 土 。<br />

2.2 方 钢 管 约 束 混 凝 土 短 柱 的 轴 压 承 载 力 公 式<br />

试 验 结 果 表 明 , 轴 压 荷 载 作 用 下 , 方 钢 管 约 束 混 凝 土 短 柱 达 到 峰 值 承 载 力 时 , 钢 管 并 未 屈 服 ,<br />

钢 管 在 荷 载 - 位 移 曲 线 的 下 降 段 屈 服 。 因 此 在 计 算 方 钢 管 约 束 混 凝 土 短 柱 的 轴 压 承 载 力 时 不 能 假 定 钢<br />

管 屈 服 。 方 钢 管 约 束 混 凝 土 短 柱 的 轴 压 承 载 力 公 式 仍 采 用 式 (1), 但 f cc 计 算 方 法 不 同 :<br />

f<br />

cc<br />

'<br />

'<br />

= frs<br />

,<br />

frs<br />

,<br />

f ( − co<br />

1.254 + 2.254 1 + 7.94 2 )<br />

f<br />

− f<br />

(4)<br />

co<br />

co<br />

式 中 ,<br />

f — 方 钢 管 约 束 混 凝 土 的 有 效 约 束 应 力 :<br />

'<br />

rs ,<br />

f = k f = 0.635 f<br />

(5)<br />

'<br />

rs , e rs , rs ,<br />

式 中 , f<br />

rs ,<br />

— 方 钢 管 对 混 凝 土 的 约 束 应 力 :<br />

f<br />

'<br />

rs ,<br />

=<br />

2tσ<br />

hp<br />

D−<br />

2t<br />

(6)<br />

式 中 , σ — 峰 值 荷 载 点 处 方 钢 管 的 横 向 应 力 :<br />

hp<br />

σ = (7)<br />

0.5<br />

hp<br />

0.1( D/ t)<br />

f215<br />

式 中 ,f 215 —Q235 钢 材 的 设 计 强 度 ,f 215 =215MPa。<br />

图 3 为 方 钢 管 约 束 混 凝 土 短 柱 轴 压 承 载 力 试 验 结 果 与 本 文 公 式 计 算 结 果 的 对 比 , 吻 合 较 好 。 本<br />

文 的 公 式 可 用 于 计 算 方 钢 管 约 束 素 混 凝 土 、 方 钢 管 约 束 钢 筋 混 凝 土 和 方 钢 管 约 束 型 钢 混 凝 土 短 柱 的<br />

轴 压 承 载 力 公 式 。<br />

5000<br />

4000<br />

Nuc(kN)<br />

3000<br />

2000<br />

1000<br />

本 文<br />

文 献 [19] [23]<br />

0<br />

3<br />

0 1000 2000 3000 4000 5000<br />

N ue (kN)<br />

图 3 方 钢 管 约 束 混 凝 土 短 柱 计 算 结 果 与 试 验 结 果 的 对 比<br />

3 钢 管 约 束 混 凝 土 的 抗 剪 承 载 力 公 式<br />

根 据 试 验 结 果 , 本 文 建 议 采 用 如 下 公 式 计 算 方 钢 管 约 束 混 凝 土 柱 的 抗 剪 承 载 力 :<br />

V= ⎛<br />

+<br />

N ⎞<br />

f D + f A+ f<br />

A<br />

h + t h f<br />

⎝ ⎠<br />

2<br />

sv<br />

0.17 1 0.56<br />

⎜<br />

co 215 t yv 0 w w av<br />

14A<br />

⎟<br />

g<br />

s<br />

式 中 ,V— 抗 剪 承 载 力 ;<br />

N— 柱 子 上 作 用 的 轴 力 , 压 力 为 正 , 当 为 拉 力 时 取 N=0;<br />

(8)<br />

-220-


A g — 混 凝 土 毛 截 面 面 积 ;<br />

f co — 混 凝 土 抗 压 强 度 ;<br />

D— 方 形 截 面 边 长 ;<br />

f 215 —Q235 钢 材 的 设 计 强 度 ,f 215 =215MPa;<br />

A t — 钢 管 两 侧 腹 板 总 面 积 ;<br />

f yv — 箍 筋 屈 服 强 度 ;<br />

A sv — 箍 筋 各 肢 的 全 部 截 面 面 积 ;<br />

s — 箍 筋 间 距 ;<br />

h 0 — 核 心 钢 筋 混 凝 土 柱 的 截 面 有 效 高 度 ,h 0 =D-a,a 为 纵 筋 合 力 点 至 截 面 近 边 的 距 离 。<br />

t w ,h w — 分 别 为 型 钢 腹 板 厚 度 和 高 度 ; 当 型 钢 为 十 字 形 截 面 时 , 计 算 抗 剪 承 载 力 公 式 时 应 计 入<br />

与 受 剪 方 向 一 致 的 所 有 钢 骨 板 材 面 积 。<br />

f aw — 型 钢 腹 板 强 度 设 计 值 。<br />

公 式 (8) 中 , 第 一 项 为 混 凝 土 对 抗 剪 承 载 力 的 贡 献 , 包 括 轴 压 力 对 抗 剪 承 载 力 的 有 利 影 响 [7] ; 第<br />

二 项 为 外 包 钢 管 对 抗 剪 承 载 力 的 影 响 , 其 中 包 括 钢 管 腹 板 的 抗 剪 作 用 和 钢 管 的 受 拉 作 用 对 抗 剪 承 载<br />

力 的 贡 献 ; 第 三 项 和 第 四 项 为 箍 筋 和 型 钢 腹 板 的 抗 剪 作 用 [8 。 对 于 圆 形 截 面 , 可 简 化 为 等 面 积 的 方<br />

形 截 面 进 行 抗 剪 承 载 力 计 算 [7] 。<br />

Vuc(kN)<br />

图 4 为 本 文 公 式 计 算 结 果 与 试 验 结 果 的 对 比 ; 由 图 中 可 见 , 本 文 公 式 可 较 准 确 的 计 算 钢 管 约 束<br />

混 凝 土 柱 的 抗 剪 承 载 力 ; 公 式 适 用 于 钢 管 约 束 钢 筋 混 凝 土 和 钢 管 约 束 型 钢 混 凝 土 柱 。<br />

4 钢 管 约 束 混 凝 土 M-N 相 关 曲 线 计 算<br />

500<br />

450<br />

400<br />

350<br />

300<br />

250<br />

200<br />

150<br />

100<br />

50<br />

本 文<br />

2<br />

0<br />

0 100 200 300 400 500<br />

V ue (kN)<br />

图 4 抗 剪 承 载 力 计 算 结 果 与 试 验 结 果 的 对 比<br />

4.1. 钢 管 约 束 钢 筋 混 凝 土 柱 M-N 相 关 曲 线 计 算 方 法<br />

对 于 钢 管 约 束 钢 筋 混 凝 土 压 弯 构 件 , 由 于 核 心 混 凝 土 受 到 钢 管 的 有 效 约 束 , 混 凝 土 的 峰 值 应 变<br />

和 极 限 压 应 变 增 大 , 即 使 是 约 束 混 凝 土 的 峰 值 应 变 也 远 大 于 素 混 凝 土 的 极 限 压 应 变 0.003。 由 于 我<br />

国 混 凝 土 规 范 和 美 国 混 凝 土 规 范 ACI 中 规 定 非 约 束 混 凝 土 边 缘 纤 维 极 限 压 应 变 为 0.0033 和 0.003,<br />

约 为 非 约 束 混 凝 土 峰 值 压 应 变 (0.002 左 右 ) 的 1.5 倍 ; 因 此 本 文 在 采 用 ACI 的 计 算 方 法 时 假 定 钢 筋<br />

混 凝 土 截 面 的 压 区 边 缘 纤 维 应 变 为 约 束 混 凝 土 峰 值 应 变 ε cc 的 1.5 倍 , 然 后 根 据 平 截 面 假 定 和 截 面 平<br />

衡 条 件 计 算 截 面 的 极 限 抗 弯 承 载 力 。<br />

本 文 建 议 在 进 行 钢 管 约 束 钢 筋 混 凝 土 压 弯 构 件 设 计 时 , 对 于 构 件 的 弯 矩 - 轴 力 相 关 曲 线 , 可 先 根<br />

据 公 式 (2) 和 (4) 求 得 核 心 约 束 混 凝 土 的 抗 压 强 度 , 最 后 采 用 规 范 的 计 算 方 法 验 算 试 件 的 承 载 力 [19] 。<br />

4.2. 钢 管 约 束 型 钢 混 凝 土 柱 M-N 相 关 曲 线 计 算 方 法<br />

本 文 建 议 在 计 算 钢 管 约 束 型 钢 混 凝 土 压 弯 构 件 的 弯 矩 - 轴 力 相 关 曲 线 时 , 可 先 根 据 公 式 (2) 和 (4)<br />

-221-


求 得 核 心 约 束 混 凝 土 的 抗 压 强 度 , 然 后 根 据 欧 洲 规 范 EC4 [20] 的 截 面 全 塑 性 方 法 验 算 构 件 的 截 面 强 度<br />

承 载 力 和 构 件 整 体 稳 定 承 载 力 。<br />

4 方 钢 管 的 端 部 加 劲 方 式<br />

对 于 方 钢 管 约 束 混 凝 土 柱 , 由 于 方 钢 管 对 核 心 混 凝 土 的 约 束 效 果 不 均 匀 , 相 对 于 圆 钢 管 约 束 混<br />

凝 土 柱 , 方 钢 管 约 束 混 凝 土 柱 的 延 性 稍 差 , 因 此 在 工 程 实 践 中 , 对 于 方 钢 管 约 束 混 凝 土 柱 需 增 大 方<br />

钢 管 的 宽 厚 比 以 保 证 柱 子 的 延 性 满 足 规 范 要 求 。 可 采 用 图 5 的 两 种 方 钢 管 的 端 部 塑 性 铰 区 加 劲 方 式<br />

以 增 加 钢 管 在 端 部 塑 性 铰 区 的 钢 管 宽 厚 比 。 图 5 中 , 加 劲 板 的 厚 度 不 宜 小 于 钢 管 厚 度 , 加 劲 板 与 钢<br />

管 的 焊 接 连 接 应 保 证 为 等 强 连 接 ; 加 劲 板 在 柱 上 下 两 端 设 置 , 高 度 等 于 柱 子 直 径 D。 加 劲 板 上 宜 设<br />

孔 以 保 证 混 凝 土 柱 的 整 体 性 , 孔 直 径 不 应 大 于 加 劲 板 宽 度 的 一 半 D r /2; 孔 边 与 加 劲 板 边 缘 距 离 不 应<br />

大 于 D r /4 且 孔 净 距 不 宜 小 于 D r /2 ( 图 5c); 斜 加 劲 板 边 与 钢 管 边 缘 的 距 离 取 为 钢 管 边 长 的 1/4~1/3( 当<br />

为 1/3 时 , 加 劲 板 之 间 的 钢 管 宽 厚 比 相 同 , 约 束 效 果 最 好 )。 设 置 加 劲 板 后 , 柱 子 上 不 必 再 设 箍 筋 ;<br />

但 纵 筋 外 保 护 层 厚 度 和 纵 筋 净 距 应 满 足 现 行 《 混 凝 土 结 构 设 计 规 范 》 要 求 , 以 保 证 钢 筋 与 混 凝 土 之<br />

间 有 充 分 的 粘 结 作 用 。 采 用 加 劲 板 的 方 钢 管 约 束 混 凝 土 柱 , 应 保 证 加 劲 板 之 间 的 钢 管 宽 厚 比 小 于 70。<br />

加 劲 板<br />

加 劲 板<br />

D<br />

D/4~ D/3 D/4~ D/3<br />

D r<br />

D<br />

Dr=D/2 Dr=D/2<br />

t 1<br />

D/4~ D/3 D/4~ D/3<br />

D r =D/2 D r =D/2<br />

D<br />

D<br />

(a) 斜 加 劲 板<br />

(b) 直 加 劲 板<br />

≥D r /4 ≥D r /4 ≥D r /4 ≥D r /4<br />

D<br />

t 2<br />

≥Dr/2<br />

t 1 > t 2<br />

Dr/2<br />

t 1<br />

D<br />

≥Dr/2<br />

≥Dr/2<br />

D<br />

图 6 方 钢 管 端 部 加 厚<br />

D r<br />

D r<br />

(c) 加 劲 板 详 图<br />

图 5 方 钢 管 端 部 加 劲 板 设 置 图<br />

在 工 程 实 践 中 , 也 可 采 用 图 6 中 钢 管 端 部 加 厚 的 方 法 以 保 证 钢 管 在 塑 性 铰 区 对 核 心 混 凝 土 产 生<br />

足 够 的 约 束 作 用 。 如 图 6 中 所 示 , 在 钢 管 两 端 一 倍 边 长 范 围 内 采 用 厚 度 较 大 的 钢 管 ; 而 在 柱 子 中 部<br />

采 用 较 薄 钢 管 , 因 为 柱 跨 中 段 弯 矩 很 小 , 柱 中 的 钢 管 主 要 抵 抗 剪 力 作 用 , 而 非 对 核 心 混 凝 土 产 生 横<br />

向 的 约 束 作 用 。<br />

5 典 型 工 程 实 践<br />

-222-


5.1. 大 连 市 体 育 馆 钢 管 约 束 钢 筋 混 凝 土 柱 应 用<br />

大 连 市 体 育 馆 是 大 连 市 重 点 工 程 ,2013 年 全 运 会 场 馆 之 一 ; 图 7 为 大 连 市 体 育 馆 的 建 筑 外 观 图 。<br />

大 连 市 体 育 馆 的 结 构 体 系 分 为 两 部 分 , 即 顶 部 的 大 跨 度 弦 支 穹 顶 结 构 和 底 部 的 钢 筋 混 凝 土 框 架 结<br />

构 。 顶 部 的 大 跨 度 弦 支 穹 顶 屋 盖 形 状 为 椭 圆 形 , 椭 圆 的 短 轴 为 116m, 长 轴 为 140m。 由 于 屋 盖 的 跨<br />

度 很 大 , 且 屋 盖 顶 部 采 用 了 双 层 屋 面 以 解 决 排 水 问 题 , 同 时 屋 盖 下 部 的 吊 挂 设 备 很 重 , 导 致 屋 盖 的<br />

荷 载 较 大 , 因 此 支 承 上 部 大 跨 度 屋 盖 的 框 架 柱 截 面 较 大 , 截 面 尺 寸 为 1.0×1.8m。 结 构 整 体 分 析 结<br />

果 表 明 , 水 平 地 震 作 用 下 , 由 于 支 承 上 部 大 跨 度 屋 盖 的 框 架 柱 截 面 尺 寸 显 著 大 于 其 它 框 架 柱 , 因 此<br />

其 承 担 的 水 平 剪 力 将 显 著 高 于 其 它 框 架 柱 。 在 结 构 的 一 层 , 柱 子 净 高 为 5.4m 左 右 , 则 一 层 框 架 柱<br />

的 剪 跨 比 一 般 都 在 1.5 作 用 , 为 超 短 柱 , 如 果 采 用 普 通 钢 筋 混 凝 土 柱 , 则 难 以 满 足 结 构 整 体 的 罕 遇<br />

地 震 设 计 要 求 。<br />

图 7 大 连 市 体 育 馆 建 筑 外 观 图<br />

(a) 钢 管 现 场 吊 装<br />

(b) 混 凝 土 完 成 浇 筑<br />

图 8 大 连 市 体 育 馆 钢 管 约 束 钢 筋 混 凝 土 柱 施 工 照 片<br />

为 解 决 结 构 一 层 的 钢 筋 混 凝 土 超 短 柱 抗 震 性 能 不 足 的 问 题 , 本 工 程 的 结 构 专 业 项 目 组 进 行 了 多<br />

种 方 案 的 综 合 对 比 。 采 用 传 统 的 型 钢 混 凝 土 柱 可 解 决 超 短 柱 抗 震 性 能 不 足 的 问 题 。 但 在 本 工 程 中 ,<br />

由 于 框 架 梁 为 钢 筋 混 凝 土 , 因 此 梁 柱 节 点 连 接 复 杂 , 施 工 难 度 大 。 且 如 果 在 一 层 采 用 型 钢 混 凝 土 柱 ,<br />

则 型 钢 混 凝 土 柱 需 再 向 上 和 向 下 分 别 延 伸 一 层 作 为 过 渡 层 [21] , 进 一 步 增 加 了 施 工 难 度 , 用 钢 量 也 将<br />

显 著 增 加 。 在 本 工 程 的 一 层 采 用 钢 管 混 凝 土 代 替 钢 筋 混 凝 土 柱 也 可 有 效 提 高 超 短 柱 的 抗 震 性 能 , 但<br />

钢 管 混 凝 土 柱 与 钢 筋 混 凝 土 梁 的 连 接 节 点 仍 存 在 施 工 复 杂 和 造 价 偏 高 的 问 题 。 综 合 考 虑 大 连 体 育 馆<br />

结 构 的 施 工 问 题 和 工 程 造 价 , 结 构 专 业 项 目 组 决 定 采 用 钢 管 约 束 钢 筋 混 凝 土 以 解 决 一 层 超 短 柱 的 抗<br />

-223-


震 问 题 , 并 避 免 传 统 组 合 框 架 柱 与 钢 筋 混 凝 土 框 架 梁 连 接 复 杂 和 造 价 偏 高 的 问 题 。 图 8 为 钢 管 吊 装<br />

图 和 混 凝 土 浇 筑 后 的 钢 管 图 。 施 工 单 位 反 映 , 钢 管 约 束 混 凝 土 柱 的 外 包 薄 壁 钢 管 的 安 装 过 程 简 单 方<br />

便 , 与 普 通 钢 筋 混 凝 土 柱 的 模 板 安 装 过 程 基 本 相 同 。<br />

5.2. 大 连 中 国 石 油 大 厦 钢 管 约 束 钢 筋 混 凝 土 柱 应 用<br />

大 连 中 国 石 油 大 厦 为 中 国 石 油 总 公 司 东 北 地 区 销 售 总 部 。 建 筑 总 体 包 括 群 房 部 分 和 主 体 结 构 部<br />

分 ; 群 房 部 分 地 下 4 层 , 地 上 4 层 , 主 体 结 构 部 分 地 下 4 层 , 地 上 43 层 , 主 体 结 构 与 裙 房 之 间 设<br />

缝 。 图 9 为 大 连 中 国 石 油 大 厦 的 效 果 图 , 图 10 为 主 体 结 构 立 面 布 置 图 。 主 体 结 构 地 上 部 分 高 176m,<br />

属 超 高 层 结 构 。 主 体 结 构 内 筒 为 钢 筋 混 凝 土 筒 体 ; 结 构 外 筒 地 下 4 层 至 地 上 3 层 为 框 架 ,4 层 至 43<br />

层 为 交 叉 桁 架 筒 体 , 即 结 构 在 地 下 4 层 至 地 上 3 层 为 框 架 - 核 心 筒 结 构 , 而 4 层 至 43 层 为 筒 中 筒 结<br />

构 ; 外 围 结 构 从 第 4 层 开 始 由 框 架 结 构 体 系 转 换 为 桁 架 筒 体 结 构 体 系 , 因 此 本 结 构 属 超 高 层 复 杂 结<br />

构 [22] 。<br />

交 叉 桁 架 斜 柱<br />

斜 柱 轴 力<br />

斜 柱 轴 力<br />

水 平 分 力<br />

预 应 力 筋<br />

水 平 拉 力<br />

预 应 力 RC 梁<br />

转 换 超 短 柱<br />

转 换 短 柱<br />

图 9 大 连 中 国 石 油 大 厦 建 筑 外 观 图<br />

图 10 大 连 中 国 石 油 大 厦 主 体 结 构 立 面 布 置 图<br />

大 连 中 国 石 油 大 厦 的 初 始 结 构 方 案 为 钢 管 混 凝 土 外 筒 - 钢 筋 混 凝 土 内 筒 , 楼 面 体 系 采 用 钢 - 混 凝<br />

土 组 合 梁 。 但 本 工 程 的 建 设 单 位 提 出 , 采 用 钢 结 构 外 筒 加 组 合 楼 盖 的 结 构 方 案 经 济 性 不 好 , 结 构 成<br />

本 太 高 ; 因 此 建 设 单 位 要 求 设 计 单 位 采 用 钢 筋 混 凝 土 外 筒 和 钢 筋 混 凝 土 楼 盖 。 经 工 程 设 计 单 位 对 比 ,<br />

采 用 普 通 钢 筋 混 凝 土 外 筒 加 混 凝 土 楼 盖 方 案 可 节 省 结 构 方 面 的 投 资 2000 万 元 以 上 。 但 分 析 结 果 表<br />

明 , 为 满 足 建 筑 的 立 面 要 求 , 外 交 叉 桁 架 筒 体 的 斜 柱 截 面 尺 寸 不 能 过 大 , 为 满 足 斜 柱 的 轴 压 比 限 值<br />

要 求 , 桁 架 筒 体 的 底 部 斜 柱 需 采 用 钢 - 混 凝 土 组 合 构 件 。<br />

由 图 10 可 见 , 在 交 叉 桁 架 筒 体 中 , 梁 柱 节 点 处 , 除 梁 柱 相 交 外 , 桁 架 筒 的 斜 柱 也 在 此 相 交 。<br />

交 叉 桁 架 外 筒 节 点 处 共 有 6 根 杆 件 相 交 , 钢 筋 非 常 密 集 , 若 斜 柱 采 用 传 统 的 型 钢 混 凝 土 或 钢 管 混 凝<br />

土 柱 , 则 梁 柱 节 点 将 非 常 复 杂 , 施 工 难 度 很 大 。 且 由 图 10 可 见 , 当 桁 架 筒 斜 柱 在 筒 体 边 缘 相 交 时 ,<br />

两 斜 柱 的 轴 力 必 然 产 生 水 平 分 力 ; 普 通 钢 筋 混 凝 土 梁 难 以 承 担 斜 柱 的 巨 大 水 平 拉 力 , 因 此 需 在 混 凝<br />

-224-


土 梁 中 设 立 直 线 预 应 力 钢 筋 以 平 衡 斜 柱 的 水 平 分 力 。 如 果 斜 柱 采 用 传 统 的 型 钢 混 凝 土 或 钢 管 混 凝<br />

土 , 则 节 点 区 除 需 处 理 普 通 钢 筋 的 连 接 外 , 预 应 力 钢 筋 也 需 穿 过 柱 节 点 区 核 心 , 进 一 步 加 大 了 节 点<br />

处 理 的 难 度 , 很 难 保 证 节 点 区 的 施 工 质 量 。 经 过 本 工 程 结 构 专 业 项 目 组 的 研 究 对 比 , 项 目 组 最 终 决<br />

定 采 用 钢 管 约 束 钢 筋 混 凝 土 柱 , 解 决 普 通 钢 筋 混 凝 土 斜 柱 轴 压 比 不 满 足 要 求 的 问 题 , 同 时 避 免 传 统<br />

组 合 柱 节 点 处 理 复 杂 的 问 题 。<br />

由 图 10 可 见 , 当 结 构 外 筒 由 交 叉 桁 架 筒 体 变 为 普 通 框 架 时 , 由 于 两 根 斜 柱 并 为 一 根 直 柱 , 导<br />

致 柱 截 面 很 大 ; 在 3 层 转 换 层 处 , 框 架 柱 基 本 都 是 剪 跨 比 小 于 1.5 的 超 短 柱 , 而 在 1 层 和 2 层 , 框<br />

架 柱 也 是 剪 跨 比 小 于 2 的 短 柱 。 由 于 上 部 交 叉 桁 架 筒 体 的 刚 度 远 大 于 底 部 框 架 , 因 此 底 部 框 架 是 整<br />

个 结 构 的 薄 弱 层 , 特 别 是 底 部 框 架 柱 都 是 短 柱 甚 至 超 短 柱 , 抗 震 性 能 较 差 , 难 以 满 足 结 构 的 大 震 设<br />

计 要 求 。 在 底 部 框 架 短 柱 中 采 用 传 统 的 型 钢 混 凝 土 或 钢 管 混 凝 土 柱 可 有 效 提 高 结 构 的 整 体 抗 震 性<br />

能 , 但 由 于 框 架 梁 为 钢 筋 混 凝 土 , 梁 柱 节 点 处 理 复 杂 ; 因 此 项 目 组 在 底 部 框 架 短 柱 中 仍 采 用 钢 管 约<br />

束 钢 筋 混 凝 土 柱 , 避 免 了 传 统 组 合 框 架 柱 与 钢 筋 混 凝 土 框 架 梁 连 接 复 杂 的 问 题 。<br />

经 本 工 程 结 构 专 业 项 目 组 的 分 析 论 证 , 决 定 在 本 工 程 的 地 下 4 层 至 地 上 3 层 的 竖 直 框 架 柱 以 及<br />

地 上 4 层 至 8 层 的 桁 架 筒 斜 柱 采 用 钢 管 约 束 钢 筋 混 凝 土 柱 ; 柱 截 面 配 筋 形 式 、 构 造 措 施 及 梁 柱 节 点<br />

处 理 见 图 11。 为 防 止 薄 壁 钢 管 在 混 凝 土 浇 筑 过 程 中 涨 模 , 本 工 程 中 的 钢 管 仍 采 取 了 对 拉 钢 筋 的 防 涨<br />

模 措 施 。<br />

6 结 语<br />

图 11 大 连 中 国 石 油 大 厦 钢 管 约 束 钢 筋 混 凝 土 柱 截 面 配 筋 与 梁 柱 节 点<br />

钢 管 约 束 混 凝 土 柱 的 承 载 力 高 , 抗 震 和 抗 火 性 能 优 越 , 梁 柱 节 点 连 接 方 便 , 是 一 种 具 有 广 泛 推<br />

-225-


广 价 值 的 新 型 组 合 结 构 形 式 。 本 文 作 者 对 钢 管 约 束 混 凝 土 构 件 进 行 了 系 统 的 研 究 , 并 将 这 种 新 型 组<br />

合 结 构 形 式 在 “ 大 连 市 体 育 馆 ”、“ 大 连 市 体 育 场 ”、“ 大 连 中 国 石 油 大 厦 ”、“ 丹 东 体 育 场 ”、“ 淮 南 体<br />

育 场 ”、“ 黑 龙 江 省 博 物 馆 ”、“ 哈 尔 滨 科 技 大 厦 ”、“ 哈 尔 滨 投 资 大 厦 ” 等 重 点 工 程 中 推 广 应 用 , 取 得<br />

了 显 著 的 经 济 效 益 和 综 合 效 益 。 可 以 预 见 , 钢 管 约 束 混 凝 土 结 构 以 其 承 载 力 高 、 抗 震 性 能 优 越 、 抗<br />

火 性 能 好 、 在 高 强 材 料 应 用 方 面 优 势 明 显 、 梁 柱 连 接 节 点 施 工 方 便 等 突 出 优 点 , 将 在 高 层 或 超 高 层<br />

结 构 、 大 跨 空 间 结 构 的 下 部 支 承 结 构 、 大 型 复 杂 公 共 建 筑 、 大 跨 度 桥 梁 的 桥 墩 、 大 型 重 载 复 杂 工 业<br />

厂 房 和 城 市 大 型 地 下 建 筑 等 大 型 复 杂 结 构 中 得 到 广 泛 应 用 , 取 得 良 好 的 经 济 效 益 和 显 著 的 社 会 效<br />

益 。<br />

参 考 文 献<br />

[1] 周 绪 红 , 张 素 梅 , 刘 界 鹏 . 钢 管 约 束 钢 筋 混 凝 土 压 弯 构 件 滞 回 性 能 试 验 研 究 与 分 析 . 建 筑 结 构 学 报 , 2008, 29(5):<br />

19-28.<br />

[2] Zhou, X.H.and Liu, J.P. Seismic behavior and strength of tubed steel reinforced-concrete (SRC) short columns. Journal of<br />

Constructional Steel Research. 2010, 66(7):885-896.<br />

[3] Tomii, M., Sakino, K., Xiao, Y. and Watanabe, K. Earthquake resisting hysteretic behavior of reinforced concrete short<br />

columns confined by steel tube. Proceeding of the international speciality conference on concrete filled steel tubular structures,<br />

Harbin, China, August 1985, 119-125.<br />

[4] Xiao, Y., Tomii, M. and Sakion, K. Experimental study on design method to prevent shear failure of reinforced concrete<br />

short circular columns by confining in steel tube. Transactions of Japan Concrete Institute, 1986, 8: 535-542.<br />

[5] Aboutaha, R. S. and Machado, R. I. Seismic Resistance of Steel-Tubed High-Strength Reinforced-Concrete Columns.<br />

Journal of Structural Engineering. 1999, 125 (5): 485-494.<br />

[6] 肖 岩 , 郭 玉 荣 , 何 文 辉 , 吴 徽 . 局 部 加 劲 钢 套 管 加 固 钢 筋 混 凝 土 柱 的 研 究 . 建 筑 结 构 学 报 , 2003, 24(6):79-86.<br />

[7] 肖 岩 , 何 文 辉 . 约 束 钢 管 混 凝 土 结 构 柱 的 开 发 研 究 . 中 国 钢 结 构 协 会 钢 - 混 凝 土 组 合 结 构 分 会 , 第 九 次 年 会 论 文<br />

集 . 哈 尔 滨 工 业 大 学 学 报 ( 增 刊 ), 2003,35: 40-42.<br />

[8] 肖 岩 . 套 管 钢 筋 混 凝 土 结 构 的 发 展 与 展 望 . 土 木 工 程 学 报 , 2004, 37(4): 8-12.<br />

[9] Priestley. M. J. N. Seible. F, Calvi. M. Seismic design and retrofit of bridges. John Wiley & Sons, 1996.<br />

[10] 刘 界 鹏 , 张 小 冬 , 张 素 梅 . 圆 钢 管 约 束 钢 骨 高 强 混 凝 土 短 柱 轴 压 力 学 性 能 . 建 筑 结 构 学 报 , 2010, 31(2):242-248.<br />

[11] 周 绪 红 , 刘 界 鹏 , 张 素 梅 . 方 钢 管 约 束 型 钢 混 凝 土 短 柱 抗 震 性 能 试 验 研 究 . 建 筑 结 构 学 报 , 2010, 31(7):49-55.<br />

[12] 周 绪 红 , 刘 界 鹏 , 张 素 梅 . 钢 管 约 束 型 钢 高 强 混 凝 土 压 弯 构 件 的 抗 震 性 能 研 究 . 土 木 工 程 学 报 , 2010, 43(9):1-11.<br />

[13] 周 绪 红 , 刘 界 鹏 . 钢 管 约 束 混 凝 土 柱 的 性 能 与 设 计 . 科 学 出 版 社 , 2010.<br />

[14] O’Shea, M. D. and Russsell, Q. Bridge. Design of circular thin-walled concrete filled steel tubes. Journal of Structural<br />

Engineering, 2000, 126(11): 1295-1303.<br />

[15] Lahlou, K. Comportement des colonnes courtes en beton à hautes performances confine dans des tubes circulaires en acier<br />

soumises à des efforts de compression. PhD thesis, Universite de Sherbrooke, Sherbrooke, PQ, Canada (in French), 1994.<br />

[16] 张 素 梅 , 刘 界 鹏 , 马 乐 , 邢 涛 . 圆 钢 管 约 束 高 强 混 凝 土 轴 压 短 柱 的 试 验 研 究 与 承 载 力 分 析 . 土 木 工 程 学 报 , 2007,<br />

40(3): 24-31.<br />

[17] Mei, H., Kiousis, P. D., Ehsani, M. R. and Saadatmanesh, H. Confinement effects on high-strength concrete. ACI<br />

Structural Journal, 2001, 98(4): 548-553.<br />

[18] Fam, A., Qie, F. S. and Rizkalla, S. Concrete-filled steel tubes subjected to axial compression and lateral cyclic loads.<br />

Journal of Structural Engineering, 2004, 130(4): 631-640.<br />

[19] 周 绪 红 , 张 素 梅 , 刘 界 鹏 . 钢 管 约 束 钢 筋 混 凝 土 压 弯 构 件 滞 回 性 能 试 验 研 究 与 分 析 . 建 筑 结 构 学 报 , 2008, 29(5):<br />

19-28.<br />

[20] Eurocode4 (EC4). Design of composite steel and concrete structure----Part1.1: General rules and rules for buildings.<br />

1994.<br />

[21] YB 9082-2006. 钢 骨 混 凝 土 结 构 技 术 规 程 . 北 京 , 冶 金 工 业 出 版 社 , 2003.<br />

[22] 张 小 冬 , 刘 界 鹏 . 大 连 中 国 石 油 大 厦 结 构 方 案 优 化 设 计 . 建 筑 结 构 学 报 . 2009, 30 (s1): 27-33.<br />

-226-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

钢 - 混 凝 土 组 合 结 构 桥 梁 研 究 新 进 展<br />

聂 建 国 陶 慕 轩 吴 丽 丽 聂 鑫 李 法 雄 雷 飞 龙<br />

( 清 华 大 学 土 木 工 程 系 土 木 工 程 安 全 与 耐 久 教 育 部 重 点 实 验 室 , 北 京 100084)<br />

摘 要 : 钢 - 混 凝 土 组 合 结 构 桥 梁 近 年 来 在 我 国 得 到 了 迅 速 的 发 展 。 在 传 统 桥 梁 结 构 形 式 的 基 础<br />

上 , 发 展 了 多 种 新 型 组 合 结 构 桥 梁 形 式 , 拓 宽 了 组 合 结 构 桥 梁 的 应 用 领 域 。 本 文 介 绍 了 近 年 来 在 钢<br />

- 混 凝 土 组 合 结 构 桥 梁 方 面 的 最 新 研 究 进 展 , 内 容 包 括 波 形 钢 腹 板 组 合 梁 桥 、 槽 型 钢 - 混 凝 土 组 合 梁<br />

桥 、 钢 - 混 凝 土 组 合 刚 构 桥 、 双 重 组 合 作 用 连 续 组 合 梁 桥 和 大 跨 斜 拉 桥 组 合 桥 面 系 。 通 过 对 传 统 结 构<br />

形 式 的 改 进 和 发 展 , 可 充 分 发 挥 组 合 结 构 桥 梁 的 综 合 优 势 , 研 究 结 果 表 明 , 钢 - 混 凝 土 组 合 结 构 桥 梁<br />

具 有 广 阔 的 推 广 应 用 前 景 。<br />

关 键 词 : 钢 - 混 凝 土 组 合 结 构 桥 梁 波 形 钢 腹 板 槽 型 组 合 梁 组 合 刚 构 桥 双 重 组 合 组<br />

合 桥 面 系<br />

RESEARCH ADVANCES OF STEEL-CONCRETE COMPOSITE BRIDGES<br />

J. G. Nie 1 , M. X. Tao 1 , L. L. Wu 1 , X. Nie 1 , F. X. Li 1 and F. L. Lei 1<br />

1 Key Laboratory of Civil Engineering Safety and Durability of China Education Ministry,<br />

Department of Civil Engineering, Tsinghua <strong>University</strong>, Beijing 100084, China<br />

Abstract: Steel-concrete composite bridges have been developed rapidly in recent years in China. Several new<br />

types of composite bridges have been developed on the basis of traditional structures to broaden the application<br />

area of composite bridges. In this paper, some advances in research of steel-concrete composite bridges in recent<br />

years are summarized. <strong>The</strong> main research work includes composite girder bridges with corrugated steel web,<br />

channel-shaped steel-concrete composite girder bridges, steel-concrete composite rigid frame bridges,<br />

continuous composite bridges with double composite action and composite deck systems for large-span<br />

cable-stayed bridges. Through the improvement and development of the traditional structural forms, the<br />

comprehensive advantages of composite bridges can be fully displayed, which demonstrates a good prospect of<br />

application and extension for steel-concrete composite bridges.<br />

Keywords: steel-concrete composite structure, bridge, corrugated steel web, channel-shaped composite girder,<br />

composite rigid frame bridge, double composite, composite deck system.<br />

一 、 引 言<br />

钢 - 混 凝 土 组 合 结 构 桥 梁 ( 简 称 组 合 桥 ) 是 指 将 钢 梁 与 混 凝 土 桥 面 板 通 过 抗 剪 连 接 件 连 接 成 整<br />

体 并 考 虑 共 同 受 力 的 桥 梁 结 构 形 式 。 相 对 于 不 按 组 合 结 构 设 计 的 纯 钢 桥 , 组 合 桥 可 以 有 效 减 小 结 构<br />

高 度 , 提 高 结 构 刚 度 , 减 小 结 构 在 活 荷 载 下 的 挠 度 。 通 过 抗 剪 连 接 件 的 连 接 作 用 , 混 凝 土 桥 面 板 对<br />

钢 梁 受 压 翼 缘 起 到 约 束 作 用 , 从 而 增 强 了 钢 梁 的 稳 定 性 , 有 利 于 材 料 强 度 的 充 分 发 挥 。 截 面 高 度 的<br />

降 低 , 使 结 构 外 形 更 加 纤 巧 , 改 善 桥 梁 的 景 观 效 果 , 有 利 于 增 加 桥 下 净 空 或 降 低 桥 面 高 程 。 组 合 桥<br />

相 对 于 混 凝 土 桥 , 上 部 结 构 高 度 降 低 、 自 重 减 轻 、 地 震 作 用 减 小 , 结 构 延 性 提 高 、 基 础 造 价 降 低 。<br />

同 时 , 组 合 桥 便 于 工 厂 化 生 产 、 现 场 安 装 质 量 高 、 施 工 费 用 低 、 施 工 速 度 快 , 并 可 以 适 用 于 传 统 砖<br />

基 金 项 目 : 长 江 学 者 和 创 新 团 队 发 展 计 划 资 助 项 目 (IRT00736)<br />

作 者 简 介 : 聂 建 国 (1958-), 男 , 湖 南 衡 阳 人 , 教 授 , 博 导 , 从 事 钢 - 混 凝 土 组 合 结 构 的 教 学 、 科 研 和 工 程 实 践 工 作 。<br />

-227-


石 及 混 凝 土 结 构 难 以 应 用 的 情 况 [1] 。<br />

组 合 桥 自 20 世 纪 50 年 代 之 后 得 到 了 迅 速 的 发 展 , 从 20~25m 跨 径 的 中 小 跨 径 梁 桥 到 跨 径 近 千<br />

米 的 斜 拉 桥 , 都 有 组 合 结 构 的 应 用 [2] , 近 年 来 , 除 常 用 的 组 合 板 梁 桥 和 组 合 箱 梁 桥 之 外 , 相 继 研 发<br />

了 波 形 钢 腹 板 组 合 梁 桥 、 组 合 桁 梁 桥 、 组 合 刚 构 桥 等 一 系 列 新 的 结 构 形 式 , 拓 宽 了 组 合 桥 的 应 用 领<br />

域 。 而 在 国 内 , 随 着 道 路 等 级 的 不 断 提 高 和 建 设 规 模 的 扩 大 , 桥 梁 呈 现 出 跨 径 不 断 增 大 、 桥 型 不 断<br />

丰 富 、 结 构 不 断 轻 型 化 的 发 展 趋 势 , 同 时 对 桥 梁 建 设 的 经 济 性 和 综 合 效 益 也 越 来 越 重 视 。 在 这 种 背<br />

景 和 需 求 下 , 传 统 的 桥 梁 结 构 形 式 在 许 多 情 况 下 已 经 不 能 满 足 设 计 、 建 造 和 使 用 的 要 求 。 钢 - 混 凝<br />

土 组 合 结 构 桥 梁 由 于 兼 有 钢 桥 和 混 凝 土 桥 的 优 点 , 适 合 我 国 基 本 建 设 的 国 情 , 近 20 年 来 已 得 到 迅<br />

速 发 展 。<br />

本 文 重 点 介 绍 了 笔 者 所 在 的 研 究 团 队 近 年 来 在 钢 - 混 凝 土 组 合 结 构 桥 梁 方 面 完 成 的 一 些 最 新 研<br />

究 和 实 践 工 作 , 内 容 主 要 包 括 波 形 钢 腹 板 组 合 梁 桥 、 槽 型 钢 - 混 凝 土 组 合 梁 桥 、 钢 - 混 凝 土 组 合 刚 构<br />

桥 、 双 重 组 合 作 用 连 续 组 合 梁 桥 和 大 跨 斜 拉 桥 组 合 桥 面 系 , 通 过 对 传 统 结 构 形 式 的 改 进 和 发 展 , 可<br />

充 分 发 挥 组 合 结 构 桥 梁 的 综 合 优 势 。 研 究 结 果 表 明 , 组 合 结 构 桥 梁 具 有 广 阔 的 应 用 前 景 。<br />

二 、 波 形 钢 腹 板 组 合 梁 桥<br />

减 轻 大 跨 度 预 应 力 混 凝 土 桥 梁 上 部 结 构 的 重 量 是 桥 梁 结 构 技 术 革 新 的 一 个 重 要 方 向 。 对 于 预 应<br />

力 混 凝 土 箱 梁 而 言 , 由 于 需 要 在 腹 板 内 布 筋 和 使 预 应 力 筋 转 向 , 必 须 增 加 腹 板 厚 度 , 腹 板 面 积 可 占<br />

总 截 面 面 积 的 25%~35%。 因 此 , 减 小 腹 板 厚 度 对 减 轻 箱 梁 自 重 和 减 少 预 应 力 筋 用 量 是 非 常 有 效 的<br />

途 径 之 一 。 为 减 小 腹 板 厚 度 , 法 国 首 先 提 出 了 用 平 面 钢 腹 板 来 代 替 传 统 箱 梁 的 混 凝 土 腹 板 , 并 通 过<br />

箱 形 截 面 内 的 体 外 预 应 力 筋 来 施 加 预 应 力 。 经 工 程 实 践 , 用 平 钢 腹 板 代 替 传 统 箱 梁 的 混 凝 土 腹 板 后 ,<br />

自 重 减 轻 达 到 25%~35% [3] , 但 由 于 顶 板 和 底 板 的 混 凝 土 因 徐 变 、 收 缩 而 产 生 的 变 形 受 到 钢 腹 板 的 约<br />

束 , 使 得 顶 板 和 底 板 内 的 预 应 力 有 向 钢 腹 板 转 移 的 趋 势 , 后 者 承 担 了 大 约 20%~25% 的 预 应 力 , 不<br />

但 降 低 了 预 应 力 的 使 用 效 率 , 同 时 也 要 求 在 钢 腹 板 上 增 设 纵 向 或 竖 向 加 劲 肋 来 防 止 屈 曲 。 为 解 决 由<br />

于 钢 腹 板 的 约 束 作 用 而 造 成 的 截 面 预 应 力 损 失 , 法 国 Campenon Bernard 公 司 于 1975 年 提 出 了 用 波 形<br />

钢 腹 板 来 代 替 平 面 钢 腹 板 的 设 想 。<br />

传 统 的 波 形 钢 腹 板 组 合 梁 构 造 形 式 如 图 1a 所 示 , 与 传 统 的 预 应 力 混 凝 土 腹 板 箱 梁 桥 相 比 , 通 过<br />

用 波 形 钢 腹 板 替 换 混 凝 土 腹 板 , 可 减 轻 自 重 , 有 效 改 善 结 构 受 力 性 能 、 施 工 性 能 以 及 经 济 性 能 ; 充<br />

分 发 挥 材 料 潜 能 , 提 高 材 料 效 率 ; 改 善 结 构 的 施 工 性 能 、 正 常 使 用 性 能 和 长 期 性 能 。 与 纵 向 加 劲 的<br />

平 面 钢 腹 板 组 合 梁 桥 相 比 , 这 种 波 形 钢 腹 板 组 合 梁 桥 能 有 效 减 小 腹 板 对 混 凝 土 翼 缘 板 的 约 束 , 提 高<br />

预 应 力 导 入 度 , 提 高 材 料 效 率 ; 大 大 提 高 腹 板 的 剪 切 屈 曲 强 度 , 降 低 其 对 几 何 初 始 缺 陷 的 敏 感 性 ;<br />

充 分 利 用 波 形 钢 腹 板 的 三 维 挠 曲 特 性 , 方 便 施 工 , 改 善 拼 装 条 件 。 然 而 , 传 统 的 波 形 钢 腹 板 组 合 梁<br />

在 下 翼 缘 混 凝 土 板 浇 筑 时 , 需 现 场 搭 设 满 堂 红 脚 手 架 以 支 撑 模 板 , 存 在 现 场 作 业 及 施 工 难 度 较 大 等<br />

问 题 , 同 时 下 翼 缘 混 凝 土 板 和 下 翼 缘 钢 板 的 结 合 部 由 于 浇 注 空 间 狭 小 , 混 凝 土 浇 注 质 量 难 以 保 证 ,<br />

影 响 结 合 部 的 受 力 性 能 ; 另 一 方 面 , 下 翼 缘 混 凝 土 板 通 过 连 接 件 悬 挂 于 波 形 钢 腹 板 下 方 , 在 正 弯 矩<br />

作 用 下 容 易 过 早 开 裂 , 从 而 降 低 结 构 的 承 载 力 及 刚 度 , 影 响 结 构 的 耐 久 性 , 这 些 问 题 都 给 该 类 结 构<br />

的 推 广 应 用 造 成 了 一 定 的 困 难 。<br />

针 对 上 述 传 统 波 形 钢 腹 板 的 不 足 , 笔 者 通 过 改 进 构 造 , 提 出 了 多 种 新 型 波 形 钢 腹 板 组 合 梁 桥 形<br />

式 , 如 图 1b~f 所 示 。<br />

图 1b 是 一 种 下 翼 缘 改 进 型 波 形 钢 腹 板 组 合 梁 桥 , 相 比 传 统 的 波 形 钢 腹 板 组 合 梁 桥 , 主 要 优 势 包<br />

括 如 下 几 个 方 面 :(1) 通 过 将 下 翼 缘 混 凝 土 板 置 于 下 翼 缘 钢 板 之 上 , 下 翼 缘 钢 板 可 以 为 下 翼 缘 混 凝<br />

土 板 的 浇 注 提 供 可 靠 的 模 板 支 撑 , 从 而 避 免 现 场 搭 设 满 堂 红 脚 手 架 , 降 低 混 凝 土 湿 作 业 工 作 量 和 施<br />

工 难 度 , 缩 短 施 工 周 期 , 加 快 施 工 进 度 。 整 个 施 工 过 程 可 以 不 中 断 桥 下 交 通 , 施 工 快 捷 , 综 合 经 济<br />

效 益 好 , 是 对 传 统 波 形 钢 腹 板 组 合 箱 梁 的 明 显 改 进 ;(2) 下 翼 缘 钢 与 混 凝 土 结 合 部 构 造 简 单 , 混 凝<br />

土 浇 注 空 间 大 , 浇 注 质 量 易 于 保 证 , 结 合 部 的 受 力 性 能 得 到 有 效 改 善 ;(3) 下 翼 缘 钢 板 置 于 下 翼 缘<br />

-228-


混 凝 土 板 外 侧 , 受 力 更 为 合 理 , 下 翼 缘 钢 板 不 仅 自 身 能 更 充 分 地 参 与 结 构 受 力 , 发 挥 抗 拉 性 能 好 的<br />

特 点 , 同 时 对 下 翼 缘 混 凝 土 板 提 供 部 分 约 束 , 从 而 提 高 结 构 的 承 载 力 、 刚 度 和 抗 裂 性 能 , 改 善 结 构<br />

的 耐 久 性 ;(4) 下 翼 缘 混 凝 土 板 为 上 翼 缘 混 凝 土 板 的 浇 筑 提 供 了 支 撑 , 减 少 了 上 翼 缘 混 凝 土 板 的 模<br />

板 工 程 和 混 凝 土 湿 作 业 工 作 量 ;(5) 下 翼 缘 改 进 型 波 形 钢 腹 板 组 合 箱 梁 相 比 传 统 波 形 钢 腹 板 组 合 箱<br />

梁 在 受 力 性 能 、 施 工 性 能 等 方 面 得 到 显 著 改 善 的 同 时 , 用 钢 量 几 乎 没 有 增 加 , 经 济 性 能 较 优 。<br />

(a) 传 统 型 波 形 钢 腹 板 组 合 梁 (b) 下 翼 缘 改 进 型 波 形 钢 腹 板 组 合 梁 (c) 采 用 波 形 钢 腹 板 的 普 通 组 合 梁<br />

(d) 采 用 波 形 钢 腹 板 叠 合 板 组 合 梁 (e) 下 弦 开 敞 桁 式 波 形 钢 腹 板 组 合 梁 (f) 下 弦 开 敞 桁 式 波 形 钢 腹 板 叠 合 板 组 合 梁<br />

图 1 不 同 形 式 的 波 形 钢 腹 板 组 合 梁 桥<br />

图 1b 所 示 的 下 翼 缘 改 进 型 波 形 钢 腹 板 主 要 适 用 于 连 续 梁 桥 , 对 于 简 支 梁 桥 , 可 以 采 用 图 1c 所 示<br />

的 结 构 形 式 , 取 消 下 翼 缘 混 凝 土 板 , 这 样 上 翼 缘 混 凝 土 受 压 , 下 翼 缘 钢 板 受 拉 , 可 充 分 发 挥 材 料 的<br />

强 度 , 避 免 了 下 翼 缘 混 凝 土 开 裂 的 问 题 。 这 种 波 形 钢 腹 板 组 合 梁 也 可 以 看 作 采 用 波 形 钢 腹 板 的 普 通<br />

组 合 梁 , 因 此 它 除 了 具 有 普 通 组 合 梁 的 优 点 外 ( 刚 度 大 、 承 载 力 高 、 施 工 方 便 、 自 重 轻 、 抗 震 性 能<br />

好 等 ), 还 能 有 效 减 小 收 缩 徐 变 、 温 度 效 应 等 对 结 构 的 不 利 影 响 。 如 果 将 图 1c 的 现 浇 混 凝 土 板 改 为 叠<br />

合 混 凝 土 板 , 就 形 成 了 图 1d 所 示 的 采 用 波 形 钢 腹 板 的 叠 合 板 组 合 梁 , 采 用 叠 合 混 凝 土 板 后 , 不 仅 能<br />

保 留 现 浇 混 凝 土 板 整 体 性 好 的 优 点 , 同 时 预 制 混 凝 土 板 可 作 为 现 浇 混 凝 土 层 的 施 工 模 板 , 现 场 施 工<br />

可 以 不 用 模 板 和 脚 手 架 , 降 低 混 凝 土 湿 作 业 工 作 量 和 施 工 难 度 , 缩 短 施 工 周 期 。<br />

图 1e 所 示 为 下 弦 开 敞 桁 式 波 形 钢 腹 板 组 合 梁 。 对 于 简 支 组 合 梁 , 下 弦 钢 管 可 施 加 预 应 力 , 以 提<br />

高 结 构 的 承 载 力 , 对 于 连 续 组 合 梁 , 正 弯 矩 区 下 弦 钢 管 仍 可 施 加 预 应 力 , 负 弯 矩 区 下 弦 钢 管 可 灌 注<br />

混 凝 土 , 从 而 提 高 负 弯 矩 截 面 的 刚 度 和 承 载 力 , 同 时 下 弦 混 凝 土 施 工 十 分 方 便 , 不 需 要 模 板 脚 手 架 ,<br />

钢 管 外 包 混 凝 土 , 从 根 本 上 解 决 了 传 统 波 形 钢 腹 板 组 合 梁 下 翼 缘 混 凝 土 施 工 困 难 、 易 于 开 裂 、 耐 久<br />

性 不 足 等 问 题 。 此 外 , 下 弦 水 平 支 撑 和 斜 撑 可 为 上 翼 缘 混 凝 土 板 的 浇 筑 提 供 支 撑 , 减 少 了 模 板 工 程<br />

和 混 凝 土 湿 作 业 工 作 量 和 支 模 工 序 , 工 业 化 程 度 高 , 钢 结 构 部 分 主 要 为 工 厂 预 制 , 质 量 易 于 保 证 。<br />

同 样 的 , 可 以 在 图 1e 的 基 础 上 进 一 步 将 上 翼 缘 现 浇 混 凝 土 板 替 换 为 叠 合 混 凝 土 板 , 简 化 施 工 工 艺 ,<br />

缩 短 施 工 工 期 , 减 小 施 工 对 周 围 环 境 的 不 利 影 响 。<br />

波 形 钢 腹 板 和 混 凝 土 翼 缘 板 的 有 效 结 合 是 波 形 钢 腹 板 组 合 梁 设 计 的 关 键 问 题 之 一 。 传 统 的 结 合<br />

方 式 为 采 用 图 2a 所 示 的 开 孔 板 和 栓 钉 连 接 件 共 同 使 用 的 混 合 连 接 件 。 采 用 这 种 形 式 的 连 接 件 主 要 存<br />

-229-


在 如 下 不 足 :(1) 开 孔 板 和 栓 钉 连 接 件 传 力 分 担 比 例 不 清 ;(2) 开 孔 板 连 接 件 的 抗 剪 钢 筋 布 设 困 难 ;<br />

(3) 桥 面 板 横 向 钢 筋 布 设 困 难 ;(4) 桥 面 混 凝 土 板 被 开 孔 板 分 割 , 整 体 性 差 。 而 采 用 如 图 2b 所 示 的 栓<br />

钉 连 接 件 形 式 , 能 十 分 有 效 克 服 上 述 四 点 不 足 。<br />

(a) 开 孔 板 和 栓 钉 混 合 连 接 件<br />

图 2 两 种 连 接 件 形 式<br />

(b) 栓 钉 连 接 件<br />

为 了 进 一 步 研 究 传 统 及 改 进 后 的 波 形 钢 腹 板 组 合 梁 的 受 力 性 能 , 进 行 了 大 比 例 模 型 试 验 。 试 验<br />

共 设 计 了 三 个 试 件 , 其 截 面 和 构 造 形 式 如 图 3 所 示 。 试 件 A 为 传 统 波 形 钢 腹 板 组 合 梁 , 上 翼 缘 采 用 开<br />

孔 板 和 栓 钉 混 合 连 接 件 , 下 翼 缘 采 用 开 孔 板 连 接 件 ( 图 3a); 试 件 B 为 连 接 件 改 进 型 波 形 钢 腹 板 组 合<br />

梁 , 和 试 件 A 相 比 , 连 接 件 形 式 改 为 栓 钉 连 接 件 ( 图 3b); 试 件 C 为 下 翼 缘 改 进 型 波 形 钢 腹 板 组 合 梁 ,<br />

连 接 件 形 式 为 栓 钉 连 接 件 ( 图 3c)。 三 个 试 件 的 下 翼 缘 预 应 力 水 平 及 用 钢 量 基 本 保 持 一 致 , 使 试 验 结<br />

果 具 有 可 比 性 。 图 4 所 示 为 三 个 试 件 的 制 作 情 况 , 由 于 对 于 下 翼 缘 混 凝 土 进 行 了 较 大 的 优 化 , 试 件 C<br />

的 加 工 难 度 明 显 小 于 试 件 A 和 B。<br />

(a) 传 统 型 ( 试 件 A) (b) 连 接 件 改 进 型 ( 试 件 B) (c) 下 翼 缘 改 进 型 ( 试 件 C)<br />

图 3 三 个 试 件 的 截 面 及 构 造 形 式<br />

(a) 传 统 型 ( 试 件 A) (b) 连 接 件 改 进 型 ( 试 件 B) (c) 下 翼 缘 改 进 型 ( 试 件 C)<br />

图 4 三 个 试 件 的 制 作<br />

-230-


试 验 采 用 跨 中 两 点 对 称 加 载 直 至 试 件 破 坏 , 三 个 试 件 都 表 现 为 延 性 受 弯 破 坏 , 破 坏 时 上 翼 缘 混<br />

凝 土 压 溃 , 下 翼 缘 钢 筋 或 钢 板 屈 服 , 如 图 5 所 示 , 可 见 采 用 栓 钉 连 接 件 可 以 很 好 地 保 证 波 形 钢 腹 板<br />

和 混 凝 土 翼 缘 板 的 共 同 工 作 , 下 翼 缘 改 进 型 波 形 钢 腹 板 组 合 梁 具 有 良 好 的 受 力 性 能 。<br />

图 6 所 示 为 三 个 试 件 的 主 要 试 验 结 果 对 比 。 图 6a 为 荷 载 - 跨 中 位 移 实 测 曲 线 , 三 个 试 件 都 表 现 出<br />

良 好 的 延 性 , 下 翼 缘 改 进 型 波 形 钢 腹 板 组 合 梁 ( 试 件 C) 相 比 其 他 试 件 具 有 更 高 的 使 用 刚 度 和 承 载<br />

能 力 , 但 其 用 钢 量 并 不 比 其 他 两 个 试 件 高 。 图 6b 为 荷 载 - 下 翼 缘 混 凝 土 板 最 大 裂 缝 宽 度 实 测 曲 线 ,<br />

在 相 同 的 荷 载 水 平 下 , 下 翼 缘 改 进 型 波 形 钢 腹 板 组 合 梁 的 最 大 裂 缝 宽 度 远 小 于 其 他 两 个 试 件 , 可 见<br />

通 过 改 进 下 翼 缘 构 造 , 下 翼 缘 钢 板 可 有 效 地 限 制 混 凝 土 板 裂 缝 的 发 展 , 使 改 进 后 的 波 形 钢 腹 板 组 合<br />

梁 具 有 更 好 的 抗 裂 性 能 , 这 对 提 高 结 构 的 耐 久 性 具 有 重 要 意 义 。<br />

(a) 传 统 型 ( 试 件 A) (b) 下 翼 缘 改 进 型 ( 试 件 C)<br />

图 5 试 件 整 体 破 坏 形 态<br />

260<br />

240<br />

220<br />

200<br />

荷 载 / kN<br />

180<br />

160<br />

140<br />

120<br />

传 统 型 A<br />

连 接 件 改 进 型 B<br />

下 翼 缘 改 进 型 C<br />

100<br />

0.10 0.15 0.20 0.25<br />

最 大 裂 缝 宽 度 / mm<br />

(a) 荷 载 - 跨 中 位 移 曲 线<br />

三 、 槽 型 钢 - 混 凝 土 组 合 梁 桥<br />

图 6 主 要 试 验 结 果<br />

(b) 荷 载 - 最 大 裂 缝 宽 度 曲 线<br />

轨 道 交 通 桥 梁 的 结 构 形 式 直 接 关 系 到 轨 道 交 通 的 建 设 费 用 、 城 市 景 观 协 调 、 减 振 降 噪 等 问 题 。<br />

因 此 , 合 理 选 择 桥 梁 结 构 形 式 是 轨 道 交 通 的 一 个 重 要 课 题 。 目 前 , 我 国 已 经 修 建 和 正 在 修 建 的 城 市<br />

轨 道 高 架 桥 一 般 采 用 上 承 式 箱 型 截 面 , 如 图 7a 所 示 , 由 于 城 市 轨 道 高 架 桥 大 多 位 于 城 市 或 城 市 组 团<br />

的 中 心 地 区 , 上 承 式 桥 梁 不 仅 对 城 市 景 观 影 响 较 大 , 对 地 面 人 群 造 成 压 抑 感 , 而 且 会 对 附 近 居 民 产<br />

生 较 大 的 噪 音 干 扰 , 同 时 箱 形 梁 梁 高 及 梁 重 较 大 , 常 采 用 的 现 浇 施 工 方 法 具 有 施 工 成 本 高 和 工 期 长<br />

等 缺 点 , 施 工 时 对 周 围 环 境 干 扰 较 大 。 图 7b 所 示 的 预 应 力 混 凝 土 槽 型 梁 是 为 适 应 城 市 高 架 轨 道 交<br />

通 的 需 要 , 由 法 国 索 菲 图 公 司 首 先 提 出 的 一 种 桥 梁 结 构 型 式 , 其 主 要 特 点 是 将 高 架 桥 结 构 承 载 、 降<br />

噪 、 防 撞 与 运 营 等 功 能 有 机 结 合 成 整 体 , 同 时 达 到 降 低 结 构 总 体 高 度 的 目 的 [4,5] 。 然 而 在 我 国 , 一 方<br />

面 由 于 对 预 应 力 槽 型 梁 的 研 究 成 果 尚 不 系 统 , 多 针 对 具 体 工 程 , 且 较 少 直 接 应 用 于 城 市 轨 道 交 通 ,<br />

距 离 指 导 工 程 实 践 还 有 一 定 差 距 , 研 究 工 作 的 滞 后 制 约 了 这 种 结 构 型 式 的 发 展 和 应 用 ; 另 一 方 面 受<br />

预 应 力 混 凝 土 自 身 特 点 的 限 制 , 特 别 是 过 于 复 杂 的 施 工 工 艺 和 混 凝 土 的 开 裂 问 题 [6,7] , 给 槽 型 梁 的 应<br />

-231-


用 带 来 很 多 障 碍 , 因 此 我 国 自 80 年 代 开 始 试 制 槽 型 梁 以 来 , 始 终 未 能 推 广 应 用 [8] 。<br />

针 对 预 应 力 混 凝 土 槽 型 梁 存 在 的 上 述 问 题 , 笔 者 将 钢 - 混 凝 土 组 合 原 理 应 用 于 槽 型 梁 结 构 中 ,<br />

提 出 了 槽 型 钢 - 混 凝 土 组 合 梁 的 新 结 构 方 案 , 如 图 7c。 槽 型 钢 - 混 凝 土 组 合 梁 的 构 造 形 式 如 图 8 所 示 。<br />

首 先 加 工 制 作 U 型 截 面 的 钢 梁 , 钢 梁 安 装 就 位 后 现 场 浇 筑 混 凝 土 , 钢 板 与 混 凝 土 通 过 抗 剪 连 接 件 组<br />

合 成 整 体 共 同 工 作 , 可 以 充 分 发 挥 钢 材 抗 拉 、 混 凝 土 抗 压 性 能 好 的 材 料 特 点 。 通 过 对 钢 材 与 混 凝 土<br />

两 种 材 料 的 有 效 组 合 , 可 以 在 保 留 传 统 槽 型 梁 结 构 优 点 的 基 础 上 , 大 大 改 善 其 施 工 性 能 和 使 用 性 能 ,<br />

明 显 提 高 结 构 的 综 合 经 济 效 益 [9,10] 。<br />

(a) 预 应 力 混 凝 土 箱 梁 (b) 预 应 力 混 凝 土 槽 型 梁 (c) 槽 型 钢 - 混 凝 土 组 合 梁<br />

图 7 用 于 轨 道 交 通 的 桥 梁 结 构 形 式<br />

槽 形 钢 - 混 凝 土 组 合 梁 除 具 有 预 应 力 混 凝 土 槽 形 梁 的 一 些 优 点 外 , 如 降 低 轨 道 标 高 、 降 低 噪 声<br />

污 染 等 , 还 具 有 以 下 优 点 :(1) 钢 结 构 部 分 为 工 厂 制 作 , 易 于 保 证 质 量 ; 混 凝 土 直 接 利 用 钢 板 进 行<br />

浇 筑 , 可 以 无 模 板 或 使 用 较 少 模 板 进 行 施 工 , 易 于 控 制 质 量 ;(2) 钢 结 构 部 分 重 量 较 轻 , 给 安 装 就<br />

位 带 来 很 大 便 利 ;(3) 结 构 受 拉 区 外 包 钢 板 , 避 免 混 凝 土 裂 缝 暴 露 , 便 于 维 护 , 同 时 外 包 钢 板 能 在<br />

一 定 程 度 上 缓 解 目 前 频 频 发 生 的 超 高 车 辆 撞 击 桥 梁 而 使 桥 梁 受 损 所 产 生 的 严 重 后 果 , 因 此 结 构 的 安<br />

全 性 和 耐 久 性 较 好 ;(4) 构 造 简 单 。 对 于 承 受 轨 道 及 列 车 荷 载 而 横 向 受 弯 的 底 板 , 下 层 钢 板 在 横 向<br />

可 以 充 分 发 挥 抗 拉 作 用 , 避 免 板 底 纵 向 开 裂 ; 作 为 纵 向 主 梁 的 下 翼 缘 , 底 层 钢 板 又 可 以 在 纵 向 充 分<br />

发 挥 抗 拉 作 用 , 因 此 相 对 于 预 应 力 混 凝 土 结 构 可 大 大 简 化 构 造 , 减 少 钢 筋 绑 扎 、 焊 接 以 及 多 向 预 应<br />

力 张 拉 的 困 难 。<br />

图 8 槽 型 钢 - 混 凝 土 组 合 梁 构 造 图<br />

-232-


为 了 深 入 了 解 槽 型 钢 - 混 凝 土 组 合 梁 这 种 新 型 结 构 的 受 力 性 能 , 开 展 了 试 验 研 究 。 图 9 所 示 为<br />

槽 型 组 合 梁 试 件 的 制 作 过 程 , 主 要 分 三 个 步 骤 , 先 制 作 钢 结 构 , 焊 接 栓 钉 , 然 后 绑 扎 钢 筋 , 最 后 浇<br />

注 混 凝 土 。 图 10 所 示 为 槽 型 组 合 梁 的 最 终 破 坏 形 态 , 槽 型 组 合 梁 表 现 出 延 性 弯 曲 破 坏 形 态 , 上 翼<br />

缘 混 凝 土 压 溃 , 下 翼 缘 钢 板 受 拉 屈 服 , 侧 面 钢 腹 板 略 有 鼓 屈 。<br />

(a) 制 作 钢 结 构 , 焊 接 栓 钉 (b) 制 作 钢 筋 网 (c) 浇 注 混 凝 土<br />

图 9 槽 型 组 合 梁 试 件 制 作 过 程<br />

图 10 槽 型 组 合 梁 的 破 坏 形 态<br />

在 试 验 研 究 的 基 础 上 , 笔 者 发 展 了 适 用 于 槽 型 组 合 梁 全 过 程 受 力 性 能 模 拟 的 三 种 理 论 分 析 模<br />

型 , 包 括 条 带 模 型 、 壳 有 限 元 模 型 ( 图 11a) 和 实 体 有 限 元 模 型 ( 图 11b), 其 模 拟 结 果 和 试 验 结 果<br />

的 对 比 情 况 如 图 11c 所 示 , 可 见 这 些 理 论 模 型 均 有 较 高 的 精 度 , 实 体 模 型 承 载 力 计 算 结 果 略 有 偏 大 。<br />

利 用 这 些 理 论 模 型 , 笔 者 重 点 研 究 了 槽 型 组 合 梁 的 截 面 优 化 和 选 型 [9] 、 剪 力 滞 后 效 应 和 空 间 受 力 性<br />

能 、 组 合 板 稳 定 承 载 力 及 栓 钉 设 计 方 法 等 [11-13] , 为 槽 型 组 合 梁 的 应 用 和 推 广 提 供 了 理 论 基 础 。<br />

四 、 钢 - 混 凝 土 组 合 刚 构 桥<br />

图 11 槽 型 钢 - 混 凝 土 组 合 梁 理 论 分 析 模 型<br />

刚 构 桥 具 有 桥 下 净 空 大 、 节 省 支 座 、 造 型 美 观 、 桥 面 平 顺 性 好 和 抗 震 性 能 高 等 优 点 , 是 适 用 于<br />

-233-


跨 越 深 阔 河 谷 的 优 良 桥 梁 结 构 形 式 。 刚 构 桥 将 墩 - 梁 固 结 后 , 在 竖 向 荷 载 作 用 下 , 主 梁 端 部 将 产 生<br />

负 弯 矩 , 从 而 可 减 小 跨 中 正 弯 矩 , 梁 跨 中 截 面 相 比 其 它 梁 式 桥 型 可 相 应 减 小 [14] 。 相 对 于 梁 式 桥 , 刚<br />

构 桥 可 以 有 效 避 免 强 震 下 发 生 落 梁 破 坏 。 此 外 , 墩 - 梁 固 结 后 如 采 用 悬 臂 施 工 , 不 仅 可 省 略 一 般 梁<br />

式 桥 施 工 时 的 临 时 锚 固 措 施 , 而 且 结 构 体 系 在 施 工 过 程 中 的 受 力 状 态 与 成 桥 后 的 受 力 状 态 也 基 本 一<br />

致 [15] 。 但 是 , 刚 构 桥 的 超 静 定 次 数 高 , 因 此 对 基 础 变 位 、 常 年 温 差 和 日 照 温 差 等 较 为 敏 感 。<br />

将 钢 - 混 凝 土 组 合 结 构 与 刚 构 桥 二 者 的 优 势 进 行 结 合 所 形 成 的 钢 - 混 凝 土 组 合 刚 构 桥 , 在 受 力 性<br />

能 、 施 工 性 能 、 综 合 造 价 以 及 耐 久 性 等 方 面 具 有 很 多 优 势 。 组 合 刚 构 桥 是 指 在 混 凝 土 刚 构 桥 的 基 础<br />

上 , 将 钢 - 混 凝 土 组 合 梁 与 混 凝 土 桥 墩 或 组 合 结 构 桥 墩 相 固 结 所 形 成 的 桥 梁 结 构 形 式 , 如 图 12 所 示 。<br />

组 合 刚 构 桥 除 具 备 刚 构 桥 原 有 的 优 势 之 外 , 通 过 与 组 合 技 术 的 结 合 , 还 具 有 以 下 优 点 :(1) 结<br />

构 自 重 轻 、 刚 度 大 、 跨 越 能 力 强 ;(2) 延 性 高 , 抗 震 性 能 好 ;(3) 工 厂 化 制 造 , 环 境 影 响 小 ;(4) 良<br />

好 的 抗 裂 性 能 ;(5) 具 有 较 好 的 长 期 受 力 性 能 ;(6) 施 工 方 便 快 捷 。<br />

图 12 钢 - 混 凝 土 组 合 刚 构 桥<br />

主 梁 与 桥 墩 固 结 所 形 成 的 组 合 刚 构 桥 , 具 有 不 设 置 支 座 的 优 点 , 但 是 如 何 保 证 桥 面 系 承 担 的 荷<br />

载 能 够 有 效 地 传 给 桥 墩 , 则 是 设 计 施 工 的 关 键 问 题 。 组 合 刚 构 桥 墩 - 梁 结 合 部 的 设 计 除 保 证 有 足 够<br />

的 强 度 和 刚 度 外 , 还 应 具 有 良 好 的 耐 久 性 和 延 性 以 抵 抗 温 度 和 地 震 作 用 的 影 响 [16] 。<br />

目 前 已 有 的 组 合 刚 构 桥 墩 - 梁 结 合 部 构 造 有 许 多 种 , 如 图 13 所 示 , 这 些 构 造 形 式 中 应 用 最 多 的<br />

是 钢 筋 锚 固 式 结 合 部 ( 图 13a) 以 及 钢 柱 式 结 合 部 ( 图 13b)。 钢 筋 锚 固 式 结 合 部 依 靠 墩 中 主 筋 锚 固<br />

在 主 梁 内 传 力 , 需 要 主 筋 较 长 的 锚 固 长 度 , 施 工 过 程 复 杂 且 节 点 核 心 区 混 凝 土 的 浇 筑 质 量 难 以 保 证 ;<br />

钢 柱 式 结 合 部 在 钢 主 梁 上 焊 接 钢 梁 牛 腿 , 将 钢 牛 腿 埋 入 墩 中 形 成 整 体 受 力 , 其 整 体 工 作 性 能 好 , 但<br />

施 工 较 为 复 杂 , 节 点 区 钢 筋 、 钢 骨 交 错 , 影 响 了 节 点 核 心 区 混 凝 土 的 浇 筑 质 量 。<br />

(a) 钢 筋 锚 固 式 结 合 部 (b) 钢 柱 式 结 合 部 (c) 锚 杆 式 结 合 部<br />

(d) 局 部 承 压 式 结 合 部 (e) 钢 板 锚 固 式 结 合 部 (f) 外 置 式 结 合 部<br />

图 13 已 有 的 墩 - 梁 结 合 部 构 造<br />

-234-


笔 者 针 对 已 有 结 合 部 构 造 的 不 足 , 提 出 了 一 种 新 型 钢 套 筒 式 墩 - 梁 结 合 部 构 造 , 如 图 14 所 示 。<br />

该 结 合 部 中 的 钢 梁 和 钢 套 筒 在 工 厂 预 制 , 现 场 安 装 于 钢 筋 混 凝 土 墩 上 , 浇 注 结 合 部 混 凝 土 , 为 提 高<br />

负 弯 矩 区 组 合 梁 的 刚 度 和 承 载 力 , 防 止 负 弯 矩 区 钢 梁 底 板 的 局 部 屈 曲 , 可 在 负 弯 矩 区 底 板 上 方 浇 注<br />

混 凝 土 , 最 后 施 工 混 凝 土 桥 面 板 。<br />

钢 套 筒 式 墩 - 梁 结 合 部 构 造 形 式 相 比 现 有 的 结 合 部 构 造 形 式 具 有 以 下 优 点 :(1) 通 过 设 置 钢 套<br />

筒 , 确 保 钢 梁 与 混 凝 土 墩 可 靠 连 接 , 同 时 构 造 简 单 , 施 工 方 便 ;(2) 钢 套 筒 对 节 点 核 心 区 混 凝 土 具<br />

有 约 束 效 应 , 可 提 高 结 合 部 强 度 及 延 性 ;(3) 墩 顶 钢 套 筒 可 在 施 工 阶 段 作 为 浇 筑 混 凝 土 的 模 板 , 施<br />

工 方 便 ;(4) 在 靠 近 结 合 段 的 钢 梁 底 板 上 浇 筑 混 凝 土 , 可 有 效 降 低 箱 梁 底 板 压 应 力 并 防 止 其 局 部 屈<br />

曲 。<br />

图 14 新 型 钢 套 筒 墩 - 梁 结 合 部 构 造<br />

为 了 研 究 传 统 和 新 型 墩 - 梁 结 合 部 的 受 力 性 能 , 进 行 了 钢 筋 锚 固 式 结 合 部 以 及 新 型 钢 套 筒 结 合<br />

部 在 滞 回 往 复 荷 载 作 用 下 的 对 比 模 型 试 验 , 如 图 15 所 示 。 同 时 , 还 开 展 了 组 合 刚 构 桥 组 合 梁 剪 力<br />

滞 效 应 、 负 弯 矩 区 混 凝 土 桥 面 板 防 开 裂 措 施 、 组 合 刚 构 桥 的 长 期 性 能 以 及 全 桥 分 析 和 设 计 优 化 方 法<br />

等 研 究 。<br />

(a) 试 件 制 作 情 况 (b) 钢 筋 锚 固 式 结 合 部 (c) 新 型 钢 套 筒 结 合 部<br />

图 15 组 合 刚 构 桥 墩 - 梁 结 合 部 试 验<br />

五 、 双 重 组 合 作 用 连 续 组 合 梁 桥<br />

在 大 跨 连 续 组 合 梁 桥 中 , 正 弯 矩 区 混 凝 土 板 和 钢 梁 通 过 抗 剪 连 接 件 相 连 , 能 充 分 发 挥 两 者 的 组<br />

合 作 用 , 但 在 支 座 附 近 , 由 于 组 合 梁 承 受 负 弯 矩 , 若 仍 采 用 和 正 弯 矩 区 相 同 的 组 合 方 式 , 就 会 导 致<br />

混 凝 土 板 受 拉 和 钢 梁 下 翼 缘 板 受 压 , 混 凝 土 板 开 裂 和 钢 梁 下 翼 缘 板 失 稳 就 会 成 为 突 出 的 问 题 , 为 解<br />

决 该 问 题 , 往 往 需 要 在 负 弯 矩 区 桥 面 板 布 置 大 量 的 预 应 力 束 , 并 在 负 弯 矩 区 钢 梁 底 板 上 焊 接 加 劲 肋<br />

或 增 加 底 板 厚 度 , 这 种 设 计 施 工 工 艺 复 杂 、 预 应 力 效 率 低 、 对 于 负 弯 矩 截 面 的 刚 度 提 高 十 分 有 限 。<br />

-235-


针 对 负 弯 矩 区 组 合 梁 的 受 力 特 点 , 可 以 通 过 改 变 钢 与 混 凝 土 的 组 合 方 式 来 改 善 负 弯 矩 区 组 合 梁<br />

的 受 力 性 能 。 具 体 做 法 为 :(1) 在 负 弯 矩 区 混 凝 土 板 和 钢 梁 不 组 合 , 钢 与 混 凝 土 之 间 只 设 置 稀 疏 的<br />

构 造 栓 钉 , 这 样 混 凝 土 板 和 钢 梁 能 自 由 变 形 , 从 而 有 效 释 放 混 凝 土 板 中 因 收 缩 徐 变 、 温 度 效 应 以 及<br />

汽 车 荷 载 引 起 的 拉 应 力 , 并 提 高 负 弯 矩 区 桥 面 板 纵 向 预 应 力 的 施 加 效 率 , 改 善 桥 面 系 的 抗 裂 性 能 ;<br />

(2) 在 负 弯 矩 区 附 近 钢 梁 底 板 上 方 浇 注 混 凝 土 , 形 成 钢 - 混 凝 土 组 合 板 , 充 分 发 挥 混 凝 土 材 料 抗 压 性<br />

能 好 的 优 点 , 显 著 改 善 钢 梁 下 翼 缘 受 压 稳 定 性 能 , 经 济 地 实 现 增 大 负 弯 矩 区 结 构 刚 度 的 目 的 。 这 样 ,<br />

在 连 续 梁 桥 的 负 弯 矩 区 形 成 了 一 个 倒 置 的 组 合 截 面 , 这 样 的 组 合 梁 桥 就 称 为 双 重 组 合 作 用 连 续 组 合<br />

梁 桥 , 如 图 16a 所 示 。 双 重 组 合 技 术 还 可 用 于 连 续 组 合 桁 梁 桥 , 主 要 的 区 别 是 在 负 弯 矩 区 下 弦 杆 内<br />

灌 注 混 凝 土 , 如 图 16b 所 示 。<br />

(a) 连 续 组 合 梁 桥<br />

(b) 连 续 组 合 桁 梁 桥<br />

图 16 双 重 组 合 作 用 连 续 组 合 梁 桥<br />

(a) 架 设 临 时 支 撑 和 钢 梁<br />

(b) 浇 注 负 弯 矩 区 底 板 混 凝 土 和 正 弯 矩 区 混 凝 土 桥 面 板 ( 完 全 组 合 )<br />

(c) 拆 除 临 时 支 撑<br />

(d) 浇 注 负 弯 矩 区 混 凝 土 桥 面 板 ( 不 组 合 )<br />

(e) 安 装 桥 面 铺 装<br />

(f) 张 拉 负 弯 矩 区 混 凝 土 桥 面 板 预 应 力<br />

(g) 施 工 预 留 槽 混 凝 土 桥 面 板 及 铺 装 , 成 桥<br />

图 17 双 重 组 合 作 用 连 续 组 合 梁 桥 施 工 工 序<br />

双 重 组 合 作 用 连 续 组 合 梁 桥 只 有 配 合 合 理 的 施 工 工 序 , 才 能 充 分 发 挥 其 综 合 效 益 , 从 而 有 效 改<br />

善 结 构 的 受 力 状 态 、 简 化 施 工 工 艺 、 提 高 材 料 利 用 效 率 、 降 低 结 构 施 工 成 本 和 周 期 等 。 图 17 所 示<br />

为 合 理 的 双 重 组 合 作 用 连 续 组 合 梁 桥 施 工 工 序 , 主 要 分 为 7 个 步 骤 : 首 先 架 设 临 时 支 撑 和 钢 梁 ; 其<br />

次 浇 注 负 弯 矩 区 底 板 混 凝 土 和 正 弯 矩 区 混 凝 土 桥 面 板 , 分 别 形 成 正 弯 矩 组 合 截 面 和 负 弯 矩 组 合 截<br />

-236-


面 , 浇 注 的 混 凝 土 和 相 应 的 钢 结 构 通 过 栓 钉 形 成 完 全 组 合 作 用 ; 然 后 拆 除 临 时 支 撑 , 这 样 正 弯 矩 区<br />

混 凝 土 板 能 预 存 压 应 力 ; 接 着 浇 注 负 弯 矩 区 混 凝 土 板 , 混 凝 土 板 和 钢 梁 不 组 合 , 因 此 收 缩 徐 变 、 温<br />

度 效 应 等 都 不 会 引 起 混 凝 土 板 的 拉 应 力 ; 接 着 安 装 铺 装 再 张 拉 负 弯 矩 区 混 凝 土 桥 面 板 预 应 力 , 这 样<br />

负 弯 矩 区 混 凝 土 板 需 要 预 存 的 压 应 力 只 需 抵 抗 活 载 作 用 引 起 的 拉 应 力 即 可 ; 最 后 施 工 预 留 槽 混 凝 土<br />

和 铺 装 , 成 桥 。<br />

为 了 检 验 双 重 组 合 作 用 对 结 构 性 能 的 影 响 , 笔 者 对 潍 坊 市 东 绕 城 跨 济 青 高 速 立 交 桥<br />

66.04+96+66.04m 三 跨 一 联 变 截 面 钢 - 混 凝 土 连 续 组 合 梁 桥 进 行 现 场 实 测 , 该 桥 采 用 了 双 重 组 合 技<br />

术 。 该 桥 主 梁 为 双 箱 单 室 钢 - 混 凝 土 组 合 箱 梁 结 构 , 钢 箱 梁 采 用 开 口 截 面 形 式 , 该 桥 主 梁 立 面 和 横<br />

截 面 尺 寸 如 图 18 所 示 , 在 负 弯 矩 区 截 面 钢 箱 梁 底 板 上 方 浇 注 了 混 凝 土 , 从 而 形 成 双 重 组 合 作 用 连<br />

续 组 合 梁 。 图 19 所 示 为 立 交 桥 现 场 施 工 、 成 桥 以 及 现 场 加 载 试 验 的 情 况 。<br />

(b) 主 梁 立 面 图<br />

I-I 截 面<br />

II-II 截 面<br />

(b) 主 梁 横 断 面 图<br />

图 18 潍 坊 市 东 绕 城 跨 济 青 高 速 立 交 桥<br />

(a) 施 工 过 程 (b) 成 桥 (c) 加 载 试 验<br />

图 19 立 交 桥 现 场 实 测<br />

-237-


με με<br />

με<br />

l<br />

l<br />

/<br />

/<br />

0<br />

0<br />

M M<br />

L c<br />

L 0<br />

f f<br />

Lc<br />

/ L<br />

0<br />

(a) 工 况 三 挠 度 计 算 和 实 测 结 果 对 比<br />

(b) II-II 截 面 钢 箱 梁 底 板 应 变<br />

图 20 现 场 实 测 和 有 限 元 分 析 结 果<br />

(c) 内 力 和 变 形 随 L c /L 0 的 变 化<br />

图 20a 所 示 为 组 合 梁 挠 度 实 测 结 果 和 有 限 元 分 析 结 果 的 对 比 情 况 , 其 中 FEM1 表 示 考 虑 双 重 组<br />

合 作 用 的 计 算 值 ,FEM2 表 示 不 考 虑 双 重 组 合 作 用 的 计 算 值 , 可 见 , 考 虑 双 重 组 合 作 用 的 计 算 值 和<br />

实 测 结 果 更 吻 合 , 双 重 组 合 作 用 能 有 效 提 高 结 构 刚 度 。 图 20b 所 示 为 负 弯 矩 区 II-II 截 面 钢 箱 梁 底 板<br />

应 变 实 测 和 有 限 元 结 果 对 比 情 况 , 考 虑 双 重 组 合 作 用 的 计 算 值 明 显 小 于 不 考 虑 双 重 组 合 作 用 的 计 算<br />

值 , 而 且 和 实 测 结 果 也 更 为 吻 合 。 图 20c 为 连 续 梁 内 力 和 变 形 随 负 弯 矩 区 底 板 混 凝 土 浇 注 长 度 的 变<br />

化 情 况 , 图 中 M l /M 0 以 及 f l /f 0 分 别 为 考 虑 双 重 组 合 作 用 和 不 考 虑 双 重 组 合 作 用 的 弯 矩 和 挠 度 比 值 ,<br />

可 见 考 虑 双 重 组 合 作 用 后 , 连 续 组 合 梁 桥 支 点 截 面 刚 度 增 加 , 负 弯 矩 增 加 , 跨 中 正 弯 矩 减 小 , 中 跨<br />

挠 度 也 减 小 。 以 上 现 场 实 测 以 及 有 限 元 分 析 结 果 充 分 表 明 , 双 重 组 合 作 用 能 有 效 提 高 连 续 组 合 梁 负<br />

弯 矩 截 面 以 及 结 构 的 刚 度 、 调 整 结 构 的 内 力 、 改 善 结 构 的 受 力 性 能 , 可 以 在 计 算 中 充 分 考 虑 这 些 有<br />

利 因 素 。<br />

六 、 大 跨 斜 拉 桥 组 合 桥 面 系<br />

组 合 梁 斜 拉 桥 是 在 混 凝 土 斜 拉 桥 以 及 钢 斜 拉 桥 基 础 上 发 展 起 来 的 一 种 斜 拉 桥 形 式 , 它 除 具 有 自<br />

重 轻 , 施 工 便 捷 等 优 点 外 , 由 于 其 采 用 廉 价 的 混 凝 土 桥 面 板 代 替 昂 贵 的 正 交 异 性 钢 桥 面 板 , 利 用 混<br />

凝 土 桥 面 板 抵 抗 轴 力 , 能 够 有 效 节 约 用 钢 量 , 且 其 刚 度 及 抗 风 稳 定 性 优 于 钢 斜 拉 桥 。 当 斜 拉 桥 跨 度<br />

在 300m~700m 之 间 时 , 预 应 力 混 凝 土 斜 拉 桥 因 自 重 大 失 去 了 竞 争 优 势 , 而 钢 斜 拉 桥 方 案 造 价 过 于<br />

昂 贵 , 且 正 交 异 性 钢 桥 面 板 疲 劳 性 能 及 铺 装 耐 久 性 问 题 突 出 , 因 此 组 合 梁 斜 拉 桥 成 为 一 种 合 理 的 选<br />

择 。<br />

图 21a、b 是 目 前 工 程 中 常 用 的 大 跨 斜 拉 桥 组 合 桥 面 系 的 构 造 形 式 , 主 梁 可 以 采 用 工 字 型 钢 梁 或<br />

扁 平 箱 形 钢 梁 , 由 于 采 用 实 腹 式 平 钢 板 作 为 主 梁 腹 板 , 钢 梁 对 混 凝 土 桥 面 板 的 收 缩 徐 变 效 应 具 有 约<br />

束 作 用 , 容 易 引 起 桥 面 板 开 裂 , 从 而 降 低 结 构 的 承 载 力 及 刚 度 , 影 响 结 构 的 耐 久 性 。 这 些 问 题 都 给<br />

该 类 结 构 的 推 广 应 用 造 成 了 一 定 的 困 难 。<br />

(a) 工 字 型 主 梁 组 合 桥 面 系<br />

(b) 箱 形 主 梁 组 合 桥 面 系<br />

-238-


(c) 波 形 钢 腹 板 工 字 型 主 梁 组 合 桥 面 系 (d) 波 形 钢 腹 板 箱 型 主 梁 组 合 桥 面 系 (e) 全 波 形 钢 腹 板 工 字 型 梁 组 合 桥 面 系<br />

图 21 大 跨 斜 拉 桥 组 合 桥 面 系 的 构 造 形 式<br />

为 了 改 善 组 合 桥 面 系 的 受 力 性 能 , 笔 者 提 出 几 种 用 于 大 跨 斜 拉 桥 的 新 型 组 合 桥 面 系 构 造 形 式 ,<br />

如 图 21c、d、e 所 示 , 这 些 新 型 组 合 桥 面 系 的 主 要 特 点 是 采 用 波 形 钢 腹 板 作 为 腹 板 。 新 型 组 合 桥 面<br />

系 的 混 凝 土 桥 面 板 受 压 , 下 部 钢 板 受 拉 , 通 过 组 合 作 用 充 分 发 挥 混 凝 土 与 钢 材 的 特 性 , 且 充 分 发 挥<br />

波 形 钢 腹 板 优 越 的 受 力 性 能 及 其 对 混 凝 土 收 缩 徐 变 效 应 和 温 度 效 应 的 释 放 作 用 , 有 效 缓 解 了 混 凝 土<br />

桥 面 板 因 收 缩 徐 变 及 温 度 效 应 引 起 的 混 凝 土 开 裂 现 象 。 相 对 于 斜 拉 桥 传 统 组 合 桥 面 系 而 言 , 斜 拉 桥<br />

波 形 钢 腹 板 组 合 桥 面 系 具 有 承 载 力 高 、 施 工 方 便 、 自 重 轻 、 抗 震 性 能 好 、 结 构 耐 久 性 好 等 优 点 , 是<br />

对 传 统 组 合 桥 面 系 的 重 要 改 进 。<br />

针 对 这 些 新 型 组 合 桥 面 系 , 重 点 开 展 了 剪 力 滞 效 应 及 长 期 性 能 研 究 , 图 22 所 示 为 新 型 组 合 桥<br />

面 系 剪 力 滞 效 应 的 试 验 研 究 情 况 。<br />

图 22 组 合 桥 面 系 剪 力 滞 效 应 的 试 验 研 究<br />

七 、 结 语<br />

钢 - 混 凝 土 组 合 结 构 桥 梁 具 有 上 部 结 构 高 度 小 、 自 重 轻 、 地 震 作 用 小 , 结 构 延 性 好 、 基 础 造 价<br />

低 、 便 于 工 厂 化 生 产 、 现 场 安 装 质 量 高 、 施 工 费 用 低 、 施 工 速 度 快 等 优 点 , 近 年 来 在 我 国 得 到 了 迅<br />

速 的 发 展 。 本 文 介 绍 了 在 组 合 结 构 桥 梁 方 面 完 成 的 几 项 最 新 研 究 和 实 践 工 作 , 内 容 包 括 波 形 钢 腹 板<br />

组 合 梁 桥 、 槽 型 钢 - 混 凝 土 组 合 梁 桥 、 钢 - 混 凝 土 组 合 刚 构 桥 、 双 重 组 合 作 用 连 续 组 合 梁 桥 和 大 跨 斜<br />

拉 桥 组 合 桥 面 系 。 研 究 通 过 对 传 统 结 构 形 式 进 行 改 进 和 发 展 , 进 一 步 发 挥 了 组 合 结 构 桥 梁 的 综 合 优<br />

势 , 本 文 工 作 可 为 组 合 结 构 桥 梁 的 应 用 和 推 广 提 供 参 考 和 借 鉴 。<br />

参 考 文 献<br />

[1] 樊 健 生 , 聂 建 国 . 钢 - 混 凝 土 组 合 桥 梁 研 究 与 应 用 新 进 展 . 建 筑 钢 结 构 进 展 , 2003, 8(5): 35-39.<br />

[2] 项 海 帆 . 世 界 桥 梁 发 展 中 的 主 要 技 术 创 新 . 广 西 交 通 科 技 , 2003, 28(5): 1-7.<br />

[3] 周 一 桥 , 钱 建 漳 , 编 译 . 折 叠 形 钢 腹 板 预 应 力 混 凝 土 箱 梁 . 国 外 桥 梁 , 2000, (2): 46-51.<br />

[4] 贺 恩 怀 . 槽 型 梁 在 城 市 轨 道 交 通 工 程 中 的 应 用 . 铁 道 工 程 学 报 , 2003(2): 13-16.<br />

[5] Shepherd, B. and Gibbens, B. <strong>The</strong> evolution of the concrete channel bridge system and its application to road and rail<br />

bridges. Concrete Structures: the Challenge of Creativity, CEB-FIB Symposium, 2004.<br />

[6] 张 志 刚 . 预 应 力 混 凝 土 槽 型 梁 产 生 裂 缝 的 原 因 浅 析 . 铁 道 建 筑 , 1998(5): 24-26.<br />

-239-


[7] Durham, S. A., Heymsfield, E. and Schemmel, J. J. Structural evaluation of precast concrete channel beams in bridge<br />

superstructures. Transportation Research Record, 2003: 79-87.<br />

[8] 称 文 艳 , 顾 民 杰 . 轨 道 交 通 双 线 槽 型 梁 的 研 究 . 地 铁 与 轻 轨 , 2003(2): 15-19.<br />

[9] 聂 建 国 , 吴 丽 丽 , 樊 健 生 , 吕 坚 锋 . 槽 形 钢 - 混 凝 土 组 合 梁 及 其 应 用 前 景 初 探 . 土 木 工 程 学 报 , 2008, 41(11): 78-85.<br />

[10] 聂 建 国 . 钢 - 混 凝 土 组 合 结 构 原 理 与 实 例 , 北 京 : 科 学 出 版 社 , 2009.<br />

[11] 吴 丽 丽 , 聂 建 国 . 四 边 简 支 钢 - 混 凝 土 组 合 板 的 弹 性 局 部 剪 切 屈 曲 分 析 . 工 程 力 学 , 2010, 27(1): 52-57.<br />

[12] 聂 建 国 , 李 法 雄 . 钢 - 混 凝 土 组 合 板 的 弹 性 弯 曲 与 稳 定 性 分 析 . 工 程 力 学 , 2009, 26(10): 59-66.<br />

[13] 聂 建 国 , 李 法 雄 . 钢 - 混 凝 土 组 合 板 单 向 受 压 稳 定 性 分 析 , 中 国 铁 道 科 学 . 2009, (6): 27-32.<br />

[14] 邬 晓 光 , 邵 新 鹏 , 万 振 江 . 刚 架 桥 , 北 京 : 人 民 交 通 出 版 社 , 2000.<br />

[15] 王 武 勤 . 大 跨 度 桥 梁 施 工 技 术 , 北 京 : 人 民 交 通 出 版 社 , 2000.<br />

[16] 刘 玉 擎 . 组 合 结 构 桥 梁 , 北 京 : 人 民 交 通 出 版 社 , 2005.<br />

-240-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

RATE DEPENDENCE OF ULTRA HIGH TOUGHNESS CEMENTITIOUS<br />

COMPOSITE IN TENSION<br />

S. L. Xu and H. D. Li<br />

Institute of Advanced Engineering Structures and Materials,<br />

Zhejiang <strong>University</strong>, Mainland, China.<br />

ABSTRACT<br />

This paper examines the effect of strain rates on the tensile properties of ultra high toughness cementitious<br />

composite (UHTCC) under direct tension tests with the rates varying in the range of 4×10 -6 s -1 ~ 1×10 -1 s -1 . <strong>The</strong><br />

experimental results showed that UHTCC possesses significant strain hardening and excellent multiple cracking<br />

properties under both static and seismic loading rates. It showed that the ultimate tensile strain lies in the range<br />

of 3.7%~4.1% and is almost immune to the change of strain rates; the ultimate tensile strength and energy<br />

absorption capability tends to continuously increase when the rate exceeds 1×10 -4 s -1 ; energy absorption<br />

capability of UHTCC is about 1000~3000 times that of concrete under seismic loading rate; the average<br />

ultimate crack widths were all well controlled below 0.1mm under all strain rates. <strong>The</strong>se results revealed that<br />

UHTCC is an ideal construction material for earthquake resistant applications. Several fitting formulas on the<br />

relations between different mechanical properties and strain rates were given according to the experimental<br />

results.<br />

KEYWORDS<br />

Ultra high toughness cementitious composite, strain rate, direct tension, strain hardening, multiple cracking.<br />

-241-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

超 1000 米 建 筑 结 构 与 岩 土 工 程<br />

研 究 与 探 讨 初 步<br />

陈 祥 福<br />

( 中 国 建 筑 工 程 总 公 司 )<br />

2010 年 4 月 7 日<br />

提 纲<br />

• 前 言<br />

• 一 . 世 界 最 高 建 筑 进 化 史<br />

• 二 . 中 国 古 代 名 塔<br />

• 三 . 正 在 建 设 的 世 界 最 高 建 筑<br />

• 四 . 世 界 最 高 建 筑 竞 争 地 转 移<br />

• 五 . 世 界 最 高 建 筑 新 建 信 息<br />

• 六 . 中 国 可 建 超 1000 米 建 筑<br />

• 七 . 超 1000 米 建 筑 设 计 最 复 杂 的 问 题<br />

• 八 . 超 1000 米 建 筑 平 面 和 结 构 设 计 方 案 研 究<br />

• 九 . . 超 1000 米 建 筑 岩 土 工 程 探 讨<br />

• 结 语<br />

-242-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

DESIGN OF WIND AND STRUCTURAL HEALTH MONITORING SYSTEM FOR<br />

STONECUTTERS BRIDGE<br />

K.Y. Wong<br />

B.Sc., M.Sc., Ph.D., Eur Ing, CEng, MICE, MHKIE<br />

Bridges & Structures Division, Highways Department,<br />

<strong>The</strong> Government of <strong>Hong</strong> <strong>Kong</strong> Special Administrative Region<br />

E-mail: sebta.bstr@hyd.gov.hk.<br />

ABSTRACT<br />

A wind and structural health monitoring system (WASHMS) is deployed by Highways Department to monitor<br />

the structural performance of Stonecutters Bridge and its environment under the designated performance criteria<br />

at the serviceability limit state. This system at this stage is composed of two major systems, i.e., the structural<br />

health monitoring system (SHMS) and the structural health data management system (SHDMS). <strong>The</strong> former<br />

system is devised to carry out the monitor works on four categories of physical and chemical quantities in<br />

fifteen types of measurands; whereas the latter system is devised to carry out data operation, data storage, data<br />

retrieval and data interfacing for semis-automatic processing and analysis of data. <strong>The</strong> four categories of<br />

physical and chemical quantities considered for monitoring are: (i) environment loads and status monitoring<br />

which includes the measurands of wind and weather, temperatures, seismic and corrosion status (in structural<br />

concrete only), (ii) operation loads monitoring which includes the measurands of highway traffics, ship<br />

impacting and permanent loads, (iii) bridge features monitoring which includes the measurands of static and<br />

dynamic features of Stonecutters Bridge, and (iv) bridge responses monitoring which includes the measurands<br />

of stay forces, tendon forces, displacements, stress/force distribution, fatigue damage and articulation responses.<br />

This paper first briefly outlines the equipment types of the installed hardware system and facilities such as<br />

sensory systems, data acquisition systems, cabling network systems and computer systems in SHMS. It then<br />

describes the functional details of the customized software tools for processing and analysis of the<br />

above-mentioned fifteen types of measurands. <strong>The</strong> customized software tools for SHMS are customized basing<br />

on six aspects of programming consideration, i.e. (i) type of input data, (ii) type of data processing, (iii) type of<br />

data classification, (iv) type of data analysis, (v) type of derived parameters, and (vi) type of plots and/or outputs.<br />

<strong>The</strong> details of these six aspects of programming consideration for each type of measurands are graphically<br />

interpreted. As the deployment of sensory systems has a strong influence on the methods or strategies of data<br />

processing and analysis, the functional description of the customized software tools are therefore incorporated<br />

with the deployment strategies of the sensory systems. Some typical examples of the monitoring reports are also<br />

included to show the output functions and effectiveness of these customized software tools, and the operation<br />

strategy of the SHMS is also briefly outlined. <strong>The</strong> hardware and software details of the SHDMS, which is<br />

devised to carry out the operation and management of the huge amount of data generated from the SHMS, are<br />

also briefly described. Finally the conclusions of structural health monitoring are discussed in three aspects, i.e.<br />

summary of SHMS design process, performance criteria for monitoring and nature and functions of SHMS.<br />

KEYWORDS<br />

Structural health monitoring, environmental and status monitoring, operation loads monitoring, bridge features<br />

monitoring and bridge responses monitoring, Stonecutters Bridge.<br />

1. INTRODUCTION<br />

<strong>The</strong> Stonecutters Bridge is a two-cable plane cable stayed bridge and carries dual-3 highway traffics. It is<br />

currently the world’s second cable-stayed bridge with the main/central span length exceeding 1 km. It has a<br />

9-span configuration of 69.25m+70m+70m+79.75m+1018m+79.75m+70m+70m+69.25m (=1596m). <strong>The</strong> key<br />

structural features of Stonecutters Bridge are: (i) separated and streamlined steel twin-box girders at central span,<br />

(ii) hybrid (steel/concrete) separated box girders at two side span, (iii) single tower-leg, in which upper part is a<br />

hybrid (steel/concrete) structure and lower part is a concrete structure, and (iv) floating deck connections at<br />

towers and rigid deck connections at all side span piers.<br />

-243-


A wind and structural health monitoring system (WASHMS) is deployed by Highways Department to monitor<br />

the structural performance of Stonecutters Bridge and its environment under the designated performance criteria<br />

at serviceability limit state. <strong>The</strong> performance limits at the serviceability limit state are selected as the base-line<br />

of structural health monitoring. This is because the serviceability load combinations specified in most limit state<br />

design codes for bridgeworks are reasonable estimates of combinations of loads likely to cause maintenance<br />

problem or affect the performance of the structural system and its components, and the probability of<br />

exceedance (POE) for the serviceability load combination is approximately 0.528% per year or 63.36% in<br />

120-year return period basing on 120-year design life (i.e, when the return period is equal to the design life, the<br />

POE = 63.36%). Such value of probability of exceedance is normally considered as the standard for load and<br />

load-effect design at serviceability limit state in most bridge design codes and standards such as BS5400 and<br />

AASHTO.<br />

2. SCOPE OF SCB-WASHMS<br />

<strong>The</strong> SCB-WASHMS is the acronym of the Wind And Structural Health Monitoring System for Stonecutters<br />

Bridge. It is ultimately composed of four systems, namely, structural health monitoring system (SHMS) [Ref.<br />

15], structural health evaluation system (SHES) [Ref. 32], structural health rating system (SHRS) [Ref. 31] and<br />

structural health data management system (SHDMS) [Ref. 15]. <strong>The</strong>y are devised to carry out the respective<br />

works of monitoring, evaluation, rating and data management. A flow diagram of the SCB-WASHMS is shown<br />

in Figure 1. <strong>The</strong> SHES and SHRS, which are currently under development, are arranged under separated<br />

contracts other than the SCB-WASHMS, and therefore they will not be presented in this paper.<br />

3. FUNCTIONS AND COMPONENTS OF SHMS<br />

3.1 Functions of SHMS<br />

<strong>The</strong> SHMS is devised to monitor the structural performance and the environment of the Stonecutter Bridge and<br />

its environment under its in-service condition by collection, processing and analysis of the measured data<br />

obtained from its on-structure instrumentation system. <strong>The</strong> structural performance and the environment are<br />

referring to the designated performance criteria at the serviceability limit state (SLS). <strong>The</strong>se SLS performance<br />

criteria are expressed in terms of the following four categories of physical and chemical quantities, i.e., (i)<br />

environment loads and status monitoring which includes the measurands of wind and weather, temperatures,<br />

seismic and corrosion status, (ii) operation loads monitoring which includes the measurands of highway traffics,<br />

ship impacting and permanent loads, (iii) bridge features monitoring which includes the measurands of static<br />

and dynamic features of Stonecutters Bridge, and (iv) bridge responses monitoring which includes the<br />

measurands of stay forces, tendon forces, displacements, stress/force distribution, fatigue damage and<br />

articulation responses. <strong>The</strong> operation flow diagram of SHMS is illustrated in Figure 2.<br />

In SHMS, both direct data acquisition and indirect data acquisition are adopted. <strong>The</strong> former refers to the usage<br />

of the measured data (such as temperatures and strains/stresses) directly in the interpretation of structural<br />

performance such as temperatures variations and strain/stress status in structural components; whereas the latter<br />

refers to the further processing and analysis of measured data (such as displacements, accelerations and loads)<br />

by analytical models or methods in interpretation of structural performance such as global bridge dynamic<br />

features variations, global bridge geometry profiles variation, etc.<br />

3.2 Hardware of SHMS<br />

<strong>The</strong> hardware of SHMS is composed of sensory system, portable data acquisition system, data acquisition<br />

system, cabling network system, data processing and control system, and portable inspection and maintenance<br />

system. <strong>The</strong> schematic hardware layouts for one-dimensional signals and two-dimensional (video) signals are<br />

given in respective Figure 3 and Figure 4. <strong>The</strong> brief functional description of these six modular systems is<br />

outlined in following paragraphs.<br />

3.2.1 Sensory System (SS)<br />

<strong>The</strong> sensory system, which refers to the transducers and their interfacing devices for detecting and collecting the<br />

physical/chemical signals, is the most important part of the SHMS as all hardware and software are devised and<br />

deployed as a results of the selected sensory system. <strong>The</strong> sensory system should be deployed to achieve six<br />

functions, i.e., (i) able to measure bridge responses at both local and global levels, (ii) able to integrate the<br />

measured results with the analyzed/predicted results, (iii) able to acquire data in a consistence and retrievable<br />

-244-


manner for statistical analysis, (iv) able to use contemporary commercially available sensors, (v) able to carry<br />

out additional measurements by removable/portable sensors, and (vi) able to carry out cross-calibration of<br />

different sensors at same (key) locations. <strong>The</strong> criteria for sensor selection should be based on: (i) operational<br />

bandwidth, (ii) magnitude and frequency response within the operational bandwidth, (iii) accuracy and<br />

resolution, (iv) type of input or excitation energy, (v) type of output energy, (vi) output signal type, (vii) physical<br />

dimensions, weight and materials, (viii) further signal conditioning requirements, (ix) operating environment<br />

conditions, (x) installation constricts, (xi) special contractual requirements on component being monitored, and<br />

(xii) costs.<br />

A total number of 1781 sensors (of which, 400 nos. of them has been used since the construction stage) in<br />

fifteen types are deployed to monitor the structural performance of Stonecutters Bridge and its environment.<br />

<strong>The</strong>se 1781 sensors are classified into 15 types and they are: anemometers, barometers, rainfall gauges,<br />

temperature sensors, hygrometers, corrosion cells, accelerometers, dynamic strain gauges, static strain gauges,<br />

global positioning systems, tiltmeters, dynamic weigh-in-motion stations, video cameras, articulation sensors<br />

and tensio-magnetic gauges. <strong>The</strong> layout of sensory systems in Stonecutters Bridge is shown in Figure 5. <strong>The</strong><br />

details and locations of sensors for monitoring are tabulated in Tables 1 to 4. <strong>The</strong> arrangement and/or<br />

deployment of different types of sensors are made reference to their intrinsic characteristics of measurements<br />

and requirements of data processing, analysis and /or derivation. Detailed discussions are given in Paragraph<br />

3.3.2.<br />

3.2.2 Portable Data Acquisition System (PDAS)<br />

<strong>The</strong> portable data acquisition system is composed of two portable PC-based data logging systems. <strong>The</strong> first (or<br />

PDAS-1) is used for the collection, processing and analysis of time-series acceleration and dynamic strain data<br />

received from the portable servo-type accelerometers (a total of 48 nos.) installed on stay cables during vibration<br />

measurements and the dynamic strain gauges (a total of 400 nos. – which had been used for construction<br />

monitoring during the bridge construction stage and are deployed for re-usage in the bridge operation stage)<br />

installed at ten locations inside five steel deck segments at main span. <strong>The</strong> PDAS-1, which is a 24-bit data<br />

logging system with 48 nos. of simultaneous sampling data channels (at 200 kHz per data channel), is equipped<br />

with Dewetron products, i.e. Dewe-5000 Industrial PC as the PCI-Controller plus Dewesoft 6 Professional<br />

Edition as the application software for data triggering and management, ORION-1624-S1 as the<br />

analogue-to-digital converter for data storage and transmission, DAQP-LV-S1 as the signal conditioner for<br />

acceleration data filtering and amplification, DAQP-Bridge-A-S1 as the signal conditioner for dynamic strain<br />

data filtering and amplification, and associated facilities such strain gauge adapter, rack, power cables, signal<br />

cables, batteries and battery charger. Dewetron products are selected because they had comparative better<br />

performance than other products (in terms of 24-bit simultaneous data acquisition capability and more compact<br />

in design – 16 nos. of simultaneous sampling data channels per card-device) during the system installation<br />

design stage (~2005/2006).<br />

<strong>The</strong> second (or PDAS-2) is used for the collection and analysis of corrosion data received from corrosion cells<br />

(a total of 33 nos.) installed in concrete tower-bases, pier-base, and concrete cross-girder in side-span. It is<br />

composed of two identical portable corrosion monitoring data loggers, and each logger is equipped with Camur<br />

II System for the measurement of open circuit potential, macrocell current, concrete resistivity, concrete<br />

temperature and concrete relative humidity and Gamry G300 with DC105 Corrosion Techniques Software for<br />

linear polarization resistance measurement.<br />

3.2.3 Data Acquisition System (DAS)<br />

<strong>The</strong> data acquisition system is composed of eight types, namely, (i) DAS for Static and Dynamic Signals, (ii)<br />

DAS for DWIMS (dynamic weigh-motion stations) Signals, (iii) DAS for GPS Signals, (iv) DAS for Fiber<br />

Optic Signals, (v) DAS for Vibrating-Wire Strain Gauges, (vi) DAS for Tensio Magnetic Gauges, (vii) DAS for<br />

corrosion signals, and (viii) DAS for Video Signals (or the video signal converters). Installation design is<br />

required for the first type only. As the remaining types of DAS are proprietary DAS and are supplied directly<br />

from the sensor-suppliers, the installation design is therefore not required, and only the DAS for Static and<br />

Dynamic Signals is described here.<br />

<strong>The</strong> first DAS is composed of 8 nos. data acquisition units. Of which, 4 nos. are installed at upper level of the<br />

cross girders in main span and the remaining 4 nos. are installed inside the tower-shafts). Due to the same<br />

reasons as mentioned in the paragraph for PDAS-1, Dewetron products are also applied to the four major types<br />

of hardware devices in all data acquisition units. <strong>The</strong>y are : (i) PCI-Controller – Dewe-808 Industrial PC, (ii)<br />

-245-


Data Triggering and Management Software – Dewesoft 6 Professional Edition, (iii) Analogue-to-Digital (A/D)<br />

Converter – ORION-1624-S1, and (iv) Signal Conditioner – DAQP-LV-S1 for all accelerometer and tiltmeter<br />

data, DAQP- Bridge-A-S1 for all dynamic strain gauge data, DAQP-LV-S2 for propeller type anemometer data,<br />

and PAD-TH8-P for all platinum resistance thermometry type temperature sensor data. Other sensory data such<br />

as ultrasonic anemometers, barometers and rainfall gauges are digital data and therefore no A/D converter and<br />

signal conditioner are required.<br />

3.2.4 Cabling Network System (CNS)<br />

<strong>The</strong> cabling network system is composed of two categories, i.e., local cabling network system (LCNS) and<br />

global cabling network system (GCNS). <strong>The</strong> LCNS refers to the cabling network (shield instrumentation<br />

cables – BS 5708 or equivalent) connecting the sensory system (except GPS) and individual DAS. It is used to<br />

transmit the signals received from the sensory system to related DAS. <strong>The</strong> LCNS for GPS, TMU-5 and video<br />

cameras is single mode fibre optic cables (9/125 μm). <strong>The</strong> GCNS refers to the single mode fiber optic cabling<br />

networks (with a transmission capacity of 1Gbps) installed inside deck-sections, tower-shafts and pier-shafts. It<br />

is used to collect and transmit the digitized data from the five types of DAS as mentioned above to the DPCS in<br />

the Structural Health Control Room, as shown in Figure 6, located at the Administration Building (or West<br />

Control Building). It is composed of two separated ring-type backbone network (i.e., GCNS-1 and GCNS-2)<br />

with dual-redundancy for data transmission. <strong>The</strong> GCNS-1 is used for the transmission of all one-dimensional<br />

signals collected from all sensory systems; whereas the GCNS-2 is used solely for the transmission of<br />

two-dimensional (or video signals) collected from the video cameras. Both GCNS-1 and GCNS-2 include<br />

subnets that connect individual DASs and VSCs to the two backbone networks, i.e. GCNS-1 and GCNS-2.<br />

3.2.5 Data Processing and Control System (DPCS)<br />

<strong>The</strong> data processing and control system is composed of 2 sets of computers, i.e. DPSC1 and DPCS-2. <strong>The</strong><br />

former is for the control and operation of the one-dimensional data collected through the GCNS-1; whereas the<br />

latter is for the control and operation of the two-dimensional data collected through the GCNS-2. Each set of<br />

computers is composed of one server and two workstations. <strong>The</strong> server is 64-bit Itanium server equipped with<br />

two dual-core processors, 128 GB RAM of main memory and 2300 GB hard disk storage system; whereas the<br />

workstation is 64-bit Xeon console equipped with a quad-core processor, 64 GB RAM of main memory, 1500<br />

GB hard disk storage system and a high resolution graphical system. Since the server cannot support high<br />

resolution graphics, the workstations equipped with high resolution graphics are therefore used for analysis and<br />

display of the measured results. <strong>The</strong> DPCS is devised to carry out four basic functions: (i) control and display of<br />

system operation, (ii) monitoring and display of bridge operation including the early warning of parameters<br />

exceeding the designated performance limit (or at serviceability limit state), (iii) post-processing and analysis of<br />

all measured data, and (iv) dual redundancy for control and display of system and bridge operation. <strong>The</strong>se two<br />

sets of computers are normally referred to as the “SHMS-Server”.<br />

3.2.6 Portable Inspection and Maintenance System (PIMS)<br />

<strong>The</strong> portable inspection and maintenance system is composed of two Lenovo Thinkpad W700ds Mobile<br />

Workstations and associated maintenance tool-box (including soldering and de-soldering tools) for carrying out<br />

the inspection and maintenance works on the sensory system, data acquisition system, cabling networks. Each<br />

mobile workstation is equipped with customized software tools for detection and rectification of faulty<br />

equipment and facilities.<br />

3.3 Software of SHMS<br />

<strong>The</strong> software of SHMS is composed of the three major software systems, i.e., (i) the software system for<br />

instrumentation system control and display (customized by LabVIEW software tools), (ii) the software system<br />

for bridge monitoring and display (customized by MATALB software tools), and (iii) the software system for<br />

data interfacing and analysis execution (customized by JAVA software tools). <strong>The</strong> first and second software<br />

systems are installed and operated in respective DPCS-severs and DPCS-workstations; whereas the third<br />

software system is installed and operated in SHDMS-server. <strong>The</strong> flow diagram of these three major software<br />

systems is shown in Figure 7. <strong>The</strong> key issues regarding the customization of these three major software systems<br />

are discussed in the following Paragraphs 3.3.1, 3.3.2 and 3.3.3.<br />

-246-


3.3.1 Software System for Instrumentation System Control and Display<br />

<strong>The</strong> software system for instrumentation system control and display is deployed in the DPCS for carrying out<br />

the operation control and operation display of the instrumentation system. It is devised to carry out two major<br />

functions; (i) to control the operation of all fixed data acquisition systems regarding data acquisition, signal<br />

conditioning, data pre-processing, data temporary storage and data transmissions (from individual data<br />

acquisition system to DPCS and from DPCS to SHDMS) and (ii) to display the operation status of all sensors,<br />

the key components in the fixed DAS-Static & Dynamic or the eight data acquisition units (such as the<br />

PCI-controller, the signal conditioner, the A/D converter, the serial interfacing device, the hard disk storage<br />

status, the network interfacing device, the air conditioner, etc.), the key components in other fixed data<br />

acquisition systems which are not connected to the DAS-Static & Dynamic (such as DAS-DWIMS, DAS-GPS,<br />

DAS-TMU-5 and DAS-Video – as shown in Figures 3 and 4) and all the key components in cabling network<br />

systems such as the media access unit and the input/output fiber optic repeater, etc. <strong>The</strong> required items for<br />

control and display of instrumentation system are shown in Figure 8. Figure 9 illustrates a typical display form<br />

of the software system for instrumentation system.<br />

3.3.2 Software System for Bridge Monitoring and Display<br />

<strong>The</strong> software system for bridge monitoring and display is deployed in the DPCS for carrying out the operation<br />

monitoring and operation display of the bridge structural system and its environment. It is devised to have the<br />

capabilities in processing and analysis of the afore-mentioned fifteen types of measurands and to generate the<br />

corresponding monitoring reports, i.e., (i) the wind and weather monitoring report, (ii) the temperature<br />

monitoring report, (iii) the seismic monitoring report, (iv) the corrosion status monitoring report, (v) the<br />

highway traffic monitoring report, (vi) ship impacting monitoring report, (vii) the permanent load monitoring<br />

report, (viii) the static bridge features monitoring report, (ix) the dynamic bridge features monitoring report, (x)<br />

the stay cables monitoring report, (xi) the tendons monitoring report, (xii) the displacements monitoring report,<br />

(xiii) the stress/force distribution monitoring report, (xiv) the fatigue damage monitoring report, and (xv) the<br />

articulation monitoring report. All software tools for processing, analysis and reports generation are customized<br />

in accordance with the six aspects of programming consideration, i.e. type of input data, type of data processing,<br />

type of data classification, type of data analysis, type of derived parameters, and type of plots and/or outputs. As<br />

the deployment of sensory systems has a strong influence on the methods or strategies of data processing and<br />

analysis, the functional description of the customized software tools are therefore incorporated with the<br />

deployment strategies of sensory systems. <strong>The</strong> following paragraphs (A to O) therefore outline the strategies<br />

and/or methods of sensors arrangement and software tools customization.<br />

A. Customized Software Tools for Wind and Weather Monitoring [Ref. 2, 3, 5, 11, 12, 13, 14, 15, 17, 20,<br />

and 31]<br />

Stonecutters Bridge is identified as a wind-sensitive structure because over twenty numbers of its natural<br />

frequencies are falling within the normal wind frequency range of 0-1 Hz (approximately), as shown in Figure<br />

10. Wind loads and load-effects monitoring is therefore an essential part of the SHMS. <strong>The</strong> wind responses of<br />

the bridge are composed of two parts, namely the along-wind responses and the across-wind responses. <strong>The</strong><br />

former is predominated by steady wind components; whereas the latter is predominated by the fluctuating wind<br />

components. <strong>The</strong> steady wind is normally monitored by measuring the mean wind characteristics of wind speeds<br />

and directions. <strong>The</strong> fluctuating wind is monitoring by measuring/deriving the turbulent wind characteristics of<br />

wind turbulence intensities, wind turbulence length scales, wind turbulence auto spectra, wind turbulence<br />

co-spectra and wind turbulence coherence spectra. <strong>The</strong> wind load-effects are monitoring by correlation analysis<br />

of wind speeds and responses. As various types of aeroelastic instabilities such as vortex shedding, galloping<br />

and fluttering have been designed not to occur or suppressed to be occurred at higher wind speeds exceeding the<br />

performance requirements at ultimate limit state, the wind load-effects monitoring are therefore referring to the<br />

monitoring of the steady wind responses and the buffeting responses at the serviceability limit state.<br />

For wind speeds and wind directions measurements, three types of anemometers are used, i.e. 3D ultrasonic type<br />

anemometers (12 nos. which are digital sensors – Ultrasonic Grill R3-50, each with a sampling rate of 50 Hz),<br />

2D ultrasonic type anemometers (10 nos. which are digital sensors – Ultrasonic Enercorp 2D, each with a<br />

sampling rate of 10 Hz) and 2D propeller type anemometers (2 nos. which are analogue sensors – RM Young<br />

05106, and each of them is configured with a sampling rate of 50 Hz). <strong>The</strong> first type is deployed for the<br />

monitoring of incidence and horizontal winds speeds below 45 m/s (which is the maximum wind speed to be<br />

measured by the current 3D ultrasonic anemometers) at deck-levels and at tower-tops and for the direct<br />

-247-


correlation plots of wind vs acceleration and/or dynamic strain results. <strong>The</strong> second type is deployed for the<br />

monitoring of horizontal winds above 45 m/s and for the correlation analysis of wind vs displacement and/or<br />

static strain results. <strong>The</strong> third type is deployed for the monitoring/calibration of the horizontal wind directions<br />

determined by the ultrasonic anemometers. <strong>The</strong> first and second types are also deployed for the monitoring of<br />

the wind turbulence coherence characteristics by installing the anemometers around 20~30 meters apart from<br />

each other (longitudinally – or along the bridge-deck alignment). <strong>The</strong> barometers (RM Young 61202) and<br />

rainfall gauges (Casella 103800D) are deployed for the calibration of and/or making references to the wind<br />

speeds taken by the anemometers, in particular under the condition of extreme heavy rains.<br />

<strong>The</strong> functions of the software tools for wind and weather monitoring are : (i) to derive the relevant wind and<br />

weather monitoring parameters such as wind turbulence characteristics, the frequency response functions and<br />

the correlation plots of wind vs responses, and (ii) to output the measured/derived results in the standardized<br />

format of four –in-one format, i.e. 3 graphs and 1 table on one page. <strong>The</strong> 3 graphs are location plan of bridge<br />

site, location of sensors in bridge, graphical report of monitoring results, and a legend-table to monitoring results,<br />

as shown in Figures 32-39 and 43-45. <strong>The</strong> flow diagram of the software tools for wind and weather monitoring<br />

is shown in Figure 11. In Figure 11, the fundamental wind data analysis should be carried out first before the<br />

execution of other types of analysis. <strong>The</strong> details of this fundamental wind data analysis are tabulated in Table 5,<br />

as an example of fundamental wind data analysis for wind data collected from 3D ultrasonic type anemometers.<br />

B. Customized Software Tools for Temperature Monitoring [Ref. 15, 17 and 24]<br />

<strong>The</strong> temperature monitoring in Stonecutters Bridge is devised to monitor the temperature in five types of<br />

materials, i.e., structural steel sections (steel deck sections and steel tower skin sections), structural concrete<br />

sections (concrete deck sections and concrete tower sections), cable steel sections, air (un-shaded, shaded, and<br />

inside section) and asphalt pavement. Due to different installation requirements and measuring materials, the<br />

temperature sensors are classified as seven types, i.e. TMU-1, TMU-2, TMU-3, TMU-4, TMU-5, TMU-6 and<br />

TMU-7. TMU-1 and TMU-2 are both for structural steel temperature measurement, in which the former is bolt<br />

fixed type and the latter is bond mount type. TMU-3 and TMU-4 are both for structural concrete temperature<br />

measurement, in which the former is for tower and the latter is for deck. TMU-5 is for cable steel temperature<br />

measurement. TMU-6 is for air temperature and relative humidity (thermo transmitter type) measurement.<br />

TMU-7 is for asphalt pavement temperature measurement. All temperature sensors are Class A Platinum<br />

Resistance <strong>The</strong>rmometry except TMU-5 which is distributed optical fiber sensors (based on Raman scattering<br />

process).<br />

Two major parameters are required for temperature monitoring of structural steel sections and they are the<br />

effective temperature and the differential temperature. In order to derive these two parameters in the orthotropic<br />

steel deck section, the temperature sensors should be distributed along the top and bottom deck plate-trough<br />

sections and the web plates. <strong>The</strong> computational method for effective and differential temperatures basing on<br />

such arrangement of temperature sensors is shown in Figure 12. In a similar way, the effective temperature and<br />

differential temperatures in structural concrete sections are determined. Differential temperature in steel cable,<br />

tower skin and asphalt pavement is not required (as the respective diameter, thickness and depth are comparative<br />

small) and only the average (or effective) temperature is required. <strong>The</strong> effective temperature is normally used to<br />

estimate the thermal movement of the structural component and to calibrate/eliminate thermal effects for other<br />

aspects of monitoring such as stay forces monitoring. <strong>The</strong> flow diagram of the software tools for temperature<br />

monitoring is shown in Figure 13.<br />

C. Customized Software Tools for Seismic Monitoring [Ref. 10, 15 and 17]<br />

In an earthquake, no actual force is applied to the bridge. Instead, the ground moves back and forth (and/or up<br />

and down) and this movement induces inertial forces that then deform the bridge. It is the displacements in the<br />

bridge, relative to the moving base, that impose deformations on the bridge. Through the elastic properties, these<br />

deformations cause elastic forces to develop in the individual members and connections. Earthquake ground<br />

motions usually are imposed through the use of the ground acceleration records. In Stonecutters Bridge, if an<br />

earthquake with a Modified Mercalli Scale of 5 or more is occurred, the time-series acceleration data at<br />

tower-base will be used to generate the response spectrum of the bridge at a certain level of damping (which is<br />

5% according to Design Memorandum of Stonecutters Bridge). <strong>The</strong> spectrum is obtained by repeated solving<br />

the dynamic equilibrium equations (expressed in terms of bridge’s circular frequency) by Newmark’s method<br />

with the input of the recorded time-series acceleration data (in a time increment of 0.02 seconds) for the bridge<br />

with various circular frequencies and then plotting the peak displacement obtained for that circular frequency (ω)<br />

versus the frequency for which the displacement was obtained. <strong>The</strong> velocity (or pseudo-velocity) response<br />

-248-


spectrum is obtained in an approximate manner by assuming that the displacement response is harmonic and<br />

hence that the velocity at each circular frequency is equal to the frequency time the displacement. <strong>The</strong><br />

acceleration (or pseudo-acceleration) response spectrum is obtained from the displacement response spectrum<br />

by multiplying by the circular frequencies squared. <strong>The</strong> pseudo-acceleration represents the total acceleration in<br />

the bridge while the pseudo-velocity and the displacement are relative quantities. Upon obtaining the response<br />

spectrum diagrams, the seismic forces in the bridge can be obtained by the CQC (complete quadratic<br />

combination) method in combining the effects from different vibration mode shapes considered. <strong>The</strong> flow<br />

diagram of the software tools for seismic monitoring is shown in Figure 14.<br />

D. Customized Software Tools for Corrosion Status Monitoring [Ref. 1, 9, 15, 22, 26, 27 and 30]<br />

Corrosion of reinforcement in structural concrete is normally caused by two major factors, carbonation of the<br />

concrete cover and the penetration of chlorides providing from the marine atmosphere or from chemicals in<br />

contact with concrete. <strong>The</strong> former generally results in uniform corrosion of reinforcement while the latter<br />

induces localized corrosion. Both types of corrosion are of electrochemical nature. <strong>The</strong>refore electrochemical<br />

techniques, or corrosion cells, are used to monitor the electrochemical corrosion activity of the metallic<br />

reinforcement with the four aims: (i) to identify the initiation period or the time until the steel reinforcement<br />

becoming depassivation either by the presence of chloride salts or by carbonation, (ii) to estimate crack<br />

initiation and propagation, and (iii) to estimate the time to corrosion crack of concrete cover.<br />

<strong>The</strong> corrosion cells used for corrosion monitoring are the anode-ladder system from S+R Sensortec GmBH. A<br />

typical corrosion cell is composed of seven components, i.e., one anode ladder assembly, one platinized titanium<br />

cathode, one PT 1000 (Platinum-type) temperature sensor, one manganese dioxide reference electrode, one<br />

reinforcement connection bracket and two relative humidity sensors. A picture of corrosion cell is shown in<br />

Figure 15. <strong>The</strong> corrosion cells are used to measure six types of parameters, i.e., open circuit potential, macrocell<br />

current, concrete resistivity, linear polarization resistance, concrete temperature, and concrete relative humidity<br />

in tower-bases, pier-base and concrete cross-girders in side-span.<br />

<strong>The</strong> first and second parameters provide the information on the time-to-corrosion, i.e., the time required for the<br />

reinforcement changing from the passive state to the active state. <strong>The</strong> critical values of voltage and electric<br />

current for no corrosion, as advised by sensor supplier, are >-150 mV and 100 kΩ-cm. <strong>The</strong> fourth parameter is the only electrochemical parameter<br />

with quantitative ability regarding the corrosion rate of reinforcement. In the measurement of this parameter, a<br />

small perturbation potential is applied to the anode or reinforcement (by means of the counter electrode or the<br />

platinized titanium cathode and the reference electrode or the manganese dioxide reference electrode), and the<br />

resulting current response is measured. This small potential perturbation potential should be applied step-wise,<br />

starting below the free corrosion potential and terminating above the free corrosion potential. <strong>The</strong> linear<br />

polarization resistance is the ratio of the applied potential and the resulting current response and is inversely<br />

related to the uniform corrosion rate. <strong>The</strong> critical value of linear polarization resistance for no corrosion is > 250<br />

kΩ-cm 2 . <strong>The</strong> fifth and sixth parameters are used to provide respective concrete temperature and relative<br />

humidity in concrete for references in concrete resistivity measurement. <strong>The</strong> Camur II System equipped with<br />

Gamry G300 and DC105 Corrosion Techniques Software is deployed for the measurement of these six types of<br />

parameters.<br />

Based on the measurement results of linear polarization resistance and concrete resistivity, the parameter of<br />

corrosion rate is derived basing on the Stern Geary equation. <strong>The</strong> critical value of corrosion rate for no corrosion<br />

is < 0.1 μA per cm 2 . <strong>The</strong> derived corrosion rate will then be used to estimate the reinforcement deterioration<br />

(including crack formation and crack width opening) and the time to corrosion cracking of concrete cover. <strong>The</strong><br />

software tools for the report of corrosion monitoring results are customized to output the measured/derived<br />

results in the standardized format required by the monitoring report and to derive the corrosion rate, the<br />

reinforcement deterioration and the time to corrosion cracking of concrete cover. <strong>The</strong> flow diagram of the<br />

software tools for corrosion status monitoring is shown in Figure 16.<br />

E. Customized Software Tools for Highway Traffics Monitoring [Ref. 4, 15 and 25].<br />

In highway traffics monitoring, two major traffic events of daily flow traffics (which result in cyclic load-effects)<br />

and jammed traffics (which result in maximum static load-effects) are considered. <strong>The</strong> former will induce<br />

-249-


potential durability (fatigue) problems on the steel deck and its connections with stay cables and the latter will<br />

induce potential stability (static) problems on global bridge structural system. Three types of sensors are<br />

deployed to monitor these two types of potential problems, i.e. (i) dynamic weigh-in-motion stations(or DWIMS)<br />

installed in each traffic lane on both side spans, (ii) dynamic strain gauges on the steel deck plates, webs and<br />

deck trough sections in steel deck sections at towers, 1/4, 1/2 and 3/4 main span, and (iii) video cameras at<br />

tower-tops, mid-height of towers and at deck-levels near the DWIMS. <strong>The</strong> fatigue problem is monitored by the<br />

DWIMS and dynamic strain gauges and the stability problem is monitored by video cameras, DWIMS and<br />

dynamic strain gauges.<br />

In daily traffic flow monitoring or fatigue damage monitoring, the DWIMS measures two basic data, i.e.<br />

axle-loads and axle-speeds. <strong>The</strong> software provide by DWIMS would process these two basic data into the<br />

highway loading spectrum diagram on each traffic lane. A software tool, which is customized based on Clauses<br />

8.4 (based on measured traffic data), 11.1, 11.2, 11.3 and 11.4 of BS5400: Part 10: 1980, is used to estimate the<br />

fatigue induced-damage or the fatigue life of a particular structural component basing on the measured<br />

axle-loads and axle-speeds from the DWIMS. This software tool has two options, i.e. simplified vehicular<br />

classification or the 6- vehicle type approach (which is the default in software application) and detailed<br />

vehicular classification or the 20-vehicle type approach. <strong>The</strong> simplified option is devised to provide quick and<br />

concise information for daily monitoring works and the detailed option is devised to produce more accuracy<br />

evaluation for quarterly (or special event) monitoring reports. <strong>The</strong> procedures of fatigue damage monitoring are<br />

illustrated in Tables 6, 7 and 8, which are based on the simplified vehicular classification. Fatigue damage<br />

monitoring by strain data are discussed in Paragraph N below.<br />

In traffic jams monitoring or static stability monitoring, the video cameras records the traffic flow and a<br />

software tool is required to classify the recorded traffic flow into 3 conditions, i.e. Condition 1 refers free-flow<br />

traffic, Condition 2 refers to Jammed Traffics with a jammed length of less than 100m, and Condition 3 refers to<br />

Jammed Traffics with a jammed length of greater than or equal to 100m. Further analysis works are solely<br />

required for Condition 3. Another software tool is required to carry out such analysis works, which include: (i)<br />

determining the highway loading spectrum diagram, (ii) determining the positions and vehicle-weights of goods<br />

vehicles in main span, and (iii) analyzing the load-effects on each pre-defined key locations. <strong>The</strong> flow diagram<br />

of the software tools for highway traffics monitoring is shown in Figure 17.<br />

F. Customized Software Tools for Ship Impacting Monitoring [Ref. 7 and 15]<br />

In Stonecutters Bridge, as the foundation of the towers are located within the reclaimed land at approximately<br />

10m from the cope-line of the seawalls, ship impacting monitoring on towers is therefore required. Ship<br />

impacting monitoring on back-span piers is not required as they are sufficiently remote to be not directly subject<br />

to ship collision loads. <strong>The</strong> approach for ship impacting monitoring is similar to that in seismic monitoring<br />

except Duhamel Integral is used to determine the shock response spectra of relative displacement, relative<br />

velocity and absolute acceleration rather than the Newmark’s method. <strong>The</strong> flow diagram of the software tools<br />

for ship impacting monitoring is shown in Figure 18.<br />

G. Customized Software Tools for Permanent Loads Monitoring [Ref. 6, 15 and 24]<br />

<strong>The</strong> permanent loads monitoring refers to the monitoring of the load-effects on the tower-bases due to dead<br />

loads and super-imposed dead loads. This monitoring work has been started since the beginning of the tower<br />

construction, and the purpose is to monitor the long-term effects of shrinkage and creep on tower deformation.<br />

Static strain gauges (Geokon 4200 – vibrating wire type strain gauge with a measuring range of ±3500με at an<br />

accuracy of 0.1% of full scale and a resolution of 1με) are installed at tower-bases, and in order to have a<br />

sampling rate of 20 Hz, Campbell CR5000 is used as the data acquisition system for all static strain gauges. <strong>The</strong><br />

monitoring is carried out by comparing the estimated total strain with the measured strain at tower-bases. <strong>The</strong><br />

estimated strain is estimated in terms of structural strains (due to the dead loads and super-imposed dead loads),<br />

the creep strains and the shrinkage strains, which are estimated basing on Structures Design Manual and BS5400:<br />

Part 4: 1990. Both estimated and measured should be calibrated to the same reference temperature (or 20°C used<br />

in this comparison) for comparison. <strong>The</strong> flow diagram of the software tools for permanent loads monitoring is<br />

shown in Figure 19.<br />

H. Customized Software Tools for Global Static Features Monitoring [Ref. 15]<br />

<strong>The</strong> global static features monitoring is the measurement of stress and displacement at instrumented points of<br />

GPS and dynamic strain gauges under static and moving vehicular load-trials. This monitoring is scheduled to<br />

-250-


e carried out once per maintenance contract period (which is a 6-year period at current stage). <strong>The</strong> first<br />

measurement, which had been carried out two weeks before the opening of the bridge to public traffics (24<br />

December 2009), was composed of four loading types, i.e. (i) longitudinal line load – a line load of nine 5-axle<br />

trucks (each of ~41 tons) with a loaded length of ~145m in a traffic lane, (ii) uniformly distributed load – a<br />

pattern load of 3 x 3 5-axle trucks in three traffic lanes, (iii) transverse line load – a transverse line load of eight<br />

5-axle trucks distributed across the 6 traffic lanes and 2 hard shoulders, and (iv) moving load – influence line<br />

testing of one or two 5-axle trucks travelling on traffic lanes. <strong>The</strong> first three loading types are static load trials<br />

and form the basic load-patterns and loading-positions used in the simulation (by super-position of appropriate<br />

measured results in each load test) of adverse load-effects under traffic jams with much heavy goods vehicles.<br />

<strong>The</strong> loading centers of these three loading types were set at 1/4, 1/2 and 3/4 of main span. <strong>The</strong> last loading type<br />

was used to calibrate/update the analyzed influence lines at instrumented points. <strong>The</strong>se influence lines will form<br />

the basis to estimate the stress ranges for fatigue damage monitoring. <strong>The</strong> results of this monitoring are also<br />

used to calibrate/update the static computational features of the finite element bridge model. <strong>The</strong><br />

calibrated/updated model will form the basis to compute the stress ranges at non-instrumented locations for<br />

fatigue assessment. <strong>The</strong> flow diagram of the software tools for global static features monitoring is shown in<br />

Figure 20.<br />

I. Customized Software Tools for Global Dynamic Features Monitoring [Ref. 15, 16, 18, 19, 21 and 28]<br />

<strong>The</strong> global dynamic features monitoring, which is a fundamental monitoring requirement in SHMS for<br />

long-span bridge with steel deck structure, is the extraction of the modal frequencies and modal vectors (or<br />

modes shapes) of the global bridge structural system by the approach of operational modal (or ambient vibration)<br />

analysis. <strong>The</strong> extracted global dynamic characteristics will be used to calibrate/update the 3D global finite<br />

element bridge model. <strong>The</strong> calibrated/updated bridge model will be used to estimate the responses at locations<br />

with and/or without sensory systems.<br />

For global dynamic features monitoring, with the consideration of saving equipment cost, both fixed and<br />

removal types of accelerometers are deployed, and they are installed in the two towers (i.e., at tower-top, 2/3<br />

and 1/3 height of tower above deck-level), in the steel deck at main span (i.e., at 1/8 main span spacing) and in<br />

the concrete deck at two side-spans (i.e., at mid-span between two side-span piers). All accelerometers installed<br />

are removal except those accelerometers at the five fixed locations (i.e., two tower-tops and 3 deck-levels at 1/4,<br />

1/2 and 3/4 main span). <strong>The</strong> fixed accelerometers are mainly for instant modal frequencies monitoring. All<br />

accelerometers used are tri-axial servo-type accelerometers, and each is made of three uni-axial QA 700 Q-Flex<br />

accelerometers from Honeywell. <strong>The</strong> total number of accelerometers deployed is 58 nos., of which 48 nos. are<br />

removal (which are also used for stay forces monitoring) and 10 nos. (including the 2 nos. at tower-based for<br />

seismic loads monitoring) are fixed. <strong>The</strong> sampling rate of the accelerometer is devised to be reconfigurable in 3<br />

rates, i.e, 50 Hz (default), 100 Hz (under ambient vibration measurements and typhoons), and 1000 Hz (under<br />

load-trials or special events).<br />

For global dynamic features monitoring, the tri-axial accelerometers inside the deck section should be installed<br />

in accordance with Figure 21. <strong>The</strong> distance between the two tri-axial accelerometers should be kept constant and<br />

the accelerometers (both removal-type and fixed type) installed along the bridge-deck alignment should be kept<br />

as a continuous line at the same horizontal plane and be parallel to the centroid of the bridge-deck section. <strong>The</strong><br />

measured time-series acceleration data are converted to four types of processed time-series acceleration data, i.e.,<br />

vertical deck acceleration, lateral deck acceleration, torsional deck acceleration and longitudinal deck<br />

acceleration, as shown in formulas given in Figure 21. Similarly, the arrangement of accelerometers and<br />

processing of acceleration data for tower-sections are given in Figure 22, which results in three types of<br />

time-series acceleration data, i.e., longitudinal tower acceleration, lateral tower acceleration and torsional tower<br />

acceleration. Customized software tools are devised to extract the seven types of predominating vibration mode<br />

shapes from the above mentioned seven types of processed time-series acceleration data by power spectrum<br />

analysis (for frequencies extraction), linear spectrum analysis and curve fitting analysis (for mode shapes<br />

determination). <strong>The</strong> Fast Fourier Transform (FFT) is used to transform the processed time-series acceleration<br />

data into power spectrum and linear spectrum. <strong>The</strong> FFT, due to its intrinsic deficiency of spectrum leakage for<br />

random signals, is applied together with Hanning Window technique (which is used to minimize the spectrum<br />

leakage induced by FFT) and 50% data overlap processing (which is used to improve the accuracy of spectral<br />

estimation or overcoming the attenuation of useful data by the Hanning Window on two ends of each data<br />

block/frame).<br />

<strong>The</strong> seven types of vibration mode shapes required for extraction are: (i) vertical flexural deck modes, (ii) lateral<br />

flexural deck modes, (iii) longitudinal flexural deck modes, (iv) torsional flexural deck modes, (v) lateral<br />

-251-


flexural tower modes, (vi) longitudinal flexural tower modes and (vii) torsional flexural tower modes. From<br />

these types of vibration mode shapes and their modal participation factors, the major modes and frequencies of<br />

vibration under different types of dynamic loads such as wind and seismic can be identified for<br />

observation/monitoring. <strong>The</strong> customized software tools are also required to compare the measured results with<br />

the analyzed results. <strong>The</strong> flow diagram of the software tools for global dynamic features monitoring is shown in<br />

Figure 23.<br />

J. Customized Software Tools for Stay Forces Monitoring [Ref.15 and 29]<br />

<strong>The</strong> stay forces monitoring refers to the monitoring of the dynamic features of stay cables, hence the tensile<br />

forces and stability against wind-rain-induced vibrations. This monitoring requires the execution of field<br />

vibration measurement works on the stay cables. <strong>The</strong> equipment and facilities required are the PDAS-1, the<br />

removal accelerometers and the associated portable power and signal cables. During measurement, two<br />

accelerometers (one for measurement and the other for control/calibration) should be used and in order to<br />

minimize the measurement errors induced by the stay cable deck anchorage, the first accelerometer should be<br />

fixed as a distance of no less than six meters away from the edge of the stay cable deck anchorage. Customized<br />

software tools are required to: (i) extract the in-plane flexural frequencies, out-of-plane frequencies and<br />

longitudinal frequencies of the stay cable, hence determine the tensile force in stay cable under ambient<br />

vibration measurement, (ii) determine the logarithmic decrement, hence determine the Scruton number basing<br />

on manual-excited forced-free vibration measurement, and (iii) plot the analyzed results and together with the<br />

corresponding analysis/design/as-built values. <strong>The</strong> flow diagram of the software tools for stay forces monitoring<br />

is shown in Figure 24.<br />

K. Customized Software Tools for Tendon Forces Monitoring [Ref. 15]<br />

<strong>The</strong> tendon forces monitoring refers to the monitoring of the forces in the eight tendons per deck-girder-box,<br />

which are designed for enhancing the deck strength at the interfacing regions between steel/concrete in the<br />

side-span. A total of 32 tendons (each of 37 strands) are used and therefore 32 nos. of tensio magnetic gauges<br />

are deployed, i.e. one for each tendon. <strong>The</strong> tensio magnetic gauge used is EM sensor (EM-T-160) with<br />

dimension 400mm length x 220 mm overall diameter for 160 mm strand bundle at a measuring stress range of<br />

1770 MPa and the data logger equipped is PowerStress P-500-08, which is a 8-channel data logger with a<br />

sampling rate of 0.1 Hz. <strong>The</strong> customized software tools for tendon forces monitoring are comparatively simply<br />

as solely time-average statistics analysis is required. <strong>The</strong> flow diagram of the software tools for tendon forces<br />

monitoring is shown in Figure 25.<br />

L. Customized Software Tools for Displacements Monitoring [Ref. 15 and 19]<br />

<strong>The</strong> displacement monitoring of Stonecutters Bridge is based on GPS (rover stations and reference stations) and<br />

tiltmeters. <strong>The</strong> GPS (rover stations – for measurement of bridge motions) are installed at tower-tops and on two<br />

edges of the bridge-deck section at 1/4, 1/2 and 3/4 of main span; whereas the tiltmeters are installed in both<br />

towers at H (or tower-top), 2/3H, 1/3H and 0/3H (or deck-level), where H is the tower-height above the deck<br />

level. <strong>The</strong> tiltmeters are also installed on the internal walls of the upper side-span piers. All tiltmeters are<br />

bi-axial tiltmeters. <strong>The</strong> GPS (reference stations – 2 nos., one for master reference and one for backup reference)<br />

are installed at roof-top of East Portal Building which is around 1 km from Stonecutters Bridge. All time-series<br />

data from different types of sensors are synchronized and time-tagged to the GPS (reference stations) with a<br />

minimum accuracy of ±0.1 second. Each GPS is installed at a stiff GPS post so that the GPS installed at the<br />

post-top should have the same motion as the measurement point at the post-base. In order to minimize the<br />

multi-path effects at ground/base level and/or avoidance of obstruction induced by the passage of high-side<br />

vehicles, the heights of the GPS posts at deck-levels, tower-tops and roof-top are around 4 meters, 3 meters and<br />

1.5 meters respectively. Both GPS (digital sensors) and tiltmeters (analogue sensors) are configured to have the<br />

same sampling rate of 20 Hz. Figure 26 illustrates the arrangement of GPS (rover stations) at deck-levels.<br />

All GPS are Leica GPS, in which the GPS (rover station) is GX1230 GG GNSS receiver with AX1202 GG<br />

GNSS antenna; whereas the GPS (reference station) is GX1230 GG GNSS receiver with AT504 GG<br />

Choke-Ring GNSS antenna. <strong>The</strong> acronym of GNSS (or Global Navigation Satellite System) is a technique that<br />

supports the satellite-signals from both systems of American GPS and Russian GLONASS. <strong>The</strong> horizontal<br />

measurement accuracy of GPS is 3mm±0.5ppm (for static mode) and 10mm±1ppm (for real-time kinematic<br />

mode); whereas the vertical measurement accuracy of GPS is 6m±1ppm (for static mode) and 20mm±1ppm (for<br />

real-time kinematic mode). <strong>The</strong> tiltmeter used is Model 716-2A Wall Mount Tiltmeter from Applied<br />

Geomechanics, which has a resolution of 0.1 μ-radian.<br />

-252-


<strong>The</strong> monitoring work is composed of two approaches, i.e., direct and indirect. <strong>The</strong> direct approach refers to the<br />

direct interpretation of the measured displacement results by statistical analysis and/or correlation analysis with<br />

other measured results such as wind and temperature at instrumented points or locations. <strong>The</strong> indirect approach<br />

refers to the input of the measured displacement results (by the finite element interfacing software tool:<br />

FEMtools) into the pre-built and pre-calibrated 3D mixed finite element bridge model (based on the commercial<br />

finite element software: MSC/NASTRAN) and to re-generate the global bridge deformed/displaced shapes with<br />

the output of corresponding stress conditions at key monitoring or instrumented locations. <strong>The</strong> indirect approach<br />

has the advantages of interpreting the locations or sections without sensors. <strong>The</strong> flow diagram of the software<br />

tools for displacements monitoring is shown in Figure 27.<br />

M. Customized Software Tools for Stress/Force Distribution Monitoring [Ref. 15]<br />

Stress/force distribution monitoring can provide an overall indication of the responses of the instrumented<br />

structural components or locations under live loads. <strong>The</strong> monitoring is divided into two parts, i.e. dynamic<br />

strain/stress distribution monitoring and static strain/force distribution monitoring. Dynamic strain/stress<br />

distribution monitoring applies to longitudinal steel girders (deck sections at towers and 1/2 main span), cross<br />

steel girders (deck section at Tsing Yi Tower, 1/8 and 3/8 main span), and tower steel skins (upper, middle and<br />

lower of steel anchor box)). Static strain/force distribution monitoring applies to upper tower-sections (at 299.60<br />

mPD, 237.00 mPD and 279 mPD), lower tower-sections (at tower-base and ~170 mPD), pier-sections in back<br />

spans (pier-base and pier-top of Piers 1W, 1E, 4W and 4E) and deck-sections in back spans (one at Tsing Yi<br />

back span and two at Stonecutters back span). Dynamic strain is monitored by weldable foil type strain gauges<br />

(Vishay LWK-06-W250B-350 – arranged either in single, in pair or in rosette forms); whereas the static strain is<br />

monitored by vibrating wire type strain gauges (Geokon 4200 – arranged either in single or rosette forms).<br />

<strong>The</strong> flow diagram of the software tools for stress/force distribution monitoring is shown in Figure 28. For strain<br />

gauges with single or parallel arrangement, the stress demand ratio is defined as the ratio of the measured stress<br />

to the design live load stress at serviceability limit state (or σ SLS ). For strain gauges with rosette arrangement,<br />

the measured strains should first be converted to principal stresses, and then to Von-Mises Stress. <strong>The</strong> stress<br />

demand ratio for rosette gauges should be σ Von-Mises /σ SLS .<br />

N. Customized Software Tools for Fatigue Damages Monitoring [Ref. 4 and 15]<br />

Fatigue damage monitoring is primarily carried out by dynamic strain gauges, i.e., weldable foil type strain<br />

gauges (Vishay LWK-06-W250B-350 – arranged either in single or in pair forms) installed in steel deck troughs<br />

(at 1/4 and 1/2 main span), deck-cable anchorages (at 1/4, 1/2 and 3/4 main span), and tower-cable anchorages<br />

(at upper, middle and lower of the anchor box). This fatigue damage monitoring is different from that in<br />

Paragraph E, which deals with highway live loads only, because it includes all types of live loads acting on the<br />

bridge. <strong>The</strong> fatigue monitoring approach developed is basing on Clause 11 of BS5400: Part 10: 1980. <strong>The</strong> flow<br />

diagram of the software tools for fatigue damage monitoring is shown in Figure 29.<br />

O. Customized Software Tools for Articulation Responses Monitoring [Ref. 15]<br />

<strong>The</strong> articulation responses monitoring in Stonecutters Bridge refers to the monitoring the performance of the<br />

longitudinal buffers and lateral bearings at deck-tower interfacing locations. <strong>The</strong> monitoring items are: (i) the<br />

force transfer between the superstructure (steel deck) and the substructure (concrete tower) under lateral<br />

dynamic loads such as wind and seismic, and (ii) the longitudinal movement of the steel deck at main span<br />

under temperature loads. All these sensors, under the Stonecutters Bridge Contract, were supplied and installed<br />

by the buffer and bearing manufacturers.<br />

Each set of lateral bearing sensor (i.e., 4 rosette-type dynamic strain gauges in each lateral bearing) is required<br />

to provide four sets of data, i.e. lateral (Y-axis) bearing displacement, lateral bearing force, vertical bearing<br />

moment (Y-Z plane) and horizontal bearing moment (X-Z plane), where X-axis is the horizontal axis parallel to<br />

the bridge-deck alignment, Y-axis is the horizontal axis perpendicular to the bridge-deck alignment and Z-axis is<br />

the vertical axis perpendicular to the bridge-deck alignment. <strong>The</strong> flow diagram of the software tools for lateral<br />

bearing responses monitoring is shown in Figure 30 (upper part).<br />

Each set of longitudinal buffer sensor (i.e., 2 for pressure, 1 for displacement and 1 for temperature<br />

measurements in one longitudinal buffer) is required to provide longitudinal (X-axis) force, longitudinal<br />

displacement and temperature of hydraulic oil in longitudinal buffer (for calibration of the derived force and<br />

-253-


displacement). <strong>The</strong> flow diagram of the software tools for longitudinal buffer responses monitoring is shown<br />

in Figure 30 (lower part).<br />

3.3.3 Software System for Data Interfacing and Analysis Execution<br />

<strong>The</strong> software system for data interfacing and analysis execution, which is deployed in SHDMS-Server, is<br />

customized for five functions: (i) retrieval of data and information from the rolling storage buffer (or the raw<br />

data database as shown in Figure 45) in the SHDMS-Server to the DPCS-workstations in accordance with user’s<br />

request, (ii) matching the retrieval data and information to each assigned customized software tools (i.e. the<br />

software system for bridge structural system as described in Paragraph 3.3.2), (iii) execution of the requested<br />

processing, analysis and plotting works, (iv) generation and/or storage of relevant monitoring reports. In the<br />

application of this software system, the required input data and information by the user are: the tag-name of<br />

sensors, the time durations for processing and analysis, the types of processing, analysis and plots required. This<br />

approach not only minimizes the time for data retrieval, but also speeds up the works of data processing,<br />

analysis and reporting works. <strong>The</strong> required types of data and tools for data interfacing and analysis execution are<br />

shown in Figure 31.<br />

3.4. Examples of Monitoring Results<br />

In order to limit the scope of presentation, only eight examples of monitoring examples are presented. <strong>The</strong>y are<br />

selected because they have comparative large variations and therefore some of the major features of the<br />

measurands and/or the bridge can be extracted. <strong>The</strong>se eight examples are : (i) wind and weather monitoring –<br />

basic steady wind and turbulence wind monitoring, (ii) highway traffic monitoring – highway loading spectrum,<br />

(iii) static features monitoring – stress influence line determination, (iv) global dynamic features monitoring –<br />

vertical flexural deck frequencies extraction, (v) stay force monitoring – stay cable forces distribution, (vi)<br />

displacements monitoring – indirect approach of re-generating the global bridge deformed geometry, (vii)<br />

fatigue damage monitoring, and (viii) permanent loads monitoring.<br />

As all the customized software systems for SCB-WASHMS till now (May 2011) have not yet been completed<br />

and handed over to Highways Department for operation and in order to avoid any contractual controversy, it is<br />

declared that the following monitoring results are produced based on the measured data in Stonecutters Bridge<br />

with the application of similar software tools used in the SHMS for Tsing Ma Bridge, Kap Shui Mun Bridge and<br />

Ting Kau Bridge [Ref. 8, 34 and 35].<br />

3.4.1 Wind and Weather Monitoring (Figures 32-33)<br />

In this example, two outputs are presented. <strong>The</strong> first output, as shown in Figure 32, is the plot of wind rose<br />

diagram basing on the monthly wind data in February 2011 obtained from the wind sensor: Ane(U3D) which is<br />

located at the Southeast side of mid main span. <strong>The</strong> second output, as shown in Figure 32, is the plots of wind<br />

power spectrum diagram and the integral length scale (based on the direct integration of the auto-covariance<br />

function of the measured wind speeds) of the 2-hour wind data (21 October 2010 at 08:00-10:00 am) from the<br />

wind sensor: Ane(U3D) which is located at Northeast of mid main span. This 2-hour wind data has a 10-minute<br />

mean wind speed of 6.27 m/s with a turbulence intensity of 20.1%. <strong>The</strong> measured peak value of the spectrum is<br />

at 0.0047 Hz; whereas the derived integral length scale is 75.2m, which is much less than the criteria for<br />

buffeting response check.<br />

3.4.2 Highway Traffics Monitoring (Figures 34-35)<br />

Figure 34 illustrates the highway traffic composition monitoring result in February 2011, and Figure 35<br />

illustrates the corresponding estimated the fatigue damages in individual traffic lane. <strong>The</strong> sum of damage<br />

(fatigue) ratios in six traffic lanes is 0.1714, which has been projected to 1-year’s damage ratio, and the<br />

corresponding estimated fatigue life is 700 years (=120 years/0.1714, where 120-year is the design life of the<br />

bridge), which is greater than the Engineer’s design fatigue life of 200~500 years at the fatigue assessment<br />

point.<br />

3.4.3 Global Static Features Monitoring (Figures 36-37)<br />

<strong>The</strong> static features monitoring refers to the determination of the static influence displacement and stress<br />

coefficients. Before, the opening of the Stonecutters Bridge to public traffic, load trials (including influence<br />

lines measurements) had been carried [Ref. 37]. Figure 36 and 37 illustrate the tested and analyzed results of<br />

-254-


displacement influence lines and stress influence line. <strong>The</strong> measured displacement range at GPS instrument<br />

point: GP04 is 115 mm (or 1.4 mm per ton), and the corresponding measured stress range at strain instrument<br />

point: SW60X is 16 MPa (or 0.195 MPa per ton).<br />

3.4.4 Global Dynamic Features Monitoring (Figures 38-39)<br />

Four hours of time-series acceleration data obtained from the two nos. of accelerometers installed inside the<br />

steel deck-section at mid main span are used. <strong>The</strong> time duration for one data frame used is 30 minutes and the<br />

total number of data points per data frame is 180,000 which are computed on a sampling rate of 100 Hz. <strong>The</strong><br />

frequency resolution for extraction is therefore 0.000556 Hz. <strong>The</strong> total number of data frames considered in<br />

average is fifteen and the corresponding wind speed and air temperature references are respective 2.05 m/s<br />

(10-miniute mean) and 16.4~17.5°C. <strong>The</strong> result of vertical flexural deck frequencies (the first 20 nos.) extracted<br />

from the power spectrum is shown in Figures 38. Of these 20 measured frequencies, 16 of them are<br />

corresponding to the analyzed results. <strong>The</strong> comparison between the measured results and analyzed results (by<br />

normal mode analysis of a 3D space frame finite element model – built and analyzed by ANSYS/Multiphysics<br />

[Ref. 38]) is shown in Figure 39. <strong>The</strong> comparison shows good correlation between measured and analyzed<br />

results, and the model is reliable and accuracy for the analysis of highway load-effects.<br />

3.4.5 Stay Forces Monitoring (Figures 40-41)<br />

<strong>The</strong> example of stay forces monitoring is refers to the ambient vibration works which was requested by the<br />

Engineer for checking the as-built stay forces with their computed stay forces under their reference (completion)<br />

conditions [Ref. 36]. A total of 224 stay cables were measured in the seven nights during the period of 4-12<br />

November 2009. Figure 40 illustrates the installation of accelerometer on stay cable and corresponding<br />

procedures of data processing and analysis. Figure 41 illustrates the results of stay forces in the main span of<br />

East Tower – North side obtained from the Engineer, the Contractor and Highways Department.<br />

3.4.6 Displacements Monitoring (Figures 42-43)<br />

<strong>The</strong> displacements monitoring refers to the input of the GPS data into the pre-built finite element bridge model<br />

and re-generate the deflected bridge geometry profile/profiles and stress contours with stresses output at<br />

key/instrumented structural sections/locations. Figure 42 illustrates the derived displacement profile (with<br />

colour contours) of Stonecutters Bridge, and Figure 43 shows the corresponding derived stress in deck trough<br />

sections.<br />

3.4.7 Fatigue Damages Monitoring (Figures 44-45)<br />

<strong>The</strong> example of fatigue damages monitoring is based on the results of dynamic strain gauges installed in the<br />

deck trough section at bottom of north longitudinal box at mid main span. Figure 44 shows the stress demand<br />

ratio plot, which is used to detect whether the measured stress range at deck trough section exceeding the<br />

maximum allowable stress range of 35 MPa (Class F2). If the measured stress range exceeds the maximum<br />

allowable limit, then micro-cracks might be occurred and the Palmgren-Miner rule cannot be applied in fatigue<br />

assessment. As Figure 44 shows that the maximum measured stress range at the deck trough section considered<br />

is 9.99 MPa, which is much less than the allowable stress range of 35 MPa (F2 Class), fatigue assessment can<br />

therefore be proceeded. Figure 45 illustrates the fatigue damage assessment result. <strong>The</strong> estimate fatigue life at<br />

the fatigue assessment point considered is 713 years, which is greater than the design fatigue life of 200~500<br />

years.<br />

3.4.8 Permanent Loads Monitoring (Figures 46-47)<br />

<strong>The</strong> results of permanent loads monitoring at Stonecutters tower-base is shown in Figures 46 and 47. <strong>The</strong> total<br />

measured compressive strains in Locations A and B of Figure 46 are 626 με and 651 με respectively at a<br />

reference concrete temperature of 20°C. <strong>The</strong> strain analysis shows that the percentages of structural strain, creep<br />

strain and shrinkage strain are respective 44%, 20% and 36% (upon the bridge opening to public traffics). <strong>The</strong><br />

strain due to permanent loads is around 44% of the total measured compressive strains or 281 με (=44%<br />

x(626+651)/2). This strain corresponds to 8.486 MPa ≈ 8.392 MPa which is estimated basing on self-weight of<br />

concrete (71250 tons), stainless steel (1400 tons), steel reinforcement (5500 tons), self-weight of steel deck and<br />

stay cables (20100 tons per tower), pavement (2126 tons), street furniture (277 tons) and a tower-area of 119.94<br />

m 2 at 5.8 mPD or the strain instrument level.<br />

-255-


4. OPERATION STRATEGY OF SHMS<br />

<strong>The</strong> operation of SHMS is in three stages, routine monitoring stage, sub-critical monitoring stage and critical<br />

monitoring stage. <strong>The</strong> routine monitoring stage refers to the condition that if the measured data which is less<br />

than or equal to 75% of Q SLS , then no additional monitoring work is required and routine monitoring continues,<br />

where Q SLS is the designated performance criteria at the serviceability limit state. <strong>The</strong> sub-critical monitoring<br />

stage refers to the condition that if the measured data which is following within 76% to 100% of Q SLS , then<br />

components or locations associated with this measurement value will be concentrated for monitoring. <strong>The</strong><br />

critical monitoring stage refers to the condition that if the measured data is exceeding 100% of Q SLS , then<br />

structural evaluation works (to be carried out by the SHES) on the relevant components/locations or whole<br />

structural system will be carried out. <strong>The</strong> 75% of Q SLS is used as the limit or alarm limit for no additional<br />

monitoring works because the measured data implies that there still has a safe factor of 1.3333 (=1/75%) or a<br />

reserved capacity of 25% in the instrumented location or structural component.<br />

In the routine monitoring, a quarterly monitoring report is produced. <strong>The</strong> customized software tools as<br />

mentioned in Paragraphs 3.3.1 – 3.3.3 are used to generate this quarterly report semi-automatically and it, under<br />

the normal condition, includes the monitoring results of all measurands except seismic, ship impacting and<br />

global static features – because they are either under extreme condition or special arrangement.<br />

5. FUNCTIONS AND COMPONENTS OF SHDMS<br />

5.1 System Overview<br />

<strong>The</strong> SHDMS is composed of hardware and software systems. <strong>The</strong> hardware system refers to the SHDMS-Server,<br />

which is located in the Bridge Health Monitoring Room at the West Control (Administration) Building. This<br />

server is a 64-bit Itanium server and is equipped with 4 dual-core processors, 64 GB RAM of main memory, a<br />

120-day or 3000 GB rolling-buffer storage system for raw data, 2000 GB permanent storage system for<br />

processed and analyzed data and two 600 GB DLT tape drivers as permanent storage system for raw data. <strong>The</strong><br />

software system refers to the Data Warehouse Management System (IBM DB2 Data Warehouse Enterprise<br />

Edition under UNIX operation platform). This software is equipped with relational database management<br />

system and on-line analytical processing (OLAP) tools for integrating enterprise-wide corporate data into a<br />

single repository from which users can run queries, perform analysis and produce reports. <strong>The</strong> data flow<br />

diagram for SHDMS is given in Figure 48.<br />

5.2 Basic Data Operation Requirements in SHDMS<br />

<strong>The</strong> data operation in SHDMS is devised to fulfill the four basic types of data operation requirements, i.e., (i)<br />

subject-oriented data operation, (ii) integrated data operation, (iii) time-variant data operation and (iv)<br />

non-volatile data operation. <strong>The</strong> first requirement means that the data warehouse is organized around the major<br />

subjects of bridge monitoring works or the fifteen types of measurands as stated in Paragraph 3.3.2 above. <strong>The</strong><br />

second requirement means that the data from different sensory systems and application software systems at<br />

different hardware platforms are having different data formats, and these inconsistent data are integrated to a<br />

unified format for viewing, storage, processing and analysis. <strong>The</strong> third requirement means that the data in the<br />

data warehouse is accurate and valid at some point in time or over some time interval, and the time-variance of<br />

the data warehouse is also shown in the extended time that represents a series of snapshots. <strong>The</strong> last requirement<br />

means that as the data is not updated in real-time but is refreshed from operational systems on a regular basis,<br />

hence new data is always added as a supplement to the data warehouse rather than a replacement – the data<br />

warehouse is therefore required to continually absorb the new data and incrementally integrate it with the<br />

previous data.<br />

5.3 Functions of the Customized Software Tools in SHDMS<br />

<strong>The</strong> customized software tools in SHDMS are required to carry out four major functions of (i) data<br />

administration and management, (ii) operational data storage, (iii) data manipulation, and (iv) multi-dimensional<br />

view of data. <strong>The</strong> first function ensures the data warehouse having the capability of executing the tasks of (a)<br />

monitoring data loading from multiple sources, (b) data quality and integrity checks, (c) managing and updating<br />

metadata, (d) monitoring databases performance to ensure efficiency query response times and resource<br />

utilization, (e) auditing data warehouse usage to provide user chargeback information, (f) replicating, sub-setting<br />

and distributing data, (f) maintaining efficient data storage management, (g) purging data, (h) archiving and<br />

backing-up data, (i) implementing recovery following failure, and (j) security management. <strong>The</strong> second function<br />

-256-


efers to data processing and storage includes the tasks of systematic cleansing, reconciling, deriving, matching,<br />

standardizing, transforming, conforming, storing and exporting data and information received from DPCS,<br />

PDAS and PIMS for display, processing, analysis and interrogation. <strong>The</strong> third function refers to the data<br />

interfacing capability for semi-automatic operation of data retrieval from SHDMS-Server, downloading the<br />

retrieved data into relevant consoles and workstations, and execution of data analysis by the customized<br />

software tools as mentioned in Paragraph 3.3.2. This function is operated through a customized user-interfacing<br />

software tool which links the SHDMS-Server, consoles and workstations together for data communication. <strong>The</strong><br />

last function refers to the application of OLAP tools for getting more information such as development trends<br />

from the processed and analyzed data through multi-dimensional views.<br />

5.4 Data Manipulation Requirements<br />

As data processing, analysis and reporting are the key activities of the SHMS and in order to keep such activities<br />

in high efficiency basing on available hardware and software system, a time performance requirement is<br />

therefore set to execute three activities. This time performance requirement is that the time duration for retrieval,<br />

processing, analysis and reporting of any 120 data channels in 86.4 MB data (which is equivalent to the 2-hour<br />

of data collected from 40 nos. of sensors – each at a sampling rate of 100 Hz) should be no more than 6 minutes.<br />

<strong>The</strong> word “retrieval” means the download of the data from the rolling storage buffer in the SHDMS-server to<br />

the DPCS-workstations. <strong>The</strong> words “processing, analysis and reporting” refers to the execution of the<br />

customized software tools in the software system for bridge structural system, as mentioned in Paragraph 3.3.2,<br />

in the DPCS-workstations.<br />

6. CONCLUSIONS<br />

6.1 Summary of SHMS Design Process<br />

In the design of a SHMS, the types of measurands required should first be identified and the corresponding<br />

sensory layout should be devised to achieve the six functions as stated in Paragraph 3.2.1. <strong>The</strong> sensor type is<br />

then selected in accordance with the 12 criteria as stated in Paragraph 3.2.1. Upon completion of sensory system<br />

deployment and selection, the layouts and/or locations of the data acquisition systems and associated cabling<br />

networks should be deployed. Selection of the data acquisition equipments and associated cabling network<br />

systems for data acquisition and transmission should comply with the technical specification of the selected<br />

sensory system. After this stage, the system design work would primarily be on the software systems rather than<br />

the hardware systems. <strong>The</strong> software system design includes the design of: (i) the software system for control and<br />

display of the instrumentation system, (ii) the software system for monitor and display of the bridge, (iii) the<br />

software system for data interfacing and analysis execution, and (iv) the software system for data operation and<br />

data management. <strong>The</strong> first software system is essential in keeping the instrumentation system in good operating<br />

condition because though control and display, the operator can quickly identify the malfunction equipment and<br />

promptly arrange for rectification. <strong>The</strong> second software system is the solely system providing tools to convert<br />

the measured data into codified information so that bridge engineers can make use of this information for<br />

planning and scheduling of their works in design, construction, maintenance and rehabilitation. <strong>The</strong> third<br />

software system is devised to interface the measured data with the bridge monitoring and display tools and to<br />

execute the requested analysis in the DPCS-Workstations. <strong>The</strong> fourth software system is devised for the works<br />

of data operation, data storage and data retrieval. <strong>The</strong> third and the fourth software system are in fact the basis<br />

for semi-automatic processing and analysis of the measured data. <strong>The</strong> first and second software systems are<br />

installed in the DPCS, whereas the third and fourth systems are installed in the SHDMS.<br />

6.2 Performance Criteria for Monitoring<br />

<strong>The</strong> monitoring criteria for bridge performance and its environment should be based on design criteria in bridge<br />

codes (such as BS5400 or appropriate Eurocodes) at the serviceability limit state. This approach makes the<br />

monitoring results be acceptable by bridge engineers because its design and operation are complied with the<br />

codified requirements for bridge design, construction, maintenance and rehabilitation, which also forms the<br />

legal basis for monitoring. Under such an approach, all software tools for data processing and analysis should<br />

therefore be customized to transform the measured data into the parameters specified in the bridge design codes.<br />

6.3 Nature and Functions of SHMS<br />

SHMS is in fact a civil-mechatronic system which integrates the knowledge and experience in the fields of<br />

sensing and instrumentation, modeling and simulation, data interrogation and management, and bridgeworks of<br />

design, construction, maintenance and rehabilitation. It should be devised to carry out the functions of collection,<br />

-257-


processing and analysis of bridge data and information for: (i) monitoring the structural performance of the<br />

bridge and its environment under the serviceability limit state, and (ii) provision of alarms when the monitoring<br />

criteria in certain structural components or locations are exceeded.<br />

ACKNOWLEDGEMENT<br />

<strong>The</strong> author wishes to express his thanks to the Director of Highways for his permission to publish this article.<br />

Any opinions expressed or conclusions reached in the text are entirely those of the author.<br />

REFERENCES<br />

[1] Stern, M. and Geary, A. L., “Electrochemical Polarization – A <strong>The</strong>oretical Analysis of Shape of<br />

Polarization Curves”, Journal of the Electrochemical Society, Vol. 104, 1957, pp.56063.<br />

[2] Ferry-Borges, J and Castanheta, M., “Structural Safety”, LNEC, Lisbon, 1971.<br />

[3] Engineering Sciences Data Unit, “Characteristics of atmospheric turbulence near the ground – Part I :<br />

Definitions and General Information”, ESDU 74030, October 1976.<br />

[4] British Standard Institution, “BS5400 : Part 10 : 1980 – Code of practice for fatigue”, British Standard<br />

for steel, concrete and composite bridges.<br />

[5] Augusti, G., Baratta, A. and Casciati, F., “Probabilistic Methods in Structural Engineering”, Chapman<br />

and Hall Ltd., 1984.<br />

[6] British Standard Institution, “BS5400: Part 4 : 1990 – Code of practice for design of concrete bridges”,<br />

British Standard for steel, concrete and composite bridges.<br />

[7] Rubin, S., “Concepts in Shock Data Analysis”, Chapter 23 of Shock and Vibration Handbook, 4 th<br />

Edition, McGraw-Hill Co. Inc., 1996.<br />

[8] Highways Department, “Wind and Structural Health Monitoring system for Lantau Fixed Crossing and<br />

Ting Kau Bridge”, Consultancy Agreement No. CE 75/94, the Government of <strong>Hong</strong> <strong>Kong</strong>, 1997.<br />

[9] ASTM, C876-91, “Standard Test Method for Half-Cell Potential of Uncoated Reinforcing Steel in<br />

Concrete”, 1999 pages 446 to 451.<br />

[10] Chopra, A.K., “Dynamics of Structures – <strong>The</strong>ory and Applications to Earthquake Engineering”, 2 nd<br />

Edition, Prentice Hall Inc., 2001.<br />

[11] Engineering Sciences Data Unit, “Characteristics of atmospheric turbulence near the ground – Part II :<br />

Single point data for strong winds (neutral atmosphere)”, ESDU 85020, August 2001.<br />

[12] Engineering Sciences Data Unit, “Characteristics of atmospheric turbulence near the ground – Part<br />

III : Variations in space and time for strong winds (neutral atmosphere)”, ESDU 86010, August 2001.<br />

[13] Engineering Sciences Data Unit, “Strong winds in the atmospheric boundary layer – Part 1 :<br />

hourly-mean wind speeds”, ESDU 82026, March 2002.<br />

[14] Engineering Sciences Data Unit, “Strong winds in the atmospheric boundary layer – Part 2 : discrete<br />

gust speeds”, ESDU 83045, March 2002.<br />

[15] Highways Department, “Wind And Structural Health Monitoring System”, Section 33 of the Particular<br />

Specification for the Construction of Stonecutters Bridge, Highway Contract No. HY/2002/26, the<br />

Government of the <strong>Hong</strong> <strong>Kong</strong> Special Administrative Region, 2002.<br />

[16] Wong, K.Y., “Structural Identification of Tsing Ma Bridge”, Transactions of <strong>Hong</strong> <strong>Kong</strong> Institution of<br />

Engineers, Volume 10, Number 1, March 2003, pp. 38-47.<br />

[17] Highways Department, “Design Memorandum for Stonecutters Bridge Design and Construction<br />

Assignment”, under Agreement No. CE 61/2000 (between Highways Department and Ove Arup<br />

Partners <strong>Hong</strong> <strong>Kong</strong> Ltd), 7 th Issue, 16 June 2004.<br />

[18] Lyons, R., “Understanding Digital Signal Processing”, Prentice Hall, 2004.<br />

[19] MSC Software, “MSC.Nastran Reference Manual”, 2004.<br />

[20] Larsen, S. and Damsgaard, A., “Full Aeroelastic Model Tests for Stonecutters Bridge, <strong>Hong</strong> <strong>Kong</strong>”, A<br />

Report submitted to Highways Department – Major Works Project Management Office, 1 November<br />

2005.<br />

[21] Wong, K.Y., “Modeling and Simulation of Cable-Supported Bridges by MSC/NASTRAN”, MSC<br />

Software Virtual Product Development Conference, organized by MSC Software corporation, 3-4<br />

November 2005, Mountain Emei, Chengdu, China.<br />

[22] FIB (2006) : “Model Code for Service Life Design”, Model Code, Bulletin 34.<br />

[23] SAMCO, Final Report 2006, “F09 Report on Bridge Management”, SAMCO, March 2006.<br />

[24] Highways Department, “Structures Design Manual for Highways and Railways”, 3 rd Edition, the<br />

Government of <strong>Hong</strong> <strong>Kong</strong> Special Administrative Region, August 2006.<br />

[25] Wong, K.Y., “Highway Traffic Loads Monitoring in Stonecutters Bridge”, Proceedings of International<br />

Conference on Bridge Engineering, held in <strong>Hong</strong> <strong>Kong</strong>, during 1-3 November 2006, organized by<br />

-258-


<strong>Hong</strong> <strong>Kong</strong> Institution of Engineers.<br />

[26] Song, H.W. and Saraswathy, V., “Corrosion Monitoring of Reinforced Concrete Structures – A Review”,<br />

International Journal of Electrochemical Science, 2 (2007) 1-28.<br />

[27] Hansson, C.M., Poursaee, A. and Jaffer, S.J., “Corrosion of Reinforceing Bars in Concrete”, PCA<br />

R&D Serial No. 3013, Portland Cement Association, 2007.<br />

[28] Schwarz, B. and Richardson, M., “Using a De-Convolution Window for operating Modal Analysis”,<br />

paper presented at the Internal Operational Modal Analysis Conference (IOMAC), 30 April – 2 May,<br />

2007.<br />

[29] Kumarasena S., Jones, N. P., Irwin, P. and Taylor, P., “Wind-Induced Vibration of Stay Cables”,<br />

Publication No. FHWA-HRT-05-083, Federal Highway Administration, US Department of<br />

Transportation, August 2007.<br />

[30] Markeset, G. and Myrdal R., “Modelling of Reinforcement Corrosion in Concrete – State of the Art”,<br />

COIN Project Report 7 – 2008.<br />

[31] Wong, K.Y., Chan, K.W.Y. and Chong, W.K., “Monitoring of Wind-Effects in Tsing Ma Bridge”, the<br />

Proceedings of 4 th International Conference on Advances in Wind and Structures (AWAS’08), in Jeju,<br />

Korea, 29-31 May 2008.<br />

[32] Highways Department, “Development of Structural Prognosis Tools for Evaluation of Stonecutters<br />

Bridge under In-Service Condition”, an assignment brief of research collaboration project for<br />

WASHMS development, the Government of the <strong>Hong</strong> <strong>Kong</strong> Special Administrative Region, March<br />

2009.<br />

[33] Highways Department, “Development of Structural Health Rating System for Inspection and<br />

Maintenance of Stonecutters Bridge under In-Service Condition”, an assignment brief of research<br />

collaboration project for WASHMS development, the Government of the <strong>Hong</strong> <strong>Kong</strong> Special<br />

Administrative Region, March 2009.<br />

[34] Wong, K.Y. and Ni,Y.Q., “Structural Health Monitoring of Cable-Supported Bridges in <strong>Hong</strong> <strong>Kong</strong>”,<br />

Chapter 12 of the book: Structural Health Monitoring of Civil Infrastructure Systems, edited by V.<br />

Karbhari and F. Ansari, Woodhead Publishing, Cambridge, UK, 371-411, 2009.<br />

[35] Wong, K.Y., Yau, D.M.S. and Hui, M.C.H., “Structural Health Monitoring of Stonecutters Bridge<br />

under the Construction Stage”, Proceedings of the 5 th International Conference on Bridge Maintenance,<br />

Safety and Management, Philadelphia, Pennsylvania, USA, 11-15 July 2010.<br />

[36] Bridge & Structures Division, “Report of Stay Forces Measurements in Stonecutters Bridge basing on<br />

Frequency Method during 4-12 November 2009”, an internal report submitted to Major Work Project<br />

Management Office, Highways Department, the Government of the <strong>Hong</strong> <strong>Kong</strong> Special Administrative<br />

Region, 25 November 2009.<br />

[37] Bridges & Structures Division, “Report of Load Trials on Stonecutters Bridge during 7-11 December<br />

2009”, an internal report submitted to Major Work Project Management Office, Highways<br />

Department, the Government of the <strong>Hong</strong> <strong>Kong</strong> Special Administrative Region, April 2011.<br />

[38] <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong>, “3D Space Frame (Implicit) Finite Element Model (Final Issue)”,<br />

a report submitted to Major Work Project Management Office, Highways Department, the Government<br />

of the <strong>Hong</strong> <strong>Kong</strong> Special Administrative Region, April 2011.<br />

-259-


Monitoring<br />

Category<br />

Environmental<br />

Loads and Status<br />

Monitoring<br />

Table 1 Details of Environmental Loads and Status Monitoring<br />

Measurands<br />

Sensor Type<br />

Location<br />

Name<br />

Notation<br />

Bi-axial Propeller Anemometer Ane (P2D) • Deck-Levels<br />

Bi-axial Ultrasonic Anemometer Ane (U2D) • Tower-Tops<br />

Wind &<br />

• Deck-Levels<br />

Weather Tri-axial Ultrasonic Anemometer Ane (U3D) • Tower-Tops<br />

• Deck-Levels<br />

Barometer Bar • Deck-Levels<br />

Rainfall Gauge Rfg • Deck-Levels<br />

Structural Steel Temperature TMU-1 • Steel Decks<br />

Sensor – Bolt Fixed Type<br />

• Upper Towers<br />

Structural Steel Temperature TMU-2 • Steel Decks<br />

Sensor – Bond Mount Type<br />

• Upper Towers<br />

Structural Concrete Temperature TMU-3 • Lower Towers<br />

Sensor – Tower<br />

• Upper towers<br />

Structural Concrete Temperature TMU-4 • Concrete Decks<br />

Temperature<br />

Sensor – Deck<br />

Cable Steel Temperature Sensor TMU-5 • Stay Cables<br />

– Fibre Optic Type<br />

Air Temperature and Relative TMU-6 • Concrete Decks<br />

Humidity<br />

Asphalt Pavement Temperature<br />

Sensor<br />

Seismic Tri-axial Servo Type<br />

Accelerometers<br />

TMU-7<br />

Corrosion<br />

Status Anode Ladder Corrosion Cells Cor<br />

Acc-F (3D)<br />

• Concrete Towers<br />

• Steel Deck Pavement<br />

• Concrete Deck<br />

Pavement<br />

• Deck-Levels<br />

• Tower-Tops<br />

• Tower-Bases<br />

• Concrete Decks<br />

• Concrete Towers<br />

• Piers<br />

Monitoring<br />

Category<br />

Operation Loads<br />

Monitoring<br />

Table 2 Details of Operation Loads Monitoring<br />

Measurands<br />

Sensor Type<br />

Name<br />

Notation<br />

Dynamic Weigh-in-Motion DWIMS<br />

Highway Stations<br />

Traffics<br />

Fixed Video Cameras<br />

VC-F<br />

Ship Impacting<br />

Permanent<br />

Loads<br />

Location<br />

• Concrete Decks<br />

• Towers<br />

• Concrete Decks<br />

• East Portal Building<br />

Portable Video Cameras VC-P • Deck-Levels<br />

Fixed Tri-axial Servo Type Acc-F (3D)<br />

Accelerometers<br />

Single Dynamic (Weldable Foil<br />

Type) Strain Gauges<br />

Rosette Dynamic (Weldable Foil<br />

Type) Strain Gauges<br />

Single Static (Vibrating Wire<br />

Type) Strain Gauges<br />

Rosette Static (Vibrating Wire<br />

Type) Strain Gauges<br />

D-Str (S)<br />

D-Str (R)<br />

S-Str (S)<br />

S-Str (R)<br />

• Deck-Levels<br />

• Tower-Tops<br />

• Tower-Bases<br />

• Steel Longitudinal<br />

Box Girders<br />

• Steel Orthotropic<br />

Deck Trough Sections<br />

• Steel Longitudinal<br />

Box Girders<br />

• Steel Orthotropic<br />

Deck Trough Sections<br />

• Tower-Bases<br />

• Pier-Bases<br />

-260-


Monitoring<br />

Category<br />

Bridge Features<br />

Monitoring<br />

Table 3 Details of Bridge Features Monitoring<br />

Measurands<br />

Sensor Type<br />

Name<br />

Notation<br />

Location<br />

• Deck-Levels<br />

• Tower-Tops<br />

• East Portal Building<br />

Global Global Positioning Systems GPS<br />

Static<br />

Features Bi-axial Tiltmeters Til (2D) • Concrete Decks<br />

Global<br />

Dynamic<br />

Features<br />

Fixed Tri-axial Servo Type<br />

Accelerometers<br />

Removal Tri-axial Servo Type<br />

Accelerometers<br />

Acc-F (3D)<br />

Acc-R (3D)<br />

• Towers<br />

• Deck-Levels<br />

• Tower-Tops<br />

• Stay Cables<br />

• Towers<br />

• Decks<br />

• Piers<br />

Monitoring<br />

Category<br />

Bridge<br />

Responses<br />

Monitoring<br />

Measurands<br />

Table 4 Details of Bridge Responses Monitoring<br />

Sensor Type<br />

Location<br />

Name<br />

Notation<br />

Stay Forces Removal Tri-axial Servo Acc-R (3D) • Stay Cables<br />

Type Accelerometers<br />

Tendon Forces Tensio Magnetic Gauge Ten • Tendons<br />

GPS (Rov) • Deck-Levels (Main-Span)<br />

Displacements Global Positioning Systems<br />

• Tower-Tops<br />

GPS (Ref) • East Portal Building<br />

Bi-axial Tiltmeters Til (2D) • Towers<br />

• Piers<br />

• Concrete Decks<br />

Single Static (Vibrating S-Str (S) • Upper Piers<br />

Wire Type) Strain Gauges<br />

• Lower Piers<br />

• Upper Towers<br />

• Lower Towers<br />

Stress/Force<br />

Distribution<br />

Fatigue Damage<br />

Rosette Static (Vibrating<br />

Wire Type) Strain Gauges<br />

Single Dynamic (Weldable<br />

Foil Type) Strain Gauges<br />

Rosette Dynamic<br />

(Weldable Foil Type) Strain<br />

Gauges<br />

Single Dynamic (Weldable<br />

Foil Type) Strain Gauges<br />

S-Str (R)<br />

D-Str (S)<br />

D-Str (R)<br />

D-Str (S)<br />

• Concrete Decks<br />

• Upper Piers<br />

• Lower Piers<br />

Rosette Dynamic D-Str (R)<br />

(Weldable Foil Type) Strain<br />

Gauges<br />

Articulation Hydraulic Buffer Sensors Buf (9AS) • Towers<br />

Responses Lateral Bearing Sensors Brg (4AS) • Towers<br />

• Steel Longitudinal Box Girder<br />

• Steel Cross Box Girder<br />

• Steel Orthotropic Deck Trough<br />

Sections<br />

• Steel Cable Anchorage in Steel<br />

Girders<br />

• Steel Skin in Upper Tower<br />

• Steel Cable Anchorage in Upper<br />

Tower<br />

• Steel Longitudinal Box Girders<br />

• Steel Cable Anchorage in Steel<br />

Girders<br />

• Steel Longitudinal Box Girder<br />

• Steel Cross Box Girder<br />

• Steel Orthotropic Deck Trough<br />

Sections<br />

• Steel Cable Anchorage in Steel<br />

Girders<br />

• Steel Cable Anchorage in Upper<br />

Tower<br />

• Steel Longitudinal Box Girders<br />

• Steel Cable Anchorage in Steel<br />

Girders<br />

-261-


Table 5 Fundamental Wind Data Analysis for 3D Ultrasonic Anemometers<br />

No Type of Processing Measurands/Parameter Remarks<br />

.<br />

s<br />

1 Display of <strong>The</strong> measurands required • U is the along-wind component.<br />

Time-Series Raw Data are: U, V and W. • V is the across-wind component.<br />

2 Display of<br />

Time-Series Processed<br />

Data<br />

3 Derivation of<br />

Resultants & Angles<br />

<strong>The</strong> processed<br />

measurands required are:<br />

U, V and W.<br />

<strong>The</strong> derived wind<br />

parameters required are:<br />

H, R, β and θ.<br />

4 Basic Statistics <strong>The</strong> 11 statistical<br />

Analysis<br />

parameters required are:<br />

Probability Density<br />

Function, Mean, Mean<br />

Square, Root Mean<br />

Square, Variance,<br />

Standard Deviation,<br />

Skewness, Kurtosis,<br />

Maximum, Minimum and<br />

Gust Factor.<br />

4 Diurnal and Nocturnal <strong>The</strong> processed<br />

Wind Characteristic measurands and derived<br />

Plots<br />

wind parameters required<br />

are: U, V, W, H, R, β and<br />

θ.<br />

5 Histogram and <strong>The</strong> processed<br />

Cumulative Histogram measurands and derived<br />

Plots (Bin Size = 1 wind parameters required<br />

m/s or 1°, where are: U, V, W, H, R, β and<br />

appropriate)<br />

θ.<br />

6 Wind Rose Diagrams <strong>The</strong> processed<br />

measurands required are:<br />

U and V.<br />

7 Wind Incidence<br />

Diagrams<br />

8 Division of processed<br />

wind data according to<br />

16 horizontal sectors<br />

(each of 22.5°)<br />

<strong>The</strong> derived wind<br />

parameters required are:<br />

R-θ.<br />

<strong>The</strong> 16 divisions are:<br />

Sector 1, Sector 2, Sector<br />

3, Sector 4, Sector 5,<br />

Sector 6, Sector 7, Sector<br />

8, Sector 9, Sector 10,<br />

Sector 11, Sector 12,<br />

Sector 13, Sector 14,<br />

Sector 15 and Sector 16.<br />

• W is the vertical-wind component.<br />

• Same as above, but wind data with<br />

spikes and noises removed.<br />

• H is the resultant of U and V.<br />

• R is the resultant of U, V and W.<br />

• β is the horizontal angle of H with<br />

reference to bridge alignment which is 0°<br />

at NW side (+ve for clockwise)<br />

• θ is the vertical angle of R with reference<br />

to horizontal plane which is 0° (+ve for<br />

upward)<br />

• <strong>The</strong> analysis is conducted for those<br />

parameters as listed in Nos. 2 and 3 of<br />

this table.<br />

• <strong>The</strong> time interval for analysis should be<br />

selectable as 1-sec, 3-sec, 1-min, 10-min<br />

and 1-hr.<br />

• <strong>The</strong> time-intervals for plots are selectable<br />

as 10-min and 1-hr.<br />

• <strong>The</strong> time-intervals for plots are selectable<br />

as 1-sec, 3-sec, 1-min, 10-min and 1-hr.<br />

• <strong>The</strong> time-intervals for plots are selectable<br />

as 1-sec, 3-sec, 1-min, 10-min and 1-hr.<br />

• <strong>The</strong> time-intervals for plots are selectable<br />

as 1-sec, 3-sec, 1-min, 10-min and 1-hr.<br />

WNW<br />

W<br />

WSW<br />

NW<br />

V(t)<br />

SW<br />

15<br />

14<br />

NNW<br />

16<br />

13<br />

02<br />

01<br />

12<br />

W(t)<br />

11<br />

N<br />

03<br />

10<br />

NNE<br />

NE<br />

04<br />

U(t)<br />

05<br />

09<br />

08<br />

06<br />

07<br />

SE<br />

ENE<br />

E<br />

ESE<br />

SSW<br />

S<br />

SSE<br />

Bridge Alignment<br />

Notes:<br />

1. <strong>The</strong> angle of Bridge Alignment away from North (N) is 129.8º (in Clockwise Direction).<br />

2. Each Sector is a 22.5° Sector.<br />

-262-


9 Combination of Above<br />

Divided Sectors<br />

<strong>The</strong> nine combinations<br />

are:<br />

CB01, CB02, CB03,<br />

CB04, CB05, CB06,<br />

CB07, CB08 and CB09.<br />

Notes:<br />

1. <strong>The</strong> time-interval/duration of data processing and analysis should first be determined.<br />

2. <strong>The</strong> default time-interval should be 10-min.<br />

• CB01 – Any 1 Sector<br />

• CB02 – Any 2 Continuous Sectors<br />

• CB03 – Any 3 Continuous Sectors<br />

• CB04 – Any 4 Continuous Sectors<br />

• CB05 – Any 5 Continuous Sectors<br />

• CB06 – Any 6 Continuous Sectors<br />

• CB07 – Any 7 Continuous Sectors<br />

• CB08 – Any 8 Continuous Sectors<br />

• CB09 – All 16 Sectors<br />

Table 6 Determination of Stress Range (σ r ) and Length (L) in Each Typical Structural Component (Location)<br />

considered in Fatigue Assessment<br />

Stonecutters Bound<br />

Tsing Yi Bound<br />

VT Vehicle EUL<br />

Slow Mid Lane Fast Lane Fast Lane Mid Lane Slow<br />

Weight (kN)<br />

Lane<br />

Lane<br />

(kN)<br />

σ r L σ r L σ r L σ r L σ r L σ r L<br />

1 30 – 55 42.5 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1<br />

2 55 – 240 147.5 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1<br />

3 240 – 440 340 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1<br />

4 40 – 240 140 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1<br />

5 500 – 1500 1000 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1 #1<br />

Note:<br />

1. VT = Vehicle Type, EUL = Equivalent Unit Load and σ r is in MPa and L is in m.<br />

2. Vehicle with a vehicle weight of less than 30 kN is not considered for fatigue damage assessment.<br />

3. #1 = <strong>The</strong> values should be determined from the results of influence lines analysis basing on the equivalent unit load as<br />

specified in the 3 rd column.<br />

Table 7 Determination of Applied Stress Cycles (n i ) and Failure Stress Cycles (N i ) in accordance with the results<br />

from Table 6 above<br />

Stonecutters Kau Bound<br />

Tsing Yi Bound<br />

VT Vehicle EUL<br />

Slow Mid Lane Fast Lane Fast Lane Mid Lane Slow<br />

Weight (kN)<br />

Lane<br />

Lane<br />

(kN)<br />

n i N i n i N i n i N i n i N i n i N i n i N i<br />

1 30 – 55 42.5 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2<br />

2 55 – 240 147.5 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2<br />

3 240 – 440 340 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2<br />

4 40 – 240 140 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2<br />

5 500 – 1500 1000 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2<br />

Notes:<br />

1. VT = Vehicle Type and EUL = Equivalent Unit Load<br />

2. #1 = n i should be obtained from the annual highway traffic flow spectrum.<br />

3. #2 = N i should be obtained from Clause 11.2 of BS 5400 : Part 10 based on the computed value of σ r .<br />

4. i = 1, …….., 20.<br />

Table 8 Determination of Fatigue Damage in accordance with the results from Table 7 above<br />

Stonecutters Bound<br />

Tsing Yi Bound<br />

V Vehicle EU<br />

Slow Mid Lane Fast Lane Fast Lane Mid Lane Slow<br />

T Weight (kN)<br />

Lane<br />

Lane<br />

(kN)<br />

n i / N i n i / N i n i / N i n i / N i n i / N i n i / N i<br />

1 30 – 55 42.5 #1 #2 #3 #4 #5 #6<br />

2 55 – 240 147.5 #1 #2 #3 #4 #5 #6<br />

3 240 – 440 340 #1 #2 #3 #4 #5 #6<br />

4 40 – 240 140 #1 #2 #3 #4 #5 #6<br />

5 500 – 1500 1000 #1 #2 #3 #4 #5 #6<br />

Sum of Damage Ratio due<br />

to Individual Traffic Lane<br />

#7 #7 #7 #7 #7 #77<br />

-263-


Sum of Damage Ratio due #8<br />

to All Traffic Lanes<br />

Notes:<br />

1. VT = Vehicle Type and EUL = Equivalent Unit Load<br />

2. #1 = <strong>The</strong> damage ratio of n i / N i should be computed basing on the results of the Slow Lane Column under Stonecutters<br />

Bound Traffic in Table 7.<br />

3. #2 = <strong>The</strong> damage ratio of n i / N i should be computed basing on the results of the Mid Lane Column under Stonecutters<br />

Bound Traffic in Table 7.<br />

4. #3 = <strong>The</strong> damage ratio of n i / N i should be computed basing on the results of the Fast Lane Column under Stonecutters<br />

Bound Traffic in Table 7.<br />

5. #4 = <strong>The</strong> damage ratio of n i / N i should be computed basing on the results of the Fast Lane Column under Tsing Yi<br />

Bound Traffic in Table 7.<br />

6. #5 = <strong>The</strong> damage ratio of n i / N i should be computed basing on the results of the Mid Lane Column under Tsing Yi<br />

Bound Traffic in Table 7.<br />

7. #6 = <strong>The</strong> damage ratio of n i / N i should be computed basing on the results of the Slow Lane Column under Tsing Yi<br />

Bound Traffic in Table 7.<br />

8. #7 = <strong>The</strong> value should be determined by summing up the damage ratios obtained from vehicle type 1 to vehicle type 5.<br />

9. #8 = <strong>The</strong> value should be determined by summing up the damage ratios obtained at individual traffic lanes.<br />

-264-


Back to Routine Monitoring<br />

Not<br />

Exceeded<br />

Not<br />

Exceeded<br />

BIMT<br />

SHMS<br />

Health<br />

Monitoring<br />

SHES<br />

Exceeded<br />

Safety<br />

Evaluation<br />

Exceeded<br />

Evaluation Report<br />

Download of<br />

Monitoring Indices<br />

Call for<br />

Evaluation Indices<br />

Download of<br />

Evaluation Indices<br />

Call for<br />

Monitoring Indices<br />

SHRS<br />

SHDMS<br />

Download of Monitoring Data from SHMS<br />

Download of Inspection & Maintenance Results<br />

Figure 1 Flow Diagram of SCB-WASHMS for Stonecutters Bridge<br />

(Note : BIMT = Bridge Inspection & Maintenance Team)<br />

SHMS<br />

Routine Monitoring<br />

SHMS-<br />

Server<br />

Upload or Download<br />

SHDMS-<br />

Server<br />

Structure<br />

Inventory Module<br />

HIID-DB<br />

SHMD-DB<br />

Environmental<br />

Loads & Status<br />

Monitoring<br />

Operational<br />

Loads<br />

Monitoring<br />

Structural<br />

Features<br />

Monitoring<br />

Structural<br />

Responses<br />

Monitoring<br />

SHED-DB<br />

SHRD-DB<br />

No Exceedance<br />

Is the measured<br />

values exceeding<br />

75% of the Design Value<br />

at SLS<br />

Download Criteria<br />

Note: SLS = Serviceability Limit State<br />

Exceedance<br />

Figure 2 Operation Flow Diagram of SHMS (Structural Health Monitoring System)<br />

-265-


Ane (P2D)<br />

Bar & Rfg<br />

Rfg<br />

Buf & Brg<br />

Til<br />

Ane (U2D) &<br />

Ane (U3D)<br />

Acc-F<br />

(3D)<br />

D-Str<br />

TMU 1-4,<br />

6 and 7<br />

S-Str<br />

Ten<br />

GPS-Antennae<br />

(Base Reference<br />

Stations)<br />

GPS-LCNS<br />

DSG-LDAUs<br />

TAUs<br />

SSG-LDAUs<br />

DAS-Ten<br />

DAS-GPS<br />

DAS-S-Str<br />

GPS-LCNS<br />

Subnets<br />

Individual LCNS<br />

WCNS<br />

LCNS<br />

DAS -<br />

Static & Dynamic<br />

Acc-R<br />

(3D)<br />

PDAS-1<br />

D-Str<br />

GCNS-1<br />

Cor<br />

PDAS-2<br />

(or DAS-Cor)<br />

WCNS<br />

LCNS<br />

DPCS-1<br />

Server<br />

DAS-GPS DAS-TMU5- DAS-DWIMS<br />

GPS-LCNS<br />

GPS-Antennae<br />

(Rover)<br />

LCNS LCNS LCNS<br />

LCNS<br />

TMU-5<br />

LCNS<br />

Bending Plates &<br />

Induction Loops<br />

Notes:<br />

1. PDAS-1 is for collection of data from Portable 3-D<br />

Accelerometers.<br />

2. PDAS-2 is for collection of data from Corrosion cells.<br />

3. <strong>The</strong> DAS for non-proprietary sensory systems has a<br />

total of 8 numbers of Data Acquisition Units.<br />

4. DPCS-1 is located in the Structural Health<br />

Monitoring Room inside the Administration Building.<br />

Figure 3 Schematic Network Layout of One Dimensional Signals<br />

4 Video Cameras<br />

at Deck Levels<br />

4 Video Cameras<br />

at Tower-Tops<br />

4 Video Cameras<br />

at Mid-Tower Height<br />

2 Video Cameras<br />

at East Portal Building<br />

Individual LCNS<br />

DAS-VC<br />

(or Video Signals<br />

Converters for Video Cameras)<br />

Video Cameras<br />

Management System<br />

GCNS-2<br />

Notes:<br />

1. <strong>The</strong> video cameras management system (VCMS) is used for the control and operation of<br />

video cameras and also used as a rolling buffer for the storage and retrieval of video signals<br />

2. <strong>The</strong> DPCS-2 is used for the traffic composition and traffic load-effects analyses.<br />

3. Both VCMS and DPCS-2 are located in the Structural Health Monitoring Room inside the<br />

Administration Building.<br />

DPCS-2<br />

Server<br />

Figure 4 Schematic Network Layout of Two Dimensional (Video) Signals<br />

-266-


Tsing Yi Stonecutters<br />

Figure 5 Layout of Sensory System in Stonecutters Bridge<br />

-267-


Figure 6 Photo of Structural Health Monitoring Room for Stonecutters Bridge<br />

DAS<br />

DAS for Static &<br />

Dynamic Signals<br />

DAS for Dynamic Weigh-<br />

In-Motion Station Signals<br />

SHMS<br />

DPCS<br />

Software System for<br />

Instrumentation<br />

System<br />

(DPCS-Server)<br />

SHDMS<br />

Permanent<br />

Storage<br />

System<br />

DAS for GPS Signals<br />

DAS for Fiber Optic<br />

Signals<br />

Rolling Buffer<br />

Storage System in<br />

SHDMS-Server<br />

DAS for Vibrating Wire<br />

Type Strain Gauge Signals<br />

DAS for Tensio Magnetic<br />

Gauge Signals<br />

Storage Systems<br />

in DPCS-Servers<br />

Data Warehouse<br />

Management<br />

System<br />

DAS for Corrosion<br />

Cell Signals<br />

DAS for Video Signals<br />

Software System for<br />

Bridge Structural<br />

System<br />

(DPCS-Workstations)<br />

Software System for<br />

Data Interfacing &<br />

Analysis Execution<br />

Figure 7 Flow Diagram of the 3 Major Software Systems in DPCS of SHMS<br />

-268-


Continuous Data Acquisition<br />

GPS Time Synchornization<br />

of all time-series data<br />

Operation Status of Sensors &<br />

Related Equipments in DAU1<br />

Operation Status of Sensors &<br />

Related Equipments in DAU2<br />

Heading of all time-series data<br />

Operation Status of Sensors &<br />

Related Equipments in DAU3<br />

Instrumentation System Items for Control<br />

Scaling of Sensors Readings<br />

Generation of Data Files<br />

Statistical Summary Processing<br />

Traceable Event Log with<br />

Time-Tags<br />

Automatic Data Transmission<br />

with manual options<br />

Reconfiguration of Ring Buffer<br />

Storage Time (Default: 40 days)<br />

Reconfiguration of Data<br />

Transmission Rate<br />

Reconfiguration of<br />

Alarm Status (Default: 10 sec.)<br />

Reconfiguration of Sampling<br />

Rates for all Acc & D-Str<br />

Automatic Archiving Media<br />

Management for Storage &<br />

Retrieval of Data<br />

Automatic Data Archiving,<br />

Filing and Backup<br />

Display Control of<br />

Six Plasma Monitors<br />

Software System<br />

for<br />

Instrumentation<br />

System<br />

(Operating in<br />

DPCS-Server)<br />

Operation Status of Sensors &<br />

Related Equipments in DAU4<br />

Operation Status of Sensors &<br />

Related Equipments in DAU5<br />

Operation Status of Sensors &<br />

Related Equipments in DAU6<br />

Operation Status of Sensors &<br />

Related Equipments in DAU7<br />

Operation Status of Sensors &<br />

Related Equipments in DAU8<br />

Operation Status of Sensors &<br />

Loggers in DAS-GPS<br />

Operation Status of Sensors &<br />

Loggers in DAS-TMU-5<br />

Operation Status of Sensors &<br />

Loggers in DAS-S-Str<br />

Operation Status of Sensors &<br />

Loggers in DAS-Ten<br />

Operation Status of Sensors &<br />

Loggers in DAS-DWIMS<br />

Operation Status of Sensors &<br />

Loggers in DAS-VC<br />

Operation Status of Storage<br />

Systems in DPCS-Servers<br />

Instrumentation System Items for Display<br />

Figure 8 Required items for Control and Display carried out by DPCS-Servers<br />

Sensor drawing<br />

Sensor photo<br />

Real-time display of data (wind )<br />

Technical Specification of Sensor (Anemometer)<br />

Figure 9 A Typical Display Form of the Software System for Instrumentation System<br />

-269-


Figure 10<br />

Spectral Densities of Wind Velocities and Earthquake Acceleration, compared with Natural Frequencies<br />

of Major Structures [Ref. 2 and 5]<br />

Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

Ane (U3D)<br />

Ane (U2D)<br />

Ane (P2D)<br />

Acc-F(3D)<br />

GPS<br />

D-Str<br />

TMU-1<br />

TMU-2<br />

Bar<br />

Rfg<br />

TMU-6<br />

Spikes & Noises<br />

Removal<br />

Wind Data<br />

Division &<br />

Combination<br />

Wind Data<br />

Classification<br />

Selection of<br />

Data at the same<br />

time period as<br />

the wind data<br />

considered<br />

Data Editing &<br />

Classification<br />

Selection of<br />

Data at the same<br />

time period as<br />

the wind data<br />

considered<br />

Data Editing &<br />

Classification<br />

Typhoon<br />

Winds<br />

Monsoon<br />

Winds<br />

Other<br />

Winds<br />

Acceleration<br />

Displacement<br />

Dynamic<br />

Stress<br />

Steel<br />

Temperature<br />

Air Pressure<br />

Air<br />

Temperature<br />

Rainfall<br />

Relative<br />

Humidity<br />

Fundamental<br />

Wind Data<br />

Analysis<br />

Covariance<br />

Analysis<br />

Spectral<br />

Analysis<br />

Curve Fitting<br />

Analysis<br />

Extreme Value<br />

Analysis<br />

Frequency<br />

Response<br />

Analysis<br />

Correlation<br />

Analysis<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Wind Speeds &<br />

Wind Directions<br />

Terrain Factors<br />

Extreme Winds<br />

Wind Rose Diagrams<br />

Wind Incidences<br />

Wind Turbulence<br />

Intensities<br />

Wind Turbulence<br />

Length Scales<br />

Wind Turbulence<br />

Auto Spectra<br />

Wind Turbulence<br />

Co-Spectra<br />

Wind Turbulence<br />

Coherence Spectra<br />

Frequency<br />

Response Functions<br />

(Wind-Acceleration)<br />

Wind–Acceleration<br />

Correlation<br />

Wind–Displacement<br />

Correlation<br />

Wind–Stress<br />

Correlation<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Histogram & Cumulative<br />

Histogram Plots of Wind<br />

Speeds & Wind Directions<br />

Plots of Wind Rose Diagrams<br />

in both Contour-Line Format<br />

& Telescope Format<br />

Plots of Extreme Winds with<br />

comparison of design winds in<br />

120- & 2400-year return periods<br />

Plots of Integral Turbulence<br />

Length Scale by Direct<br />

Integration with comparison of<br />

Spectral Peal Length Scale by<br />

Von Karman Model<br />

Plots of Wind Power Spectrum<br />

with fitting of Von Karman Curve<br />

& determination of frequency<br />

at Peak Spectral value<br />

Plots of Wind Speed Profiles &<br />

Wind Turbulence Intensity<br />

Profiles based on Power Law<br />

Wind Transfer Function Plots<br />

(i.e., Effective Mass &<br />

Dynamic Stiffness)<br />

X-Y and/or X-Y-Z Plots of<br />

Wind Responses<br />

Histogram Plots<br />

References of Air Pressure,<br />

Air Temperature, Rainfall<br />

& Relative Humidity<br />

Figure 11 Flow Diagram of Customized Software Tools for Wind and Weather Monitoring<br />

-270-


Y(+ve)<br />

Y T<br />

Temperature Distribution<br />

in Deck Cross Section<br />

=<br />

+ +<br />

Measured<br />

Temperatures<br />

(Elevation)<br />

Effective<br />

Bridge<br />

Temperature<br />

Linear<br />

Differential<br />

Temperature<br />

T T E<br />

T L<br />

( )<br />

Self-Equilibrating<br />

Temperature<br />

With reference to above diagram, let the measured mean temperature in a part of the cross-section of A i<br />

having its centroid distance Y i<br />

from the neutral axis of the cross-section be T i<br />

, the material of the part<br />

having coefficient of thermal expansion and modulus of elasticity be respective α i<br />

and E i<br />

.<br />

<strong>The</strong> Effective Temperature is computed as:<br />

T<br />

E<br />

=<br />

n<br />

∑<br />

i=<br />

1<br />

n<br />

∑ ( E i α i A i<br />

)<br />

i=<br />

1<br />

E<br />

i<br />

α<br />

i<br />

A<br />

i<br />

T<br />

i<br />

<strong>The</strong> Differential Temperature is computed as:<br />

T<br />

E<br />

=<br />

n<br />

∑ ( Ei<br />

αi<br />

Ai<br />

( Ti<br />

− TE<br />

) Yi<br />

)<br />

i=<br />

1<br />

n<br />

∑<br />

i=<br />

1<br />

⎛<br />

⎜<br />

Ei<br />

αi<br />

⎝<br />

2<br />

y ⎞<br />

i<br />

A ⎟<br />

i<br />

Y<br />

T ⎠<br />

<strong>The</strong> Self-Equilibrating Temperature is computed as:<br />

T<br />

SE<br />

=<br />

( T − T )<br />

i<br />

E<br />

Yi<br />

− TL<br />

Y<br />

Figure 12 Formulas of Major Temperature Parameters<br />

T<br />

Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

TMU-1<br />

TMU-2<br />

TMU-3<br />

TMU-4<br />

TMU-5<br />

TMU-6<br />

TMU-7<br />

GPS<br />

Buf<br />

Data Editing<br />

Temperature<br />

Data<br />

Classification<br />

Selection of<br />

Data at the same<br />

time period as<br />

the Temperature<br />

Data considered<br />

Data Editing &<br />

Classification<br />

Steel Deck<br />

Temperature<br />

Steel Tower<br />

Skin<br />

Temperature<br />

Concrete<br />

Tower<br />

Temperature<br />

Concrete<br />

Deck<br />

Temperature<br />

Stay Cable<br />

Temperature<br />

Air<br />

Temperature<br />

& Relative<br />

Humidity<br />

Asphalt<br />

Pavement<br />

Temperature<br />

Vertical<br />

Displacement<br />

of Bridge-Deck<br />

Longitudinal<br />

Displacement<br />

of Bridge-Deck<br />

Longitudinal<br />

Displacement<br />

of Tower-Top<br />

Sectional<br />

Temperature<br />

Distribution<br />

Analysis<br />

in terms of<br />

Effective<br />

Temperature,<br />

Differential<br />

Temperature &<br />

Self-Equilibrating<br />

Temperature<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Effective Temperature<br />

in Steel Deck<br />

Effective Temperature<br />

in Steel Tower Skin<br />

Effective Temperature<br />

in Concrete Tower<br />

Effective Temperature<br />

in Concrete Deck<br />

Effective Temperature<br />

in Stay Cable<br />

Effective Temperature<br />

in Asphalt Pavement<br />

Differential Temperature<br />

in Steel Deck<br />

Differential Temperature<br />

in Concrete Deck<br />

Differential Temperature<br />

in Concrete Tower<br />

Air Temperature at<br />

Above, Inside & Below<br />

Deck Section<br />

Air Temperature at Inside<br />

& Outside Tower Section<br />

Vertical Deck Displacement<br />

at ¼, ½, & ¾ Main-Span<br />

Longitudinal Deck<br />

Displacement at Tower<br />

Longitudinal Tower<br />

Displacement at Tower-Top<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Hourly, Daily, Monthly &Yearly<br />

Variation Plots of Effective<br />

Temperatures in respective<br />

Steel Deck, Steel Tower Skin,<br />

Concrete Tower, Concrete<br />

Deck, Stay Cable & Asphalt<br />

Pavement with associated Air<br />

Temperatures as references<br />

Plots of Maximum & Minimum<br />

Variations of Differential<br />

Temperatures in respective<br />

Steel Deck, Concrete Tower<br />

& Concrete Deck<br />

Plots of Maximum, Mean &<br />

Minimum Variations of<br />

Vertical Deck Displacement at<br />

¼, ½ & ¾ Main-Span with<br />

derivations of <strong>The</strong>rmal<br />

Displacement in mm per °C<br />

Plots of Maximum, Mean &<br />

Minimum Variations of<br />

Longitudinal Deck Displacement<br />

at Towers with derivations<br />

of <strong>The</strong>rmal Displacement<br />

in mm per °C<br />

Plots of Maximum, Mean &<br />

Minimum Variations of<br />

Longitudinal Tower<br />

Displacement at Tower-Tops<br />

with derivations of <strong>The</strong>rmal<br />

Displacement in mm per °C<br />

Figure 13 Flow Diagram of Customized Software Tools for Temperature Monitoring<br />

-271-


Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Checking/Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

Acc-F(3D)<br />

at<br />

Tower-<br />

Bases<br />

Data Editing<br />

Acceleration<br />

Data<br />

Classification<br />

Longitudinal<br />

Acceleration<br />

Lateral<br />

Acceleration<br />

Vertical<br />

Acceleration<br />

Peak Acceleration<br />

≧ MM5<br />

Response<br />

Spectrum Method<br />

(Newmark’s Method)<br />

Displacement<br />

Response Spectra in 3<br />

orthogonal directions<br />

Pseudo-velocity<br />

Response Spectra in 3<br />

orthogonal directions<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Global Bridge Deformed Shape<br />

of individual mode &<br />

Sum of modes considered<br />

Eigen<br />

Analyzed<br />

Results<br />

GPS<br />

Til<br />

Selection of<br />

Data at the same<br />

time period as<br />

the acceleration<br />

data considered<br />

Data Editing &<br />

Classification<br />

GPS at<br />

Tower-Top<br />

GPS at<br />

Deck-Level<br />

Til along<br />

Tower-Height<br />

Displacement<br />

Input Type Finite<br />

Element Analysis<br />

Pseudo-acceleration<br />

Response Spectra in 3<br />

orthogonal directions<br />

Bridge Deformed<br />

Shape and Force in<br />

each mode considered<br />

Derived Stress/Force Outputs<br />

at Each Dynamic Strain<br />

Instrumented Deck Sections<br />

or Pre-defined Key Locations<br />

Derived Stress/Force Outputs<br />

at Each Static Strain<br />

Instrumented Tower & Pier<br />

Sections or Pre-defined<br />

Key Locations<br />

TMU-1<br />

TMU-2<br />

TMU-3<br />

TMU-4<br />

TMU-6<br />

Ane(U3D)<br />

Ane(U2D)<br />

Selection of<br />

Data at the same<br />

time period as<br />

the acceleration<br />

data considered<br />

Data Editing &<br />

Classification<br />

Temperatures<br />

of steel deck,<br />

concrete deck,<br />

concrete<br />

tower & air<br />

Wind at<br />

Deck-Level &<br />

Tower-Top<br />

D-Str at<br />

Steel Deck<br />

Near Tower<br />

S-Str at<br />

Tower-Base<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Same Type of Derived<br />

Parameters as that<br />

in Displacement<br />

Monitoring Report<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Same Type of Plots as that in<br />

Displacement Monitoring Report<br />

References of Temperatures,<br />

Winds, Dynamic Strains &<br />

Static Strains<br />

Note: Due to the extreme heavy computational works, this monitoring work solely<br />

executes when the measured accelerations at tower-bases are equal to or exceeding<br />

0.02 g (i.e. at or above Modified Mercalli Intensity Scale 5).<br />

Figure 14 Flow Diagram of Customized Software Tools for Seismic Monitoring<br />

Platinized Titanium<br />

Cathode<br />

Anode Ladder (six<br />

Anodes) and One<br />

PT1000 Temperature<br />

Sensor<br />

Reinforcement<br />

Connection<br />

Manganese Dioxide<br />

Reference Electrode<br />

Note: <strong>The</strong> two relative humidity sensors after concreting and are not shown here.<br />

Figure 15 Typical Layout of Corrosion Cell<br />

-272-


Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Measurements<br />

& Analysis<br />

Type of Measured/Derived<br />

Parameters<br />

Type of Plots<br />

Cor<br />

Data Editing<br />

Corrosion<br />

Data<br />

Classification<br />

Corrosion Data<br />

At Tower-Bases<br />

Corrosion Data<br />

At Pier-Bases<br />

Corrosion Data<br />

At Concrete<br />

Decks<br />

Open Circuit<br />

Potential<br />

Corrosion<br />

Current<br />

AC -Resistance<br />

Linear<br />

Polarization<br />

Resistance<br />

Temperature &<br />

Relative Humidity<br />

In Concrete<br />

Corrosion Induced<br />

Deterioration<br />

Model of Concrete<br />

Open Circuit<br />

Potential<br />

Corrosion<br />

Current<br />

Concrete<br />

Resistivity<br />

Linear Polarization<br />

Resistance (LPR)<br />

Concrete<br />

Temperature<br />

Relative Humidity<br />

in Concrete<br />

Corrosion Rate<br />

(based on LPR)<br />

Corrosion Rate<br />

(Tafel Plots)<br />

Reinforcement<br />

Deterioration<br />

Crack Initiation &<br />

Propagation<br />

Time to Corrosion<br />

Cracking of Concrete<br />

Cover<br />

Plots of Corrosion Potential<br />

with indication of Critical<br />

Potential for Corrosion<br />

Plots of Corrosion Current<br />

with indication of Critical<br />

Current for Corrosion<br />

Plots of Concrete Resistivity<br />

with indication of Critical<br />

Resistance for Corrosion<br />

Plots of Linear Polarization<br />

Resistance with indication of<br />

Critical Range for Corrosion<br />

Plots of Corrosion Rate<br />

with indication of Critical<br />

Range for Corrosion<br />

Plots of Relative Humidity<br />

inside Concrete with indication<br />

of Critical Range for Corrosion<br />

Output of Corrosion Rate<br />

(based on LPR)<br />

Output of Corrosion Rate<br />

(based on Tafel Plots)<br />

Output of Reinforcement<br />

Deterioration<br />

Output of Crack Initiation &<br />

Propagation<br />

Output of Time to Corrosion<br />

Cracking of Concrete Cover<br />

Figure 16 Flow Diagram of Customized Software Tools for Corrosion Status Monitoring<br />

Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

DWIMS<br />

Influence<br />

Lines<br />

Analyzed<br />

Results &<br />

BS 5400<br />

Part 10<br />

VC<br />

GPS<br />

D-Str<br />

Data Editing<br />

Traffic Data<br />

Classification<br />

Selection of<br />

Data at the same<br />

time period as<br />

the DWIMS data<br />

considered<br />

Traffic Data<br />

Classification<br />

Conversion of<br />

Video Data into<br />

Numerical Data<br />

Selection of<br />

Data at the same<br />

time period as<br />

the DWIMS data<br />

considered<br />

Data Editing<br />

Axle-Speed<br />

Axle-Weight<br />

Condition 3 :<br />

Traffic Jammed<br />

Length ≧100 m<br />

Condition 2 :<br />

Traffic Jammed<br />

Length < 100 m<br />

Condition 1 :<br />

Free-Flow<br />

Traffic<br />

GPS Data at<br />

Deck-Levels<br />

Dynamic Strain<br />

Data at<br />

Deck-Levels<br />

Detailed Vehicular<br />

Classification<br />

Simplified Vehicular<br />

Classification<br />

Fatigue Damage<br />

Analysis based<br />

on Traffic Data<br />

Traffic Data<br />

Analysis<br />

Stress Ranges<br />

Analysis of<br />

Simplified or<br />

Detailed<br />

Vehicular Types<br />

Traffic Composition<br />

Analysis<br />

Goods Vehicles<br />

Positions Analysis<br />

Traffic Load-<br />

Effects Analysis<br />

Analysis<br />

Not Required<br />

Time-Averaging<br />

Statistics Analysis<br />

Rainflow Counts of<br />

Dynamic Strains<br />

Highway Loading<br />

Spectrum – Detailed<br />

Vehicular Types<br />

Highway Loading<br />

Spectrum – Simplified<br />

Vehicular Types<br />

Fatigue damage on<br />

each traffic lane and sum<br />

of fatigue damage in<br />

Dual-3 carriageway<br />

Percentage of<br />

Commercial Vehicles<br />

Total Daily Flow Rate<br />

(in terms of passenger<br />

car unit or pcu)<br />

Over-Weighed Axles<br />

Highway Loading<br />

Spectrum under<br />

Jammed Traffics<br />

Positions and Vehicle<br />

Weights of Goods<br />

Vehicles in Main Span<br />

Load-Effects on each<br />

pre-defined key locations<br />

Status of displacement,<br />

stress distribution &<br />

fatigue damage in<br />

pre-defined key location<br />

Highway Loading<br />

Spectrum Diagram<br />

(Detailed Classification)<br />

Highway Loading<br />

Spectrum Diagram<br />

(Simplified Classification)<br />

Fatigue Damage on each<br />

traffic lane and overall fatigue<br />

Damage on Dual-3 carriageway<br />

Plots of Monthly Maximum<br />

Total Daily Flow Rates<br />

on Each Traffic Lanes with<br />

indication of percentage of<br />

Commercial Vehicles<br />

Plots of Axle Flow in Each<br />

Traffic Lane & Carriageway with<br />

the indication of the % of<br />

over-weighed axles<br />

Highway Loading Spectrum<br />

Diagram under<br />

Jammed Traffics<br />

Tabulation of Positions and<br />

Vehicle Weights of Goods<br />

Vehicles in Main Span<br />

Output of Load-Effects on predefined<br />

key locations for each<br />

Identified Traffic Condition<br />

References of Displacement,<br />

Stress Distribution & Fatigue<br />

Damage Status in Pre-defined<br />

Key Locations<br />

Figure 17 Flow Diagram of Customized Software Tools for Highway Traffics Monitoring<br />

-273-


Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Checking/Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

Acc-F(3D)<br />

In<br />

Towers<br />

Data Editing<br />

Acceleration<br />

Data<br />

Classification<br />

Acc at<br />

Tower-Bases<br />

Acc at<br />

Tower-tops<br />

Shock Response<br />

Spectrum by<br />

Duhamel Integral<br />

Relative Displacement<br />

Spectra at Tower-Tops<br />

and Tower-Bases<br />

Relative Velocity<br />

Spectra at Tower-Tops,<br />

and Tower-Bases<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Global Bridge Deformed Shape<br />

of individual mode &<br />

Sum of modes considered<br />

Eigen<br />

Analyzed<br />

Results<br />

GPS<br />

Til<br />

Selection of<br />

Data at the same<br />

time period as<br />

the acceleration<br />

data considered<br />

Data Editing &<br />

Classification<br />

GPS at<br />

Tower-Top<br />

GPS at<br />

Deck-Level<br />

Til along<br />

Tower-Height<br />

Displacement<br />

Input Type Finite<br />

Element Analysis<br />

Absolute Acceleration<br />

Spectra at Tower-Tops,<br />

and Tower-Bases<br />

Bridge Deformed<br />

Shape and Force in<br />

each mode considered<br />

Derived Stress/Force Outputs<br />

at Each Dynamic Strain<br />

Instrumented Deck Sections<br />

or Pre-defined Key Locations<br />

Derived Stress/Force Outputs<br />

at Each Static Strain<br />

Instrumented Tower & Pier<br />

Sections or Pre-defined<br />

Key Locations<br />

TMU-1<br />

TMU-2<br />

TMU-3<br />

TMU-4<br />

TMU-6<br />

Ane(U3D)<br />

Ane(U2D)<br />

Selection of<br />

Data at the same<br />

time period as<br />

the acceleration<br />

data considered<br />

Data Editing &<br />

Classification<br />

Temperatures<br />

of steel deck,<br />

concrete deck,<br />

concrete<br />

tower & air<br />

Wind at<br />

Deck-Level &<br />

Tower-Top<br />

D-Str at<br />

Steel Deck<br />

Near Tower<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Same Type of Derived<br />

Parameters as that<br />

in Displacement<br />

Monitoring Report<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Same Type of Plots as that in<br />

Displacement Monitoring Report<br />

References of Temperatures,<br />

Winds, Dynamic Strains &<br />

Static Strains<br />

S-Str at<br />

Tower-Base<br />

Figure 18 Flow Diagram of Customized Software Tools for Ship Impacting Monitoring<br />

Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

S-Str<br />

(Tower-<br />

Bases),<br />

TMU-3<br />

SDM &<br />

BS 5400 :<br />

Part 4<br />

Data Editing<br />

Strain Data<br />

Classification<br />

Strains in<br />

Tsing Yi<br />

Tower<br />

Concrete<br />

Temperature<br />

in Tsing Yi<br />

Tower<br />

Strains in<br />

Stonecutters<br />

Tower<br />

Concrete<br />

Temperature in<br />

Stonecutters<br />

Tower<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Structural<br />

Strain<br />

Estimation<br />

Creep<br />

Strain<br />

Estimation<br />

Shrinkage<br />

Strain<br />

Estimation<br />

Concrete Tower<br />

Temperature<br />

Estimated<br />

Structural Strain<br />

Estimated<br />

Creep Strain<br />

Estimated<br />

Shrinkage Strain<br />

Total Estimated<br />

Strain with<br />

Temperature<br />

Calibration<br />

Total<br />

Estimated Shortening<br />

Total Measured<br />

Strain with<br />

Temperature<br />

Calibration<br />

Total<br />

Measured Shortening<br />

Estimated Total<br />

Creep & Shrinkage<br />

Strains in Tower<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Output of the Bi-yearly Results<br />

of Total Estimated Strains<br />

Expressed in terms of<br />

Structural Strains, Creep<br />

Strains & Shrinkage Strains at<br />

20°C Reference Temperature<br />

Output of the Bi-yearly Results<br />

of Total Measured Strains with<br />

Calibration at 20°C<br />

Reference Temperature<br />

Determine the Estimated &<br />

Measured Total<br />

Shortening in Tower<br />

Output the Bi-yearly Results of<br />

the Total Creep & Shrinkage<br />

Strains in Tower<br />

DWIMS<br />

Selection of<br />

Data at the same<br />

time period as<br />

the Strain<br />

data considered<br />

Data Editing &<br />

Classification<br />

Notes:<br />

1. SDM = Structures Design Manual published by Highways Department.<br />

2. <strong>The</strong> basic concrete data for estimation:<br />

(a) Concrete Grade 60 (i.e. Fcu = 60 MPa.<br />

(b) Instantaneous Elastic Modulus = 30200 MPa (from Table 22 of SDM).<br />

(c) Design uni-axial compression stress at SLS = 26.8 MPa (at 1518 micro-strains).<br />

Highway Traffics Condition<br />

as References<br />

Figure 19 Flow Diagram of Customized Software Tools for Permanent Loads Monitoring<br />

-274-


Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Measurements<br />

& Analysis<br />

Type of Measured/Derived<br />

Parameters<br />

Type of Plots<br />

Testing<br />

Program<br />

Truck<br />

Data<br />

GPS<br />

D-Str<br />

Analyzed<br />

Results<br />

Selection of<br />

Data at<br />

the same<br />

time period of<br />

the load test<br />

Data<br />

Classification<br />

Longitudinal<br />

Line Load<br />

Uniformly<br />

Distributed Load<br />

Transverse<br />

Line Load<br />

Moving Load<br />

Longitudinal<br />

Line Load<br />

Effects at towertops,<br />

¼, ½, & ¾<br />

of main span<br />

Uniformly<br />

Distributed Load<br />

Effects at towertops,<br />

¼, ½, & ¾<br />

of main span<br />

Knife Edge Load<br />

Effects at towertops,<br />

¼, ½, & ¾<br />

of main span<br />

Super-position of<br />

Analyzed Results<br />

Stress/forces<br />

Influence Lines<br />

Displacement<br />

Influence Lines<br />

Single Load Effects<br />

of Stresses &<br />

Displacements at<br />

tower-tops, ¼, ½, & ¾<br />

of main span for<br />

each load test<br />

(Measurement)<br />

Multi-Load Effects<br />

of Stresses &<br />

Displacements at<br />

tower-tops, ¼, ½, & ¾c<br />

of main span for<br />

combined load-tests<br />

(by Super-position<br />

-Simulation)<br />

Displacement Ranges<br />

at Instrumented Points<br />

Stress Ranges<br />

at Instrumented Points<br />

Load-Deflection Plots of<br />

Longitudinal Line Load-Effects<br />

at respective tower-tops,<br />

¼,1/2, & ¾ main span<br />

Load-Deflection Plots of<br />

Uniformly Distributed Load-<br />

Effects at respective tower-tops,<br />

¼, ½, &3/4 main span<br />

Load-Deflection Plots of<br />

Transverse Line Load-Effects<br />

at respective tower-tops,<br />

¼,1/2, & ¾ main span<br />

Load-Stress/Moment Plots of<br />

Longitudinal Line Load-Effects<br />

In deck sections at towers,<br />

¼,1/2, & ¾ main span<br />

Load-Stress/Moment Plots of<br />

Uniformly Distributed Load-<br />

Effects in deck sections at<br />

Towers, ¼, ½, &3/4 main span<br />

Load-Stress/Moment Plots of<br />

Transverse Line Load-Effects<br />

In deck sections at towers,<br />

¼,1/2, & ¾ main span<br />

Displacement Influence Lines<br />

at each GPS and comparison<br />

with Analyzed Results<br />

Stress Influence Lines at<br />

each D-Str (in deck-trough<br />

section) and comparison<br />

with Analyzed Results<br />

Figure 20 Flow Diagram of Customized Software Tools for Global Static Features Monitoring<br />

A A<br />

3<br />

6<br />

A 2<br />

A 5<br />

A 1 A 4<br />

Accelerometer<br />

B<br />

Vertical Acceleration =<br />

Lateral Acceleration =<br />

A3 + 6<br />

A<br />

2<br />

A2 + 5<br />

A<br />

2<br />

Torsional Acceleration = TAN -1 | A A6|<br />

B<br />

3 −<br />

Accelerometer<br />

Longitudinal Acceleration =<br />

A 4<br />

1+ A<br />

2<br />

A<br />

i = Direction of time-series acceleration, where i = 1, 2, 3, 4, 5, 7 & 6.<br />

Note: All 3D-Accelerometers at a deck section must be installed in the same vertical plane.<br />

Figure 21 Arrangement of Accelerometers in Steel Deck Section<br />

-275-


Bridge-Deck<br />

Alignment<br />

A6<br />

A5<br />

A4<br />

D<br />

A3<br />

A2<br />

A1<br />

Bridge-Deck<br />

Alignment<br />

Longitudinal Acceleration of Tower =<br />

Lateral Acceleration of Tower =<br />

Torsional Acceleration of Tower =<br />

Vertical Acceleration of Tower =<br />

A<br />

1<br />

+ A 4<br />

2<br />

A<br />

2<br />

+ A 5<br />

2<br />

TAN-1<br />

A<br />

2<br />

- A<br />

5<br />

D<br />

A<br />

3<br />

+ A 6<br />

2<br />

(when the data is<br />

significant)<br />

A i = Acceleration time-history data record where i = 1, 2, 3, 4, 5, and 6.<br />

Note: All 3D-Accelerometers at a deck section must be installed in the same vertical plane.<br />

Figure 22 Arrangement of Accelerometers in Tower Section<br />

Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

Vertical<br />

Acceleration<br />

Data<br />

Linear<br />

Spectrum<br />

Analysis<br />

Vertical Flexural<br />

Deck Modes<br />

Acc-F(3D)<br />

at one or<br />

more<br />

Sections<br />

Eigen<br />

Analyzed<br />

Results<br />

TMU-1<br />

TMU-2<br />

TMU-3<br />

TMU-4<br />

TMU-6<br />

TMU-7<br />

Ane (U3D)<br />

Ane (U2D)<br />

Data Editing<br />

Acceleration<br />

Data Arithmetic<br />

Processing<br />

Selection of<br />

Data at the same<br />

time period as<br />

the acceleration<br />

data considered<br />

Data Editing &<br />

Classification<br />

Lateral<br />

Acceleration<br />

Data<br />

Torsional<br />

Acceleration<br />

Data<br />

Longitudinal<br />

Acceleration<br />

Data<br />

Steel<br />

Deck-Section<br />

Temperature<br />

Concrete<br />

Deck-Section<br />

Temperature<br />

Concrete<br />

Tower-Section<br />

Temperature<br />

Asphalt<br />

Pavement<br />

Temperature<br />

Air<br />

Temperature<br />

Curve<br />

Fitting<br />

Analysis<br />

Power<br />

Spectrum<br />

Analysis<br />

Coherence<br />

Spectrum<br />

Analysis<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Lateral Flexural<br />

Deck Modes<br />

Torsional Flexural<br />

Deck Modes<br />

Longitudinal<br />

Flexural Deck Modes<br />

Lateral Flexural<br />

Tower Modes<br />

Torsional Flexural<br />

Tower Modes<br />

Longitudinal<br />

Flexural Tower Modes<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Linear Spectrum Diagrams<br />

(Real-Part)<br />

Linear Spectrum Diagrams<br />

(Imaginary-Part)<br />

Power Spectrum Diagrams<br />

(with Frequencies Extraction<br />

by Peak-Picking)<br />

Coherence Spectrum Diagrams<br />

(between 2 Sections)<br />

Mode Shapes Determination<br />

(by Curve Fitting of the Peaks<br />

of the Imaginary Part of the<br />

Frequency Response Function)<br />

Comparison Plots of Measured<br />

Frequencies and Analyzed<br />

Frequencies (by Finite Element<br />

Analysis – either ANSYS /<br />

Multiphysics or MD/NASTRAN)<br />

Temperature References in<br />

Steel Deck-Section, Concrete<br />

Deck-Section, Concrete<br />

Tower-Section, Asphalt<br />

Pavement and Air<br />

Wind References at<br />

Deck-Levels and Tower-Tops<br />

Deck &<br />

Tower Winds<br />

Figure 23 Flow Diagram of Customized Software Tools for Global Dynamic Features Monitoring<br />

-276-


Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

Ambient<br />

Vibration<br />

Analysis<br />

In-Plane<br />

Flexural<br />

Frequencies<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Acc-R(3D)<br />

at one<br />

or two<br />

Sections<br />

Design &<br />

As-Built<br />

Stay Cable<br />

Values<br />

TMU-5<br />

TMU-6<br />

Ane (U3D)<br />

Ane (U2D)<br />

Data Editing<br />

Acceleration<br />

Data Arithmetic<br />

Processing<br />

Selection of<br />

Data at the same<br />

time period as<br />

the acceleration<br />

data considered<br />

Data Editing &<br />

Classification<br />

Vertical<br />

Acceleration<br />

Data<br />

Lateral<br />

Acceleration<br />

Data<br />

Longitudinal<br />

Acceleration<br />

Data<br />

Stay Cable<br />

Temperature<br />

Air<br />

Temperature<br />

Deck &<br />

Tower Winds<br />

Power<br />

Spectrum<br />

Analysis<br />

Coherence<br />

Spectrum<br />

Analysis<br />

Force-Free<br />

Vibration<br />

Analysis<br />

Decay Motion<br />

Identification<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Out-of-Plane<br />

Flexural<br />

Frequencies<br />

Longitudinal<br />

Flexural<br />

Frequencies<br />

Logarithmic<br />

Decrement<br />

Scruton<br />

Number<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Power Spectrum Diagrams<br />

(with Frequencies Extraction<br />

by Peak-Picking)<br />

Coherence Spectrum Diagrams<br />

(between 2 Sections)<br />

Plots of Derived Stay Cable<br />

Frequencies along the Bridge-<br />

Spans & Comparison with<br />

Design & As-Built Values for<br />

checking of Parametric<br />

Effects with Deck & Towers<br />

Plots of Derived Stay Cable<br />

Forces along the Bridge-<br />

Spans & Comparison with<br />

Design & As-Built Values<br />

Plots of Stay Cable<br />

Damping Ratios along the<br />

Bridge-Spans<br />

Plots of Stay Cable<br />

Scruton Numbers along the<br />

Bridge-Spans for checking of<br />

Stay Cable Stability against<br />

Rain-Wind Induced Vibrations<br />

Temperature References in<br />

Stay Cable and Air<br />

Wind References at<br />

Deck-Levels and Tower-Tops<br />

Figure 24 Flow Diagram of Customized Software Tools for Stay Forces Monitoring<br />

Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

Ten<br />

Data Editing<br />

Ten Data<br />

Classification<br />

Tsing Yi<br />

Data<br />

Stonecutters<br />

Data<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Correlation<br />

Analysis with<br />

Wind Speeds &<br />

Temperature<br />

Tendon Force<br />

Correlation of Tendon<br />

Force & Wind Speeds<br />

or Temperature<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Tendon Force Distribution<br />

in Tsing Yi Deck<br />

Tendon Force Distribution<br />

in Stonecutters Deck<br />

Correlation Plots of Tendon<br />

Force and Wind Speeds<br />

Tendon<br />

Design<br />

Values<br />

Correlation Plots of Tendon<br />

Force and Temperature<br />

Hourly, Daily, Monthly &<br />

Annual Variation of<br />

Individual Tendon<br />

TMU-4<br />

TMU-6<br />

Ane (U3D)<br />

Ane (U2D)<br />

Selection of<br />

Data at the same<br />

time period as<br />

the acceleration<br />

data considered<br />

Data Editing &<br />

Classification<br />

Air & Deck<br />

Temperature<br />

Deck &<br />

Tower Winds<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Temperature References in<br />

Air and Concrete Deck<br />

Wind References at<br />

Deck-Levels and Tower-Tops<br />

Figure 25 Flow Diagram of Customized Software Tools for Tendon Forces Monitoring<br />

-277-


D 3<br />

D 1<br />

D 2<br />

D 6<br />

D 4<br />

D 5<br />

B<br />

Vertical Displacement =<br />

Lateral Displacement =<br />

Torsional Displacement = TAN -1<br />

Longitudinal Displacement =<br />

D3 + D6<br />

2<br />

D2 + S5<br />

2<br />

| D3 −D6|<br />

B<br />

D1 + D4<br />

2<br />

D i<br />

Note: All GPS at a deck section must be installed in the same vertical plane.<br />

= Direction of time-series GPS data, where i = 1, 2, 3, 4, 5, 7 & 6.<br />

Figure 26 Arrangement of GPS in Steel Deck Section<br />

Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

GPS,<br />

Til<br />

Influence<br />

Lines<br />

Analyzed<br />

Results<br />

Correction of<br />

Abrupt Changes<br />

Spikes & Noises<br />

Removal<br />

Displacement<br />

Data<br />

Classification<br />

Deck at ¼<br />

Main Span<br />

Deck at ½<br />

Main Span<br />

Deck at ¾<br />

Main Span<br />

Tower at<br />

1 H<br />

Tower at<br />

2/3 H<br />

Tower at<br />

1/3 H<br />

Tower at<br />

0/3 H<br />

Upper Piers<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Curve Fitting<br />

Analysis of<br />

Deck & Tower<br />

Profiles<br />

Displacement<br />

Input Type<br />

Finite Element<br />

Analysis<br />

Displacement<br />

Influence Lines<br />

Validation<br />

in Load Trials<br />

Hourly, Daily, Monthly<br />

Yearly Variations of<br />

Displacements at<br />

GPS and Til<br />

Instrumented Sections<br />

Deformed Line-Profiles<br />

of Deck and Towers<br />

Global Bridge<br />

Deformed Shape with<br />

Strain and Stress<br />

Contours<br />

Derived Stresses at<br />

Dynamic Strain<br />

Instrumented<br />

Steel Deck Sections<br />

Derived Stresses at<br />

Static Strain<br />

Instrumented<br />

Tower-Base and<br />

Pier-Base Sections<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Deformed Line-Profiles of<br />

Deck and Towers<br />

Global Bridge Deformed<br />

Shape with Strain and Stress<br />

Contours Illustration<br />

Derived Stress/Force Outputs<br />

at Each Dynamic Strain<br />

Instrumented Deck Sections<br />

or Pre-defined Key Locations<br />

Derived Stress/Force Outputs<br />

at Each Static Strain<br />

Instrumented Tower & Pier<br />

Sections or Pre-defined<br />

Key Locations<br />

(under Strong Lateral Loads)<br />

Plots of Measured & Analyzed<br />

Displacement Influence Lines<br />

at GPS Instrumented<br />

Deck & Tower Sections<br />

Ane (U3D)<br />

Ane (U2D)<br />

Ane (P2D)<br />

TMU-6<br />

D-Str<br />

S-Str<br />

Acc-F(3D)<br />

Selection of<br />

Data at the same<br />

Time period as<br />

the GPS & Til<br />

data considered<br />

Data Editing &<br />

Classification<br />

Deck &<br />

Tower Winds<br />

Air<br />

Temperature<br />

Dynamic &<br />

Static Strains<br />

Deck<br />

Accelerations<br />

Tower<br />

Accelerations<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Correlation of<br />

Measured & Analyzed<br />

Displacement<br />

Influence Lines at<br />

GPS instrumented<br />

Deck & Tower Sections<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Comparison Plots of<br />

Derived Stresses &<br />

Measured Stresses<br />

Wind References at<br />

Deck-Levels & Tower-Tops<br />

Acceleration References at<br />

Deck & Tower Sections<br />

Temperature References<br />

in Air<br />

Figure 27 Flow Diagram of Customized Software Tools for Displacements Monitoring<br />

-278-


Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

D-Str<br />

Data Editing<br />

Spikes & Noises<br />

Removal<br />

Strain Data<br />

Classification<br />

Longitudinal<br />

Steel Girder<br />

Cross<br />

Steel Girder<br />

Tower<br />

Steel Skins<br />

Strain<br />

Conversions<br />

Stress Demand<br />

Ratios<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Correlation<br />

Analysis of<br />

Strain/Stress<br />

vs Wind Speeds<br />

Vs GPS or Til<br />

Individual Stress<br />

History Dialog<br />

Stress Distribution<br />

Across a Component<br />

or Section<br />

Stress Demand Ratio<br />

Correlation of<br />

Stress – Wind – GPS<br />

Or Til<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Individual Stress History<br />

Dialog Plot<br />

Plot of Stress Distributions<br />

Stress Demand Ratio Plots<br />

Correlation Plots of Stress vs<br />

Wind vs GPS/Til<br />

Correlation Plots of Stress vs<br />

Acceleration<br />

Correlation<br />

Analysis of<br />

Strain/Stress<br />

vs Acceleration<br />

Correlation of<br />

Stress –Acceleration<br />

Ane (U3D)<br />

Ane (U2D)<br />

Ane (P2D)<br />

TMU-1<br />

TMU-2<br />

TMU-6<br />

GPS<br />

Til,<br />

Acc-3D(F)<br />

Selection of<br />

Data at the same<br />

time period<br />

as the Strain<br />

data considered<br />

Data Editing &<br />

Classification<br />

Air<br />

Temperature<br />

Steel Deck<br />

Temperature<br />

Deck &<br />

Tower Winds<br />

GPS & Til<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Wind References at<br />

Deck-Levels & Tower-Tops<br />

Temperature References<br />

in Air and Steel Deck<br />

Figure 28 Flow Diagram of Customized Software Tools for Stress/Force Distribution Monitoring<br />

Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

Steel Deck<br />

Trough<br />

Strain<br />

Conversions<br />

D-Str<br />

Data Editing<br />

Spikes & Noises<br />

Removal<br />

Deck-Cable<br />

Anchorage<br />

Stress Demand<br />

Ratios<br />

Stress Levels<br />

Selection<br />

Stress Level<br />

Distribution<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Stress Demand Ratio Plot<br />

Strain Data<br />

Classification<br />

Tower-Cable<br />

Anchorage<br />

Determination<br />

of Endurance<br />

N Cycles for<br />

each Stress<br />

Level selected –<br />

by Clause 11<br />

of BS 5400 :<br />

Part 10<br />

Fatigue Damage<br />

Derived Fatigue Life<br />

& check with<br />

Design Value<br />

Fatigue Damage and Fatigue<br />

Life Estimation Plots<br />

Rainflow<br />

Count<br />

Ane (U3D)<br />

Ane (U2D)<br />

Ane (P2D)<br />

TMU-1<br />

TMU-2<br />

TMU-6<br />

Selection of<br />

Data at the same<br />

time period<br />

as the D-Str<br />

data considered<br />

Data Editing &<br />

Classification<br />

Air<br />

Temperature<br />

Steel Deck<br />

Temperature<br />

Deck &<br />

Tower Winds<br />

<strong>The</strong> Palmgren-<br />

Miner Rule<br />

Fatigue Damage<br />

& Fatigue Life<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Wind References at<br />

Deck-Levels & Tower-Tops<br />

Temperature References<br />

in Air and Steel Deck<br />

Figure 29 Flow Diagram of Customized Software Tools for Fatigue Damage Monitoring<br />

-279-


Type of<br />

Input Data<br />

Type of Data<br />

Processing<br />

Type of Data<br />

Classification<br />

Type of Data<br />

Analysis<br />

Type of Derived<br />

Parameters<br />

Type of Plots<br />

Brg<br />

Data Editing<br />

Bearing Data<br />

Classification<br />

Lateral<br />

Displacements<br />

Bearing<br />

Force, N<br />

Bearing<br />

Vertical<br />

Moment, M V<br />

Bearing<br />

Horizontal<br />

Moment, M H<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Displacement &<br />

Effective Friction<br />

Coefficient<br />

Analysis<br />

Correlation<br />

Analysis with<br />

Typhoon Data<br />

Lateral Displacements<br />

Lateral Forces Transfer<br />

between Tower & Deck<br />

Effective Friction<br />

Coefficient<br />

Correlation of Wind<br />

Speeds vs Lateral<br />

Displacement &<br />

Lateral Force<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Plots of Lateral Displacement<br />

vs Wind Speeds<br />

Plots of Lateral Force Transfer<br />

Vs Wind Speeds<br />

Output of Effective Friction<br />

Coefficient<br />

Buf<br />

Ane (U3D)<br />

Ane (U2D)<br />

Ane (P2D)<br />

TMU-1<br />

TMU-2<br />

TMU-6<br />

Data Editing<br />

Buffer Data<br />

Classification<br />

Selection of<br />

Data at the same<br />

time period<br />

as Brg & Buf<br />

data considered<br />

Data Editing &<br />

Classification<br />

Pressures in<br />

Buffers<br />

Pressures in<br />

Accumulator<br />

Tank<br />

Piston-<br />

Positions<br />

in Buffers<br />

Temperatures<br />

Inside Buffers<br />

Air<br />

Temperature<br />

Steel Deck<br />

Temperature<br />

Deck &<br />

Tower Winds<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Correlation<br />

Analysis with<br />

Temperatures<br />

Time-Averaging<br />

Statistics<br />

Analysis<br />

Buffer Forces in<br />

Main Span Buffers<br />

Buffer Forces in<br />

Back Span Buffers<br />

Pressure at<br />

Accumulator Tank<br />

Buffer Displacements<br />

in Main Span<br />

Temperatures inside<br />

Buffers<br />

Maximum, Mean,<br />

Minimum, Standard<br />

Deviation, etc.<br />

Time-Series Plots of All Raw<br />

& Processed Data<br />

Plots of Buffer Forces in<br />

Main & Side Spans<br />

Plots of Longitudinal Displace-<br />

-ments in Main & Side Spans<br />

Plot of Temperatures<br />

inside Buffers<br />

Plot of Buffer Displacements vs<br />

Steel Deck Temperatures<br />

Wind References at<br />

Deck-Levels & Tower-Tops<br />

Temperature References<br />

in Air and Steel Deck<br />

Figure 30 Flow Diagram of Customized Software Tools for Articulation Responses Monitoring<br />

Data Stored in Rolling Buffer Storage System controlled by SHDMS-Server<br />

Wind & Weather Data<br />

Temperature Data<br />

Seismic Data<br />

Corrosion Status Data<br />

Highway Traffics Data<br />

Ship Impacting Data<br />

Permanent Loads Data<br />

Global Static Features Data<br />

Global Dynamic Features Data<br />

Stay Forces Data<br />

Tendon Forces Data<br />

Displacements Data<br />

Stress/Force Distribution Data<br />

Fatigue Damage Data<br />

Articulation Responses Data<br />

Software System<br />

for Data<br />

Interfacing &<br />

Analysis Execution<br />

(operating in<br />

SHDMS-Server)<br />

Wind & Weather<br />

Monitoring Software Tools<br />

Temperature<br />

Monitoring Software Tools<br />

Seismic<br />

Monitoring Software Tools<br />

Corrosion Status<br />

Monitoring Software Tools<br />

Highway Traffics<br />

Monitoring Software Tools<br />

Ship Impacting<br />

Monitoring Software Tools<br />

Permanent Loads<br />

Monitoring Software Tools<br />

Global Static Features<br />

Monitoring Software Tools<br />

Global Dynamic Features<br />

Monitoring Software Tools<br />

Stay Forces<br />

Monitoring Software Tools<br />

Tendon Forces<br />

Monitoring Software Tools<br />

Displacements<br />

Monitoring Software Tools<br />

Stress/Force Distribution<br />

Monitoring Software Tools<br />

Fatigue Damage<br />

Monitoring Software Tools<br />

Articulation Responses<br />

Monitoring Software Tools<br />

Software System for Bridge Structural System in DPCS-Workstations<br />

Figure 31 Required types of data and tools for data interfacing and analysis execution<br />

-280-


Tsing Yi<br />

Stonecutters<br />

Stonecutters<br />

Bridge<br />

4W 3W 2W 1W<br />

Rambler Channel<br />

North<br />

SCB Location Plan<br />

Ane (U3D)_AVS_SD(TP)_6<br />

South<br />

Wind Speeds at POE=2.5%<br />

Wind Speeds at POE=5.0%<br />

Wind Speeds at POE=10%<br />

315<br />

Wind Speeds at POE=20%<br />

Wind Speeds at POE=40%<br />

300<br />

Wind Speeds at POE=60%<br />

Wind Speeds at POE=80%<br />

Bridge Alignment<br />

285<br />

West 270<br />

330<br />

345<br />

North<br />

0<br />

15<br />

30<br />

45<br />

60<br />

75<br />

2 4 6 8 10m/s<br />

Key Information Extracted:<br />

Wind Rose Diagram for 3-<br />

second gust wind<br />

Wind Duration for the whole<br />

Feb 2011<br />

Mean wind speed: 2.93 m/s<br />

Prevailing wind direction: SSW<br />

Design wind speed: 50 m/s<br />

(hourly)<br />

90 East<br />

Sensor Location :<br />

Deck Level<br />

Sensor Tag-Name :<br />

Ane(U3D)_AVS_SD(TP)_7<br />

Time Reference of Time-Series Date: 2011-02-01~2011-02-<br />

Data Collected :<br />

28<br />

Time Started: 00:00:00<br />

Time Ended: 23:59:59<br />

Duration of Time-Series Wind 28 days<br />

Data Analyzed :<br />

Location of Time-Series Data Deck Segment No. 35<br />

Collected :<br />

Sampling Rate :<br />

50 Hz<br />

Number of Raw Data Points : 12,096,000<br />

Type of Time Average :<br />

3-second<br />

255<br />

105<br />

Number of Averaged Data<br />

Points :<br />

Mean Wind Speed :<br />

806,400<br />

2.9289 m/s<br />

240<br />

120<br />

Prevailing Wind Direction :<br />

SSW<br />

225<br />

135<br />

210<br />

195<br />

180<br />

South<br />

165<br />

150<br />

POE=Probability of Exceedence<br />

Figure 32 Wind Rose Diagram<br />

from a Ane (U3D) in Southwest deck at mid main span<br />

Wind Power Spectrum and Integral Length Scale<br />

Key Information Extracted:<br />

Duration: 2010-10-21 08:00 to 10:00<br />

Mean Wind Speed, U= 6.27 m/s (10-min.)<br />

Turbulence Intensity, y Iu= 20.1%<br />

Integral Length Scale, y Lu = 75.2 m<br />

Criteria for Checking Buffeting Response<br />

L Alarm = 300m<br />

f M =0.0047Hz<br />

f VK =0.0133Hz<br />

Figure 33 Wind Power Spectrum and corresponding Integral Length Scale determination<br />

(from a Ane(U3D) in Northwest deck at mid main span)<br />

-281-


DWIMS on<br />

Tsing Yi Side<br />

DWIMS on<br />

Stonecutters Side<br />

Highway Traffic Loading Spectrum Diagram in Individual Traffic Lane<br />

SCB Location Plan<br />

180,000<br />

160,000<br />

140,000<br />

120,000<br />

No. of Vehicles<br />

100,000<br />

80,000<br />

60,000<br />

40,000<br />

20,000<br />

0<br />

slow lane middle lane fast lane fast lane middle lane slow lane<br />

Stonecutters Bound<br />

Tsing Yi Bound<br />

Figure 12 Highway Traffic Composition Monitoring<br />

Figure 34 Highway Traffic Loading Spectrum in Individual Traffic Lane<br />

DWIMS on<br />

Tsing Yi Side<br />

DWIMS on<br />

Stonecutters Side<br />

Equivalent (Fatigue) Damage Ratio in Individual Traffic Lane<br />

SCB Location Plan<br />

Damage Ratio Contributed by Individual Traffic Lane<br />

0.400<br />

0.350<br />

0.300<br />

0.250<br />

0.200<br />

0.150<br />

0.100<br />

0.050<br />

0.000<br />

0.380<br />

0.104<br />

0.011<br />

0.000<br />

0.137<br />

0.288<br />

slow lane middle lane fast lane fast lane middle lane slow lane<br />

Stonecutters Bound<br />

fatigue point<br />

for assessment<br />

Sum of Damage Ratios<br />

in 6 Traffic Lanes,<br />

6<br />

∑<br />

where i = no. of traffic lanes<br />

i<br />

=<br />

1<br />

F<br />

i =<br />

0 .1714<br />

Tsing Yi Bound<br />

Figure 35 Equivalent (Fatigue) Damage Ratio in Individual Traffic Lane with reference to the Fatigue<br />

Point for Assessment<br />

-282-


青 衣<br />

Tsing Yi<br />

昂 船 洲<br />

Stonecutters<br />

昂 船 洲 大 桥<br />

Stonecutters<br />

Bridge<br />

at North Side<br />

4W<br />

3W<br />

2W<br />

Traveling<br />

direction<br />

1W<br />

藍 巴 勒 海 峽<br />

Rambler Channel<br />

GPS04<br />

SCB Location Plan<br />

Basic Information<br />

Vehicle used<br />

41 Ton<br />

No. of vehicle 2<br />

Speed<br />

15 kph<br />

Lane used 7 and 8<br />

106 mm<br />

Figure 36<br />

Measured and analysis displacement influence lines at GPS instrumented point: GPS04<br />

obtained in the load-trial carried out before the bridge opening to public traffics<br />

青 衣<br />

Tsing Yi<br />

昂 船 洲<br />

Stonecutters<br />

昂 船 洲 大 桥<br />

Stonecutters<br />

Bridge<br />

at North Side<br />

+ +<br />

_ _<br />

Longitudinal Stress (MPa)<br />

4W<br />

16<br />

14<br />

12<br />

10<br />

8<br />

6<br />

4<br />

2<br />

0<br />

-2<br />

-4<br />

3W<br />

2W<br />

Traveling<br />

direction<br />

1W<br />

藍 巴 勒 海 峽<br />

Rambler Channel<br />

Distance from SEGMENT midspan(m) NO.32<br />

South<br />

-800 -700 -600 -500 -400 -300 -200 -100 0 100 200 300 400 500 600 700 800<br />

SW60X<br />

78.00<br />

68.25<br />

58.50<br />

48.75<br />

39.00<br />

29.25<br />

19.50<br />

-19.50<br />

3470 3570 3670 3770 3870 3970 4070 4170 4270 4370 4470 4570 4670 4770 4870 4970 5070<br />

Chainage(m)<br />

Test<br />

Analysis<br />

North<br />

9.75<br />

0.00<br />

-9.75<br />

SCB Location Plan<br />

Basic Information<br />

Vehicle used<br />

Longitudinal Strain(με)<br />

No. of vehicle 2<br />

Speed<br />

41 Ton<br />

15 kph<br />

Lane used 7 and 8<br />

Figure 37<br />

Measured and analysis stress influence lines at dynamic strain gauge instrumented point:<br />

SW60X obtained in the load-trial carried out before the bridge opening to public traffics<br />

-283-


青 衣<br />

Tsing Yi<br />

昂 船 洲<br />

Stonecutters<br />

昂 船 洲 大 桥<br />

Stonecutters<br />

Bridge<br />

at North &<br />

South Side<br />

4W<br />

3W<br />

2W<br />

1W<br />

藍 巴 勒 海 峽<br />

Rambler Channel<br />

North<br />

SCB Location Plan<br />

SEGMENT NO.32<br />

South<br />

Sensor Location :<br />

Sensor Tag-Name :<br />

Deck Level<br />

Deck32_N+Deck32N_S - FDAU<br />

Channel<br />

2/s4 /Hz)<br />

Date of time-series data collected: 2011-03-24<br />

Duration: 10:39:00 to 14:39:00 (4 hours)<br />

Duration per frame = 30 mins<br />

Time Reference of Time-Series<br />

Data Collected :<br />

Duration of Time-Series<br />

Acceleration Data Analyzed :<br />

Location of Time-Series Data<br />

Collected :<br />

Date: 2011-03-24 Time Started:<br />

10:39:00<br />

Time Ended: 14:39:00<br />

4 Hours<br />

Chainage : 42+6390 Mid Main Span<br />

Deck Segment No. 32<br />

Vertical (Z) Power Spectrum (m<br />

Type of Window for Data<br />

Hanning Window<br />

Sampling :<br />

Data Ov erlapping :<br />

50%<br />

Duration of 1 Frame :<br />

30 Minutes<br />

Sampling Rate :<br />

102.4 Hz<br />

Number of Data Points Per Data 184,320<br />

Frame :<br />

Number of Data Frames<br />

15<br />

Considered :<br />

Data Filtering :<br />

De-trendi ng<br />

Frequency Resolution (= Ite m 8 / 0.000555556 Hz<br />

Item 9) :<br />

Method of Acceleration Data Fast Fourier Transform (FFT)<br />

Analysis :<br />

Method of Frequencies<br />

Extraction :<br />

Power Spectrum with Peak-Picking<br />

Mean Wind Speed on Deck-Level : 2.05 m/s (10-Minute Mean)<br />

Mean Air Temperature / Mean<br />

Steel Temperature :<br />

16.4 Deg. C / 17.5 Deg. C<br />

Analysed<br />

Frequency (Hz)<br />

Measured<br />

Frequency (Hz)<br />

A01<br />

M01<br />

A03<br />

A02 A04<br />

M02<br />

M03<br />

M04<br />

A05<br />

M05<br />

A07 A08<br />

A13 A15<br />

A06 A09 A10 A11 A12 A14 A16<br />

M06<br />

M07<br />

M08<br />

M09<br />

M10<br />

M11<br />

M12<br />

A17 A18 A19<br />

M13<br />

A20<br />

M14<br />

A21 A23<br />

A22<br />

M15<br />

Mean Relative Humidity :<br />

Software Tools :<br />

Name of Acceleration Data<br />

Logger Used :<br />

98.4 %<br />

MATLAB Signal Processing &<br />

Analysis<br />

TMB-FDAU (16-Bi t Sequential Data<br />

Logger)<br />

Figure 38 Vertical Flexural Deck Frequencies Extraction by Power Spectrum<br />

青 衣<br />

Tsing Yi<br />

昂 船 洲 大 桥<br />

Stonecutters<br />

Bridge<br />

at North &<br />

South Side<br />

昂 船 洲<br />

Stonecutter<br />

s<br />

4W<br />

3W<br />

2W<br />

1W<br />

藍 巴 勒 海 峽<br />

Rambler Channel<br />

North<br />

SCB Location Plan<br />

Sensor Location :<br />

Deck Level<br />

SEGMENT NO.32<br />

South<br />

Sensor Tag-Name :<br />

Time Reference of Time-Series<br />

Data Collected :<br />

Deck32_N+Deck32N_S - FDAU<br />

Channel<br />

Date: 2011-03-24 Time Started:<br />

10:39:00<br />

Time Ended: 14:39:00<br />

Duration of Time-Series<br />

Acceleration Data Analyzed :<br />

4 Hours<br />

Location of Time-Series Data<br />

Collected :<br />

Chainage : 42+6390 Mid Main Span<br />

Deck Segment No. 32<br />

Type of Window for Data<br />

Sampling :<br />

Hanning Window<br />

Data Ov erlapping :<br />

50%<br />

Duration of 1 Frame :<br />

30 Minutes<br />

Sampling Rate :<br />

102.4 Hz<br />

Number of Data Points Per Data<br />

Frame :<br />

184,320<br />

Number of Data Frames<br />

Considered :<br />

15<br />

Data Filtering :<br />

De-trendi ng<br />

Frequency Resolution (= Ite m 8 /<br />

Item 9) :<br />

0.000555556 Hz<br />

Method of Acceleration Data<br />

Analysis :<br />

Fast Fourier Transform (FFT)<br />

Method of Frequencies Extraction : Power Spectrum with Peak-Picking<br />

Mean Wind Speed on Deck-Level : 2.05 m/s (10-Minute Mean)<br />

Mean Air Temperature / Mean<br />

Steel Temperature :<br />

Mean Relative Humidity :<br />

Software Tools :<br />

16.4 Deg. C / 17.5 Deg. C<br />

98.4 %<br />

MATLAB Signal Processing &<br />

Analysis<br />

Name of Acceleration Data Logger TMB-FDAU (16-Bi t Sequential Data<br />

Used :<br />

Logger)<br />

Figure 39 Comparison of Measured and Analyzed Vertical Flexural Deck Frequencies<br />

-284-


Accelerometer<br />

Fixing of Portable<br />

Accelerometer in<br />

Stay Cable<br />

Figure 40 Stay Cable with Accelerometer installed and Measurement Results<br />

Tsing Yi Tower – Stay Cables in North Side of Main Span<br />

Engineer’s computed cable force at completion state in kN<br />

Highways Department’s measured cable force at completion state in kN<br />

Contractor’s computed cable force at completion state in kN<br />

Contractor’s measured cable force at completion state in kN<br />

Figure 41 Comparison of Measured Results with Computed Results<br />

-285-


Stonecutters Bridge Elevation View<br />

Figure 42 Derived Displacement Profile with Contours in Stonecutters Bridge basing on GPS Input Data<br />

Stonecutters Bridge Elevation View<br />

Figure 43 Derived Stress in Deck Trough Sections in Mid Main Span of Stonecutters Bridge based on<br />

GPS Input Data<br />

-286-


Tsing Yi<br />

Stonecutters<br />

Bridge<br />

Stonecutt<br />

ers<br />

at North Side<br />

4<br />

W<br />

3<br />

W<br />

2<br />

W<br />

1<br />

W<br />

Rambler Channel<br />

North<br />

SCB Location Plan<br />

SEGMENT NO.32<br />

South<br />

+ + Tension<br />

_ _<br />

Compression<br />

Percentage Ocurrance<br />

Computed Stress Point / Location<br />

Period of time-series data collected: Mar 2011<br />

0 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2 0.225 0.25 0.275<br />

Demand Ratio<br />

0 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0 13.5 15.0 16.5<br />

Stress (MPa)<br />

Figure 44 Stress Demand Ratio Plots<br />

Tsing Yi<br />

Stonecutters<br />

Stonecutters<br />

Bridge<br />

at North Side<br />

4W<br />

3W<br />

2W<br />

1W<br />

Rambler Channel<br />

North<br />

SCB Location Plan<br />

SEGMENT NO.32<br />

South<br />

BS 5400 : Part 10 σ r – N curves :<br />

Class F<br />

Class F2<br />

Class G<br />

Half-Cycles Counts<br />

Measured Result<br />

CALTRANS Allowable Fatigue Stress Envelope<br />

Fatigue Evaluation Point / Location<br />

Period of time-series data collected: Mar 2011<br />

STRESS<br />

STRESS (MPa)<br />

Figure 45 Fatigue Induced Damage Monitoring<br />

-287-


Locations of Static Strain<br />

Gauges Considered<br />

East Tower<br />

Pre-camber<br />

Direction of<br />

Stonecutters<br />

Tower under<br />

construction<br />

Main Span Side<br />

B<br />

Stonecutters<br />

Tower<br />

Stonecutters Side<br />

A<br />

Main Span Side<br />

(West)<br />

Stonecutters Side<br />

(East)<br />

Figure 46 Arrangement of Static Strain Gauges for Permanent Loads Monitoring in Stonecutters Tower<br />

Measurement and Predict Strains at Stonecutter Tower-Bases<br />

C om pressive S train in M icro-S trains<br />

(at 20 degrree C)<br />

800<br />

700<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

Location A (MS) Location B (MS) Location A (AS) Location B (AS)<br />

Completion of Tower<br />

(13 Mar 2008)<br />

0 10 20 30 40 50 60 70<br />

Closure of Bridge Deck<br />

(19 Mar 2009)<br />

Number of Months after Concreting<br />

Openning of Bridge to Public Traffic<br />

(20 Dec 2009)<br />

Note: MS = Measured Strain<br />

AS = Analyzed Strain<br />

Figure 47 Concrete Strain Histories at Locations A and B (as shown in Figure 46 above) at Stonecutters<br />

Tower Base<br />

-288-


Structural Health Data Management System<br />

Data<br />

Sources<br />

DPSC-1<br />

Server<br />

DPCS-2<br />

Server<br />

PDAS-1<br />

Extract<br />

Extract<br />

Extract<br />

Data Warehouse<br />

Operational<br />

Data Store<br />

Load<br />

Data Operation<br />

• Data Storage<br />

•Relational<br />

•Fast Retrieval<br />

Databases<br />

1. Database for<br />

Raw Data<br />

2. Database for<br />

Processed &<br />

Analyzed Data<br />

• Data Processing<br />

PDAS-2<br />

•Cleansing<br />

3. Database for<br />

•Reconciling<br />

Extract<br />

Monitoring<br />

•Deriving<br />

Reports<br />

•Matching<br />

•Standardizing<br />

PIMS-1<br />

Extract •Transforming<br />

4. Database for<br />

•Conforming<br />

System<br />

•Exporting<br />

Maintenance<br />

Records<br />

PIMS-2<br />

Extract<br />

Feed & Load<br />

Feed & Load<br />

User Access Tools<br />

•Reporting &<br />

Query Tools<br />

•OLAP Tools<br />

•Data Mining Tools<br />

Feed & Load<br />

Users<br />

• SHES-Workstations<br />

• SHES-Console<br />

• DPCS-Consoles<br />

Figure 48 Operation Flow Diagram of SHDMS (Structural Health Data Management System)<br />

-289-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

A RELIABILITY BASED SIMULATION, MONITORING AND CODE CALIBRATION<br />

OF VEHICLE EFFECTS ON EXISTING BRIDGE PERFORMANCE<br />

C.S. Cai 1 , Wei Zhang 1 , Miao Xia 1 and Lu Deng 1<br />

Dept. of Civil and Environmental Engineering, Louisiana State <strong>University</strong>, Baton Rouge, Louisiana 70803,<br />

cscai@lsu.edu.<br />

ABSTRACT<br />

This paper summarizes the recent work by the first author’s research group related to vehicle effects on existing<br />

bridges, using reliability based performance assessment. <strong>The</strong> first part presents a simulation framework of<br />

fatigue reliability assessment for existing bridges considering the effects of vehicle speed, road roughness<br />

condition, equivalent stress range and the constant amplitude fatigue threshold. <strong>The</strong> second part develops a<br />

framework to estimate the extreme structural response due to live load in a mean recurrence interval based on<br />

short-term monitoring of existing bridges. <strong>The</strong> Gumbel distribution of the extreme values was derived from<br />

extreme value theory and Monte Carlo Simulation. Studies have shown that the vehicle-induced dynamic<br />

allowance IM value prescribed by the AASHTO LRFD code may be underestimated under poor road surface<br />

conditions (RSCs) of some existing bridges. In addition, the IM, which is a random variable with certain<br />

statistical properties, was modeled as a deterministic constant in the code calibration process. In the third part of<br />

this paper, the reliability indices of a selected group of prestressed concrete girder bridges are calculated by<br />

modeling the IM explicitly as a random variable for different RSCs. Appropriate IM values are suggested for<br />

different RSCs in order to achieve a consistent target reliability index and a reliable load rating for existing<br />

bridges.<br />

KEYWORDS<br />

Vehicle effects, existing bridges, calibrations, reliability, extreme values.<br />

INTRODUCTION<br />

Due to the progressive deteriorations and accumulated fatigue damages of structures under dynamic loads such<br />

as vehicles, it is essential to ensure the structure safety and calculate fatigue reliabilities. In fatigue design, the<br />

load-induced fatigue effect should be less than the nominal fatigue resistance. Naturally, the fatigue requirement<br />

can also be stated as the fatigue life consumed by the load being less than the available fatigue life of the bridge<br />

detail. However, the vehicle speed and road roughness conditions are not considered in a typical fatigue analysis<br />

such as in the American Association of State Highway and Transportation Officials (AASHTO) specifications<br />

(2007). It has been proven that these two factors have significant effects on the dynamic responses of short span<br />

bridges (Deng and Cai 2010; Shi et al. 2008). In the present study, the deterioration of the road surface due to<br />

environmental corrosions and passages of truckloads is considered. In addition, the road may be resurfaced<br />

when the road surface becomes unacceptable. <strong>The</strong>refore, the road roughness coefficient is considered to change<br />

periodically with time. A limit state function is defined in the present study to calculate the fatigue reliability.<br />

Besides the random vehicle speeds and road profiles, the traffic increase rate is taken into account in the present<br />

study to compare their effects on the fatigue reliability index and fatigue life.<br />

Recently developed structural health monitoring (SHM) systems of bridges can provide useful information with<br />

which bridge performance can be assessed and predicted more precisely. Instead of measuring the weight of<br />

every vehicle passing through bridges, SHM techniques can more conveniently record structural response such<br />

as strains under routine service traffic loads. <strong>The</strong> procedure of reliability assessment using monitoring data<br />

includes collecting survey data, identifying the distribution type of live load effects and estimating its<br />

distribution parameters by curve fitting. <strong>The</strong> extreme load effects are derived from the response of a bridge<br />

directly including each possible combination of the number of loaded lanes multiplied by a corresponding<br />

multiple presence factor to account for the probability of simultaneous lane occupation (Orcesi and Frangopol<br />

2010). For reliability calculations, with the help of the monitored data, one can minimize the uncertainties and<br />

apply fewer assumptions in structural analysis, thus make the reliability calculation more rational. One of the<br />

aims of this study is to develop a methodology that can establish the maximum live load effects distribution for<br />

-290-


a mean recurrence interval with extreme value theories based on short-term monitoring data of structural<br />

response.<br />

VEHICLE-INDUCED FATIGUE RELIABILITY ASSESSMENT OF EXISTING BRIDGES<br />

Prediction of vehicle induced bridge vibration<br />

To predict vehicle-induced vibrations for fatigue assessment, the vehicle is modeled as a combination of several<br />

rigid bodies connected by several axle mass blocks, springs and damping devices (Cai and Chen 2004). <strong>The</strong> tires<br />

and suspension systems are idealized as linear elastic spring elements and dashpots. <strong>The</strong> equations of motion for<br />

the vehicle-bridge coupled system are as follows:<br />

⎡Mb ⎤⎧⎪d&&<br />

⎫<br />

b⎪ ⎡Cb + Cbb Cbv⎤⎧⎪d&<br />

⎫ K<br />

b⎪<br />

⎡ b<br />

+ Kbb Kbv⎤⎧db⎫ ⎧ Fb<br />

⎫<br />

⎢ G<br />

M<br />

⎥⎨ ⎬+ ⎢<br />

v d C<br />

v<br />

vb<br />

C<br />

⎥⎨ ⎬+ ⎢ ⎨ ⎬=<br />

⎨ ⎬<br />

v d K<br />

v<br />

vb<br />

K<br />

⎥<br />

⎣ ⎦⎪ &&<br />

⎪ ⎣ ⎦⎪ &<br />

(1)<br />

⎩ ⎭ ⎩ ⎪⎭<br />

⎣ v ⎦⎩dv⎭ ⎩Fc + Fv<br />

⎭<br />

where, [M b ] is the mass matrix, [C b ] is the damping matrix and [K b ] is the stiffness matrix of the bridge, {F b } is<br />

wheel-bridge contact forces on bridge; [M v ], is the mass matrix, [C v ], is the damping matrix and [K v ] is the<br />

stiffness matrix of the vehicle; {F G v } is the self-weight of vehicle; and {F c }is the vector of wheel-road contact<br />

forces acting on the vehicle. <strong>The</strong> terms C bb , C bv , C vb , K bb , K bv , K vb , F b and F v in Eq. (1) are due to the interactions<br />

between the bridge and vehicles.<br />

To demonstrate the methodology, a 12m long and 13 m wide slab-on-girder bridge is analyzed, which is designed<br />

in accordance with AASHTO LRFD bridge design specifications (AASHTO 2007). In the present study, after<br />

conducting a sensitivity studying by changing the meshing, 27543 solid elements and 43422 nodes are used to<br />

build the finite element model of the bridge. <strong>The</strong> damping ratio is assumed to be 0.02. <strong>The</strong> present study focuses<br />

on the fatigue analysis at the longitudinal welds located at the conjunction of the web and the bottom flange at<br />

the mid-span. Since the design live load for the prototype of the bridge is HS20-44 truck, this three-axle truck is<br />

chosen as the prototype of the vehicle in the present study. In addition, only one vehicle in one lane is<br />

considered to travel along the bridge for fatigue analysis due to its short span length.<br />

Road surface roughness is an important parameter that causes dynamic effect and fatigue problem. It is<br />

generally defined as an expression of irregularities of the road surface and it is the primary factor affecting the<br />

dynamic response of both vehicles and bridges (Deng and Cai 2010; Shi et al. 2008). Road roughness condition<br />

is classically quantified using Present Serviceability Rating (PSR), Road Roughness Coefficient (RRC) or<br />

International Roughness Index (IRI). Various correlations have been developed between the indices (Paterson<br />

1986; Shiyab 2007). Based on the studies carried out by Dodds and Robson (1973) and Honda et al. (1982), the<br />

road surface roughness was assumed as a zero-mean stationary Gaussian random process and it could be<br />

generated based on the RRC through an inverse Fourier transformation as (Wang and Huang 1992):<br />

N<br />

rx ( ) = ∑ 2 φ( nk) Δ ncos(2 πnx<br />

k<br />

+ θk)<br />

(2)<br />

k = 1<br />

where θ k is the random phase angle uniformly distributed from 0 to 2π; φ()<br />

is the power spectral density (PSD)<br />

function (m 3 /cycle/m) for the road surface elevation; n k is the wave number (cycle/m). Wang and Huang (1992)<br />

also suggested a concise power-spectrum-density function that was used in the present study:<br />

n −2<br />

φ( n) = φ( n0<br />

)( )<br />

(3)<br />

n0<br />

where φ( n)<br />

is the PSD function (m 3 /cycle) for the road surface elevation; n is the spatial frequency (cycle/m);<br />

n 0 is the discontinuity frequency of 1/2π (cycle/m); and φ( n0<br />

) is the RRC (m 3 /cycle) and its value is chosen<br />

depending on the road condition.<br />

In order to consider the road surface damages, a progressive deterioration model for road roughness is necessary.<br />

More specifically, it is essential to have such a model for RRC in order to generate the random road profile.<br />

<strong>The</strong>refore, the RRC at any time after construction is predicted using the progressive deterioration model for IRI<br />

and the relationship between the IRI and RRC(Paterson 1986; Shiyab 2007):<br />

− 9 η<br />

( { 5 ( ) }<br />

6<br />

0) 6.1972 10 exp 1.04 t<br />

− −<br />

φ n<br />

t<br />

= × × ⎡<br />

⎣ e ⋅ IRI0<br />

+ 263(1 + SNC) CESAL ⎤/ 0.42808 + 2×<br />

10<br />

t ⎦ (4)<br />

-291-


Fatigue Assessment<br />

Based on Miner’s rule, the accumulated fatigue damages could lead to bridge failures. According to Miner’s rule,<br />

the accumulated damage is<br />

ni<br />

ntc<br />

Dt () = ∑ =<br />

(5)<br />

i Ni<br />

N<br />

where n i is number of observations in the predefined stress-range bin S ri , N i is the number of cycles to failure<br />

corresponding to the predefined stress-range bin; n tc is the total number of stress cycles and N is the number of<br />

cycles to failure under an equivalent constant amplitude loading (Kwon and Frangopol 2010):<br />

m<br />

N = A⋅ S −<br />

re<br />

(6)<br />

where S re is the equivalent stress range and A is the detail constant taken from Table 6.6.1.2.5-1 in AASHTO<br />

(AASHTO, 2007). Either using the Miner’s rule or Linear Elastic Fracture Mechanics (LEFM) approach, the<br />

equivalent stress range for the whole design life is obtained through the following equation (Chung 2006):<br />

n<br />

1/ m<br />

⎛<br />

m ⎞<br />

Sre = ⎜∑ αi ⋅Sri<br />

⎟<br />

(7)<br />

⎝ i=<br />

1 ⎠<br />

where α i is the occurrence frequency of the stress-range bin, n is the total numbers of the stress-range bin and m is<br />

the material constant that could be assumed as 3.0 for all fatigue categories (Keating and Fisher 1986).<br />

Since each truck passage might induce multiple stress cycles, two correlated parameters were essential to<br />

calculate the fatigue damages done by each truck passage, i.e. the equivalent stress range and numbers of cycle<br />

per truck passage. In the present study, a new single parameter, S w , is introduced for simplifications to combine<br />

the two parameters on a basis of equivalent fatigue damage; namely, the fatigue damage of multiple stress cycles<br />

is the same as that of a single stress cycle of S w . For truck passage j, the revised equivalent stress range is<br />

defined and derived as:<br />

j j<br />

( ) 1/m j<br />

Sw = Nc ⋅ Sre<br />

(8)<br />

where N j c is the number of stress cycles due to the j th j<br />

truck passage, and S<br />

re<br />

is the equivalent stress range of the<br />

stress cycles by the j th truck.<br />

When D(t) is 1, the structure approaches to fatigue failure based on the Miner’s rule. Correspondingly, the limit<br />

state function (LSF) is defined (Nyman and Moses 1985):<br />

g( X ) = Df<br />

− D( t)<br />

(9)<br />

where D f is the damage to cause failure and is treated as a random variable with a mean value of 1; D(t) is the<br />

accumulated damage at time t; and g is a failure function such that g


<strong>The</strong> present study is concerned with the fatigue cracks that may develop at the longitudinal welds located at the<br />

conjunction of the web and the bottom flange at the mid-span, which falls into fatigue detail Category B<br />

(AASHTO 2007). When the fatigue detail coefficient A is assumed to follow a lognormal distribution, the mean<br />

and COV values can be calculated based on the test results of welded bridge details. Based on the tests<br />

performed by Keating and Fisher (1986), the mean and COV are calculated as 7.83×10 10 and 0.34.<br />

S w , the revised equivalent stress range, is calculated for given combinations of vehicle speed and road roughness<br />

condition. In the present study, the chi-square goodness-of-fit test is used to check the distribution type of the<br />

parameter S w for each combination of vehicle speed and road roughness condition. Based on Sturges’ rule, the<br />

number of intervals is seven for a 50-data bin and the degree of freedom is four. If a 5% significance level is<br />

chosen, the test limit for the Chi-Square test is calculated as C 0.95,4 =9.488. As demonstrations in the present<br />

study, six out of the total thirty groups of cases are employed to verify the distribution type. Both normal and<br />

lognormal are acceptable for revised equivalent stress range S w based on the Chi-Square test.<br />

Based on the assumption that all the variables (i.e. A, D f and S wj ) follow a certain distribution, the fatigue<br />

reliability index is obtained using the method in the literature (Estes and Frangopol 1998). Based on such a<br />

method, an arbitrary initial design point can be chosen and the solving process for the complex equation of<br />

g()=0 can be avoided. After several iterations, convergence can be achieved without forcing every design points<br />

to fall on the original failure surface.<br />

In order to investigate the effects of the vehicle speed and road roughness condition on the fatigue reliability, the<br />

fatigue reliability are calculated for all the 30 combinations of 6 vehicle speeds from 10m/s to 60m/s and 5 road<br />

roughness conditions from very good to very poor. <strong>The</strong> limit state function for a given combination of vehicle<br />

speed and road roughness condition is simplified as:<br />

m<br />

ntr<br />

⋅ Sw<br />

g( X ) = Df<br />

−<br />

(11)<br />

A<br />

Based on the LSF function, the fatigue reliability index, β, can be derived, assuming that all random variables<br />

(i.e. A, D f and S w ) are lognormal, as follows:<br />

λD + λ ( ln )<br />

f A<br />

− m⋅ λS + n<br />

w tr<br />

β =<br />

(12)<br />

2 2 2<br />

ζ + ζ + ( m⋅ζ<br />

)<br />

λ<br />

D f<br />

Df<br />

A Sw<br />

where , λ<br />

A , λ<br />

S and ζ<br />

w<br />

D , ζ<br />

f A , ζ<br />

S denote the mean value and standard deviation of ln(A), ln(D<br />

w<br />

f ) and ln(S w ),<br />

respectively.<br />

Generally, the fatigue reliability indices are found to decrease with the increase of vehicle speed and road<br />

roughness coefficient. If a 5% failure probability, i.e., a 95% survival probability is assumed, the corresponding<br />

reliability index is 1.65 (Kwon and Frangopol 2010). When the vehicle increase rate is 0%, the reliability index<br />

in all the thirty cases is larger than the target index of 1.65. Accordingly, the survival probability of all the thirty<br />

cases is larger than 95%. When the vehicle increase rate is 3% or 5%, seven or eight cases are found with a<br />

fatigue reliability index less than 1.65, i.e., the probability of fatigue failure for the bridge is larger than 5%<br />

during its 75 years life. In addition, when the vehicle increase rate is 5% and the road roughness condition is<br />

very poor, the reliability index is negative, which suggests the bridge would be more likely to suffer fatigue<br />

failure than to survive in its 75 years life. In general, the higher vehicle speed, the smaller reliability index and<br />

the higher probability of failure the structure will have in most cases. <strong>The</strong> deteriorated road surface seems to<br />

accelerate fatigue damages due to the dynamic effects from vehicles, which implies the importance of road<br />

surface maintenances for existing bridges.<br />

Α<br />

Table 1 Fatigue reliability index and Fatigue life<br />

β (for Fatigue life of 75<br />

years)<br />

Fatigue life for target β=1.65<br />

(unit: years)<br />

0 % 6.5 1906<br />

3 % 6.0 136<br />

5 % 4.4 93<br />

In the real circumstances, vehicle speeds vary for different trucks and road surface conditions deteriorate with<br />

time. By treating the vehicle speed and road roughness as random variables as discussed earlier, the fatigue<br />

reliability index is calculated and listed in Table 1 for different traffic increase rate, i.e., α= 0%, 3% and 5%. <strong>The</strong><br />

-293-


fatigue life corresponding to a target fatigue reliability of 1.65 is also presented in the table. As expected, the<br />

predicted fatigue life is longer than that if we assume the road roughness is poor or very poor and shorter than<br />

that if we assume the road roughness is good or very good. For the current modeling of the vehicle speed and<br />

road surface deteriorations, the fatigue life of the bridge components is comparable with the case with a 60m/s<br />

vehicle speed and an average road-roughness condition.<br />

ESTIMATE THE EXTREME STRUCTURAL RESPONSE<br />

This section is focus on the estimation of the extreme structural response based on monitoring data. At present,<br />

the live load applied to calculate the reliability index was computed based on the live load models that were<br />

developed by Nowak (1993) and used in the calibration of the early version of AASHTO LRFD Bridge Design<br />

Specifications (AASHTO 1994). <strong>The</strong> models were derived from the available statistical data on 9,250 selected<br />

truck surveys, and weigh-in-motion measurements. For a structure with a given capacity, its reliability index is<br />

only related to the maximum load effect distribution corresponding to the structure’s service life. Assuming a<br />

normal distribution for the individual truck load, the maximum live load effects (moment or shear) for various<br />

time periods were determined by extrapolation as follows.<br />

Let the live load effects following certain distribution Ω, and the number of trucks in the surveying interval<br />

is . <strong>The</strong>n, the total number of trucks passing through the bridge in an expected service life ,<br />

will be<br />

(13)<br />

<strong>The</strong> maximum live load effects<br />

corresponding to any expected bridge service life is<br />

where is the inverse of the standard distribution function. According to AASHTO specifications, the<br />

expected service life for a new bridge is 75 years.<br />

According to Orcesi and Frangopol (2010), the extreme value distribution of SHM data is assumed to approach<br />

a Gumbel probability distribution (Gumbel 1958).<br />

(15)<br />

where is the cumulative distribution function of Gumbel probability distribution, both μ and σ are<br />

constants to be determined from the measured data. Thus, the extreme values of the SHM data in a mean<br />

recurrence interval, , ( is longer than the monitoring period), can be predicted as<br />

(16)<br />

Using these methods to estimate the extreme value of load effects in a mean recurrence interval, the information<br />

of the number of the trucks running through the bridge must be available. However, sometimes it is not easy to<br />

identify the number of trucks only by dealing with the recorded strain data. Even the number of the trucks is<br />

known, some cases whose maximum structural response is induced by multiple presences of vehicles side by<br />

side or one after another in a same span are still excluded in the reliability calculation.<br />

<strong>The</strong> aim of this part is to develop a methodology that can establish the maximum live load effects distribution<br />

for a mean recurrence interval with extreme value theories based on short-term monitoring data of structural<br />

response.<br />

Extreme Value <strong>The</strong>ory<br />

Since only the maximum structure response is concerned, it is reasonable to use extreme theories to estimate the<br />

long-term maximum response from the short-term records of structure response monitoring. In this study, the<br />

Gunbel distribution is used to model the extreme values of long (finite) sequences of independent, identically<br />

distributed random variables. Gumbel distributions also known as Type I extreme value distribution within the<br />

extreme value theory. Let the variable be the maximum of independent random variables . Since<br />

the inequality implies for all i , it follows that<br />

<strong>The</strong> probabilities are referred to as the initial distributions of the variables . In the particular case in<br />

(14)<br />

(17)<br />

-294-


which all the variables have the same probability distribution , the probability distribution of<br />

becomes<br />

If the number becomes large enough, the cumulative distribution of the largest values approach limits<br />

known as Type I or Type II extreme value distributions if the initial distributions are of the exponential or of the<br />

Cauchy type, respectively (Simiu and Scanlan 1986).<br />

Probabilistic Modeling of Maximum Live Load Effects for a Mean Recurrence Interval<br />

For simplification, let’s use the yearly maximum live load effects as a demonstration. <strong>The</strong> yearly maximum live<br />

load effect is the extreme value of live load effects the structure is subjected in a year. A year can be divided into<br />

time segments; the maximum live load effects in each time segment, , is a variable and can be obtained<br />

with bridge health monitoring system. <strong>The</strong> extreme values are assumed to be independent with each other and<br />

have the same distribution with a cumulative distribution function . <strong>The</strong> probabilities, , is<br />

referred to as the initial distribution. <strong>The</strong> distribution of yearly maximum live load effect can be derived<br />

according to Eqs. (17) and (18) as:<br />

<strong>The</strong> accuracy of the estimation of the yearly maximum live load effects probability distribution depends on the<br />

accuracy of the distribution of the initial population and the number of the intervals. For different number of<br />

time segments and (or different length of time segments, seg_1 and seg_2), initial distribution<br />

and<br />

can be obtained. <strong>The</strong> distribution of the extreme live load effects in a mean<br />

recurrence interval can be derived in terms of and as follow:<br />

For example, the yearly extreme structure response distribution can be derived from initial distributions based<br />

on time segments of an hour or a minute as:<br />

where and are initial distributions and they represent the maximum structure response<br />

cumulative distribution function for 1 hour and 1 minute, respectively. <strong>The</strong> initial distribution can be derived by<br />

distribution fitting of the monitored data. It shows in Eq. (20) and Eq. (21) that the yearly maximum live load<br />

distribution is unique; the initial distributions corresponding to different length of time segments are not unique.<br />

To estimate a reasonable number of intervals or to determine the length of the time segment requires that the<br />

following two principles are satisfied (Duan et al. 2002):<br />

1. <strong>The</strong> time segment is long enough so that the maximum live load in every interval satisfies independence<br />

requirement.<br />

2. <strong>The</strong> length of the time segment is reasonable so that the maximum live loads in every interval follow the<br />

same distribution.<br />

<strong>The</strong>se two conditions require the length of the time segments is long enough so that the structural response<br />

record in every time segment is a stationary and ergodic process. <strong>The</strong> property of stationarity of a stochastic<br />

process always refers to the process being unchanged when shifting along the time axis (Lutes and Sarkani<br />

2004). For an ergodic process, its statistical properties (such as its mean and variance) can be deduced from a<br />

single, sufficiently long sample (realization) of the process. In other words, statistical properties obtained from a<br />

single time-series will approach definite limits independent of the particular series as the length of the series<br />

increases. Once the initial distribution of the parent population is determined, the maximum distribution in any<br />

mean recurrent intervals can be derived from Eq. (20) and Eq. (21).<br />

(18)<br />

(19)<br />

(20)<br />

(21)<br />

-295-


Case Study<br />

With time segment of 8s<br />

0.05<br />

With time segment of 50s<br />

Probability Density Function<br />

0.2<br />

0.15<br />

0.1<br />

0.05<br />

Probabilily Density Function<br />

0.045<br />

0.04<br />

0.035<br />

0.03<br />

0.025<br />

0.02<br />

0.015<br />

0.01<br />

0<br />

0 20 40 60 80 100 120 140 160<br />

Strain (1.0E-6)<br />

0.005<br />

0<br />

20 40 60 80 100 120 140 160<br />

Strain (1.0E-6)<br />

0.025<br />

With time segment of 90s<br />

0.012<br />

With time segment of 300s<br />

Probability Density Function<br />

0.02<br />

0.015<br />

0.01<br />

0.005<br />

Probability D ensity Function<br />

0.01<br />

0.008<br />

0.006<br />

0.004<br />

0.002<br />

0<br />

0 20 40 60 80 100 120 140 160<br />

Strain (1.0E-6)<br />

0<br />

0 20 40 60 80 100 120 140 160<br />

Strain (1.0E-6)<br />

Figure 1 Initial distribution fitting using Gumbel distribution: time segment of (a) 8s; (b) 50s; (c) 90s; (d) 300s<br />

Time history record of strains at the mid-span of a steel girder in the CORIBM Bridge on route LA 70 in District<br />

61, Assumption Parish, Louisiana, were recorded. Since only three hours monitoring data is available, it is not<br />

advisable to estimate extreme strain distribution for a very long mean recurrence interval. <strong>The</strong> extreme strain<br />

distribution (Gumbel distribution) for mean recurrence intervals of 1 day, 10 days, 30 days, 180 days and one<br />

year were estimated using maximum likelihood estimation method as shown in Fig . 1. <strong>The</strong> yearly extreme<br />

strain distribution can be derived from the initial distributions with time segments of 90s or 300s as<br />

(22)<br />

where the initial distributions and have been obtained from distribution fitting previously. It is<br />

difficult to prove directly that Eq. (22) is tenable and it is difficult to derive the parameters of yearly extreme<br />

response distribution through an analytical method. An alternative method is to generate samples using Monte<br />

Carlo simulation following the distribution functions on the right side of Eq. (22), and then, fit the generated<br />

samples with the selected distribution function, Gumbel distribution (maximum cases). <strong>The</strong> factors derived from<br />

distribution fitting were shown in Fig. 2. From Fig. 2, it is seen that both the location factor μ and the scale factor<br />

σ show converging property as the length of time segment increases. For different mean recurrence intervals, μ<br />

converges to different values, but σ converges to a fixed value. <strong>The</strong> location factor μ determines the mode value of<br />

the distribution while the shape factor σ determines the variance or the standard deviation of the distribution. Its<br />

location shifts to the right direction as the mean recurrence interval increases. <strong>The</strong> distributions have different<br />

mode values but same variance for different mean recurrence intervals. <strong>The</strong> PDF of extreme strain distribution for<br />

mean recurrence intervals of 1 day, 10 days, 30 days and one year are shown in Fig. 3.<br />

-296-


450<br />

35<br />

400<br />

30<br />

350<br />

300<br />

25<br />

( 1.0E- 6)<br />

m<br />

250<br />

200<br />

150<br />

100<br />

50<br />

1 day<br />

10 days<br />

30 days<br />

one year<br />

σ<br />

20<br />

15<br />

10<br />

5<br />

1 day<br />

10 days<br />

30 days<br />

one year<br />

0<br />

0<br />

0 50 100 150 200 250 300<br />

0 50 100 150 200 250 300<br />

Tim e Segm ent (s)<br />

Tim e Segm ent (s)<br />

Figure 2 Extreme strain distributions parameter, μ, σ, derived from different initial distributions<br />

Probability Density Function<br />

0.012<br />

0.01<br />

0.008<br />

0.006<br />

0.004<br />

0.002<br />

Mean recurrence interval of 1 day<br />

Mean recurrence interval of 10 days<br />

Mean recurrence interval of 30 days<br />

Mean recurrence interval of 180 days<br />

Mean recurrence interval of one year<br />

0<br />

100 200 300 400 500 600 700 800 900 1000<br />

Extreme Strain (1.0E-6)<br />

Figure 3 PDF of extreme strain distribution for mean recurrence intervals of 1 day, 10 days, 30 days and one year<br />

CALIBRATION OF IMPACT FACTOR FOR EXISTING BRIDGES<br />

Impact factors of existing bridges may be significantly different from code specified values. <strong>The</strong>refore, in the<br />

present study the impact factor of existing bridges is calibrated through a reliability approach. <strong>The</strong> following<br />

design equation, recommended by the current LRFD code (AASHTO 2004) for considering situations with only<br />

dead loads and vehicle live loads, was used in the calibration process.<br />

φ R ≥ 1.25DC<br />

+ 1.50DW<br />

+ 1.75LLl + 1.75(1 + IM ) LL t<br />

(23)<br />

where R = nominal value for the resistance, DC = effects due to design dead loads excluding the weight of<br />

wearing surface, DW = effects due to wearing surface weight, LL<br />

l = effects due to design lane load, LL<br />

t =<br />

effects due to the most demanding between truck and tandem load, IM = LL t , dyn<br />

LLt<br />

-1= 0.33 denotes the<br />

dynamic load allowance(impact factor), in which LL<br />

t , dyn = effects due to the most demanding between truck<br />

and tandem load including the dynamic effects, and φ = resistance factor. For prestressed-concrete bridges,<br />

the current AASHTO LFRD code suggests a resistance factor of 1.0 and 0.85 for moment and shear, respectively,<br />

based on the study by Nowak (1995).<br />

In order to calibrate the impact factors for existing bridges, a group of seven prestressed concrete girder bridges<br />

were selected as benchmark bridges in this study. <strong>The</strong> detailed configurations and properties of the seven<br />

prestressed concrete girder bridges are described in Deng and Cai (2010). By assuming certain statistical<br />

properties of the impact factors under different road surface conditions (RSCs) and explicitly modeling the IM<br />

as an individual random variable, calibrations can be performed with respect to different RSCs (Deng et al.<br />

2011).<br />

<strong>The</strong> proposed IMs have been chosen with the aim of balancing three different needs, i.e., (1) obtaining reliability<br />

indices consistent with the target reliability index of 3.5 for all bridges under any RSC, (2) ensuring as much as<br />

possible consistency with the AASHTO LRFD code (2004), and (3) minimizing the modifications to the IM<br />

values suggested by the AASHTO LRFR manual (2003).<br />

Table 2 provides, for all the RSCs considered, (1) the minimum values of IM required to reach the target<br />

-297-


eliability index of 3.5 for both moment and shear strength limit states for all the bridges considered here, (2) the<br />

proposed values of IM for rating of existing bridges, and (3) the values of IM proposed by two of the authors in<br />

a previous study (Deng and Cai 2010). From Table 2, it is observed that the IMs proposed in the present study<br />

are close to those proposed in Deng and Cai (2010), except for the case of very poor RSC.<br />

Table 2 Comparison between the proposed dynamic load allowance IMs in the present study and those by Deng<br />

and Cai (2010)<br />

Deng and Cai (2010)<br />

Present study<br />

Road Surface<br />

<strong>The</strong>oretical Minimum<br />

Condition<br />

Proposed IMs<br />

Proposed IMs<br />

IMs<br />

Very Poor 1.99 2.50 2.40<br />

Poor 0.99 1.00 0.90<br />

Average 0.50 0.50 0.33<br />

Good 0.33 0.15 0.20<br />

Very Good 0.23 0.05 0.10<br />

To verify whether the proposed IM can lead to a consistent reliability index of 3.5 for all RSCs, the reliability<br />

indices are recalculated for the seven bridges designed following Eq. (23) while employing the proposed IM<br />

values. <strong>The</strong>se results confirm that using the proposed IM values produce more consistent reliability indices that<br />

are very close to the target reliability index of 3.5, regardless of what the actual RSC is.<br />

It is noteworthy that, when the RSC is below average, the proposed IM values are greater than the value of 0.33<br />

suggested by the current LRFD code. In particular, the IM value proposed for very poor RCS is significantly<br />

larger than 0.33 (i.e., more than seven times larger). Since a significant portion (i.e., more than 17%) of all<br />

bridge decks in the United States belong to fair or worse RSCs, it is concluded that the use of different IMs,<br />

which account for the different RSCs, is necessary for accurate performance evaluation of existing bridges.<br />

SUMMARY AND CONCLUSIONS<br />

This paper summarizes the research group’s work related to vehicle effects on existing bridges, using reliability<br />

based performance assessment. <strong>The</strong> first part presents a simulation framework of fatigue reliability assessment<br />

for existing bridges considering the effects of vehicle speed, road roughness condition, equivalent stress range<br />

and the constant amplitude fatigue threshold. <strong>The</strong> second part develops a framework to estimate the extreme<br />

structural response due to live load in a mean recurrence interval based on short-term monitoring of existing<br />

bridges. <strong>The</strong> Gumbel distribution of the extreme values was derived from extreme value theory and Monte Carlo<br />

Simulation. In the third part of this paper, the reliability indices of a selected group of prestressed concrete girder<br />

bridges are calculated by modeling the IM explicitly as a random variable for different RSCs. Appropriate IM<br />

values are suggested for different RSCs in order to achieve a consistent target reliability index and a reliable<br />

load rating for existing bridges.<br />

REFERENCES<br />

American Association of State Highway and Transportation Officials (AASHTO) (2004). LRFD bridge design<br />

specifications, Washington, DC.<br />

American Association of State Highway and Transportation Officials (AASHTO). (2007). LRFD bridge<br />

design specifications, Washington, DC.<br />

Cai, C. S., and Chen, S. R. (2004). “Framework of vehicle-bridge-wind dynamic analysis”, Journal of Wind<br />

Engineering and Industrial Aerodynamics, 92(7-8), 579-607.<br />

Chung, H.-Y., Manuel, L., and Frank, K. H. (2006). “Optimal inspection scheduling of steel bridges using<br />

nondestructive testing techniques”, Journal of Bridge Engineering, 11(3), 305-319.<br />

Deng, L., and Cai, C. S. (2010). “Development of dynamic impact factor for performance evaluation of existing<br />

multi-girder concrete bridges”, Engineering Structures, 32(1), 21-31.<br />

Deng, L., Cai, C.S., and Barbato, M. (2011). “Reliability-based dynamic impact factor for prestressed concrete<br />

girder bridges”, Journal of Bridge Engineering, ASCE. In press.<br />

doi:10.1061/(ASCE)BE.1943-5592.0000178.<br />

Dodds, C. J., and Robson, J. D. (1973). “<strong>The</strong> description of road surface roughness”, Journal of Sound and<br />

Vibration, 31(2), 175-183.<br />

Duan,Z., Ou,J., and Zhou, D. (2002). “Optimal probability distributions of extreme wind speeds”, China Civil<br />

Engineering Journal, 35(5), 11-16.<br />

Estes, A. C., and Frangopol, D. M. (1998). “RELSYS: a computer program for structural system reliability”,<br />

-298-


Structural Engineering and Mechanics, 6(8), 901.<br />

Gumbel, E. J. (1958). Statistics of extremes, Columbia <strong>University</strong> Press, New York.<br />

Honda, H., Kajikawa, Y., and Kobori, T. (1982). “Specta of road surface roughness on bridges”, Journal of the<br />

Structural Division, 108(ST-9), 1956-1966.<br />

Keating, P. B., and Fisher, J. W. (1986). “Evaluation of fatigue tests and design criteria on welded details”,<br />

NCHRP Report 286, Transportation Research Board, Washington, D.C.<br />

Kwon, K., and Frangopol, D. M. (2010). “Bridge fatigue reliability assessment using probability density<br />

functions of equivalent stress range based on field monitoring data”, International Journal of Fatigue, 32(8),<br />

1221-1232.<br />

Lutes, L. D., and Sarkani, S. (2004). Random vibrations: analysis of structural and mechanical systems,<br />

Elsevier, Amsterdam ; Boston.<br />

Nowak, A. S. (1993). “Live load model for highway bridges”, Structural Safety, 13(1-2), 53-66.<br />

Nowak, A. S. (1995). “Calibration of LRFD bridge code”, Journal of Structural Engineering, 121(8),<br />

1245-1251.<br />

Nyman, W. E., and Moses, F. (1985). “Calibration of bridge fatigue design model”, Journal of Structural<br />

Engineering, 111(6), 1251-1266.<br />

Orcesi, A. D., and Frangopol, D. M. (2010). “Inclusion of crawl tests and long-term health monitoring in bridge<br />

serviceability analysis”, Journal of Bridge Engineering, 15(3), 312-326.<br />

Paterson, W. D. O. (1986). “International roughness index: relationship to other measures of roughness and<br />

riding quality”, Transportation Research Record, 1084, Washington, D.C.<br />

Rackwitz, R. and Fiessler, B. (1978). “Structural reliability under combined random load sequences”, Computer<br />

and Structures, 9, 489-494.<br />

Shi, X., Cai, C. S., and Chen, S. (2008). “Vehicle induced dynamic behavior of short-span slab bridges<br />

cconsidering effect of approach slab condition”, Journal of Bridge Engineering, 13(1), 83-92.<br />

Shiyab, A. M. S. H. (2007). “Optimum use of the flexible pavement condition indicators in pavement<br />

management system”, Ph.D Dissertation, Curtin <strong>University</strong> of Technology.<br />

Simiu, E., and Scanlan, R. H. (1986). Wind effects on structures: an introduction to wind engineering, Wiley,<br />

New York.<br />

Wang, T.-L., and Huang, D. (1992). “Computer modeling analysis in bridge evaluation”, Florida Department of<br />

Transportation, Tallahassee, FL.<br />

-299-


Parallel Session –<br />

Geotechnical I


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

A STUDY ON DISPLACEMENT-BASED EARTHQUAKE LOSS ASSESSMENT<br />

ADOPTING EQUIVALENT STIFFNESS LINEARIZATION METHOD<br />

Jitao Shi and Liang Su<br />

Department of Civil Engineering, Zhejiang <strong>University</strong>, Hangzhou, China, E-mail: suliang@zju.edu.cn<br />

ABSTRACT<br />

A simplified displacement-based earthquake loss assessment methodology (DBELA) on an urban scale is<br />

discussed, which can be used to determine the damage distribution under strong ground shaking for a single<br />

earthquake scenario. <strong>The</strong> paper focuses on the deterministic framework part of DBELA, and the accuracy of the<br />

loss assessment results used in the reinforced concrete frames are checked by corresponding inelastic dynamic<br />

time-history analyses both in the yield limit state and in the post-yield limit states. Since the secant stiffness<br />

method used herein during the equivalent linearization procedures is under suspicion, an equivalent stiffness<br />

method is chosen to replace the secant stiffness method, and finally the loss assessment outcomes using both<br />

methods are tested and compared.<br />

KEYWORDS<br />

Displacement-based, loss assessment, reinforced concrete frame, earthquake.<br />

INTRODUCTION<br />

<strong>The</strong> earthquakes occurred in recent decades have caused devastating impacts in both the social and economic<br />

arenas. <strong>The</strong> purpose of a vulnerability assessment is to obtain the probability of a given level of damage to a<br />

given building type due to a scenario earthquake. Loss assessment models on regional basis are aimed to<br />

predicting the economical impact of future earthquakes and to be for risk mitigation.<br />

Various methods for vulnerability assessment are developed in the past, and most of them can be divided into<br />

three categories: empirical, analytical and hybrid. However, the tradition procedures are mainly based on<br />

macroseismic intensity or peak ground acceleration, and the relationship between the frequency content of the<br />

ground motion and the dominant period of the buildings is not considered. What’s more, empirical procedures<br />

such as damage probability matrices or vulnerability functions are based on a large degree of subjectivity.<br />

Hybrid analyses combine the analytical results with the damages observation in the past events, so many<br />

intensity levels can be considered in this method which also allows calibration of the analytical models. But<br />

when the characteristics of building stocks change, the hybrid procedure cannot be easily transplanted.<br />

Fully analytical method for earthquake risk assessment has recently been proposed. <strong>The</strong> capacity spectrum<br />

method, as an embranchment of the analytical method, has been applied in the HAZUS loss assessment<br />

methodology and it has been adopted by many researchers for loss assessment studies. <strong>The</strong> main disadvantage<br />

of HAZUS is that its comprehensive methodologies require huge amounts of time and cost and research and<br />

computer power to perform a complete study.<br />

Another group of mechanical procedures adopt displacement to model the demand and capacity of the building<br />

class. Displacement capacity of classes of buildings in DBELA is described by mechanics-derived equations<br />

which are given in terms of the levels of limit states, material and geometrical properties. <strong>The</strong> displacement<br />

demand is predicted from the displacement response spectrum of a single earthquake scenario. With the<br />

relationship between height of building and limit state period, capacity curve and demand curve thus can be<br />

compared in the same period-versus-displacement coordinate. <strong>The</strong> original concept is illustrated in Figure 1,<br />

whereby the performance points are obtained and transformed into a range of heights using the aforementioned<br />

relationship between limit state period and height. This range of height is then superimposed onto the<br />

cumulative distribution function (CDF) of building stock to find the proportion of building failing the given<br />

limit state.<br />

-300-


However, the results of this simplified loss assessment method DBELA have never been checked, so we first<br />

focused on the deterministic framework part of DBELA in this paper. Since a secant stiffness model, which is<br />

proved inappropriate by many researchers, is used in DBELA’s methodologies proposed by Crowley and Pinho,<br />

an equivalent stiffness model will be adopted to replace the secant stiffness model. And the results using both<br />

equivalent linearization methods are also checked by a set of nonlinear dynamic time-history analysis in this<br />

paper.<br />

LS3 LS2 LS1<br />

η LS1<br />

Cumulative<br />

frequency<br />

Displacement<br />

η LS2<br />

η LS3<br />

P LS1<br />

P LS2<br />

P LSi – percentage of<br />

buildings failing LSi<br />

Demand<br />

spectra<br />

P LS3<br />

H LS3 H LS2 H LS1 Height<br />

T LS3 T LS2 T LS1<br />

Effective<br />

period<br />

(a)<br />

(b)<br />

Figure 1 DBELA method: (a) displacement demand/capacity curves (b) CDF of building stocks [S. Glaister<br />

and R. Pino, 2003]. LS stands for limit state.<br />

DBELA<br />

Displacement capacity of reinforced concrete building<br />

H LSi =f(T LSi , LSi)<br />

Before we go further, some important concepts should be noticed here: when damages happened to the<br />

structural (load-bearing) system of building class, different limit states then can be defined. <strong>The</strong> first and the<br />

most important limit state is defined as the yield point of structure. And the other post-yield limit states can be<br />

attained by the suggested sectional steel and concrete strains. And a building class of reinforced concrete<br />

beam-sway resisting frames is mainly discussed in this paper.<br />

<strong>The</strong> fundamental of the direct displacement-based seismic methodology is the representation of the<br />

multi-degree-of-freedom (MDOF) structure by an equivalent single-degree-of-freedom (SDOM). <strong>The</strong>refore, the<br />

SDOF substitute structure with given period and damping must give the displacement capacity at the centre of<br />

seismic force of the original structure.<br />

F<br />

m e<br />

Δ<br />

H e<br />

θ<br />

Figure 2 SDOF simulation structure modeling the first inelastic mode of response.<br />

<strong>The</strong> displacement capacity at the center of the seismic force can be obtained either by the limit state base<br />

rotation/drift or the roof deformation, and the former is adopted in structure displacement capacity equations<br />

-301-


provided in DBELA, a base rotation can be mechanics-derived for both beam- and column-sway frames and the<br />

displacement at the center of seismic force is calculated by multiplying this rotation θ by an effective height H e<br />

as is shown in Figure.2.<br />

<strong>The</strong> displacement capacity can be expressed as follow:<br />

Δ = θ ⋅ H e<br />

(1)<br />

where<br />

H<br />

e<br />

=<br />

n<br />

∑( miΔi<br />

Hi<br />

) ∑<br />

i=<br />

1<br />

n<br />

i=<br />

1<br />

( m Δ )<br />

(2a)<br />

and m i is the mass at the height H i associated with displacement Δ i . For regular beam-sway frames, the effective<br />

height coefficient ef h defined as the ratio of the height of the SDOF substitute structure (H e ) to the total height of<br />

the original building (H T ), expressed as a function of the number of storey n, suggested by Priestley[1997]. <strong>The</strong>n<br />

H e can also be computed as Equitation (2b)<br />

H = ef ⋅ H<br />

(2b)<br />

e<br />

h<br />

ef h = 0.64<br />

n ≤ 4<br />

ef h = 0.64<br />

− 0.0125( n − 4)<br />

4 < n < 20<br />

(3)<br />

ef h = 0.44<br />

n ≥ 20<br />

Researches in deformational behavior of RC members show that the yield curvature of beams and columns are<br />

independent of reinforce ratio and strength, thus storey yield drift of frames may be represented by the material<br />

and geometrical properties of the buildings only. For a concrete frame, ignoring strain-hardening and using<br />

empirical coefficients to account for shear and joint deformation, the yield base rotation of a beam-sway frames<br />

is shown in Equation (4).<br />

lb<br />

θ y = 0. 5ε y<br />

(4)<br />

hb<br />

where h b is the height of the beam section, l b is the length of the beam and ε y is the yield strain of the<br />

reinforcement steel.<br />

Combining the equations above, the yield displacement capacity (Δ y ) formula to define capacity displacement in<br />

the yield limit state for beam-sway frames presented in Equation (4) is altered as following.<br />

lb<br />

Δ y = 0.5ε yefhHT<br />

(5)<br />

hb<br />

From the definition of displacement ductility (Equation 6), displacement capacity of other limit states can easily<br />

be derived when the ductility levels are known.<br />

Δ LSi<br />

μ LSi =<br />

(6)<br />

Δ<br />

Displacement demand<br />

Displacement response spectra, which present the displacement demand to SDOF oscillators with a range of<br />

periods of vibration, are used in DBELA to represent the input from the earthquake to the building class under<br />

consideration. Normally response spectra provide information on the peak elastic response for a specified elastic<br />

damping ratio (typically 5%), and are plotted against the elastic period.<br />

T<br />

<strong>The</strong> concept viscous damping is generally used to account for the energy dissipated by structures in the elastic<br />

range and this viscous damping is often taken as 5% for reinforced concrete structures an 2% for steel structures.<br />

In order to account for non-linear behavior, the hysteretic damping included into the viscous damping term<br />

leading to so-called equivalent viscous damping ξ eq . <strong>The</strong> equivalent viscous damping is related to the ductility (μ)<br />

of the structural system in many studies and can be expressed in Equation (7) for concrete frame building.<br />

⎛ μ −1⎞<br />

ξeq = 0.05 + 0.565<br />

⎜<br />

⎟<br />

(7)<br />

⎝ μπ ⎠<br />

<strong>The</strong> equivalent viscous damping values obtained through Equation (7), for different ductility levels, can then be<br />

combined with Equation (8), proposed by Bommer et al. [2000] and currently implemented in EC8 [CEN, 2003],<br />

to compute a reduction factor η to be applied to the 5% damped spectra at periods from the beginning of<br />

y<br />

i<br />

i<br />

-302-


acceleration plateau to the displacement plateau, and thus can be used as the demand curve in DBELA.<br />

0.1<br />

η =<br />

0.05+<br />

ξ<br />

(8)<br />

Actually a Takede “Fat” (TF) model is used in Equation (7), and several other hysteretic models, such as Takede<br />

“Thin” (TT), Elasto-plastic (EPP), Bi-linear (BI) and Ramberg-Osgood (RO) hysteretic models, were suggested<br />

to solve the equivalent viscous damping. Usually the equivalent viscous damping is dependent on the equating<br />

the energy absorbed by hysteretic steady-state cyclic response to a given displacement level. Though different<br />

hysteretic model gives different equivalent viscous damping ratio, but for a set of much similar structure classes,<br />

for example, concrete structures, the EPP, BI, RO and TF models give viscous damping ratio without much<br />

difference. And when the equivalent viscous damping is used in Equation (8), the reduction factor doesn’t differ<br />

much. A comparison of the reduction factor deduced from EPP, BI, RO hysteretic models respectively with from<br />

the TF model shows that 5.2% difference at most. At least we could say that demand displacement isn’t so<br />

sensitive to these hysteretic models mentioned in Figure 3.<br />

Relative displacement error to<br />

TF hysteretic model<br />

0.06<br />

0.05<br />

0.04<br />

0.03<br />

0.02<br />

0.01<br />

0.00<br />

EPP<br />

BI,r=0.2<br />

RO<br />

2 4 6 8 10<br />

Displacement Ductility Factor (μ)<br />

Figure 3 Comparisons of EPP, BI, and RO hysteretic models to TT model when applied to the reduction factor η<br />

Uniform the coordinate of displacement capacity and displacement demand<br />

<strong>The</strong> displacement capacity equations obtained in Section 2.2 are actually functions of the total height of the<br />

original structure, and a reliable relationship between the limit state period and height of the building is<br />

fundamentally requirement in DBELA, in this way the displacement capacity formulae can be accurately<br />

defined in terms of limit state period and directly compared with the displacement demand curve.<br />

Simple empirical relationships are available in many force-based design codes to relate the fundamental period<br />

of vibration of a building to its height, however, these conservative period-height relationship would not<br />

appropriate for displacement-based loss assessment. So a relationship between yield period T y and height of<br />

reinforced concrete frames has been studied and shown in Equation (9)<br />

Ty = 0. 1H T<br />

(9)<br />

<strong>The</strong>n the yield displacement capacity equation in form of period can be expressed as following Equation (10) for<br />

beam -sway reinforced concrete frames.<br />

lb<br />

Δ y = 5.0ε yefhTy<br />

(10)<br />

hb<br />

In the post-yield limit states, the limit state period of the substitute structure can be deduced from the yield<br />

period and the ductility. A secant stiffness equivalent linearization method is adopted by Pinho in DBELA:<br />

TLS1 = T y μLSi<br />

(11)<br />

However, Miranda and Ruiz-Carcia found that the equivalent linearization method based on secant stiffness<br />

produces large error between the maximal displacements of the SDOF oscillator and the real structure, and the<br />

error gets more pronounced when periods are lower. Iwan introduced a two-dimensional minimization method<br />

to compute equivalent period and equivalent viscous damping, and proposed a set of effective linear parameters<br />

based on the response of hysteretic systems to earthquake excitations.<br />

ξ −ξ<br />

= A μ −1 2 + B μ −<br />

μ < 4.<br />

(12a)<br />

( ) ( ) 3<br />

eq 0 1<br />

0<br />

-303-


ξ<br />

( μ − ) 3<br />

eq −ξ0 = C + D 1<br />

4.0<br />

≤ 6. 5<br />

T<br />

T<br />

T<br />

T<br />

eq<br />

y<br />

eq<br />

y<br />

( μ −1) 2 + F( μ 1) 3<br />

−1 = E<br />

−<br />

( μ )<br />

−1 = G + H −1<br />

≤ μ (12b)<br />

μ < 4.0<br />

(13a)<br />

4.0<br />

≤ μ ≤ 6.5<br />

(13b)<br />

<strong>The</strong>n the displacement capacity equations during post-yield limit should be summarized as follow:<br />

lb<br />

Δ LSi = 5.0efhε<br />

y μLSiT<br />

( Ty,<br />

μ)<br />

(14)<br />

hb<br />

And also the reduction factor to the displacement demand spectrum is substituted in Equation (15).<br />

NUMERICAL ANALYSIS<br />

Yield limit state<br />

η =<br />

0.10<br />

0.05 + ξ<br />

( ) μ LSi<br />

<strong>The</strong> deterministic analysis stage of DBELA can easily be carried out by simply comparing the capacity and<br />

demand displacement spectra deduced from equations in the yield limit state. An example using DBELA is<br />

provided herein to illustrate the workings of the deterministic method on a building class of beam-sway frames,<br />

and then the accuracy of the assessment result would be checked.<br />

<strong>The</strong> capacity displacement spectra can be generated from Equation 10, and the parameters are available in Table<br />

1. Twenty artificially seismic waves, of which the acceleration and displacement response spectra are shown in<br />

Figure 4, are selected. And we chose the average displacement spectra of the twenty waves as the demand<br />

displacement curve.<br />

Table 1 Parameter of the RC frames in yield limit state displacement for beam-sway frames<br />

Parameter<br />

Value<br />

Beam length, l b<br />

6.0m<br />

Height of beam section, h b<br />

0.5m<br />

Storey height<br />

3m<br />

Steel yield strain, ε y<br />

0.15%<br />

Mass at top height<br />

50tone<br />

Mass at ordinary-storey 40tone<br />

(15)<br />

Acceleration (cm/s/s)<br />

700<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

average acc. spectrum<br />

target acc. spectrum<br />

Ta=0.15s<br />

Tb=0.50s<br />

Tc=2.5s<br />

PGA=0.25g<br />

Displacement (cm)<br />

30<br />

25<br />

20<br />

15<br />

10<br />

5<br />

average disp. spectrum<br />

target disp. spectrum<br />

Ta=0.15s<br />

Tb=0.5s<br />

Tc=2.5s<br />

PGA=0.25g<br />

0<br />

0 1 2 3 4 5 6<br />

Period (seconds)<br />

0<br />

0 1 2 3 4 5 6<br />

Period (seconds)<br />

Figure 4 Acceleration and displacement response spectra of 20 artificially waves used in the yield limit state<br />

Comparison of the capacity and demand curves is carried out as shown in Figure 5. Demand curve exceeds the<br />

capacity curve between the period 0.91s and 3.13s, which means that, RC frames with height between 9.1m and<br />

31.3m according to Equation (9) have failed in the yield limit state.<br />

-304-


In order to check the loss assessment results, a series of dynamic time-history analysis are carried out by a<br />

nonlinear fiber program “SeismoStruct”. Four 2D reinforced concrete frames with two bays are built in fiber<br />

element, 5-storey (H T =15m, T y =1.5s), 8-storey (H T =24m, T y =2.4s), 10-storey (H T =30m, T y =3.0s) and 12-storey<br />

(H T =36m, T y =3.6s) respectively. Geometry parameters of the model frames are used the same with that in Table<br />

1. Since fiber elements model only flexural response. Shear strength and shear deformation are generally not<br />

modeled. Column-beam joints and foundation joints are modeled by elastic non-fiber spring elements in order to<br />

take the joint rotation and shear deformation into account. <strong>The</strong> analysis outcomes are shown in Figure 6 and the<br />

maximum equivalent displacement is calculated as Equation (16)<br />

Δ<br />

eq<br />

=<br />

n<br />

∑<br />

i=<br />

1<br />

2<br />

i i<br />

m Δ<br />

n<br />

∑<br />

i=<br />

1<br />

m Δ<br />

where m i is the mass at the height H i associated with maximum displacement Δ i during the time-history analysis.<br />

i<br />

i<br />

(16)<br />

Displacement (cm)<br />

25<br />

20<br />

15<br />

10<br />

PGA=0.25g<br />

Ta=0.15s<br />

Tb=0.5s<br />

Tc=2.5s<br />

T=3.13s<br />

H=31.3m<br />

Capacity Displacement<br />

5 T=0.91s<br />

H=9.1m Demand Displacement<br />

0<br />

0 1 2 3 4 5 6<br />

Period (seconds)<br />

Figure 5 Beam-sway RC frame yield capacity and demand curves<br />

Displacement (m)<br />

0.27<br />

max disp. of a single wave<br />

0.24<br />

Average Displacement<br />

0.21<br />

Demand Displacement<br />

0.18<br />

Capacity displacement<br />

0.15<br />

0.12<br />

0.09<br />

0.06<br />

0.03<br />

5-storey RC frame<br />

PGA=0.25g<br />

0.00<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially wave<br />

Displacement (m)<br />

0.27 8-storey RC frame<br />

0.24<br />

PGA=0.25g<br />

0.21<br />

0.18<br />

0.15<br />

0.12<br />

Max disp. of a single wave<br />

0.09<br />

Average Displacement<br />

0.06<br />

Demand Displacement<br />

0.03<br />

Capacity Displacement<br />

0.00<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially waves<br />

Displacement (m)<br />

0.27<br />

10-storey RC frame<br />

0.24<br />

PGA=0.25g<br />

0.21<br />

0.18<br />

0.15<br />

0.12<br />

Max disp. of a single wave<br />

0.09<br />

Average displacement<br />

0.06<br />

Demand displacement<br />

0.03<br />

Capacity displacement<br />

0.00<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially waves<br />

Displacement (m)<br />

0.27<br />

0.24<br />

0.21<br />

0.18<br />

0.15<br />

12-storey RC frame<br />

PGA=0.25g<br />

0.12<br />

Max disp. of a single wave<br />

0.09<br />

Average displacement<br />

0.06<br />

Demand displacement<br />

0.03<br />

Capacity displacement<br />

0.00<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially waves<br />

Figure 6 Maximum equivalent displacement of time-history analysis<br />

According to the result of the DBELA, the 5-storey and 8-storey reinforced concrete frames have been failed in<br />

the yield limit state, which agrees well with the result of dynamic time-history analysis. All the maximum<br />

-305-


equivalent displacements exceed the capacity displacements, but however the average equivalent displacement<br />

may not exactly the same with the demand displacement due to random character of the seismic waves. Also we<br />

should notice that the 5- or 8-storey frame have already passed the yield limit state, the actual period should be<br />

longer than the yield period, and the viscous damping is no more 5% as well. But the result of DBELA, the<br />

critical height will be hardly affected. <strong>The</strong> 10-storey frame with the height of 30m is well around but a little less<br />

than the critical height (31.3m) and thirteen twentieth analytical points exceed the capacity displacement and<br />

seven below. <strong>The</strong> 12-storey frame doesn’t reach the yield limit state according to DBELA, and most of the<br />

analytical equivalent displacements are less than the capacity displacement, both fit well with the result of<br />

DBELA.<br />

From the analysis above, we can see that DBELA gives sound assessment results during the yield limit state.<br />

What’s more, we should realized that this simplified loss assessment method gives limited accuracy in that the<br />

height of the building increased by the storey height, so the errors less than half of the storey height are<br />

acceptable.<br />

Post-yield limit states<br />

A DBELA analysis on RC frame then will be carried out in the post-yield limit states which is very similar to<br />

the procedures in the yield limit state. First we choose the same reinforce concrete frame class as in the yield<br />

limit state and the geometric parameters are listed in Table 1. However, a group of more powerful artificially<br />

seismic waves with a PGA=0.75g are selected as the seismic input in the post-yield limit states. <strong>The</strong> average<br />

acceleration and displacement elastic response spectra (with a 5% damping ratio) of the 20 artificially waves are<br />

shown in Figure 7.<br />

Acceleration (cm/s/s)<br />

2000<br />

1500<br />

1000<br />

500<br />

average acc. spectrum<br />

target acc. spectrum<br />

PGA=0.75g<br />

Ta=0.15s<br />

Tb=0.5s<br />

Tc=4.0s<br />

0<br />

0 1 2 3 4 5 6<br />

Period (seconds)<br />

Displacement (cm)<br />

140<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

average disp. spectrum<br />

target disp. spectrum<br />

PGA=0.75g<br />

Ta=0.15s<br />

Tb=0.5s<br />

Tc=4.0s<br />

0<br />

0 1 2 3 4 5 6<br />

Period (seconds)<br />

Figure 7 Acceleration and displacement response spectra of 20 artificially waves used in the post-yield limit<br />

state<br />

Displacement (cm)<br />

70 μ=3.5<br />

60<br />

PGA=0.75g<br />

50<br />

40<br />

30<br />

T=5.2s<br />

H=27.79m<br />

20<br />

Capacity displacement<br />

10 T=0.78s<br />

Demand displacement<br />

H=4.17m<br />

0<br />

0 1 2 3 4 5 6 7 8<br />

Period (seconds)<br />

Displacement (cm)<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

μ=4<br />

PGA=0.75g<br />

T=1.59s<br />

H=7.95m<br />

Capacity dispalcement<br />

0<br />

0 1 2 3 4 5 6 7 8<br />

Period (seconds)<br />

T=4.75s<br />

H=23.75m<br />

Demand displacement<br />

-306-


Displacement (cm)<br />

70<br />

60<br />

50<br />

40<br />

30<br />

μ=4.5<br />

PGA=0.75g<br />

T=2.75s<br />

H=12.96m<br />

T=4.29s<br />

H=20.22m<br />

20<br />

Capacity displacement<br />

10<br />

Demand displacement<br />

0<br />

0 1 2 3 4 5 6 7 8<br />

Period (seconds)<br />

Displacement (cm)<br />

70 μ=5<br />

60<br />

PGA=0.75g<br />

50<br />

40<br />

30<br />

20<br />

10<br />

T=3.85s<br />

H=17.22m<br />

T=3.32s<br />

H=14.85m<br />

Capacity displacement<br />

Demand displacement<br />

0<br />

0 1 2 3 4 5 6 7 8<br />

Period (seconds)<br />

Figure 8 Results of DBELA using secant stiffness method under different ductility levels<br />

As we have already discussed, different equivalent linearization procedures can be used to generate the capacity<br />

and demand displacement curves during the post-yield limit states. For the first, a group of DBELA procedures<br />

using secant stiffness methodology suggested by H. Crowley and R. Pinho are carried out and the loss<br />

assessment results with the ductility factors increasing from 3.5 to 5.0 are plotted in Figure 8.<br />

Another DBELA procedures using equivalent stiffness linearization method are conducted with the ductility<br />

factors varies from 2.5 to 4.0. A Stiffness Degrading Model (STDG) is chosen to represent the RC frame, and<br />

the coefficients for equivalent linear parameters used in Equation (12)-(13) are listed in Table 2. <strong>The</strong> post-yield<br />

stiffness ratio α is obtained from a pushover analysis in this paper, then a linear interpolation is carried out to get<br />

the coefficients’ value. Finally, the results of DBELA which use this equivalent linearization method are plotted<br />

in Figure 9.<br />

Table 2 Coefficients for equivalent linear parameters for STDG, Equation (12)-(13).<br />

α stands for post-yield stiffness ratio<br />

α A B C D E F G H<br />

0% 0.1916 -0.0429 0.0652 -0.0179 0.0796 0.1820 0.1000 0.0129<br />

5% 0.1886 -0.0447 0.0654 -0.0180 0.1233 0.1502 0.1004 0.0132<br />

10% 0.1871 -0.0470 0.0682 -0.0193 0.1577 0.1257 0.0881 0.0164<br />

20% 0.1520 -0.0363 0.0646 -0.0180 0.1538 0.0924 0.1088 0.0075<br />

Displacement (cm)<br />

70<br />

μ=2.5<br />

PGA=0.75g<br />

60<br />

50<br />

40<br />

30<br />

T=5.99s<br />

H=46.77m<br />

20<br />

Capacity displacement<br />

10<br />

Demand displacement<br />

T=0.46s<br />

H=3.59m<br />

0<br />

0 1 2 3 4 5 6 7 8<br />

Period (seconds)<br />

Displacement (cm)<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

μ=3.0<br />

PGA=0.75g<br />

T=1.36s<br />

H=9.65m<br />

T=4.86s<br />

H=34.46m<br />

Capacity displacement<br />

Demand displacement<br />

0<br />

0 1 2 3 4 5 6<br />

Period (second)<br />

-307-


Displacement (cm)<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

μ=3.5<br />

PGA=0.75g<br />

T=2.87s<br />

H=19.07m<br />

T=4.08s<br />

H=27.11m<br />

Capacity displacement<br />

Demand displacement<br />

Displacement (cm)<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

μ=4<br />

PGA=0.75g<br />

Capacity displacement<br />

Demand displacement<br />

0<br />

0 1 2 3 4 5 6<br />

Period (seconds)<br />

0<br />

0 1 2 3 4 5 6<br />

Period (seconds)<br />

Figure 9 Results of DBELA using equivalent stiffness method (STDG)<br />

However, we should notice that the equivalent viscous damping ratio used in the secant stiffness procedure is<br />

deduced from a TF hysteretic model, but the latter procedure adopts a bilinear STDG model. That’s because<br />

solving the equivalent linear parameters is quite a complex and time-consuming procedure, even when a new<br />

statistical approach utilizing the technique of particle swarm optimization (PSO) is developed to determine the<br />

parameters of equivalent linearization method, very limited hysteretic models are available. Since the TT, STDG<br />

and PB hysteretic model have their area of complete cycle of stabilized force-displacement responses not differ<br />

much with each other. Actually they give much close outcomes which again prove that DBELA is not so<br />

sensitive to hysteretic models mentioned herein.<br />

<strong>The</strong>n the two kinds of loss assessment results using the secant and equivalent stiffness linearization method are<br />

compared together, and obliviously they give much different results and a simple comparison showed in Table 3.<br />

<strong>The</strong> results then will be checked by a group of inelastic dynamic time-history analysis the same as in the yield<br />

limit state.<br />

Table 3 Comparison of the DBELA results between using the secant and equivalent stiffness methodology<br />

Number of floors 4 5 6 7 8 9<br />

Total height 12m 15m 18m 21m 24m 27m<br />

Result of DBELA Secant stiffness 4~4.5 >5 4.5~5 4~4.5 3.5~4 3.5~4<br />

( μ ) Equivalent stiffness 3~3.5 3~3.5 3~3.5 3.5~4 3.5~4 3~3.5<br />

Ductility factor μ<br />

5.0<br />

4.5<br />

4.0<br />

3.5<br />

3.0<br />

2.5<br />

2.0<br />

1.5<br />

1.0<br />

0.5<br />

Average ducitility factor<br />

4-storey RC frame<br />

PGA=0.75g<br />

0.0<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially waves<br />

Ductility factor μ<br />

5.0<br />

4.5<br />

4.0<br />

3.5<br />

3.0<br />

2.5<br />

2.0<br />

1.5<br />

1.0<br />

0.5<br />

Average ductility factor<br />

5-storey RC frame<br />

PGA=0.75g<br />

0.0<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially waves<br />

-308-


Ductility factor μ<br />

5.0<br />

4.5<br />

4.0<br />

3.5<br />

3.0<br />

2.5<br />

2.0<br />

1.5<br />

1.0<br />

0.5<br />

Averge ductility factor<br />

6-storey RC frame<br />

PGA=0.75g<br />

0.0<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artifically number<br />

Ductility factor μ<br />

5.0<br />

4.5<br />

4.0<br />

3.5<br />

3.0<br />

2.5<br />

2.0<br />

1.5<br />

1.0<br />

0.5<br />

Average ductility factor<br />

7-storey RC frame<br />

PGA=0.75g<br />

0.0<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially waves<br />

Ductility factor μ<br />

7.0<br />

6.5<br />

6.0<br />

5.5<br />

5.0<br />

8-storey RC frame<br />

PGA=0.75g<br />

Average ductility factor<br />

4.5<br />

4.0<br />

3.5<br />

3.0<br />

2.5<br />

2.0<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially waves<br />

Ductility factor μ<br />

5.0<br />

4.5<br />

4.0<br />

3.5<br />

3.0<br />

2.5<br />

2.0<br />

1.5<br />

1.0<br />

0.5<br />

0.0<br />

Average ductility factor<br />

9-storey RC frame<br />

PGA=0.75g<br />

0 2 4 6 8 10 12 14 16 18 20<br />

Serial number of artificially waves<br />

Figure 10 Ductility levels of RC frames during inelastic dynamic time-history analysis<br />

In Figure 10 we could see that when the RC frames are getting taller, the results of time-history analysis are<br />

getting more scattered. In the 4-storey RC frame case that most ductility factors μ fall into the region between<br />

3.0 and 3.5 which is exactly the same with the results of DBELA using the equivalent stiffness method. In the<br />

5-storey RC frame case about half of the μ fall in the 3.0~3.5 region level, the average ductility level is a little<br />

lower than 3.0, which is, though, disagree with either of the method, however is much closer with the result of<br />

equivalent stiffness adopted method. <strong>The</strong> average ductility factors in the 6-storey frame case is exactly the same<br />

with the results obtained from the equivalent stiffness method and result of the 7-storey is much close to the<br />

results of equivalent stiffness method which condition is the much same with in the 5-storey frame case. Both of<br />

the two methods give the same loss assessment outcome (ductility factor from 3.5 to 4.0) to an 8-storey RC<br />

frame, and the average ductility factor is, the same with the 5 and 7-storey situation, a little lower than the<br />

DBELA’s result level. <strong>The</strong> analysis result of 9-storey frame shows that the equivalent stiffness adopted method<br />

give a more reasonable loss assessment outcome.<br />

From the figure and augments above, we could say that, during the post-yield limit states, DBELA using a<br />

secant stiffness equivalent linearization method gives a rather conservative loss assessment. And an equivalent<br />

stiffness method gives a more reasonable loss assessment outcome than that using the secant stiffness. Even<br />

though in some cases, such as 5- and 7-storey RC frame, the loss assessment result using equivalent stiffness<br />

method are kind of little conservative due to the random character of the seismic wave , the frames’ geometrical<br />

and material properties. DBELA gives sound loss assessment result in the post-yield limit states when the<br />

equivalent stiffness linearization is adopted.<br />

CONCLUSION<br />

DBELA is a simplified method to carry out the earthquake loss assessment in an urban scale. <strong>The</strong> results are<br />

checked by a series of time-history analysis, and DBELA gives sound assessment results during the yield limit<br />

state. Since a secant stiffness methodology is adopted in the equivalent linearization procedure, DBELA gives<br />

conservative results in the post-yield limits states. However, an equivalent stiffness method is chosen to replace<br />

the secant stiffness method. Corresponding inelastic dynamic time-history analyses are also carried out to judge<br />

the accuracy of DBELA using the equivalent stiffness method. Though there is a little offset of the results in<br />

some cases due to the random character of the seismic wave and the frames’ geometrical and material properties,<br />

-309-


it give rather reasonable results, or at least much more reasonable than that using a secant stiffness method.<br />

ACKNOWLEDEGMENT<br />

<strong>The</strong> project is sponsored by National Natural Science Foundation of China (No. 50908206).<br />

REFERENCES<br />

Bommer, J.J. and Elnashai, A.S. (1999). “Displacement spectra for seismic design”. Journal of Earthquake<br />

Engineering. 3(1), 1-32.<br />

Bommer, JJ., Elnashai, A.S. and Weir, A.G. (2000). “Compatible acceleration and displacement response spectra<br />

for seismic design codes”, Proceedings of 12 th World Conference on Earthquake Engineering, Auckland,<br />

Paper no.0207.<br />

Crowley, H. and Pinho, R. (2004). “Period-height relationship for existing European reinforced concrete<br />

buildings”. Journal of Earthquake Engineering, 8(Special Issue 1), 93-119.<br />

Crowley, H., Pinho, R., Bommer J.J. and Bird, J. F. (2006). Development of a Displacement-Based Method for<br />

Earthquake Loss Assessment, IUSS Press, Pavia, Italy<br />

Glaister, S. and Pinho, R. (2003). “Development of a simplified deformation-based method for seismic<br />

vulnerability assessment”. Journal of Earthquake Engineering, 7(Special Issue 1), 107-140.<br />

Iwan D. (1980). “Estimating inelastic response spectra from elastic spectra”. Earthquake Engineering and<br />

Structural Dynamics, 8(4), 375-388.<br />

Priestley, M. J. N. (1997). “Displacement-based seismic assessment of reinforced concrete buildings”. Journal<br />

of Earthquake Engineering, 1(1), 157-192.<br />

Priestley, M. J. N. Calvi, G.. M. and Kowalsky, M. J. (2007). Displacement-based seismic design of structures,<br />

IUSS Press, Pavia, Italy.<br />

Calvi. G. M. (1999). “A displacement-based approach for vulnerability evaluation of classes of buildings”.<br />

Journal of Earthquake Engineering, 3(3), 411-438<br />

SeismoSoft. (2005).“SeismoStruct – a computer program for static and dynamic nonlinear analysis of framed<br />

structures,” (online). Available from URL: www.seismosoft.com.<br />

Wang. Y. (2011). Research on equivalent linearization method applied in earthquake engineering, Zhejiang<br />

university.<br />

-310-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

IMPACT OF SPATIAL VARIABILITY ON SOIL SHEAR STRENGTH<br />

J. Ching 1 , K. K. Phoon 2 and Y. G. Hu 3<br />

1<br />

Associate Professor, Dept of Civil Engineering, National Taiwan <strong>University</strong>, Taipei, Taiwan.<br />

Email: jyching@gmail.com. Phone: 886-2-33664328. Fax: 886-2-23631558.<br />

2<br />

Professor, Dept of Civil Engineering, National <strong>University</strong> of Singapore, Singapore.<br />

3<br />

Ph.D. Student, Dept of Civil Engineering, National Taiwan <strong>University</strong>, Taipei, Taiwan.<br />

ABSTRACT<br />

<strong>The</strong> purpose of this study is to understand the mechanism of the effective shear strength for a soil mass in the<br />

presence of spatial variability. Random field finite elements are used to simulate the effective shear strength and<br />

the spatially averaging for the shear strength. Based on the simulation results, Vanmarcke’s theory is consistent<br />

with the spatially averaging for the shear strength. However, the effective shear strength is found to be close to<br />

the average shear strength along the actual slip curve, rather than the spatial averaging over the entire soil mass.<br />

KEYWORDS<br />

random field finite elements, Vanmarcke’s theory, soil shear strength, spatial variability.<br />

INTRODUCTION<br />

Soil properties in the field generally exhibit spatial variability. This is true even if the soil mass is nominally<br />

homogeneous, because small scale heterogeneity are always present due to natural geologic processes that create<br />

and continuously modify the soil in-situ. One important property is the shear strength of the soil. For most<br />

foundation engineering problems, resistances provided by soil mass are the “overall” shear strengths, which are<br />

typically related to spatial averaging over a certain region. Spatial variability is usually modeled by a<br />

homogeneous (or stationary) random field that can be characterized succinctly in a second-moment sense by a<br />

mean value at a point, a variance at a point, and a scale of fluctuation. Vanmarcke (1977) showed that the<br />

averaged property of a random field over a region has a mean value identical to the point mean, while the<br />

variance is less than the point variance.<br />

Vanmarcke’s definition of a spatial average is purely based on an integral of the random field over a given<br />

prescribed volume. <strong>The</strong> rationale is that the effective shear strength of a soil mass for a particular problem is<br />

governed by the spatial average along a slip curve and this spatial average is more relevant than the value at a<br />

point. <strong>The</strong>re appears to be an implicit assumption that the spatial average defined along a prescribed slip curve is<br />

comparable to the spatial average defined along a critical slip curve that depends on mechanics (equilibrium,<br />

compatibility, and constitutive relations) and boundary conditions. By definition, the critical slip curve is the<br />

curve producing the lowest factor of safety among all possible curves. In principle, it is clear that this critical<br />

curve is fundamentally different from an arbitrary trial slip curve that is prescribed rather than emerging as an<br />

outcome of a finite element or similar analysis. However, it is unclear at this stage if this fundamental difference<br />

would produce effective strengths that are significantly different from simple Vanmarcke-type spatial average<br />

strengths.<br />

<strong>The</strong> objective of this study is to elucidate this query. <strong>The</strong>re are limited studies in the literature that explore this<br />

query systematically. <strong>The</strong> outcome of this study is of practical significance, because it is computationally<br />

intensive to identify the critical slip curve and its associated effective strength. In contrast, the second-moment<br />

statistics of a Vanmarcke-type spatial average are available in closed-form.<br />

<strong>The</strong> above comparison is conducted through random field finite element analyses. A rectangular domain is<br />

divided into finite elements with shear strengths specified by realizations of a random field. <strong>The</strong> effective shear<br />

strength of the domain is determined by conducting plane-strain compression until failure. Discrepancies<br />

between the effective shear strength and the spatial average will be discussed. As to be expected, these<br />

discrepancies are mostly related to mechanical principles.<br />

-311-


RANDOM FIELD AND SPATIAL AVERAGING<br />

Stationary random fields are widely used primarily because the amount of measured data available in most site<br />

investigations would only permit characterization under this second-moment (weak) stationarity assumption. A<br />

two dimensional stationary random field for shear strength τ f (x,z) can be characterized by its point mean value<br />

E(τ f ), point variance Var(τ f ), and auto-correlation function. <strong>The</strong> auto-correlation function of a stationary random<br />

field τ f (x,z) is defined as the correlation between two locations with lag distance of Δx and Δz:<br />

COV ( τ<br />

f<br />

( x, z), τ<br />

f<br />

( x +Δ x, z +Δz))<br />

ρ( Δx, Δz) ≡ ρ( τ<br />

f<br />

( x, z), τ<br />

f<br />

( x+Δ x, z+Δ z))<br />

=<br />

(1)<br />

Var( τ ( x, z)) ⋅ Var( τ ( x +Δ x, z +Δz))<br />

where Var denotes variation; COV denotes covariance. An auto-correlation model widely used in geotechnical<br />

engineering literature is the single exponential model:<br />

ρ( Δx, Δ z) = exp( −k Δx −k Δ z)<br />

(1)<br />

x<br />

z<br />

where the parameter k x and k z are respectively equal to 2 divided by the scales of fluctuation (SOF) in the x and<br />

z directions, denoted by SOF x and SOF z . It is clear that the correlation decreases as Δx and Δz increase. This is<br />

consistent with measurements taken from natural soils: soil properties are strongly correlated within a small<br />

interval but are weakly correlated over a large interval. <strong>The</strong> SOF is the correlation length, i.e. the length scale<br />

within which two locations are significantly correlated.<br />

Vanmarcke (1977) pointed out that the spatial average of soil properties over a region D has a mean value<br />

identical to the point mean but has a variance smaller than the point variance. Let the region D be a rectangular<br />

domain defined by [x 0 x 0 +Δx] and [z 0 z 0 +Δz]. Mathematically, the spatial average over D can be defined as<br />

τ<br />

D<br />

f<br />

1<br />

x z<br />

x0+Δ x z0+Δz<br />

= ΔΔ<br />

∫ ∫<br />

x0 z0<br />

τ<br />

f<br />

( , )<br />

x z dzdx<br />

<strong>The</strong> variance of τ D f is smaller than the point variance Var(τ f ) due to the cancellation of variability through spatial<br />

D<br />

averaging. Vanmarcke (1977) further defined a variance reduction factor which is equal to the variance of τ f<br />

divided by the point variance:<br />

2 D<br />

(4)<br />

Γ ( D)<br />

= Var( τ<br />

f ) Var( τ<br />

f )<br />

Using the single exponential model in Eq. (2), Vanmarcke (1977) showed that<br />

2 2<br />

⎡<br />

2 2Δx ⎛ 2Δx ⎞⎤ 2Δx ⎡ 2Δz ⎛ 2Δz ⎞⎤<br />

2Δz<br />

Γ ( D) = ⎢ − 1+ exp⎜− ⎟⎥ × 1 exp<br />

2 ⎢ − + ⎜−<br />

⎟⎥<br />

2<br />

(5)<br />

⎣SOFx ⎝ SOFx ⎠⎦<br />

SOFx<br />

⎣SOFz ⎝ SOFz<br />

⎠⎦<br />

SOFz<br />

Γ 2 (D) is a decreasing function of Δx/SOFx and Δz/SOFz. <strong>The</strong>se latter two terms may be interpreted as the<br />

numbers of independent segments in the x and z directions.<br />

COMPARISON BETWEEN EFFECTIVE τ f AND ITS SPATIAL AVERAGING<br />

Although Vanmarcke’s theory provides useful closed-form solutions for spatial average, it is not clear if the<br />

effective shear strength is the same as τ f D in the first instance. To verify this, random field finite element analyses<br />

(FEA) are conducted. <strong>The</strong> region D is taken to be a plane-strain 36mx10m rectangular area (Δx = 10m, Δz = 36m)<br />

with 0.1m×0.1m FEA mesh grids. <strong>The</strong> total number of plane-strain elements is 36,000. <strong>The</strong> two lateral boundaries<br />

are free, the bottom boundary is roller, and the lower-left-most node is a hinge. <strong>The</strong> upper boundary is subjected to<br />

a compression stress. <strong>The</strong> unit weights of all elements are zeros, the Young’s modulus is 40 MN/m 2 , and the<br />

Poisson ratio is 0.3. <strong>The</strong> elastic properties hardly affect the failure load and a high Young’s modulus is selected for<br />

computational efficiency.<br />

<strong>The</strong> spatially varying shear strength τ f is simulated by stationary Gaussian random fields with a point mean E(τ f ) =<br />

50 kN/m 2 , with a point standard deviation Var(τ f ) 0.5 = 10 kN/m 2 , and with SOFx = SOFz = SOF. When assigning<br />

the simulated τ f to each element, the local averaging subdivision algorithm developed by Fenton and Vanmarcke<br />

(1990) is taken to conduct local averaging within each 0.1m×0.1m element. In this initial study, the shear strength<br />

τ f is assumed to be independent of the confining pressure for simplicity, i.e., φ = 0 o .<br />

f<br />

f<br />

(3)<br />

-312-


<strong>The</strong> spatial average of τ f over region D, namely τ f D , is therefore the average of the 36,000 assigned τ f values. A<br />

single realization of the random field will give a sample of τ f D . In this study, 120 realizations are taken for the<br />

following 13 chosen SOFs: SOF = 0.01m, 0.1m, 0.3m, 1m, 3m, 10m, 20m, 40m, 100m, 300m, 1000m, 3000m,<br />

and 10000m. <strong>The</strong> case with 104 m SOF is nearly a homogeneous case given the dimension of the domain. <strong>The</strong> left<br />

plot in Figure 2 shows the so-obtained τ f D samples, 120 samples for each SOF. In this plot, τ f D samples and SOF<br />

are normalized with respect to the point mean 50kN/m 2 and the mesh width 10m, respectively.<br />

Since τ f D is the spatial average of region D, it is expected that its statistical properties can be effectively estimated<br />

by Vanmarcke’s theory, i.e., the mean value of τ f D is the same as the point mean 50kN/m 2 , and the variance of τ f<br />

D<br />

is equal to the inherent variance 100 multiplied by the variance reduction factor in Eq. (5). <strong>The</strong> right plots in<br />

Figure 1 show the sample average and sample variance for the τ f D samples, 120 samples per chosen SOF. <strong>The</strong>y are<br />

normalized by the mean and variance estimated by Vanmarcke’s theory. It is clear that Vanmarcke’s theory is<br />

indeed effective for the spatial averaging τ f D , regardless the chosen SOF.<br />

In this study, the “effective” shear strength means the overall shear strength provided by the entire soil mass. <strong>The</strong><br />

effective shear strength can be obtained by actually shearing the soil mass to failure, which can be readily<br />

achieved in FEA. A normal compression stress is exerted at the upper boundary until non-convergence of the<br />

FEA, i.e., failure. <strong>The</strong> compression stress at failure divided by 2 is taken to be the “effective” shear strength of<br />

region D, denoted by τ f FEA . This FEA simulation is similar to the unconfined compression (UC) test in laboratory,<br />

except that it is in plane strain condition. <strong>The</strong> UC test is taken here because the initial confining pressure has no<br />

effect on τ f FEA due to the φ = 0 o assumption.<br />

Figure 1 Samples of τ f D , their average values and variances under different SOFs, normalized with respect to the<br />

estimation from Vanmarcke’s theory<br />

A single realization of random field gives a sample of τ FEA f . Similarly, 120 realizations are taken for the 13<br />

chosen SOFs. <strong>The</strong> left plot in Figure 2 shows the resulting τ FEA f samples, 120 samples for each SOF. In this plot,<br />

the same normalization used previously is taken for both τ FEA f and SOF. It is clear that τ FEA f samples behave<br />

differently from τ D f samples shown in Figure 2, especially when SOF is close to the width of D. <strong>The</strong> right two<br />

plots show the sample average and sample variance of the τ FEA<br />

f samples. Vanmarcke’s theory performs<br />

-313-


satisfactorily when SOF is large but slightly overestimates the sample mean when SOF is close to the width of D,<br />

and significantly underestimates the sample variance when SOF is small.<br />

COMPARISON BETWEEN EFFECTIVE τ f AND AVERAGING ALONG SLIP CURVE<br />

At the point of non-convergence of FEA, the stress states for all elements are recorded. A sophisticated<br />

algorithm is then taken in this study to identify the actual slip curve where shear failure occurs. This algorithm<br />

employs the safety factor (SF) defined by Pham and Fredlund (2003). A search algorithm based on the particle<br />

swarm optimization (PSO) (Kennedy and Eberhart 1995) is taken to find the curve with SF = 1, i.e., the actual<br />

slip curve.<br />

Figure 2 Samples of τ f D , their average values and variances under different SOFs, normalized with respect to the<br />

estimation from Vanmarcke’s theory<br />

<strong>The</strong> search starts from finding the straight line with the minimum SF by using PSO. This initial search for the<br />

optimal straight line (stage 1) only involves two degrees of freedom, as shown in Figure 3. <strong>The</strong>n, one more<br />

degree of freedom is added to the middle of the optimal straight line (second plot from the left), and this line is<br />

taken to be the initial solution for stage 2 of PSO. <strong>The</strong> end result for the PSO in stage 2 is a two-segment curve<br />

with minimum SF. More degrees of freedom are added to this optimal curve, and a new stage (stage 3) of PSO is<br />

taken. This process is continued until full degree of freedom is reached. <strong>The</strong> final optimal curve typically has SF<br />

less than 1.01 and coincides with the plastic zone predicted by the FEA.<br />

<strong>The</strong> same algorithm is executed after each FEA simulation of the UC test. <strong>The</strong> elements that the slip curve<br />

passes through are identified, and the assigned τ f values of these elements are then averaged with weights<br />

proportional to the traversing lengths. This average value is therefore the averaging along the slip curve, denoted<br />

by τ f slip . <strong>The</strong> left plot in Figure 4 shows the resulting τ f slip samples, and the right two plots show the sample<br />

average and sample variance of the τ f slip samples, normalized by the sample mean and sample variance of τ f<br />

FEA<br />

samples. It is clear that the overall shear strength τ f FEA has mean and variance that are similar to those of τ f slip ,<br />

except the minor deviations in the variance when SOF is 0.01-0.03 times the width of D.<br />

<strong>The</strong> consistency shown in Figure 4 leads to the following conclusion: the effective/overall shear strength of a<br />

region D is close to the average shear strength along the actual slip curve, not the spatial averaging over the<br />

-314-


entire region D. Note that the latter is close to what is estimated by Vanmarcke’s theory, as seen in Figure 1.<br />

<strong>The</strong>refore, it can be concluded that Vanmarcke’s theory may not be suitable for estimating the mean and<br />

variance of the effective/overall shear strength of a region D in this simple plane strain example.<br />

<strong>The</strong> minor difference between τ f slip and τ f FEA is further explored. It is found that τ f FEA is typically larger than τ f slip ,<br />

and the magnitude of difference seems to correlate well with the irregularity of the actual slip curve. This is<br />

reasonable because the overall shear strength τ f FEA should be the composition of the average shear strength<br />

along the slip curve τ f slip and the dilation effect produced by the irregularity.<br />

Figure 3 Schematics for the various stages of the search algorithm<br />

Figure 4 Samples of τ f slip , their average values and variances under different SOFs, normalized with respect to<br />

the sample average and variance of τ f<br />

FEA<br />

CONCLUSIONS<br />

Although the spatial averaging of shear strength over D has similar statistical behaviors to those estimated by<br />

Vanmarcke’s theory, the statistical behaviors of the effective/overall shear strength are NOT consistent to those<br />

estimated by Vanmarcke’s theory. However, the statistical behaviors of the effective shear strength are close to<br />

-315-


those of the average shear strength along the actual slip curve. <strong>The</strong> effective shear strength is found to be<br />

typically slightly larger than the average shear strength along the actual slip curve. <strong>The</strong> difference is believed to<br />

be due to the asperity of the actual slip curve, i.e., the overall shear strength is the composition of the average<br />

shear strength along the slip curve and the dilation effect induced by the asperity.<br />

In the case where Vanmarcke’s theory is taken to estimate the statistical behaviors of the effective shear strength,<br />

the theory generally overestimates the mean value of the overall shear strength but underestimates the variance.<br />

<strong>The</strong> overestimation of the mean value gets worse when the scale of fluctuation is close to the size of the soil<br />

mass, while the underestimation of the variance gets worse when the scale of fluctuation is small.<br />

REFERENCES<br />

Kennedy, J. and Eberhart, R. (1995). “Particle swarm optimization”, Proceedings of IEEE International<br />

Conference on Neural Networks, Vol IV, 1942–1948.<br />

Pham, H.T.V. and Fredlund, D.G. (2003). “<strong>The</strong> application of dynamic programming to slope stability analysis”,<br />

Canadian Geotechnical Journal, 40, 830-847.<br />

Fenton, G.A. and Vanmarcke, E.H. (1990). “Simulation of random fields via local average subdivision”, Journal<br />

of Engineering Mechanics, ASCE, 116(8), 1733-1749.<br />

Vanmarcke, E.H. (1977). “Probabilistic modeling of soil profiles”, Journal of Geotechnical Engineering<br />

Division, ASCE ,103(11), 1227-1246.<br />

-316-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

ADSORPTION AND DESORPTION BEHAVIOR OF BIVALENT NICKEL AND<br />

MANGANESE IONS ON LOESS SOIL<br />

Y. Wang 1 , X. W. Tang 1, * , H. Y. Wang 1 and Z. F. Sun 1<br />

1<br />

MOE Key Laboratory of Soft Soils and Geoenvironmental Engineering, Zhejiang <strong>University</strong>, China.<br />

* Corresponding author. Email: tangxiaowu@zju.edu.cn<br />

ABSTRACT<br />

Both nickel and manganese are trace heavy metals. Toxic effect will be caused by excessive intake of these<br />

heavy metals. <strong>The</strong> control of heavy metal discharge into environment has been paid much attention recently.<br />

Loess soil was used to study the behavior of removing Ni(II) and Mn(II) from aqueous solutions. <strong>The</strong> kinetic<br />

data fit the pseudo-second order kinetics model very well, and the adsorption rate of Mn(II) was faster than that<br />

of Ni(II). <strong>The</strong> adsorption capacity of loess soil for Ni(II) and Mn(II) was determined to be 13.84 and 7.65 mg g -1<br />

respectively. <strong>The</strong> D-R model fit the isothermal data of Ni(II) and Mn(II) adsorption on loess soil best, and the<br />

main adsorption mechanism was ion exchange for both the heavy metals. NTA had contribution to desorption of<br />

heavy metals from loess soil. <strong>The</strong> maximum desorption ratio of Ni(II) and Mn(II) from loess soil was 63 % and<br />

92% respectively, and heavy metal ions could be more easily desorbed from the loess soil loaded with less<br />

heavy metal. <strong>The</strong>se results have significant reference value for loess soil as a material removing heavy metals<br />

and recycling use of heavy metal contaminated loess.<br />

KEYWORDS<br />

Loess soil, nickel, manganese, adsorption, desorption.<br />

INTRODUCTION<br />

1. Heavy metal contamination in environments has aroused a lot of public attention around the world.<br />

Accumulation of excessive amounts of trace elements (such as nickel and manganese) in soils and waters<br />

can result in severe toxic effects to human and ecosystems. Nickel and manganese can be released into<br />

environments by many different ways such as mine smelting, industrial waste disposal, electroplating and<br />

battery manufacturing processes (Kadirvelu et al. 2008). Heavy metal ions can easily be bioaccumulated and<br />

spread along the food chain once entering soil or water. <strong>The</strong> level of heavy metals in environments needs<br />

limiting to allowable value.<br />

2.<br />

Many researchers show interest in working on the technology of removing heavy metals from environment, and<br />

adsorption method is popular due to its effectiveness and convenience. Chinese loess soil with abundant clay<br />

minerals may be a potential promising material for removing heavy metals. Removal of Cu(II), Zn(II), Pb(II)<br />

and Cd(II) from aqueous solutions using Chinese loess have recently been reported and the loess shows high<br />

affinities for heavy metals (Tang et al. 2008ab; Li et al. 2009; Wang et al. 2009). In this study, adsorption<br />

behavior of loess soil towards Ni(II) and Mn(II) was investigated. In order to recycle the loess soil contaminated<br />

by heavy metals, NTA (nitrilotriacetic acid) which is environmental-friendly was used to desorb heavy metals<br />

from loess soil.<br />

MATERIALS AND METHODS<br />

Preparation of Materials<br />

Chinese loess soil samples were taken from the suburban area of Xi’an, located in northwestern China. <strong>The</strong><br />

loess soil was typical Quaternary loess on the Chinese Loess Plateau. <strong>The</strong> air-dried loess soil was oven-dried at<br />

105 o C for 24 h for removing bulk water, cooled to room temperature, and then sealed in plastic bags for storage.<br />

<strong>The</strong> basic parameters and chemical compositions of loess are summarized in Table 1 and 2. <strong>The</strong> reagent nickel<br />

chloride hexahydrate (NiCl 2·6H 2 O) and manganese chloride tetrahydrate (MnCl 2·4H 2 O) used in this study were<br />

of analytical grade. Ni(II) and Mn(II) stock solution (1 g L -1 ) was separately prepared in deionized water (DW).<br />

Erlenmeyer flasks and centrifuge tubes were pre-treated, immersed in 0.01 M HNO 3 solution for 24 h firstly, and<br />

-317-


then rinsed three times with deionized water.<br />

Table 1 Basic parameters of loess<br />

Organic<br />

surface area<br />

CEC<br />

(mg g -1 Specific density<br />

)<br />

(m 2 g -1 )<br />

(cmol(+) kg -1 )<br />

pH na pH pzc<br />

5.5 2.75 24.1 11.2 9.5 2.82<br />

Table 2 Chemical compositions of loess (percent mass content)<br />

SiO 2 Al 2 O 3 CaO MgO K 2 O Fe 2 O 3 Na 2 O FeO<br />

63.68 12.77 9.56 3.14 3.01 2.74 2.35 0.89<br />

Adsorption Experiments<br />

Kinetics experiments<br />

Batch test method was used to study the adsorption behavior of Ni(II) and Mn(II) adsorption on loess soil. <strong>The</strong><br />

adsorbent dosage was fixed at 10 g L -1 , and both the initial concentration of Ni(II) and Mn(II) was set as 50 mg<br />

L -1 . <strong>The</strong> samples were placed into a thermostatic box agitated at 160 rpm at 25 o C. <strong>The</strong> reaction times were set in<br />

the range of 15 to 1440 min. Homologous samples were taken out at specific time intervals. <strong>The</strong>n, the slurries<br />

were centrifuged at 3000 rpm for 5 min to obtain the supernatant. <strong>The</strong> atomic absorption spectroscopy (AAS)<br />

was then used to determine the concentration of Ni(II) and Mn(II).<br />

Isotherms experiments<br />

<strong>The</strong> loess dosage was fixed at 10 g L -1 and the initial solute concentration ranged from 25 to 200 mg L -1 for both<br />

Ni(II) and Mn(II). All samples were equilibrated for 24 h in the thermostatic box at 25 o C, and the equilibrium<br />

concentrations of Ni(II) and Mn(II) were measured by AAS method.<br />

Desorption experiments<br />

<strong>The</strong> concentration of Ni(II) and Mn(II)was 170 mg g -1 and 130 mg g -1 respectively. <strong>The</strong> loess dosage was 10 g<br />

L -1 . <strong>The</strong> successive desorption was conducted, and the sorption test was carried out at first, followed by a<br />

four-step desorption. At each desorption step, the concentration of NTA was 1, 2, 4 and 5 mmol L -1 . At the end<br />

of each step, the supernatants after centrifuged were sampled to test the solute concentrations.<br />

RESULTS AND DISCUSSION<br />

Adsorption Kinetics<br />

Figure 1 shows the effect of contact time on concentrations of Ni(II) and Mn(II) in aqueous solutions. <strong>The</strong> initial<br />

concentration of Ni(II) and Mn(II) was 50 and 56.2 mg L -1 separately tested by AAS method. As shown in the<br />

figure, both the concentrations of nickel and manganese decreased rapidly at the initial 15 min. In the following<br />

time, the concentration of Ni(II) decreased very slowly and basicly remained constant, while, the concentration<br />

of Mn(II) showed a little up and downs.<br />

Figure 1 Effect of contact time on the removal of Ni(II) and Mn(II) by loess soil<br />

<strong>The</strong> kinetic data obtained from the batch experiments were analyzed using two kinetic equations (i.e., the<br />

pseudo-first order kinetics and the pseudo-second order kinetics) to interpret the rate law of Ni(II) and Mn(II)<br />

-318-


adsorption on loess.<br />

<strong>The</strong> pseudo-first order equation can be written as (Ho 2004)<br />

kt 1<br />

qt<br />

= qe(1 − e − )<br />

(1)<br />

where q e (mg g -1 )and q t (mg g -1 ) are the amounts of solute adsorbed per unit mass at equilibrium and at any time<br />

t (min) respectively, and k 1 (min -1 ) is the pseudo-first order adsorption rate constant.<br />

<strong>The</strong> pseudo second order equation can be expressed as (Ho and Mckay 1999)<br />

t 1 1<br />

= + t<br />

(2)<br />

2<br />

qt k2qe qe<br />

where k 2 (g mg -1 min -1 ) is the adsorption rate constant of pseudo-second order.<br />

<strong>The</strong> fitting curves simulated by the pseudo-first order kinetics and the pseudo-second order kinetics are shoun in<br />

Figures 2 and 3. Table 3 shows the predicted kinetic parameters for adsorption of Ni(II) and Mn(II) on loess soil.<br />

<strong>The</strong> pseudo-first order kinetics and the pseudo-second order kinetics both fit the test data very well according to<br />

the high correlation coefficients. However, the pseudo-second order kinetics fit the data better than the<br />

pseudo-first order kinetics in that the correlation coefficient 0.998 and 0.993 was higher than 0.933 and 0.944<br />

respectively. <strong>The</strong> adsorption amount per unit mass (q e ) obtained by the pseudo-second order equation for Ni(II)<br />

and Mn(II) was 3.46 and 2.36 mg g -1 respectively. <strong>The</strong> adsorption rate constant k 2 for Ni(II) and Mn(II) was 0.01<br />

and 0.12 g mg -1 min -1 , respectively, indicating that the adsorption rate of Mn(II) was faster than that of Ni(II).<br />

Figure 2 Fitting curves of Ni(II) and Mn(II)<br />

adsorption on loess with the pseudo-first order<br />

kinetic model<br />

Figure 3 Fitting curves of Ni(II) and Mn(II)<br />

adsorption on loess with the pseudo-second order<br />

kinetic model<br />

Table 3 Kinetic parameters for adsorption of Ni(II) and Mn(II) on loess soil<br />

Pseudo-first order kinetics<br />

Pseudo-second order kinetics<br />

q e (mg g -1 ) k 1 (min -1 ) R 2 q e (mg g -1 ) k 2 (g mg -1 min -1 ) R 2<br />

Ni(II) 3.16 0.10 0.933 3.46 0.01 0.998<br />

Mn(II) 2.44 0.12 0.944 2.36 0.12 0.993<br />

Adsorption Isotherms<br />

Figure 4 shows adsorption isotherms of Ni(II) and Mn(II) on loess soil at 25 o C. <strong>The</strong> adsorption isotherms for<br />

Ni(II) and Mn(II) can be assigned “L” type according to the definition by Giles and Smith (1974). Both heavy<br />

metal ions adsorbed on per unit loess increased with increasing the concentration of solute in the aqueous<br />

solution. Three mathematical models (i.e., Langmuir, Freundlich and Dubinin-Radushkevich (D-R) models)<br />

were used to analyze the isothermal data.<br />

<strong>The</strong> Langmuir adsorption isotherm equation is written as (Do 1998)<br />

Ce<br />

Ce<br />

1<br />

= + (3)<br />

qe<br />

Q bQ<br />

where q e (mg g -1 ) and Q (mg g -1 ) denote the adsorption amount at equilibrium and the monolayer adsorption<br />

capacity respectively, C e (mg L -1 ) the equilibrium concentration in solution, b (L mg -1 ) the Langmuir constant.<br />

-319-


<strong>The</strong> Freundlich model can be expressed as (Do 1998)<br />

1<br />

ln qe = ln KF + ln Ce<br />

(4)<br />

n<br />

where K F (mg g -1 ) is the Freundlich constants related to the adsorption capacity and intensity, and n is the<br />

heterogeneity factor.<br />

<strong>The</strong> D-R model can be written as (Do 1998)<br />

2<br />

ln qe<br />

= ln qm<br />

− kε<br />

(5)<br />

where q m (mol g -1 )is the maximum adsorption capacity, k the model constant related to the free adsorption<br />

energy and ε is the Polanyi potential related to the equilibrium concentration as follows<br />

⎛ 1 ⎞<br />

ε = RT ln ⎜1+<br />

⎟<br />

⎝ Ce<br />

⎠<br />

where the unit of C e should be translated into mol L -1 .<strong>The</strong> mean free energy of adsorption E is calculated by<br />

E<br />

1<br />

=− (7)<br />

2k<br />

In general, the adsorption is physical adsorption when |E| is between 1.0 and 8.0 kJ mol -1 , while, the adsorption<br />

mechanism is basically surface adsorption by ion exchange when |E| is between 8.0 and 16.0 kJ mol -1<br />

(Kilislioglu and Bilgin 2003; Özcan et al. 2006).<br />

Figures 5-7 show the fitting curves simulated by the above three isotherm models and the isothermal parameters<br />

for Ni(II) and Mn(II) adsorption on loess soil are presented in Table 4. <strong>The</strong> Langmuir, Freundlich and D-R<br />

models were all found to fit the test data very well according to their high coefficients of determination. <strong>The</strong><br />

D-R model showed a relatively better fit with the test results for its highest coefficients.<br />

(6)<br />

Fiurge 4 <strong>The</strong> adsorption isotherms of Ni(II) and<br />

Mn(II) on loess<br />

Figure 6 Fitting curves of Ni(II) and Mn(II)<br />

adsorption on loess with Freundlich model<br />

Figure 5 Fitting curves of Ni(II) and Mn(II)<br />

adsorption on loess with Langmuir model<br />

Figure 7 Fitting curves of Ni(II) and Mn(II)<br />

adsorption on loess with D-R model<br />

-320-


Table 4 Predicted isothermal parameters for Ni(II) and Mn(II) adsorption on loess soil<br />

Langmuir model Freundlich model D-R model<br />

Q<br />

(mg g -1 )<br />

b<br />

(L mg -1 )<br />

R 2<br />

K F<br />

(mg g -1 )<br />

n R 2 q m<br />

(mg g -1 )<br />

K<br />

(mol 2 kJ -2 )<br />

E<br />

(kJ mol -1 )<br />

Ni(II) 13.84 0.037 0.944 1.00 1.91 0.973 39.07 0.005 -9.84 0.987<br />

Mn(II) 7.65 0.022 0.950 0.53 2.03 0.964 19.68 0.006 -9.53 0.973<br />

<strong>The</strong> monolayer adsorption capacity of loess soil for Ni(II) and Mn(II) estimated by the Langmuir model was<br />

13.84 and 7.65 mg g -1 respectively. <strong>The</strong> Freundlich constant n for Ni(II) and Mn(II) was 1.91 and 2.03<br />

respectively, greater than unity, indicating some degree of heterogeneity of adsorption system. <strong>The</strong> q m predicted<br />

with the D-R model was 39.07 and 19.68 mg g -1 for Ni(II) and Mn(II) respectively, larger than the adsorption<br />

capacity Q obtained from the Langmuir model. Because the D-R model describes an ideal state which is almost<br />

impossible to realize. <strong>The</strong> adsorption energy for Ni(II) and Mn(II) adsorption was estimated to be -9.84 and<br />

-9.53 kJ mol -1 , respectively. <strong>The</strong> absolute values of the adsorption energy |E| all lied within 8-16 kJ mol -1 ,<br />

implying an ion exchange mechanism.<br />

Desorption of Heavy Metals from Loess<br />

Figure 8 shows the variation of adsorption amount (C s ) of Ni(II) and Mn(II) on loess at different concentration<br />

(C NTA ) of NTA. <strong>The</strong> initial amount of Ni(II) and Mn(II) on loess was 13.68 and 4.48 mg g -1 respectively. Both<br />

amount of the heavy metals decreased with increasing the concentration of NTA, indicating that NTA<br />

contributed to desorption of heavy metals from loess soil. <strong>The</strong> variation of desorption efficiencies (R de ) of Ni(II)<br />

and Mn(II) from loess at different C NTA is shown in Figure 9. In general, both R de increased with increasing C NTA .<br />

<strong>The</strong> desorption ratio of Ni(II) increased to 63 % at C NTA =4 mmol L -1 , and then had a slight decrease at C NTA =5<br />

mmol L -1 . While, the maximum desorption ratio of Mn(II) was 92 % at C NTA =5 mmol L -1 , much more than that<br />

of Ni(II). As can be seen from the result, the initial amount of heavy metal loaded on loess had effect on the<br />

desorption efficiency. <strong>The</strong> maximum desorption ratio of Mn(II) was greater than that of Ni(II), but the initial<br />

amount of Ni(II) was more than that of Mn(II). <strong>The</strong>refore, heavy metal ions could be more easily desorbed from<br />

the loess soil loaded with less heavy metal.<br />

R 2<br />

Figure 8 <strong>The</strong> amount of Ni(II) and Mn(II) on<br />

loess at different concentration of NTA<br />

Figure 9 Desorption efficiency of Ni(II) and Mn(II)<br />

from loess at different concentration of NTA<br />

CONCLUSIONS<br />

(1) <strong>The</strong> pseudo-second order kinetics fit the data of Ni(II) and Mn(II) adsorption on loess soil, and the<br />

adsorption rate of Mn(II) was faster than that of Ni(II).<br />

(2) <strong>The</strong> adsorption capacity of loess soil for Ni(II) and Mn(II) was 13.84 and 7.65 mg g -1 respectively.<br />

(3) <strong>The</strong> D-R model fit the isothermal data of Ni(II) and Mn(II) adsorption on loess soil best, and the main<br />

adsorption mechanism was ion exchange for both the heavy metals.<br />

(4) <strong>The</strong> maximum desorption ratio of Ni(II) and Mn(II) from loess soil was 63 % and 92% respectively,<br />

and heavy metal ions could be more easily desorbed from the loess soil loaded with less heavy metal.<br />

ACKNOWLEDGEMENTS<br />

<strong>The</strong> authors would like to express their sincere gratitude to Research Fund for the Doctoral Program of Higher<br />

Education of China (20090101110075) and Scholarship Award for Excellent Doctoral Student granted by<br />

-321-


Ministry of Education for their financial support of this study.<br />

REFERENCES<br />

Do, D. D. (1998). “Adsorption Analysis: Equilibrium and Kinetics”, Imperial College Press, London.<br />

Giles, C.H. and Smith, D.A. (1974). “A general treatment and classification of the solute sorption isotherms”,<br />

<strong>The</strong>oretical. Journal of Colloid and Interface Science, 47, 755-765<br />

Ho, Y. S. (2004). “Citation review of Lagergren kinetic rate equation on adsorption reactions”, Scientometrics,<br />

59, 171-177.<br />

Ho, Y. S. and McKay, G. (1999). “Batch lead (II) removal from aqueous solution by peat: equilibrium and<br />

kinetics”, Process Safety and Environmental Protection B, 77, 165-173.<br />

Kadirvelu, K., Thamaraiselvi, K. and Namasivayam, C. (2001). “Adsorption of nickel(II) from aqueous solution<br />

onto activated carbon prepared from coirpith”, Separation and Purification Techenology, 24, 497-505.<br />

Kilislioglu, A. and Bilgin, B. (2003). “<strong>The</strong>rmodynamic and kinetic investigations of uranium adsorption on<br />

amberlite IR-118H resin”, Applied Radiation and Isotopes, 58, 155-160.<br />

Li, Z. Z., Tang X.W., Chen Y.M. and Wang Y. (2009). “Sorption behavior and mechanism of Pb(II) on Chinese<br />

loess”, Journal of Environmental Engineering-ASCE, 135, 58-67.<br />

Özcan, A., Öncü, E. M. & Özcan, A. S. (2006). “Kinetics, isotherm and thermodynamic studies of adsorption of<br />

Acid Blue 193 from aqueous solutions onto natural sepiolite”, Colloids and Surfaces A-Physicochemical and<br />

Engineering Aspects, 277, 90-97.<br />

Tang, X. W., Li, Z. Z. and Chen, Y. M. (2008a). “Behaviour and mechanism of Zn(II) adsorption on Chinese<br />

loess at dilute slurry concentrations”, Journal of Chemical Technology and Biotechnology, 83, 673-682.<br />

Tang, X. W., Li, Z. Z., Chen, Y. M. and Wang, Y. (2008b). “Removal of Cu(II) from aqueous solution by<br />

adsorption on Chinese Quaternary loess: Kinetics and equilibrium studies”, Journal of Environmental<br />

Science Part A-Toxic/Hazardous Substances & Environmental Engineering, 43, 779-791.<br />

Wang, Y., Tang, X.W., Chen, Y. M., Zhan, L.T., Li, Z.Z. and Tang, Q. (2009). “Adsorption behavior and<br />

mechanism of Cd(II) on loess soil from China”, Journal of Hazardous Materials, 172, 30-37.<br />

-322-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

UNCERTAINTY IN NONLINEAR SEISMIC GROUND RESPONSE ANAYLYSES<br />

O. L. A. Kwok<br />

Department of Civil Engineering,<br />

National Taiwan <strong>University</strong>, Taipei, Taiwan. Email: annieonleikwok@ntu.edu.tw<br />

ABSTRACT<br />

One-dimensional seismic ground response analyses are seldom performed using nonlinear procedures because<br />

parameter selection and code usage protocols are poorly documented and understood, the effect of parametric<br />

variability on the analysis results is generally unknown, and the benefits of nonlinear analysis relative to the<br />

widely-used equivalent-linear analysis are un-quantified and unclear. In this paper, effect of parametric<br />

variability on predictions using nonlinear ground response analysis will be discussed. Sources of variability<br />

considered include material properties such as dynamic material curves and velocity profiles. Various vertical<br />

array sites are used in the study. It is found that the standard deviation of predictions is dominated by material<br />

curve variability at small periods and velocity variability at medium to large periods respectively.<br />

KEYWORDS<br />

nonlinear, ground response analysis, one-dimensional, uncertainty.<br />

INTRODUCTION<br />

<strong>The</strong>oretical modeling of one-dimensional site response can generally be accomplished using equivalent linear or<br />

nonlinear analysis. Equivalent-linear ground response modeling is commonly performed in practice as it<br />

requires the specification of well-understood and physically meaningful input parameters (shear-wave velocity,<br />

unit weight, modulus reduction and damping). Nonlinear ground response analyses provide a more accurate<br />

characterization of the true nonlinear soil behavior, but their implementation in practice has been limited, which<br />

is principally a result of poorly documented and unclear parameter selection and code usage protocols.<br />

Moreover, previous studies have thoroughly investigated the sensitivity of site response results to the input<br />

parameters in equivalent-linear analyses (e.g. Roblee et al., 1996), but this level of understanding is not<br />

available for the input parameters used in nonlinear analyses. <strong>The</strong> objective of current study is to evaluate the<br />

uncertainties in predictions by nonlinear ground response modeling due to various sources of variability,<br />

including material properties and modeling schemes. Vertical array sites are used in this study as the uncertainty<br />

due to input ground motion can be eliminated and predictions from nonlinear analyses can be compared directly<br />

to the data (recordings).<br />

SOURCES OF VARIABILITY<br />

Modeling Schemes<br />

<strong>The</strong> ground response analyses are performed by using various nonlinear codes including DEEPSOIL (Hashash<br />

and Park 2001, 2002; Park and Hashash 2004), D-MOD_2 (Matasovic 2006), a ground response module in the<br />

OpenSees simulation platform (Ragheb 1994; Parra 1996; Yang 2000; McKenna and Fenves 2001;<br />

opensees.berkeley.edu), SUMDES (Li et al. 1992), and TESS (Pyke 2000). All of these codes are time-domain<br />

analysis methods which allow soil properties within a given layer to change with time as the strains in that layer<br />

change. Like a structure, the layered soil column is idealized either as a multiple degree of freedom lumped<br />

mass system or a continuum discretized into elements with distributed mass with appropriate boundary<br />

conditions. Soil material models employed range from relatively simple cyclic stress-strain relationships<br />

(DEEPSOIL, D-MOD_2 and TESS) to advanced plasticity models (OpenSees and SUMDES).<br />

Material Properties<br />

<strong>The</strong> four vertical array sites considered in this study include KGWH02 (from the Japanese Kiknet network of<br />

strong motion stations), Lotung (Taiwan), Turkey Flat and La Cienega (both of them are in California of the<br />

-323-


United States.) <strong>The</strong> available site data include velocity logs, boring logs and site-specific material curves<br />

(except the Kiknet site).<br />

Shear Wave Velocity Profile<br />

<strong>The</strong> median velocity profiles are estimated based on available geophysical measurements and used as the<br />

baseline geotechnical model. Variability in the velocity profiles is estimated either from data (if multiple<br />

measurements at the same site are available) or based on the empirical model of Toro (1997). <strong>The</strong> Toro model is<br />

based on 176 velocity profiles from the Savannah River site, and is a statistical model that can be used to<br />

estimate the standard deviation and correlation in velocity profiles for both generic (broad geographic region)<br />

and site-specific conditions. In current analyses, the alternative velocity profiles are obtained by adding and<br />

subtracting to the baseline profile 3 times the depth-dependent site-specific standard deviation. <strong>The</strong> 3<br />

factor is used because the theoretical optimal three-point representation of a normal distribution involves<br />

sampling the distribution at the mean (μ) and μ ± 3σ<br />

, and then providing the samples with weights of 2/3 and<br />

1/6 (twice). Those sample points and weights preserve the first (mean), second (variance), and fourth central<br />

moments of the underlying distribution (Rosenblueth, 1975; Ching et al., 2006).<br />

Modulus Reduction and Damping Curves<br />

<strong>The</strong> modulus reduction and damping curves are used to model the nonlinear soil behavior for different depth<br />

ranges. <strong>The</strong> median (baseline) curves used in current analyses are based on either material-specific laboratory<br />

testing, inference of in-situ material property from vertical array data or Darendeli (2001) statistical model.<br />

Variability in material curves is considered by adding and subtracting to the baseline curves 3 times the<br />

strain-dependent standard deviation. This standard deviation is taken either from Darendeli (2001) or estimated<br />

from in-situ data.<br />

RESULTS AND DISCUSSIONS<br />

Ground response analyses are performed for the vertical array sites using various nonlinear ground response<br />

analysis codes with both baseline (median) and alternative (taking into account the variability in material<br />

properties) geotechnical models.<br />

To evaluate the model-to-model variability for the ground surface results for period T, the median estimate<br />

ln( S a<br />

( T ))<br />

is first evaluated from the five nonlinear model predictions. Model variability, σ m , is then calculated<br />

from the variance as follows:<br />

∑<br />

⎡ln ( S ) ( ) 2<br />

a( T) − ln S ( )<br />

,<br />

a<br />

T ⎤<br />

⎣<br />

pre i<br />

⎦<br />

2<br />

i<br />

σ<br />

m( T) = Var( Sa( T)<br />

) =<br />

(1)<br />

pre<br />

N −1<br />

where N = number of predictions (five).<br />

<strong>The</strong> response variability due to material uncertainty is assessed using DEEPSOIL predictions only. To calculate<br />

the standard deviation due to velocity, ground motions are predicted based on two non-baseline velocity profiles<br />

(mean + 3 standard deviation velocities and mean - 3 standard deviation velocities). <strong>The</strong> standard<br />

deviation of the ground motions due to the variability in velocity (denoted σ v ) is estimated according to the<br />

FOSM, method (Baker and Cornell, 2003; Melchers, 1999), as follows:<br />

3<br />

2<br />

σ<br />

v<br />

= ∑ wi(ln( Sa( T) i<br />

−ln( Sa( T))<br />

(2)<br />

where<br />

i=<br />

1<br />

Sa( T ) = S ( T )<br />

Sa( T ) = S ( T )<br />

1 a Vs<br />

: μ<br />

2 a Vs: μ+<br />

3σVs<br />

Sa( T ) = S ( T )<br />

3 a Vs: μ−<br />

3σV s<br />

3<br />

a<br />

= ∑ i a i<br />

i=<br />

1<br />

ln( S ( T)) w ln( S ( T) )<br />

w1 = 2/3; w2 = w3<br />

= 1/6<br />

<strong>The</strong> standard deviation due to the variability in material curves (denoted σ G ) is estimated similarly to σ v .<br />

In Figure 1, uncertainties in predictions due to different sources of variability for all four vertical array sites are<br />

(3)<br />

-324-


plotted as a function of period (left frame) and period ratio (right frame, period ratio = T/T s where T s = elastic<br />

site period). Variability of predictions due to material curve uncertainty seems to be most pronounced at periods<br />

less than 0.5 sec and has no clear association with the site period. Moreover, material curve uncertainty only<br />

produces significant response variability for relatively thick site profiles – it is not a significant issue for the<br />

Turkey Flat, which is a shallow soil site.<br />

<strong>The</strong> effect of velocity variability can have a strong influence on the predictions near the elastic site period.<br />

However, this strong influence is only observed for sites with large impedance contrast (Turkey Flat and<br />

KGWH02), which dominates the site response in those cases. This is shown in Figure 1 by a peak in the σ v term<br />

near T/T 1 = 1.0. Such a peak does not occur for the La Cienega or Lotung sites, which have a gradual variation<br />

of velocity with depth and no pronounced impedance contrast.<br />

Model-to-model variability is most pronounced at low periods, where the differences result principally from<br />

different damping formulations. Given the modest ground motions at the investigated sites, it is expected that<br />

variations in the viscous damping formulations are principally driving this variability. This large variability at<br />

low periods may be due to the fact that the predictions from SUMDES, which had only the simplified Rayleigh<br />

damping formulation at the time these predictions were made, are much lower than the predictions from codes<br />

with full Rayleigh damping formulation.<br />

1<br />

Overall Variability, σ<br />

(ln unit)<br />

Material Curve Variability, σG<br />

(ln unit)<br />

Velocity Variability, σ V<br />

(ln unit)<br />

Model Variability, σ m<br />

(ln unit)<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0.01 0.1 1 10<br />

Period, T (sec)<br />

Turkey Flat<br />

Kiknet KGWH02<br />

La Cienega<br />

Lotung<br />

0.001 0.01 0.1 1 10 100<br />

Period Ratio, T/Ts<br />

Figure 1 Comparison of variabilities across three vertical array sites<br />

CONCLUSIONS<br />

In this paper, the effect of variability in material properties and modelling scheme on predictions by nonlinear<br />

analysis codes is illustrated. It is found that the standard deviation of predictions is dominated by material curve<br />

variability at small periods. <strong>The</strong> effect of velocity variability can be significant at medium to large periods,<br />

especially when the sites have significant impedance contrast.<br />

-325-


ACKNOWLEDGMENTS<br />

Financial support for this work was provided by PEER Lifelines project No. 2G02, which is sponsored by the Pacific<br />

Earthquake Engineering Research Center’s Program of Applied Earthquake Engineering Research of Lifeline Systems.<br />

<strong>The</strong> PEER Lifelines program, in turn, is supported by the State Energy Resources Conservation and Development<br />

Commission and the Pacific Gas and Electric Company. This work made use of Earthquake Engineering Research<br />

Center’s Shared Facilities supported by the National Science Foundation under Award No. EEC-9701568. In addition,<br />

the support of the California Department of Transportation’s PEARL program is acknowledged.<br />

REFERENCES<br />

Baker, J.W. and Cornell ,C.A. (2003). “Uncertainty specification and propagation for loss estimation using<br />

FOSM methods”, Proceedings, 9th International Conference on Applications of Statistics and Probability<br />

in Civil Engineering, ICASP9, San Francisco, California: Millpress., Rotterdam, Netherlands, 8p.<br />

Ching, J., Porter, K.A. and Beck, J.L. (2004). “Uncertainty propagation and Feature selection for loss estimation<br />

in performance-based earthquake engineering”, EERL Report 2004-02, Earthquake Engineering Research<br />

Laboratory, California Institute of Technology, Pasadena, California.<br />

Darendeli, M. (2001). “Development of a new family of normalized modulus reduction and material damping<br />

curves.” Ph.D. <strong>The</strong>sis, Dept. of Civil Eng., Univ. of Texas, Austin.<br />

Hashash, Y.M.A. and Park, D. (2001). “Non-linear one-dimensional seismic ground motion propagation in the<br />

Mississippi embayment”, Eng. Geology, Amsterdam, 62(1-3), 185-206.<br />

Hashash, Y.M.A. and Park, D. (2002). “Viscous damping formulation and high frequency motion propagation in<br />

nonlinear site response analysis”, Soil Dynamics and Earthquake Engrg., 22(7), 611-624.<br />

Li, X.S., Wang, Z.L. and Shen, C.K. (1992). SUMDES: A nonlinear procedure for response analysis of<br />

horizontally-layered sites subjected to multi-directional earthquake loading, Dept. of Civil Eng., Univ. of<br />

Calif., Davis.<br />

Matasovic, N. (2006). “D-MOD_2 – A Computer Program for Seismic Response Analysis of Horizontally<br />

Layered Soil Deposits, Earthfill Dams, and Solid Waste Landfills.” User’s Manual, GeoMotions, LLC,<br />

Lacey, Washington, 20 p. (plus Appendices).<br />

McKenna, F. and Fenves, G.L. (2001). “<strong>The</strong> OpenSees command language manual, version 1.2.” Pacific<br />

Earthquake Engrg. Research Center, Univ. of Calif., Berkeley. (http://opensees.berkeley.edu).<br />

Melchers, R.E. (1999). “Structural Reliability Analysis and Prediction”, John Wiley and Sons, Chichester.<br />

Park, D. and Hashash, Y.M.A. (2004). “Soil damping formulation in nonlinear time domain site response<br />

analysis”, Journal of Earthquake Engineering,8(2), 249-274.<br />

Parra, E. (1996). “Numerical modeling of liquefaction and lateral ground deformation including cyclic mobility<br />

and dilation response in soil systemsm”, PhD <strong>The</strong>sis, Dept. of Civil Engrg., Rensselaer <strong>Polytechnic</strong> Institute,<br />

Troy, NY.<br />

Pyke, R.M. (2000). “TESS: A computer program for nonlinear ground response analyses”, TAGA Engineering<br />

Systems & Software, Lafayette, Calif.<br />

Ragheb, A. M. (1994). “Numerical analysis of seismically induced deformations in saturated granular soil<br />

strata”, PhD <strong>The</strong>sis, Dept. of Civil Eng., Rensselaer <strong>Polytechnic</strong> Institute, Troy, NY.<br />

Roblee, C.J., Silva, W.J., Toro, G.R. and Abrahamson, N.A. (1996). “Variability in site-specific seismic ground<br />

motion design predictions”, ASCE Geotech. Special Publication No. 58, Uncertainty in the Geologic<br />

Environment: From <strong>The</strong>ory to Practice, C.D. Shakelford, P.P. Nelson (eds.), 2, 1113-1133.<br />

Rosenblueth, E. (1975). “Point estimates for probability moments”, Proc. Nat. Acad. of Sci., 72(10), 3812-3814<br />

Toro, G. R. (1997). ‘‘Probabilistic models of shear-wave velocity profiles at the Savannah River site, South<br />

Carolina”, Rep. to Pacific Engineering and Analysis, El Cerrito, Calif.<br />

Yang, Z. (2000). “Numerical modeling of earthquake site response including dilation and liquefaction”, PhD<br />

<strong>The</strong>sis, Dept. of Civil Eng. and Eng. Mech., Columbia <strong>University</strong>, NY, New York.<br />

-326-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

AN ACCURATE GEOLOGICAL MODEL IS AN ESSENTIAL REQUIREMENT FOR<br />

GEOTECHNICAL ENGINEERING<br />

--- A CASE STUDY ON THE GEOLOGY OF TUEN MUN TO TIN SHUI WAI AREA,<br />

HONG KONG<br />

K. W. Lai 1 and H. M. S. Chan 2<br />

1 previous <strong>Hong</strong> <strong>Kong</strong> Geological Survey, Civil Engineering Department, Government of HKSAR, China<br />

2 Housing Department, Government of <strong>Hong</strong> <strong>Kong</strong> SAR, China<br />

ABSTRACT<br />

In geotechnical engineering practice, an accurate geological model is an essential requirement for safe and<br />

economical geotechnical engineering design. After a review of geological investigation reports and geological<br />

publications in the Tuen Mun to Tin Shui Wai area, it can be seen that the current adopted rock classification is<br />

flawed. For instance, the tuff breccia has been misinterpreted as conglomerate; the marble clast-bearing andesite<br />

has been misidentified as siltstone with layered marble. Such misinterpretations may lead to the formation of an<br />

unreliable geological model. As there will be substantial development of mega-projects in the area, it is prudent<br />

to address the field work method and problematic rock classifications. Correction of these rock classifications<br />

would enable cost saving and shortening of development programme.<br />

KEYWORDS<br />

Tuen Mun Formation, andesite, plug, vent breccia.<br />

INTRODUCTION<br />

To gain a thorough understanding of pyroclastic rocks in the Tuen Mun to Tin Shui Wai area is a lengthy and oscillating<br />

process due to the complexity in the formation of this rock type. <strong>The</strong> study of the Tuen Mun Formation, especially the<br />

marble clast-bearing tuff breccia, can be divided into three stages:<br />

Groping Stage.<br />

From 1971 to mid-1989, Allen and Stephens (1971) divided the strata of Tuen Mun to Tin Shui Wai area into<br />

two formations. <strong>The</strong> Repulse Bay Formation in the south comprised mainly sedimentary rock and water-lain<br />

volcanic rock. <strong>The</strong> Lok Ma Chau Formation in the north composed meta-sedimentary and volcanic rock. Due to<br />

the insufficient borehole data and rock outcrops, tuff breccia has not been recognized for a long period. <strong>The</strong><br />

geology of Tin Shui Wai still retained the Lok Ma Chau Formation name up to mid-1989, such as Geological<br />

Map of San Tin of 1:20,000 scale (GCOa 1989) Geological Maps of Tin Shui Wai of 1:5,000 scale (GCOb<br />

1989). Arthurton et al (GCO 1988) studied the strata in the area and defined the andesitic strata as Tuen Mun<br />

Formation which is fundamentally correct, but he misidentified the tuff breccia as conglomerate and assigned<br />

the strata as Tsing Shan Formation which is flawed (Figure 1).<br />

Understanding Stage<br />

From mid-1989 to 2000, after some 40 drillholes were sunk in the area by GCO, the marble clast-bearing tuff<br />

breccia of the Tin Shui Wai area was confirmed as volcanic in origin. This resulted in the renaming of the Lok<br />

Ma Chau Formation to Tuen Mun Formation (Darigo 1990), (Frost 1990) and (GEO 1994) Subsequently, Irfan<br />

(GCO1990) conducted a rock strength properties study; he also confirmed the volcanic origin. After further<br />

detailed study, the Tsing Shan Formation has been merged into Tuen Mun Formation as they both belong to the<br />

same rock type (GEO 2000) (Figure 2).<br />

Re-examination Stage.<br />

<strong>The</strong> updated geological result was included in “<strong>The</strong> Pre-Quaternary Geology of <strong>Hong</strong> <strong>Kong</strong>” and its related geological map.<br />

<strong>The</strong> preparation of the geological map was undertaken by Kirk et al (GEO 2000) whilst other geologists prepared the written<br />

report. Normally, report writing should refer to the updated geological map. Unfortunately, under the description of stratum,<br />

-327-


the author has quoted the outdated Tsing Shan Formation data (GCO 1988) (Figure 5), such as the rock in lower part of<br />

Tuen Mun Formation sedimentary rock predominates (Sewell et al 2000.p69); Also. Figure 3.5 of the report misinterprets<br />

the pyroclastic rock in the Tin Shui Wai area as Lok Ma Chau Formation. Both of these examples are contradictory with the<br />

more accurate geological mapping represented in the 1:100,000 scale geological map of <strong>Hong</strong> <strong>Kong</strong> and the 1:20,000 scale<br />

Solid Geological Map of San Tin (GEO 1994). We regretted that we have not found these problems before publishing the<br />

report. Up to 2003 a considerable borehole logs of Tin Shui Wai area has misidentified the pyroclastic rock as sedimentary<br />

rock. It has seriously affected the geotechnical engineering design in the study area. Those problems have been pointed out<br />

by the authors and other geologists in some published papers such as (Chan, J. et al 2005), (Chan & Kwong 2009), (Lai et al<br />

2004) and (Li 2010). Nevertheless, the problems were not rectified. Subsequently, misinterpretations of rock classifications<br />

have also been carried over into other geological publications. For instance, in “Engineering Geological Practice in <strong>Hong</strong><br />

<strong>Kong</strong>” the tuff breccia is misidentified as conglomerate (GEO 2007, Figs. 3.2.3, 5.5.1 and 6.2.7) (Figures 3-4) and the<br />

“<strong>Hong</strong> <strong>Kong</strong> Geology Guide Book” (GEO 2008 p52-59) misinterpreted the tuff breccia of Tsing Shan Monastery as<br />

comprising conglomerates and breccias which may have formed following the collapse of a volcano or by fluvial activity on<br />

the margin of a crater lake. Such errors seriously affect the geological model and subsequent geotechnical engineering<br />

design.<br />

COMMON CHARACTERISTICS OF VOLCANIC PLUGS<br />

In order to develop a thorough understanding of the Tuen Mun Formation volcanic rocks and plugs, detailed fieldwork<br />

and sample inspection has been conducted including re-mapping of five plugs, collection of 180 rock samples,<br />

identification of 105 thin sections and chemical analysis of 20 rock samples. In addition to the local study, plugs and<br />

volcanoes in other parts of the world were also visited, such as Ship Rock and Smith Rock in America (Figures 6-7), the<br />

Pinatubo and Taal volcanoes in Philippines, Changbaishan (Figure 8) and Tengchong in China, as well as volcanoes in<br />

Italy, Argentina, Chile, New Zealand and Iceland. Important characteristics of andesite plugs have been identified as<br />

follows:<br />

i<br />

ii<br />

iii<br />

iv<br />

v<br />

vi<br />

<strong>The</strong> shape of plug is circular or elliptical in plan, and cylindrical in cross-section forming positive landform.<br />

<strong>The</strong> plug is composed of tuff breccia and/or lava.<br />

Lithic clasts are cemented by lava forming an altered rim between the lithic clast and the lava.<br />

Autobreccia and vertical lava flow occur within the plug.<br />

Forming vertical, irregular tabular joints parallel to the upstanding plug body or forming concentric circle joints.<br />

A chain of plugs controlled by a concealed fault forming a photolineament.<br />

REGIONAL GEOLOGY OF TUEN MUN TO TIN SHUI WAI AREA<br />

<strong>The</strong> Tuen Mun Formation occurs in the NW New Territories between two NNE trending faults from Tuen Mun passing<br />

through Tin Shui Wai towards Shenzhen. <strong>The</strong> granite occurs in the west whereas the Carboniferous sedimentary rock<br />

outcrops in the east. <strong>The</strong>y are composed of basaltic andesite, dacite lava, tuff breccia and tuff with minor tuffaceous<br />

sandstone. <strong>The</strong> Ar-Ar dating of andesite indicates ages of 181±3 Ma and 182±3.5 Ma respectively which are of Middle<br />

Jurassic age. Since most plugs in the world belong to the Cenozoic Era, this would make the Mesozoic plugs some of the<br />

oldest in the world. <strong>The</strong>se rocks have been subject to weathering for a long period. Several hundred metres of volcanic<br />

rock have been eroded with only the bottom and root parts of the volcanoes remaining. Based on the volcanic facies, they<br />

can be divided into:<br />

i<br />

ii<br />

iii<br />

iv<br />

v<br />

Explosive facies – violent explosion with accumulated tuff breccia and lapilli tuff surrounding the crater.<br />

Effusive facies – gentle volcanic eruption mainly forming lava flow in the valley and plain.<br />

Vent facies – including plug and dyke formed by tuff breccia and lava vein.<br />

Airfall facies – pyroclastic fragments fallen from eruption cloud forming stratovolcanic deposits.<br />

Eruptive-sedimentary facies – minor tuffaceous sandstone occurring in the ridge of west Por Lo Shan.<br />

CHARACTERISTICS OF PYROCLASTIC ROCK AND PLUGS OF TUEN MUN FORMATION<br />

<strong>The</strong> Tuen Mun Formation occurs in the Tuen Mun valley. It contains a total of nine plugs ranging from 20m to 120m in<br />

diameter extending NNW from Tsing Shan Monastery to north Por Lo Shan. <strong>The</strong> misidentification of these rocks is due<br />

to lacking study on the mode of occurrence of rock body; no chemical analysis on the cement material of tuff breccia and<br />

failing to differentiate the layered conglomerate and vent-shaped volcanic plug.<br />

Tsing Shan Monastery Plug.<br />

Recently the temple has been refurbished and 11 boreholes were sunk. After studying the borehole data and re-mapping<br />

-328-


Figure 1 A layer of Tsing Shan Formation<br />

conglomerate shown in the 1:20,000 scale geological<br />

map (GCO 1988)<br />

Figure 2 <strong>The</strong> Tsing Shan Formation was deleted in<br />

the 1:100,000 scale geological map (GEO 2000)<br />

Figure 3 In “Engineering geological practice in <strong>Hong</strong><br />

<strong>Hong</strong> <strong>Kong</strong>” (GEO 2007 p.35) the vent breccia is<br />

misinterpreted as conglomerate<br />

Figure 4 A series of volcanic plugs is shown in the<br />

Tsing Shan - Por Lo Shan area (Lai 2005, 2009)<br />

-329-


Figure 5 <strong>The</strong> geological map of 2000 applied<br />

obsoleted data resulting in the difference from the<br />

map of 1994<br />

Figure 6 Volcanic plug of Ship Rock, New<br />

Mexico, .S. A<br />

Figure 7 Volcanic plug of Ship Rock is composed<br />

of vent breccia and lava<br />

Figure 8 Vent breccia and lava occur in the Tianchi<br />

Crater, Changbaishan, Jilin Province, China<br />

Figure 9 Volcanic plug of Tsing Shan Monastery<br />

Figure 10 <strong>The</strong> vent breccia intruded the crystal tuff<br />

Tsing Shan Monastery<br />

the area, the temple is situated in a 120m long and 50m wide rain drop shaped plug. It is composed of andesine lava and<br />

vent breccia/tuff breccia. <strong>The</strong> main minerals are plagioclase and hornblende with minor biotite, augite, quartz and sphene<br />

as well as secondary minerals of epidote, chlorite and sericite. Most rocks are aphanitic texture, occasionally the prism<br />

-330-


plagioclase is up to 3mm long. Lithic clasts contain tuff and andesite of an early eruptive stage, sandstone, siltstone and<br />

marble of country rock. <strong>The</strong>y are subangular to subrounded, 10mm to 300mm across and cemented by lava. <strong>The</strong> results<br />

of chemical analysis of lava in the plug indicated that two samples are basaltic trachyandesite and three samples are<br />

dacite. Contact metamorphism took place between the lava and lithic clast. <strong>The</strong> ferromagnesian clast formed<br />

chloritization or epidotization, the marble clast formed skarnization. <strong>The</strong> lava flow structure has been misinterpreted as<br />

evidence of bedding or banded strata. Surrounding the plug is crystal ash tuff as confirmed by thin section identification.<br />

<strong>The</strong> relation between the plug and tuff is an intrusive contact (Figures 9-15).<br />

Shan King Estate Plug.<br />

This plug is located at 600m west of Shan King Estate. It is elliptical in shape and cylindrical in cross-section, up to 80m<br />

long and 50m wide forming a 50m high steep cliff. It is a good exposure composed of tuff breccia and lava vein. <strong>The</strong> tuff<br />

breccia represents the violent eruption whereas the lava vein indicates the quiescent eruption. <strong>The</strong> lithic clast is cemented<br />

by andesite lava with an altered rim between their contact. <strong>The</strong> plug is associated with two parasitic plugs and some<br />

dykes in nearby area. <strong>The</strong> lava flow structure and very fine degassing pipes are parallel to the vertical plug. <strong>The</strong> vertical<br />

cracks produced an irregular columnar or tabular jointing (Figure 18). <strong>The</strong> plug has been misinterpreted as conglomerate<br />

and breccia strata which were originally laid down on the flanks of a volcano and have since been orientated vertically by<br />

deformation associated with movement on the Tsing Shan Fault (<strong>Hong</strong> <strong>Kong</strong> Geology Guide book, GEO 2008 p59).<br />

Shan King Estate Parasitic Plug.<br />

Also referred to as the smaller plug. This plug forms a 50m diameter, 20m high sub-circle, located 80m west of Shan<br />

King Estate and 40m apart from Shan King Estate Plug. <strong>The</strong> plug is composed of lava and tuff breccia. <strong>The</strong> lithic clast is<br />

greyish white subrounded to rounded, up to 800mm across and is cemented by lava. Chemical analysis indicated that the<br />

lava is mainly basaltic andesite and the lithic clast is rhyolitic ash tuff congealed in early stage. Tuff breccia occurs in the<br />

core of the plug. A special feature is that the plug reveals a concentric circle joint pattern (Figure 16-17).<br />

South Por Lo Shan Plug.<br />

Two plugs are situated 550m southwest of Por Lo Shan. <strong>The</strong> south one is sub-elliptical in shape and is 85m long, 50m<br />

wide and approximately 55m high. This plug comprises lava and tuff breccia of multiple eruptive stages. (Figures 19-21).<br />

<strong>The</strong> chemical analysis result indicates the lava is mainly andesitic dacite while the lithic clasts have andesite congealed in<br />

early stage, sandstone, siltstone and marble. When the lithic clast contacts the hot magma, the margin of marble may alter<br />

to skarn, tremolite or silicified marble. When subject to ductile deformation, the lithic clast may be flattened or elongated.<br />

<strong>The</strong> chemical analysis also shows the country rock surrounding the plug is rhyolitic ash tuff.<br />

North Por Lo Shan.<br />

<strong>The</strong> plug is composed of multiple eruptive tuff breccia and andesite -dacite lava. It forms a 95m long, 20m high ellipse<br />

which is 50m apart from the South Por Lo Shan Plug. <strong>The</strong> lithic clasts contain marble, sandstone and lava of early stage<br />

which were cemented by andesite. <strong>The</strong> plug intruded the ash crystal tuff forming a steep cliff. Lava flow structure is<br />

evident. This plug has been misinterpreted as conglomerate (Figures 22-23) in “Engineering geological practice in<br />

<strong>Hong</strong> <strong>Kong</strong>” (GEO 2007 p40 and Figure.6.2.7.).<br />

Stratovolcanic deposits at Ling To Monastery.<br />

Interbedded tuff breccias, tuff, and fine-grained lava with localized tuffite form cyclic deposits. A large boulder reveals<br />

five decomposition cycles with inclined bedding. <strong>The</strong> rocks contain the angular, subrounded and shuttle shaped lithic<br />

clasts. <strong>The</strong> lava vein and lava spatter show flow structure. <strong>The</strong> shuttle shaped clast and lava spatter suggest that those<br />

rocks are airfall deposits on the flanks of a nearby crater comprising parts of a stratovolcano. Some colluvial boulders<br />

with stratovolcanic strata can be seen from Por Lo Shan to Ling To Monastery (Figures 24-26) but these are difficult to<br />

be found in the in-situ outcrop.<br />

PROBLEMS IN TUEN MUN – TIN SHUI WAI SITE INVESTIGATION BOREHOLE LOGGING<br />

More than 1000 borehole logs from the Tuen Mun to Tin Shui Wai area were reviewed in 2003 of which, 60% borehole<br />

logs are believed to have problems in rock identification. <strong>The</strong> major problems were misclassifying the marble<br />

clast-bearing pyroclastic rock as layered marble interbedded with siltstone. This rock formation adjoins the marble strata<br />

of Yuen Long Formation, which contains sink hole and karst features which can seriously affect the safety of<br />

foundations. <strong>The</strong>refore, within marble area, it is necessary to drill at least 20m into sound marble rock during site<br />

investigation. On the contrary, if the volcanic rock is encountered, the requirement is to sink borehole to at least 5m into<br />

grad III or better rock.<br />

-331-


Figure 11 Rock cores of vent breccia and andesite<br />

lava, Tsing Shan Monastery<br />

Figure 12 Altered rim between andesite and marble<br />

clast from core sample at Tsing Shan Monastery<br />

Figure 13 Andesite in thin section under the<br />

polarization microscope<br />

Figure 14 Vent breccia cemented by andesite within<br />

a plug which is not conglomerate<br />

Figure 15 Vent breccia with lava flow in a plug were<br />

misjudged as fine-grained banded strata between<br />

breccia/conglomerate<br />

Figure 16 Circular joints in a parasitic plug, Shan<br />

King Estate<br />

-332-


Figure 17 Subrounded tuff clasts cemented by lava<br />

within the core of a parasitic plug, Shan King Estate<br />

Figure 18 Vertical joints and lava flow structure of<br />

the Shan King Estate Plug<br />

Figure 19 South Por Lo Shan Plug composed of<br />

vent breccia and lava veins<br />

Figure 20 Intrusion of lava vein into the vent<br />

breccia in South Por Lo Shan Plug<br />

Figure 21 Lava flow structure in south Por Lo Shan<br />

Plug<br />

Figure 22 Conglomerate was misjudged in North Por<br />

Lo Shan<br />

-333-


Due to the misidentification of rock type, a total of 9,809m drillcores of 553 boreholes are believed to have exceeded the<br />

drilling requirement. <strong>The</strong>se would have made a substantial cost and time implication in the development areas. Also,<br />

since the marble clasts were cemented by lava, the marble clast would be dissolved and the lava will remain which only<br />

forms the honeycomb structure, unlike the sink holes in carboniferous marble which were formed as a result of the<br />

complete dissolution of marble. <strong>The</strong> engineering geological properties of both layered marble and marble clast-bearing<br />

tuff breccia are very different.<br />

CHEMICAL COMPOSITION OF TUEN MUN FORMATION VOLCANIC ROCKS<br />

<strong>Hong</strong> <strong>Kong</strong> Geological Survey has conducted 11 chemical analyses on the rocks of Tuen Mun Formation (Langford<br />

1989 and Sewell 1997). <strong>The</strong>se analyses were carried out on samples collected from outside of craters; no samples were<br />

collected from plugs. Due to the aphanitic texture, the rock is very difficult to be identified with naked eyes...<br />

In accordance with the recommendation of International Union of Geological Sciences, 19 rock samples were collected<br />

from different volcanic facies including: plug, dyke, airfall and effusive facies. All chemical results were plotted in the<br />

TAS (total alkali-silica) diagram to classify different rock types (Figure 27). With respect to the determination of accurate<br />

chemical composition, only clean rock samples which were free of metamorphism, weathering, hydrothermal alteration<br />

and mixed lithic clast could be used for chemical analysis. Laboratories in <strong>Hong</strong> <strong>Kong</strong>, Guangzhou and England were<br />

used to conduct the analysis.<br />

<strong>The</strong> results of the chemical analyses indicate the pattern of volcanic eruption. <strong>The</strong> tuff breccia in the plug indicates the<br />

violent eruption, whereas the andesite lava is widespread in the valley and plain representing the quiescent eruption. <strong>The</strong><br />

crystal ash tuff and tuff breccia were found in the airfall facies surrounding the crater forming the stratovolcanic<br />

volcanoes. <strong>The</strong> primary magma is probably the basalt rising from the mantle which becomes more silica rich by<br />

contamination of the basalt with crustal rock of higher silica content forming andesite to dacite.<br />

GEOMECHANICAL BEHAVIOR OF TUEN MUN FORMATION<br />

It is easy to misidentify the tuff breccia as conglomerate due to the intense rock decomposition. Nevertheless, the two<br />

rock types are very different in terms of geomechanical characteristics. Some geomechanics classification were<br />

conducted on the Tuen Mun Formation by Irfan (GCO 1990) and Chan and Kwong (2009) in which the uniaxial<br />

compressive strength and the elastic modulus of both rocks were very different and they would therefore have different<br />

load deformation behavior. <strong>The</strong> uniaxial compressive strength of both rock types are tabulated in Table 1 for reference.<br />

<strong>The</strong> marble clast-bearing tuff breccia acquired a UCS about 5 to 10 times stronger than that of sedimentary rock, and 2 to<br />

3 times stronger than that of marble. <strong>The</strong> UCS of siltstone is approximately 30 MPa which is about 3 to 6 times smaller<br />

than that of fine ash tuff. <strong>The</strong> errors made in rock identification could therefore result in a major impact in the<br />

engineering design as these rocks have different geomechanical behavior. It is therefore strongly suggested to classify the<br />

rock types carefully with reference to geomechanical characterization as well.<br />

Table 1 Uniaxial Compression Test Results for Intact Rock Samples<br />

Marble clast Calcareous Clayey<br />

Fine ash crystal Marble<br />

-bearing tuff conglomerate Conglomerate Tuff<br />

breccia (MPa) (MPa)<br />

(MPa)<br />

(MPa)<br />

(MPa)<br />

a. 150 – 296 1 9.3 – 31.2 3 4.0 -27.4 4 c. 111 - 194 2 d. 65 - 138 1<br />

b. 195 - 329 2<br />

Origin: a. Tin Shui Wai b. Tsing Shan Temple c. Tuen Mun d. Yuen Long<br />

Source of Origin: 1 GCO ( 1990),<br />

2<br />

Chan and Kwong (2009), 3 Fugro Rock Mechanics Lab.<br />

4<br />

FIGG Rock Mechanics Lab<br />

Only intact rock samples were tested<br />

CONCLUSIONS<br />

Over the past 40 years, many geologists have spent much of their precious time to conduct detailed field mapping and<br />

survey in Tuen Mun, Yuen Long and Tin Shui Wai areas. <strong>The</strong> knowledge of volcanic rock has gradually expanded to<br />

reach a comprehensive level of understanding. A change in rock type from stratified conglomerate to cylindrical shape<br />

tuff breccia has been identified. Finally, a chain of palaeovolcanic plugs has been exposed and verified. <strong>The</strong> experience<br />

of reviewing the rock types in this area, highlights the importance of carrying out detailed field mapping to accurately<br />

observe the mode of occurrence of such rocks and conducting chemical analysis for aphanitic rocks determine their<br />

chemical composition as recommended by IUGS, since identification by naked eye alone is extremely difficult.<br />

-334-


Figure 23 Tuff breccia and lava vein forming a<br />

volcanic plug in North Por Lo Shan<br />

Figure 24 Sketch showing volcanic plug and<br />

stratovolcanic deposits<br />

Figure 25 Stratovolcanic deposits at Ling To<br />

Monastery<br />

Figure 26 Sketch of the stratovolcanic deposit of<br />

Ling To Monastery<br />

Figure 27 Chemical composition of volcanic rocks in the Tuen Mun Formation<br />

-335-


ACKNOWLEDGEMENT<br />

<strong>The</strong> author would offer sincere thanks to Profs. K Y Tao K Y & G F Xing of Nanjing Institute of Geology &<br />

Mineral Resources and Prof. P. Robinson of Dalhousie <strong>University</strong>, Canada who have assisted to conduct field<br />

survey and thin section studies. Special thanks would also be given to L M Lee, B Y Wang, and J.Chan of <strong>Hong</strong><br />

<strong>Kong</strong> <strong>University</strong> who have helped to prepare thin sections and chemical analyses. and also thanks to R J Sewell<br />

for providing chemical analysis data. <strong>The</strong> author would also appreciate the assistance provided by Dr R Wong<br />

of <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong>, Mr Chung of <strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> to conduct geomechanicws tests to<br />

rock samples and J. Cunningham for her valuable opinion.<br />

REFERENCES<br />

Allen, P.M., and Stephens, E.A. (1971). Report on the geological survey of <strong>Hong</strong> <strong>Kong</strong>, Government Printer, 107p.<br />

Arthurton, R.S., Lai, K.W., and Shaw, R. (1988). Solid and superficial geology map of Tsing Shan (Castle Peak), Map<br />

HGM20 Sheet 5. 1:20,000, Geotechnical Control Office, <strong>Hong</strong> <strong>Kong</strong> Government.<br />

Chan, S.H.M., and Kwong, A.K.L. (2009). “Enhanced and systematic classification system for rock types in Tuen Mun<br />

and Tin Shui Wai Area, <strong>Hong</strong> <strong>Kong</strong>”, Proceedings of ISRM-sponsored International Symposium on Rock Mechanics,<br />

Rock Characterization, Modeling and Engineering Design Methods.<br />

Chan, J. Lai, K.W., and Chan L.S. (2005). “Origin of a chain of clast-bearing cone-shaped rocks within the Tuen Mun<br />

andesite in Tsing Shan, Tuen Mun, <strong>Hong</strong> <strong>Kong</strong>.”, Abstracts of the Conference on Recent Advance sin Geologically<br />

Research of <strong>Hong</strong> <strong>Kong</strong> and the Pearl River Mouth Region, edited by Switzer, A.D. and Duzgoren-Aydin, N.S.<br />

<strong>Hong</strong> <strong>Kong</strong> <strong>University</strong>. 21-22.<br />

Darigo, N.J. (1988). Deep Bay: Solid Geology, HGP 5B, Sheet 2-SW-C. 1:5,000 Geotechnical Control Office, Civil<br />

Engineering Services Department.<br />

Darigo, N.J. (1990). “Marble-bearing Jurassic volcanics of the western New Territories, <strong>Hong</strong> <strong>Kong</strong>”, Karst geology in<br />

<strong>Hong</strong> <strong>Kong</strong> Geological Society of <strong>Hong</strong> <strong>Kong</strong> Bulletin, edited by Langford R. D., Hansen, A., and Shaw, R., No.4,<br />

61-72.<br />

Fletcher, C. J. N. (2004). Geology of Site Investigation boreholes from <strong>Hong</strong> <strong>Kong</strong>, Double Helix. <strong>Hong</strong> <strong>Kong</strong>. 132 p.<br />

Frost, D.V. (1988). Shan Pui: Solid Geology, HGP5B, Sheet 2-SW-D. 1:5,000 Geotechnical Control Office, Civil<br />

Engineering Services Department.<br />

Frost, D.V. (1990). “Geological structure and stratigraphy of the Yuen Long area, <strong>Hong</strong> <strong>Kong</strong>”, Karst geology in <strong>Hong</strong><br />

<strong>Kong</strong> Geological Society of <strong>Hong</strong> <strong>Kong</strong> Bulletin, edited by Langford R. D., Hansen, A., and Shaw, R., No.4, 49-60.<br />

Geotechnical Control Office (1989a). Solid and Superficial Geological Map of San Tin, 1:20,000 Map HGM20, Sheet 2.<br />

Geotechnical Control Office (1989b). Tin Shui Wai: Solid Geology, HGP 5B, Sheet 6-NW-A 1:5,000, Civil Engineering<br />

Services Department.<br />

Geotechnical Control Office (1990). Foundation Properties of Marble and Other Rocks in the Yuen Long–Tuen Mun<br />

Area, GCO Publication No.2/90, Civil Engineering Department, <strong>Hong</strong> <strong>Kong</strong>, 117p.<br />

Geotechnical Engineering Office (1994). Solid Geological Map of San Tin, Sheet 2, <strong>Hong</strong> <strong>Kong</strong> Geological Survey,<br />

1:20,000, Series HGM20S, Civil Engineering Department. <strong>Hong</strong> <strong>Kong</strong> Government.<br />

Geotechnical Engineering Office (2000). Geological Map of <strong>Hong</strong> <strong>Kong</strong>, 1:100,000 Series HGM20S, Compiled by Kirk,<br />

P.A. Sewell, R.J. Campbell, C.D.G. Fletcher, C.J.N. Lai, K.W. and Li, X.C., <strong>Hong</strong> <strong>Kong</strong> Geological Survey, Civil<br />

Engineering Department, <strong>Hong</strong> <strong>Kong</strong> SAR Government.<br />

Geotechnical Engineering Office (2007). Engineering Geological Practice in <strong>Hong</strong> <strong>Kong</strong>, GEO Publication No.1/2007,<br />

Civil Engineering and Development Department, <strong>Hong</strong> <strong>Kong</strong> SAR Government, 278p.<br />

Geotechnical Engineering Office (2008). <strong>Hong</strong> <strong>Kong</strong> Geology Guide Book, Geotechnical Engineering Office, Civil<br />

Engineering Department, <strong>Hong</strong> <strong>Kong</strong>.173p.<br />

Lai, K.W. Chan, H.H.K. Choy, C.S.M., and Tsang, A L.Y. (2004). “<strong>The</strong> characteristics of marble clast-bearing volcanic<br />

rock and its influence on foundation in <strong>Hong</strong> <strong>Kong</strong>.”, Proceedings of Conference on foundation Practice in <strong>Hong</strong><br />

<strong>Kong</strong>, E1-E10.<br />

Langford, R.L. Lai, K.W. Arthurton, R.S., and Shaw, R. (1989). Geology of the Western New Territories, <strong>Hong</strong> <strong>Kong</strong><br />

Geological Survey Memoir No.3, 140 Geotechnical Control Office, Civil Engineering Services Department,<br />

Government Printer, 140p.<br />

. Li, Q.W. (2010). “Discovery of palaeovolcanic plug group of the Middle Jurassic age in <strong>Hong</strong> <strong>Kong</strong> and its significance”,<br />

Proceedings on the Geosciences and Engineering of Guangdong, <strong>Hong</strong> <strong>Kong</strong> and Shenzhen, Shenzhen Geological Society,<br />

288p. (in Chinese)<br />

Sewell, R.J., Campbell, S.D.G. Fletcher, C.J.N. Lai, K.W., and Kirk, P.A. (2000). <strong>The</strong> Pre-Quaternary Geology of <strong>Hong</strong><br />

<strong>Kong</strong>, Geotechnical Engineering Office, Civil Engineering Department, 181p.<br />

-336-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

UPPER BOUND LIMIT ANALYSIS OF SOIL SLOPE STABILITY BASED ON RPIM<br />

MESHLESS METHOD<br />

F. T. Liu 1 , J. D. Zhao 1 , Y. F. Fan 2 , and J. H. Yin 3<br />

1<br />

Department of Civil and Environmental Engineering,<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> of Science and Technology, <strong>Hong</strong> <strong>Kong</strong>, China.<br />

2<br />

School of Natural Sciences and Humanities,<br />

Harbin Institute of Technology Shen Zhen Graduate School, Shen Zhen, China. Email: yhfan@hit.edu.cn<br />

3<br />

Department of Civil and Structural Engineering,<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong>, <strong>Hong</strong> <strong>Kong</strong>, China.<br />

ABSTRACT<br />

Limit analysis is widely used to evaluate the stability of structures in civil engineering. In comparison with<br />

elasto-plastic analysis, limit analysis can avoid the complicated computation of incremental analysis. A solution<br />

procedure based on radial point interpolation method for upper bound limit analysis of structures is presented.<br />

For evaluating the integrations of the external work rate and internal power dissipation rate, a new meshless<br />

integration technique based on Cartesian Transformation Method (CTM) was used to transform the domain<br />

integral into a boundary integral and a 1D integral. Finally, the nonlinear optimization problem derived from the<br />

upper bound limit analysis can be solved based on distinguishing rigid/plastic zones. And some examples of<br />

stability analysis show that this approach is a valid and simple technique.<br />

KEYWORDS<br />

Slope stability; Upper bound limit analysis; Radial point interpolation method; nonlinear programming;<br />

Cartesian Transformation Method (CTM)<br />

INTRODUCTION<br />

Limit analysis is a powerful method for stability analysis and limit bearing capacity of engineering structures. In<br />

geotechnical engineering, upper bound limit analysis is widely used to analyze the slope stability. Drucker (1952)<br />

firstly presented limit analysis based on plastic limit theorem, and then Chen (1975) introduced limit analysis<br />

into the geotechnical engineering for analyzing the bearing capacity, earth pressure on retaining wall and slope<br />

stability. It takes advantage of the lower and upper theorems of classical plasticity to bracket the true solution<br />

from a lower bound to an upper bound. However, it is difficult to obtain analytical solution for practical<br />

engineering, and numerical approaches are often required for limit analysis. In the past three decades, many<br />

studies have been devoted to developing numerical methods of limit analysis.<br />

Many researchers (Lysmer 1970; Anderheggen and Knopfel 1972; Bottero et al. 1980; Sloan 1988, 1989; Sloan<br />

and Kleeman 1995) constructed numerical limit analysis based on finite element method and linear<br />

programming theory, where the general yield criterion often was linearized to a convex polyhedron, and the<br />

nonlinear inequalities were approximated by a set of linear inequalities. Especially for the slope stability<br />

analysis, following the related work by Sloan and Kleeman (1995), some researchers (Yu et al. 1998; Kim et al.<br />

1999; Kim et al. 2002, etc) have applied the lower and upper bound approach to evaluate the slope stability. On<br />

the other hand, following the work of Zouain et al. (1993), Lyamin and Sloan (2002) proposed a nonlinear<br />

numerical method to perform upper and lower bound limit analysis based on linear finite elements and nonlinear<br />

programming. <strong>The</strong> results showed that their approach is vastly superior to a widely used linear programming<br />

formulation, especially for large scale applications. However, this approach has a potential difficulty in applying<br />

these formulations is that special stress or displacement finite elements need to be used. <strong>The</strong>refore, an<br />

alternative nonlinear technique which named the direct iterative algorithm is used to perform limit analysis of<br />

non-frictional materials (Zhang et al., 1991; Liu et al., 1995; Capsoni and Corradi, 1997). Following these ideas,<br />

Li and Yu (2006) extended the direct iterative algorithm to calculate plastic collapse loads of 2D and 3D<br />

structures obeying the ellipsoid yield criterion. In these approaches, upper bound limit analysis of structures is<br />

formulated as a nonlinear optimization problem with a single equality constraint, and a technique based on<br />

distinguishing rigid/plastic zones was adopted to solve this special nonlinear constrained optimization problem.<br />

-337-


Recently, as the development of finite element method, a so-called ‘meshless’ method has attracted more<br />

attention in the field of numerical method. Recently, Chen et al. (2008) and Le et al. (2009, 2010) constructed<br />

lower and upper formulation of limit analysis based element-free Galerkin (EFG) method which first proposed<br />

by Belytschko et al. (1994). Although the EFG method has been successfully applied to the lower and upper<br />

bound approaches, two issues are still not well studied: 1) difficulties in the enforcement of essential boundary<br />

conditions. This is because its shape function which calculated based on the moving least square method (MLS)<br />

is lack of Kronecher delta function property, i.e., where is the Kronecker delta function; 2) complexity in<br />

numerical algorithms for calculating shape function and its derivatives. For two issues, the one of approach is<br />

so-called radial point interpolation method (RPIM) proposed by Wang and Liu (2002). <strong>The</strong> RPIM shape<br />

functions have the Kronecker delta function property and partitions of unity. <strong>The</strong>refore, the essential boundary<br />

conditions could be easily enforced. Furthermore, the accuracy of RPIM is higher than that of the MLS (Liu and<br />

Gu 2005). On the other hand, regarding the integral strategy, some truly meshless methods commonly rely on<br />

the nodal integration technique. However, direct nodal integration is unstable because of under-integration and<br />

vanishing derivatives of shape functions at the nodes. To overcome this difficulty, Beissel and Belytschko (1996)<br />

added a residual of the equilibrium equation terms to the potential energy functional for stabilizing nodal<br />

integration. However, Beissel and Belytschko (1996) stated that the accuracy of this method is less than that of<br />

the original EFG method. Up-to-date stabilized conforming nodal integration technique is proposed by Chen et<br />

al. (2001, 2002), they modified the shape functions prior to nodal integration, even though this method is<br />

considered as a robust integration technique, it is based on the construction of a Voronoi diagram. So it cannot<br />

be considered a true meshless method. Recently, Khosravifard and Hematiyan (2010) proposed a new meshless<br />

integration technique based on Cartesian Transformation Method (CTM), in their method a domain integral is<br />

transformed into a boundary integral and a 1D integral.<br />

In this paper, we reformulated the upper bound limit analysis of structures using the nonlinear programming<br />

theory and the RPIM method, and the new integration technique proposed by Khosravifard and Hematiyan<br />

(2010) to calculate the internal dissipation power and external work rate. And the present method was used to<br />

calculate the limit loading parameter of a vertical slope. <strong>The</strong> layout of this paper is as follows: Section 2 briefly<br />

describes the upper bound limit analysis formulation for an ellipsoid yield function using RPIM method and<br />

CTM integration. A direct iterative algorithm based on Lagrange method is used to solve the nonlinear<br />

programming problem in Section 3. Numerical example for vertical slope is provided in Section 4 to illustrate<br />

the validity of the present method.<br />

NUMERICAL FORMULATION FOR UPPER BOUND APPROACH BASED ON RPIM MESHLESS<br />

METHOD<br />

<strong>The</strong> upper bound theorem<br />

<strong>The</strong> upper bound theorem of limit analysis states: among all kinematically admissible velocities (that is the<br />

plastic admissible strains), the real one yields the lowest rate of plastic dissipation power (Drucker and Prager<br />

1952)<br />

* T * T *<br />

D ε& dV ≥ TudΓ+<br />

λ f udV<br />

(1)<br />

∫ ( ) ∫ ∫<br />

V<br />

Γσ<br />

V<br />

where λ is the limit load multiplier, T is the basic load vector of surface tractions, f is the body force vector, u * is<br />

the kinematic admissible velocity vector, D( ε& * ) denotes the function for the rate of the plastic dissipation power<br />

*<br />

in terms of the admissible strain rate ε& , Γ σ denotes the traction boundary, and V denotes the space domain of<br />

the structure. Here, the kinematic admissible velocity vector u * must satisfy the following two conditions (Chen<br />

2002):<br />

1) Compatibility and velocity bound conditions<br />

1<br />

ε& * = ( ∇ u * + u<br />

* ∇)<br />

in V<br />

(2a)<br />

2<br />

*<br />

u = u on Γ u<br />

(2b)<br />

2) <strong>The</strong> yield criteria function<br />

f ( σ) = 0<br />

(3)<br />

<strong>The</strong> above two conditions can be related by the associated flow rule, i.e.,<br />

* f<br />

ε& = μ ∂<br />

(4)<br />

∂σ<br />

where Γ u denotes the displacement boundary; μ is the non-negative plastic multiplier. <strong>The</strong>refore, the solution<br />

of limit load multiplier based on upper bound theorem can be formulated as the following mathematical<br />

-338-


programming problems (the surface traction is omitted)<br />

*<br />

λ = min D ε&<br />

dV<br />

∫<br />

∫<br />

V<br />

( )<br />

T *<br />

st .. f u dV = 1<br />

V<br />

(5)<br />

* 1 * *<br />

ε&<br />

= ( ∇ u + u ∇)<br />

in V<br />

2<br />

*<br />

u = u<br />

on Γu<br />

From the optimum of limit loading multiplier λ opt , the limit loading can be computed according to the following<br />

equation:<br />

f = λ ⋅f (6)<br />

lim<br />

opt<br />

Nonlinear optimization problems for the plane strain von-Mises yield criterion<br />

In general, the slope stability problems are treated as the plain strain problems in geotechnical engineering. For<br />

the plain strain condition, the von-Mises (or Tresca) yield criterion can be written as (Pastor, 2000)<br />

f ( σ ) =<br />

1<br />

( ) 2 2<br />

σx − σ<br />

y<br />

+ τxy<br />

− c = 0<br />

4<br />

(7)<br />

where c is the cohesion. According to the associated flow rule, the power of dissipation can be formulated as a<br />

function of strain rates as (Capsoni and Corradi, 1997)<br />

D ( ε& ) =<br />

T<br />

ε& Θε& (8)<br />

where<br />

2 2<br />

⎡ c −c<br />

0 ⎤<br />

⎢ 2 2 ⎥<br />

Θ = ⎢ −c<br />

c 0 ⎥<br />

(9)<br />

⎢<br />

2<br />

0 0 c ⎥<br />

⎣<br />

⎦<br />

<strong>The</strong>refore, the mathematical programming problem (5) of finding the upper bound solution of limit loading<br />

multiplier can be formulated as the following nonlinear optimization problem:<br />

T<br />

λ = min εΘε & & dV<br />

V<br />

T *<br />

.. ∫ f u dV = 1<br />

V<br />

st<br />

∫<br />

* 1 * *<br />

ε&<br />

= ( ∇ u + u ∇)<br />

2<br />

in V<br />

*<br />

u = u<br />

on Γ<br />

Radial point interpolation method<br />

u<br />

<strong>The</strong> approximation of the field variables of interest x using radial point interpolation method (RPIM) can be<br />

expressed in the following form (Liu and Gu, 2005):<br />

T<br />

T<br />

−1<br />

⎧Us<br />

⎫<br />

u( x)<br />

= ⎡<br />

⎣Rq ( x) Pm ( x) ⎤<br />

⎦G ⎨ ⎬=<br />

Φ( x)<br />

Us<br />

(7)<br />

⎩ 0 ⎭<br />

where u(x) is the function of field variables, U s ={u 1 , u 2 , …, u n } T is the vector of function values, Ф(x) is the<br />

RPIM shape functions corresponding to the nodal value and given by<br />

Φ ( x T<br />

T<br />

−1<br />

) = ⎣<br />

⎡ R ( x ) P q m ( x ) ⎤<br />

⎦ G = ⎡⎣φ1( x ) φ2( x ) L φn( x ) ⎤⎦<br />

(8)<br />

in which, R q is the moment matrix of the radial basis function (RBF) given by<br />

⎡R1( x1) R2( x1) L Rn<br />

( x1)<br />

⎤<br />

⎢<br />

⎥<br />

⎢<br />

R1( x2) R2( x2) L Rn<br />

( x2)<br />

R<br />

q<br />

= ⎥<br />

(9)<br />

⎢ M M O M ⎥<br />

⎢<br />

⎥<br />

⎢⎣R1( xn) R2( xn) L Rn( xn)<br />

⎥⎦n×<br />

n<br />

and P m the polynomial moment matrix is defined as follows<br />

(10)<br />

-339-


( x ) ( x ) L<br />

m ( x )<br />

( x ) ( x ) L ( x )<br />

⎡p p p<br />

⎢<br />

p p p<br />

P ⎢<br />

M M O M<br />

⎢<br />

⎢⎣p p p<br />

and the matrix G is defined as<br />

⎡Rq<br />

Pm⎤<br />

G = ⎢ T ⎥<br />

⎣Pm<br />

0 ⎦<br />

1 1 2 1 1<br />

1 2 2 2 m 2<br />

m<br />

= ⎢ ⎥<br />

( x ) ( x ) L ( x )<br />

⎤<br />

⎥<br />

⎥<br />

⎥<br />

⎥⎦<br />

1 n 2 n m n n×<br />

m<br />

(10)<br />

(11)<br />

<strong>The</strong>refore, the k-th element of shape function can be expressed as follows,<br />

n<br />

m<br />

φ<br />

k i j<br />

+<br />

i= 1 j=<br />

1<br />

( ) = ∑R ( ) G( , )<br />

+<br />

ik ∑ p ( ) G( n jk , )<br />

x x x (12)<br />

where G (i,k) is the element of matrix G -1 . A classical RBF is multiquadric basis (MQ), which has the following<br />

form (Gu and Liu, 2005):<br />

2 2 2<br />

q<br />

R ( x i ) = ⎡( x− xi) + ( y− yi) + ( αcdc)<br />

⎤<br />

⎣<br />

⎦<br />

(13)<br />

where α c and q are two shape parameters, d c is the character length that relates to the nodal spacing in the local<br />

support domain. In addition, the complete polynomial basis of order p for two-dimensional domains can be<br />

written in the following form (Gu and Liu, 2005):<br />

2 2<br />

P T x = 1 x y x xy y L x p y<br />

p<br />

(14)<br />

( ) { }<br />

If the function u(x) stands for the displacement field for two-dimensional domains, it can be interpolated from<br />

the vectors of nodal function value and RPIM shape function corresponding to the nodal value, i.e.,<br />

n<br />

0<br />

n<br />

ux ⎡φI<br />

⎤⎡uI⎤<br />

( ) = ∑⎢ I<br />

( )<br />

I<br />

I= 1 0 φ<br />

⎥⎢ =<br />

I<br />

v<br />

⎥ ∑ Φ xu (15)<br />

⎣ ⎦⎣ I ⎦ I=<br />

1<br />

where Ф I (x) is the matrix of shape function of node I, and u I is the nodal displacements. And, the derivatives of<br />

the RPIM shape functions can be formulated as follows (Wang and Liu 2002):<br />

n<br />

m<br />

∂φk<br />

∂R ∂p<br />

i<br />

j<br />

= ∑ G<br />

( ik , )<br />

+ ∑ G<br />

( n+<br />

jk , )<br />

(17a)<br />

∂x ∂x ∂x<br />

i= 1 j=<br />

1<br />

∂φ<br />

p<br />

k<br />

∂R ∂<br />

= G +<br />

∂y ∂y ∂y<br />

n<br />

m<br />

i<br />

j<br />

( ik , ) G<br />

( n+<br />

jk , )<br />

i= 1 j=<br />

1<br />

∑ ∑ (17b)<br />

According to the approximation of displacement field function, the plastic admissible strains can be expressed<br />

as<br />

⎡ ∂ ⎤ ⎡<br />

1<br />

0<br />

u ∂φ<br />

∂φ<br />

⎤<br />

n<br />

⎢ ⎥<br />

⎡ 1⎤ ⎢ 0 L 0 ⎥⎡u1⎤<br />

⎢∂x<br />

⎥<br />

⎢<br />

v<br />

⎥ ⎢ ∂x<br />

∂x<br />

⎥⎢ 1<br />

v<br />

⎥<br />

1<br />

⎢ ∂ ⎥⎡φ1 ( x) 0 L φn<br />

( x) 0 ⎤⎢ ⎥ ⎢ ∂φ1<br />

∂φ<br />

⎥⎢ ⎥<br />

n<br />

ε = Lu( x) = ⎢ 0 ⎥⎢ 0 0<br />

y 0 φ1<br />

( ) 0 φn<br />

( )<br />

⎥⎢ M ⎥ = ⎢ L ⎥⎢ M ⎥ = Bu (18)<br />

⎢ ∂ ⎥⎣ x L<br />

x ⎦⎢ ⎥ y y<br />

u ⎢ ∂ ∂ ⎥⎢ ⎥<br />

n<br />

un<br />

⎢ ∂ ∂ ⎥<br />

⎢ ⎥ ⎢∂φ1 ∂φ1<br />

∂φn<br />

∂φ<br />

⎥⎢ ⎥<br />

⎢<br />

n<br />

⎢ ⎥ v ⎥ ⎢<br />

n ⎢<br />

⎥ v ⎥<br />

n<br />

∂y ∂x ⎣ ⎦ L<br />

∂y ∂x ∂y ∂x<br />

⎣ ⎦<br />

⎣ ⎦ ⎣ ⎦<br />

where B is the strain matrix. <strong>The</strong>refore, substituting Eq. (18) into nonlinear programming problem, the<br />

discretized formulation of upper bound approach based on RPIM meshless method can be expressed as follows:<br />

T T<br />

λ = min uBΘBu dV<br />

∫<br />

V<br />

T<br />

st .. f Φu dV = 1<br />

V<br />

∫<br />

I<br />

I<br />

u = u<br />

on Γ<br />

A new integration method based on CTM<br />

I<br />

u<br />

In the numerical formulation of nonlinear programming problem (19) based on RPIM, the main task is to<br />

calculate the integration in the objective function and constrained equations. Recently, Khosravifard and<br />

Hematiyan (2010) proposed a new meshless integration technique based on Cartesian Transformation Method<br />

(19)<br />

-340-


(CTM), in their method a domain integral is transformed into a boundary integral and a 1D integral. According<br />

to the CTM integral technique, the integral can be calculated in terms of the following formulation, for 2D<br />

problems:<br />

n G<br />

2D<br />

2D<br />

I = W ⋅ F =∑W ( xi) f ( x<br />

i)<br />

(20)<br />

where<br />

i=<br />

1<br />

( x )<br />

W = J ⋅J ⋅w ⋅w<br />

2D x y x y<br />

i i i i i<br />

and x i is the Gaussian points, i=1,…,n G , where n G is the number of the Gaussian points. Furthermore, by<br />

introducing the transformation matrix C e , the nodal velocity vector u I for each node can be expressed by the<br />

global nodal velocity vector U for the slope, i.e<br />

u = C U (21)<br />

I<br />

e<br />

<strong>The</strong>refore, the objective function of nonlinear programming problem (19) can be reformulated as follows:<br />

I<br />

where<br />

n G<br />

2D<br />

T<br />

= ∑W<br />

( xi)<br />

U KiU (22)<br />

i=<br />

1<br />

K = C D ΘDC<br />

T T<br />

i e e<br />

On the other hand, the integral of external work rate can be calculated by using 1-D Gaussian quadrature method,<br />

and the nodal traction force vector F can be expressed as follows<br />

nct<br />

nGt<br />

T<br />

T<br />

F = ∑∑Jilw( xi ) ⋅f( xi ) Φ ⋅C e<br />

(23)<br />

l i=<br />

1<br />

<strong>The</strong>refore, the RPIM formulation of nonlinear programming (19) can be finally expressed as follows:<br />

n G<br />

2D<br />

T<br />

λ = min ∑W<br />

( xi)<br />

⋅ U KiU<br />

i=<br />

1<br />

st ..<br />

T<br />

F ⋅ U=1<br />

KU<br />

v<br />

= 0<br />

u = 0 on Γu<br />

(24)<br />

where K v U=0 is the plastic incompressibility should be satisfied for the materials with von Mises’ or Hill’s yield<br />

criterion, the matrix K v can be expressed as follows:<br />

Kv = DvC e; Dv = [ Dv 1<br />

Dv2<br />

L Dvn]<br />

; vi<br />

= ⎡<br />

⎣φi , x<br />

φ ⎤<br />

i, y ⎦; i = 1, , n<br />

In addition, it should be pointed out the velocity boundary conditions can be imposed by means of the<br />

conventional finite element technique due to the use of radial point interpolation shape function in this study.<br />

THE DIRECT ITERATIVE METHOD<br />

For the nonlinear programming problem (24), there is a calculation of square root which could make the<br />

objective function unsmooth and nondifferentiable. This causes some difficulties in solving the nonlinear<br />

programming problem. Following the work of Li and Yu (2006), it can be overcome using an iterative algorithm<br />

for distinguishing rigid/plastic zones. At first, the NLP (24) are transformed into an unconstrained optimization<br />

problem using Lagrange method. <strong>The</strong> Lagrange function is the following form:<br />

n<br />

G<br />

( ) 2 G<br />

D T<br />

( ) 2 D T<br />

T<br />

, μ = ∑<br />

i<br />

⋅<br />

i<br />

+ α∑<br />

( i) ⋅ ( v ) v<br />

+ 2μ( 1− ⋅ )<br />

i= 1 i=<br />

1<br />

n<br />

L U W x UKU W x KU KU F U (25)<br />

where μ is the Lagrange multiplier. Following the work of Li and Yu (2005), an iterative control parameter ω ICP<br />

was defined as follows:<br />

ICP T<br />

ω = UKU<br />

i<br />

(26)<br />

And then, the Lagrange function can be reformulated as:<br />

( ) ( ) ( ) n<br />

T<br />

G UKU<br />

n<br />

2 G<br />

D i<br />

, 2 D T<br />

T<br />

L U μ = ∑W xi ⋅ + α W ( i) ⋅ ( v ) v<br />

+ 2μ( 1− ⋅<br />

ICP ∑ x K U K U F U )<br />

(27)<br />

ω<br />

i= 1 i=<br />

1<br />

For finding all rigid regions, the following iterative process is needed.<br />

-341-


Step1: initializing the nonlinear objective function<br />

Let iterative control parameter ω ICP =1, then, the initial nodal displacement velocity can be estimated by solving<br />

the following equation system:<br />

n<br />

⎧ G<br />

2D<br />

T<br />

⎪ ∑W ( xi) ⋅ ( Ki + αKvKv)<br />

U0 = μ0F<br />

⎨ i=<br />

1<br />

(28)<br />

⎪ T<br />

⎩f U0 = 1<br />

and the initial load multiplier can be calculated by using:<br />

n G<br />

2D<br />

T<br />

∑W<br />

( xi)<br />

U KiU (29)<br />

λ = ⋅<br />

0 0 0<br />

i=<br />

1<br />

Step k+1 (k=0, 1, 2 …): distinguishing the nondifferentiable areas to revise the objective function<br />

Based on the results at step k, the value of ω ICP need to be calculated at very Gaussian integral point of CTM,<br />

k 1<br />

then the Gaussian integral point set S will be subdivided into two subsets: the subset S +<br />

r<br />

where the object<br />

k 1<br />

function is not differentiable and the subset S +<br />

P<br />

where the object function is differentiable, i.e.,<br />

k + 1 ICP k + 1<br />

ICP<br />

S = i∈ S, ω = 0 ; S = i∈S, ω ≠ 0<br />

(30)<br />

r<br />

{ } p { }<br />

For ω ICP =0, the original optimization problem can be solved in terms of the following problem:<br />

nG<br />

T<br />

2D Uk+ 1KiUk+<br />

1<br />

min ∑ W ( xi<br />

) ⋅<br />

U k+<br />

1<br />

T<br />

i∈S<br />

p UKU<br />

T<br />

st .. f U = 1<br />

k + 1<br />

( KU ) = 0 ( i∈S)<br />

U<br />

v k+<br />

1 i<br />

K U<br />

= 0 ∈<br />

k i k<br />

k+<br />

1<br />

( i S )<br />

T<br />

k+ 1 i k+<br />

1<br />

r<br />

<strong>The</strong> revised NLP problem can be solved in terms of the following equation system:<br />

⎧ 2D KU<br />

i k+<br />

1<br />

2D 2D T<br />

⎪ ∑ W ( xi) ⋅ + ξ ∑ W ( xi) ⋅ KiUk+ 1+ α ∑ W ( xi)<br />

⋅ KvKvUk+ 1<br />

= μk+<br />

1F<br />

k+ 1 T<br />

k+ 1 k+<br />

1<br />

⎨i∈Sp UKU<br />

k i k<br />

i∈Sr i∈Sr<br />

⎪ T<br />

⎩FUk<br />

+ 1<br />

= 1<br />

(31)<br />

(32)<br />

By solving the Eq. (32), we can obtain the nodal velocity and limit load multiplier at this step<br />

n G<br />

λ<br />

k + 1<br />

=<br />

i<br />

⋅<br />

k + 1 i k + 1<br />

i=<br />

1<br />

2D<br />

T<br />

∑W<br />

( x ) U K U (33)<br />

<strong>The</strong> above iterative process is repeated until the following convergence criteria are satisfied<br />

λk+ 1<br />

−λk Uk+<br />

1−Uk<br />

≤η1;<br />

≤η2<br />

λk+ 1<br />

Uk+<br />

1<br />

where η 1 and η 2 are the computational error tolerances.<br />

(34)<br />

UPPER BOUND FOR THE HEIGHT LIMIT OF A VERTICAL SLOPE<br />

<strong>The</strong> height limit of a vertical slope is a classical problem of limit analysis or yield design theory. <strong>The</strong> vertical<br />

slope (See Figure 1) is subjected only to own weight. <strong>The</strong> soil is homogeneous and isotropic, and its cohesion is<br />

c, unit weight is γ. According to the research by Pastor et al. (2000), the new bounds of limit loading parameter<br />

is as follows:<br />

γ H<br />

3.7603 ≤ Qγ<br />

= ≤ 3.7859<br />

(35)<br />

c<br />

-342-


H<br />

γ<br />

Soil 1 #<br />

Soil c (kN/m 2 ) γ (kN/m 3 )<br />

1 # 30 18<br />

Figure 1 <strong>The</strong> model of critical height of a vertical slope<br />

(a)<br />

(b)<br />

Figure 2 <strong>The</strong> layout of the field nodes: (a) regular; (b) irregular<br />

For this test problem, two types of layout of field nodes as shown in Figure 2 can be used to discretize the<br />

domain. And then, the RPIM shape function can be constructed based on the discretization of field nodes. For a<br />

reliable RPIM shape function construction, a T2L-Scheme proposed by Liu (2010) is used to select local<br />

supporting nodes. On the other hand, the interpolation accuracy of RPIM can also be affected by the<br />

dimensionless shape parameters α c , q and numbers of field nodes. <strong>The</strong>refore, these parameters should be<br />

analysed one by one.<br />

Firstly, the shape parameters α c =4 and q=0.5 are fixed for analysing the effect of nodal layout on the limit<br />

loading parameter Q γ . In addition, the optimal parameters for the direct iterative algorithm can be chose<br />

according to the research of Li and Yu (2006). And the computational error tolerances η 1 =η 2 =0.001 are fixed.<br />

Limit Loading Multiplier<br />

3<br />

2.8<br />

2.6<br />

2.4<br />

2.2<br />

2<br />

95 nodes, 576 integral points<br />

95 nodes, 1024 integral points<br />

141 nodes, 576 integral points<br />

141 nodes, 1024 integral points<br />

186 nodes, 576 integral points<br />

186 nodes, 1024 integral points<br />

224 nodes, 576 integral points<br />

224 nodes, 1024 integral points<br />

264 nodes, 576 integral points<br />

264 nodes, 1024 integral points<br />

303 nodes, 576 integral points<br />

303 nodes, 1024 integral points<br />

376 nodes, 576 integral points<br />

376 nodes, 1024 integral points<br />

389 nodes, 576 integral points<br />

389 nodes, 1024 integral points<br />

518 nodes, 1024 integral points<br />

577 nodes, 1024 integral points<br />

701 nodes, 1024 integral points<br />

1.8<br />

1.6<br />

0 10 20 30 40 50 60 70<br />

Iteration Step<br />

Figure 3 <strong>The</strong> convergence sequence of limit loading multiplier with iterative steps for irregular nodal layout<br />

-343-


Table 1 <strong>The</strong> results of limit load multiplier for irregular nodal layout (576 integral points)<br />

Nodes 95 141 186 224 264 303 376 389<br />

λ 2.496 2.242 2.163 2.102 1.954 1.921 1.814 1.784<br />

Q γ 9.985 8.968 8.650 8.407 7.825 7.682 7.257 7.137<br />

errors 58% 42% 37% 33% 23% 21% 15% 13%<br />

Runtime (s) -- -- -- 29 48 92 195 171<br />

Table 2 <strong>The</strong> results of limit load multiplier for irregular nodal layout (1024 integral points)<br />

Nodes 95 141 186 224 264 303 376 389 518 577 701<br />

λ 2.509 2.293 2.219 2.157 2.086 2.084 2.007 2.025 1.847 1.837 1.662<br />

Q γ 10.037 9.170 8.877 8.627 8.345 8.335 8.028 8.098 7.387 7.348 6.650<br />

errors 59% 45% 41% 37% 32% 32% 27% 28% 17% 16% 0.5%<br />

Runtime (s) -- -- -- 42 73 112 171 171 701 902 1360<br />

With the above parameters, the optimal value of limit loading multiplier can be found using the direct iterative<br />

algorithm based on regular and irregular nodal layout. <strong>The</strong> convergence of limit loading multiplier with iterative<br />

steps for irregular nodal layout is shown in Figure 3. <strong>The</strong> optimal value of limit multiplier and the corresponding<br />

limit loading parameter for 576 and 1024 integral points are shown in Table 1 and Table2 respectively. As the<br />

same numbers of integral point, the accuracy of limit loading parameter will increase with the increasing<br />

numbers of field node (see Figure 5a). For regular nodal layout, the convergence sequence of limit multiplier<br />

with iterative steps is shown in Figure 4. And the optimal value of limit multiplier and the corresponding limit<br />

loading parameter for different integral points are shown in Table 3. In this case, the number of integral points<br />

dependents on that of field nodes, hence, the accuracy of limit loading parameter is analysed just for the<br />

different numbers of field node. From the Figure 5b and Table3, the accuracy of limit loading parameter will<br />

also increase with the increasing numbers of field node.<br />

Limit loading Multiplier<br />

3.4<br />

3.2<br />

3<br />

2.8<br />

2.6<br />

2.4<br />

2.2<br />

2<br />

83 field nodes<br />

293 field nodes<br />

446 field nodes<br />

631 field nodes<br />

1.8<br />

1.6<br />

1.4<br />

0 5 10 15 20 25 30 35 40 45<br />

Iteration Step<br />

Figure 4 <strong>The</strong> convergence sequence of limit loading multiplier with iterative steps for regular nodal layout<br />

Limit Loading Multiplier<br />

2.6<br />

2.5<br />

2.4<br />

2.3<br />

2.2<br />

2.1<br />

2<br />

1.9<br />

1.8<br />

1.7<br />

576 Integral points<br />

1024 Integral points<br />

Limit Loading Multiplier<br />

3.2<br />

3<br />

2.8<br />

2.6<br />

2.4<br />

2.2<br />

2<br />

1.8<br />

1.6<br />

1.6<br />

0 100 200 300 400 500 600 700 800<br />

Numbers of Field Node<br />

1.4<br />

0 100 200 300 400 500 600 700<br />

Numbers of Field Node<br />

(a)<br />

(b)<br />

Figure 5 <strong>The</strong> limit loading multiplier with different numbers of field nodes: (a) irregular nodal layout, and (b)<br />

regular nodal layout<br />

-344-


Table 3 <strong>The</strong> results of limit load multiplier for irregular nodal layout<br />

Nodes 83 293 446 631<br />

λ 3.108 2.203 2.215 1.575<br />

Q γ 12.432 8.813 8.500 6.287<br />

errors 97% 40% 35% 0.04%<br />

Runtime (s) 0.1 35 194 1169<br />

By comparing the results of limit load parameter listed in Tables1, 2 and 3 (see Figure 5), it is very apparent that<br />

the accuracy of limit load parameter for regular nodal layout is higher than that of irregular layout. However, the<br />

reason of difference between two nodal layouts is not analysed here, and it will be further studied in the<br />

following research works.<br />

CONCLUSIONS AND DISCUSSIONS<br />

In this paper, a new formulation of upper bound approach based on RPIM and nonlinear programming is<br />

proposed. In the present method, the CTM integration method is used to calculate the internal dissipation, and<br />

the direct iterative algorithm is used to solve nonlinear programming for finding the optimal value of limit<br />

loading parameter of vertical slope. By the classical vertical slope stability problem, the validity of the present<br />

method is verified in this paper. <strong>The</strong> accuracy of limit loading parameter mainly depends on the number of field<br />

nodes and integral points. On the other hand, the accuracy of limit loading parameter for regular nodal layout is<br />

higher than that of irregular layout. <strong>The</strong> reasons of different accuracy need to be further studied. It may be<br />

carried out from the following two aspects, i.e., the CTM integration method and interpolation of RPIM shape<br />

function.<br />

ACKNOWLEDGMENTS<br />

<strong>The</strong> first author appreciates the useful comments from Prof H.S. Yu of <strong>University</strong> of Nottingham. <strong>The</strong> work was<br />

partly supported by Research Grants Council of <strong>Hong</strong> <strong>Kong</strong> (under grant No. 623609).<br />

REFERENCES<br />

Anderheggen, E., Knöpfel, H. (1972). “Finite element limit analysis using linear programming”, International<br />

Journal of Solids and Structures, 8, 1413-1431.<br />

Beissel, S., Belytschko T. (1996). “Nodal integration of the element-free Galerkin method”, Computer Methods<br />

in Applied Mechanics and Engineering, 139, 49-74.<br />

Belytschko, T., Lu, Y.Y., Gu, L. (1994). “Element-free Galerkin methods”, International Journal for Numerical<br />

Methods in Engineering, 37, 229-256.<br />

Bottero, A., Negre, R., Pastor, J., Turgeman, S. (1980). “Finite Element Method and Limit Analysis <strong>The</strong>ory for<br />

Soil Mechanics Problems”, Computer Methods in Applied Mechanics and Engineering, 22, 131-149.<br />

Capsoni, A., Corradi, L. (1997). “ A finite element formulation of the rigid plastic limit analysis problem”,<br />

International Journal for Numerical Methods in Engineering, 40, 2063-2086.<br />

Chen, J.S., Wu, C.T., Yoon, S., You, Y. (2001). “A stabilized conforming nodal integration for Galerkin<br />

mesh-free methods”, International Journal for Numerical Methods in Engineering, 50, 435-466.<br />

Chen, J.S, Yoon, S., Wu, C.T. (2002). “Non-linear version of stabilized conforming nodal integration for<br />

Galerkin mesh-free methods”, International Journal for Numerical Methods in Engineering, 53,<br />

2587-2615.<br />

Chen, W.F. (1975). “Limit Analysis and Soil Plasticity”, New York: Elsevier Scientific Publishing Co.<br />

Chen, S., Liu, Y., Cen, Z. (2008). “Lower-bound limit analysis by using the EFG method and non-linear<br />

programming”, International Journal for Numerical Methods in Engineering, 74, 391-415.<br />

Chen, Z. (2002). “Limit analysis of the classic problems of soil mechanics”, Chinese Journal of Geotechnical<br />

Engineering, 24(1), 1-11. (in Chinese)<br />

Ciria, H., Peraire, J., Bonet, J. (2008). “Mesh adaptive computation of upper and lower bounds in limit analysis”,<br />

International Journal for Numerical Methods in Engineering, 75, 899-944.<br />

Drucker, D.C., Prager, W. (1952). “Soil mechanics and plastic analysis or limit design”, Quarterly of Applied<br />

Mathematics, 10(2), 157-165<br />

Kim, J., Salgado, R., Yu, H.S. (1999). “Limit analysis of soil slopes subjected to pore-water pressures”, Journal<br />

of Geotechnical and Geoenvironmental Engineering, ASCE, 125(1), 49-58.<br />

Kim, J., Salgado, R., Lee, J. (2002). “Stability analysis of complex soil slopes using limit analysis”, Journal of<br />

Geotechnical and Geoenvironmental Engineering, ASCE, 128(7), 546-557.<br />

-345-


Khosravifard, A., Hematiyan, M.R. (2010). “A new method for meshless integration in 2D and 3D Galerkin<br />

meshfree methods”, Engineering Analysis with Boundary Elements, 34, 30-40.<br />

Krabbenhøft, K., Lyamin, A.V., Sloan, S.W. (2007). “Formulation and solution of some plasticity problems as<br />

conic programs”, International Journal of Solids and Structures, 44, 1533-1549.<br />

Krabbenhøft, K., Lyamin, A.V., Sloan, S.W. (2008). “Three-dimensional Mohr-Coulomb limit analysis using<br />

semidefinite programming”, Communications in Numerical Methods in Engineering, 24, 1107-1119.<br />

Le, C.V., Gilbert, M., Askes, H. (2009). “Limit analysis of plates using the EFG method and second-order cone<br />

programming”, International Journal for Numerical Methods in Engineering, 78, 1532-1552.<br />

Le, C.V., Nguye-Xuan, H., Nguyen-Dang, H. (2010a). “Upper and lower bound limit analysis of plates using<br />

FEM and second-order cone programming”, Computer and Structures, 88, 65-73.<br />

Le, C.V., Nguye-Xuan, H., Askes, H., Bordas, S. P. A., Rabczuk, T., Nguyen-Vinh, H. (2010b). “A cell-based<br />

smoothed finite element method for kinematic limit analysis”, International Journal for Numerical Methods<br />

in Engineering, 83, 1651-1674.<br />

Li, H.X., Yu, H.S. (2006). “Limit analysis of 2-D and 3-D structures based on an ellipsoid yield criterion”, Acta<br />

Geotechnica, 1, 179-193.<br />

Liu, G.R. (2010). “Meshfree method: moving beyond the finite element method, 2 nd ed.”, CRC press: Boca<br />

Raton, USA.<br />

Liu, G.R., Gu, Y.T. (2005). “An introduction to meshfree methods and their programming”, Spring.<br />

Liu, Y.H., Cen, Z.Z., Xu, B.Y. (1995). “A numerical method for plastic limit analysis of 3-D structures”,<br />

International Journal of Solids and Structures, 32(12), 1645-1658.<br />

Lyamin, A.V., Sloan, S.W. (2002). “Upper bound limit analysis using linear finite elements and non-linear<br />

programming”, International Journal for Numerical and Analytical Methods in Geomechanics, 26,<br />

181-216.<br />

Lysmer, J. (1970). “Limit analysis of plane problems in soil mechanics”, Journal of the Soil Mechanics and<br />

Foundations Division, ASCE, 96(4), 1311-1334.<br />

Makrodimopoulos A., Martin, C.M. (2006). “Lower bound limit analysis of cohesive-frictional materials using<br />

second-order cone programming”, International Journal for Numerical Methods in Engineering, 66,<br />

604-634.<br />

Makrodimopoulos A., Martin, C.M. (2007). “Upper bound limit analysis using simplex strain elements and<br />

second-order cone programming”, International Journal for Numerical and Analytical Methods in<br />

Geomechanics, 31, 835-865.<br />

Makrodimopoulos A., Martin, C.M. (2008a). “Upper bound limit analysis using discontinuous quadratic<br />

displacement fields”, Communications in Numerical Methods in Engineering, 24, 911-927.<br />

Martin, C.M., Makrodimopoulos A. (2008b). “Finite-Element Limit Analysis of Mohr-Coulomb Materials in 3D<br />

Using Semidefinite Programming”, Journal of Engineering Mechanics, ASCE, 134(4), 339-347.<br />

Makrodimopoulos A. (2010). “Remarks on some properties of conic yield restrictions in limit analysis”,<br />

International Journal for Numerical Methods in Biomedical Engineering, 2010, 26, 1449-1461.<br />

Pastor, J., Thai, T.H., Francescato, P. (2000). “New bounds for the height limit of a vertical slope”, International<br />

Journal for Numerical and Analytical Methods in Geomechanics, 24, 165-182.<br />

Sloan, S.W. (1988). “Lower bound limit analysis using finite elements and linear programming”, International<br />

Journal for Numerical and Analytical Methods in Geomechanics, 12, 61-77.<br />

Sloan, S.W. (1989). “Upper bound limit analysis using finite elements and linear programming”, International<br />

Journal for Numerical and Analytical Methods in Geomechanics, 13, 263-282.<br />

Sloan, S.W., Kleeman, P.W. (1995). “Upper bound limit analysis using discontinuous velocity fields”, Computer<br />

Methods in Applied Mechanics and Engineering, 127, 293-314.<br />

Wang, J.G., Liu, G.R. (2002). “A point interpolation meshless method based on radial basis functions”,<br />

International Journal for Numerical Methods in Engineering, 54 (11), 1623-1648.<br />

Yu, H.S., Salgado, R., Sloan, S.W., Kim, J.M. (1998). “Limit analysis versus limit equilibrium for slope<br />

stability”, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 124(1), 1-11.<br />

Zhang, P., Lu, M., Hwang, K. (1991). “A mathematical programming algorithm for limit analysis”, ACTA<br />

MECHANICA SINICA, 23(4), 433-442. (in Chinese)<br />

Zouain, N., Herskovits, J., Borges, L.A., Feijóo, R.A. (1993). “An iterative algorithm for limit analysis with<br />

nonlinear yield functions”, International Journal of Solids and Structures, 30, 1397-1417.<br />

-346-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

5.12 汶 川 地 震 震 损 水 坝 特 征 及 震 害 影 响 因 素 分 析<br />

梁 海 安 , 景 立 平 , 李 永 强 , 刘 春 辉<br />

( 中 国 地 震 局 工 程 力 学 研 究 所 , 黑 龙 江 哈 尔 滨 150080)<br />

摘 要 : 中 国 四 川 省 1998 年 5.12 汶 川 地 震 造 成 了 水 坝 的 严 重 震 害 。 本 文 通 过 对 汶 川 严 重 震 害 水<br />

坝 的 原 始 数 据 系 统 调 查 和 分 析 , 对 汶 川 地 震 中 水 坝 震 害 类 型 、 分 布 特 点 以 及 它 们 的 影 响 因 素 等 进 行<br />

了 定 量 和 定 性 的 分 析 , 认 为 汶 川 地 震 中 , 震 损 水 坝 数 量 多 、 分 布 广 , 震 害 种 类 多 , 以 地 震 波 造 成 的<br />

震 害 为 主 , 震 害 形 式 以 坝 体 裂 缝 为 主 , 滑 坡 、 震 陷 、 渗 漏 等 震 害 普 遍 。 指 出 汶 川 地 震 中 , 地 震 烈 度 、<br />

施 工 质 量 、 水 坝 坝 基 、 筑 坝 材 料 、 水 库 管 理 水 平 等 因 素 对 汶 川 地 震 中 水 坝 的 震 损 情 况 有 重 要 的 影 响 ,<br />

并 在 此 基 础 上 提 出 水 坝 抗 震 和 震 害 评 估 的 一 些 建 议 。<br />

关 键 字 : 5.12 汶 川 Ms8.0 地 震 水 坝 震 害 震 损 特 征 震 害 因 素<br />

CHARATERISTICS OF DAMAGED DAMS AND INFLUENCING FACTORS IN THE<br />

WENCHUAN 5.12 EARTHQUAKE<br />

H.A Liang,L.P Jing,Y.Q Li,C.H Liu<br />

Institute of Engineering Mechanics, China Earthquake Administration, Harbin 150080.china<br />

Abstract: <strong>The</strong>re were serious damaged dams caused by 5.12 Wenchuan Ms 8.0 earthquake in Sichuan province.<br />

A lot of raw data of seismic damadeg dams, especically severe damged cases were be analyzied. Quantitative<br />

and qualitative analysis is presented for the typical damaged types, distribution and their impact factors of dams<br />

based on large sets of original data. <strong>The</strong> number of seismic damaged dams is huge and the severe damaged dam<br />

almost are small dams in Wenchuan earthquake.<strong>The</strong> types of damaged dams are simultaneously, dams damaged<br />

by seismic wave are more and serious.Intensity, quality of construction, dam base, dam material, reservoir<br />

management and other factors were believed have significant impact on dams seismic safety. Finally, some<br />

recommendation were proposed to reduce the damage of dams in earthquake according to the analysis.<br />

Keywords: 5.12 Ms 8.0 Wenchuan earthquake, Seimic damage of Dam, Characteristic of damaged dams,<br />

Influence factors.<br />

一 、 前 言<br />

水 坝 震 害 是 大 震 后 容 易 发 生 的 震 害 形 式 之 一 , 威 胁 到 人 们 群 众 的 生 命 财 产 。 水 坝 在 地 震 中 受 损<br />

后 , 不 仅 是 承 灾 体 , 更 是 危 险 的 次 生 灾 害 源 , 一 旦 溃 决 将 造 成 严 重 的 后 果 。 汶 川 2008 年 5.12Ms8.0<br />

级 地 震 中 , 大 量 水 坝 受 损 , 国 内 有 八 个 省 水 坝 不 同 程 度 受 到 破 坏 , 各 类 受 损 水 坝 总 数 达 2666 座 [1-2] 。<br />

其 中 以 四 川 省 水 库 震 害 最 为 严 重 , 根 据 《 中 国 水 利 年 鉴 》 统 计 , 至 2008 年 初 四 川 省 有 水 坝 6717 座 ,<br />

其 中 大 型 水 库 8 座 、 中 型 水 库 105 座 , 小 型 水 库 6604 座 。 根 据 四 川 农 业 水 利 局 的 统 计 通 报 : 截 止<br />

2008.6.12 日 汶 川 地 震 四 川 省 共 有 1996 座 水 库 震 损 , 占 全 省 水 库 总 数 的 29.71%。 震 损 水 库 按 受 损 程<br />

度 分 为 溃 坝 险 情 、 高 危 险 情 和 次 高 危 险 情 三 类 。 根 据 核 定 ,1996 座 震 损 水 库 中 有 69 座 属 溃 坝 险 情<br />

水 库 ,310 座 水 库 属 高 危 险 情 水 库 ,1617 座 水 库 属 于 次 高 危 和 一 般 险 情 水 库 [4-7] 。 受 损 严 重 的 的 69<br />

座 溃 坝 险 情 水 库 全 部 属 于 小 型 水 库 , 分 布 在 四 川 省 绵 阳 、 德 阳 、 广 元 和 成 都 四 个 市 , 其 中 绵 阳 市 38<br />

座 , 德 阳 市 24 座 , 广 元 市 6 座 , 成 都 市 1 座 。 四 川 省 所 辖 18 个 地 级 市 、3 个 自 治 州 中 , 甘 孜 州 和<br />

____________________________<br />

基 金 项 目 : 地 震 行 业 专 项 : 大 震 生 命 线 工 程 灾 害 损 失 评 估 技 术 研 究 (2010008005)<br />

中 央 级 公 益 性 研 究 所 基 本 科 研 业 务 费 专 项 基 金 项 目 资 助 (051800309)<br />

作 者 简 介 : 梁 海 安 (1980-) 男 , 博 士 研 究 生 。 主 要 从 事 岩 土 工 程 抗 震 与 水 利 工 程 震 害 损 失 研 究 ,E-mail:lianghaian@foxmail.com<br />

-347-


阿 坝 州 没 有 水 库 , 凉 山 州 没 有 水 库 受 损 , 攀 枝 花 市 没 有 水 库 受 损 , 严 重 震 损 的 水 坝 均 位 于 汶 川 地 震<br />

发 震 断 层 的 西 北 侧 。 上 述 水 库 中 有 均 质 土 坝 65 座 , 土 石 混 合 坝 1 座 , 心 墙 石 渣 坝 和 粘 土 心 墙 坝 各<br />

一 座 。 由 于 地 震 后 , 采 用 应 急 处 理 方 式 及 时 得 当 , 上 述 69 座 水 坝 虽 然 震 损 严 重 , 甚 至 一 些 水 坝 在<br />

震 后 被 拆 毁 后 重 建 , 但 在 汶 川 地 震 中 无 一 溃 坝 。<br />

表 1 震 损 水 库 分 类<br />

序 号 水 坝 震 损 等 级 分 级 原 则<br />

水 库 大 坝 及 其 主 体 工 程 发 生 漫 溢 、 出 现 较 大 贯 穿 性 裂 缝 、 上<br />

1 溃 坝 险 情<br />

2 高 危 险 情<br />

3 次 高 危 险 情<br />

下 游 坝 坡 大 面 积 滑 坡 、 大 流 量 集 中 渗 流 等 情 况 , 短 时 间 内 有<br />

导 致 溃 坝 的 险 情 ;<br />

水 库 大 坝 及 其 主 体 工 程 发 生 上 述 险 情 , 可 能 直 接 影 响 大 坝 及<br />

主 要 建 筑 物 安 全 的 险 情<br />

水 库 大 坝 及 其 主 体 工 程 发 生 上 述 险 情 , 但 不 影 响 主 体 工 程 安<br />

全 运 行 的 险 情<br />

二 、 汶 川 地 震 土 石 坝 震 损 特 征<br />

2.1. 震 损 水 坝 数 量 多 、 分 布 广 , 小 型 水 坝 震 害 重<br />

汶 川 地 震 水 库 震 损 分 布 范 围 之 广 前 所 未 有 , 表 2 对 今 年 来 世 界 因 地 震 造 成 损 坏 的 水 坝 数 量 进 行<br />

了 比 较 , 汶 川 地 震 2666 座 震 损 水 库 , 比 1995 年 的 日 本 1983 Nihon-kai OKI 地 震 水 库 数 量 多 出 了 一 倍 ,<br />

是 唐 山 地 震 震 损 水 坝 数 量 的 6 倍 。<br />

表 2 世 界 近 年 地 震 水 坝 震 损 统 计 表<br />

地 震 名 称 国 家 震 级 (MS) 震 损 水 坝 数 ( 座 )<br />

1970 通 海 地 震 中 国 7.7 64<br />

1974 溧 阳 地 震 中 国 6.0 94<br />

1975 海 城 地 震 中 国 7.3 54<br />

1976 格 林 尔 地 震 中 国 6.3 75<br />

1976 唐 山 地 震 中 国 7.8 433<br />

1983 Nihon-kai OKI 地 震 日 本 7.7 238<br />

1995Hyogo-ken nanbu 地 震 日 本 7.2 1362<br />

1996 丽 江 地 震 中 国 7.0 53<br />

2000 姚 安 地 震 中 国 6.5 70<br />

2001Bhuj 地 震 印 度 7.7 240<br />

2008Iwate-Miyagi Nairiku 地 震 日 本 7.2 134<br />

2008 汶 川 地 震 中 国 8.0 2666<br />

-348-


表 3 汶 川 地 震 及 余 震 震 损 水 坝 数 量 统 计 表<br />

序<br />

号<br />

省 市<br />

水 坝 总 数<br />

( 座 )<br />

震 损 水 坝<br />

( 座 )<br />

占 各 省 ( 市 ) 水 坝<br />

总 数 比 例 (%)<br />

占 震 损 水 坝 总 数 比<br />

例 (%)<br />

险 情 分 类<br />

溃 坝 险 情 高 危 险 情 次 高 危 险 情<br />

1 四 川 6717 1996 29.71 74.87 69 310 1617<br />

2 重 庆 2824 352 12.46 13.20 2 350<br />

3 陕 西 1036 126 12.16 4.73 17 109<br />

4 云 南 5422 51 0.94 1.91 2 49<br />

5 甘 肃 297 81 27.27 3.04 81<br />

6 贵 州 2105 12 0.57 0.45 12<br />

7 湖 北 5804 25 0.43 0.94 25<br />

8 湖 南 11435 23 0.20 0.86 23<br />

合 计 35640 2666 69 331 2266<br />

如 表 3 所 示 。 汶 川 地 震 震 损 水 坝 涉 及 八 个 省 , 四 川 省 受 损 水 库 占 绝 大 多 数 。 汶 川 地 震 地 震 中 共<br />

有 69 座 水 坝 处 于 溃 坝 险 情 , 均 为 小 型 水 库 ;310 座 水 库 属 高 危 险 情 水 库 中 , 无 大 型 水 库 , 仅 12 座 中<br />

型 水 坝 , 其 余 298 座 均 为 小 型 水 坝 。 溃 坝 险 情 的 89 座 水 库 中 , 10 米 以 下 和 10-20 范 围 水 坝 各 占 震 损 溃<br />

坝 水 库 总 数 的 27.53% 和 57.97%, 坝 高 大 有 20 米 的 水 坝 仅 占 14.48%, 小 型 水 坝 的 震 害 程 度 远 远 超 出 了<br />

大 中 型 水 坝 。<br />

2.2. 汶 川 地 震 震 害 种 类 多 , 以 地 震 波 造 成 的 震 害 多 、 破 坏 重<br />

汶 川 地 震 中 震 损 的 水 坝 震 害 现 象 众 多 , 主 要 表 现 为 坝 体 裂 缝 、 坝 身 滑 坡 、 渗 漏 管 涌 、 沉 陷 、 隆<br />

起 变 形 、 冒 水 冒 砂 , 防 浪 墙 倒 塌 、 护 坡 裂 缝 滑 塌 、 附 属 设 施 震 损 等 等 。 其 中 以 坝 体 裂 缝 、 隆 起 变 形 、<br />

坝 体 滑 坡 、 隆 起 变 形 和 沉 陷 对 坝 体 安 全 威 胁 最 大 。 以 四 川 省 1996 座 震 损 水 库 为 例 , 如 图 1 所 示 ,<br />

险 情 主 要 表 现 为 大 坝 裂 缝 (1425 座 )、 所 占 比 例 为 71.89%; 大 坝 塌 陷 (687 座 ) 所 占 比 例 34.42%、<br />

滑 坡 (354 座 ) 所 占 比 例 为 17.74、 渗 漏 (428 座 ) 所 占 比 例 为 21.44%, 均 属 地 震 波 造 成 的 破 坏 ; 由<br />

附 属 设 施 破 破 坏 的 类 型 有 : 起 闭 设 施 损 坏 (161 座 ) 占 8.07%, 其 他 放 水 设 施 、 溢 洪 道 、 管 理 房 不<br />

同 程 度 震 损 (422 座 ) 等 形 式 占 21.14%, 没 有 水 坝 因 涵 管 震 损 造 成 渗 漏 冲 毁 坝 体 的 情 况 。<br />

震 损 比 例 (%)<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

汶 川 地 震 水 坝 震 害 类 型 及 震 损 比 例<br />

震 损 比 例<br />

坝 体 裂 缝 坝 体 沉 陷 坝 体 滑 坡 坝 体 渗 漏 启 闭 设 施 损 坏 其 他<br />

震 害 类 型<br />

图 1 汶 川 地 震 水 坝 震 害 类 型 及 震 损 比 例<br />

2.3. 重 灾 区 震 损 水 坝 数 量 沿 发 震 断 层 展 布 , 大 体 随 震 中 距 的 增 大 数 量 减 少<br />

汶 川 地 震 中 受 损 水 坝 绝 大 部 分 位 于 发 震 断 裂 的 东 南 侧 , 大 体 沿 北 东 向 展 布 [6-11] 。 由 图 2 可 以 看 到<br />

-349-


汶 川 地 震 中 属 于 溃 坝 险 情 的 69 座 水 坝 基 本 位 于 山 前 盆 地 边 缘 的 过 渡 带 , 处 于 发 震 断 裂 西 南 侧 。 地 震<br />

震 中 位 置 的 西 部 及 北 部 基 本 没 有 严 重 震 损 水 坝 , 震 害 水 坝 基 本 位 于 震 中 的 东 南 方 向 。 统 计 数 据 显 示 :<br />

震 中 距 50-200 公 里 范 围 内 集 中 了 绝 大 多 数 溃 坝 险 情 水 坝 。 但 在 震 中 距 250 公 里 以 外 , 仍 有 溃 坝 险 情 的<br />

水 坝 。<br />

101 102 103 104 105 106 107 108 E109<br />

35<br />

Ⅵ<br />

Ⅴ<br />

34<br />

Ⅵ<br />

Ⅶ<br />

Ⅵ<br />

Xian<br />

33<br />

Ⅷ<br />

Ⅶ<br />

32<br />

31<br />

30<br />

29<br />

Wenchuan<br />

雅 安<br />

Ⅷ<br />

成 都<br />

Ⅹ<br />

眉 山<br />

德 阳<br />

绵 阳<br />

资 阳<br />

内 江<br />

自 贡<br />

宜 宾<br />

Ⅵ<br />

Ⅸ<br />

遂 宁<br />

泸 州<br />

广 元<br />

南 充<br />

巴 中<br />

广 元<br />

重 庆<br />

达 州<br />

Ⅵ<br />

N28<br />

Ⅵ<br />

溃 坝 险 情 水 坝<br />

图 2 汶 川 地 震 中 溃 坝 险 情 水 坝 分 布 图<br />

2.4. 余 震 对 水 坝 震 害 发 展 影 响 大<br />

汶 川 地 震 后 16 天 , 国 家 对 外 发 布 1803 座 水 库 受 损 , 随 着 汶 川 地 震 及 其 余 震 的 发 生 , 经 部 、 省 、<br />

市 、 县 各 级 联 合 排 查 , 截 止 6.12 日 , 四 川 全 省 水 库 共 有 1996 座 水 库 发 生 震 损 , 较 5.28 日 增 加 193 座 ,<br />

主 要 原 因 为 余 震 和 核 查 深 入 而 新 增 的 。 汶 川 地 震 发 生 后 , 余 震 活 动 频 繁 , 截 止 2009-3.12 日 共 发 生 四<br />

级 以 上 地 震 313 次 , 约 80% 发 生 在 2008.5-6 月 。 频 繁 的 余 震 一 方 面 增 加 了 震 损 水 库 的 数 量 , 另 一 方 面<br />

加 剧 了 众 多 震 损 水 库 的 震 害 程 度 。 具 体 表 现 为 : 坝 体 裂 缝 数 量 和 规 模 增 加 , 坝 体 渗 流 加 重 、 引 发 坝<br />

体 滑 坡 及 附 属 设 施 震 害 等 。 以 绵 阳 市 安 县 丰 收 水 库 为 例 ,5.12 主 震 后 , 大 坝 出 现 了 较 为 严 重 的 坝 体<br />

裂 缝 ( 如 图 3), 在 2 次 6 级 以 上 余 震 加 重 水 库 震 害 , 引 起 坝 体 滑 坡 ( 如 图 4)。 岐 山 水 库 在 5.12 日 地 震 后 在<br />

坝 顶 出 现 90 米 , 最 宽 21 厘 米 的 纵 向 裂 缝 , 至 5.25 日 , 在 余 震 作 用 下 裂 缝 长 达 132 米 , 宽 26 厘 米 , 裂 缝<br />

两 侧 出 现 位 错 达 10 厘 米 , 坝 顶 出 现 约 7 厘 米 的 沉 陷 。<br />

图 13 丰 收 水 库 坝 体 震 害 概 况 ( 摄 于 2008.5.15) 图 14 丰 收 水 库 震 害 概 况 ( 摄 于 2008.8.16)<br />

-350-


2.5 多 数 溃 坝 险 情 水 坝 由 于 坝 体 裂 缝 造 成 , 液 化 造 成 溃 坝 险 情 比 例 很 低 。<br />

四 川 省 1996 座 震 损 水 库 中 , 险 情 主 要 表 现 大 坝 裂 缝 的 计 有 1425 座 , 占 总 数 的 71.39%。 表 4<br />

对 69 座 溃 坝 险 情 水 坝 震 害 情 况 行 了 统 计 分 析 , 由 统 计 数 据 可 知 : 共 有 68 座 出 现 裂 缝 , 占 98.55%。<br />

唯 一 没 有 明 显 裂 缝 震 害 的 德 阳 市 庙 儿 嘴 水 库 大 坝 , 判 断 是 由 于 坝 基 轻 微 液 化 隔 震 , 从 而 减 轻 了 坝 体<br />

本 身 的 震 害 。 水 坝 最 容 易 产 生 破 坏 的 位 置 是 坝 顶 。 无 论 横 向 裂 缝 和 纵 向 裂 缝 , 在 坝 顶 出 现 的 比 例 是<br />

最 大 的 。 上 游 坝 坡 产 生 纵 向 裂 缝 破 坏 的 情 况 要 略 重 于 下 游 坝 坡 。 横 向 裂 缝 在 较 低 烈 度 区 产 生 的 比 例<br />

远 低 于 较 高 烈 度 地 区 , 贯 穿 性 横 向 裂 缝 主 要 出 现 在 坝 顶 , 尤 其 是 在 坝 肩 处 和 主 坝 拐 弯 处 最 容 易 出 现 ,<br />

此 类 裂 缝 在 库 水 位 较 高 的 情 况 下 是 极 为 危 险 的 , 短 时 间 内 可 能 造 成 库 水 渗 流 的 通 道 , 引 起 冲 刷 和 管<br />

涌 , 导 致 溃 坝 。<br />

地<br />

震<br />

烈<br />

度<br />

严 重<br />

震 害<br />

的 土<br />

坝 总<br />

数<br />

( 座 )<br />

坝 数<br />

( 座 )<br />

表 4 溃 坝 险 情 水 坝 害 类 型 统 计 表<br />

裂 缝 滑 坡 沉 陷<br />

所 占<br />

百 分<br />

数<br />

(%)<br />

坝 数<br />

( 座 )<br />

所 占<br />

百 分<br />

数<br />

(%)<br />

坝 数<br />

( 座 )<br />

所 占<br />

百 分<br />

数<br />

(%)<br />

漏 水 或 漏 水 量<br />

增 加<br />

所 占<br />

坝 数 百 分<br />

( 座 ) 数<br />

(%)<br />

鼓 包 变 形<br />

坝 数<br />

( 座 )<br />

所 占<br />

百 分<br />

数<br />

(%)<br />

VI 2 2 2.89 0 0 1 1.44 1 1.44 1 1.44<br />

VI<br />

I<br />

20 20 28.98 2 2.89 5 7.24 12 17.39 3 4.34<br />

I 8 27 39.13 5 7.24 15 21.73 7 10.14 9 13.04<br />

IX 19 19 27.53 9 13.04 14 20.28 5 7.24 2 2.89<br />

总<br />

计<br />

69 68 98.55 16 23.18 35 50.72 25 36.23 15 21.73<br />

此 外 发 生 严 重 滑 坡 的 水 坝 共 16 座 , 其 中 13 座 水 坝 为 上 游 坝 坡 滑 坡 , 另 外 有 两 座 位 于 9 度 烈 度<br />

区 的 水 库 发 生 背 水 坡 滑 坡 , 剩 余 一 座 背 水 坡 滑 坡 位 于 6 度 区 , 有 坝 顶 渠 道 震 裂 , 渗 水 侵 入 坝 体 遭 受<br />

局 部 滑 坡 。 另 14 座 水 坝 存 在 潜 在 滑 坡 迹 象 。 汶 川 地 震 中 鼓 包 变 形 位 置 多 存 在 于 水 坝 背 水 坡 , 伴 随<br />

着 水 坝 裂 缝 和 滑 坡 同 时 发 生 。 由 于 大 多 数 震 损 水 坝 在 震 后 较 短 时 间 内 采 取 了 放 低 库 水 位 的 应 急 措<br />

施 , 水 坝 渗 漏 现 象 并 不 明 显 , 很 多 水 坝 渗 漏 在 放 低 库 水 位 后 减 轻 或 者 停 止 渗 漏 。 由 于 四 川 省 水 坝 多<br />

位 于 山 区 或 者 丘 陵 地 区 , 坝 体 和 坝 基 土 料 少 有 饱 和 粉 土 或 砂 土 , 坝 体 液 化 现 在 在 本 次 地 震 中 罕 见 。<br />

三 、 汶 川 地 震 土 石 坝 震 害 影 响 因 素<br />

3.1. 烈 度 的 影 响<br />

5.12 汶 川 地 震 中 , 震 源 特 性 及 地 质 构 造 、 场 地 条 件 等 决 定 了 了 汶 川 地 震 的 烈 度 分 布 情 况 。 汶 川<br />

地 震 存 在 两 个 震 中 Ⅺ 度 区 , 近 似 于 龙 门 山 断 裂 平 行 , 震 中 位 置 处 于 两 个 Ⅺ 度 区 的 西 南 方 。 震 后 统 计<br />

数 据 表 明 :Ⅴ 度 区 水 坝 基 本 没 有 严 重 的 震 害 ,Ⅵ、Ⅶ、Ⅷ 度 区 有 溃 坝 险 情 的 水 坝 数 位 2,20,28 座 ,<br />

占 全 部 溃 坝 险 情 水 坝 的 2.89%、28.98%、39.13%。9 度 区 面 积 不 足 Ⅷ 度 区 的 1/3, 但 震 损 比 例 占 全 部<br />

溃 坝 险 情 的 27.53%。 随 着 烈 度 的 增 加 , 水 坝 震 害 类 型 也 逐 步 增 多 ,Ⅶ 度 区 震 害 以 裂 缝 居 多 , 个 别 或<br />

少 数 兼 沉 陷 、 渗 漏 、 以 及 鼓 包 变 形 , 随 着 烈 度 的 增 大 , 震 害 类 型 也 随 着 增 多 。 至 Ⅸ 度 区 时 , 各 种 震<br />

害 伴 随 发 生 , 除 了 普 遍 或 多 数 仍 以 裂 缝 为 主 要 震 害 形 式 外 , 滑 坡 、 渗 漏 、 沉 陷 、 鼓 包 变 形 等 震 害 算<br />

占 的 比 例 都 明 显 增 加 。 可 见 随 着 地 震 烈 度 的 增 加 , 震 损 水 坝 的 数 量 和 震 害 类 型 都 随 着 显 著 提 高 。<br />

3.2. 坝 基 的 影 响<br />

总 的 来 说 , 基 岩 坝 基 的 水 坝 震 害 要 比 非 基 岩 坝 基 或 者 软 弱 地 基 轻 。 汶 川 地 震 中 水 坝 多 分 布 于 中 、<br />

底 山 区 或 者 丘 陵 地 区 , 覆 盖 层 厚 度 很 薄 , 坝 基 情 况 差 异 不 大 。 软 弱 地 基 造 成 的 震 害 多 为 渗 漏 和 管 涌 ,<br />

在 烈 度 不 是 很 高 的 地 区 也 能 造 成 危 重 的 震 害 。 例 如 德 阳 八 一 水 库 , 处 于 7 度 区 , 震 中 距 约 70 公 里 。 坝<br />

基 下 存 在 饱 和 粉 细 砂 层 。 震 后 大 坝 上 游 砼 护 坡 裂 缝 加 长 、 加 宽 并 增 多 , 共 发 现 横 向 缝 20 条 , 总 长 250m,<br />

-351-


缝 宽 约 2mm~5mm; 纵 向 断 裂 缝 5 条 , 总 长 100m, 缝 宽 约 2mm。 同 时 坝 体 散 浸 及 漏 水 、 坝 后 管 涌 等 。<br />

被 认 定 该 水 库 安 全 问 题 严 重 , 存 在 溃 坝 风 险 。 汶 川 地 震 中 , 紫 坪 铺 坝 址 距 震 中 约 17km, 烈 度 在 Ⅸ 度 以<br />

上 , 在 地 震 中 震 损 轻 微 , 良 好 的 坝 基 为 大 坝 安 全 起 了 重 要 的 作 用 。<br />

3.3. 施 工 质 量 的 影 响<br />

施 工 质 量 的 影 响 最 主 要 体 现 在 两 个 方 面 , 一 是 施 工 过 程 中 对 坝 体 碾 压 质 量 的 控 制 , 二 是 施 工 中 对<br />

坝 基 及 坝 肩 部 位 的 清 理 程 度 。 汶 川 地 震 中 大 量 的 水 坝 属 于 “ 三 边 ” 工 程 , 施 工 质 量 较 差 。 一 方 面 坝 体 碾<br />

压 遍 数 达 不 到 要 求 , 造 成 坝 体 强 度 不 足 , 震 前 便 是 病 险 水 库 , 地 震 时 可 想 而 知 。 如 长 河 水 库 震 前 上 游<br />

边 坡 便 已 发 生 塌 滑 , 坝 坡 陡 立 , 震 后 有 形 成 新 的 滑 塌 和 裂 缝 。 另 一 方 面 , 由 于 没 有 彻 底 的 清 理 坝 基 和<br />

坝 肩 , 造 成 水 库 运 行 后 便 出 现 坝 基 渗 漏 和 坝 肩 漏 水 , 地 震 左 右 下 这 些 现 象 更 加 严 重 。 汶 川 地 震 中 严 重<br />

震 损 的 基 本 都 是 小 型 水 坝 , 大 、 中 型 水 坝 受 损 轻 微 , 究 其 原 因 , 施 工 质 量 起 了 很 重 要 的 作 用 , 大 型 水<br />

库 对 坝 料 的 选 择 , 施 工 质 量 的 监 督 较 为 规 范 , 加 之 在 坝 体 自 重 下 坝 料 的 固 结 程 度 均 要 优 于 小 型 水 坝 ,<br />

因 此 造 成 的 震 害 也 就 轻 微 的 多 。<br />

3.4. 坝 料 的 影 响<br />

汶 川 地 震 中 震 损 的 大 多 数 为 小 型 水 坝 , 坝 料 多 就 地 取 材 , 选 择 当 地 的 粘 土 、 亚 粘 土 或 壤 土 或 风 化<br />

残 积 土 为 筑 坝 材 料 。 坝 型 以 均 质 土 坝 为 主 , 根 据 坝 料 土 工 实 验 资 料 , 多 数 震 损 严 重 的 水 坝 坝 体 材 料 处<br />

于 欠 密 实 、 很 湿 或 较 湿 状 态 , 坝 料 干 密 度 多 处 于 1.55g/cm3 左 右 , 有 些 坝 体 材 料 干 密 度 才 1.48 g/cm3。<br />

在 地 震 作 用 下 , 坝 体 产 生 裂 缝 和 塌 陷 就 不 足 为 奇 了 。 坝 料 就 地 取 材 , 筑 坝 材 料 的 级 配 控 制 就 难 以 控 制 ,<br />

本 来 就 疏 松 的 坝 体 , 在 级 配 比 较 差 的 情 况 下 , 容 易 发 生 渗 漏 , 乃 至 发 生 管 涌 造 成 溃 坝 危 险 。<br />

3.5. 水 库 维 护 管 理 水 平 影 响<br />

汶 川 地 震 中 69 座 溃 坝 险 情 水 库 , 绝 大 多 数 都 存 在 各 类 病 险 问 题 , 除 去 施 工 质 量 造 成 的 渗 漏 滑 坡 问<br />

题 、 坝 体 变 形 裂 缝 问 题 外 , 近 一 半 水 库 有 白 蚁 病 害 、 上 游 坝 坡 由 于 淘 刷 以 及 虾 蟹 寄 居 出 现 陡 坎 洞 穴 。<br />

还 有 一 些 水 库 溢 洪 道 堵 塞 、 启 闭 设 施 运 转 不 灵 , 这 些 疏 于 维 护 的 水 坝 在 地 震 侵 袭 时 , 更 容 易 造 成 严 重<br />

的 破 坏 。 此 外 , 汶 川 地 震 中 , 大 多 数 震 损 水 坝 建 造 于 上 世 纪 70 年 代 , 没 有 或 很 少 考 虑 抗 震 问 题 和 采 取<br />

必 要 的 抗 震 结 构 和 工 程 措 施 , 很 多 坝 体 几 何 尺 寸 不 合 理 , 坝 体 过 于 单 薄 , 坝 坡 过 陡 , 而 且 许 多 没 有 考<br />

虑 排 水 结 构 和 设 置 护 坡 , 致 使 这 类 水 坝 出 现 不 同 程 度 的 震 害 。 但 对 于 这 一 年 代 的 水 坝 , 一 定 要 加 强 平<br />

时 的 维 护 管 理 , 确 保 其 安 全 , 使 之 继 续 发 挥 本 该 起 到 的 作 用 。<br />

综 上 所 述 , 汶 川 地 震 中 , 由 地 震 震 级 、 频 谱 特 征 、 持 时 等 地 震 要 素 确 定 了 本 次 地 震 的 烈 度 区 域 ,<br />

震 损 水 坝 的 震 损 比 例 、 震 害 类 型 随 这 烈 度 区 的 等 级 的 增 加 而 增 大 。 坝 基 的 类 型 、 水 坝 的 施 工 质 量 和 坝<br />

料 的 好 坏 对 水 坝 的 震 损 情 况 起 了 关 键 性 的 作 用 。 水 坝 平 时 的 维 护 管 理 对 地 震 时 水 坝 的 震 损 有 影 响 , 另<br />

外 , 震 前 水 库 的 库 水 位 对 水 坝 地 震 时 的 安 全 性 应 当 有 重 要 的 影 响 , 但 汶 川 地 震 中 益 于 对 震 损 水 库 安 全<br />

的 高 度 重 视 , 措 施 及 时 有 力 , 库 水 位 的 作 用 没 有 得 当 充 分 显 现 。<br />

四 、 结 论<br />

5.12 汶 川 地 震 造 成 的 水 坝 震 害 数 量 大 , 范 围 广 , 震 害 类 型 多 , 以 小 型 水 库 震 害 为 主 。 从 影 响 汶 川<br />

地 震 水 坝 震 害 特 点 和 影 响 因 素 来 看 , 我 们 可 以 得 出 以 下 结 论 :<br />

1) 水 坝 的 场 地 和 地 基 方 面 , 我 们 一 定 要 做 好 勘 察 , 避 开 和 远 离 断 层 。 从 汶 川 地 震 大 型 水 库 的 震 害<br />

经 验 来 看 , 只 有 水 坝 选 址 没 有 问 题 , 地 震 中 受 到 的 损 害 是 有 限 的 , 土 石 坝 本 身 是 有 很 大 的 抗 震 潜 能 的 。<br />

地 基 选 择 上 应 当 尽 可 能 的 选 择 基 岩 坝 基 或 者 以 良 好 的 持 力 层 为 地 基 。 尽 量 避 让 软 弱 土 层 地 基 。<br />

2) 设 计 和 施 工 质 量 方 面 , 采 用 设 计 合 理 的 坝 体 几 何 尺 寸 , 对 所 用 筑 坝 土 料 进 行 合 理 的 筛 选 ,<br />

控 制 坝 体 各 部 分 土 料 的 级 配 , 采 取 必 要 的 抗 震 结 构 和 工 程 措 施 , 在 水 坝 开 始 施 工 时 重 视 坝 基 和 坝 肩<br />

的 清 理 工 作 , 不 留 隐 患 , 这 是 保 证 和 增 强 水 坝 抗 震 性 能 的 关 键 性 措 施 。<br />

3) 水 坝 的 坝 顶 、 水 坝 坝 肩 处 及 水 坝 拐 弯 处 , 是 贯 穿 性 横 向 裂 缝 易 发 部 位 。 应 当 重 点 加 强 这 些<br />

-352-


部 位 的 施 工 质 量 。<br />

4) 小 型 水 库 是 地 震 中 敏 感 而 脆 弱 的 承 灾 体 , 小 型 水 库 的 土 石 坝 病 险 多 , 管 理 水 平 差 , 得 益 于<br />

对 震 损 水 库 安 全 的 高 度 重 视 , 措 施 及 时 有 力 本 次 地 震 没 有 发 生 溃 坝 , 但 应 对 小 型 水 库 的 日 常 维 护 和<br />

管 理 应 当 更 加 重 视 。<br />

5) 评 估 水 坝 的 震 损 情 况 不 应 过 分 强 调 微 观 震 中 距 影 响 , 结 合 多 方 面 情 况 综 合 评 估 。 大 震 中 发<br />

震 断 层 规 模 较 大 , 等 震 线 通 常 不 为 规 则 的 圆 形 , 评 估 震 害 过 分 强 调 水 坝 与 震 中 的 关 系 , 会 造 成 评 估<br />

结 果 的 偏 差 。<br />

参 考 文 献<br />

[1] 宋 胜 武 , 蒋 峰 , 陈 万 涛 . 汶 川 地 震 灾 区 大 中 型 水 电 工 程 震 损 特 征 初 步 分 析 . 四 川 水 利 发 电 2009, 28(2): 1-7.<br />

[2] 马 文 涛 , 李 , 杨 主 恩 , 陈 桂 华 , 陈 献 程 , 杨 清 源 , 邓 志 辉 , 孙 谦 . 汶 川 Ms8.0 地 震 对 四 川 省 水 电 水 利 工 程 场 地<br />

安 全 性 评 价 结 果 的 检 验 . 地 震 地 质 2008,30,(3):796-803.<br />

[3] 景 立 平 , 陈 国 兴 , 李 永 强 , 等 . 汶 川 8.0 级 地 震 水 坝 震 害 调 查 . 地 震 工 程 与 工 程 振 动 , 2009, 29(1): 14-23.<br />

[4] 吴 世 泽 . 高 烈 度 区 土 ( 石 ) 坝 震 损 特 征 与 成 因 初 步 研 究 . 人 民 长 江 , 2008, 39(22): 63-66.<br />

[5] 文 松 霖 . 汶 川 地 震 绵 阳 市 水 库 震 损 现 象 及 发 生 机 理 初 探 . 水 利 水 电 技 , 2009, 40(2): 70-73.<br />

[6] 景 立 平 , 陈 国 兴 , 李 永 强 , 王 宝 光 . 汶 川 地 震 中 江 油 市 水 坝 震 害 调 查 与 分 析 [J]. 世 界 地 震 工 程 2009,25,(2):1-10.<br />

[7] 陈 国 兴 , 景 立 平 , 周 新 贵 , 等 . 汶 川 8.0 级 地 震 绵 阳 市 游 仙 区 水 库 震 害 调 查 与 分 析 . 防 灾 减 灾 工 程 学 报 , 2009,<br />

29(3): 347-355.<br />

[8] 陈 运 泰 , 张 冯 许 周 , 2008 年 汶 川 大 地 震 的 时 空 破 裂 过 程 . 中 国 科 学 D 辑 : 地 球 科 学 , 2008, 38(10): 1186-1194.<br />

[9] 李 志 强 , 袁 一 凡 , 李 晓 丽 , 何 萍 . 对 汶 川 地 震 宏 观 震 中 和 极 震 区 的 认 识 . 地 震 地 质 , 2008, 30(3): 768-777.<br />

[10] 于 海 英 , 王 栋 , 杨 永 强 , 等 . 汶 川 8.0 级 地 震 强 震 动 加 速 度 记 录 的 初 步 分 析 . 地 震 工 程 与 工 程 振 动 , 2009,<br />

29(1): 1-13.<br />

[11] 中 国 地 震 震 害 防 御 司 , 汶 川 8.0 级 地 震 未 校 正 加 速 度 记 录 , 北 京 : 地 震 出 版 社 , 2008.<br />

-353-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

LESSONS LEARNED FROM 2011 TOHOKU EARTHQUAKE AND TSUNAMI<br />

K.T. Chau<br />

Department of Civil and Structural Engineering,<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong>, <strong>Hong</strong> <strong>Kong</strong>, China.<br />

ABSTRACT<br />

<strong>The</strong> largest earthquake ever recorded in Japan’s history occurred on March 11, 2011 off shore of Sendai area.<br />

This earthquake measured 9.0 in moment magnitude, and led to a death toll of 14,616 (up to May 4, 2011), and<br />

11,111 people missing and over 5,278 injured. <strong>The</strong> earthquake caused structural damages to many buildings<br />

in Sendai area and the tsunami is even more destructive, leading to nearly complete destruction at the cities of<br />

Miyako, Kamaishi, Ofunato, Rikuzentakata, Kesenuma, Minamisanriku, Ishinomaki, Natori, and Soma. This<br />

is probably is the most destructive tsunami in Japan history. <strong>The</strong> earthquake also induced a number of<br />

explosions at the Fukushima I Nuclear Power Plant. Radioactive iodine, cesium and plutonium had been<br />

detected both in Japan and in most countries in the North hemisphere. This paper summarizes a number issues<br />

raised by this huge earthquake, including the inadequacy of our seismic hazard estimation technique, the<br />

uncertainty in focal mechanism estimation, and inadequacy of tsunami mitigation strategies.<br />

KEYWORDS<br />

2011 Tohoku earthquake, tsunami, seismic hazard estimation.<br />

INTRODUCTION<br />

<strong>The</strong> largest earthquake recorded in the history of Japan occurred on March 11, 2011 offshore of the Sendai area<br />

of Miyagi Prefecture. This earthquake measured 9.0 in moment magnitude (United States Geological Survey<br />

(USGS) raised the moment magnitude from an initial of 7.9 to 8.8, 8.9 and then finally to 9.0), and led to the<br />

numbers of death, injured and missing of 14,616, 11,111 and 5,278 up to May 4, 2011. It was estimated that over<br />

240,000 people still live in temporary camps and a total of 19,000 buildings were destroyed (mainly by the<br />

associated tsunami). <strong>The</strong> initial earthquake magnitude issued by the Japan Meteorological Agency is 8.4 and<br />

was then raised to 8.8 and finally 9.0, which is the same as the final magnitude assigned by USGS. This is the<br />

fourth largest earthquake ever recorded since the invention of modern seismograph. It is only smaller than the<br />

1960 9.5 Chile earthquake, 1964 9.2 Alaska earthquake and the 2004 9.1 Sumatra earthquake, and equal by the<br />

1952 9.0 Kamchatka earthquake (all magnitudes are in moment magnitude).<br />

SEISMIC HAZARD ESTIMATION<br />

Cornell’s (1968) Method<br />

<strong>The</strong> most commonly adopted method for seismic hazard estimation is called Cornell’s (1968) method. <strong>The</strong> full<br />

details of this method are referred to Cornell (1968) and to Chau et al. (2000). In short, historical earthquake data<br />

within a region of concern have to be analyzed to generate a Richter-Gutenberg type recurrence relation, which<br />

predicts the occurrence frequency of earthquake as a function of magnitude. <strong>The</strong> local attenuation relation for<br />

ground motions have to be established by historical or instrumental isoseismal maps. <strong>The</strong> seismic sources around<br />

the site of interest have to be divided into potential seismic sources (normally in the shape of a strip resembling the<br />

underlying fault direction). <strong>The</strong>n, a maximum credible earthquake magnitude has to be assigned for each source<br />

zone, based on either local geological structures that are conducive to earthquake occurrence or by historical<br />

records. By assuming earthquake occurrence is an independent process (i.e. one earthquake would not influence<br />

the occurrence of the next earthquake), we can use Poisson distribution of probability of occurrence to get the<br />

probability of shaking at a particular site through the integration of both magnitude ranks and seismic sources.<br />

Typically, a magnitude rank of ½ is normally used to conduct the summation process for earthquakes starting from<br />

magnitude 4.5 to the maximum credible earthquake magnitude of the seismic source (say 7.5), and summation is<br />

also done for spatially distributed seismic sources. Because of ground motion attenuation, the contribution from<br />

far sources is generally smaller than that of the nearby ones.<br />

-354-


Japan’s Seismic Hazard Map and the 2011 Tohoku Earthquake<br />

Although various government agencies had modified or fine-tuned the Cornell’s (1968) method, the process of<br />

hazard estimated is essentially same as those adopted by Cornell. As shown in Figure 1, the Earthquake<br />

Research Committee of the Headquarters for Earthquake Promotion (hereafter, referred to as Earthquake<br />

Research Committee (ERC)) in 2010 published the guidebook for the Japan’s “National Seismic Hazard Maps”<br />

and predicted that there is 99% chance of earthquake occurrence at the Miyagi-Oki source region in the next 30<br />

years (the source zone for the 2011 9.0 Tohoku earthquake and tsunami) and the predicted magnitude is M7.5<br />

(ERC, 2010). In a sense, this 2011 Tohoku earthquake is expected but the magnitude is unexpected.<br />

Figure 1 <strong>The</strong> Japanese hazard map for the off-shore seismic sources, including the Miyagi-Oki (the source of the<br />

2011 Tohoku earthquake) (after ERC, 2010)<br />

This hazard map illustrates a major problem of the currently-adopted seismic hazard analysis. That is, the<br />

maximum credible earthquake of a source zone is generally limited by the largest historical earthquake of this<br />

source. Actually, Cornell’s idea of hazard analysis is in fact based on past earthquake records (both the<br />

earthquake reoccurrence frequency and the maximum magnitude) but this can be unreliable. As summarized<br />

by Kanamori et al. (2006), the earthquake sequence from the offshore of Miyagi prefecture, Japan, was<br />

recognized by the ERC. In particular, the following sequence of earthquakes has been documented: 1793 (8.2),<br />

1835 (7.3), 1861 (7.4), 1897 (7.4), 1936 (7.4), 1978 (7.4), 2005 (7.2) (numbers in brackets are earthquake<br />

magnitudes). <strong>The</strong>re is a striking regularity in return period of 37±7 years and all earthquakes are of comparable<br />

magnitude of 7.3-8.2, and this leads to an illusion that Reid’s (1910) “elastic rebound model” works perfectly<br />

for this source zone. <strong>The</strong> 2011 Tohoku earthquake wakes us up that nature is much more complicated than<br />

what we thought. This also casts doubts on the recurrence model of Richter and Gutenberg that we commonly<br />

adopted for seismic hazard analysis. Recently, Chau (2010a) also questioned the use of past earthquake<br />

records for seismic hazard estimation even before the 2011 Tohoku Earthquake and Tsunami.<br />

-355-


Problem with Seismic Hazard Analysis<br />

Actually, the 2011 Tohoku earthquake is not the sole example that defies the results of our current adopted<br />

seismic hazard analysis. In fact, most of the notable earthquakes occurred in the last 30 years had repeatedly<br />

illustrated that our seismic hazard analysis is inadequate. For example, the Chelungpu fault before the 1999<br />

Mw 7.6 Chi-Chi Earthquake is largely inactive. So, it came as a surprise in Taiwan when it killed 2415 people<br />

on September 21, 1999. <strong>The</strong> largest historical earthquake induced by the Awaji fault is M=6 in 1916, yet it<br />

produced the 1995 7.2 Kobe Earthquake that killed 5,094. Kanamori (1995) reported that the Awaji fault,<br />

which generated the 1995 Kobe earthquake, is not included in the 1994 review of potential seismic hazard the<br />

Kinki district by the “Committee of Earthquake Observation and Research” in the Kansai Area. <strong>The</strong>refore,<br />

1995 kobe Earthquake also came as a surprise. <strong>The</strong> more recent 2008 Wenchuan earthquake occurred on the<br />

Longmenshan fault, at which the largest historical earthquake is the 1657 April 21 Wenchuan Earthquake (M =<br />

6.5) (Chau, 2009). <strong>The</strong> estimated seismic hazard level by the China Earthquake Administration at Yingxiu and<br />

Beichuan (the worst hit towns) before the 2008 Wenchuan earthquake is 0.1g which is even lower than 0.15g of<br />

<strong>Hong</strong> <strong>Kong</strong>. <strong>The</strong> 2010 Mw 6.9 Yushu Earthquake happened on a segment of the Ganzi-Yushu fault with a<br />

largest historical earthquake of 6.5 (Chau, 2010b; Zhou et al., 1997). I think it is obvious that we have major<br />

problem with our current terminology in estimating seismic hazard. Probably, it is time to revolutionize the<br />

idea of Cornell’s (1968) method.<br />

UNCERTAINTY IN ESTIMATING THE FOCAL MECHANISM<br />

<strong>The</strong> focal mechanism of the huge 2011 Tohoku earthquake has been studied by various researchers. Figure 2<br />

compiled six different versions of the focal mechanism. <strong>The</strong>se versions were proposed by: (a) USGS (United<br />

States Geological Survey); (b) NIED (National Institute for Earth Science and Disaster Prevention); (c) Caltech<br />

(California Institute of Technology); (d) Nagoya <strong>University</strong>; (e) Tsukuba <strong>University</strong>; and (f) GFZ (German<br />

Research Center for Geosciences). <strong>The</strong> activated fault plane is normally interpreted from the distribution of<br />

the aftershocks. <strong>The</strong>refore, the fault plane is rather consistent amount these different models. However, the<br />

slip distributions as well as the maximum slips are substantially different. <strong>The</strong> maximum slips are 18m, 25m,<br />

23m, 10m, 30m, and 30m for the models by USGS, NIED, Caltech, Nagoya <strong>University</strong>, Tsukuba <strong>University</strong>,<br />

and GFZ respectively. Clearly, they are completely different. Except for the NIED model, all these back<br />

analyses were based on the far field velocity data observed at seismic stations thousands of kilometres from the<br />

epicentre. <strong>The</strong> locations and numbers of seismic stations used in back analyses vary from one model to the<br />

other. <strong>The</strong> sensitivity of slip distribution is clearly not high from far field data. Unfortunately, seismologists<br />

and geophysicists did not seek for a better back analysis technique. To be fair, most of their results are<br />

questionable since all of them fit real data well and it also means that none of them are really reliable.<br />

(a)<br />

(c)<br />

(e)<br />

(b)<br />

(d)<br />

(f)<br />

Figure 2 Focal mechanisms in terms of slip distribution given by (a) USGS, (b) NIED, (c) Caltech, (d) Nagoya<br />

<strong>University</strong>, (e) Tsukuba <strong>University</strong>, (f) GFZ.<br />

-356-


Mathematically speaking, finding the focal mechanism (in terms of slip distribution) of an earthquake is an<br />

inverse problem that the boundary condition on the ground surface is not well posed. That is, the boundary<br />

condition in terms of ground shaking is not completely given, or in other words we only have finite records at<br />

discrete points (our seismic stations). <strong>The</strong> solution of this kind of ill-posed problems is not unique, or there are<br />

many solutions that pretty much give the same results but none of them is exact (e.g. Yeih et al, 1993).<br />

Mechanism based on Strong Ground Motion<br />

A simple way to improve the situation is to use the data of acceleration time history recorded locally to fine tune<br />

the back analysis. <strong>The</strong>y are much more sensitive to the slip distribution and maximum slip. For example,<br />

Figure 3 shows the NS-component of the acceleration time histories along the coastline from north to south.<br />

<strong>The</strong> stations from top to bottom are the IWTH08, IWT016, IWT009, MYG003, FKS004, FKS013 and CHB012.<br />

Clearly, there are two distinct phases of ground shaking, suggesting two areas of large slips (NIED, 2011). In<br />

fact, similar observation was made during the 2008 Wenchuan earthquake (Wen et al., 2010). In the north, the<br />

first phase dominates, while this phase is invisible in the south. In the middle region, both phases are clearly<br />

visible. This suggests a very complicated slip process (both temporal and spatial). Actually, the slip<br />

distribution shown in Figure 2(b) is obtained by using the near far acceleration, and clearly differs significantly<br />

from other slip distributions shown in Figure 2. However, there is no comparison of the simulated ground<br />

motions compared to those given in Figure 3, thus the reliability of this slip distribution is still questionable.<br />

<strong>The</strong> validity of slip distribution shown in Figure 2(b) needs to be further demonstrated by using this slip<br />

distribution to generate tsunami to compare to that observed along the coastline. Both time of arrival and the<br />

tsunami run-up should be used to further validate the proposed slip mechanism given in Figure 2(b). So far as I<br />

know, such analysis had not been done.<br />

Figure 3 Focal mechanism in terms of slip distribution given by (a) USGS, (b) NIED, (c) Caltech, (d) Nagoya<br />

<strong>University</strong>, (e) Tsukuba <strong>University</strong>, (f) GFZ (modified from NIED, 2011)<br />

-357-


INADEQUACY OF TSUNAMI MITIGATION STRATEGIES<br />

Tsunami Barriers<br />

Although the 2011 Tohoku tsunami is not the first tsunami occurred in this region, the tsunami hazard along the<br />

coastline of Iwate and Miyagi Prefectures is clearly underestimated and inadequate mitigation strategies had<br />

been employed. According to the records compiled by the Japan Meteorological Agency, tsunami induced by<br />

submarine earthquake in the “Off Miyagi” and “Off Sanriku” sources (source zone of the 2011 Tohoku<br />

earthquake) had caused fatality of 1,000 in 869 (M~8.3), 50 in 1,611 (M~ 8.1), 44 in 1973 (M~8.2), 27,122 in<br />

1861 (M~7.4), 3,000 in 1915 (M~7.5), and 28 in 1936 (M~7.4). In 1896, a M ~7.6 earthquake had created<br />

tsunami waves as high as 38 m, killing 22,000. In 1933, a massive earthquake of M ~ 8.6 produced tsunami<br />

waves as high as 29 m on the Sanriku coast and caused more than 3,000 deaths. Yet, the countermeasures<br />

against tsunami are clearly not enough. For example, all tsunami barriers built along the coastline of Miyagi<br />

and Iwate Prefecture were not high enough to stop the tsunami surges and thus actually provided a false sense of<br />

security along the coastal cities. For example, Figure 4 shows the tsunami barrier built along the coastline of<br />

Miyako City which is clearly not high enough in stopping the tsunami. Another example is the tsunami barrier<br />

shown on the left of Figure 5 at Taro, Miyagi Prefecture. <strong>The</strong> devastation in the Taro City after the tsunami is<br />

shown on the right of Figure 5.<br />

Figure 4 Over flowing of the tsunami surges over the barrier wall at Miyako City.<br />

Figure 5 Tsunami barrier at Taro and devastation at the city after the 2011 Tohoku tsunami<br />

-358-


Elevated Structures for Evacuation<br />

<strong>The</strong> 2011 Tohoku tsunami taught us that tsunami barrier shown in Figures 4-5 are not effective in mitigating<br />

tsunami hazard. An alternative is to use elevated structure along the coastline that can withstand the impact of<br />

tsunami surges (including floating debris or ships). Such elevated platform can be used for vertical evacuation.<br />

<strong>The</strong> one shown in Figure 6 is used as a scenic vista point for tourists at Okushiri Island, Japan, which was<br />

impacted by a huge tsunami on July 12, 1993 and led to 239 deaths. Another example is given in Figure 7<br />

which is an elevated RC structure for vertical evacuation during storm surges in Bangladesh. In fact, a<br />

guidebook was published by FEMA ⎯FEMA P646 “Guidelines for Design of Structures for vertical Evacuation<br />

from Tsunamis” ⎯after the 2004 Indian Ocean Tsunami. <strong>The</strong>refore, vertical evacuation may provide a potential<br />

solution for tsunami mitigation. But, after all we have to know the tsunami run-up under the worst scenario.<br />

Figure 8 shows a car on the top of a 3-story building at Minamisanriku town, Miyagi Prefecture. Probably, we<br />

need to build vertical structures taller than those shown in Figures 6 and 7 for the 2011 Tohoku Tsunami.<br />

Figure 6 Elevated platform used for tsunami evacuation and for scenic vista point at Okushiri Island, Japan.<br />

CONCLUSIONS<br />

Figure 7 Elevated structures in Bangladesh for storm surge evacuation (after Emanuel, 2005)<br />

This paper discusses three important issues that raised by the 2011 Tohoku Earthquake and Tsunami. <strong>The</strong>y are<br />

the inadequacy of our seismic hazard estimation technique, the uncertainty in focal mechanism estimation, and<br />

inadequacy of tsunami mitigation strategies. More specifically, we propose to revolutionize the well-adopted<br />

seismic hazard technique proposed by Cornell (1968) and to drop the idea of based on past earthquake records.<br />

<strong>The</strong> dynamic earth is clearly a vibrant system that change is the only constant, and we should not blindly<br />

follows the over-simplified so-called “elastic rebound theory” proposed by Reid. History has repeatedly<br />

-359-


demonstrated that huge earthquake can happen at places that we believe are earthquake free. <strong>The</strong> customary<br />

wisdom of back-analyzing slip mechanism by using far field velocity time histories is challenged. It is argued<br />

that local strong ground motion needs to be used to study the focal mechanism of each earthquake. In case of<br />

earthquake-induced-tusnami, tsunami simulation should also be used to confirm the slip distribution. Finally,<br />

tsunami barrier built along the coastline of Miyagi and Iwate Prefectures are argued inadequate. <strong>The</strong> use of<br />

vertical evacuation structure can be considered, provided that a reliable of tsunami run-up can be estimated.<br />

ACKNOWLEDGMENTS<br />

<strong>The</strong> author gratefully acknowledge the financial support provided by the Research Grants Council of <strong>Hong</strong><br />

<strong>Kong</strong> Special Administrative Region, China (Project No: PolyU 5002/08P), and <strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong><br />

<strong>University</strong> through Project No. 1-BBZF.<br />

Figure 8 <strong>The</strong> car was carried to the top of a 3-story building at Minamisanriku town, Miyagi Prefecture<br />

-360-


REFERENCES<br />

Cornell, C.A. (1968). “Engineering Seismic Analysis”, Bulletin of the Seismological Society of America, 58,<br />

1583-1606.<br />

Chau, K.T. (2009). “Some geohazards associated with the 8.0 Wenchuan Earthquake on May 12, 2008",<br />

Keynote Lecture, International Symposium on Prediction and Simulation Methods for Geohazard<br />

Mitigation, May 25-27, 2009, Kyoto, Japan. In Prediction and Simulation Methods for Geohazard<br />

Mitigation, ed. Oka F., Murakami A, and Kimoto S. pp 65-72, CRC Press, Boca Raton.<br />

Chau, K.T. (2010a). “How reliable is our existing ‘seismic hazard analysis’?”, U.S.-Taiwan Workshop on the<br />

Advancement of Societal Responses to Mega-Disasters afflicting Mega-Cities (US-Taiwan MC/MD<br />

Workshop), Taipei, Taiwan, May 6-8, 2010 (abstract), pp. 10-11.<br />

Chau, K.T. (2010b). “Surface Ruptures of the 2008 Wenchuan Earthquake and 2010 Qinghai Yushu Earthquake”,<br />

<strong>The</strong> first young scholars across the Taiwan Strait Earthquake Engineering Symposium ( 第 一 屆 地 震 工 程 海<br />

峽 兩 岸 青 年 學 者 研 討 會 ), Taipei, Taiwan, China, October 19-24, 2010 (full paper in CD-ROM).<br />

Chau, K.T., Chen, Y. and Wong, T.F. (2000). Earthquake Mechanics and Seismic Hazard Analysis. Course notes<br />

of a 3-day short course conducted at <strong>Hong</strong> <strong>Kong</strong> <strong>Polytechnic</strong> <strong>University</strong>.<br />

Emanuel, K. (2005). Divine Wind: <strong>The</strong> History and Science of Hurricanes, Oxford <strong>University</strong> Press, USA.<br />

ERC (2010). Map Guide, National Seismic Hazard Maps for 2010, Headquarters for Earthquake Research<br />

Promotion, http://www.jishin.go.jp/main/chousa/10_yosokuchizu/index.htm<br />

Kanamori, H. (1995). “<strong>The</strong> Kobe (Hyogo-ken Nanbu), Japan, Earthquake of January 16, 1995”, Seismological<br />

Research Letters, 66(2), 6-10.<br />

Kanamori, H., Miyazawa, M. and Mori, J. (2006). “Investigation of the earthquake sequence off Miyagi<br />

prefecture with historical seismograms”, Earth Planets Space, 58, 1533-1541.<br />

NIED (National Research Institute for Earth Science and Disaster Prevention), (2011). 2011 Off the Pacific<br />

Coast of Tohoku Earthquake, Strong Ground Motion, Emergency Meeting of Headquarters for Earthquake<br />

Research Promotion. http://www.k-net.bosai.go.jp/k-net/topics/TohokuTaiheiyo_20110311/nied_kyoshin2e.<br />

pdf.<br />

Reid, H.F. (1910). <strong>The</strong> Mechanics of the Earthquake, <strong>The</strong> California Earthquake of April 18, 1906, Report of the<br />

State Investigation Commission, Vol.2, Carnegie Institution of Washington, Washington, D.C. 1910.<br />

Wen, Z.P., Xie, J.J., Gao, M.T., Hu, Y.X. and Chau K.T. (2010). “Near-source strong ground motion<br />

characteristics of the 2008 Wenchuan Earthquake”, Bulletin of the Seismological Society of America,<br />

100(B5), 2425-2439.<br />

Yeih, W. Koya, T. and Mura, T. (1993). “An inverse problem in elasticity with partially overprescribed boundary<br />

conditions, Part I: <strong>The</strong>oretical approach”, Journal of Applied Mechanics ASME, 60, 595-600.<br />

Zhou, R., Wen X., Chai, C. and Ma, S. (1997). “Recent earthquakes and assessment of seismic tendency on the<br />

Ganzi - Yushu fault zone”, Seismology and Geology, 19(2), 115-124.<br />

-361-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

同 一 建 筑 刚 性 桩 与 柔 性 桩 基 础 沉 降 差 过 大 事 故 处 理<br />

1,<br />

张 忠 苗<br />

2 , 张 乾 青<br />

(1. 浙 江 大 学 软 弱 土 与 环 境 土 工 教 育 部 重 点 实 验 室 , 浙 江 杭 州 310058;<br />

2. 浙 江 大 学 岩 土 工 程 研 究 所 , 浙 江 杭 州 310058)<br />

1, 2<br />

摘 要 : 本 文 以 同 一 建 筑 刚 性 桩 与 柔 性 桩 基 础 沉 降 差 过 大 的 事 故 为 例 , 详 细 分 析 了 单 层 粮 库 刚 柔<br />

两 种 桩 基 础 沉 降 差 过 大 事 故 的 原 因 , 并 提 出 了 具 体 加 固 方 案 。 分 析 表 明 , 对 位 于 淤 泥 层 较 厚 场 地 的<br />

单 层 粮 库 来 说 , 不 能 采 用 刚 性 桩 位 于 较 好 持 力 层 而 柔 性 桩 位 于 淤 泥 层 的 设 计 方 法 。 对 于 同 一 建 筑 刚<br />

性 桩 与 柔 性 桩 基 础 沉 降 差 过 大 的 事 故 , 可 采 用 钻 孔 灌 注 桩 加 梁 板 筏 式 基 础 的 加 固 方 案 。 这 种 加 固 方<br />

案 可 显 著 减 小 粮 库 地 面 沉 降 和 粮 库 主 体 结 构 部 分 与 粮 库 地 面 沉 降 差 。<br />

关 键 词 : 刚 性 桩 柔 性 桩 沉 降 差 加 固<br />

ACCIDENT TREATMENT ON LARGE DIFFERENTIAL SETTLEMENT BETWEEN<br />

RIGID AND FLEXIBLE PILE FOUNDATION IN THE SAME BUILDING<br />

Zhongmiao Zhang 1, 2 and Qianqing Zhang 1, 2<br />

1 MOE Key Laboratory of Soft Soils and Geoenvironmental Engineering, Zhejiang <strong>University</strong>, Hangzhou<br />

310058, China; 2 Institute of Geotechnical Engineering, Zhejiang <strong>University</strong>, Hangzhou 310058, China<br />

Abstract: This paper analyzes an accident which is caused by large differential settlement between rigid and<br />

flexible pile foundation in a grain bin, and proposes a detailed reinforcement scheme. <strong>The</strong> analyses show that in<br />

a site with a large depth of silt, the rigid and flexible pile should be embedded into a stiff bearing stratum with<br />

the same depth, and the rigid and flexible piles are not supposed to be drilled into stiff bearing stratum and silt,<br />

respectively. It is concluded that bored pile and beam-slab raft foundation can be used to treat the accident of<br />

large differential settlement between rigid and flexible pile foundation. This reinforcement scheme can<br />

significantly decrease the settlement of ground and the differential settlement between ground and main body<br />

structure of grain bin.<br />

Keywords: Rigid pile, flexible pile, differential settlement, solidification.<br />

一 、 前 言<br />

随 着 经 济 的 发 展 和 社 会 的 需 求 , 高 层 和 超 高 层 建 筑 不 断 涌 现 。 目 前 , 我 国 拥 有 高 492 m 的 上 海<br />

环 球 金 融 中 心 , 高 565.6 m 的 上 海 中 心 大 厦 和 高 约 600 m 的 天 津 117 大 厦 正 在 建 造 中 , 我 国 土 木 工<br />

作 者 积 累 了 高 层 和 超 高 层 设 计 和 建 设 的 成 功 经 验 。 然 而 , 低 层 建 筑 、 单 层 厂 房 、 粮 食 仓 库 等 建 筑 因<br />

其 上 部 荷 载 较 小 , 结 构 简 单 往 往 不 能 引 起 设 计 和 施 工 人 员 的 足 够 重 视 , 反 而 引 发 了 一 些 工 程 事 故 ,<br />

继 而 也 积 累 了 一 些 处 理 桩 基 事 故 的 成 功 经 验 [1-3] 。 本 文 选 取 同 一 建 筑 采 用 刚 柔 不 同 桩 基 础 而 引 起 两 种<br />

不 同 桩 型 沉 降 差 过 大 的 典 型 案 例 , 详 细 分 析 了 单 层 粮 库 刚 柔 两 种 桩 沉 降 不 协 调 的 原 因 , 并 提 出 了 具<br />

体 解 决 方 法 。 笔 者 希 望 此 工 程 案 例 能 引 起 广 大 土 木 工 作 者 的 注 意 , 在 设 计 和 施 工 过 程 中 避 免 此 类 事<br />

件 再 次 发 生 。<br />

基 金 项 目 : 国 家 自 然 科 学 基 金 资 助 项 目 (51078330)<br />

作 者 简 介 : 张 忠 苗 (1961-), 男 , 浙 江 宁 海 人 , 教 授 , 博 导 , 主 要 从 事 桩 基 础 、 基 础 工 程 和 地 质 工 程 的 教 学 与 研 究 工 作 。<br />

-362-


二 、 工 程 概 况 和 场 地 地 质 条 件<br />

温 州 某 粮 食 储 备 仓 库 共 有 6 幢 ,6 幢 粮 库 总 建 筑 面 积 为 3060 m 2 , 建 筑 占 地 面 积 为 3060 m 2 , 单<br />

幢 粮 库 的 建 筑 面 积 为 510 m 2 , 屋 面 为 人 字 梁 结 构 , 屋 架 部 分 的 有 效 高 度 为 6.8 m, 设 计 储 粮 高 度 为<br />

6.5 m, 实 际 储 粮 部 分 的 地 面 面 积 为 12.26 m×30.52 m, 单 幢 粮 库 设 计 总 仓 储 为 331 万 市 斤 , 室 内 地<br />

面 设 计 堆 粮 荷 载 约 为 35.75 kPa。 该 场 地 位 于 温 州 软 土 地 区 , 场 地 各 土 层 的 物 理 力 学 参 数 见 表 1。<br />

表 1 场 地 各 土 层 的 物 理 力 学 参 数<br />

土 层 名 称<br />

含 水<br />

压 缩 地 基 承 载 钻 孔 灌 注 桩<br />

厚 度<br />

重 度<br />

量<br />

模 量 力 特 征 值<br />

(m)<br />

(kN/m 3 )<br />

侧 阻 力 特 端 阻 力 特<br />

(%)<br />

(MPa) (kPa)<br />

征 值 (kPa) 征 值 (kPa)<br />

粉 质 粘 土 1.6-2.5 28.3 18.8 5 90 18<br />

淤 泥 未 打 穿 , 最 大 勘 探 厚 度 为 12.5- 9.4 68 15.2 1.24 50 5.5<br />

粘 土 夹 粉 砂 43.5* 17.3* 2.74* 80* 13*<br />

砾 砂 350* 42* 1500*<br />

需 要 说 明 的 是 , 勘 探 单 位 提 供 的 本 工 程 地 质 剖 面 图 显 示 , 在 6 个 粮 食 储 备 仓 库 场 地 上 勘 探 深 度<br />

都 不 足 ( 勘 探 孔 深 15.0m~20.2m), 没 有 一 个 勘 探 孔 打 穿 淤 泥 层 , 这 显 然 是 不 合 理 的 。 表 1 中 带 *<br />

的 参 数 也 是 相 邻 工 地 土 层 的 参 数 。<br />

粮 库 的 主 体 结 构 部 分 基 础 设 计 采 用 预 应 力 管 桩 双 桩 承 台 梁 式 基 础 , 预 应 力 管 桩 直 径 为 400(60)<br />

mm, 其 有 效 桩 长 约 为 26 m, 设 计 要 求 管 桩 单 桩 竖 向 抗 压 承 载 力 特 征 值 为 400 kN, 单 幢 粮 库 主 体 结<br />

构 部 分 共 布 置 预 应 力 管 桩 60 根 ( 具 体 布 置 方 式 可 参 见 5.2 节 中 的 图 6)。 粮 库 室 内 堆 粮 地 面 设 计 采<br />

用 水 泥 搅 拌 桩 进 行 加 固 , 水 泥 搅 拌 桩 设 计 桩 径 为 500 mm, 设 计 有 效 桩 长 ≥10 m, 水 泥 掺 入 量 为 15%。<br />

设 计 要 求 水 泥 搅 拌 桩 90 天 龄 期 的 无 侧 限 抗 压 强 度 为 1.5 MPa, 复 合 地 基 承 载 力 特 征 值 为 100 kPa。<br />

开 挖 结 果 显 示 , 实 际 施 工 过 程 中 , 水 泥 搅 拌 桩 顶 上 覆 盖 有 厚 约 1.65 m 的 杂 填 土 , 杂 填 土 上 为 厚 约<br />

10 cm 的 素 混 凝 土 层 , 素 混 凝 土 上 直 接 堆 放 粮 食 。 单 桩 粮 库 的 布 桩 型 式 和 场 地 地 质 条 件 如 图 1 所 示<br />

( 图 1 中 淤 泥 的 厚 度 是 参 照 相 邻 工 地 的 数 据 得 到 的 )。<br />

单 层 粮 库<br />

主 体 屋 架<br />

回 填 土<br />

粮 库 室 内 地 面<br />

±0.00<br />

-1.75<br />

粉 质 粘 土<br />

-3.86<br />

水 泥 搅 拌 桩<br />

设 计 桩 长 10m<br />

淤 泥<br />

预 应 力 管 桩<br />

设 计 桩 长 26m<br />

-26.70<br />

砾 砂<br />

图 1 单 幢 粮 库 的 桩 基 础<br />

-363-


三 、 粮 库 室 内 地 面 沉 降 观 测<br />

6 幢 仓 库 中 的 6# 粮 食 仓 库 为 空 仓 库 ,1~5# 粮 库 均 储 有 粮 食 。2010 年 1 月 份 储 粮 工 人 发 现 粮 库<br />

仓 储 地 面 ( 水 泥 搅 拌 桩 基 础 ) 储 粮 后 出 现 了 过 大 的 沉 降 ( 见 图 2), 最 大 约 为 40cm 且 沉 降 还 有 不 断<br />

增 大 的 趋 势 , 已 严 重 影 响 了 粮 库 的 正 常 使 用 。 但 采 用 预 应 力 管 桩 的 粮 库 主 体 结 构 部 分 基 础 沉 降 很 小 ,<br />

粮 库 的 外 立 面 基 本 完 好 , 采 用 刚 性 预 应 力 管 桩 基 础 的 主 体 结 构 与 采 用 柔 性 水 泥 桩 基 础 的 室 内 地 面 沉<br />

降 差 过 大 。 同 时 对 还 未 储 粮 的 6# 仓 库 现 场 观 测 发 现 ,6# 空 粮 库 室 内 地 面 出 现 了 不 同 程 度 的 沉 降 裂 缝 ,<br />

粮 库 地 面 混 凝 土 横 梁 被 拉 裂 , 见 图 3。<br />

主 体 结 构<br />

图 2 粮 库 储 粮 后 的 地 面 沉 降<br />

地 面 裂 缝<br />

粮 库 地 面<br />

横 梁 拉 裂<br />

图 3 未 储 粮 粮 库 的 地 面 沉 降<br />

为 弄 清 储 粮 地 面 的 具 体 沉 降 情 况 , 有 必 要 对 储 粮 地 面 进 行 现 场 实 测 。 为 配 合 测 量 , 建 设 单 位 对<br />

4# 粮 库 的 粮 食 进 行 了 腾 空 。 笔 者 利 用 水 准 仪 对 腾 空 后 的 4# 粮 食 仓 库 地 面 进 行 了 详 细 的 沉 降 观 测 。 观<br />

测 时 共 在 4# 粮 食 仓 库 地 面 上 设 置 沉 降 观 测 点 54 个 , 以 东 面 门 口 地 梁 作 为 基 准 点 来 测 试 各 个 沉 降 观<br />

测 点 的 相 对 沉 降 , 测 得 的 相 对 沉 降 结 果 见 图 4。<br />

-364-


北<br />

0 2.5 5 7.5 10 12.5<br />

0<br />

2.5<br />

5<br />

7.5<br />

10<br />

12.5<br />

15<br />

17.5<br />

20<br />

22.5<br />

25<br />

27.5<br />

30<br />

32.5<br />

图 4 4# 粮 库 卸 载 后 地 面 沉 降 等 值 线 图 ( 平 面 基 础 尺 寸 单 位 :m; 沉 降 数 据 单 位 :mm)<br />

从 图 4 可 以 看 出 ,4# 粮 库 粮 食 卸 载 后 地 面 沉 降 最 大 值 出 现 在 粮 库 的 中 心 位 置 , 沉 降 呈 “ 锅 底 形 ”。<br />

北 面 至 室 内 地 面 中 心 点 的 沉 降 差 约 为 350 mm, 其 倾 斜 率 约 为 2.15%; 西 面 至 室 内 地 面 中 心 点 的 沉 降<br />

差 约 为 360 mm, 其 倾 斜 率 约 为 5.05 %。 测 试 结 果 表 明 4# 粮 库 的 累 计 中 心 沉 降 量 和 倾 斜 值 均 超 过 了<br />

规 范 要 求 [4] 。 需 要 说 明 的 是 , 图 4 中 的 沉 降 数 据 是 粮 食 卸 载 后 ( 地 面 会 产 生 回 弹 ) 的 实 测 沉 降 数 据 ,<br />

粮 食 堆 满 时 的 室 内 地 面 实 际 沉 降 值 肯 定 要 比 图 4 中 数 据 大 。1#~5# 粮 库 储 粮 后 室 内 地 面 沉 降 情 况 也<br />

与 4# 粮 库 类 似 , 同 时 6# 空 粮 库 室 内 地 面 也 发 现 了 自 重 固 结 引 起 的 沉 降 裂 缝 , 这 说 明 6 幢 粮 库 均 属<br />

于 建 筑 桩 基 事 故 , 且 这 6 幢 粮 食 仓 库 均 不 能 正 常 使 用 , 如 果 继 续 使 用 有 可 能 会 出 现 储 粮 地 面 整 体 滑<br />

移 破 坏 而 导 致 整 幢 粮 库 倾 覆 , 必 须 采 取 有 效 措 施 进 行 加 固 处 理 。<br />

四 、 两 种 桩 基 础 沉 降 差 过 大 原 因 分 析<br />

为 弄 清 该 粮 库 刚 性 桩 与 柔 性 桩 基 础 沉 降 差 过 大 的 原 因 , 有 必 要 从 多 个 方 面 进 行 详 细 分 析 。<br />

(A) 经 水 泥 搅 拌 桩 处 理 后 的 粮 库 室 内 地 面 沉 降 过 大 , 有 必 要 对 水 泥 搅 拌 桩 的 桩 长 和 桩 身 质 量 进<br />

行 检 测 。 为 此 , 甲 方 委 托 第 三 方 对 6# 空 粮 库 地 面 下 的 7 根 水 泥 搅 拌 桩 进 行 了 钻 探 取 芯 检 验 。 钻 探 取<br />

芯 检 测 结 果 表 明 , 水 泥 搅 拌 桩 桩 长 的 严 重 不 足 , 水 泥 土 强 度 的 参 差 不 齐 。7 根 水 泥 搅 拌 桩 实 际 桩 长<br />

只 有 3.2~4.7 m, 均 未 达 到 设 计 要 求 的 10 m 有 效 桩 长 。 水 泥 土 芯 样 的 抗 压 强 度 为 0.51~7.63 MPa, 实<br />

测 强 度 差 异 很 大 , 且 有 两 根 桩 的 芯 样 无 侧 限 抗 压 强 度 未 达 到 1.5 MPa。<br />

(B) 甲 方 委 托 第 三 方 对 6# 空 粮 库 地 面 3 个 不 同 的 地 点 进 行 了 复 合 地 基 平 板 载 荷 试 验 。 试 验 结 果<br />

显 示 ,2 个 地 点 的 复 合 地 基 承 载 力 特 征 值 约 为 110 kPa, 满 足 复 合 地 基 承 载 力 特 征 值 为 100 kPa 的 设<br />

计 要 求 。1 个 地 点 的 复 合 地 基 承 载 力 特 征 值 约 为 90 kPa, 略 低 与 复 合 地 基 承 载 力 特 征 值 为 100 kPa<br />

的 设 计 要 求 。<br />

(C) 笔 者 根 据 浙 江 省 标 准 建 筑 地 基 基 础 设 计 规 范 (DB 33/1001-2003) [5] 对 陶 山 粮 库 地 面 沉 降 进<br />

行 了 理 论 计 算 。 计 算 中 取 设 计 地 面 堆 粮 荷 载 为 35.75 kPa, 搅 拌 桩 上 的 回 填 土 和 素 混 凝 土 垫 层 厚 度 按<br />

照 6# 粮 库 的 开 挖 结 果 取 为 1.75 m, 即 回 填 土 的 荷 载 约 为 35.0 kPa。 计 算 分 为 4 种 情 况 , 取 水 泥 搅 拌<br />

桩 总 桩 数 为 226 根 ( 实 际 施 工 水 泥 搅 拌 桩 桩 数 ) 和 244 根 ( 设 计 水 泥 搅 拌 桩 桩 数 ) 分 别 计 算 。 同 时<br />

沉 降 计 算 时 又 分 设 计 10 m 长 水 泥 搅 拌 桩 和 实 际 约 4 m 水 泥 搅 拌 桩 ( 根 据 钻 探 取 芯 结 果 取 值 ) 两 种<br />

情 况 分 别 进 行 计 算 , 计 算 结 果 见 表 3。 需 要 说 明 的 是 , 表 2 中 回 填 土 计 算 时 考 虑 了 设 计 地 面 堆 粮 荷<br />

载 引 起 的 压 缩 沉 降 但 没 有 考 虑 回 填 土 对 水 泥 搅 拌 桩 的 负 摩 阻 力 和 回 填 土 自 重 作 用 下 的 固 结 沉 降 ( 回<br />

填 土 参 数 不 足 )。 由 于 6 幢 粮 库 场 地 勘 探 孔 勘 探 深 度 不 足 ( 最 大 勘 探 深 度 约 20 m), 没 有 20 m 以 下<br />

-365-


土 层 的 物 理 力 学 参 数 , 故 计 算 水 泥 搅 拌 桩 桩 端 下 沉 降 时 只 考 虑 现 有 勘 探 深 度 范 围 内 桩 端 下 淤 泥 层 的<br />

压 缩 量 。 其 桩 端 沉 降 计 算 结 果 会 比 按 照 相 关 规 范 确 定 的 桩 端 下 压 缩 层 范 围 内 压 缩 量 的 计 算 结 果 要<br />

小 。<br />

表 2 不 同 情 况 下 的 室 内 地 面 沉 降 计 算 值<br />

基 础 型 式<br />

荷 载 (kPa)<br />

工<br />

况<br />

总 桩 数<br />

( 根 )<br />

地 面 回 填<br />

土 沉 降<br />

理 论 最 大 沉 降 值 (mm)<br />

搅 拌 桩 复 合 土 桩 端 下 卧 层<br />

层 的 沉 降 沉 降<br />

累 计 沉<br />

降<br />

10 m 长 水 泥 搅<br />

拌 桩 复 合 地 基<br />

4 m 长 水 泥 搅<br />

拌 桩 复 合 地 基<br />

设 计 地 面 堆 载 + 回 填 土 荷 工 况 1 226 19.7 44.7 237.5 301.9<br />

载 (35.75+35) kPa 工 况 2 244 19.7 41.8 237.5 299.0<br />

设 计 地 面 堆 载 + 回 填 土 荷 工 况 3 226 19.7 17.0 338.7 375.4<br />

载 (35.75+35) kPa 工 况 4 244 19.7 16.0 338.7 374.4<br />

从 表 2 中 可 以 看 出 :<br />

(1) 在 堆 粮 和 回 填 土 荷 载 作 用 下 , 设 计 10 m 长 水 泥 搅 拌 桩 室 内 地 面 理 论 沉 降 计 算 值 约 为 299.0<br />

mm( 设 计 244 根 水 泥 搅 拌 桩 ) 和 301.9 mm ( 实 际 施 工 226 根 水 泥 搅 拌 桩 ), 因 水 泥 搅 拌 桩 的 下 卧<br />

层 为 淤 泥 , 其 压 缩 量 大 。<br />

(2) 在 堆 粮 和 回 填 土 荷 载 作 用 下 ,4 m 长 水 泥 搅 拌 桩 室 内 地 面 理 论 沉 降 计 算 值 达 374.4 mm( 设<br />

计 244 根 水 泥 搅 拌 桩 ) 和 375.4 mm( 实 际 施 工 226 根 水 泥 搅 拌 桩 ), 这 说 明 桩 长 缩 短 后 更 严 重 地 加<br />

大 了 室 内 地 面 的 累 计 沉 降 量 。 同 时 , 由 于 实 际 水 泥 搅 拌 桩 取 芯 结 果 桩 长 为 3.2~4.7 m 不 等 , 且 桩 身<br />

强 度 为 0.51~7.63 MPa 参 差 不 齐 , 更 加 严 重 地 导 致 了 粮 库 室 内 地 面 的 不 均 匀 沉 降 。<br />

(3) 表 2 中 的 计 算 结 果 只 是 按 浙 江 省 标 准 建 筑 地 基 基 础 设 计 规 范 (DB 33/1001-2003) 进 行 的<br />

理 论 沉 降 计 算 结 果 且 粮 库 桩 端 下 土 的 压 缩 变 形 计 算 时 只 考 虑 勘 探 深 度 内 淤 泥 层 的 压 缩 , 它 与 实 际 沉<br />

降 会 有 一 定 的 误 差 。 粮 库 室 内 地 面 理 论 沉 降 计 算 值 只 能 作 为 分 析 参 考 。<br />

(4) 粮 库 主 体 结 构 采 用 26 m 的 刚 性 预 应 力 管 桩 桩 基 础 沉 降 很 小 , 现 场 观 测 沉 降 不 足 1 mm。<br />

因 此 , 表 2 中 的 沉 降 计 算 数 据 可 近 似 看 成 是 同 一 粮 库 刚 性 预 应 力 管 桩 与 柔 性 水 泥 搅 拌 桩 沉 降 差 。<br />

(D) 本 工 程 粮 库 室 内 地 面 过 大 沉 降 的 原 因 总 结 如 下 :<br />

(1) 由 于 本 工 程 粮 库 所 在 场 地 具 有 深 厚 淤 泥 层 , 设 计 粮 库 室 内 地 面 下 水 泥 搅 拌 桩 的 有 效 桩 长<br />

只 有 10 m, 柔 性 搅 拌 桩 桩 长 不 足 导 致 下 卧 淤 泥 层 的 压 缩 沉 降 过 大 ; 而 粮 库 主 体 结 构 采 用 26 m 的 刚<br />

性 预 应 力 管 桩 桩 基 础 沉 降 很 小 ( 现 场 观 测 沉 降 不 足 1 mm), 导 致 同 一 粮 库 主 体 结 构 刚 性 基 础 与 室 内<br />

仓 储 地 面 柔 性 基 础 的 差 异 沉 降 过 大 。 这 种 设 计 方 法 欠 合 理 。<br />

(2) 施 工 单 位 偷 工 减 料 , 水 泥 搅 拌 桩 实 际 桩 长 远 小 于 设 计 有 效 桩 长 且 强 度 参 差 不 齐 , 是 造 成<br />

粮 库 室 内 地 面 沉 降 过 大 的 一 个 重 要 原 因 。<br />

(3) 粮 库 水 泥 搅 拌 桩 顶 上 实 际 回 填 有 厚 约 1.65 m 的 回 填 土 及 厚 约 0.1 m 的 素 混 凝 土 垫 层 。 现<br />

粮 库 中 的 粮 食 直 接 堆 放 在 这 层 素 混 凝 土 垫 层 面 上 ( 原 设 计 先 分 级 堆 载 预 压 , 待 沉 降 稳 定 后 再 做 永 久<br />

性 粮 库 地 面 )。 回 填 土 不 仅 加 大 了 地 面 荷 载 , 而 且 回 填 不 密 实 带 来 了 自 身 的 固 结 沉 降 , 同 时 也 给 水<br />

泥 搅 拌 桩 带 来 了 负 摩 阻 力 , 这 都 会 加 大 粮 库 室 内 地 面 的 沉 降 。<br />

(4) 设 计 要 求 地 面 仓 储 粮 食 堆 载 要 分 级 堆 载 预 压 , 而 在 使 用 过 程 中 采 用 一 次 性 堆 载 , 造 成 沉<br />

降 速 率 加 快 。<br />

五 两 种 桩 基 础 沉 降 差 过 大 事 故 加 固 方 法<br />

本 工 程 加 固 方 案 设 计 中 要 考 虑 以 下 几 个 方 面 :<br />

(1) 主 体 结 构 的 室 内 净 高 为 6.8 m, 加 固 方 案 设 计 中 要 考 虑 施 工 机 械 的 操 作 高 度 。<br />

(2) 粮 库 现 室 内 仓 储 地 面 约 有 0.1 m 厚 的 素 砼 和 1.65 m 厚 的 回 填 土 , 回 填 土 中 有 大 小 不 等 的<br />

-366-


石 块 和 泥 土 , 要 考 虑 施 工 机 械 的 施 工 可 行 性 。<br />

(3) 加 固 设 计 中 要 考 虑 今 后 粮 库 仓 储 地 面 加 固 后 的 沉 降 应 与 已 建 粮 库 主 体 结 构 基 础 的 沉 降 协<br />

调 问 题 。 观 测 表 明 , 采 用 长 26 m 预 应 力 管 桩 的 主 体 结 构 部 分 基 础 沉 降 很 小 且 外 立 面 基 本 完 好 , 粮<br />

库 主 体 结 构 部 分 现 在 是 基 本 安 全 的 , 不 需 要 加 固 ( 只 需 要 局 部 内 墙 裂 缝 修 补 )。 所 以 , 主 要 加 固 室<br />

内 仓 储 地 面 的 过 大 沉 降 , 为 了 变 形 协 调 , 加 固 也 采 用 刚 性 桩 方 案 较 好 。<br />

(4) 由 于 粮 食 仓 库 勘 探 孔 深 度 不 够 , 所 以 施 工 图 加 固 设 计 前 必 须 进 行 补 充 勘 探 , 勘 探 孔 的 深<br />

度 要 钻 至 好 的 土 层 。<br />

(A) 加 固 方 案 的 种 类 及 选 择<br />

针 对 上 述 情 况 粮 库 室 内 地 面 加 固 方 案 无 疑 都 要 采 用 桩 基 础 , 具 体 有 以 下 几 个 方 案 可 供 选 择 :<br />

(1) 补 打 旋 喷 桩 的 加 固 方 案 。 优 点 是 成 本 略 低 , 但 缺 点 是 回 填 土 要 挖 除 或 部 分 挖 除 。 本 场 地<br />

有 厚 约 20 m 的 淤 泥 土 层 , 若 施 工 桩 长 不 够 还 是 会 带 来 仓 储 地 面 的 过 大 沉 降 , 同 时 由 于 旋 喷 桩 属 柔<br />

性 桩 , 所 以 采 用 该 方 案 加 固 , 仓 储 地 面 沉 降 与 主 体 结 构 基 础 的 沉 降 协 调 仍 无 法 保 证 。<br />

(2) 树 根 桩 的 加 固 方 案 。 在 本 场 地 厚 约 20 m 的 淤 泥 层 中 施 工 树 根 桩 时 施 工 质 量 无 法 保 证 , 若<br />

桩 身 质 量 存 在 问 题 还 是 会 带 来 室 内 仓 储 地 面 的 过 大 沉 降 。<br />

(3) 采 用 直 径 500 mm 的 钻 孔 灌 注 桩 ( 桩 长 约 为 27 m, 持 力 层 为 砾 砂 , 桩 顶 标 高 约 为 -0.70 m)<br />

加 梁 板 筏 式 基 础 的 加 固 方 案 。 施 工 可 采 用 地 质 勘 察 的 小 钻 机 成 孔 或 简 易 取 土 钻 成 孔 , 该 方 案 可 以 保<br />

证 桩 穿 越 回 填 土 层 和 深 厚 淤 泥 土 层 且 成 桩 质 量 基 本 有 保 证 , 同 时 钻 孔 灌 注 桩 和 预 应 力 管 桩 同 是 刚 性<br />

桩 且 打 到 同 一 持 力 层 , 这 样 储 粮 仓 库 地 面 基 础 与 粮 库 主 体 结 构 部 分 的 基 础 沉 降 都 较 小 且 能 协 调 , 是<br />

较 合 理 的 加 固 设 计 方 案 。<br />

(B) 钻 孔 灌 注 桩 加 固 方 案 中 的 平 面 布 桩 方 式<br />

原 设 计 中 的<br />

屋 架 双 桩 承 台<br />

补 桩 有 效 桩 长 27 m<br />

架 空 层 楼 盖 板 厚 250mm<br />

图 5 单 幢 粮 库 地 面 加 固 方 案 ( 单 位 :mm)<br />

加 固 方 案 采 用 钻 孔 灌 注 桩 加 梁 板 的 基 础 型 式 。 加 固 设 计 时 单 幢 粮 库 室 内 地 面 堆 粮 的 设 计 荷 载 按<br />

原 设 计 取 为 35.75 kPa, 堆 粮 的 总 荷 载 331 万 市 斤 即 1665 吨 , 加 固 梁 板 自 重 荷 载 约 为 398 吨 , 即 总<br />

荷 载 约 为 2061 吨 。 由 于 架 空 梁 板 做 在 回 填 土 上 , 所 以 不 需 要 考 虑 回 填 土 荷 载 。 加 固 设 计 钻 孔 桩 单<br />

桩 抗 压 承 载 力 特 征 值 取 为 400 kN, 考 虑 到 室 内 粮 库 储 粮 过 程 中 人 员 和 运 输 机 械 等 荷 载 作 用 以 及 堆 粮<br />

局 部 偏 载 作 用 ( 考 虑 偏 心 荷 载 作 用 时 , 桩 的 数 量 约 增 加 20% [6] ), 单 幢 粮 食 仓 库 室 内 地 面 加 固 需 要<br />

布 置 的 平 面 总 桩 数 取 为 65 根 。 桩 位 布 置 图 见 图 5, 承 台 梁 板 布 置 剖 面 见 图 6。 需 要 注 意 的 是 , 根 据<br />

-367-


笔 者 对 4# 粮 库 的 沉 降 观 测 资 料 , 发 现 仓 储 地 面 沉 降 规 律 是 中 间 大 四 周 小 。 因 此 , 平 面 布 桩 的 方 式 应<br />

为 粮 库 地 面 中 间 桩 稍 密 一 些 , 外 围 略 稀 一 些 。 同 时 钻 孔 灌 注 桩 布 桩 时 要 避 开 粮 库 原 主 体 结 构 承 台 的<br />

预 应 力 管 桩 。<br />

±0.00<br />

-0.70m<br />

-1.25m<br />

原 有 承 台<br />

直 径 500mm<br />

的 钻 孔 桩 ,<br />

桩 长 约 27 m<br />

现 浇 板<br />

现 浇 板<br />

原 有 承 台<br />

图 6 地 面 加 固 方 案 剖 面 (1-1 剖 面 ; 单 位 :mm)<br />

六 、 结 论<br />

本 文 以 同 一 建 筑 刚 性 桩 与 柔 性 桩 基 础 沉 降 差 过 大 的 事 故 为 例 , 详 细 分 析 了 单 层 厂 房 刚 柔 两 种 桩<br />

基 础 沉 降 差 过 大 事 故 的 原 因 , 并 提 出 了 具 体 加 固 方 案 。 分 析 表 明 :<br />

(1) 对 于 淤 泥 层 厚 度 较 厚 的 场 地 , 采 用 刚 柔 复 合 桩 基 时 , 要 特 别 注 意 刚 性 桩 与 柔 性 桩 变 形 协 调<br />

的 问 题 。 笔 者 认 为 在 此 类 场 地 中 采 用 刚 柔 桩 复 合 桩 基 时 两 种 桩 型 要 选 择 同 一 较 好 的 持 力 层 , 以 使 得<br />

两 种 桩 型 沉 降 协 调 。 实 际 工 程 中 可 采 用 变 桩 径 、 变 桩 距 的 设 计 方 法 , 不 宜 采 用 变 桩 长 的 设 计 思 想 。<br />

(2) 本 工 程 中 水 泥 搅 拌 桩 实 际 桩 长 远 小 于 设 计 有 效 桩 长 且 强 度 参 差 不 齐 , 是 造 成 粮 库 室 内 地 面<br />

沉 降 过 大 的 一 个 重 要 原 因 。 同 时 粮 库 水 泥 搅 拌 桩 顶 上 回 填 土 较 厚 , 回 填 土 不 仅 加 大 了 地 面 荷 载 , 而<br />

且 回 填 不 密 实 带 来 了 自 身 的 固 结 沉 降 , 同 时 也 给 水 泥 搅 拌 桩 带 来 了 负 摩 阻 力 , 这 都 会 加 大 粮 库 室 内<br />

地 面 的 沉 降 。 地 面 仓 储 粮 食 堆 载 没 有 按 设 计 要 求 分 级 堆 载 预 压 , 而 是 一 次 性 堆 载 , 造 成 沉 降 速 率 加<br />

快 。<br />

(3) 对 于 该 工 程 中 同 一 建 筑 刚 性 桩 与 柔 性 桩 基 础 沉 降 差 过 大 的 事 故 , 可 采 用 钻 孔 灌 注 桩 加 梁 板<br />

筏 式 基 础 的 加 固 方 案 。 该 方 案 这 种 加 固 方 案 可 保 证 粮 库 地 面 沉 降 较 小 且 粮 库 主 体 结 构 部 分 与 粮 库 地<br />

面 沉 降 差 较 小 。<br />

参 考 文 献<br />

[1] 张 忠 苗 . 桩 基 工 程 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 2007.<br />

[2] 周 青 春 , 李 海 波 , 刘 亚 群 . 用 换 填 法 及 刚 性 桩 复 合 地 基 基 础 处 理 桩 基 质 量 事 故 . 岩 土 力 学 , 2003, 24(5): 845-848.<br />

[3] 冯 立 明 . 锚 杆 静 压 桩 技 术 在 某 住 宅 小 区 2^# 楼 桩 基 事 故 中 的 应 用 . 安 徽 建 筑 , 2006, 13(4): 125-126, 130.<br />

[4] 中 华 人 民 共 和 国 建 设 部 . 建 筑 地 基 基 础 设 计 规 范 (GB 50007-2002), 北 京 : 中 国 建 筑 工 业 出 版 社 , 2002.<br />

[5] 浙 江 省 建 设 厅 . 建 筑 地 基 基 础 设 计 规 范 (DB 33/1001-2003), 杭 州 : 浙 江 大 学 出 版 社 , 2003.<br />

[6] 中 华 人 民 共 和 国 建 设 部 . 建 筑 桩 基 技 术 规 范 (GBJ94-2008), 北 京 : 中 国 建 筑 工 业 出 版 社 , 2008.<br />

-368-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

关 于 国 标 中 主 动 土 压 力 系 数 公 式 的 讨 论<br />

陆 新 , 王 永 甫 , 郑 颖 人<br />

( 后 勤 工 程 学 院 建 筑 工 程 系 , 重 庆 400041)<br />

摘 要 : 对 于 挡 土 墙 背 不 是 光 滑 、 垂 直 , 墙 后 填 土 面 不 是 水 平 , 且 填 土 是 粘 性 土 时 , 土 压 力 的 计<br />

算 大 多 采 用 广 义 库 仑 土 压 力 理 论 。 现 行 的 国 家 标 准 也 推 荐 采 用 广 义 库 仑 土 压 力 理 论 计 算 。 对 于 支 挡<br />

结 构 后 缘 存 在 较 陡 峻 的 稳 定 岩 石 坡 面 时 , 国 标 也 提 出 了 相 应 的 公 式 , 同 时 规 定 稳 定 岩 石 坡 面 与 填 土<br />

间 的 摩 擦 角 , 当 无 试 验 资 料 时 , 可 取 δ R =0.33ϕ; 或 取 δ R =0.5ϕ。 公 式 在 实 际 使 用 过 程 中 , 存 在 一<br />

定 的 不 足 之 处 , 发 现 土 压 力 的 值 随 δ R 减 小 而 增 加 很 多 。 经 过 计 算 与 比 较 , 认 为 该 公 式 在 使 用 时 , 应<br />

将 参 数 调 正 为 δ R =(0.5-0.8)ϕ。<br />

关 键 词 : 国 家 标 准 挡 土 墙 土 压 力 系 数 摩 擦 角 δ R<br />

DISCUSSION OF ACTIVE EARTH PRESSURE’S COEFFICIENT FORMULA<br />

ABOUT THE NATIONAL STANDARD<br />

X. Lu, Y. F. Wang, Y. R. Zheng<br />

(Department of Architecture and Civil Engineering, Logistical Engineering <strong>University</strong>, Chongqing 400041,<br />

P.R.China)<br />

Abstract: <strong>The</strong> calculation of active earth pressure mostly adopted by broad sense Coulomb’s earth pressure<br />

theory, the retaining wall back is not smooth and perpendicular, fill behind the wall is not level, and is clayey<br />

soil. <strong>The</strong> current national standard also recommends the broad sense Coulomb’s earth pressure theory. <strong>The</strong><br />

standard provides corresponding formula, a stable steep rock existed behind the retaining structure. Friction<br />

angle δ R =0.33%,or δ R =0.5%, between the stable steep rock and fill, is been recommended in the current national<br />

standard . <strong>The</strong>re are several shortages when the formula used. <strong>The</strong> active earth pressure increased quickly when<br />

friction angle δ R decreased. Through calculation and comparison, friction angle δ R should be δ R =(0.5-0.8) ϕ in<br />

the future national standard formula used.<br />

Keywords: national standard, retaining wall, active earth pressure’s coefficient, friction angle δ R .<br />

一 引 言<br />

在 边 坡 工 程 与 支 挡 工 程 中 , 土 压 力 计 算 是 十 分 重 要 的 工 作 , 土 压 力 的 计 算 结 果 , 直 接 影 响 到 支<br />

挡 结 构 的 设 计 , 也 关 系 到 支 挡 结 构 的 安 全 。 目 前 常 用 的 土 压 力 计 算 方 法 是 朗 金 土 压 力 理 论 与 库 仑 土<br />

压 力 理 论 。 但 它 们 都 有 一 定 的 假 定 与 适 用 条 件 。 对 于 挡 土 墙 背 不 是 光 滑 、 垂 直 , 墙 后 填 土 面 不 是 水<br />

平 , 且 填 土 是 粘 性 土 时 , 以 上 两 种 理 论 都 不 适 用 。 大 多 采 用 广 义 库 仑 土 压 力 理 论 进 行 计 算 [1] 。 现 行<br />

的 国 家 标 准 《 建 筑 地 基 基 础 设 计 规 范 》(GB50007-2002)( 以 下 简 称 GB50007 规 范 ) [2] 、《 建 筑 边 坡<br />

工 程 技 术 规 范 》(GB50330-2002)( 以 下 简 称 GB50330 规 范 ) [3] 也 推 荐 采 用 广 义 库 仑 土 压 力 理 论 计 算 。<br />

对 于 支 挡 结 构 后 缘 存 在 较 陡 峻 的 稳 定 岩 石 坡 面 时 , 国 标 也 提 出 了 相 应 的 公 式 。 这 些 公 式 的 提 出 , 为<br />

工 程 设 计 提 供 便 利 。 但 在 实 际 使 用 过 程 中 , 这 些 公 式 存 在 一 定 的 不 足 之 处 , 现 通 过 实 际 工 程 的 例 子 ,<br />

进 行 了 计 算 与 比 较 , 发 现 这 几 个 公 式 在 使 用 时 , 有 些 参 数 需 要 作 调 正 。<br />

作 者 简 介 : 陆 新 (1959-), 男 , 上 海 崇 明 人 , 教 授 、 博 士 , 从 事 岩 土 工 程 教 学 、 科 研 与 工 程 实 践 工 作 。<br />

-369-


二 国 标 中 主 动 土 压 力 系 数 的 计 算 公 式<br />

目 前 国 标 GB50007 规 范 与 GB50330 规 范 中 规 定 了 主 动 土 压 力 计 算 公 式 , 计 算 公 式 分 一 般 情 况<br />

下 及 挡 墙 后 土 体 破 裂 面 以 内 有 较 陡 的 稳 定 岩 石 坡 面 的 二 种 情 况 , 主 动 土 压 力 的 计 算 公 式 是 相 同 的 ,<br />

主 动 土 压 力 系 数 是 不 一 样 的 , 分 别 如 下 式 :<br />

一 般 情 况 下 主 动 土 压 力 系 数 为 :<br />

K<br />

a<br />

sin( α + β)<br />

= + − + + −<br />

α α β ϕ δ<br />

2 2<br />

sin sin ( + − − )<br />

{ Kq[<br />

sin( α β)sin( α δ) sin( ϕ δ)sin( ϕ β)<br />

]<br />

+ 2η sinαcosϕcos( α + β −ϕ−δ) − 2 ⎡<br />

⎣( Kq<br />

sin( α + β)sin( ϕ− β) + ηsinαcos ϕ)<br />

( K sin( α − δ)sin( ϕ+ δ) + ηsinαcos ϕ)<br />

q<br />

0.5<br />

] }<br />

(1)<br />

式 中 : K<br />

q<br />

2qsinα<br />

cosβ<br />

= 1+<br />

γ H sin( α + β)<br />

2c<br />

η = γ H<br />

有 限 范 围 填 土 时 主 动 土 压 力 系 数 为 :<br />

K<br />

a<br />

sin( α + β)<br />

⎡sin( α + θ)sin( θ −δ ) cosδR⎤<br />

= −η<br />

α − δ + θ −δ θ −β ⎣ α α ⎥<br />

⎦<br />

R<br />

2<br />

sin( )sin( ) ⎢<br />

R<br />

sin sin<br />

(2)<br />

式 中 :θ ── 稳 定 岩 石 坡 面 的 倾 角 , 度 ;<br />

δ<br />

R<br />

── 稳 定 岩 石 坡 面 与 填 土 间 的 摩 擦 角 , 度 ; 宜 根 据 试 验 确 定 。 当 无 试 资 料 时 , 粘 性 土 与 粉 土<br />

可 取 δ = 0.33ϕ<br />

; 砂 性 土 与 碎 石 土 可 取 δ = 0.5ϕ<br />

。<br />

R<br />

R<br />

公 式 (1) 在 二 本 国 标 中 是 相 同 的 , 公 式 (2) 在 二 本 国 标 中 略 有 不 同 , 在 GB50330 中 考 虑 了 粘<br />

聚 力 C 的 影 响 , 同 时 稳 定 岩 石 坡 面 与 填 土 间 的 摩 擦 角 也 有 所 不 同 。<br />

三 对 公 式 进 行 计 算 与 探 讨<br />

为 了 探 讨 公 式 的 应 用 情 况 , 按 工 程 中 常 用 的 例 子 及 数 据 进 行 计 算 , 来 探 讨 国 标 中 公 式 的 适 用 情<br />

o<br />

o<br />

o<br />

o<br />

况 。 设 某 个 边 坡 的 参 数 为 H = 6 m, α = 70 , β = 15 , γ = 18 kN/m3, δ = 15 , ϕ = 35 , q = 0 。<br />

求 :1 当 θ =60°、62.5°、70°、80° 时 由 公 式 (1)、(2) 得 到 的 主 动 土 压 力 值 ;2 当 c =0、5、<br />

10、15、20kPa 时 由 公 式 (1)、(2) 得 到 的 主 动 土 压 力 值 ;3 当 α =90°, β =0°,δ =0° 与 δ =15°<br />

时 由 公 式 (1) 得 到 的 计 算 值 及 朗 金 主 动 土 压 力 计 算 值 与 静 止 土 压 力 值 。 经 过 计 算 得 到 如 下 结 果 。<br />

-370-


表 1 当 θ=62.5° 时 , 不 同 的 c 值 、 δ 值 时 主 动 土 压 力 计 算 结 果 汇 总<br />

Table 1 <strong>The</strong> values of active earth pressure with different c ,<br />

计 算 方 法<br />

c/kPa<br />

R<br />

δ<br />

R<br />

(θ=62.5°)<br />

0 5 10 15 20<br />

公 式 1 E1 170.28 135.21 101.67 69.17 37.40<br />

公 式 2<br />

(θ=<br />

62.5°)<br />

计 算 方 法<br />

δ<br />

R<br />

=0.2ϕ<br />

δ<br />

R<br />

=0.33ϕ<br />

δ<br />

R<br />

=0.4ϕ<br />

δ<br />

R<br />

=0.5ϕ<br />

δ<br />

R<br />

=0.6ϕ<br />

δ<br />

R<br />

=0.8ϕ<br />

δ<br />

R<br />

=1.0ϕ<br />

E 2 321.61 275.90 230.19 184.48 138.77<br />

E 2 / E 1 1.889 2.041 2.264 2.667 3.710<br />

E 2 295.23 251.28 207.32 163.36 119.41<br />

E 2 / E 1 1.734 1.858 2.039 2.362 3.193<br />

E 2 281.54 238.50 195.45 152.41 109.36<br />

E 2 / E 1 1.653 1.764 1.922 2.203 2.924<br />

E 2 262.45 220.68 178.90 137.13 95.35<br />

E 2 / E 1 1.541 1.632 1.760 1.983 2.549<br />

E 2 243.77 203.24 162.71 122.17 81.64<br />

E 2 / E 1 1.432 1.503 1.600 1.766 2.183<br />

E 2 207.04 168.95 130.87 92.78 54.69<br />

E 2 / E 1 1.216 1.250 1.287 1.341 1.462<br />

E 2 170.24 134.60 98.96 63.32 27.67<br />

E 2 / E 1 1.000 0.995 0.973 0.915 0.740<br />

注 :1 土 压 力 单 位 为 kN/m, 下 同 。 2 θ=62.5° 是 公 式 (1) 的 理 论 破 裂 面 。<br />

表 2 当 θ=70° 时 , 不 同 的 c 值 、 δ 值 时 主 动 土 压 力 计 算 结 果 汇 总<br />

Table 2 <strong>The</strong> values of active earth pressure with different c ,<br />

c/kPa<br />

R<br />

δ<br />

R<br />

(θ=70°)<br />

0 5 10 15 20<br />

公 式 1 E 1 170.28 135.21 101.67 69.17 37.40<br />

公 式 2<br />

(θ=70°)<br />

δ<br />

R<br />

=0.2ϕ<br />

δ<br />

R<br />

=0.33ϕ<br />

δ<br />

R<br />

=0.4ϕ<br />

δ<br />

R<br />

=0.5ϕ<br />

δ<br />

R<br />

=0.6ϕ<br />

δ<br />

R<br />

=0.8ϕ<br />

δ<br />

R<br />

=1.0ϕ<br />

E 2 289.45 245.8 202.16 158.51 114.87<br />

E 2 / E 1 1.700 1.818 1.988 2.292 3.071<br />

E 2 266.44 224.97 183.51 142.04 100.58<br />

E 2 / E 1 1.565 1.664 1.805 2.053 2.689<br />

E 2 254.71 214.36 174 133.65 93.30<br />

E 2 / E 1 1.496 1.585 1.711 1.932 2.495<br />

E 2 238.60 199.77 160.95 122.12 83.30<br />

E 2 / E 1 1.401 1.477 1.583 1.766 2.227<br />

E 2 223.10 185.74 148.39 111.03 73.67<br />

E 2 / E 1 1.310 1.374 1.460 1.605 1.970<br />

E 2 193.37 158.83 124.29 89.75 55.21<br />

E 2 / E 1 1.136 1.175 1.222 1.298 1.476<br />

E 2 164.52 132.71 100.91 69.11 37.30<br />

E 2 / E 1 0.966 0.982 0.993 0.999 0.997<br />

注 : 当 θ=70° 时 , 表 明 稳 定 岩 石 坡 面 在 公 式 (1) 的 理 论 破 裂 面 以 内 , 应 按 公 式 (2) 计 算 。 公<br />

式 (1) 的 结 果 是 作 比 较 用 的 。<br />

-371-


表 3 c =0kPa 时 不 同 的<br />

δ 、θ 值 主 动 土 压 力 计 算 结 果<br />

Table 3 <strong>The</strong> values of active earth pressure with different<br />

公 式 1 170.28<br />

R<br />

δ<br />

R<br />

,θ ( c =0kPa)<br />

θ<br />

60 62.5 70 80<br />

公 式 2<br />

δ<br />

R<br />

0.2ϕ 332.53 321.61 289.45 244.72<br />

0.33ϕ 304.71 295.23 266.44 224.79<br />

0.4ϕ 290.18 281.54 254.71 214.92<br />

0.5ϕ 269.84 262.45 238.60 201.66<br />

0.6ϕ 249.81 243.77 223.10 189.21<br />

0.8ϕ 210.13 207.04 193.37 166.17<br />

1.0ϕ 169.94 170.24 164.52 144.79<br />

注 : 当 θ=60° 时 不 适 用 于 公 式 (2) 计 算 , 计 算 值 尽 供 讨 论 用 。 表 4、5 相 同 。<br />

表 4 c =10kPa 时 不 同 的<br />

δ<br />

R<br />

、θ 值 主 动 土 压 力 计 算 结 果<br />

δ<br />

R<br />

,θ (c =10kPa)<br />

Table 4 <strong>The</strong> values of active earth pressure with different<br />

公 式 1 101.67<br />

θ<br />

60 62.5 70 80<br />

公 式 2<br />

δ<br />

R<br />

0.2ϕ 238.65 230.19 202.16 156.32<br />

0.33ϕ 214.09 207.32 183.51 142.38<br />

0.4ϕ 201.27 195.45 174 135.47<br />

0.5ϕ 183.31 178.90 160.95 126.20<br />

0.6ϕ 165.63 162.71 148.39 117.49<br />

0.8ϕ 130.60 130.87 124.29 101.37<br />

1.0ϕ 95.11 98.96 100.91 86.41<br />

表 5 c =20kPa 时 不 同 的<br />

δ<br />

R<br />

、θ 值 主 动 土 压 力 计 算 结 果<br />

δ<br />

R<br />

,θ (c =20kPa)<br />

Table 5 <strong>The</strong> values of active earth pressure with different<br />

公 式 1 37.40<br />

θ<br />

60 62.5 70 80<br />

公 式 2<br />

δ<br />

R<br />

0.2ϕ 144.77 138.77 114.87 67.92<br />

0.33ϕ 123.47 119.41 100.58 59.97<br />

0.4ϕ 112.35 109.36 93.30 56.03<br />

0.5ϕ 96.77 95.35 83.30 50.73<br />

0.6ϕ 81.44 81.64 73.67 45.77<br />

0.8ϕ 51.06 54.69 55.21 36.57<br />

1.0ϕ 20.29 27.67 37.30 28.04<br />

-372-


表 6 当 α = 90<br />

, β = 0<br />

, 变 化 c 、δ 值 , 对 比 公 式 1 与 朗 金 主 动 土 压 力 计 算 结 果<br />

Table 6 <strong>The</strong> values comparison between Rankine’s active earth pressure and equation(1) with<br />

different c , δ ( α = 90<br />

, β = 0<br />

)<br />

c/kPa<br />

计 算<br />

方 法<br />

0 10<br />

朗 金 主 动 87.80 36.44<br />

公 式 1<br />

o<br />

δ = 0<br />

o<br />

δ = 5<br />

o<br />

δ = 10<br />

o<br />

δ = 15<br />

o<br />

δ = 0<br />

o<br />

δ = 5<br />

o<br />

δ = 10<br />

o<br />

δ = 15<br />

87.80 84.37 81.91 80.28 25.33 24.33 23.58 23.05<br />

从 表 6 计 算 结 果 知 道 , 当 c =0、δ =0 时 , 朗 金 主 动 土 压 力 计 算 结 果 与 公 式 (1) 的 计 算 结 果 相<br />

同 。<br />

经 计 算 , 静 止 土 压 力 计 算 值 为 138kPa。<br />

从 以 上 计 算 结 果 知 道 :<br />

(1) 从 表 1、2 的 计 算 结 果 看 , 土 的 内 聚 力 c 与 稳 定 岩 石 坡 面 与 填 土 之 间 的 摩 擦 角 δ<br />

R 对 主 动 土<br />

压 力 计 算 值 影 响 很 大 。 当 θ=62.5°, δ R =1.0ϕ,c =0 时 , 公 式 (1) 与 公 式 (2) 的 值 相 同 ; 当 δ<br />

R =1.0ϕ<br />

时 , 随 c 、θ 值 的 增 大 , 公 式 (2) 的 值 比 公 式 (1) 的 值 小 。 其 余 情 况 下 , 公 式 (2) 的 值 比 公 式 (1)<br />

的 值 大 。 如 果 取 c =0, 其 结 果 为 最 大 , 随 填 土 的 内 聚 力 c 的 增 加 , 主 动 土 压 力 值 明 显 减 小 。 同 样 ,<br />

随 稳 定 岩 石 坡 面 与 填 土 之 间 的 摩 擦 角 δ<br />

R 的 增 加 , 主 动 土 压 力 值 也 明 显 减 小 。<br />

(2) 从 表 3~5 的 计 算 结 果 看 , 随 稳 定 岩 石 坡 面 角 度 θ 的 增 加 , 主 动 土 压 力 的 值 会 有 所 减 小 ,<br />

但 影 响 值 明 显 比 稳 定 岩 石 坡 面 与 填 土 之 间 的 摩 擦 角 δ<br />

R 影 响 小 。 实 际 上 , 在 表 3~5 中 的 θ=60° 的 情<br />

况 是 不 适 用 于 公 式 (2) 的 , 表 中 列 出 主 要 是 为 了 探 讨 土 压 力 随 θ 角 的 影 响 。<br />

(3) 从 表 6 的 计 算 结 果 看 , 当 α = 90<br />

o , β = 0<br />

o ,c =0、δ =0 时 公 式 (1) 与 朗 金 主 动 土 压 力 计<br />

算 结 果 是 相 同 的 , 说 明 公 式 (1) 符 合 经 典 土 压 力 计 算 理 论 。 随 挡 土 墙 背 的 摩 擦 角 δ 的 增 加 , 主 动<br />

土 压 力 相 应 有 所 减 小 , 但 其 变 化 值 不 大 。 当 c =10kPa 时 , 公 式 (1) 的 计 算 结 果 与 朗 金 土 压 力 计 算 结 果<br />

不 一 致 , 其 值 要 比 朗 金 土 压 力 理 论 计 算 值 略 小 。 静 止 土 压 力 计 算 结 果 要 比 规 范 中 公 式 计 算 的 大 , 这<br />

符 合 一 般 的 土 压 力 计 算 理 论 。<br />

四 对 国 标 公 式 的 讨 论<br />

经 过 对 国 标 中 不 同 主 动 土 压 力 系 数 公 式 的 计 算 及 分 析 , 在 特 殊 情 况 下 公 式 (1) 与 古 典 土 压 力<br />

计 算 理 论 相 同 , 该 公 式 有 一 定 的 可 信 性 。 当 稳 定 岩 石 坡 面 等 于 θ=45°+ϕ /2, 且 δ<br />

R =1.0ϕ 时 , 公 式<br />

(1) 计 算 结 果 等 于 公 式 (2)。 所 以 我 们 认 为 国 标 中 的 计 算 公 式 基 本 具 有 一 致 性 。 但 当 δ<br />

R 0 时 , 比 值


的 值 比 公 式 (1) 的 值 小 外 , 其 余 的 值 都 比 公 式 (1) 的 值 大 。 根 据 有 的 学 者 研 究 认 为 , 当 存 在 较 陡<br />

峻 岩 石 坡 面 时 的 土 压 力 值 应 界 于 主 动 土 压 力 与 静 止 土 压 力 之 间 [4]。 因 此 , 综 合 分 析 认 为 原 规 范 规 定<br />

的 δ<br />

R =0.33ϕ 或 δ<br />

R =0.5ϕ 时 , 得 到 的 土 压 力 值 明 显 偏 大 , 甚 至 大 于 静 止 土 压 力 。 所 以 , 笔 者 认 为 δ<br />

R<br />

应 作 适 当 调 整 。 根 据 以 上 分 析 , 认 为 δ =(0.5~0.8)ϕ 基 本 合 理 。<br />

R<br />

五<br />

结 论<br />

(1) 国 标 规 范 中 的 2 个 公 式 , 当 条 件 处 于 特 定 情 况 时 , 其 值 是 基 本 一 致 的 , 且 与 朗 金 土 压 力 、<br />

库 仑 土 压 力 计 算 理 论 基 本 一 致 。 因 此 规 范 的 公 式 具 有 一 定 的 理 论 基 础 。<br />

(2) 国 标 规 范 中 , 当 坡 后 存 在 较 陡 峻 稳 定 岩 石 坡 面 时 , 原 规 范 规 定 的 δ<br />

R =0.33ϕ 或 δ<br />

R =0.5ϕ<br />

的 公 式 (2) 计 算 结 果 明 显 偏 大 。 根 据 计 算 分 析 认 为 δ<br />

R =(0.5~0.8)ϕ 比 较 合 适 。 建 议 在 规 范 修 订 时<br />

作 相 应 的 调 正 。<br />

参 考 文 献<br />

[1] 赵 明 华 . 土 力 学 与 基 础 工 程 ( 第 2 版 ). 武 汉 理 工 大 学 出 版 社 , 2004.<br />

Zhao, M.H. (2004). Soil mechanics and foundation engineering. Wuhan <strong>University</strong> of Technology Press.<br />

[2] 中 华 人 民 共 和 国 国 家 标 准 . GB50007-2002 建 筑 地 基 基 础 设 计 规 范 , 2002.<br />

National standard of the People’s Republic of China. GB 50007-2002 Code for design of building foundation, 2002.<br />

[3] 中 华 人 民 共 和 国 国 家 标 准 . GB50330-2002 建 筑 边 坡 工 程 技 术 规 范 , 2002.<br />

National standard of the People’s Republic of China. GB 50330-2002 Technical code for building slope engineering, 2002.<br />

[4] 黄 求 顺 , 张 四 平 , 胡 岱 文 . 边 坡 工 程 . 重 庆 大 学 出 版 社 , 2004.<br />

Huang, Q.S., Zhang, S.P., Hu, and D.W. Slope Engineering. Chongqing <strong>University</strong> Press, 2004.<br />

-374-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

麒 麟 山 超 大 断 面 隧 道 施 工 期 变 形 规 律 及 控 制 分 析<br />

1<br />

李 彰 明<br />

2<br />

冯 强<br />

2<br />

朱 新 浩<br />

1<br />

巫 峰<br />

1<br />

万 灵<br />

3<br />

欧 加 利<br />

1<br />

广 东 工 业 大 学 岩 土 工 程 技 术 开 发 中 心 , 广 东 广 州 510090<br />

2<br />

深 圳 市 宝 安 区 建 筑 工 务 局 , 广 东 深 圳 518101<br />

3<br />

中 交 第 三 航 务 工 程 局 深 圳 分 公 司 , 广 东 深 圳 518067<br />

摘 要 : 针 对 双 向 八 车 道 、V 级 围 岩 、 小 间 距 、 单 孔 开 挖 最 大 外 轮 廓 面 积 为 241.95m 2 超 大 断 面<br />

麒 麟 山 双 孔 隧 道 , 利 用 原 位 监 控 量 测 , 对 隧 道 施 工 期 变 形 规 律 与 变 形 控 制 进 行 总 结 及 分 析 讨 论 。 结<br />

果 表 明 : 该 类 隧 道 采 用 双 侧 壁 导 坑 多 台 阶 法 分 部 开 挖 可 较 好 控 制 围 岩 变 形 ; 每 个 台 阶 开 挖 时 应 及 时<br />

进 行 支 护 , 防 止 上 一 台 阶 初 期 支 护 拱 脚 滑 移 以 确 保 稳 定 ; 及 时 有 效 的 初 期 支 护 对 整 个 隧 道 的 开 挖 至<br />

关 重 要 , 尽 早 施 作 仰 拱 形 成 封 闭 环 对 于 改 善 结 构 受 力 和 抑 制 隧 道 变 形 非 常 有 效 ; 在 洞 口 施 作 大 管 棚<br />

进 行 预 支 护 , 对 洞 周 围 岩 进 行 系 统 径 向 注 浆 加 固 , 再 用 超 前 小 导 管 加 固 工 作 面 前 方 围 岩 能 够 有 效 抑<br />

制 进 出 口 段 过 大 变 形 。 这 些 结 论 可 资 类 似 工 程 借 鉴 并 为 进 一 步 研 究 提 供 基 础 。<br />

关 键 词 : 变 形 规 律 变 形 控 制 分 析 超 大 断 面 双 孔 隧 道 施 工 期<br />

DEFORMATION CHARACTER AND CONTROL ANALYSIS FOR A<br />

LARGE-SECTION TWIN TUNNEL IN CONSTRUCTION<br />

Z. M. Li 1 , Q. Feng 2 , X. H. Zhu 2 , F. Wu 1 , L. Wan 1 and J. L. Ou 3<br />

1 Center of Geotechnical Engineering, Guangdong <strong>University</strong> of Technology,Guangzhou 510090,China<br />

2<br />

Baoan Architectural Works Bureau, Shenzhen 518100, China<br />

3<br />

Shenzhen Branch, CCCC Third Harbor Engineering Co., LTD, Shenzhen 518067, China<br />

Abstract: For two-way eight-lane, V-class surrounding rock, the small spacing, the large single hole section<br />

excavated 241.95m 2 of Unicorn Mountain twin tunnel in Shenzhen, the deformation character and control of the<br />

tunnel in construction period are analyzed and discussed upon in situ monitoring. <strong>The</strong> results showed that the<br />

double side heading multi-step construction method for this kind of tunnel can well control the surrounding rock<br />

deformation of the tunnel; each bench excavation support should be timely to prevent the foot arch support slip<br />

to ensure stability; timely and effective initial support to the excavation of the tunnel is essential, constructing<br />

inverted arch as early as possible to form a closed loop is very effective for improving the structure and limiting<br />

tunnel deformation; supporting with the big pipe roof in the hole at the entrance to advance protection,<br />

systematic radial grouting around the surrounding rock, and then strengthening ahead of rock with a small tube<br />

advance consolidation section can inhibit the large deformation at the entrance and exit of the tunnel. <strong>The</strong>se<br />

conclusions could be referenced for similar projects and provide a basis for further study.<br />

Keywords: Deformation character, control, analysis, a large cross-section twin tunnel, construction.<br />

一 、 前 言<br />

为 适 应 快 速 发 展 的 高 等 级 公 路 建 设 , 各 地 已 建 成 或 正 在 建 设 越 来 越 多 的 大 跨 度 隧 道 。 与 普 通 隧<br />

道 相 比 , 这 些 隧 道 断 面 大 且 施 工 方 法 更 为 多 样 化 、 变 形 更 加 复 杂 。 业 界 在 大 跨 隧 道 变 形 及 稳 定 性 研<br />

-375-


[1]<br />

究 等 方 面 都 有 了 新 的 发 展 与 提 高 , 并 取 得 了 一 定 的 研 究 成 果 。 王 鑫 等 验 证 了 浅 埋 、 大 跨 隧 道 地 表<br />

[2]<br />

控 制 对 隧 道 洞 口 稳 定 的 重 要 性 ; 瞿 万 波 等 根 据 监 测 数 据 , 提 出 洞 桩 施 工 方 法 有 效 控 制 了 浅 埋 大 跨<br />

[3]<br />

隧 道 地 表 沉 降 和 地 层 变 形 ; 王 者 超 等 通 过 变 形 监 测 给 出 了 分 岔 隧 道 施 工 过 程 中 表 现 出 来 的 动 态 响<br />

[4]<br />

应 特 点 及 空 间 效 应 特 点 ; 夏 才 初 等 分 析 大 断 面 小 净 距 隧 道 围 岩 和 支 护 系 统 变 形 及 受 力 特 点 ; 周 丁<br />

[5]<br />

恒 等 提 出 不 同 围 岩 情 况 下 特 大 断 面 公 路 隧 道 控 制 大 变 形 的 工 程 措 施 。<br />

目 前 , 虽 然 对 于 大 跨 隧 道 取 的 研 究 得 了 一 定 成 就 , 但 是 对 于 大 断 面 双 孔 隧 道 施 工 期 间 变 形 规 律<br />

及 控 制 研 究 还 不 够 。 本 文 以 双 向 八 车 道 特 大 断 面 麒 麟 山 隧 道 工 程 为 背 景 , 以 现 场 监 控 量 测 为 手 段 ,<br />

对 该 类 隧 道 施 工 期 变 形 规 律 及 控 制 进 行 分 析 讨 论 , 以 资 类 似 工 程 借 鉴 并 为 进 一 步 研 究 提 供 基 础 。<br />

二 、 工 程 特 点<br />

麒 麟 山 ( 车 行 ) 隧 道 位 于 中 国 深 圳 宝 安 区 松 岗 镇 。 该 隧 道 所 在 地 区 的 原 始 地 貌 为 低 丘 , 自 然 地 形<br />

坡 度 35°~40°, 地 面 标 高 在 7.0~40.0m 之 间 , 地 形 起 伏 较 大 , 场 地 范 围 内 主 要 的 节 理 为 原 生 裂<br />

隙 , 基 本 呈 微 ~ 张 开 状 态 , 个 别 节 理 面 延 伸 长 8~10m。 裂 隙 产 状 为 :( 倾 向 ∠ 倾 角 ) 层 理 裂 隙 83~<br />

88°∠33~38°,122°∠61°,3 条 /m;182°∠75°,2 条 /m;280°∠56°,3 条 /m。 据 裂 隙 性<br />

质 与 线 路 关 系 分 析 , 在 隧 道 出 口 处 岩 石 层 理 裂 隙 倾 向 、 倾 角 与 边 坡 倾 向 一 致 ; 岩 石 层 理 多 泥 质 和 砂<br />

质 胶 结 , 呈 微 张 状 态 , 为 隧 道 出 口 的 潜 在 滑 动 面 。 拟 建 工 程 线 路 内 分 布 且 隧 道 穿 越 的 土 ( 岩 ) 为 自<br />

上 而 下 分 别 为 第 四 系 杂 填 土 、 植 物 层 、 坡 残 积 亚 粘 土 、 残 积 亚 粘 土 , 侏 罗 系 砂 岩 、 粉 砂 质 泥 岩 。<br />

该 工 程 主 要 特 点 为 : 上 覆 重 要 建 筑 物 ( 深 圳 输 电 枢 纽 之 高 压 电 塔 )、 超 浅 埋 深 ( 最 大 埋 深 为<br />

13.1m)、 超 大 跨 度 ( 开 挖 时 单 洞 21.6m、 双 洞 47.6m)、 开 挖 断 面 最 大 高 度 为 13.96m、 单 孔 开 挖 最 大<br />

外 轮 廓 面 积 为 241.95m 2 , 两 孔 净 距 小 ( 开 挖 时 中 间 土 柱 厚 4.4~4.7m) 及 左 、 右 线 并 行 ; 且 其 围 岩 大 部<br />

分 划 为 极 破 碎 的 V 级 。 鉴 于 上 述 条 件 极 易 造 成 隧 道 开 挖 面 及 上 覆 电 塔 不 稳 定 , 因 而 采 用 双 侧 壁 导 坑<br />

法 ( 上 、 中 、 下 ) 多 台 阶 法 开 挖 , 两 次 衬 砌 完 成 施 工 。<br />

三 、 支 护 结 构 设 计 及 施 工 流 程<br />

该 隧 道 的 设 计 和 施 工 采 用 新 奥 法 思 想 。 充 分 利 用 围 岩 自 承 能 力 , 以 大 管 棚 或 超 前 注 浆 小 导 管 、<br />

锚 杆 、 钢 筋 网 、 喷 混 凝 土 、 钢 拱 架 为 初 期 支 护 ; 然 后 根 据 隧 道 监 控 量 测 数 据 确 定 二 次 衬 砌 合 理 的 施<br />

作 时 间 , 以 确 保 隧 道 安 全 有 序 地 进 行 开 挖 。 隧 道 支 护 参 数 选 择 是 根 据 围 岩 类 别 、 工 程 地 质 与 水 文 地<br />

质 条 件 、 地 形 、 埋 深 、 跨 度 以 及 施 工 方 法 等 以 工 程 类 比 方 法 确 定 , 并 结 合 施 工 过 程 的 模 拟 数 值 分 析<br />

与 原 位 监 测 所 获 得 的 信 息 反 馈 , 及 时 调 整 、 修 改 支 护 参 数 , 进 行 支 护 衬 砌 的 再 设 计 。 初 期 支 护 及 二<br />

次 衬 砌 的 具 体 支 护 参 数 见 表 1。<br />

表 1 隧 道 复 合 式 衬 砌 支 护 参 数<br />

断 面 型 式 围 岩 级 别 初 支 参 数 二 衬 参 数<br />

Ⅴ 级<br />

复 合 加 强<br />

Ⅴ<br />

φ108×6 大 管 棚 +R25 中 空 注 浆 系 统 锚 杆 +310mm<br />

厚 C25 喷 砼 +Ⅰ25a 型 钢 钢 架 ( 间 距 0.5m)+φ8、<br />

200×200mm 钢 筋 网<br />

φ42×3.5 超 前 注 浆 小 导 管 +R25 中 空 注 浆 系 统 锚<br />

杆 +310mm 厚 C25 喷 砼 +Ⅰ25a 型 钢 钢 架 ( 间 距<br />

0.5m)+φ8、200×200mm 钢 筋 网<br />

700mm 厚 C30、S8<br />

钢 筋 砼 , 带 仰 拱<br />

Ⅴ 级 复 合<br />

Ⅴ<br />

650mm 厚 C30、S8<br />

钢 筋 砼 , 带 仰 拱<br />

由 于 隧 道 围 岩 条 件 差 , 开 挖 面 不 稳 定 , 故 采 用 双 侧 壁 导 坑 法 多 台 阶 法 开 挖 。 其 工 序 为 : 施 作 超<br />

前 长 管 棚 或 小 导 管 ( 测 量 放 线 、 打 入 长 管 棚 或 小 导 管 、 做 止 浆 墙 封 闭 掌 子 面 、 注 浆 ) → 开 挖 上 台 阶 及<br />

左 、 右 侧 导 洞 → 第 一 次 柔 性 衬 砌 ( 初 喷 5cm 混 凝 土 , 立 边 墙 钢 架 、 挂 网 、 打 预 应 力 锚 杆 , 复 喷 至 设 计<br />

厚 度 ) → 施 作 底 部 横 撑 → 开 挖 下 一 台 阶 ( 过 程 如 上 ), 初 支 封 闭 成 环 → 开 挖 核 心 土 → 施 作 二 次<br />

衬 砌 ; 整 个 过 程 实 施 包 括 变 形 在 内 的 各 种 监 测 。 主 要 施 工 工 序 见 图 1。<br />

-376-


1<br />

2<br />

5<br />

6<br />

7<br />

3<br />

4<br />

图 1 隧 道 施 工 工 序 示 意 图<br />

图 2 变 形 监 测 断 面 布 置 图<br />

四 、 监 测 方 案 设 计<br />

该 特 大 断 面 隧 道 两 洞 之 间 间 距 小 , 左 右 洞 施 工 相 互 影 响 较 大 , 各 洞 施 工 工 序 也 直 接 影 响 着 隧 道<br />

围 岩 和 结 构 的 稳 定 性 , 特 别 是 中 间 核 心 土 , 受 力 及 变 形 尤 为 复 杂 ; 而 目 前 可 供 借 鉴 的 类 似 工 程 经 验<br />

较 少 , 设 计 和 施 工 方 法 也 都 还 不 成 熟 。 因 此 , 加 强 监 控 量 测 是 保 证 隧 道 安 全 施 工 极 为 重 要 的 手 段 。<br />

依 据 公 路 隧 道 施 工 规 范 的 基 本 要 求 , 针 对 该 隧 道 的 结 构 特 点 、 施 工 工 艺 及 地 质 情 况 , 确 定 隧 道 各 类<br />

监 测 项 目 达 11 种 ; 因 篇 幅 有 限 , 本 文 在 此 着 重 介 绍 对 稳 定 性 起 控 制 作 用 的 变 形 监 测 , 相 关 项 目 及 量<br />

测 方 法 见 表 2。<br />

各 断 面 变 形 监 测 测 点 具 体 布 设 情 况 见 图 2。<br />

表 2 隧 道 变 形 监 测 主 要 项 目 及 量 测 方 法<br />

围 岩 级 别 监 测 项 目 监 测 仪 器 测 点 ( 测 线 ) 数 / 个<br />

Ⅴ<br />

水 平 净 空 收 敛 收 敛 计 10( 左 线 )<br />

拱 顶 下 沉 精 密 水 准 仪 、 钢 挂 尺 , 全 站 仪 15<br />

浅 埋 地 段 下 沉 精 密 水 准 仪 , 铟 钢 尺 , 全 站 仪 47<br />

围 岩 内 位 移 ( 洞 内 设 点 ) 洞 内 钻 孔 中 安 设 多 点 式 位 移 计 17<br />

五 、 监 测 结 果 及 分 析<br />

5.1. 浅 埋 地 表 沉 降 分 析<br />

根 据 相 关 规 范 及 开 挖 断 面 、 埋 深 及 岩 性 情 况 , 特 别 是 隧 道 左 线 处 上 覆 高 压 电 塔 稳 定 性 要 求 , 进<br />

行 了 地 表 沉 降 监 测 以 掌 握 地 表 累 积 沉 降 量 及 沉 降 速 率 。 经 过 对 该 隧 道 浅 埋 段 典 型 断 面 地 表 沉 降 监<br />

测 , 绘 出 浅 埋 典 型 断 面 隧 道 施 工 过 程 中 地 表 沉 降 随 时 间 的 变 化 关 系 曲 线 , 各 断 面 变 化 曲 线 如 图 3~<br />

图 5 所 示 。<br />

该 隧 道 典 型 断 面 地 表 沉 降 观 测 数 据 分 析 表 明 :<br />

(1) 浅 埋 段 隧 道 施 工 过 程 中 , 虽 然 受 相 邻 隧 道 施 工 的 影 响 , 但 各 地 表 沉 降 均 是 在 隧 道 中 心 线<br />

或 其 附 近 沉 降 量 最 大 , 两 侧 沉 降 量 随 测 点 距 离 隧 道 中 心 距 离 的 增 加 而 逐 渐 减 小 ; 同 一 隧 道 断 面 上 沿<br />

隧 道 横 断 面 地 表 沉 降 曲 线 表 现 大 致 呈 抛 物 线 形 。<br />

(2) 就 同 一 断 面 而 言 , 隧 道 洞 身 正 上 方 测 点 Z1、Z2、Z3、Z4 处 地 表 下 沉 量 最 大 , 此 时 , 隧 道 顶<br />

拱 中 点 正 上 方 Z4 测 点 , 最 大 地 表 沉 降 量 约 48.9mm; 在 隧 道 洞 身 正 上 方 以 外 的 其 它 测 点 处 地 表 下 沉 量<br />

相 对 较 小 , 其 中 隧 道 左 侧 测 点 Z3 下 沉 约 为 22.9mm, 而 测 点 Z5、Y1、Y3 地 表 沉 降 量 均 小 于 20mm; 进<br />

口 截 水 天 沟 Y2 点 所 在 处 为 新 填 土 层 , 从 而 导 致 该 处 沉 降 很 大 ; 另 外 R1、R2、R3 受 旁 边 人 行 隧 道 开 挖<br />

的 影 响 , 有 一 定 的 沉 降 量 。<br />

-377-


(3) 根 据 对 隧 道 典 型 断 面 观 测 数 据 的 分 析 , 在 隧 道 埋 深 较 浅 时 ( 小 于 1.5B), 隧 道 开 挖 横 向 影<br />

响 区 域 主 要 在 位 于 “ 隧 道 顶 埋 深 H0+ 隧 道 宽 度 B” 范 围 以 内 。 因 此 , 横 向 影 响 范 围 较 小 , 而 在 影 响 范<br />

围 内 的 测 点 沉 降 量 较 大 。<br />

(4) 上 覆 高 压 电 塔 基 础 的 总 沉 降 及 沉 降 差 均 在 稳 定 性 及 设 计 要 求 的 允 许 范 围 内 。<br />

5.2. 拱 顶 下 沉 监 测 分 析<br />

对 于 左 线 侧 导 洞 拱 顶 沉 降 的 研 究 , 因 为 左 线 ZK16+465 左 导 洞 离 高 压 电 塔 非 常 近 , 为 了 监 测 隧<br />

道 开 挖 过 程 对 电 塔 的 影 响 , 所 以 选 取 该 断 面 所 在 的 两 个 侧 导 洞 的 拱 顶 沉 降 进 行 研 究 ; 对 于 左 线 核 心<br />

土 拱 顶 沉 降 进 行 监 测 ; 又 因 进 口 位 置 埋 深 在 整 个 隧 道 中 最 浅 , 地 质 情 况 很 复 杂 , 各 种 隐 形 病 害 都 进<br />

行 监 测 ; 又 因 进 口 位 置 埋 深 在 整 个 隧 道 中 最 浅 , 地 质 情 况 很 复 杂 , 各 种 隐 形 病 害 都 可 能 存 在 , 故 选<br />

取 离 进 口 最 近 的 断 面 ZK16+420 所 在 的 核 心 土 的 沉 降 进 行 监 测 。 经 过 对 麒 麟 山 隧 道 左 线 拱 顶 下 沉 近 3<br />

个 月 监 测 , 绘 出 隧 道 施 工 过 程 中 拱 顶 下 沉 随 时 间 变 化 关 系 曲 线 , 各 断 面 变 化 曲 线 如 图 6~ 图 7 所 示 。<br />

累 积 沉 降 量 ( mm)<br />

90.0<br />

80.0<br />

70.0<br />

60.0<br />

50.0<br />

40.0<br />

30.0<br />

20.0<br />

10.0<br />

0.0<br />

-10.0<br />

-20.0<br />

2009/12/6<br />

2009/12/20<br />

2010/1/3<br />

2010/1/17<br />

2010/1/31<br />

2010/2/14<br />

2010/2/28<br />

2010/3/14<br />

2010/3/28<br />

2010/4/11<br />

2010/4/25<br />

2010/5/9<br />

2010/5/23<br />

2010/6/6<br />

2010/6/20<br />

2010/7/4<br />

2010/7/18<br />

2010/8/1<br />

2010/8/15<br />

2010/8/29<br />

2010/9/12<br />

2010/9/26<br />

2010/10/10<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

R1#<br />

R2#<br />

R3#<br />

Z1#<br />

Z2#<br />

Z3#<br />

Z4#<br />

Z5#<br />

Y1#<br />

Y2#<br />

Y3#<br />

35.0<br />

30.0<br />

25.0<br />

20.0<br />

15.0<br />

10.0<br />

5.0<br />

0.0<br />

-5.0<br />

累 积 沉 降 量 ( m m )<br />

2 0 1 0 / 5 / 1<br />

2 0 1 0 / 5 / 1 5<br />

2 0 1 0 / 5 / 2 9<br />

2 0 1 0 / 6 / 1 2<br />

2 0 1 0 / 6 / 2 6<br />

2 0 1 0 / 7 / 1 0<br />

2 0 1 0 / 7 / 2 4<br />

2 0 1 0 / 8 / 7<br />

2 0 1 0 / 8 / 2 1<br />

2 0 1 0 / 9 / 4<br />

2 0 1 0 / 9 / 1 8<br />

2 0 1 0 / 1 0 / 2<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

R1#<br />

R2#<br />

R3#<br />

Z1#<br />

Z2#<br />

Z3#<br />

Z4#<br />

Z5#<br />

Y1#<br />

Y2#<br />

Y3#<br />

图 3 进 口 沿 截 水 天 沟 地 表 沉 降 时 程 曲 线<br />

图 4 出 口 沿 截 水 天 沟 地 表 沉 降 时 程 曲 线<br />

累 计 沉 降 量 ( mm)<br />

20.0<br />

15.0<br />

10.0<br />

5.0<br />

0.0<br />

-5.0<br />

-10.0<br />

2010/3/20<br />

2010/4/3<br />

2010/4/17<br />

2010/5/1<br />

2010/5/15<br />

2010/5/29<br />

2010/6/12<br />

2010/6/26<br />

2010/7/10<br />

2010/7/24<br />

2010/8/7<br />

2010/8/21<br />

2010/9/4<br />

2010/9/18<br />

2010/10/2<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

R1#<br />

R2#<br />

R3#<br />

Z1#<br />

Z2#<br />

Z3#<br />

Z4#<br />

Z5#<br />

Y1#<br />

Y2#<br />

Y3#<br />

图 5 离 进 口 25m 处 典 型 地 表 沉 降 测 点 时 程 曲 线<br />

根 据 现 场 监 测 数 据 及 拱 顶 下 沉 时 程 关 系 曲 线 图 可 知 , 隧 道 里 程 ZK16+465 左 、 右 侧 导 洞 的 拱 顶<br />

沉 降 最 终 值 分 别 为 5.82mm,5.39mm; 核 心 土 的 拱 顶 沉 降 的 最 大 值 是 19.72mm, 平 均 值 为 18.70mm。<br />

因 为 ZK16+465 左 右 侧 导 洞 拱 顶 下 沉 测 点 是 在 两 侧 导 洞 打 通 后 才 布 设 的 , 故 有 一 定 的 滞 后 性 , 但 是<br />

从 后 期 的 监 测 结 果 来 看 两 侧 导 洞 的 拱 顶 下 沉 量 较 小 , 也 符 合 某 种 变 形 规 律 。<br />

-378-


拱 顶 沉 降 (mm)<br />

0.00<br />

-1.00<br />

-2.00<br />

-3.00<br />

-4.00<br />

-5.00<br />

-6.00<br />

-7.00<br />

2010/8/19<br />

2010/8/26<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

2010/9/2<br />

2010/9/9<br />

K16+465 左<br />

K16+465 右<br />

2010/9/16<br />

2010/9/23<br />

2010/9/30<br />

2010/10/7<br />

拱 顶 沉 降 (mm)<br />

0.00<br />

-5.00<br />

-10.00<br />

-15.00<br />

-20.00<br />

-25.00<br />

2010/7/5<br />

2010/7/12<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

2010/7/19<br />

2010/7/26<br />

2010/8/2<br />

2010/8/9<br />

2010/8/16<br />

2010/8/23<br />

2010/8/30<br />

2010/9/6<br />

2010/9/13<br />

K16+420 左 K16+420 中 K16+420 右<br />

2010/9/20<br />

图 6 ZK16+465 断 面 侧 导 洞 拱 顶 下 沉 时 程 曲 线 图 7 ZK16+420 断 面 核 心 土 部 位 拱 顶 下 沉 时 程 曲 线<br />

由 图 6~ 图 7 可 知 拱 顶 下 沉 随 核 心 土 施 工 掌 子 面 的 变 化 可 明 显 区 分 为 以 下 三 个 阶 段 :(1) 当 核<br />

心 土 施 工 掌 子 面 距 测 线 处 距 离 约 2 倍 隧 道 开 挖 宽 度 时 , 此 时 拱 顶 下 沉 小 于 总 变 形 量 的 10%, 这 主 要<br />

是 由 于 工 作 面 的 开 挖 对 土 体 的 扰 动 而 引 起 的 变 形 , 处 于 小 变 形 阶 段 ;(2) 随 着 隧 道 的 开 挖 , 核 心 土<br />

掌 子 面 不 断 向 前 推 进 , 拱 顶 下 沉 速 率 加 速 增 强 。 该 阶 段 变 化 主 要 是 由 于 核 心 土 掌 子 面 开 挖 导 致 地 应<br />

力 场 及 边 界 条 件 发 生 改 变 , 对 覆 土 土 体 产 生 较 大 扰 动 , 从 而 引 起 应 力 场 的 重 分 布 。 由 麒 麟 山 隧 道 典<br />

型 断 面 观 测 数 据 可 见 , 当 掌 子 面 通 过 ZK16+465, 该 断 面 两 侧 拱 顶 下 沉 量 最 大 值 达 到 约 4.2mm, 各 测<br />

点 下 沉 量 占 总 下 沉 量 的 60% 以 上 , 在 随 后 的 1~1.5 倍 的 隧 道 开 挖 宽 度 范 围 内 时 变 化 仍 然 较 大 , 约 为<br />

总 下 沉 量 的 30% 左 右 , 整 个 该 阶 段 属 于 变 形 急 速 增 大 阶 段 ;(3) 当 开 挖 掌 子 面 离 测 点 在 1.5 倍 隧 道<br />

开 挖 宽 度 以 后 , 变 形 速 率 开 始 减 缓 , 变 形 量 增 长 缓 慢 直 至 趋 于 稳 定 , 此 阶 段 变 形 量 占 总 变 形 的 10<br />

% 左 右 , 该 阶 段 属 于 缓 慢 变 形 逐 渐 趋 于 稳 定 阶 段 。<br />

5.3. 水 平 收 敛 监 测 分 析<br />

各 断 面 洞 周 水 平 收 敛 测 线 布 置 如 图 2 所 示 , 左 线 Ⅴ 级 围 岩 周 边 收 敛 量 测 结 果 见 表 3。ZK16+465<br />

断 面 水 平 收 敛 位 移 时 程 曲 线 如 图 8 所 示 。<br />

由 于 隧 道 洞 口 浅 埋 段 岩 性 较 差 , 且 施 工 开 挖 复 杂 、 扰 动 大 , 使 得 收 敛 变 形 相 对 较 大 , 量 测 的 最<br />

大 水 平 收 敛 位 移 为 12.67mm , 但 相 对 收 敛 均 小 于 0.2%, 洞 身 段 收 敛 测 线 变 形 总 体 来 看 相 对 较 小 ,<br />

平 均 为 9.67mm 且 收 敛 稳 定 时 间 较 快 。<br />

根 据 表 3 周 边 收 敛 统 计 结 果 , 发 现 断 面 ZK16+465 和 ZK16+477 的 周 边 收 敛 值 较 大 , 最 大 值 为<br />

14.66mm, 平 均 值 为 13.13mm, 而 现 场 实 际 情 况 表 明 这 两 个 断 面 的 上 覆 土 是 最 厚 的 , 故 可 以 认 为 变 形<br />

主 要 是 由 山 体 两 侧 围 岩 水 平 挤 压 引 起 的 。 从 表 3 发 现 , 周 边 收 敛 的 初 期 变 形 速 率 相 对 来 说 较 快 , 初<br />

期 支 护 及 时 施 作 以 后 , 变 形 速 率 得 到 有 效 抑 制 , 从 而 避 免 了 围 岩 破 坏 , 使 施 工 得 以 安 全 、 顺 利 进 行 。<br />

从 图 8 可 知 , 经 过 2 个 月 左 右 的 监 测 时 间 , 该 断 面 水 平 收 敛 基 本 稳 定 ,2010 年 9 月 27 日 后 有<br />

一 段 非 常 明 显 的 突 变 , 主 要 是 由 于 那 段 时 间 在 拆 核 心 土 部 分 的 横 竖 撑 引 起 的 , 与 实 际 情 况 相 符 。<br />

表 3 左 线 Ⅴ 围 岩 周 边 收 敛 量 测 结 果<br />

测 点 位 置 ( 左 线 ) 围 岩 级 别 观 测 天 数 /d 收 敛 时 间 /d 初 期 变 形 速 率 /(mm.d -1 ) 最 大 收 敛 /mm<br />

ZK16+422 Ⅴ 63 35 0.70 6.53<br />

ZK16+442 Ⅴ 63 34 0.87 6.78<br />

ZK16+447 Ⅴ 57 28 0.49 5.64<br />

ZK16+465 Ⅴ 96 50 0.28 14.66<br />

ZK16+477 Ⅴ 92 40 0.15 11.60<br />

ZK16+492 Ⅴ 92 63 0.13 10.25<br />

-379-


水 平 收 敛 位 移 (mm)<br />

12.00<br />

10.00<br />

8.00<br />

6.00<br />

4.00<br />

2.00<br />

0.00<br />

边 墙<br />

拱 脚<br />

水 平 收 敛 位 移 (mm)<br />

16.00<br />

14.00<br />

12.00<br />

10.00<br />

8.00<br />

6.00<br />

4.00<br />

2.00<br />

0.00<br />

边 墙<br />

拱 脚<br />

2010/6/24<br />

2010/7/1<br />

2010/7/8<br />

2010/7/15<br />

2010/7/22<br />

2010/7/29<br />

2010/8/5<br />

2010/8/12<br />

2010/8/19<br />

2010/8/26<br />

2010/9/2<br />

2010/9/9<br />

2010/9/16<br />

2010/9/23<br />

2010/9/30<br />

2010/10/7<br />

2010/6/24<br />

2010/7/1<br />

2010/7/8<br />

2010/7/15<br />

2010/7/22<br />

2010/7/29<br />

2010/8/5<br />

2010/8/12<br />

2010/8/19<br />

2010/8/26<br />

2010/9/2<br />

2010/9/9<br />

2010/9/16<br />

2010/9/23<br />

2010/9/30<br />

2010/10/7<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

(a) 左 导 洞<br />

图 8 ZK16+465 水 平 收 敛 位 移 时 程 曲 线<br />

(b) 右 导 洞<br />

5.4. 围 岩 内 部 位 移 监 测 分 析<br />

在 隧 道 洞 内 同 一 断 面 埋 设 多 点 位 移 计 , 对 开 挖 过 程 中 洞 周 围 岩 及 左 、 右 线 间 中 岩 柱 的 位 移 实 时<br />

监 测 。 本 文 以 ZK16+430 断 面 左 右 侧 导 洞 中 台 阶 位 置 各 一 套 多 点 位 移 计 为 例 ( 其 中 左 导 洞 的 位 移 计<br />

对 洞 周 围 岩 位 移 进 行 研 究 , 右 导 洞 的 多 点 位 移 计 对 左 右 线 间 的 中 岩 柱 位 移 进 行 研 究 ), 分 析 围 岩 内<br />

部 位 移 规 律 , 多 点 位 移 计 所 测 位 移 时 程 曲 线 如 图 9 所 示 。<br />

位 移 (mm)<br />

7.000<br />

6.000<br />

5.000<br />

4.000<br />

3.000<br />

2.000<br />

1.000<br />

0.000<br />

2010/5/10<br />

1.43m#<br />

2.43m#<br />

3.43m#<br />

2010/5/17<br />

2010/5/24<br />

2010/5/31<br />

2010/6/7<br />

2010/6/14<br />

2010/6/21<br />

2010/6/28<br />

2010/7/5<br />

2010/7/12<br />

2010/7/19<br />

位 移 (mm)<br />

5.000<br />

4.000<br />

3.000<br />

2.000<br />

1.000<br />

0.000<br />

2010/5/10<br />

1.73m#<br />

2.73m#<br />

3.73m#<br />

2010/5/17<br />

2010/5/24<br />

2010/5/31<br />

2010/6/7<br />

2010/6/14<br />

2010/6/21<br />

2010/6/28<br />

2010/7/5<br />

2010/7/12<br />

2010/7/19<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

监 测 时 间 ( 年 / 月 / 日 )<br />

(a) 右 导 洞<br />

(b) 左 导 洞<br />

图 9 ZK16+430 断 面 围 岩 位 移 ( 多 点 位 移 计 测 ) 时 程 曲 线 图<br />

由 图 9 可 知 , 隧 道 开 挖 初 期 , 围 岩 内 部 各 测 点 位 移 很 小 ; 当 洞 内 不 同 开 挖 工 序 的 掌 子 面 到 达 量<br />

测 断 面 时 , 由 其 引 起 围 岩 松 动 位 移 。 这 两 套 多 点 位 移 计 围 岩 时 程 曲 线 存 在 两 处 明 显 突 变 , 一 处 是 从<br />

2010 年 5 月 12 日 至 5 月 20 日 这 段 时 间 发 生 的 , 该 段 时 间 内 左 右 侧 导 洞 中 台 阶 错 开 15m 的 距 离 不<br />

断 向 前 推 进 , 变 形 是 由 各 侧 导 洞 中 台 阶 开 挖 时 土 体 受 扰 动 , 从 而 导 致 应 力 场 重 分 布 引 起 ; 当 中 台 阶<br />

开 到 离 起 始 下 台 阶 大 概 30m 的 时 候 (5 月 20 日 ), 继 续 开 挖 中 台 阶 和 回 头 开 挖 下 台 阶 并 在 下 台 阶 开<br />

挖 完 后 及 时 施 作 仰 拱 , 所 以 第 二 处 突 变 是 由 下 台 阶 和 中 台 阶 开 挖 综 合 引 起 的 , 但 主 要 还 是 由 下 台 阶<br />

的 开 挖 引 起 。 另 外 从 稳 定 时 间 和 空 间 上 看 , 当 仰 拱 开 挖 完 毕 时 , 测 点 位 移 达 最 大 位 移 的 90% 左 右 ,<br />

而 当 仰 拱 面 通 过 量 测 断 面 大 约 40m 时 , 各 点 位 移 已 基 本 达 到 稳 定 。<br />

分 析 图 9 还 可 以 得 知 , 离 洞 周 壁 面 最 近 测 点 位 移 最 大 , 离 洞 周 越 远 位 移 越 小 , 即 同 一 位 置 测 点<br />

埋 设 越 深 测 点 位 移 越 小 ; 同 一 断 面 , 中 岩 柱 围 岩 位 移 明 显 大 于 洞 周 围 岩 位 移 , 故 加 强 中 岩 柱 围 岩 位<br />

移 的 监 测 至 关 重 要 。<br />

六 、 变 形 控 制 措 施<br />

隧 道 在 施 工 过 程 中 , 侧 导 洞 和 核 心 土 均 发 生 过 较 大 的 变 形 , 尤 其 在 开 挖 中 部 核 心 土 的 时 候 , 左<br />

右 侧 导 洞 所 加 部 分 横 撑 发 生 了 较 大 的 变 形 , 个 别 横 撑 受 力 已 超 过 设 计 允 许 值 。 另 外 在 隧 道 的 进 出 口<br />

-380-


段 , 由 于 围 岩 地 质 条 件 较 差 , 地 表 沉 降 及 洞 内 支 护 变 形 都 较 大 。 通 过 变 形 监 测 反 馈 的 信 息 , 施 工 中<br />

采 取 了 以 下 措 施 达 到 控 制 大 变 形 的 效 果 :<br />

( 1 ) 针 对 进 出 口 段 导 洞 开 挖 过 程 中 变 形 较 大 的 情 况 , 在 洞 口 施 作 大 管 棚 进 行 预 支 护 , 对 洞 周<br />

围 岩 进 行 系 统 径 向 注 浆 加 固 , 再 用 超 前 小 导 管 加 固 工 作 面 前 方 围 岩 , 以 改 善 支 护 的 受 力 条 件 , 限<br />

制 进 出 口 段 过 大 变 形 。<br />

( 2 ) 严 格 控 制 每 循 环 进 尺 和 台 阶 长 度 , 施 工 中 尽 量 减 少 对 围 岩 的 扰 动 , 在 洞 口 段 开 挖 方 式 以 机<br />

械 ( 炮 机 ) 及 人 工 ( 风 镐 ) 为 主 , 辅 以 微 震 、 松 动 爆 破 , 洞 身 段 采 用 微 震 动 “ 光 面 爆 破 ” 技 术 , 微<br />

差 爆 破 的 安 全 速 度 控 制 在 2.0cm/s 以 下 , 以 最 大 限 度 保 护 周 边 岩 体 的 完 整 性 和 靠 近 隧 道 左 线 高 压 电<br />

塔 的 安 全 。<br />

(3) 采 用 双 侧 壁 导 坑 法 分 部 开 挖 。 导 洞 及 核 心 土 开 挖 采 用 上 中 下 台 阶 法 施 工 , 相 邻 洞 室 掌 子<br />

面 开 挖 错 开 15m 以 上 。 开 挖 成 形 后 及 时 进 行 初 期 支 护 , 紧 扣 工 序 衔 接 , 尽 早 施 作 仰 拱 , 形 成 封 闭 环 。<br />

七 、 结 语<br />

该 隧 道 断 面 超 大 , 跨 度 超 大 、 双 孔 净 距 非 常 小 、 扁 平 率 值 很 小 , 到 目 前 为 止 , 隧 道 两 次 衬 砌 已<br />

施 工 完 成 , 稳 定 性 非 常 好 , 也 是 一 个 零 事 故 的 安 全 工 程 , 可 以 得 出 以 下 几 点 :<br />

(1) 从 地 表 下 沉 和 洞 周 变 形 监 测 结 果 看 ,V 级 围 岩 隧 道 采 用 双 侧 壁 导 坑 多 台 阶 法 分 部 开 挖 是 可<br />

行 而 有 效 的 , 尽 管 其 施 工 工 序 较 复 杂 , 但 可 以 较 好 地 控 制 围 岩 变 形 。<br />

(2) 从 监 测 结 果 来 看 , 及 时 有 效 的 初 期 支 护 对 整 个 隧 道 的 开 挖 至 关 重 要 ; 同 时 , 隧 道 每 个 台<br />

阶 开 挖 时 要 及 时 进 行 支 护 , 以 确 保 上 一 台 阶 初 期 支 护 拱 脚 的 稳 定 , 防 止 产 生 滑 移 , 危 害 隧 道 整 体 稳<br />

定 性 。<br />

(3) 对 于 大 跨 小 间 距 隧 道 , 尽 早 施 作 仰 拱 形 成 封 闭 环 对 于 改 善 结 构 受 力 和 抑 制 隧 道 变 形 非 常<br />

有 效 。 当 隧 道 进 出 口 围 岩 岩 性 很 差 时 , 在 洞 口 施 作 大 管 棚 进 行 预 支 护 , 对 洞 周 围 岩 进 行 系 统 径 向 注<br />

浆 加 固 , 再 用 超 前 小 导 管 加 固 工 作 面 前 方 围 岩 能 够 有 效 抑 制 进 出 口 段 过 大 变 形 。<br />

(4) 采 用 新 奥 法 的 思 想 , 进 行 及 时 变 形 监 测 并 反 馈 到 施 工 实 施 过 程 中 且 加 以 必 要 的 工 艺 调 整<br />

再 次 被 证 明 是 十 分 必 要 和 确 有 成 效 的 。<br />

参 考 文 献<br />

[1] 王 鑫 , 宋 战 平 , 冯 志 焱 等 . 芙 蓉 山 公 路 隧 道 地 表 沉 降 监 测 技 术 及 应 用 . 水 利 与 建 筑 工 程 学 报 , 2008, 6(3): 93-112.<br />

[2] 瞿 万 波 , 刘 新 荣 , 黄 瑞 金 等 . 浅 埋 大 跨 洞 桩 隧 道 变 形 监 测 与 控 制 分 析 . 土 木 建 筑 与 环 境 工 程 , 2009, 31(1): 38-43.<br />

[3] 王 者 超 , 李 术 才 , 陈 卫 忠 等 . 分 岔 隧 道 变 形 监 测 与 施 工 对 策 研 究 . 岩 土 力 学 , 2007, 28(4): 785-789.<br />

[4] 夏 才 初 , 龚 建 伍 , 唐 颖 等 . 大 断 面 小 净 距 公 路 隧 道 现 场 监 测 分 析 研 究 . 岩 石 力 学 与 工 程 学 报 , 2007, 26(1): 44-50.<br />

[5] 周 丁 恒 , 曹 力 桥 , 曲 海 锋 等 . 不 同 围 岩 情 况 下 特 大 断 面 公 路 隧 道 施 工 变 形 监 测 与 控 制 . 岩 石 力 学 与 工 程 学 报 , 2009,<br />

28(12): 2510-2519.<br />

-381-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

SCALING EARTHQUAKE RECORDS FOR SEISMIC PERFORMANCE<br />

ASSESSMENT OF BUILDINGS<br />

Yin-Nan Huang 1<br />

1<br />

Department of Civil Engineering,<br />

National Taiwan <strong>University</strong>, Taipei, Taiwan.<br />

ABSTRACT<br />

<strong>The</strong> procedures used to select and scale earthquake ground motions for nonlinear response-history analysis<br />

directly affect the distributions of response of structural components and nonstructural components (secondary<br />

systems). <strong>The</strong> paper presents the results of a study of three scaling methods: 1) geometric mean scaling of pairs<br />

of ground motions, 2) spectrum matching of ground-motion components and 3) first-mode-period scaling to a<br />

target spectral acceleration. Data were developed by nonlinear response-history analysis of a large family of<br />

nonlinear SDOF oscillators that could represent first-mode-dominated fixed-base and base-isolated structures.<br />

<strong>The</strong> results show that Method 1 preserves the irregular spectral shape of recorded ground motion and retains<br />

some dispersion in the spectral demand. Method 2 underestimates the median seismic demand in nonlinear<br />

systems with ductility greater than 2.5 and cannot capture the dispersion in the structural response. Method 3<br />

provides unbiased estimates of median responses of nonlinear systems, produces dispersions of the same order<br />

as or greater than those of Method 1 for nonlinear systems with ductility greater than 3, but cannot capture the<br />

dispersion in response of elastic and near-elastic systems.<br />

KEYWORDS<br />

Ground motion, response spectra, scaling, seismic design, time-series analysis.<br />

INTRODUCTION<br />

<strong>The</strong> first-generation tools for Performance-Based Earthquake Engineering (PBEE), such as those documented in<br />

FEMA 273 and 274 (FEMA 1997), FEMA 356 (FEMA 2000b) and ASCE/SEI 41-06 (ASCE 2006), use a<br />

deterministic approach to assess structural performance. FEMA 350 (FEMA 2000a), which was drafted as part<br />

of the SAC Steel Project, extended the first generation tools and introduced probabilistic seismic assessment<br />

procedures for buildings. <strong>The</strong> ATC-58 project is developing the second generation tools and guidelines for<br />

performance-based seismic design and assessment using a probability framework, which incorporates the<br />

inherent uncertainty and variability in seismic hazard, structural and non-structural responses, damage states and<br />

economic and casualty losses (ATC 2008).<br />

One key issue in any seismic performance assessment procedure is the scaling of ground motions for nonlinear<br />

response-history analysis, which should preserve the distribution (e.g., both median and dispersion) in the<br />

earthquake shaking for the selected characterization of the hazard (e.g., a spectrum, a magnitude-distance pair,<br />

or one or a family of hazard curves) for the site of interest so that one can properly capture the distributions of<br />

structural responses and demands on secondary systems.<br />

This paper seeks to identify the influence of different scaling procedures on the distributions of structural<br />

responses and demands on nonstructural components and secondary systems. Results of analysis using three<br />

scaling procedures are presented: a spectrum-matching procedure and two amplitude-scaling procedures.<br />

Bilinear single-degree-of-freedom (SDOF) oscillators are used to represent fixed-based and base-isolated<br />

structures. Both median responses and dispersions are reported. <strong>The</strong> advantages and disadvantages of the<br />

procedures are identified.<br />

NUMERICAL MODELS AND SEED GROUND MOTIONS<br />

A large number of bilinear SDOF models were analyzed for this study. Yield strengths were set at infinity,<br />

0.40W, 0.20W, 0.10W, and 0.06W to represent, albeit simplistically, conventional and isolated (0.06W)<br />

construction, where W is the reactive weight of the structure. In this paper, only the results for yield strengths of<br />

-382-


0.40W, 0.20W and 0.06W are presented. Results for other yield strengths are available in Huang et al. (2008).<br />

For the oscillators with yield strength of 0.40W, the elastic period ranged between 0.05 second and 2 seconds;<br />

for the oscillator with yield strength of 0.06W, the post-yield (isolated) period ranged between 2 and 4 seconds.<br />

<strong>The</strong> ratio of post-yield to elastic stiffness was set to 0.1 for all oscillators.<br />

Response-history analysis was performed using the 25 pairs of near-fault (NF) earthquake histories listed in<br />

Table 1. (Studies were also performed for far-field motions; see Huang et al. (2011) for details.) <strong>The</strong> moment<br />

magnitude for the ground motions in Table 1 (the NF bin) ranges between 6.2 and 7.3 and the site-to-source<br />

distance ranges between 1 and 17.5 km. <strong>The</strong>re are a limited number of U.S. near-fault records and so records<br />

from Japan, Turkey, Chile, and Taiwan were used to augment the U.S. dataset. <strong>The</strong> first ten pairs of ground<br />

motions in the NF set were developed in the SAC Steel Project for a firm-soil site in Los Angeles and a 2%<br />

probability of exceedance in 50 years (Somerville et al. 1997).<br />

Table 1 Seed ground motions<br />

esignation Event Station M r<br />

NF1, NF2 Kobe 1995 jma 6.9 3.4<br />

NF3, NF4 Loma Prieta 1989 lgpc 7.0 3.5<br />

NF5, NF6 Northridge 1994 rrs 6.7 7.5<br />

NF7, NF8 Northridge 1994 sylm 6.7 6.4<br />

NF9, NF10 Tabas 1974 tab 7.4 1.2<br />

NF11, NF12 Elysian Park 1 (simulated) st04 7.1 17.5<br />

NF13, NF14 Elysian Park 2 (simulated) st10 7.1 10.7<br />

NF15, NF16 Elysian Park 3 (simulated) st13 7.1 11.2<br />

NF17, NF18 Palos Verdes 1 (simulated) st03 7.1 1.5<br />

NF19, NF20 Palos Verdes 2 (simulated) st06 7.1 1.5<br />

NF21, NF22 Cape Mendocino 04/25/92 18:06 89156 Petrolia 7.1 9.5<br />

NF23, NF24 Chi-Chi 09/20/99 TCU053 7.6 6.7<br />

NF25, NF26 Chi-Chi 09/20/99 TCU056 7.6 11.1<br />

NF27, NF28 Chi-Chi 09/20/99 TCU068 7.6 1.1<br />

NF29, NF30 Chi-Chi 09/20/99 TCU101 7.6 11.1<br />

NF31, NF32 Chi-Chi 09/20/99 TCUWGK 7.6 11.1<br />

NF33, NF34 Duzce 11/12/99 Duzce 7.1 8.2<br />

NF35, NF36 Erzinkan 03/13/92 17:19 95 Erzinkan 6.9 2.0<br />

NF37, NF38 Imperial Valley 10/15/79 23:16 5057 El Centro Array #3 6.5 9.3<br />

NF39, NF40 Imperial Valley 10/15/79 23:16 952 El Centro Array #5 6.5 1<br />

NF41, NF42 Imperial Valley 10/15/79 23:16 942 El Centro Array #6 6.5 1<br />

NF43, NF44 Kobe 01/16/95 20:46 Takarazu 6.9 1.2<br />

NF45, NF46 Morgan Hill 04/24/84 04:24 57191 Halls Valley 6.2 3.4<br />

NF47, NF48 Northridge 1/17/94 12:31 24279 Newhall 6.7 7.1<br />

NF49, NF50 Northridge 1/17/94 12:31 0637 Sepulveda VA 6.7 8.9<br />

-383-


THREE PROCEDURES FOR SCALING GROUND MOTIONS<br />

Method 1: Geometric-mean scaling method<br />

Method 1 involves amplitude scaling a pair of seed motions by a single factor to minimize the sum of the<br />

squared errors between the target spectral values and the geometric mean (square root of the product, hereafter<br />

termed geomean) of the spectral ordinates for the pair at periods of 0.3, 0.6, 1, 2 and 4 seconds. <strong>The</strong> resultant<br />

amplitude-scaled NF motions are denoted herein as Bin 1a motions. <strong>The</strong> solid dots of Figure 20a show the target<br />

spectral acceleration for Bin 1a motions, which was based on that developed by Somerville et al. (1997) from<br />

the USGS hazard maps for a 2% probability of exceedance in 50 years (denoted hereafter as 2/50) and a site in<br />

Los Angeles.<br />

This scaling procedure preserves some dispersion in the spectral demand at a given period and the irregular<br />

spectral shapes of the seed ground motions. Figure 20a presents the 16th, 50 th (median) and 84th percentiles of<br />

elastic spectral acceleration for Bin 1a. Figure 20b presents logarithmic standard deviation ( β ) of elastic<br />

spectral acceleration for Bin 1a. <strong>The</strong> results presented in Figure 20 were estimated using Eqs. (8) through (11)<br />

assuming that the spectral acceleration is lognormally distributed:<br />

n<br />

⎛1<br />

⎞<br />

θ = exp⎜ ∑ ln yi<br />

⎟<br />

(8)<br />

⎝n<br />

i=<br />

1 ⎠<br />

1<br />

β = −<br />

n<br />

∑ n<br />

2<br />

(ln yi<br />

ln θ )<br />

(9)<br />

−1<br />

i=<br />

1<br />

y16th<br />

y84th<br />

−<br />

= θ ⋅ e β<br />

(10)<br />

= θ ⋅ e β<br />

(11)<br />

where θ , y<br />

16th<br />

and y 84th<br />

are the 50th, 16th and 84th percentile values identified in Figure 20, respectively;<br />

y<br />

i<br />

is the spectral ordinate for the ith ground motion at a given period; n is the total number of ground<br />

motions, which is 50 in this case.<br />

Spectral acceleration, Sa (g)<br />

3<br />

2<br />

1<br />

Bin 1a<br />

Bin 1b<br />

target<br />

0<br />

0 1 2 3 4<br />

Period (sec)<br />

β of Sa<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

Bin 1a<br />

Bin 1b<br />

0<br />

0 1 2 3 4<br />

Period (sec)<br />

a. Sa, Bins 1a and 1b b. β , Bins 1a and 1b<br />

Figure 20 Elastic acceleration spectra (84 th , 50 th and 16 th percentiles) and β for Bins 1a and 1b<br />

In Figure 20a, differences of about 0.1 and 0.3 g, respectively, between the median (50th percentile) ordinate<br />

and target spectral ordinate are observed at a period of 0.6 second. This is because the shape of the median<br />

spectrum resulting from geomean scaling is independent of the target spectral ordinates. To demonstrate the<br />

insensitivity of the shape of the resultant median spectrum for Method 1 on target spectral ordinates, various<br />

geomean scalings were performed and presented in Figure 21.<br />

θ of Sa (g)<br />

2<br />

1.5<br />

1<br />

0.5<br />

Scale to 1.61 g<br />

at T=0.3 sec<br />

β of Sa<br />

1<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0 1 2 3 4<br />

Period (sec)<br />

0<br />

0 1 2 3 4<br />

Period (sec)<br />

a. Median, θ b. Dispersion, β<br />

-384-


M original data<br />

S scaled data with target at T=0.3 sec only<br />

S scaled data with target at T=0.3 and 2.0 secs<br />

S scaled data with target at T=0.3, 0.6, 1, 2 and 4 sec<br />

target spectrum<br />

Figure 21 <strong>The</strong> impact of the geomean scaling method for the θ and β of spectral acceleration<br />

<strong>The</strong> records in Bin 1a were scaled with target values at a) T = 0.3 second only, b) T = 0.3 and 2 seconds, and c)<br />

T = 0.3, 0.6, 1, 2 and 4 seconds. <strong>The</strong> median spectra for the original (or pre-scaled) motions and the three sets<br />

of the amplitude-scaled motions were normalized to the target spectrum at a period of 0.3 second and presented<br />

in Figure 21a. <strong>The</strong> dispersions β in the spectral acceleration for the four sets are presented in Figure 21b.<br />

Some key observations are:<br />

a. <strong>The</strong> shape of the median spectrum after geomean scaling of pairs of ground motions is independent of the<br />

target spectral values and is identical to the shape of the pre-scaled median spectrum. If a UHS is used for<br />

design where the period range of interest is broad (for example, if the structure has significant higher<br />

mode effects that should be considered in the analysis) and different magnitude-distance pairs dominate<br />

the UHS across the subject period range, it will be difficult to select a set of ground motions whose<br />

median spectrum closely matches the target spectrum.<br />

b. <strong>The</strong> dispersions in the spectral acceleration for the pre-scaled (seed) motions were reduced at periods<br />

smaller than 3 seconds for all three sets of target spectral ordinates of Figure 21. Whether sufficient<br />

dispersion is preserved by the geomean scaling method depends on the dispersion in the seismic hazard of<br />

interest, which can be the dispersion in the spectral demand predicted by an attenuation relationship for a<br />

magnitude-distance pair, or the epistemic or model uncertainty in the spectral acceleration at an annual<br />

frequency of exceedance associated with a family of hazard curves. <strong>The</strong> geomean scaling method does not<br />

guarantee that the dispersion preserved in the spectral ordinates is sufficient.<br />

Figure 22 presents the 16th, 50th and 84th percentiles and β of peak displacement for the oscillators<br />

introduced earlier and for Bin 1a. <strong>The</strong>se statistical representations were computed using Eqs. (8) through (11)<br />

except that y<br />

i<br />

is the peak displacement for a given oscillator subjected to the ith ground motion. <strong>The</strong> dashed<br />

line in panels a, c and e of both figures identifies the yield displacement for each oscillator. <strong>The</strong> value of β for<br />

the displacements shown in Figure 22 varies mainly between 0.3 and 0.7.<br />

<strong>The</strong> studies for Methods 2 and 3 use the median spectrum of Figure 20a as the target spectrum to scale the<br />

ground motions in Bin 1a but have different treatments for the dispersion preserved in the spectral ordinates of<br />

Figure 20a. <strong>The</strong> results of response-history analysis presented in Figure 22 using the 50 motions in Bins 1a are<br />

used to benchmark the results for Methods 2 and 3 to investigate the impact of scaling procedures on the<br />

distribution of structural responses. In Figure 22, the periods presented in the X axes in panels a through d are<br />

computed using the pre-yield stiffness and those in panels e and f are computed using the post-yield stiffness.<br />

-385-


Peak displacement (m)<br />

84, 50, 16 percentiles<br />

1.2<br />

0.8<br />

0.4<br />

LBin 1a<br />

LBin 1b<br />

Lyield displ.<br />

0<br />

0 0.4 0.8 1.2 1.6 2<br />

Period (sec)<br />

β<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0 0.4 0.8 1.2 1.6 2<br />

Period (sec)<br />

a. Peak displ., Fy<br />

= 40% W<br />

b. β , Fy<br />

= 40% W<br />

1.2<br />

0.8<br />

Peak displacement (m)<br />

84 , 5 0, 16 p erce ntile s<br />

0.8<br />

0.4<br />

0<br />

0 0.4 0.8 1.2 1.6 2<br />

Period (sec)<br />

β<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0 0.4 0.8 1.2 1.6 2<br />

Period (sec)<br />

c. Peak displ., Fy<br />

= 20% W<br />

d. β , Fy<br />

= 20% W<br />

1.2<br />

0.8<br />

Peak displacement (m)<br />

84, 50, 16 percentiles<br />

0.8<br />

0.4<br />

0<br />

2 2.4 2.8 3.2 3.6 4<br />

Period (sec)<br />

β<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

2 2.4 2.8 3.2 3.6 4<br />

Period (sec)<br />

e. Peak displ., F = 6% W<br />

f. β , F = 6% W<br />

y<br />

Figure 22 Median, 84th and 16th percentiles and β of peak displacement response for Bins 1a and 1b<br />

Method 2: Spectrum-Matching Method<br />

Spectrally matched ground motions are often used to compute seismic demands in structural framing systems.<br />

To judge the utility of spectrum-matched ground motions for predicting the response of nonlinear SDOF<br />

framing systems (such as base-isolated structures), the NF seed motions of Table 1 were modified to match the<br />

median spectrum of Bin 1a using the computer code RSPMATCH (Abrahamson 1998, Hancock et al. 2006).<br />

<strong>The</strong> bin of 50 spectrally matched motions is labeled as Bin 1b. <strong>The</strong> 16th, 50th and 84th percentiles of elastic<br />

spectral acceleration for Bin 1b are shown in Figure 20a. <strong>The</strong> median spectral accelerations for Bins 1a and 1b<br />

are virtually identical. Figure 20b presents the dispersion in the spectral accelerations for the two bins of ground<br />

motions. <strong>The</strong> dispersion in the spectral accelerations for Bin 1b is negligible.<br />

<strong>The</strong> 16th, 50th and 84th percentiles and β of the inelastic peak displacement for Bin 1b are shown in Figure<br />

22 together with those for Bin 1a. <strong>The</strong> results indicate that the use of spectrally matched ground motions<br />

underestimate the median peak displacement demand in highly nonlinear SDOF systems and cannot capture the<br />

dispersion in the displacement response. For Fy<br />

= 0.06W<br />

(see Figure 22e), the median responses are<br />

underestimated by 20%. This observation is similar to that of Bazzurro and Luco (2006) for NF motions.<br />

Earthquake ground motions that are spectrally matched to a target median spectrum should not be used for risk<br />

computations of nonlinear SDOF systems and first-mode-dominated structures because a) the median<br />

displacement response will be underestimated, and b) the dispersion in the displacement response is<br />

underestimated by a wide margin.<br />

y<br />

-386-


Method 3: Sa( T<br />

1)<br />

Scaling Method<br />

Shome et al. (1998) proposed a method for scaling ground motions that involves amplitude scaling to a specified<br />

spectral acceleration at the first mode period of the structure. Each of the 50 NF seed motions of Table 1 was<br />

scaled to match the median elastic spectral acceleration of Bin 1a at each of many periods in the range of 0.05 to<br />

4 seconds, where the oscillator period is based on the pre-yield (elastic) stiffness (including the oscillator with<br />

yield strength equal to 0.06W). <strong>The</strong> bin of amplitude-scaled motions is labeled as Bin 1c. Figure 23 presents the<br />

16th, 50th, 84th percentiles and β of peak displacement for Bins 1a and 1c. Some key observations are:<br />

a. Method 3 produces unbiased estimates of the median displacement response, relative to the geomean<br />

scaling method (Method 1), even for the weakest oscillator ( Fy<br />

= 0.06W<br />

) for which the post-yield<br />

(stiffness) period dictates the displacement response.<br />

b. <strong>The</strong> dispersion ( β ) in the peak displacement is not preserved for elastic or near-elastic systems: an<br />

expected result given that the dispersion in the ground motion was eliminated at the natural period of the<br />

oscillator. <strong>The</strong> dispersion is smaller for the Bin 1c motions than for the Bin 1a motions for oscillators with<br />

a small (displacement) ductility ratio and greater for the oscillators with a large ductility ratio. (For<br />

oscillators with F y<br />

= 0.40W and 0.20W, the ductility ratio decreases as the period of the oscillator<br />

increases.)<br />

Peak displacement (m)<br />

84, 50, 16 percentiles<br />

1.2<br />

0.8<br />

0.4<br />

LBin 1a<br />

LBin 1c<br />

Lyield displ.<br />

0<br />

0 0.4 0.8 1.2 1.6 2<br />

Period (sec)<br />

β<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0 0.4 0.8 1.2 1.6 2<br />

Period (sec)<br />

a. Peak displ., Fy<br />

= 40% W<br />

b. β , Fy<br />

= 40% W<br />

1.2<br />

0.8<br />

Peak displacement (m)<br />

84, 50, 16 percentiles<br />

0.8<br />

0.4<br />

0<br />

0 0.4 0.8 1.2 1.6 2<br />

Period (sec)<br />

β<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

0 0.4 0.8 1.2 1.6 2<br />

Period (sec)<br />

c. Peak displ., Fy<br />

= 20% W<br />

d. β , Fy<br />

= 20% W<br />

1.2<br />

0.8<br />

Peak displacement (m)<br />

84, 50, 16 percentiles<br />

0.8<br />

0.4<br />

0<br />

2 2.4 2.8 3.2 3.6 4<br />

Period (sec)<br />

β<br />

0.6<br />

0.4<br />

0.2<br />

0<br />

2 2.4 2.8 3.2 3.6 4<br />

Period (sec)<br />

o. β of Sd, F y =6%W<br />

e. Peak displ., F = 6% W<br />

f. β , F = 6% W<br />

y<br />

Figure 23 Median, 84th and 16th percentiles and β of peak displacement response for Bins 1a and 1c<br />

y<br />

-387-


CONCLUSIONS<br />

Performance-based design and assessment requires reasonable estimates to be made of the distribution (both<br />

median and dispersion) of nonlinear structural responses. Studies were performed with the goals of establishing<br />

an optimal ground motion scaling procedure for uni-directional response-history analysis and providing a<br />

technical basis for future work on selecting and scaling pairs of ground motions for bi-directional response<br />

analysis. Three ground-motion scaling methods were evaluated using nonlinear single-degree of freedom<br />

models with yield strengths ranging from 0.06W to infinity and periods ranging from 0.01 to 4 seconds. Two<br />

sets of 25 pairs of ground motions were used in the study: one set of near-fault motions and the other set of<br />

far-field motions. Only the results for the near-fault motions are presented in this paper.<br />

<strong>The</strong> key conclusions of the studies are summarized below. <strong>The</strong>se conclusions are limited to<br />

first-mode-dominated buildings with minor to moderate inelastic deformation and typical base-isolated<br />

structures. For structures in which multiple modes contribute significantly to the drift and acceleration response<br />

due to earthquake shaking (e.g., tall buildings), alternate procedures may be required.<br />

1. Method 1 (geomean scaling) preserves the irregular spectral shapes of recorded ground motion and some<br />

dispersion in the spectral demand. <strong>The</strong> shape of the median spectrum for a bin of ground motions scaled<br />

by this method depends solely on the pre-scaled shape of the median spectrum for the bin. If a wide range<br />

of periods must be addressed for analysis, and multiple magnitude-distance pairs dominate the target<br />

spectrum at different periods across the range of interest, it will be difficult to select a bin of ground<br />

motions whose median spectrum closely matches the target spectrum. For such a circumstance, multiple<br />

bins of ground motions should be considered at different periods of interest. This conclusion applies to all<br />

amplitude-scaling methods.<br />

2. Method 2, spectrum-matching scaling, underestimates the median benchmark displacement demand (from<br />

the results of Method 1) in highly nonlinear SDOF systems and cannot capture the dispersion in the<br />

structural response because the scatter in the spectral ordinates is eliminated by the matching process.<br />

Earthquake ground motions that are spectrally matched to target median spectrum should not be used to<br />

characterize a distribution of seismic responses resulting from a distribution of spectral demand because<br />

the median displacement response will be underestimated for highly nonlinear systems and the dispersion<br />

in the displacement response will be underestimated by a wide margin for all systems, regardless of<br />

whether the response is linear or nonlinear.<br />

3. Method 3, Sa( T 1)<br />

scaling, provides unbiased estimates of median benchmark responses of nonlinear<br />

systems and produces dispersions of the same order as or greater than those of Method 1 for nonlinear<br />

systems with ductility greater than 3 because the first mode period does not necessarily dictate the<br />

response of such systems. However, the method cannot capture (and was not intended to capture) the<br />

benchmark dispersion in response of elastic and near-elastic systems.<br />

ACKNOWLEDGEMENTS<br />

This research is supported by the department of Civil Engineering, National Taiwan <strong>University</strong> and a grant<br />

(Code# ’09 R&D A01) from Cutting-edge Urban Development Program funded by Ministry of Land, Transport<br />

and Maritime Affairs of Korean government.<br />

REFERENCES<br />

Abrahamson, N.A. Non-stationary spectral matching program RSPMATCH. PG&E, Internal Report, 1998.<br />

American Society of Civil Engineers, ASCE. (2006). “Minimum design loads for buildings and other structures”,<br />

ASCE/SEI 7-05, American Society of Civil Engineers, Reston, Virginia.<br />

Applied Technology Council (ATC). (2008). “Guidelines for seismic performance assessment of buildings”.<br />

ATC-58 50% Draft, Applied Technology Council, Redwood City, California.<br />

Bazzurro, P. and Luco, N. “Do scaled and spectrum-matched near-source records produce biased nonlinear<br />

structural responses”, Proceedings, 8 th U.S. National Conference on Earthquake Engineering, San<br />

Francisco, California, 2006.<br />

Federal Emergency Management Agency. NEHRP recommended provisions for seismic regulations for new<br />

buildings and other structures. Rep. No. 273/274, FEMA, Washington, D.C. ,1997.<br />

Federal Emergency Management Agency. Recommended seismic design criteria for new steel moment-frame<br />

buildings. Rep. No. 350, FEMA, Washington, D.C., 2000.<br />

-388-


Federal Emergency Management Agency. “Prestandard and commentary for the seismic rehabilitation of<br />

buildings”, Rep. No. 356, FEMA, Washington, D.C., 2000.<br />

Hancock, J., Watson-Lamprey, J., Abrahamson, N. A., Bommer, J. J., Markatis, A., McCoyh, E. and Mendis, R.<br />

“An improved method of matching response spectra of recorded earthquake ground motion using wavelets”,<br />

Journal of Earthquake Engineering, 2006, 10(S1): 67-89.<br />

Huang, Y.-N., Whittaker, A.S. and Luco, N. “Performance assessment of conventional and base-isolated nuclear<br />

power plants for earthquake and blast loadings”, MCEER-08-0029, Multidisciplinary Center for<br />

Earthquake Engineering Research, State <strong>University</strong> of New York at Buffalo, Buffalo, NY., 2008.<br />

Huang, Y.-N., Whittaker, A.S., Luco, N. and Hamburger, R.O. “Scaling earthquake ground motions for<br />

performance-based assessment of buildings”, Journal of Structural Engineering, 2011, 137(3), 311-321.<br />

Shome, N., Cornell, C. A., Bazzurro, P., and Carballo, J. E. “Earthquakes, records, and nonlinear responses”,<br />

Earthquake Spectra, 1998, 14(3), 469-500.<br />

Somerville, P., Smith, N., Punyamurthula, S. and Sun, J. “Development of ground motion time histories for<br />

phase 2 of the FEMA/SAC steel project”, Report SAC/BD-97/04, SAC Joint Venture, Sacramento,<br />

California, 1997.<br />

-389-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

MODEL TESTS STUDY ON CAST-IN-PLACE X-SECTIONAL PILE<br />

G. Q. <strong>Kong</strong> 1 , H. L. Liu 2 and M. X. Zhang 3<br />

1<br />

Key Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Geotechnical<br />

Research Institute, Hohai <strong>University</strong>, Nanjing, Jiangsu, China.<br />

2<br />

Key Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Geotechnical<br />

Research Institute, Hohai <strong>University</strong>, Nanjing, Jiangsu, China. Email: hliuhhu@163.com<br />

3<br />

Key Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Geotechnical<br />

Research Institute, Hohai <strong>University</strong>, Nanjing, Jiangsu, China.<br />

ABSTRACT<br />

Cast-in-place X-sectional reinforced concrete pile (referred as to XCC pile), which is casted by steel mould with<br />

X-shaped cross section. <strong>The</strong> main steps contain sinking of mould, injection of concrete and conservation. This<br />

pile has the advantages like high capacity, economy and convenience for construction etc. It is therefore suitable<br />

for the soft clay ground treatment on highways and railways projects. Based on the large-scale model test<br />

system located on Hohai <strong>University</strong>, the field load static tests were carried out. <strong>The</strong> load-displacement curves<br />

and bearing capacities were analyzed and discussed. <strong>The</strong> result shows that the XCC pile can augment the<br />

utilization rate of section and the resistance capacities with the same amount of concrete. <strong>The</strong>se test results are<br />

useful for the practical engineering and analysis of XCC pile in the same areas.<br />

KEYWORDS<br />

XCC pile, load transfer, full scale model test, shape effect, bearing capacity.<br />

INTRODUCTION<br />

To cater to rapid economic growth, China has undergone a massive development in their transportation<br />

infrastructure in recent years. Many roads, rail ways, seaports, and airports have been constructed. <strong>The</strong> total<br />

length of highway increased from 271 km in 1989 to 75000 km in May, 2009. It is nearly the sum total of the<br />

highways in America, which was 80000 km in May, 2009. In an effort to develop economic pile systems for<br />

supporting embankments on soft clays, there have been some interesting research works that were focused on<br />

investigation of potential benefits of using special concrete pile sections (Lei et al. 2001; Liu et al. 2007a, 2007b;<br />

Chen et al. 2010). Compared with the commonly adopted circular pile sections, these special pile cross sections<br />

have been advanced with the idea of offering higher contact areas of the pile-soil interface and the direction<br />

dependent stronger lateral stiffness, for the same amount of concrete used as that of circular section piles. <strong>The</strong>re<br />

are no uniform rules to follow to calculate its bearing capacity, and its vertical bearing capacity is determined<br />

through the static load testing in the actual project. Optimization design method through studying on the section<br />

geometric characteristics of XCC pile was discussed (Liu et al. 2009). According to the small XCC pile static<br />

load model test and finite element analysis model, a simplified calculation method of the XCC pile based on the<br />

key factors parameter sensitivity analysis of XCC pile bearing capacity (Liu 2008). Full-scale test is a good<br />

research tool. It can overcome the size effect error caused by model test, and save a lot of cost and shorten the<br />

test cycle compared with the field test. Its arrangement of test equipment and test precision are more easily<br />

controlled. <strong>The</strong>refore, through comparative tests of the XCC pile and the circular pile in the same cross-section<br />

area and the circular pile in the same perimeter in this paper, the pile bearing mechanical characteristics and load<br />

transfer mechanism of the XCC pile were discussed.<br />

CONSTRUCTION METHOD<br />

Pictures of the XCC pile construction machine and the special flap pile shoe for preventing the soils from<br />

getting into the pile mold are shown in Figure 1. <strong>The</strong> construction procedures for XCC piles can be described as<br />

follows. First, the vibratory pile driver is connected to the X-section steel casing by means of a butt plate. Next,<br />

the X-section steel casing is connected with the flap pile shoe. <strong>The</strong>n, the vibratory driver drives the X-cross<br />

section steel casing into the desired elevation. After reaching the required penetration depth, concrete is then fed<br />

through the X-section steel casing inlet mouth. Finally, the vibrating driver extracts the casing to the ground.<br />

-390-


Thus, in-situ X-section concrete pile is formed. In order to improve the quality of pile, the pulling speed is<br />

0.6-1.5 m/min, the outside equivalent diameter of casing is between 0.4-2.0 m, and the pile length is between<br />

6-40 m.<br />

Figure 1 Physical diagram of XCC pile-driving equipment, pile mold, and flappable pile shoe<br />

FULL-SCALE MODEL TEST<br />

Model Trench<br />

Large-scale pile test model trench situated at Hohai <strong>University</strong>, China (shown in Figure 2). It is steel reinforced<br />

concrete structure. Model slot size is 4 m wide, 5 m long and 7 m high. With reaction frame and the above<br />

canopy height, total height is 12.3 m.<br />

Model Piles<br />

Figure 2 Full physical diagram of large-scale model test facility<br />

<strong>The</strong>re are three test model piles, including an XCC pile, a circular pile (numbered C 1 ) in the same cross-section<br />

area and a circular pile (numbered C 2 ) in the same perimeter. <strong>The</strong> specific parameters of test piles are shown in<br />

Table 2. Each test model pile concrete strength grade is C25 and its length is 5 m.<br />

-391-


Figure 3 Physical diagrams of pile head of XCC pile, and circular pile<br />

Table 2 Parameters of test piles<br />

Parameters XCC C 1 C 2<br />

Area /m 2 0.1425 0.1425 0.2462<br />

Perimeter /m 1.7604 1.3401 1.7604<br />

Square side length /m 0.5303 - -<br />

Arc space /m 0.11 - -<br />

Radian /° 90 - -<br />

Diameter /m - 0.426 0.560<br />

Pile length /m 5.0 5.0 5.0<br />

Model Soils<br />

Test soil is divided into two layers. <strong>The</strong> upper 2.4 m layer is sand. In order to ensure density coherency, fall sand<br />

height should be controlled and each layer should be flat compaction. <strong>The</strong> lower 3.9 m layer is salty clay. <strong>The</strong><br />

soil is approximately homogeneous materials. Indoor geotechnical test is performed before filling soil, and soil<br />

specific parameters are shown in Table 1. Especially, the dry density is controlled during the whole filling<br />

process.<br />

Material<br />

RESULTS AND DISCUSSIONS<br />

Axial Compression Load Test<br />

Table 1 Mechanical parameter of soils<br />

Cohesion<br />

c /kPa<br />

Friction<br />

angle φ /°<br />

Water ratio<br />

ω /%<br />

Nature<br />

density<br />

ρ /g/cm 3<br />

Sand 17.6 25.9 5.1 1.5<br />

Clay 27.6 21.2 16.7 1.9<br />

In Figure 5, base on the China Pile Foundation Construction Technical Specification (JGJ94-2008) of the ultimate<br />

bearing capacity judgment method, it is to know that the ultimate bearing capacity of XCC pile equals 119 kN,<br />

limit bearing capacities of circular pile C 1 and C 2 equal 90 kN, 140 kN, respectively. Compared with the circular<br />

pile C 1 in the same area, the concrete consumption of the XCC pile is the same with that of C 1 , its perimeter is 1.31<br />

times that of C 1 , and its ultimate bearing capacity is 1.32 times that of C 1 , that is to say its ultimate bearing<br />

capacity than that of the circular pile C 1 in the same area increases by 32 %. Compared with the circular pile C 2 in<br />

the same perimeter, the concrete consumption of the XCC pile is only 0.58 times that of C 1 , while its ultimate<br />

bearing capacity is 0.84 times that of C 2 . Although the ultimate bearing capacity of the XCC pile is on higher than<br />

-392-


that of C 2 , its concrete consumption saved 42 % and unit volume concrete ultimate bearing capacity increased 47<br />

% than that of C 2 . From the view of saving concrete and increasing bearing capacity, the XCC pile is better than<br />

the circular pile C 2 .<br />

Pile head compressive load, Q (kN)<br />

0 25 50 75 100 125 150 175<br />

0<br />

Displacement of pile head, s (mm)<br />

20<br />

40<br />

60<br />

80<br />

100<br />

XCC pile<br />

C1 pile<br />

C2 pile<br />

Axial Uplift Load Test<br />

Figure 5 <strong>The</strong> curves of pile head compressive load versus displacement (Q-s curves)<br />

Figure 6 shows the load-displacement curves under uplift load for the two different pile sections. <strong>The</strong> failure<br />

modes of a pile under uplift load can be divided into three types: (1) failure of the pile body; (2) failure of the<br />

pile-soil interface; (3) failure of the soil around the pile. From visual inspection of the tested model piles, it was<br />

observed that failure occurred in the pile-soil interface. Using the intersection of the two trend lines of the<br />

load-displacement curve, the uplift capacity of XCC pile and circular pile was found to be -70.6 kN and -56.1 kN,<br />

respectively. <strong>The</strong> ultimate uplift capacity was improved nearly 25.8 % by changing the pile cross section from a<br />

circular section to an X-section for the same amount of concrete volume used.<br />

Pile head uplift load, Q (kN)<br />

-20 -30 -40 -50 -60 -70 -80 -90<br />

0<br />

Displacement of pile head, s (mm)<br />

-20<br />

-40<br />

-60<br />

-80<br />

XCC pile<br />

Circular pile<br />

-100<br />

Figure 6 <strong>The</strong> curves of pile head uplift load versus displacement (Q-s curves)<br />

Lateral Load Test<br />

<strong>The</strong> test result of the lateral load versus lateral deflection at pile head is plotted in Figure 7 for two different pile<br />

sections. For the XCC pile tested, the ultimate lateral load-carrying capacity calculated by the p-y curve of circular<br />

piles with the equivalent pile diameter is equal to 40.4 kN, which is close to 40 kN determined by the measured<br />

-393-


load-deflection curve at the pile head. <strong>The</strong>refore, for the same lateral capacity, the amount of concrete volume<br />

used in a XCC pile is about 6.9 % less than in a circular pile.<br />

CONCLUSIONS<br />

Lateral displacement of pile head, y 0<br />

(mm)<br />

0 10 20 30 40 50 60<br />

0<br />

10<br />

20<br />

30<br />

40<br />

50<br />

Pile head lateral load, H 0<br />

(kN)<br />

XCC pile<br />

Circular pile<br />

Figure 7 <strong>The</strong> curves of pile head lateral load versus displacement (H 0 -y 0 curves)<br />

Based on the full-scale experimental study on the shape effects of cast-in-place concrete piles, the following<br />

conclusions may be drawn. XCC pile has larger ultimate bearing capacity than circular pile with the same<br />

concrete usage; In other words, XCC pile can save concrete usage under the same working load. <strong>The</strong> ultimate<br />

compressive bearing capacity of XCC pile is 1.32 times that of C 1 , that is to say its ultimate bearing capacity<br />

than that of the circular pile C 1 in the same area increases by 32 %. Compared with the circular pile C 2 in the<br />

same perimeter, the concrete consumption of the XCC pile is only 0.58 times that of C 1 , while its ultimate<br />

bearing capacity is 0.84 times that of C 2 . <strong>The</strong> ultimate uplift bearing capacity and lateral bearing capacity of<br />

XCC pile is improved 25.8 %, and 6.9 %, respectively.<br />

ACKNOWLEDGMENTS<br />

<strong>The</strong> authors gratefully acknowledge the financial support provided by the National Science Foundation of China<br />

(No. 51008116), the Provincial Science Foundation of Jiangsu, China (No. BK2008040), the Jiangsu<br />

Postdoctoral Science Foundation Funded Project (No. 1001017B), and the Fundamental Research Funds for the<br />

Central Universities (No. 2009B14514).<br />

REFERENCES<br />

Chen, R.P., Xu, Z.Z., Chen, Y.M., Ling, D.S., and Zhu, B. (2010). “Field tests on pile-supported embankments<br />

over soft ground”, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 136 (6), 777-785.<br />

JGJ94. (2008). Technical code for building pile foundations, China Architecture and Building Press, Beijing.<br />

Lei, G.H. (2001). “Behavior of excavated rectangular piles (barrettes) in granitic saproletes”, Ph.D. thesis, <strong>The</strong><br />

<strong>Hong</strong> <strong>Kong</strong> <strong>University</strong> of Science and Technology, <strong>Hong</strong> <strong>Kong</strong>.<br />

Liu, H.L. (2007a). “In-situ X-section reinforced concrete pile construction method”, China patent<br />

ZL200710020306.3.<br />

Liu, H.L., Ng, C.W.W., and Fei, K. (2007b). “Performance of a geogrid-reinforced and pile-supported highway<br />

embankment over soft clay: case study”, Journal of Geotechnical and Geoenvironmental Engineering,<br />

ASCE, 133(12), 1483-1493.<br />

Liu, H.L., Liu, Z.P., and Wang, X.Q. (2009). “Study on the geometric characteristics of the cast-in-place X-type<br />

vibro-pile section”, China Railway Science, 30(1), 17-23.<br />

Liu, Z.P., (2008). “Research on bearing behaviour of X-section concrete pile”, Ph.D. Dissertation, Hohai<br />

<strong>University</strong>, China.<br />

-394-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

基 于 基 坑 测 斜 曲 线 动 态 预 测 邻 近 单 桩 水 平 响 应<br />

章 荣 军 郑 俊 杰 潘 玉 涛 余 舜<br />

( 华 中 科 技 大 学 岩 土 与 地 下 工 程 研 究 所 , 湖 北 武 汉 430074)<br />

摘 要 : 基 于 基 坑 围 护 结 构 现 场 测 斜 曲 线 , 推 导 出 了 基 坑 外 任 意 一 点 土 体 位 移 的 计 算 表 达 式 。 基<br />

于 现 有 的 p-y 曲 线 , 考 虑 桩 土 相 互 作 用 的 非 线 性 特 性 , 提 出 了 一 种 由 围 护 结 构 监 测 信 息 动 态 预 测 既<br />

有 邻 近 单 桩 水 平 附 加 变 形 及 附 加 内 力 的 简 单 方 法 , 并 给 出 了 采 用 增 量 法 求 解 非 线 性 方 程 组 的 计 算 过<br />

程 。 验 证 实 例 表 明 该 算 法 是 可 行 的 , 能 有 效 地 分 析 基 坑 开 挖 对 邻 近 单 桩 的 影 响 规 律 。<br />

关 键 词 : 基 坑 单 桩 水 平 响 应 p-y 曲 线 测 斜 曲 线 动 态 预 测<br />

PERFORMANCE PREDICTION OF LATERAL RESPONSE OF ADJACENT<br />

SINGLE PILE BASED ON THE INCLINOMETER CURVES<br />

R. J. Zhang, J. J. Zheng, Y. T. Pan, and S. Yu<br />

Institute of Geotechnical and Underground Engineering, Huazhong <strong>University</strong> of Science and Technology,<br />

Wuhan 430074, China<br />

Abstract: According to the inclinometer curves of foundation pits, a calculation formula which can be used to<br />

assess the free-field soil displacement of any point outside the pit was derived. Based on the existing p-y curves,<br />

governing equations of “passive” single pile were obtained, and a simplified method to analyze the additional<br />

deformation and internal force of adjacent single pile by using monitoring information of the retaining structure<br />

was proposed. By adopting incremental method, the solving process of nonlinear equations was discussed in<br />

detail. <strong>The</strong> verification results indicated that the proposed method was reasonable and could be used to analyze<br />

the problem of excavation adjacent to single piles effectively.<br />

Keywords: Foundation pit, Lateral response of single pile, p-y curve, Inclinometer curve, Performance prediction.<br />

一 、 前 言<br />

合 理 评 价 基 坑 开 挖 对 邻 近 桩 基 的 影 响 是 目 前 的 研 究 热 点 之 一 [1] 。 由 于 问 题 的 复 杂 性 , 除 整 体 有<br />

[2-4]<br />

限 元 外 , 其 余 理 论 方 法 均 进 行 了 相 应 的 简 化 。 依 据 简 化 方 式 的 不 同 , 具 体 的 分 析 方 法 有 三 种 :1<br />

基 于 Mindlin 解 的 弹 性 理 论 法 [5] , 该 法 能 考 虑 土 体 的 连 续 性 , 但 不 能 考 虑 非 线 性 特 性 , 计 算 较 复 杂 ;<br />

2 将 桩 周 土 视 为 弹 性 体 的 边 界 元 法 [6-7] , 该 法 也 难 以 扩 展 到 非 线 性 ;3 基 于 Winkler 地 基 梁 模 型 的 地<br />

基 反 力 法 [8-9] , 该 法 概 念 明 确 , 简 单 适 用 , 但 不 能 考 虑 土 体 的 连 续 性 。<br />

事 实 上 , 上 述 理 论 方 法 在 实 际 应 用 存 在 两 个 问 题 :1 鉴 于 城 市 软 土 的 复 杂 性 及 区 域 性 特 征 , 依<br />

赖 于 有 限 元 法 或 经 验 法 获 取 的 土 体 自 由 场 位 移 未 必 符 合 实 际 ;2 仅 对 最 终 状 态 进 行 评 估 , 不 能 有 效<br />

地 做 到 信 息 化 施 工 。 本 文 考 虑 桩 土 荷 载 传 递 的 非 线 性 特 性 , 给 出 一 种 由 围 护 结 构 监 测 信 息 ( 可 直 接<br />

监 测 ) 动 态 预 测 既 有 桩 基 水 平 附 加 变 形 及 附 加 内 力 ( 通 常 已 投 入 运 营 , 不 便 直 接 监 测 ) 的 方 法 , 利<br />

用 该 方 法 可 以 及 时 反 馈 邻 近 桩 基 的 预 警 状 态 及 演 变 趋 势 , 为 后 续 方 案 的 决 策 及 实 施 提 供 借 鉴 意 义 。<br />

二 、 桩 土 界 面 水 平 荷 载 传 递 p-y 曲 线<br />

基 金 项 目 : 新 世 纪 优 秀 人 才 支 持 计 划 资 助 项 目 (NCET-06-649)<br />

作 者 简 介 : 章 荣 军 (1983-), 男 , 湖 北 钟 祥 人 , 博 士 生 , 主 要 从 事 隧 道 工 程 与 桩 基 工 程 方 面 理 论 研 究 及 数 值 计 算 工 作 。<br />

-395-


在 此 选 取 Kondner [10] 及 Goh [11] 推 荐 的 双 曲 线 荷 载 传 递 模 型 , 地 基 反 力 与 相 对 位 移 之 间 的 关 系 为 :<br />

ωs( z) −ωp( z)<br />

pz ( ) =<br />

, (1)<br />

1/ k + ω ( z) −ω<br />

( z) / p ( z)<br />

ni<br />

s<br />

式 中 : pz ( ) 为 桩 侧 土 体 的 地 基 反 力 ; ω s<br />

( z)<br />

为 自 由 场 桩 身 位 置 处 土 体 水 平 位 移 ; ω p<br />

( z)<br />

为 桩 体 水 平 位<br />

移 ; k ni<br />

为 桩 周 土 初 始 地 基 反 力 系 数 ; p u<br />

( z ) 为 桩 侧 极 限 土 压 力 。<br />

对 于 砂 土 地 基 而 言 ,Vesic [12] 及 Goh [11] 建 议 采 用 下 式 计 算 初 始 地 基 反 力 系 数 :<br />

k<br />

⎛0.65<br />

⎞⎛ E ⎞ E (2 r )<br />

= ⎜ ⎟<br />

⎝ ⎠⎝ ⎠<br />

s s 0<br />

ni ⎜ ⎟<br />

12<br />

2<br />

2r0 1-ν<br />

s<br />

EpIp<br />

p<br />

u<br />

4<br />

, (2)<br />

式 中 : E s<br />

为 土 体 压 缩 模 量 ; EI p p<br />

为 桩 体 抗 弯 刚 度 ; ν s<br />

为 土 体 泊 松 比 ; r 0<br />

为 桩 半 径 。 至 于 砂 土 地 基 桩<br />

侧 极 限 土 压 力 值 ,Broms [13] 提 出 如 下 的 计 算 方 法 :<br />

⎛1+<br />

sinϕ ⎞<br />

pu<br />

( z) = 3 ⎜ ⎟γ<br />

z , (3)<br />

⎝ 1 − sin ϕ ⎠<br />

其 中 : γ 为 z 平 面 以 上 土 的 平 均 重 度 ;ϕ 为 z 深 度 处 土 体 内 摩 擦 角 。<br />

对 于 黏 性 土 地 基 ,Goh [11] 提 出 适 用 于 上 述 双 曲 线 模 型 的 初 始 地 基 反 力 系 数 的 计 算 公 式 为 :<br />

k = 3.34 E / (2 r ) , (4)<br />

ni 50 0<br />

式 中 : E 50<br />

为 三 轴 不 排 水 试 验 中 50% 极 限 应 力 所 对 应 的 割 线 模 量 , 可 由 饱 和 黏 性 土 的 不 排 水 抗 剪 强<br />

度 c<br />

u<br />

及 土 体 的 塑 性 指 数 大 致 确 定 [11] 。 而 p ( z ) u<br />

的 大 小 则 可 按 照 Reese [14] 提 出 的 方 法 确 定 :<br />

三 、 基 于 测 斜 曲 线 确 定 土 体 自 由 场 水 平 位 移<br />

( γ<br />

)<br />

⎧ p ( z) = 3 + z/ c + 0.25 z/<br />

r c<br />

⎪<br />

⎨ pu2<br />

( z) = 9.0 cu<br />

⎪<br />

⎩<br />

pu( z) = min[ pu1( z), pu2( z)]<br />

u1 u 0 u<br />

. (5)<br />

Xu 和 Poulos 基 于 Sagaseta 解 , 积 分 得 到 基 坑 外 任 意 点 (x, z) 处 水 平 位 移 表 达 式 为 [15,16] :<br />

αh + βh+ χ ⎧1 ⎛ x x ⎞ x ⎡ z( z+<br />

h)<br />

⎤⎫<br />

Sx<br />

=− − + −<br />

⎩⎪<br />

⎝ ⎠ ⎣ ⎦⎭⎪<br />

∫<br />

2<br />

H0<br />

⎪<br />

⎪<br />

⎨ ⎜ 1 d<br />

0<br />

2 2 ⎟ 2 ⎢ h<br />

2 ⎥⎬<br />

, (6)<br />

π 2 r1 r2 r2 0.5r2<br />

2 2<br />

2 2<br />

2<br />

式 中 : r1 = x + ( z− h)<br />

, r2 = x + ( z+ h)<br />

, χ = ut<br />

, α =−2(2 uc −ut − ub) / H , β = (4uc −3 ut − ub) / H 。<br />

h 为 积 分 变 量 , H 0<br />

为 围 护 结 构 深 度 , u t<br />

、 u b<br />

及 u<br />

c<br />

分 别 为 围 护 结 构 顶 部 、 底 部 和 中 部 的 水 平 位 移 。<br />

式 (6) 存 在 三 点 缺 陷 :1 采 用 二 次 抛 物 线 来 描 述 围 护 结 构 的 挠 曲 有 时 会 与 实 际 相 差 甚 远 ;2 忽 略<br />

了 围 护 结 构 以 下 的 地 层 损 失 ;3 围 护 结 构 外 的 土 体 并 非 半 空 间 , 直 接 用 Sagaseta 公 式 积 分 是 不 妥 的 。<br />

在 实 际 深 基 坑 施 工 过 程 中 , 一 般 都 会 进 行 连 续 墙 测 斜 观 测 , 因 此 直 接 用 测 斜 曲 线 f ( h ) 代 替 式 (6)<br />

2<br />

中 的 αh<br />

+ βh+ χ , 便 可 有 效 避 免 上 述 缺 陷 1。 当 围 护 结 构 插 入 比 较 小 时 , u b<br />

不 为 0, 此 时 可 根 据 围<br />

护 结 构 水 平 位 移 沿 深 度 的 发 展 趋 势 及 工 程 经 验 假 定 下 部 土 体 的 水 平 位 移 曲 线 , 必 要 时 可 在 紧 邻 围 护<br />

结 构 处 布 设 土 体 测 斜 观 测 点 ( 深 度 应 足 够 大 ), 并 以 此 确 定 f ( h ) 及 变 形 范 围 H , 并 以 f ( h ) 代 替<br />

2<br />

αh<br />

+ βh+ χ 、 H 代 替 H<br />

0<br />

, 便 可 避 免 上 述 缺 陷 2。 下 面 进 一 步 分 析 克 服 缺 陷 3 的 方 法 。<br />

如 图 1(a) 所 示 , 连 续 墙 的 挠 曲 为 f ( h )。 在 小 应 变 情 况 下 , 该 问 题 相 当 于 是 在 围 护 结 构 与 土 体 之<br />

间 形 成 一 物 理 间 隙 ( 图 1(b) 中 的 阴 影 ), 该 间 隙 最 终 由 围 护 结 构 外 侧 土 体 ( 二 分 之 一 半 空 间 ) 填 充 , 基<br />

坑 外 侧 土 体 位 移 的 分 布 规 律 只 与 f ( h ) 相 关 。 由 对 称 性 , 该 问 题 就 转 化 为 图 1(c) 所 示 的 弹 性 半 空 间 中<br />

发 生 2 f ( h ) 地 层 损 失 下 XZ 平 面 上 任 意 一 点 地 层 位 移 的 求 解 问 题 , 如 此 转 化 后 便 可 直 接 应 用 Sagaseta<br />

公 式 进 行 求 解 了 。 将 整 个 水 平 位 移 分 布 曲 线 划 分 为 若 干 微 段 , 每 个 微 段 的 面 积 为 2d h⋅<br />

f( h)<br />

, 采 用 面<br />

积 等 效 的 方 法 将 微 段 面 积 等 效 为 圆 , 代 入 Sagaseta 公 式 并 积 分 , 可 得 坑 外 一 点 (x, z) 处 水 平 位 移 为 :<br />

H 2 f( h) 禳 1 骣 x x x 轾 z( z+<br />

h)<br />

Sx<br />

= ò 镲 睚 - - - 1-<br />

2 dh<br />

0<br />

2 2 2 2<br />

p<br />

2ç r1 r ÷<br />

2<br />

r 犏<br />

铪 镲 桫<br />

2 臌 r<br />

. (7)<br />

2<br />

四 、 考 虑 桩 土 作 用 非 线 性 特 性 的 分 析 方 法<br />

-396-


建 立 如 图 2 所 示 的 单 桩 二 维 分 析 模 型 , 分 析 中 假 定 :1 桩 的 存 在 不 影 响 基 坑 的 开 挖 及 围 护 结 构<br />

的 水 平 位 移 ( 与 文 献 [1] 相 同 );2 桩 土 相 互 作 用 由 非 线 性 弹 簧 来 模 拟 , 桩 土 不 发 生 脱 离 ;3 桩 径 不<br />

变 , 桩 身 为 线 弹 性 ;4 基 坑 外 土 体 不 可 压 缩 ;5 不 考 虑 桩 身 轴 力 对 水 平 变 形 的 影 响 。<br />

f (h)<br />

坑 底<br />

围 护<br />

结 构<br />

地 表<br />

h<br />

H<br />

X<br />

小 应 变<br />

围 护<br />

结 构<br />

h<br />

地 表<br />

X<br />

f (h)<br />

H<br />

dh<br />

对 称<br />

地 表<br />

h<br />

微 单<br />

元 体<br />

dh<br />

X<br />

2f (h)<br />

面 积<br />

等 效<br />

2a<br />

L<br />

P<br />

地 表<br />

2r0<br />

桩<br />

H<br />

H 0<br />

u t<br />

uc<br />

开 挖 面<br />

u b<br />

h<br />

dh<br />

f (h)<br />

围 护<br />

结 构<br />

H 0 / 2<br />

H 0 / 2<br />

X<br />

虚 拟 节 点<br />

-2<br />

-1<br />

0...<br />

i-1<br />

i<br />

i+1<br />

n...<br />

节 点 放 大<br />

ω p,i<br />

ω s,i<br />

p i<br />

p i =f (ω s,i -ω p,i )<br />

Z<br />

Z<br />

(a) (b) (c)<br />

Z<br />

E p I p<br />

Z<br />

n+1<br />

n+2<br />

虚 拟 节 点<br />

ω s,i -ω p,i<br />

图 1 坑 外 土 体 位 移 计 算 模 型<br />

图 2 邻 近 单 桩 水 平 响 应 的 预 测 模 型<br />

由 弹 性 地 基 梁 桩 体 挠 曲 微 分 方 程 可 得 :<br />

4 4<br />

EI<br />

p pd ωp( z)/dz −2 r0[ ωs( z) − ωp( z)]/ ⎡1/ kni ωs( z) ωp( z) / pu( z) ⎤<br />

⎣<br />

+ −<br />

⎦<br />

= 0. (8)<br />

用 差 分 代 替 微 分 得 :<br />

4<br />

EI( ω − 4ω + 6ω − 4 ω + ω )/ λ = 2 r( ω − ω )/[1/ k + ω − ω / p ], (9)<br />

p p p, i+2 p, i+1 p, i p, i−1 p, i−2 0 s, i p, i ni s, i p, i u, i<br />

式 中 ,i=0,1,…,n; ω p,i<br />

为 桩 身 i 节 点 挠 曲 值 ; ω s,i<br />

为 第 i 节 点 处 自 由 场 土 体 水 平 位 移 ; k ni<br />

为 土 体<br />

的 初 始 地 基 反 力 系 数 ; p u,i<br />

为 第 i 节 点 处 的 土 体 极 限 抗 力 值 。 桩 顶 桩 底 的 边 界 条 件 为 :<br />

3<br />

⎪<br />

⎧ωp,2 − 2ωp,1 + 2 ωp, −1 − ωp, −2 = (2 HF / EpIp<br />

) λ = 0<br />

桩 顶 ( 外 荷 载 H<br />

F<br />

及 M 均 取 为 0): ⎨<br />

. (10)<br />

2<br />

⎪⎩ ωp, −1 − 2 ωp,0 + ωp,1 = ( M / EpIp<br />

) λ = 0<br />

桩 底 :<br />

⎧<br />

⎪<br />

ωp, n+2 − 2ωp, n+1 + 2ωp, n−1 − ωp, n−2<br />

= 0<br />

⎨<br />

. (11)<br />

⎪⎩ ωp, n−1 − 2ωp, n<br />

+ ωp, n+1<br />

= 0<br />

在 此 采 用 增 量 法 求 解 式 (9)。 令 δ = ωs( z) − ωp( z)<br />

, 即 δi = ωs, i<br />

− ωp,<br />

i<br />

, 将 式 (9) 写 成 增 量 形 式 , 即 :<br />

[ K ] Δ ω = [ K ] Δ δ = [ K ] Δω −Δ ω<br />

(12)<br />

其 中 ,{ ω p,i }<br />

p { p, i} s { i} s { s, i p, i}<br />

Δ 为 桩 身 水 平 位 移 增 量 列 向 量 ,{ ω s,i }<br />

Δ 为 桩 身 处 自 由 场 水 平 位 移 增 量 列 向 量 ,[ K p<br />

] 及 [ Ks<br />

]<br />

分 别 为 桩 体 刚 度 矩 阵 及 土 弹 簧 刚 度 矩 阵 。 假 定 进 行 第 m 个 增 量 步 的 求 解 时 ,[ K ] m 由 第 m-1 增 量 步<br />

求 的 { } m − 1<br />

δ i<br />

所 对 应 的 切 线 刚 度 确 定 , 则 [ K<br />

p<br />

] 及 [ K<br />

s<br />

] 可 根 据 式 (1) 及 前 一 增 量 步 的 计 算 结 果 确 定 。<br />

五 、 方 法 验 证<br />

选 取 文 献 [6] 中 的 “Basic Problem” 来 进 行 对 比 验 证 。 基 坑 围 护 结 构 深 13m, 开 挖 深 度 10m; 地<br />

层 为 饱 和 软 黏 土 , 容 重 为 20kN/m 3 , c u<br />

为 40kPa, E s<br />

为 16MPa; 桩 长 L = 22 m, 桩 径 为 500mm,<br />

E<br />

p<br />

= 30 Gpa, 桩 距 离 基 坑 x = 1 m。 计 算 中 假 定 E<br />

i<br />

为 32MPa( 约 为 1600 c u<br />

) [11] , 泊 松 比 v<br />

s<br />

取 0.5。<br />

图 3 显 示 了 Poulos 法 与 本 文 方 法 计 算 结 果 的 对 比 情 况 , 由 图 可 知 , 在 “ 挖 至 -3m” 和 “ 挖 至 -6m”<br />

两 种 工 况 下 , 二 者 结 果 几 乎 重 合 , 在 “ 挖 至 -8m” 工 况 下 , 上 半 部 分 也 完 全 吻 合 , 下 半 部 分 存 在 一 定<br />

的 差 别 。 总 体 上 看 , 两 种 方 法 计 算 得 到 的 桩 轴 线 上 自 由 场 水 平 位 移 、 桩 身 的 水 平 位 移 及 桩 身 弯 矩 等<br />

均 表 现 出 较 好 的 一 致 性 , 本 文 提 出 的 算 法 是 合 理 可 行 的 。<br />

六 、 结 语<br />

本 文 提 出 了 基 于 基 坑 测 斜 曲 线 评 估 基 坑 外 任 意 一 点 土 体 水 平 位 移 的 计 算 表 达 式 , 并 考 虑 桩 土 界<br />

面 水 平 荷 载 传 递 的 非 线 性 特 性 , 给 出 一 种 由 围 护 结 构 监 测 信 息 动 态 预 测 既 有 邻 近 单 桩 水 平 附 加 变 形<br />

s<br />

-397-


及 附 加 内 力 的 简 单 方 法 , 利 用 该 方 法 可 以 及 时 反 馈 邻 近 桩 基 的 预 警 状 态 及 演 变 趋 势 。 实 例 分 析 表 明 ,<br />

本 文 所 提 出 方 法 的 计 算 结 果 是 合 理 可 行 的 , 能 够 较 为 有 效 地 分 析 基 坑 开 挖 对 邻 近 单 桩 的 影 响 规 律 。<br />

深 度 (m)<br />

0<br />

-5<br />

-10<br />

-15<br />

-20<br />

-25<br />

Poulos 法 :<br />

挖 至 -3m<br />

-3m<br />

挖 至 -6m<br />

-6m<br />

挖 至 -8m<br />

-8m<br />

本 文 挖 方 至 法<br />

-3m<br />

:<br />

挖 至 -6m<br />

-3m<br />

挖 至 -8m<br />

-6m<br />

挖 至 -8m<br />

0.00 0.01 0.02 0.03 0.04 0.05<br />

桩 身 水 平 位 移 (m)<br />

(a) 自 由 场 水 平 位 移<br />

深 度 (m)<br />

0<br />

-5<br />

-10<br />

-15<br />

-20<br />

-25<br />

本 文 方 法 :<br />

挖 至 -3m<br />

挖 至 -6m<br />

挖 至 -8m<br />

Poulos 法 :<br />

挖 至 -3m<br />

挖 至 -6m<br />

挖 至 -8m<br />

-50 0 50 100 150<br />

桩 身 弯 矩 (kNm)<br />

(b) 桩 身 弯 矩<br />

图 3 与 Poulos 法 计 算 结 果 的 对 比<br />

参 考 文 献<br />

[1] 杜 金 龙 , 杨 敏 . 基 坑 开 挖 与 邻 近 桩 基 相 互 作 用 的 弹 塑 性 解 . 岩 土 工 程 学 报 , 2008, 30(8): 1121-1125.<br />

[2] Finno, R. J., Lawence, S. A. and Allawh, N. F. Analysis of performance of pile groups adjacent to deep excavation.<br />

Journal of Geotechnical Engineering, ASCE, 1991, 117(6): 934-955.<br />

[3] Poulos, H. G., and Chen, L. T. Pile response due to excavation- induced lateral soil movement. Journal of Geotechnical<br />

and Geoenvironmental Engineering, ASCE, 1997, 23(2): 94-99.<br />

[4] 杨 敏 , 周 洪 波 , 杨 桦 . 基 坑 开 挖 与 临 近 桩 基 相 互 作 用 分 析 . 土 木 工 程 学 报 , 2005, 38(4): 91-96.<br />

[5] Poulos, H. G. and Davis, E. H. Elastic Solution for Soil and Rock Mechanics. Sydney: John Wiley & Sons, Inc, 1973.<br />

[6] Poulos, H. G. and Chen, L. T. Pile response due to excavation induced lateral soil movement. Journal of Geotechnical and<br />

Geoenvironmental Engineering, ASCE, 1997, 123(2): 94-99.<br />

[7] Xu, K. J. and Poulos, H. G. 3-D elastic analysis of vertical piles subjected to “passive” loadings. Computers and<br />

Geotechnics, 2001, 28: 349-375.<br />

[8] 黄 茂 松 , 张 陈 蓉 . 开 挖 条 件 下 非 均 质 地 基 中 被 动 群 桩 水 平 反 应 分 析 . 岩 土 工 程 学 报 , 2008, 30(7): 1017-1023.<br />

[9] Huang, M. S. and Zhang, C. R. A simplified analysis method for the influence of tunneling on grouped piles. Tunnelling<br />

and Underground Space Technology, 2009, 24(4): 410-422.<br />

[10] Kondner, R. L. Hyperbolic stress-strain response: cohesive soils. Journal of the Soil Mechanics and Foundation<br />

Engineering Division, ASCE, 1963, 89(SM1): 115-143.<br />

[11] Goh A. T. C., Teh, C. I. and Wong, K. S. Analysis of piles subjected to embankment induced lateral soil movements.<br />

Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 1997, 123(9): 792-801.<br />

[12] Vesic, A. S. Bending of beams resting on isotropic elastic solid. Journal of the Engineering Mechanics Division, ASCE,<br />

1961, 87: 35-53.<br />

[13] Broms, B. B. <strong>The</strong> lateral resistance of piles in cohesionless soils. Journal of Soil Mechanics and Foundation Engineering,<br />

ASCE, 1964, 90(3): 123-156.<br />

[14] Reese, L. C., Cox, W. R., and Koop, F. D. Analysis of laterally loaded piles in sand. Proceedings of the 6th Annual<br />

Offshore Technology Conference, Houston, 1974: 473-483.<br />

[15] Sagaseta, C. Analysis of undrained soil deformation due to ground loss. Geotechnique, 1987, 37(3): 301-320.<br />

[16] Xu, K. J. and Poulos, H. G. <strong>The</strong>oretical study of pile behaviour induced by a soil cut. Proceedings of GeoEng 2000,<br />

Melbourne, 2000.<br />

-398-


Parallel Session –<br />

Structural I


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

A NEW RELIABILITY ANALYSIS METHOD BASED ON UNIFORM DESIGN<br />

METHOD AND SUPPORT VECTOR MACHINES<br />

X. L. Yu 1 , J. B. Yu 1 , H. B. Zheng 1 and Q. S. Yan 1<br />

1<br />

School of Civil Engineering and Transportation,<br />

South China <strong>University</strong> of Technology, Guangzhou, China. Email: XLYU1@scut.edu.cn<br />

ABSTRACT<br />

To mostly engineering structures, the performance functions can not be expressed explicitly. <strong>The</strong>ir structural<br />

response can only be obtained by numerical methods. <strong>The</strong> reliability analysis needs times of finite element<br />

calculations. It costs much time and it is difficult in calculating the failure probability. Response surface method<br />

(RSM) is an effective way to solve such problems. Over the last decade, support vector machines (SVM) with<br />

good ability of generalization obtained a widely application in data classification and regression analysis. In<br />

order to improve the computational efficiency and make RSM suitable well to large and complex engineering<br />

structures, the reliability analysis method based on uniform design method (UDM) and SVM was proposed.<br />

UDM is adopted to select training data and SVM is used as response surface. Structural reliability index is<br />

calculated in combination with the traditional reliability analysis methods (such as, the first-order reliability<br />

method (FORM), the second-order reliability method (SORM) or Monte Carlo simulation method (MCSM)).<br />

Numerical examples show that sampled with the UDM can greatly reduce the number of samples required for<br />

training by SVM model, and a very good approximation of the limit state surface can be obtained to get the<br />

failure probability.<br />

KEYWORDS<br />

Reliability, support vector machine, uniform design method, response surface method.<br />

INTRODUCTION<br />

<strong>The</strong> RSM developed in recent years is an effective way to solve the structural reliability problems with implicit<br />

performance function. <strong>The</strong> basic idea of RSM is to replace the true performance function which is originally<br />

implicit or takes long time to determine with an explicit function (usually polynomial) which is easy to deal with<br />

so as to simplify the reliability analysis. Polynomial-based RSM usually use traditional quadratic polynomial to<br />

explicitly approximate the response surface of limit state function, whose effectiveness is significantly affected<br />

by the shape of limit state function. When the function used by RSM is not a good approximation of the true<br />

structural response function, the results of RSM are sensitive to the values of parameter f (Guan et al. 2001).<br />

Artificial Neural Network (ANN) has been widely adopted in structural reliability analysis for its ability to<br />

approximate function well. However, due to the traditional Empirical Risk Minimization (ERM) principle, when<br />

employed in the training process, ANN has suffered difficulties in generalization and overfitting the data. Since<br />

some prior knowledge (such as, the determination of hidden nodes) is needed to fix the structure of the network<br />

ANN has some defects, which limit the application of ANN to reliability analysis to some extent.<br />

Based on statistical learning theory, Support Vector Machines (SVM) (Cortes et al. 1995; Vapnik 1995, 1999) have<br />

strict mathematical basis. With the characteristics of structural risk minimization, the machine learning models<br />

designed by SVM have been widely used in data classification and regression analysis over a decade. Compared<br />

with ANN, SVM has some more advantages, for there is no locally optimization problem in SVM, which<br />

improves the generalization ability of learning machine. SVM was used as response surface function (JIN et al.<br />

2007; ZHAO et al. 2008). SVM was applied to reliability analysis. In their studies, the training data were<br />

generated by interpolation iteration in the center points. In this case, SVM only fit well with limit state surface<br />

in the design point similar to the traditional RSM, which doesn’t make full use of the generalization ability of<br />

SVM. In order to improve the computational efficiency of SVM, least squares support vector machine (LS-SVM)<br />

was used as response surface function (JIN et al. 2007). LS-SVM-based reliability analysis was performed in<br />

conjunction with training data produced by orthogonal design, which needs more samples.<br />

In order to improve the computational efficiency and to apply SVM-based reliability method to practical<br />

-399-


complex structure, the reliability analysis method based on SVM with uniform design method (UDM) was<br />

proposed here. UDM is adopted to select training data and SVM is used as response surface. Structural<br />

reliability index is calculated in combination with the traditional reliability analysis methods (such as, the<br />

first-order reliability method (FORM), the second-order reliability method (SORM) or Monte Carlo simulation<br />

method (MCSM)). It is shown through numerical examples that by introducing uniform design method, the<br />

number of training data for establishing a SVM model is reduced and SVM can approximate complex response<br />

surface well. <strong>The</strong> analysis results show that the method is feasible and has a good prospect, which provides a<br />

new way for structural reliability analysis.<br />

METHOD OF SOLUTION<br />

<strong>The</strong> Uniform Design Method<br />

UDM (Fang K. T. et al.1978) considers that within the scope of the study the experimental points are evenly<br />

spread. Its mathematical theory is consistent distribution theory in number-theoretic method. UDM is ideal for<br />

the situations of testing with multi-factor, multi-level and system models completely unknown.<br />

This method develops some well-designed tables to carry out tests, taking into account the test points within test<br />

scope are uniformly dispersed. By using this experimental design method, the number of trials can be reduced<br />

s<br />

and the efficiency can be improved. <strong>The</strong> tables are named uniform design tables which are expressed as U<br />

n<br />

( q ) .<br />

Where, U means the uniform design. n is number of tests. q is number of layers. s is possible factors.<br />

<strong>The</strong>se methods, such as, good lattice point method, the Latin method, orthogonal expansion method and<br />

stochastic optimization method, are used to construct uniform design. <strong>The</strong> steps of how to construct uniform<br />

distribution design table by good lattice point method are as follows:<br />

1) Given a positive integer n , to find a positive integer h ( h < n), of which the greatest common divisor of<br />

n and h is 1. <strong>The</strong> positive integers meeting this condition make up of a vector H = ( h , h , L , h<br />

1 2 m<br />

) .<br />

2) Generate the j column of the uniform design table, u = ih ( mod n<br />

ij j<br />

) . If ih exceeds n , ih subtract<br />

j<br />

j<br />

an appropriate multiple of n , making the difference fall into [1, n ]. Thus, U = ( u ij ) is a matrix with size of<br />

n× m.<br />

D X is introduced to check the construction of uniform design tables.<br />

3) <strong>The</strong> discrepancy ( )<br />

D<br />

X<br />

N<br />

S<br />

X∈C<br />

n<br />

X (1)<br />

( ) = max X<br />

−V([0, ))<br />

Where, V ([0, X))<br />

is the volume of [0, X ) . N X<br />

is the number of x 1<br />

, x 2<br />

, L , x points which belong to<br />

m<br />

[0, X ) .<br />

S<br />

C is the domain defined.<br />

Uniform design tables constructed should be used with corresponding accessory table. <strong>The</strong> experimental points,<br />

determined by the accessory table distribute uniformly in the test area, are the most representative test points.<br />

Support Vector Machines<br />

SVM is a machine learning method based on statistical learning theory. In accordance with limited samples of<br />

information, SVM finds the best between the complexity and learning ability in the model in order to get the<br />

best generalization. SVM algorithm will eventually transform into a quadratic optimization issue.<br />

n<br />

Given a set of data points{ x , y}<br />

, i = 1,2, L , l , such that x ∈ R is an input and y ∈ R is a target output, the<br />

i i<br />

i<br />

i<br />

standard form of support vector regression can be expressed as<br />

-400-


<strong>The</strong> dual is<br />

min<br />

*<br />

w, b, ξξ ,<br />

s.t<br />

1<br />

min<br />

* αα , 2<br />

n<br />

1<br />

T<br />

*<br />

ww+ C∑( ξ + ξ<br />

i i<br />

)<br />

2<br />

i=<br />

1<br />

T<br />

w Φ + b− y ≤ ε + ξ<br />

y<br />

( x )<br />

i i i<br />

( x )<br />

−w<br />

Φ −b<br />

≤ ε + ξ<br />

T<br />

*<br />

i i i<br />

*<br />

, ≥ 0, i = 1, 2, L,<br />

i i<br />

ξ ξ<br />

l<br />

l<br />

*<br />

T<br />

* * *<br />

( α−α ) Q( α− α ) + ε∑( α + α<br />

i i ) + ∑yi( α −α<br />

i i )<br />

l<br />

∑( )<br />

i= 1 i=<br />

1<br />

α − α = ≤α α ≤ C i = L l<br />

* *<br />

s. t 0, 0 , , 1, 2, ,<br />

i i i i<br />

i=<br />

1<br />

l<br />

T<br />

Where Q is an l× l positive semi-definite matrix, Q = K( x , x ) ≡ Φ( x ) Φ( x ), and K ( x , x ) is the<br />

ij i j i j<br />

i j<br />

kernel.<br />

<strong>The</strong> approximate function is:<br />

l<br />

*<br />

f ( x) = ∑( − α + α ) K( x , x ) + b<br />

(4)<br />

i i i<br />

i=<br />

1<br />

SVM-based reliability analysis method with UDM<br />

<strong>The</strong> limit state function can be calculated by SVM as<br />

g( x) = z = f ( x )<br />

(5)<br />

Where X= {x 1 ,x 2 ,…,x n } is the vector of random variables.<br />

In FORM and SORM, the first-order and second-order derivatives of the structural response need to be<br />

computed. So the polynomial function, the Gaussian radial basis function and the sigmoid function can be used<br />

as the kernel functions in SVM-based reliability analysis.<br />

Polynomial function<br />

A polynomial mapping is a popular method for non-linear modeling:<br />

T<br />

K ( xc , ) = ( xc + 1)<br />

q<br />

(6)<br />

i<br />

i<br />

Where q ≥ 2 , and c<br />

i<br />

are the support vectors.<br />

Gaussian Radial Basis Function<br />

'<br />

q<br />

q−1<br />

∂z<br />

⎛ ⎛ ⎞ ⎞<br />

⎛ ⎞<br />

= − + ⎜ + ⎟ = +<br />

∂x<br />

j i= 1<br />

⎝ ⎝ j= 1 ⎠ ⎠<br />

i= 1 ⎝ j=<br />

1 ⎠<br />

2<br />

q−1<br />

m<br />

l<br />

m<br />

l<br />

∂ z ∂ ( j) * ⎛ ( j)<br />

⎞ ( j) ( k) * ⎛ ( j)<br />

⎞<br />

= q ci ( − α + α ) 1 ( 1) ( )<br />

1<br />

i i ⎜ x c<br />

j i<br />

+ ⎟ = qq− ∑c c − α + α xc<br />

i i i i ⎜∑<br />

+<br />

j i ⎟<br />

∂x ∂x ∂x<br />

j k k i= 1 ⎝ j=<br />

1 ⎠<br />

i= 1 ⎝ j=<br />

1 ⎠<br />

m l m l<br />

( ) ( j) ( j) ( )<br />

( j)<br />

∑⎜ α α<br />

1<br />

i i ∑x c q x<br />

j i ⎟ ∑c − α + α c<br />

i i i ⎜∑ j i<br />

1 ⎟<br />

(7)<br />

∑ ∑ (8)<br />

Gaussian radial basis function can be chosen as the kernel function.<br />

2<br />

⎛ x−<br />

c ⎞<br />

i<br />

K ( xc , ) = exp −<br />

i ⎜<br />

2 ⎟<br />

(9)<br />

⎝ 2σ<br />

⎠<br />

∂ z =− −<br />

∂x<br />

j<br />

1 ci<br />

m<br />

*<br />

( j ) x −<br />

2 ∑( − α + α )( x c ) exp<br />

i i j i ⎜−<br />

σ = 2<br />

i 1<br />

⎛<br />

⎝<br />

2<br />

2<br />

2<br />

m<br />

m<br />

∂ z ∂ ⎛ 1 ⎛<br />

* ( j)<br />

− ⎞⎞<br />

1 ⎛∂x<br />

i<br />

* j 1 ( j) ( k)<br />

⎞ ⎛ − ⎞<br />

i<br />

= −<br />

2 ( − α + α )( x ) exp<br />

2<br />

2 ( ) 2 ( )( ) exp<br />

i i j<br />

− ci<br />

− =− − α + α − x<br />

i i<br />

j ci xk<br />

ci<br />

⎜− 2<br />

xj xk x ⎜ ∑<br />

x c<br />

k<br />

σ ⎜<br />

i= 1 2σ<br />

⎟<br />

⎟ ∑ ⎜ − − ⎟<br />

x c<br />

∂ ∂ ∂ ⎝ ⎠ σ i=<br />

1 ⎝∂xk<br />

σ ⎠<br />

⎜ 2σ<br />

⎟(11)<br />

⎝ ⎠ ⎝ ⎠<br />

2<br />

σ<br />

2<br />

⎞<br />

⎟<br />

⎠<br />

q−2<br />

(2)<br />

(3)<br />

(10)<br />

-401-


Sigmoid function<br />

Another kernel, that might seem appealing, is the hyperbolic tangent kernel<br />

K xc , = tanh v( xc T<br />

) + c<br />

(12)<br />

( )<br />

i<br />

(<br />

i<br />

)<br />

Supposing y = 1+ exp( − 2 ( v( xc ) ))<br />

+ c<br />

i<br />

i<br />

z<br />

m<br />

⎛ 2<br />

( − α + α )<br />

i i<br />

1<br />

i=<br />

1<br />

i<br />

⎞<br />

, then<br />

= ∑ ⎜ − ⎟+<br />

b<br />

y<br />

(13)<br />

⎝<br />

∂y<br />

i<br />

∂x<br />

j<br />

⎠<br />

( ( ))<br />

( j ) T<br />

2 exp 2 ( ) 2 ( j<br />

vc v c vc )<br />

( y 1)<br />

i i i i<br />

=− − xc + =− −<br />

(14)<br />

m m m<br />

∂z<br />

∂z<br />

∂y<br />

1 1 1<br />

= = − − − −<br />

∂x ∂y<br />

∂x y<br />

y y<br />

i<br />

*<br />

( )( ( j ) *<br />

) ( j<br />

∑ 2∑ − α + α 2 vc ( y 1) = 4 vc<br />

)<br />

2<br />

( α α )( )<br />

i i i i ∑ − +<br />

i<br />

i i<br />

2 (15)<br />

j i= 1 i j i= 1 i i=<br />

1<br />

i i<br />

2<br />

m<br />

m<br />

∂ z ∂ ⎛<br />

( j) *<br />

1 1 ⎞<br />

( j)<br />

*<br />

∂ 1 1<br />

= ⎜4 ∑vc<br />

( − α + α )( − ) 4<br />

2<br />

( ) ( )<br />

i<br />

i i ⎟ = ∑vc<br />

− α + α −<br />

i<br />

i i<br />

2<br />

∂x ∂x ∂x j k k ⎝ i= 1 y y<br />

i i ⎠ i=<br />

1<br />

∂x y y<br />

k i i<br />

2 1 ∂y<br />

2 1<br />

= − = − − + − −<br />

SVM-based reliability analysis method with UDM<br />

m<br />

( j) 2 ( j) ( k) *<br />

4 ∑vc ( ) 8 v c c ( )( y 1)<br />

i 3 2 i i i i 3 2 i<br />

i= 1 y y ∂x i i k i=<br />

1<br />

y y<br />

i i<br />

m<br />

*<br />

i<br />

( − α + α<br />

i i<br />

) ∑ ( α α )<br />

Once the first-order and second-order derivatives of the structural response have been calculated, FORM or<br />

SORM can be used to get the failure probability.<br />

<strong>The</strong> calculation steps of SVM-based reliability method can be described as follows:<br />

1) Select the random variables, specify their probabilistic characters, and define the limit state function g ( x ) = 0;<br />

2) Generate the samples from the distribution functions of the basic variables with UDM and compute the<br />

corresponding value of the performance function g(x) and divide the samples into two groups, namely the<br />

training set and the validation set;<br />

3) Establish the SVM model of the performance function;<br />

4) Calculate the support vectors according to the training set and test SVM according to the validation set. If the<br />

tolerance defined before is met, the next step will be executed. Otherwise, repeat steps 3 through 4 until the<br />

error bellows the tolerance;<br />

5) Choose the reliability method, such as FORM, SORM or MCSM and compute failure probability according<br />

to the functional approximation by SVM.<br />

6) Calculate the reliability index β and failure probability P<br />

f<br />

.<br />

APPLICATIONS<br />

Example 1<br />

<strong>The</strong> limit state function is defined as in [17]<br />

g ( x ) = exp(0.2× x + 1.4) −x<br />

1 2<br />

(17)<br />

where x and x are assumed to be independent and have a standard normal distribution with zero mean and<br />

1 2<br />

unit standard deviation.<br />

2<br />

Select the uniform design table U<br />

25 ( 25 ). Take 20 samples which are uniformly distributed within[ − 3, 3]<br />

and compose the experimental designs with 2 factors and 20 levels. <strong>The</strong> results are shown in table 1. It can be<br />

seen that the results obtained by UDM-SVM RSM are in good agreement with reference(Kim et al. 1997; Zheng<br />

et al. 2000; Elhewy et al. 2006). It suggests that even in the case of small sample sizes, SVM is also more<br />

accurate fitting of the limit state surface. Moreover, for the same number of training samples taken (eg 25),<br />

UDM selected samples obtained better results than the orthogonal design.<br />

(16)<br />

-402-


Table 1 <strong>The</strong> results of example 1<br />

Reliability index Failure probability ( P ×10 -4 )<br />

f<br />

FORM SORM MC FORM SORM MC<br />

Number of<br />

samples<br />

Method<br />

3.3800 3.600 10 6 MC(Elhewy et al. 2006)<br />

3.3500 4.060<br />

Polynomial RSM (Zheng et al.<br />

2000)<br />

3.3810 3.400 (Kim et al. 1997)<br />

3.3398 3.3951 3.3900 4.192 3.430 3.495 25<br />

SVM RSM (selected samples with<br />

UDM)<br />

3.3248 3.3919 3.3752 4.424 3.471 3.688 25<br />

SVM RSM (selected samples with<br />

orthogonal design)<br />

Example 2<br />

Under uniformly distributed load, the maximum vertical displacement of rectangular cantilever beam can not<br />

exceed the allowable deformation L / 325 . <strong>The</strong> performance function is [18]<br />

4<br />

g x = L / 325 −qbL / (8 EI)<br />

(18)<br />

( )<br />

Where, EqIbLare , , , , modulus of elasticity, uniform load, moment of inertia, the beam width and length<br />

4<br />

respectively. L = 0.6 m, E = 2.6 × 10 MPa. <strong>The</strong> limit state function is defined as in [19]:<br />

3<br />

g x = 0.01846154 − 74.76923 × x /( x )<br />

(19)<br />

( )<br />

1 2<br />

Where, x and x are assumed to be independent and have standard normal distribution. <strong>The</strong>ir statistical<br />

1 2<br />

parameters are shown in Table 2.<br />

Table 2 <strong>The</strong> statistical parameters of variables of example 2<br />

Variables Uint Mean μ Coefficient of variation<br />

x —— q<br />

1<br />

MPa 1000 0.20<br />

x —— h<br />

2<br />

m 0.25 0.15<br />

− + and design test arrangements with uniform table ( 2<br />

27 )<br />

27<br />

( 2<br />

6<br />

U 25<br />

25 ) and orthogonal table L<br />

25 ( 5 )<br />

Choose samples within [ μ 3 σ, μ 3σ]<br />

U ,<br />

. <strong>The</strong> kernel function is also the Gaussian function. <strong>The</strong> results are<br />

shown in Table 3. It can be seen that the results obtained by UDM-SVM RSM are in good agreement with<br />

reference (Rajashekhar et al. 1993). It suggests that even in the case of small sample sizes, SVM is also more<br />

accurate fitting of the limit state surface. Moreover, for the same number of training samples taken (eg 25),<br />

UDM selected samples obtained better results than the orthogonal design.<br />

Table 3 <strong>The</strong> results of example 2<br />

Reliability index Failure probability ( P ×10 -4 )<br />

f<br />

FORM SORM MC FORM SORM MC<br />

Number of<br />

samples<br />

Method<br />

2.3049 2.3381 2.3409 1.059 0.969 0.962 27 UDM<br />

2.3658 2.3818 2.3820 0.900 0.861 0.861 25 UDM<br />

2.0595 2.0802 2.0809 1.972 1.875 1.872 25<br />

selected samples with orthogonal<br />

design<br />

2.3410 0.9607 1000<br />

Importance sampling simulation<br />

(Rajashekhar et al. 1993)<br />

-403-


Example 3<br />

6<br />

A portal frame is shown in Figure 1. A , A , P are random variables. Modulus of elasticity is E = 2.0 × 10<br />

1 2<br />

KPa. Moments of inertia of beams and columns are I = A 2<br />

/12<br />

1 1 and I 2<br />

= A /6<br />

2 2<br />

. <strong>The</strong> limit state function<br />

is defined as in [20]:<br />

g( x) = 0.01 −u<br />

( x )<br />

3<br />

(20)<br />

Where, u ( x ) is the horizontal displacement of point 3. <strong>The</strong> statistical parameters are shown in Table 4. All the<br />

3<br />

variables are assumed to be uncorrelated.<br />

P<br />

2 A 2 3<br />

A1 A1<br />

4<br />

1 4<br />

4<br />

Figure 1 <strong>The</strong> diagram of the portal frame (unit:m)<br />

Table 4 <strong>The</strong> statistical parameters of variables of example 3<br />

Coefficient of<br />

Variables Uint Mean μ<br />

variation<br />

Distribution Type<br />

A1<br />

m 2 0.36 0.036 Log-normal<br />

A2<br />

m 2 0.18 0.018 Log-normal<br />

P kN 20 5 Extreme-I<br />

<strong>The</strong>re are 3 random variables. In accordance with the distribution, samples were distributed uniformly within<br />

μ − 3 σ, μ + 3σ<br />

. In order to ensure the accuracy of response surface, uniform distribution experimental<br />

[ ]<br />

2<br />

design table ( 30<br />

30 )<br />

U is used to decide 30 tests. <strong>The</strong> failure probability is 2.299×10 -3 and the reliability<br />

index is 2.8339 calculated by UDM-SVM RSM proposed in this paper. <strong>The</strong> failure probability obtained by<br />

Monte-Carlo simulation based on importance sampling using 2000 simulations is 2.3223×10 -3 (Deng et al.<br />

2005) and the reliability index is 2.8307 . It is evident that these results correspond quite well.<br />

Example 4<br />

Figure 2 is a calculation diagram of plane frame structure with three spans and 12 layers. Elasticity Modulus of<br />

7<br />

each element is E = 2.0× 10 Pa . <strong>The</strong> relationship between inertia moment and area of element section<br />

is I = α A i2 ( i = 1,2,3,4,5) . <strong>The</strong> area A<br />

i i i<br />

of element section and external load P are chosen as random<br />

variables. To consider the condition of normal use, the maximum allowable deformation [ U ] is equal to 0.096<br />

according to the requirements of the specification. <strong>The</strong>refore, the limit state function can be expressed as:<br />

G = 0.096 − U ( A, A , A , A , A , P)<br />

(21)<br />

max 1 2 3 4 5<br />

<strong>The</strong> statistical characteristics of variables are shown in tableⅡ.<br />

-404-


P<br />

P<br />

P<br />

P<br />

P<br />

P<br />

P<br />

P<br />

P<br />

P<br />

P<br />

P<br />

4 5 4 A<br />

2 4 1 5 1 4 2<br />

2 4 1 5 1 4 2<br />

2 4<br />

1<br />

5<br />

1<br />

4 2<br />

2<br />

4<br />

1<br />

5<br />

1<br />

4<br />

2<br />

2<br />

4<br />

1<br />

5<br />

1<br />

4<br />

2<br />

2<br />

4<br />

1<br />

5<br />

1<br />

4<br />

2<br />

1 3 3<br />

4 5<br />

1<br />

4<br />

1 3 3 1<br />

4 5 4<br />

1<br />

4<br />

3<br />

5<br />

3<br />

1<br />

4<br />

1 4 3 5 3 4 1<br />

1<br />

4<br />

3<br />

5<br />

3<br />

4<br />

1<br />

1 3 3 1<br />

12x4=48<br />

7.5 3.5 7.5<br />

Figure 2 Diagram of the framework(unit:m)<br />

Table5 Statistical parameters of variables<br />

Standard Distribution<br />

Variables Unit Mean μ<br />

αi<br />

deviationσ Type<br />

A m 2 0.25 0.025 Log-normal 0.08333<br />

1<br />

A m 2 0.16 0.016 Log-normal 0.08333<br />

2<br />

A m 2 0.36 0.036 Log-normal 0.08333<br />

3<br />

A m 2 0.2 0.02 Log-normal 0.26670<br />

4<br />

A m 2 0.15 0.015 Log-normal 0.20000<br />

5<br />

P kN 30 7.5 Extreme-I<br />

<strong>The</strong>re are 6 random variables. In accordance with the distribution, samples were distributed uniformly within<br />

μ − 3 σ , μ + 3σ<br />

. In order to ensure the accuracy of response surface, uniform distribution experimental design<br />

[ ]<br />

U is used to decide 25 tests. Put these into deterministic finite element model established and then<br />

6<br />

table ( 5<br />

25 )<br />

the displacements of the structure are obtained.<br />

<strong>The</strong> reliability index from Monte-Carlo simulation based on importance sampling using 2000 simulations is<br />

1.4391 (Zhao 1996) and from traditional RSM is 1.4538(Zhao 1996). <strong>The</strong> reliability index based on Kringing<br />

simulation using 60 simulations is 1.4308 (Zhang et al. 2005). <strong>The</strong> reliability index is 1.4408 calculated by<br />

UDM-SVM RSM in this paper. It is evident that these results correspond quite well.<br />

CONCLUSIONS<br />

<strong>The</strong> reliability method based on support vector machine and uniform design method was discussed. <strong>The</strong><br />

calculation results of examples show that the SVM can well approximate the limit state surface on the case of<br />

the limit state surface with a higher degree of nonlinearity. And it has high accuracy. Training data determined<br />

by the uniform design method can significantly improve the computational efficiency of SVM. <strong>The</strong> analysis<br />

results show that the method is feasible and has a good prospect, which provides a new way for structural<br />

reliability analysis.<br />

REFERENCES<br />

Guan, X. L. and Melchers, R. E. (2001). “Effect of response surface parameter variation on structural reliability<br />

estimates”, Structural Safety, 16(3), 229-230.<br />

Cortes, C. and Vapnik, V. N. (1995). “Support vector networks”, Machine Learning, 20(3), 273-297.<br />

Vapnik, V. N. (1995). <strong>The</strong> Nature of Statistical Learning <strong>The</strong>ory, New York: Springer, Verlag.<br />

Vapnik, V. N. (1999). “An overview of statistical learning theory”, IEEE Transaction on Neural Networks, 10(5),<br />

988-998.<br />

Jin, W. L., Tang, C. X. and Chen, J. (2007). “SVM based on response surface method for structural reliability<br />

analysis”, Chinese Journal of Computational Mechanic, 24(6), 713-718.<br />

-405-


Zhao, W. and Liu, J. K. (2008). “Iterative sequence response surface method for reliability computation based<br />

on support vector regression”, Journal of Mechanical Strength, 30(6), 916-920.<br />

Jin, W. L. and Yuan, X. X. (2007). “Response surface method based on LS-SVM for structural reliability<br />

analysis”, Journal of Zhejiang <strong>University</strong> (Engineering Science), 41(1), 44-47.<br />

Kim, S. H. and Na, S. W. (1997). “Response surface method using vector projected sampling points”, Structural<br />

Safety, 19(l), 3-19.<br />

Zheng, Y. and Das, P. K. (2000). “Improved response surface method and its application to stiffened plate<br />

reliability analysis”, Engineering Structure, 22(5), 544-551.<br />

Elhewy, A. H., Mesbahi, E. and Pu, Y. (2006). “Reliability analysis of structures using neural network method”,<br />

Probabilistic Engineering Mechanics, 21, 44-53.<br />

Rajashekhar, M. R. and Ellingwood, B. R. (1993). “A new look at the response surface approach for reliability<br />

analysis”, Structural Safety, 12, 205-22.<br />

Deng, J., Gu, D. and Li, X. et al. (2005). “Structural reliability analysis for implicit performance functions using<br />

artificial neural network”, Structural Safety, 35(3), 459-467.<br />

Zhang, Q. (2005). Structural reliability analysis and optimization based on Kriging technique, Dalian university<br />

of technology.<br />

Zhao, G. F. (1996). Reliability theory and applications of engineering, Dalian university of technology.<br />

-406-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

SHAKING TABLE TEST OF SEMI-ACTIVE FRICTION TUNED MASS DAMPER<br />

FOR STRUCTURAL CONTROL<br />

Chi-Chang Lin 1 , Ging-Long Lin 1 , Yu-Bo Ho 1 and Lyan-Ywan Lu 2<br />

1 Department of Civil Engineering, National Chung Hsing <strong>University</strong>,<br />

Taichung 40227, Taiwan, ROC.<br />

2<br />

Department of Construction Engineering, National Kaohsiung First <strong>University</strong> of Science and Technology,<br />

Kaohsiung 824, Taiwan, ROC<br />

ABSTRACT<br />

A tuned mass damper (TMD) system is a device consisting of a mass, a spring and a damper attached to the<br />

structure to reduce the dynamic responses of the structure. <strong>The</strong> design and application of linear typed tuned mass<br />

damper systems are well developed. Nonlinear TMD systems are still under developing. A friction-type TMD<br />

which is a nonlinear TMD has the advantages of energy dissipation via a friction mechanism. However, a<br />

passive-friction TMD (PF-TMD) has such disadvantages as a fixed and pre-determined slip load and may lose<br />

its tuning and energy dissipation capability while in stick state. In this paper, a semi-active friction TMD<br />

(SAF-TMD) which consists of a mass and a semi-active friction device is employed to overcome these<br />

disadvantages. <strong>The</strong> friction force of the semi-active friction device is controllable. A shaking table test of the<br />

SAF-TMD system was conduced to illustrate its vibration control effectiveness. A non-sticking friction (NSF)<br />

control law was applied to control the SAF-TMD. <strong>The</strong> test results demonstrate that the dynamic responses are<br />

very consistent with the theoretical ones obtained from numerical simulation. This verifies the feasibility and<br />

efficiency of the SAF-TMD system for structural control.<br />

KEYWORDS<br />

Tuned mass damper, semi-active control, variable friction, shaking table test.<br />

INTRODUCTION<br />

Tuned mass damper (TMD) systems, which were first proposed by Frahm 1911, and is a widely utilized strategy<br />

in the area of structural control, including high-rise buildings, observatory towers, building floors, railway<br />

bridges and pedestrian bridges against natural and man-made loadings (Soong and Spencer 2002; Lin et al. 2005;<br />

Zivanovic and Pavic 2005; Chen and Wu 2008; Chang et al. 2010). A TMD consists of an added mass with<br />

properly functioning spring and damping elements to provide frequency-dependent damping in a primary<br />

structure. A TMD absorbs the vibration energy of a structure and dissipates it via the damping mechanism. <strong>The</strong><br />

design and application of traditional linear TMDs are well developed (Den Hartog 1956; Sadek et al. 1997; Lin<br />

et al. 2000; Bakre and Jangid 2007; Ueng et al. 2008; Wang et al. 2009; Lin et al. 2010a); however,<br />

nonlinear-type TMDs are still in the developmental stages.<br />

Energy dissipation using a friction mechanism is a well-known technology for vibration mitigation of seismic<br />

structures (Lu et al. 2009). Some studies have suggested using friction-type TMD systems, i.e., non-linear TMD<br />

systems, for this purpose (Ricciardelli and Vickery 1999; Almazan et al. 2007). Compared with traditional linear<br />

TMDs, friction-type TMDs have the following advantages: (1) energy dissipates via the friction mechanism and<br />

(2) a suspended mechanism is unnecessary as the space demand is significantly smaller than that of the<br />

pendulum-type TMD.<br />

However, slip force is typically a pre-determined fixed value; thus, a passive-friction TMD (PF-TMD) slips (i.e.,<br />

is activated) and dissipates seismic energy only when a frictional force exerted by seismic TMD motion exceeds<br />

the constant slip force. A PF-TMD does not differ from a mass added to a primary structure when the PF-TMD<br />

is not in a slip state. That is, a PF-TMD in a stick state may lose its tuning and energy dissipating abilities.<br />

Determining slip force magnitude is key to designing structures with PF-TMD. A PF-TMD is typically designed<br />

to withstand earthquakes of a specified intensity. <strong>The</strong>refore, such a PF-TMD may perform well during<br />

earthquakes with the assumed intensity, and perform poorly during earthquakes with different intensities.<br />

Thus, a semi-active-friction TMD (SAF-TMD), which is a semi-active TMD (Abe 1996; Aldemir 2003; Lin et<br />

-407-


al. 2005; Nagarajaiah and Sonmez 2007; Nagarajaiah 2009; Chey et al. 2010), is suggested to improve the<br />

performance of the PF-TMD. By regulating the slip force applied to the friction interface in real time, the<br />

SAF-TMD can adjust its slip force in response to structural motion and/or external excitation; thus, the<br />

SAF-TMD can respond to earthquakes with different intensities. Moreover, the SAF-TMD has the following<br />

features that differentiate it from active mass dampers (AMDs). (1) It does not require a significant amount of<br />

control energy as controlling the clamping force does not generally require a large control stroke. (2) It does not<br />

pump energy into the controlled structures; thus, the effects of control spill-over and control instability are<br />

eliminated. (3) It has a control force restraint, i.e., it can only provide a resistance (passive) force to the<br />

controlled structure. In other words, the friction force produced by a SAF-TMD is always in a direction opposite<br />

to that of the device in motion.<br />

A SAF-TMD consists of a mass block and a variable friction device (VFD) (Lin et al. 2010). <strong>The</strong> many<br />

innovative VFDs described in the structural control literature include Kannan et al. 1995, who applied hydraulic<br />

and electromagnetic driven forces to regulate the clamping force of a semi-active friction device. Other<br />

researchers (Xu et al. 2008; Chen and Chen 2004; Lu et al. 2010) controlled the clamping force of a semi-active<br />

friction device via an embedded piezoelectric stack actuator for generating high stress with very low electrical<br />

current when the piezoelectric actuator is properly confined. All these studies agree that combining a<br />

semi-active friction device with TMD systems may be a feasible technology.<br />

To provide this adaptability, an SAF-TMD implementation must include a controller with a control law for<br />

on-line determination of the clamping forces in the VFD. A major advance in friction controller technology is<br />

the use of bang-bang control laws for VFD, which was first proposed by Kannan et al. 1995. <strong>The</strong>ir methods are<br />

simple and merely require the measurement of the velocity directions of the controlled structure. However,<br />

repetitive on and off damper actions can amplify structural acceleration by causing discontinuous control forces<br />

that exert a high-frequency structural response. Inaudi et al. 1997 proposed a modulated homogenous friction<br />

control strategy for producing a slip force proportional to the prior local peak of the damper deformation.<br />

Similarly, this algorithm may also cause discontinuous control forces. In a subsequent study, Yang and Agrawal<br />

2002 modified the Inaudi’s method for controlling isolated structures subject to near-fault ground motion. Lin et<br />

al. 2010b applied a non-sticking friction (NSF) controller that is able to keep the SAF-TMD in its slip state<br />

constantly for an earthquake with any intensity.<br />

For feasibility tests of the proposed SAF-TMD, a prototype SAF-TMD was fabricated and tested dynamically<br />

via a shaking table test. Firstly, the configuration and the constituent elements of the SAF-TMD are explained,<br />

and the mathematic model and the numerical analysis method for structure with an SAF-TMD are discussed.<br />

<strong>The</strong> test setup for the shaking table test of the prototype SAF-TMD, and the test results are explained. <strong>The</strong><br />

experimental performance of the SAF-TMD is also compared to those of its uncontrolled and passive<br />

counterparts. Finally, conclusions are presented.<br />

SEMI-ACTIVE FRICTION TUNED MASS DAMPER (SAF-TMD)<br />

Configuration of SAF-TMD<br />

In order to evaluate the performance of the proposed SAF-TMD, the seismic response of a prototype SAF-TMD<br />

was tested via an experiment using a shaking table. Figure 1 is a schematic diagram of the prototype SAF-TMD,<br />

which mainly consists of a sliding platform and a piezoelectric friction damper (PFD) (Lu et al. 2010). <strong>The</strong> main<br />

components of the sliding platform are the guide rails, sliding blocks and springs. <strong>The</strong> springs provide not only<br />

stiffness and resilience, but also a tuning frequency.<br />

Mass block<br />

Sliding platform<br />

Piezoelectric friction damper (PFD)<br />

Load cell Friction pad Friction bar<br />

System<br />

dynamic response<br />

Friction bar<br />

Spring<br />

Controller<br />

(PC+ A/D card)<br />

Control voltage<br />

(DC 0-10 V)<br />

Sliding block<br />

Guide rail<br />

Figure 1 Schematic diagram of the SAF-TMD.<br />

Pre-compression screw<br />

Driving voltage<br />

(DC 0-1000V)<br />

Voltage Amplifier<br />

Figure 2 Control diagram of the PFD in the SAF-TMD.<br />

-408-


Figure 2 is a photograph of the interior layout of the PFD. <strong>The</strong> photograph shows that a piezoelectric actuator<br />

embedded in the PFD provides a controllable normal force (or called clamping force) between the friction pads<br />

and friction bar, so the friction force provided by the PFD can be regulated by a control law. When the<br />

SAF-TMD is excited during an earthquake, the friction force produced by the relative motion between the<br />

friction pad of the PFD and the friction bar will provide energy dissipation capability for the SAF-TMD, and as<br />

a result the response of primary structure and the TMD stroke can be attenuated.<br />

Control of PFD<br />

Figure 2 is a control block diagram of the PFD. <strong>The</strong> figure shows that a voltage amplifier and a controller are<br />

usually needed to control the PFD. <strong>The</strong> controller can be a simple digital controller consisting of a<br />

micro-computer (or a PC) and an analog/digital (A/D) converter card. <strong>The</strong> micro-computer calculates the control<br />

command based on the sensor measurement of the current system response whereas the A/D card converts the<br />

computed digital command to an analog signal, which is usually a DC voltage below 10V. However, the<br />

piezoelectric actuator may require a driving DC voltage up to 1000V or higher; therefore, the control voltage<br />

provided by the controller is not sufficient to drive the piezoelectric actuator directly. <strong>The</strong>refore, a voltage<br />

amplifier is needed to obtain a control voltage sufficient for controlling the piezoelectric actuator. <strong>The</strong><br />

experiment in this study used an amplifier with a gain of 100V/V to amplify a 10V control signal to a driving<br />

voltage of 1000V. On the other hand, the electric current required for the piezoelectric actuator is usually at a<br />

range of several mA, so the control energy demand for the PFD is minimal.<br />

As mentioned above, the friction force of the PFD is controlled by regulating the normal force produced by a<br />

piezoelectric actuator. <strong>The</strong> normal force, which is applied to the friction interface of the PFD, can be expressed<br />

by the following equation<br />

N( t)<br />

= N0 + C V ( t)<br />

(1)<br />

z<br />

where N(t) is total normal force, N 0 is pre-compression force (which produced by adjusting the pre-compression<br />

screw shown in Figure 2), V(t) is driving voltage for the piezoelectric actuator, and C z , which is the key<br />

parameter, is the piezoelectric coefficient of the piezoelectric actuator. Equation (1) shows that the force increase<br />

is proportional to the piezoelectric coefficient C z . In terms of physical properties, C z is the trusting force of the<br />

actuator generated by a unit driving voltage; therefore, C z provides a measure of the efficiency of the<br />

piezoelectric actuator. A larger C z implies that a higher thrusting force can be generated by the actuator with a<br />

given voltage. Since the piezoelectric actuator induced by the input voltage is elongated by only several tens μm<br />

(10 -6 meters), the value of C z is highly dependent on the confinement boundary condition of the actuator. In this<br />

study, a test was conducted to identify the actual value of the coefficient C z for the SAF-TMD.<br />

According to the Coulomb friction law, the maximum friction force u d,max (t), i.e., slip force, of the PFD should<br />

correlate with normal force N(t). By using Eq. (1), this slip force can be written as<br />

u<br />

t)<br />

= μ N ( t)<br />

= μ ( N C V ( ))<br />

(2)<br />

d , max ( d<br />

d 0 + z t<br />

where μ d denotes the friction coefficient of the PFD. Note that u d,max (t) in Eq. (2) represents the absolute value of<br />

the maximum friction. Equations (1) and (2) show that, by controlling the driving voltage V(t), the normal force<br />

N(t) as well as the slip force u d,max (t) of the PFD can be altered in a desired manner.<br />

Non-sticking friction control<br />

As noted above, an SAF-TMD generally requires an on-line control algorithm to determine the driving voltage<br />

V(t) of the piezoelectric actuator. This section explains the control law and its important role in the performance<br />

of the SAF-TMD system, and this section will explain the control law employed in this study. This control law<br />

is developed based on the non-sticking friction (NSF) controller (He et al. 2003; Ng and Xu 2007). In order to<br />

suit the control of the SAF-TMD system, whose control command is the driving voltage V(t), the NSF control<br />

method is expressed as<br />

V ( t)<br />

= Vmax tanh( β v&<br />

( t) )<br />

(3)<br />

s<br />

where β is a control parameter set by the control designer. A larger value of β leads to the voltage ratio<br />

increasing more rapidly from 0.0 to 1.0 as the sliding velocity v& s<br />

(t)<br />

increases. By employing the smooth<br />

function tanh(x), Eq. (3) is able to prevent the abrupt change of the normal force N(t), as seen in Eq. (1). A larger<br />

β leads to a function value increasing rapidly to 1.0. <strong>The</strong>refore, clamping force V(t) increases rapidly as β<br />

increases. Equation (3) was used as the control law in the shaking table test to compute the on-line command of<br />

V(t). This control law is very easily implemented since it only requires measurement of sliding velocity v&<br />

s<br />

(t)<br />

for the SAF-TMD. Finally, substituting V(t) from Eq. (3) in (1) gives the following normal force function:<br />

-409-


N t)<br />

= N + C V tanh( β v&<br />

( ) )<br />

(4)<br />

( 0 z max<br />

s t<br />

<strong>The</strong> last equation implies that, when using the control law given in Eq. (3), the normal force of the PFD has a<br />

lower and an upper bound, i.e.,<br />

N ≤ N t)<br />

≤ ( N + C )<br />

(5)<br />

0 ( 0 zVmax<br />

MODELING OF A STRUCTURE WITH A SAF-TMD<br />

Figure 3 shows the mathematical model used for numerical analysis. <strong>The</strong> figure shows that the sliding platform<br />

of the SAF-TMD is modeled by a spring of stiffness k i and a friction element with friction coefficient μ i . <strong>The</strong><br />

former simulates stiffness due to the resilient mechanism whereas the latter models the friction effect of the<br />

guide rail of the sliding platform. Figure 3, however, shows that the PFD is modeled by a variable friction<br />

element with a friction coefficient μ d , and the symbol k d denotes the axial stiffness of the PFD. <strong>The</strong> mass of the<br />

SAF-TMD and the primary structure are denoted by m s and m p , respectively. Moreover, the damping and<br />

stiffness of the primary structure are denoted by c p and k p , respectively. Relative-to-the-ground displacements of<br />

the SAF-TMD and the primary structure are denoted by x s and x p , respectively. Based on the mathematical<br />

model, the dynamic equation of the system can be rewritten as<br />

where<br />

⎧x<br />

p ( t)<br />

⎫<br />

x ( t)<br />

= ⎨ ⎬ ,<br />

⎩vs(<br />

t)<br />

⎭<br />

M & x<br />

t)<br />

+ Cx&<br />

( t)<br />

+ Kx(<br />

t)<br />

= D ( u ( t)<br />

+ u ( t))<br />

+ E & x<br />

( )<br />

(6)<br />

( 2 d i 1 g t<br />

⎧ 1 ⎫<br />

D = ⎨ ⎬ ,<br />

2<br />

⎩−1⎭<br />

⎧−<br />

m p ⎫<br />

E = ⎨ ⎬ ,<br />

1<br />

⎩−<br />

ms<br />

⎭<br />

⎡mp<br />

0 ⎤<br />

M = ⎢ ⎥<br />

,<br />

⎣ms<br />

ms<br />

⎦<br />

c p<br />

⎡ 0⎤<br />

C = ⎢ ⎥<br />

,<br />

⎣ 0 0⎦<br />

⎡k<br />

p − ki<br />

⎤<br />

K = ⎢ ⎥<br />

(7)<br />

⎣ 0 ki<br />

⎦<br />

In Eq. (6), x(t) denotes the vector containing the system responses consisting of structural displacement x p (t) and<br />

TMD stroke v s (t); & x g (t)<br />

is the ground acceleration due to an earthquake; D 2 and E 1 denote the force placement<br />

matrices for the SAF-TMD system and the excitation, respectively. <strong>The</strong> matrices M, C and K represent the mass,<br />

damping and stiffness matrices of the SAF-TMD control system, respectively. <strong>The</strong> friction forces (damper<br />

forces) of the PFD and sliding platform are denoted by u d (t) and u i (t), respectively. Moreover, the forces denoted<br />

by u d (t) and u i (t) on the right-hand side of Eq. (6) are the nonlinear friction effects of the SAF-TMD. Note that<br />

u d (t) is the controllable damper force provided by the PFD, u i (t) is an uncontrollable nonlinear force. <strong>The</strong>refore,<br />

the dynamic response of the SAF-TMD system can be attenuated by altering the force u d (t) in real time. For the<br />

purpose of numerical simulation in this study, the Coulomb friction law is assumed to govern friction material<br />

behavior in the damper and sliding platform, and the friction materials are assumed to have an equal static and<br />

dynamic friction coefficient. When friction force terms u d (t) and u i (t) are included in the model of a structure<br />

controlled by a SAF-TMD, the model becomes nonlinear, and numerical methods are generally needed to<br />

analyze the dynamic behavior of the system. This study therefore applies shear balance method (SBM), a<br />

numerical method of simulating a structure equipped with a friction-type TMD. <strong>The</strong> SBM is widely used to<br />

simulate dynamic response in structures with friction-type devices such as friction damper (Lu et al. 2006) or<br />

friction isolator (Wang et al. 1998). <strong>The</strong> SBM has proven efficient and generally accurate in analyzing structures<br />

with frictional devices. Lin et al. 2010c comprehensively discussed the formulation and application of an SBA<br />

algorithm in analysis of friction-type TMDs.<br />

v s(t)<br />

k d<br />

k i<br />

μ i<br />

μ d<br />

N(t)<br />

m s<br />

(SAF-TMD)<br />

x s(t)<br />

x p(t)<br />

m p<br />

(primary structure)<br />

c p<br />

k p<br />

Figure 3 Mathematic model of a SDOF structural system with the SAF-TMD.<br />

-410-


SHAKING TABLE TEST PROGRAM<br />

System identification of the SAF-TMD and the primary structure<br />

<strong>The</strong> performance of the proposed SAF-TMD system was evaluated by shaking table test performed at the<br />

earthquake simulation laboratory of the National Center for Research on Earthquake Engineering (NCREE),<br />

Taipei, Taiwan. Figure 4 shows the setup for the shaking table test. <strong>The</strong> figure shows that the prototype<br />

SAF-TMD was placed on the top of a primary structure. <strong>The</strong> mass of the primary structure is represented by an<br />

assembly of several mass blocks, which are connected to each other by vertical tension rods at the four corners<br />

of the blocks. <strong>The</strong> stiffness of the primary structure, however, is represented by a concave rail-wheel system<br />

mounted under the mass blocks of the primary structure (Figure 4). <strong>The</strong> concave rail-wheel system (Figure 5)<br />

was designed to provide constant oscillation period of the primary structure.<br />

Figure 4 Photo of the shaking table test.<br />

Figure 5 Concaved guide rail and roller wheel (primary structure).<br />

To obtain the system parameters, a parameter identification test was performed. In the identification test, both<br />

the SAF-TMD and the primary structure are placed directly on the shaking table. <strong>The</strong> system realization using<br />

information matrix (SRIM) identification technique (Juang 1997) is used to estimate the natural frequencies and<br />

damping ratios of the primary structure and the SAF-TMD. Based on the input (table acceleration) and output<br />

(primary structural/SAF-TMD acceleration) signals, modal frequency and damping ratio can be measured<br />

precisely. Table 1 shows the identification results for the primary structure and for the prototype SAF-TMD. <strong>The</strong><br />

period of the primary structure (T p ) is 1.90 sec, which is the typical period in a 20-floor building. <strong>The</strong> mass of<br />

SAF-TMD (m s ) was adjusted to tune the SAF-TMD frequency to that of the primary structure. Table 1 shows<br />

that the mass ratio (μ) of the SAF-TMD system is<br />

ms<br />

μ = = 4.6%<br />

(8)<br />

m<br />

p<br />

Moreover, Eq. (1) shows that C z and N 0 are the two key parameters for controlling the normal force of the PFD.<br />

Figure 6 shows the experimental relation between the normal force N(t) and the driving voltage V(t), which was<br />

plotted in a calibration test to identify the two parameters. Note that the figure also depicts a regression line.<br />

Figure 6 shows that N(t) is almost linearly proportional to V(t). Based on Eq. (1), it should be easily realized that<br />

the slope of the regression line represents the piezoelectric coefficient C z while the intersection of the regression<br />

line with the vertical axis of the coordinates depicts pre-compression force N 0 . Table 1 shows that the values of<br />

C z and N 0 , which were given in Figure 6, are 0.46 N/V and 130N, respectively. Based on the identified system<br />

parameters listed in Table 1, a parametric study is conducted, and β=4.0 is chosen for the NSF controller in the<br />

shaking table test.<br />

-411-


Table 1 System parameters of the system<br />

Component Item Value<br />

Primary structure<br />

SAF-TMD<br />

Sliding<br />

platform<br />

PFD<br />

Mass (m p)<br />

Period (T p)<br />

2100 kg<br />

1.90 sec<br />

Damping ratio (ζ s) 4.90 %<br />

Maximum story drift (x p) ± 0. 25 m<br />

Mass (m s)<br />

Spring stiffness (k i)<br />

Period (T i)<br />

96 kg<br />

1017 N/m<br />

1.93 sec<br />

Friction coefficient ( μ<br />

i<br />

) 0.009<br />

Maximum stroke (x s) ± 0. 15 m<br />

Friction coefficient of damper ( μ ) 0.20<br />

Damper stiffness (k d)<br />

Piezoelectric coefficient ( C )<br />

Pre-compression force (N 0)<br />

Range of driving voltage (V)<br />

z<br />

d<br />

10 6 N/m<br />

0.46 N/Volt<br />

130 N<br />

0-1000 Volt<br />

N 0<br />

Normal force (N)<br />

Voltage & Normal force<br />

800<br />

y = 0.46*x + 1.3e+002<br />

700<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

Experimental<br />

<strong>The</strong>oretical<br />

0<br />

0 200 400 600 800 1000<br />

Control Voltage (V)<br />

Figure 6 Force and voltage relationship of the PFD.<br />

Test setup and input ground acceleration<br />

Figure 4 depicts the test setup. One velocity sensor and one accelerometer are placed on the shaking table (the<br />

ground), on the primary structure and on the SAF-TMD. Additionally, two LVDTs were used to measure the<br />

relative-to-the-ground displacement of the primary structure and the TMD stroke. A load cell was embedded in<br />

the PFD to measure normal force N(t) (Figure 2). <strong>The</strong> 1940 El Centro earthquake was applied in the shaking<br />

table test as the ground acceleration.<br />

TEST RESULTS AND DISCUSSION<br />

Comparison of experimental and theoretical results<br />

This section reports the experimental findings of the shaking table test. To ensure accurate experimental results,<br />

the test data were first verified by the theoretical results simulated by SBM. Additionally, acceleration signals<br />

measured by an accelerometer placed directly on the shaking table were used as input to simulate ground<br />

excitations. Figure 7 compares the experimental and simulated responses of the SAF-TMD system subjected to<br />

the El Centro earthquake (PGA=500gal). Note that each figure contains six sub-figures in the following<br />

sequence: (a) time-history of the structural displacement x p (t), (b) time-history of the absolute acceleration of the<br />

primary structure, (c) time-history of the TMD, (d) time-history of the normal force N(t), (e) total shear force of<br />

the SAF-TMD S s (t) vs. TMD stroke v s (t), (f) damper force u d (t) of the PFD vs. TMD stroke v s (t). Sub-figures (e)<br />

and (f) of Figure 7 also show the hysteretic behaviors of the SAF-TMD and PFD, respectively. In Figure 7, the<br />

following observations can be made: (1) all measurement data are consistent with the predicted SAF-TMD<br />

system behavior, i.e., the test data are reliable, and numerical method is effective for analyzing SAF-TMD<br />

system dynamic response. (2) Figure 7(d) demonstrates that the normal force N(t) of the PFD damper has been<br />

altered by the embedded piezoelectric actuator. This behavior differs from the constant normal force typically<br />

observed in a PF-TMD. This also implies that the slip force of the PFD has been altered by the actuator. (3) Due<br />

to the complicated friction behavior as well as measurement noise, the discrepancy in the experimental and<br />

theoretical hysteresis loops of the PFD (Figure 7(f)) is relatively larger than that in other system responses.<br />

Nevertheless, this discrepancy does not significantly affect the global responses of the SAF-TMD because for<br />

both the displacement and acceleration responses, the experimental results are consistent with the theoretical<br />

results (sub-Figs. (a) and (b) in Figure 7).<br />

0.2<br />

0.15<br />

Structural displacement<br />

<strong>The</strong>oretical<br />

Experimental<br />

2<br />

1.5<br />

Structural acceleration<br />

<strong>The</strong>oretical<br />

Experimental<br />

0.2<br />

0.15<br />

TMD stroke<br />

<strong>The</strong>oretical<br />

Experimental<br />

Displacement (m)<br />

0.1<br />

0.05<br />

0<br />

-0.05<br />

-0.1<br />

Acceleration (m/s 2 )<br />

1<br />

0.5<br />

0<br />

-0.5<br />

-1<br />

Stroke (m)<br />

0.1<br />

0.05<br />

0<br />

-0.05<br />

-0.1<br />

-0.15<br />

-1.5<br />

-0.15<br />

-0.2<br />

0 10 20 30 40 50 60<br />

Time (s)<br />

-2<br />

0 10 20 30 40 50 60<br />

Time (s)<br />

-0.2<br />

0 10 20 30 40 50 60<br />

Time (s)<br />

(a) Structural displacement (b) Structural acceleration (c) TMD stroke<br />

-412-


Force (N)<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

Normal force<br />

<strong>The</strong>oretical<br />

Experimental<br />

total shear force (N)<br />

300<br />

200<br />

100<br />

0<br />

-100<br />

-200<br />

<strong>The</strong>oretical<br />

Experimental<br />

Total hysteresis loop<br />

Friction Force (N)<br />

200<br />

150<br />

100<br />

50<br />

0<br />

-50<br />

-100<br />

-150<br />

Damper hysteresis loop<br />

<strong>The</strong>oretical<br />

Experimental<br />

0<br />

0 10 20 30 40 50 60<br />

Time (s)<br />

-300<br />

-0.2 -0.1 0 0.1 0.2<br />

TMD Stroke (m)<br />

-200<br />

-0.2 -0.1 0 0.1 0.2<br />

TMD Stroke (m)<br />

(d) Normal force (e) Total hysteresis loop (f) Damper hysteresis loop<br />

Figure 7 Comparison of experimental and theoretical results due to El Centro earthquake (PGA=500gal,<br />

N 0 =130 N).<br />

Definition of performance indices<br />

<strong>The</strong> comparison results were quantified by performance evaluations of the SAF-TMD system in terms of the<br />

five performance indices (J 1 -J 5 ) in Table 2. <strong>The</strong> indices J 1 and J 3 in Table 2 represent the peak response ratios of<br />

the structural displacement x p (t) and structural acceleration of the SAF-TMD system, respectively. <strong>The</strong> indices<br />

J 2 and J 4 represent the root-mean-square (RMS) response ratios of x p (t) and & x p , a ( t)<br />

, respectively. Performance<br />

index J 5 represents the peak TMD stroke of the SAF-TMD. Notably, indices J 1 to J 4 were all divided by the<br />

corresponding response values of the uncontrolled system (primary structure only), which are represented by<br />

symbols with a top bar in Table 2. <strong>The</strong>refore, for indices J 1 to J 4 , a value less than one implies that the<br />

SAF-TMD system has a lower response than that of the uncontrolled system.<br />

Table 2 Definition of performance indices<br />

Response Peak structural displacement RMS structural displacement<br />

max( x<br />

p<br />

( t))<br />

w / TMD<br />

RMS(<br />

x<br />

p<br />

( t))<br />

w / TMD<br />

Index J1<br />

=<br />

J<br />

2<br />

=<br />

max( x ( t))<br />

RMS(<br />

x ( t))<br />

p<br />

w / o TMD<br />

p<br />

w / o TMD<br />

Response Peak structural acceleration RMS structural acceleration<br />

Index<br />

max( && x<br />

p,<br />

a<br />

( t))<br />

w / TMD<br />

RMS(<br />

&& x<br />

p,<br />

a<br />

( t))<br />

w / TMD<br />

J<br />

3<br />

=<br />

J<br />

4<br />

=<br />

max( && x<br />

p,<br />

a<br />

( t))<br />

w / o TMD<br />

RMS(<br />

&& x<br />

p,<br />

a<br />

( t))<br />

w / o TMD<br />

Stroke<br />

Peak TMD stroke<br />

Index J = max( v ( ))<br />

5 s<br />

t<br />

Comparison of SAF-TMD and uncontrolled system responses<br />

To demonstrate the control efficiency of the SAF-TMD, this subsection compares the experimental seismic<br />

responses of the SAF-TMD system to the simulated responses of its uncontrolled counterpart system (primary<br />

structure only). In the uncontrolled system, simulated responses were used instead of experimental ones for two<br />

reasons: (1) the uncontrolled system easily exceeds the allowable displacement of 0.25m, which limits<br />

comparison with earthquakes in which PGA is larger. (2) For comparison purposes, it would be very difficult for<br />

the shaking table to generate two exactly same ground accelerations for the SAF-TMD and uncontrolled systems.<br />

<strong>The</strong>refore, simulations of the uncontrolled system for comparison purposes used the same ground acceleration<br />

measured in the SAF-TMD test and the same system parametric values shown in Table 1.<br />

Table 3 shows the performance indices of the SAF-TMD system when data for the selected earthquake was<br />

applied. Table 3 shows that, although the SAF-TMD is less effective in reducing the peak response (J 1 and J 3 ) of<br />

the system, it reduces the RMS structural displacement by 12-18% (see J 2 ) and the RMS structural acceleration<br />

by 12-17% (see J 4 ), as compared to the RMS responses of the uncontrolled system.<br />

Table 3 Performance indices of the SAF-TMD<br />

Earthquake<br />

PGA Disp. index Acc. index<br />

(gal) J 1 J 2 J 3 J 4<br />

300 0.98 0.88 1.03 0.88<br />

El Centro 400 0.94 0.84 0.99 0.85<br />

500 0.91 0.82 0.90 0.83<br />

-413-


Comparison of SAF-TMD and PF-TMD system responses<br />

In this subsection the experimental seismic responses of the SAF-TMD system are compared to the simulated<br />

responses for two passive systems (PF-TMD-1 and PF-TMD-2). <strong>The</strong>oretically, when using the SAF-TMD<br />

model, the response of a PF-TMD can be simulated by pre-selecting an equivalent pre-compression force N 0 and<br />

then setting the control voltage V(t)=0. Table 4 lists the equivalent N 0 for these two PF-TMD systems. For the<br />

PF-TMD-1, N 0 is set to 130 N, which is exactly equal to N 0 of the SAF-TMD used in the test. Restated,<br />

PF-TMD-1 may represent an uncontrolled SAF-TMD. On the other hand, N 0 of the PF-TMD-2 is selected such<br />

that the TMD results in a peak stroke equal to that of the SAF-TMD for a given ground motion. <strong>The</strong>refore, N 0 of<br />

the PF-TMD-2 differs for different PGA levels in Table 4. Figure 8 compares the RMS responses (J 2 and J 4 ,<br />

respectively) and peak TMD stroke (J 5 ) of the three TMD systems when data for the the selected earthquake is<br />

applied with different PGA levels. Note that the plots are for the experimental data for the SAF-TMD, and only<br />

the tested PGA levels are shown. Additionally, the data for PF-TMD-1 and PF-TMD-2 in Figure 8, respectively,<br />

are simulation data. <strong>The</strong> two figures confirm that, (1) for the range of the PGA levels tested, both the<br />

displacement and acceleration performance of the PF-TMD-1 are superior to those of the SAF-TMD and<br />

PF-TMD-1and that, (2) with a lower damping force, the PF-TMD-1 has the best control performance among the<br />

three systems; however, it also results in a peak stroke much higher than those of the SAF-TMD and PF-TMD-2,<br />

especially in earthquakes with high PGA. <strong>The</strong> maximum stroke for the SAF-TMD observed in the test is only<br />

0.15m. (3) With the same peak TMD strokes, both the displacement and acceleration performance of the<br />

SAF-TMD are superior to those of the PF-TMD-2 cases. Figure 9 compares the time histories of the structural<br />

displacement and TMD stroke for the SAF-TMD and PF-TMD-2 systems subjected to the El Centro earthquake<br />

(PGA=500gal). <strong>The</strong> figure shows that the SAF-TMD is more effective in suppressing the structural<br />

displacement compared to the PF-TMD-2, which has the same peak TMD stroke (0.12m). Furthermore, the<br />

SAF-TMD can also prevent a residual TMD stroke, which is observed in the PF-TMD-2 system.<br />

Table 4 Pre-compression force (N 0 ) of the PF-TMD-1 and PF-TMD-2<br />

Earthquake<br />

PGA Pre-compression force N 0 (N)<br />

(gal) PF-TMD-1 PF-TMD-2<br />

300 130 253<br />

El Centro 400 130 313<br />

500 130 474<br />

Disp. reduction ratio<br />

1<br />

0.95<br />

0.9<br />

0.85<br />

0.8<br />

0.75<br />

0.7<br />

0.65<br />

SAF-TMD<br />

PF-TMD-1<br />

PF-TMD-2<br />

Index J2 (El Centro)<br />

Acc. reduction ratio<br />

1<br />

0.95<br />

0.9<br />

0.85<br />

0.8<br />

0.75<br />

0.7<br />

0.65<br />

SAF-TMD<br />

PF-TMD-1<br />

PF-TMD-2<br />

Index J4 (El Centro)<br />

Peak TMD stroke (m)<br />

0.6<br />

0.5<br />

0.4<br />

0.3<br />

0.2<br />

0.1<br />

SAF-TMD<br />

PF-TMD-1<br />

PF-TMD-2<br />

Index J5 (El Centro)<br />

0.6<br />

200 300 400 500 600<br />

PGA (gal)<br />

0.6<br />

200 300 400 500 600<br />

PGA (gal)<br />

0<br />

200 300 400 500 600<br />

PGA (gal)<br />

(a) Index J2 (b) Index J4 (c) Index J5<br />

Figure 8 Comparison of performances indices of various TMD systems to the El Centro earthquake with<br />

different PGA levels<br />

(a) Structural displacement<br />

(b) TMD stroke<br />

-414-


Figure 9 Comparison of SAF-TMD and PF-TMD-2 control systems due to El Centro earthquake (PGA=500gal)<br />

CONCLUSIONS<br />

To enhance the efficiency of a friction-type TMD, a prototype SAF-TMD was fabricated and tested dynamically<br />

via shaking table test. <strong>The</strong> SAF-TMD is composed of a mass block, sliding platform and a piezoelectric friction<br />

damper (PFD). Depending on the feedback signal of the SAF-TMD response, the friction force of the PFD can<br />

be regulated on-line by an embedded piezoelectric actuator driven by a controllable DC voltage. Shaking table<br />

test of the SAF-TMD showed that performance was consistent with the simulation results. This confirmed the<br />

accuracy of the test results as well as the performance of the SAF-TMD system. To evaluate control<br />

performance, the measured seismic responses of the SAF-TMD were also compared to the simulated responses<br />

of its passive counterpart. <strong>The</strong> comparison results demonstrate that the control performance of the SAF-TMD is<br />

better than that of the passive friction TMD provided that both the systems have the same peak TMD stroke. An<br />

added advantage of the SAF-TMD is the elimination of the residual TMD stroke in a passive system.<br />

ACKNOWLEDGMENTS<br />

<strong>The</strong> authors would like to thank the National Science Council of the Republic of China, Taiwan, for financially<br />

supporting this research under Contract No. NSC 95-2625-Z-005-009.<br />

REFERENCES<br />

Abe, M. (1996). “Semi-active tuned mass dampers for seismic protection of civil structures”, Earthquake<br />

Engineering and Structural Dynamics, 25(7), 743-749.<br />

Aldemir, U. (2003). “Optimal control of structures with semi-active tuned mass dampers”, Journal of Sound and<br />

Vibration, 266(4), 847-874.<br />

Almazan, J.L., Llera, J.C., Inaudi, J.A, Lopez-Garcia D., Izquierdo L.E. (2007). “A bidirectional and<br />

homogeneous tuned mass damper: A new device for passive control of vibrations”, Engineering Structures,<br />

29(7), 1548-1560.<br />

Bakre, S.V., Jangid, R.S. (2007). “Optimum parameters of tuned mass damper for damped main system”,<br />

Structural Control and Health Monitoring, 14(3), 448-470.<br />

Chang, M.L., Lin, C.C., Ueng, J.M., Hsieh, K.H., Wang, J.F. (2010). “Experimental study on adjustable tuned<br />

mass damper to reduce floor vibration due to machinery”, Structural Control and Health Monitoring, 17(5),<br />

532-548.<br />

Chen, G., Chen, C. (2004). “Semi-active control of the 20-Story benchmark building with piezoelectric friction<br />

dampers”, Journal of engineering mechanics, ASCE, 130(4), 393-400.<br />

Chen, S.R., Wu, J. (2010). “Performance enhancement of bridge infrastructure systems: Long-span bridge,<br />

moving trucks and wind with tuned mass dampers”, Engineering Structures, 30(11), 3316-3324.<br />

Chey, M.H, Chase, J.G., Mander, J.B., Carr, A.J. (2010). “Semi-active tuned mass damper building systems:<br />

Application”, Earthquake Engineering and Structural Dynamics, 39(1), 69-89.<br />

Den Hartog, J.P. (1956). Mechanical Vibrations 4 th Edtion, McGraw Hill: New York.<br />

Frahm, H. Device for damping vibration of bodies. U. S. Patent 1911, No. 989-958.<br />

He, W.L., Agrawal, A.K., Yang, J.N. (2003). “Novel semi active friction controller for linear structures against<br />

earthquakes”, Journal of Structural Engineering, ASCE, 129(7), 941-950.<br />

Inaudi, J.A. (1997). “Modulated homogeneous friction: a semi-active damping strategy”, Earthquake<br />

Engineering and Structural Dynamics, 26(3), 361-376.<br />

Juang, J.N. (1997). “System realization using information matrix”, Journal of Guidance, Control, and Dynamics,<br />

21(3), 492-500.<br />

Kannan, S., Uras, H.M., Aktan, H.M. (1995). “Active control of building seismic response by energy<br />

dissipation”, Earthquake Engineering and Structural Dynamics, 24(5), 747-759.<br />

Lin, C. C., Ueng, J.M., Huang, T.C. (2000). “Seismic response reduction of irregular buildings using passive<br />

tuned mass dampers”, Engineering Structures, 22(5), 513-524.<br />

Lin, C.C., Wang, J.F., Chen, B.L. (2005). “Train-induced vibration control of high-speed railway bridges<br />

equipped with multiple tuned mass dampers”, Journal of Bridge Engineering, ASCE, 10(4), 398-414.<br />

Lin, C.C., Wang, J.F., Lien, C.H., Chiang, H.W., Lin, C.S. (2010a). “Optimum Design and Experimental Study<br />

of Multiple Tuned Mass Dampers with Limited Stroke”, Earthquake Engineering and Structural Dynamics,<br />

39(14), 1631-1651.<br />

Lin, C.C., Lin, G.L., Wang, J.F. (2010b), “Protection of seismic structures using semi-active friction TMD”,<br />

Earthquake Engineering and Structural Dynamics, 39(6), 635-659.<br />

-415-


Lin, C.C., Lu, L.Y., Lin, G.L., Yang, T.W. (2010c). “Vibration control of seismic structures using semi-active<br />

friction multiple tuned mass dampers”, Engineering Structures, 32(10), 3404-3417.<br />

Lin, P.Y., Chung, L.L., Loh, C.H. (2005). “Semi-active control of building structures with semi-active tuned<br />

mass damper”, Computer–Aided Civil and Infrastructure Engineering, 20(1), 35-51.<br />

Lu, L.Y, Chung, L.L., Wu, L.Y., Lin, G.L. (2006). “Dynamic analysis of structures with friction devices using<br />

discrete-time state-space formulation”, Computers and Structures, 84(15-16), 1049-1071.<br />

Lu, L.Y., Lin, G.L., Lin, C.H. (2009). “A unified analysis model for energy dissipation devices used in seismic<br />

structures”, Computer-Aided Civil and Infrastructure Engineering, 24(1), 41-61.<br />

Lu, L.Y., Lin, G.L., Lin, C.Y. (2010). “Experimental verification of a piezoelectric smart isolation system”,<br />

Structural Control and Health Monitoring, published on-line, DOI: 10.1002/ stc.407.<br />

Nagarajaiah, S., Sonmez, E. (2007). “Structures with semi-active variable stiffness single/multiple tuned mass<br />

dampers”, Journal of Structural Engineering, ASCE, 133(1), 67-77.<br />

Nagarajaiah, S. (2009). “Adaptive passive, semi-active, smart tuned mass dampers: Identification and control<br />

using empirical mode decomposition, Hilbert transform, and short-term Fourier transform”, Structural<br />

Control and Health Monitoring, 16(7-8), 800-841.<br />

Ng, C.L., Xu, Y.L. (2007). “Semi-active control of a building complex with variable friction dampers”,<br />

Engineering Structures, 29(6), 1209-1225.<br />

Ricciardelli, F., Vickery, B.J. (1999). “Tuned vibration absorbers with dry friction damping”, Earthquake<br />

Engineering and Structural Dynamics, 28(7), 707-723.<br />

Sadek, F., Mohraz, B., Taylor, A.W., Chung, R.M. (1997). “Method of estimating the parameters of tuned mass<br />

dampers for seismic applications”, Earthquake Engineering and Structural Dynamics, 26(6), 617-635.<br />

Soong, T.T., Spencer, B.F. (2002). “Supplemental energy dissipation: State-of-the-art and state-of-the-practice”,<br />

Engineering Structures, 24(3), 243-259.<br />

Ueng, J.M., Lin, C.C., Wang, J.F. (2008). “Practical design issues of tuned mass dampers for torsionally-coupled<br />

buildings under earthquake loadings”, Structural Design of Tall and Special Buildings,17(3), 133-165.<br />

Wang, J.F., Lin, C.C., Lian, C.H. (2009). “Two-stage optimum design of tuned mass dampers with consideration<br />

of stroke”, Structural Control and Health Monitoring, 16(1), 55-72.<br />

Wang, Y.P., Chung, L.L., Liao, W.H. (1998). “Seismic response analysis of bridges isolated with friction<br />

pendulum bearings”, Earthquake Engineering and Structural Dynamics, 27(10), 1069-1093.<br />

Xu, Y.L., Ng, C.L. (2008). “Seismic protection of a building complex using variable friction damper:<br />

experimental investigation”, Journal of Engineering Mechanics, ASCE, 134(8), 637-649.<br />

Yang, J.N., Agrawal, A.K. (2002). “Semi-active hybrid control systems for nonlinear building against near-fault<br />

earthquakes”, Engineering Structures, 24(3), 271-280.<br />

Zivanovic, S., Pavic, A., Reynolds, P. (2005). “Vibration serviceability of footbridges under human-induced<br />

excitation: A literature review”, Journal of Sound and Vibration, 279(1-2), 1-74.<br />

-416-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

台 灣 吊 橋 調 查 與 管 理 系 統 之 建 立<br />

陳 銘 鴻 王 仲 宇 廖 慶 隆<br />

( 財 團 法 人 中 華 顧 問 工 程 司 , 台 北 台 灣 )<br />

摘 要 : 本 研 究 針 對 吊 橋 之 過 去 發 展 、 結 構 形 式 以 及 吊 橋 可 能 發 生 之 災 害 , 進 行 初 步 研 判 探 討<br />

後 , 建 立 吊 橋 現 地 勘 查 之 基 本 資 料 表 及 損 傷 評 估 項 目 表 , 初 步 研 擬 出 現 地 調 查 標 準 作 業 程 序 , 並 且<br />

派 員 進 行 現 地 調 查 填 表 、 拍 照 。 經 由 本 計 畫 之 整 理 設 計 , 人 行 吊 橋 調 查 時 之 登 錄 內 容 包 括 「 基 本 資<br />

料 」、「 幾 何 資 料 」、「 結 構 資 料 」 以 及 「 設 計 資 料 」 四 大 項 , 並 且 可 將 吊 橋 分 解 成 (1) 橋 塔 、(2) 橋 面<br />

板 、(3) 主 索 ( 含 跨 座 及 錨 碇 )、(4) 垂 吊 索 、(5) 抗 風 索 ( 含 錨 碇 )、(6) 抗 風 支 索 、(7) 護 欄 等 7 類 的<br />

構 件 。 至 於 現 地 調 查 的 工 作 則 是 著 重 在 結 構 型 式 及 損 傷 程 度 的 紀 錄 , 本 研 究 設 計 之 吊 橋 調 查 內 容 必<br />

須 包 含 吊 橋 的 整 體 照 片 , 以 及 各 構 件 之 細 部 照 片 , 利 用 這 些 照 片 配 合 文 字 之 敘 述 說 明 , 可 使 查 詢 者<br />

完 整 瞭 解 該 吊 橋 之 實 際 情 形 。<br />

為 達 到 未 來 的 吊 橋 管 理 作 業 上 之 完 整 及 便 利 性 , 本 計 畫 也 進 行 吊 橋 管 理 系 統 之 開 發 , 建 置 透 過<br />

網 路 即 可 線 上 檢 視 查 詢 之 「 台 灣 地 區 吊 橋 管 理 系 統 」 資 料 庫 資 訊 管 理 平 台 , 並 請 各 參 與 人 員 將 所 有<br />

調 查 成 果 資 料 輸 入 管 理 系 統 , 上 傳 目 前 所 有 之 吊 橋 現 地 目 視 檢 測 資 料 及 調 查 照 片 , 以 便 於 長 期 管 理<br />

使 用 。 藉 由 這 一 套 「 台 灣 地 區 吊 橋 管 理 系 統 」 之 建 置 , 將 可 使 得 已 調 查 之 61 座 吊 橋 的 管 理 單 位 可<br />

以 有 系 統 的 掌 握 其 所 管 轄 吊 橋 之 現 況 , 並 可 以 提 供 更 多 其 他 吊 橋 主 管 單 位 參 考 仿 照 此 一 調 查 作 法 及<br />

建 檔 模 式 , 對 其 所 負 責 管 轄 的 吊 橋 開 始 進 行 調 查 與 資 料 建 檔 , 以 建 構 全 國 吊 橋 之 清 冊 。<br />

關 鍵 詞 : 吊 橋 目 視 檢 測 管 理 系 統<br />

ESTABLISHMENT OF TAIWAN PEDESTRIAN SUSPENSION BRIDGE<br />

MANAGEMENT SYSTEM<br />

M. H. Chen, C. Y. Wang and C. L. Liao<br />

China Engineering Consultants, Inc., Taipei, Taiwan<br />

Abstract: <strong>The</strong> investigating data sheet for pedestrian suspension bridge was set in this study according to some<br />

field cases. Four major items as basic data, geometry description, structural type and design criteria are<br />

requested to register. Meanwhile, a bridge is decomposed into 7 components as tower, deck, catenary cable<br />

(including saddle and anchor), suspender, wind-resistance cable, wind-resistance hanger and guardrail. <strong>The</strong><br />

investigating work needs to collect photos of every component and detailed description of damaged situation,<br />

and store this information into the management system.<br />

<strong>The</strong> established Taiwan Pedestrian Suspension Bridge Management System is a web-based system, 61 bridges<br />

were studied. It acts as a platform to upload the collecting data and photos. <strong>The</strong> bridge owner/manager can also<br />

access to this platform to reach their needs. <strong>The</strong> next step will encourage more bridge managing unions copying<br />

the investigating procedures and surveying more bridges. <strong>The</strong> final goal is building an integral inventory of<br />

Taiwan pedestrian suspension bridge.<br />

Keywords: pedestrian suspension bridge, visual inspection, management system<br />

-417-


一 、 前 言<br />

人 行 吊 橋 (footbridge or pedestrian bridge) 一 般 是 指 僅 供 行 人 使 用 的 吊 橋 , 部 分 時 機 也 可 供 自 行 車<br />

通 過 使 用 , 或 是 小 型 農 用 搬 運 車 輛 使 用 , 但 不 包 括 小 客 車 以 上 之 車 輛 使 用 。 人 行 吊 橋 不 僅 因 為 其 優<br />

美 的 線 條 在 山 林 間 畫 出 迷 人 的 弧 線 , 同 時 也 在 視 覺 上 點 綴 了 原 本 相 隔 的 兩 地 因 此 有 了 串 連 , 傳 達 出<br />

兩 處 分 離 地 域 的 連 繫 意 象 。 在 世 界 各 地 尤 其 是 生 活 條 件 較 為 富 足 的 國 家 , 人 行 吊 橋 不 僅 是 交 通 功 能<br />

的 發 揮 , 並 且 也 是 美 麗 的 工 藝 與 地 景 雕 塑 的 結 合 。 至 於 在 開 發 中 國 家 而 言 , 在 較 為 貧 窮 的 鄉 下 農 村 ,<br />

一 旦 當 溪 水 暴 漲 時 民 眾 便 無 法 涉 溪 而 過 , 人 行 吊 橋 便 成 為 這 些 聚 落 裡 的 人 們 對 外 就 醫 、 求 學 以 及 採<br />

購 生 活 必 需 品 的 唯 一 聯 外 道 路 。 因 此 在 這 些 地 區 所 設 計 的 吊 橋 便 會 考 慮 到 就 地 取 材 的 可 能 , 並 以 當<br />

地 的 人 力 所 能 完 成 建 造 的 施 工 方 式 為 考 量 [1] 。<br />

由 於 台 灣 許 多 的 吊 橋 都 是 在 早 期 所 建 造 的 , 當 年 主 要 是 作 為 交 通 聯 絡 使 用 , 如 今 大 部 分 的 吊 橋<br />

都 僅 作 觀 光 用 途 , 不 再 肩 負 人 員 物 資 運 送 的 重 任 。 但 是 由 於 吊 橋 缺 乏 有 系 統 的 管 理 , 許 多 歷 史 悠 久<br />

的 吊 橋 亦 鮮 少 定 期 接 受 妥 善 的 檢 測 及 維 修 , 導 致 現 在 台 灣 的 吊 橋 面 臨 著 安 全 上 的 疑 慮 。 另 一 方 面 ,<br />

台 灣 的 人 行 吊 橋 一 直 以 來 並 沒 有 明 確 的 專 責 管 理 單 位 , 從 林 務 局 、 國 家 公 園 管 理 處 、 原 住 民 管 理 單<br />

位 、 縣 市 政 府 、 乃 至 地 方 鄉 鎮 市 公 所 都 有 。 因 此 , 有 系 統 的 建 立 一 套 「 人 行 吊 橋 管 理 系 統 」 將 有 助<br />

於 把 吊 橋 維 護 相 關 事 項 加 以 彙 整 , 也 就 容 易 將 各 項 維 護 作 業 建 立 起 完 善 的 制 度 。<br />

因 此 , 中 華 顧 問 工 程 司 橋 梁 技 術 研 究 中 心 自 行 規 劃 執 行 「 人 行 吊 橋 管 理 系 統 建 置 計 畫 」, 研 究<br />

建 立 一 套 明 確 的 吊 橋 維 護 制 度 及 管 理 系 統 , 以 利 於 把 吊 橋 維 護 相 關 事 項 加 以 彙 整 , 也 就 容 易 將 各 項<br />

維 護 作 業 建 立 起 完 善 的 制 度 。<br />

二 、 吊 橋 結 構 系 統 特 色<br />

吊 橋 基 本 上 是 一 種 主 纜 跨 過 塔 頂 , 藉 由 垂 吊 索 將 橋 面 板 懸 吊 於 空 中 的 一 種 橋 梁 型 式 , 而 主 纜 在<br />

跨 過 橋 塔 後 , 會 在 兩 端 之 適 當 地 點 加 以 錨 錠 。 主 纜 由 於 缺 乏 抗 壓 與 抗 彎 能 力 , 當 垂 吊 索 施 以 向 下 之<br />

拉 力 作 用 時 , 主 纜 會 隨 之 產 生 可 觀 的 張 力 作 用 , 而 此 張 力 作 用 沿 著 主 纜 傳 遞 到 橋 塔 , 再 由 橋 塔 承 受<br />

軸 向 壓 力 、 彎 矩 及 剪 力 等 作 用 , 而 形 成 整 體 穩 定 的 結 構 。<br />

2.1. 吊 橋 之 主 要 構 件<br />

2.1.1 橋 塔<br />

橋 塔 是 支 承 主 纜 的 重 要 構 件 , 主 要 在 承 載 橋 身 下 拉 的 張 力 , 以 及 橋 塔 本 身 上 下 軸 向 的 壓 縮 力 、<br />

彎 曲 力 矩 及 剪 力 等 作 用 , 而 橋 塔 基 礎 則 需 傳 遞 軸 向 荷 重 及 安 置 於 適 當 的 承 載 地 盤 , 同 時 也 必 須 能 抵<br />

抗 可 能 之 上 舉 力 , 以 及 轉 倒 與 滑 動 等 作 用 力 , 所 以 是 吊 橋 施 工 過 程 中 最 重 要 的 環 節 。 有 些 地 區 的 吊<br />

橋 的 橋 塔 會 採 用 木 造 材 料 , 台 灣 的 人 行 吊 橋 橋 塔 則 以 鋼 構 及 RC 結 構 居 多 。<br />

2.1.2 主 索<br />

纜 索 是 由 單 根 鋼 線 (wire) 集 束 成 鋼 絞 索 (strand), 再 由 數 股 鋼 絞 索 組 成 鋼 索 (wire rope), 若 鋼 索 之<br />

直 徑 達 到 相 當 尺 寸 時 便 稱 之 為 鋼 纜 (cable)。 橋 塔 之 間 的 水 平 距 離 稱 為 跨 距 (span), 主 索 在 空 間 中 的 最<br />

高 點 與 最 低 點 間 的 垂 直 距 離 稱 為 垂 度 (sag), 而 主 索 的 垂 度 與 跨 距 的 比 值 則 稱 為 垂 跨 比 (sag-to-span<br />

ratio)。 當 主 索 跨 坐 在 兩 個 支 承 上 時 , 在 均 佈 荷 重 的 條 件 下 主 索 會 呈 懸 垂 曲 線 (catenary curve), 而 其<br />

曲 線 的 線 型 則 取 決 於 垂 跨 比 。 大 部 分 的 吊 橋 設 計 時 假 設 主 索 的 懸 垂 線 型 為 抛 物 線 型 , 以 方 便 工 程 上<br />

的 計 算 。 在 給 定 的 跨 距 及 載 重 條 件 下 , 若 採 用 愈 大 的 垂 度 則 纜 索 所 需 承 擔 的 張 力 就 愈 小 , 然 而 要 能<br />

增 加 主 索 的 垂 度 , 則 必 須 抬 高 橋 塔 的 高 度 , 這 是 在 吊 橋 主 索 設 計 時 所 面 臨 的 兩 難 問 題 。 在 考 量 增 加<br />

主 索 垂 度 的 同 時 , 不 只 是 需 要 考 慮 到 增 加 橋 塔 高 度 的 成 本 而 已 , 同 時 也 要 將 垂 吊 索 隨 之 加 長 的 費 用<br />

一 併 加 以 考 慮 , 而 且 人 行 吊 橋 因 為 橋 體 重 量 較 輕 , 過 大 的 垂 度 也 會 容 易 導 致 較 大 的 側 向 擺 動 結 果 。<br />

因 此 綜 合 考 量 結 構 穩 定 與 經 濟 效 益 的 情 形 下 , 大 部 分 的 吊 橋 垂 跨 比 採 用 1/8 至 1/12 之 間 的 範 圍 內 [2] 。<br />

由 各 年 代 建 造 的 吊 橋 演 進 可 知 , 製 作 吊 索 材 料 由 早 期 的 籐 蔓 、 木 桿 , 進 而 演 變 成 鐵 鍊 、 鐵 棒 、<br />

-418-


鋼 線 , 近 期 更 發 展 為 預 力 鋼 絞 索 。 但 這 些 不 同 的 演 變 的 材 料 中 , 主 要 都 是 用 來 承 受 軸 向 拉 力 作 用 的<br />

目 的 , 與 一 般 土 木 工 程 常 使 用 以 受 壓 或 受 剪 為 主 的 混 凝 土 的 情 況 有 明 顯 的 不 同 。<br />

2.1.3 垂 吊 索<br />

垂 吊 索 的 任 務 是 要 將 橋 面 板 的 重 量 傳 遞 到 主 索 上 , 因 此 其 設 計 及 安 裝 上 必 須 滿 足 結 構 之 健 全 、<br />

最 少 的 連 接 裝 置 、 具 有 垂 直 向 調 整 的 空 間 、 符 合 經 濟 效 益 並 能 顧 及 行 人 的 安 全 。 垂 吊 索 一 般 常 使 用<br />

等 間 距 的 方 式 排 列 , 其 長 度 的 設 計 是 由 該 垂 吊 索 固 定 在 主 索 的 高 度 , 到 橋 面 板 上 固 定 垂 吊 索 的 橫 梁<br />

之 間 的 距 離 來 決 定 , 因 此 各 垂 吊 索 的 長 度 彼 此 並 不 相 同 , 橋 塔 外 通 常 因 為 沒 有 橋 面 板 需 要 支 撐 所 以<br />

無 垂 吊 索 。 在 抛 物 線 形 的 主 索 的 任 何 一 個 位 置 上 , 從 抛 物 線 到 切 線 間 的 距 離 , 會 與 該 位 置 和 鞍 座 間<br />

的 水 平 距 離 的 平 方 成 正 比 [2] 。 利 用 此 一 關 係 式 , 可 以 進 一 步 計 算 出 介 於 抛 物 線 主 索 與 橋 面 板 之 間 的<br />

垂 吊 索 應 有 的 長 度 。 垂 吊 索 和 主 索 及 橋 面 板 間 各 有 一 個 連 接 的 接 點 , 需 使 用 適 當 的 連 接 器 (socket)<br />

來 加 以 接 合 傳 遞 作 用 力 。 不 同 的 年 代 在 連 接 器 的 使 用 上 常 見 有 明 顯 的 差 異 , 而 且 垂 吊 索 的 上 ( 連 接 主<br />

索 ) 及 下 ( 連 接 橋 面 板 ) 兩 端 的 連 接 方 式 也 常 是 不 同 的 型 式 。 部 分 吊 橋 有 裝 設 補 強 桁 架 , 可 以 將 集 中 荷<br />

重 分 配 到 數 個 垂 吊 索 來 分 擔 , 此 一 作 用 可 減 低 橋 面 板 在 承 受 不 均 勻 載 重 時 垂 直 方 向 上 的 振 動 量 。<br />

2.1.4 橋 面 板<br />

橋 面 板 是 主 要 提 供 行 人 行 走 的 構 件 , 然 而 廣 義 的 橋 面 板 從 上 而 下 可 以 再 細 分 為 橋 面 板 、 縱 梁 、<br />

橫 梁 及 補 強 桿 。 橋 面 板 預 拱 (bridge walkway camber) 的 設 計 是 因 為 考 量 纜 索 受 拉 力 而 伸 長 , 或 在 溫 度<br />

上 升 時 可 能 產 生 之 伸 長 的 現 象 , 倘 若 橋 面 板 在 初 始 時 以 水 平 構 建 , 將 會 造 成 橋 面 板 中 間 下 垂 的 後<br />

果 。 預 拱 的 設 計 同 時 扮 演 在 橋 面 板 步 道 和 橋 台 步 道 間 的 交 界 面 , 因 為 預 拱 的 關 係 會 對 橋 面 板 產 生 垂<br />

直 向 下 的 作 用 力 , 將 使 得 橋 面 板 可 穩 定 的 “ 安 坐 ” 在 橋 台 上 。<br />

2.1.5 抗 風 索<br />

抗 風 索 位 在 吊 橋 下 方 的 左 右 兩 側 , 能 有 效 抑 制 橋 面 板 的 擺 動 , 以 減 緩 在 強 風 狀 態 下 產 生 的 搖<br />

晃 , 以 避 免 吊 橋 翻 覆 。 抗 風 索 與 橋 面 板 之 間 需 藉 由 抗 風 支 索 來 傳 遞 連 結 效 果 , 一 般 抗 風 支 索 的 一 端<br />

固 定 在 抗 風 索 上 , 另 一 端 則 連 接 到 橋 面 板 的 橫 梁 上 。 此 外 , 為 增 加 橋 面 板 之 穩 定 性 , 部 分 吊 橋 會 在<br />

橋 面 板 底 部 安 置 沿 橋 面 走 向 之 底 索 , 亦 可 提 昇 橋 面 之 穩 定 性 。<br />

2.1.6 錨 碇<br />

一 般 傳 統 的 吊 橋 主 纜 係 錨 固 於 橋 梁 外 部 的 錨 碇 座 內 , 錨 碇 最 主 要 的 任 務 就 是 將 纜 索 的 張 力 正 確<br />

無 誤 的 傳 遞 給 錨 碇 基 礎 , 並 且 必 須 要 確 保 不 能 發 生 鋼 線 或 鋼 絞 索 鬆 脫 的 現 象 , 同 時 錨 碇 的 設 計 必 須<br />

能 夠 讓 各 鋼 絞 索 的 張 力 平 均 的 分 布 到 錨 碇 本 體 上 [3] 。<br />

2.1.7 鞍 座<br />

主 索 在 繞 經 橋 塔 頂 部 時 需 藉 助 鞍 座 才 能 順 利 通 過 , 鞍 座 一 方 面 支 承 主 索 , 同 時 也 改 變 主 索 的 走<br />

向 。 就 橋 塔 底 部 為 固 定 型 式 的 條 件 下 , 在 完 全 理 想 狀 況 下 鞍 座 和 主 索 間 應 該 是 沒 有 摩 擦 作 用 的 組<br />

合 , 如 此 才 能 對 橋 塔 只 傳 遞 垂 直 向 軸 力 , 而 不 會 有 水 平 向 作 用 力 , 以 避 免 造 成 橋 塔 之 彎 矩 效 應 。 鞍<br />

座 的 設 計 重 點 之 一 為 要 提 供 主 索 適 當 的 彎 曲 半 徑 , 容 許 最 小 彎 曲 半 徑 隨 著 主 索 的 索 材 、 索 徑 及 製 造<br />

方 式 而 有 所 不 同 , 在 鞍 座 及 主 索 間 需 要 有 足 夠 的 接 觸 面 積 , 以 避 免 過 大 的 輻 射 方 向 壓 力 作 用 造 成 鞍<br />

座 或 主 索 的 損 壞 。<br />

2.2. 吊 橋 之 振 動 分 級<br />

人 行 吊 橋 的 振 動 量 普 遍 會 比 公 路 橋 梁 明 顯 , 但 其 振 動 行 為 不 應 該 對 用 橋 人 造 成 不 適 或 恐 慌 的 心<br />

理 負 擔 , 人 行 吊 橋 在 無 活 載 的 條 件 下 , 垂 直 向 的 振 動 頻 率 應 要 大 於 3.0Hz 以 上 以 避 免 第 一 主 頻 的 共<br />

振 效 應 ; 橫 向 的 振 動 頻 率 則 應 要 大 於 1.3Hz 以 上 。 但 對 於 振 動 的 問 題 目 前 還 沒 有 收 歛 到 統 一 的 評 估<br />

方 法 , 或 是 任 何 適 當 的 標 準 規 定 。 相 關 的 規 定 大 致 可 分 為 兩 大 方 向 : 一 是 限 制 吊 橋 的 振 動 頻 率 會 高<br />

於 行 人 所 造 成 的 振 動 頻 率 , 另 一 是 倘 若 振 動 頻 率 無 法 滿 足 規 定 時 , 則 利 用 限 重 以 控 制 其 振 幅 [4] 。<br />

在 開 始 進 行 動 力 分 析 時 , 首 先 要 先 釐 清 該 吊 橋 的 使 用 等 級 (traffic class), 並 以 此 決 定 該 吊 橋 之 採<br />

用 之 舒 適 層 級 (comfort level)。 依 法 國 對 人 行 路 橋 的 分 類 方 式 , 依 其 使 用 交 通 量 共 分 為 四 個 使 用 等 級 :<br />

-419-


第 四 等 級 : 不 常 使 用 , 連 接 之 兩 地 人 煙 稀 少 。<br />

第 三 等 級 : 一 般 使 用 , 偶 爾 會 有 較 多 的 人 群 通 過 。<br />

第 二 等 級 : 經 常 使 用 , 座 落 在 交 通 使 用 量 高 的 地 區 , 有 時 橋 上 會 站 滿 人 。<br />

第 一 等 級 : 密 集 使 用 , 連 結 人 群 密 度 高 的 地 區 , 經 常 會 有 密 集 的 人 潮 通 過 。<br />

當 一 座 吊 橋 被 設 計 成 不 同 的 使 用 等 級 後 , 便 有 對 應 之 振 動 量 之 限 制 , 或 者 是 對 於 使 用 等 級 較 差<br />

的 吊 橋 , 則 可 控 制 在 比 較 少 的 建 造 成 本 以 內 。 至 於 舒 適 層 級 也 可 以 分 為 四 個 不 同 的 層 級 , 分 別 為 :<br />

最 高 舒 適 層 級 (maximum comfort): 使 用 者 幾 乎 不 會 察 覺 振 動<br />

一 般 舒 適 層 級 (average comfort): 使 用 者 稍 微 感 覺 到 振 動<br />

最 低 舒 適 層 級 (minimum comfort): 使 用 者 感 覺 到 振 動 但 不 致 於 無 法 忍 受<br />

不 舒 適 層 級 (uncomfortable): 使 用 者 會 明 顯 感 到 不 適 甚 至 恐 慌<br />

所 謂 舒 適 層 級 其 實 是 一 種 十 分 主 觀 的 感 受 , 在 相 同 的 加 速 度 條 件 下 , 每 個 人 會 有 不 一 樣 的 體<br />

驗 。 決 定 要 採 用 何 種 舒 適 層 級 取 決 於 該 吊 橋 之 使 用 交 通 量 、 使 用 對 象 以 及 其 重 要 等 級 。 對 特 別 敏 感<br />

的 使 用 者 而 言 ( 例 如 學 童 、 年 長 者 及 行 動 不 便 者 ) 會 有 較 嚴 格 的 要 求 , 但 對 於 較 短 的 吊 橋 因 為 通 過 所<br />

需 的 時 間 較 短 , 因 此 可 以 略 微 放 寬 容 許 的 標 準 。 表 1 及 表 2 所 示 為 垂 直 向 與 水 平 向 加 速 度 大 小 對 應<br />

之 舒 適 層 級 , 行 人 對 於 維 持 其 平 衡 所 能 忍 受 的 橫 向 加 速 度 的 極 限 遠 低 於 垂 直 向 的 加 速 度 , 分 別 大 約<br />

是 0.1g 和 0.3g [5] 。 表 3 為 台 灣 地 區 所 劃 定 的 地 震 震 度 分 級 表 ( 針 對 水 平 加 速 度 而 言 ), 比 對 表 2 及 表 3<br />

可 發 現 , 所 謂 的 不 舒 適 層 級 即 相 當 於 5 級 震 度 的 情 形 , 兩 套 不 同 目 的 的 分 級 方 式 可 謂 不 謀 而 合 。<br />

表 1 垂 直 向 加 速 度 大 小 對 應 之 舒 適 層 級 ( 單 位 :gal)<br />

加 速 度 等 級 0 50 100 250<br />

第 一 等 級<br />

第 二 等 級<br />

第 三 等 級<br />

第 四 等 級<br />

最 高 舒 適<br />

一 般 舒 適<br />

最 低 舒 適<br />

不 舒 適<br />

表 2 水 平 向 加 速 度 大 小 對 應 之 舒 適 層 級 ( 單 位 :gal)<br />

加 速 度 等 級 0 15 30 80<br />

第 一 等 級<br />

第 二 等 級<br />

第 三 等 級<br />

第 四 等 級<br />

最 高 舒 適<br />

一 般 舒 適<br />

最 低 舒 適<br />

不 舒 適<br />

表 3 地 震 震 度 分 級 表<br />

震 度 分 級 地 動 加 速 度 cm/s 2 (gal) 人 的 感 受<br />

0 無 感 0.8 以 下 人 無 感 覺 。<br />

1 微 震 0.8~2.5 人 靜 止 時 可 感 覺 微 小 搖 晃 。<br />

2 輕 震 2.5~8.0 大 多 數 的 人 可 感 到 搖 晃 , 睡 眠 中 的 人 有 部 分 會 醒 來 。<br />

3 弱 震 8~25 幾 乎 所 有 的 人 都 感 覺 搖 晃 , 有 的 人 會 有 恐 懼 感 。<br />

4 中 震 25~80 有 相 當 程 度 的 恐 懼 感 , 部 分 的 人 會 尋 求 躲 避 的 地 方 , 睡 眠 中 的<br />

人 幾 乎 都 會 驚 醒 。<br />

5 強 震 80~250 大 多 數 人 會 感 到 驚 嚇 恐 慌 。<br />

6 烈 震 250~400 搖 晃 劇 烈 以 致 站 立 困 難 。<br />

7 劇 震 400 以 上 搖 晃 劇 烈 以 致 無 法 依 意 志 行 動 。<br />

-420-


三 、 吊 橋 基 本 資 料 表 格<br />

要 想 建 置 一 套 吊 橋 管 理 系 統 , 首 先 便 需 設 計 出 吊 橋 基 本 資 料 表 格 , 之 後 再 依 所 設 計 的 表 格 進 行<br />

實 地 調 查 與 填 表 。 根 據 [ 小 規 模 吊 橋 指 針 同 解 說 ] [6] 第 14 章 「 記 錄 」, 橋 梁 資 料 庫 應 包 含 橋 長 、 橋 寬 、<br />

設 計 荷 重 、 設 計 震 度 、 基 礎 型 式 、 地 盤 條 件 、 主 要 構 件 的 構 造 圖 、 竣 工 年 月 以 及 其 他 管 理 上 需 要 記<br />

錄 的 事 項 。 而 本 研 究 除 了 參 考 [ 小 規 模 吊 橋 指 針 同 解 說 ] 的 項 目 外 , 也 根 據 實 地 調 查 台 灣 的 資 料 收 集<br />

後 進 行 整 理 , 例 如 吊 橋 的 結 構 型 式 可 分 為 雙 塔 式 、 單 塔 式 、 吊 床 板 式 、 桁 架 式 、 多 橋 塔 式 及 其 它 等<br />

六 類 ( 如 圖 1), 再 逐 一 分 類 彙 整 , 將 整 個 基 本 資 料 表 設 計 整 理 成 表 4。<br />

基 本 資 料 表 中 主 要 將 調 查 項 目 分 為 「 基 本 資 料 」、「 幾 何 資 料 」、「 結 構 資 料 」 及 「 設 計 資 料 」 等<br />

四 大 類 。 各 構 件 之 編 碼 原 則 為 由 近 到 遠 , 自 左 至 右 。 各 構 件 的 編 碼 方 式 也 要 訂 定 統 一 的 作 法 , 其 原<br />

則 是 以 河 流 自 上 游 望 向 下 游 的 方 向 為 基 準 , 將 橋 台 區 分 為 左 側 橋 台 及 右 側 橋 台 。 主 索 及 抗 風 索 因 為<br />

是 連 貫 兩 橋 台 , 其 編 碼 方 式 便 改 依 其 位 於 上 、 下 游 側 分 別 給 予 U、D 等 編 碼 ; 垂 吊 索 及 抗 風 支 索 自<br />

左 岸 開 始 編 號 , 再 配 合 U、D 編 碼 即 可 清 楚 說 明 。 循 此 編 碼 原 則 下 來 , 則 不 同 的 檢 測 人 員 所 做 的 記<br />

載 便 可 以 對 應 到 相 同 的 描 述 對 象 , 而 不 致 於 有 混 淆 不 清 的 情 形 發 生 。<br />

再 者 , 為 避 免 不 同 的 檢 測 人 員 所 採 用 的 構 件 名 稱 彼 此 有 差 異 , 而 造 成 之 後 在 進 行 大 量 調 查 成 果<br />

統 計 時 可 能 產 生 整 理 上 的 困 擾 , 所 以 本 系 統 將 「 結 構 資 料 」 的 調 查 項 目 之 各 種 可 能 的 選 項 先 行 整 理<br />

並 詳 細 列 出 , 利 用 勾 選 的 方 式 以 方 便 檢 測 人 員 的 記 錄 , 同 時 可 以 統 一 各 檢 測 人 員 之 調 查 成 果 的 呈 現<br />

方 法 。<br />

雙 塔 式 單 塔 式 吊 床 板 式<br />

桁 架 式 多 橋 塔 式 其 它 ( 斜 張 橋 式 )<br />

圖 1 各 種 不 同 的 吊 橋 型 式<br />

-421-


吊 橋 名 稱<br />

表 4 吊 橋 基 本 資 料 表<br />

填 表 人 填 表 日 期 年 月 日<br />

基 本 資 料<br />

編 號<br />

管 理 機 關 所 在 縣 市 所 在 鄉 鎮<br />

竣 工 年 月 造 價 設 計 單 位<br />

施 工 單 位 跨 越 物 體 參 考 地 標<br />

幾 何 資 料 ( 單 位 :m)<br />

橋 梁 總 長 橋 面 淨 寬 橋 塔 跨 徑<br />

橋 塔 高 度<br />

左<br />

右<br />

橋 塔 墩 高<br />

左<br />

右<br />

谷 深<br />

GPS 經 度 GPS 緯 度 海 拔 高 度<br />

結 構 資 料 ( 單 位 :cm)<br />

吊 橋 型 式<br />

□<br />

橋 塔<br />

□ 橋 面 ( 複 選 )<br />

□<br />

□<br />

□<br />

□<br />

主 索<br />

垂 吊 索<br />

抗 風 索<br />

抗 風 支 索<br />

□ 雙 塔 式 □ 單 塔 式 □ 吊 床 板 式 □ 桁 架 式 □ 其 它<br />

材 質 □ 鋼 構 □RC □ 其 它 斷 面 尺 寸<br />

固 定 型 式<br />

材 質<br />

補 強 桿<br />

索 材<br />

□ 固 接 □ 鉸 接 □ 其 它<br />

□ 原 木 □ 仿 木 □ 鍍 鋅 格 柵<br />

□ 型 鋼 □ 其 它<br />

縱 梁<br />

□ 無 □ 底 索 □ 止 搖 索 □C 型 鋼 □ 鉻 鉬<br />

橫 梁<br />

鋼 棒 □ 其 它<br />

□ 鋼 纜 □ 披 覆 鋼 纜 □ 鋼 線<br />

□ 其 它<br />

端 頭 固 定 型 式 □ 鋼 索 夾 □ 調 節 器 □ 橋 式 結 合 □ 其<br />

它<br />

錨 碇 基 礎 型 式 □ 重 力 式 混 凝 土 座 □ 基 樁 □ 預 力 地 錨<br />

□ 其 它<br />

垂 度 (m)<br />

索 徑 (mm)<br />

防 鏽 方 式<br />

跨 座 型 式<br />

索 材 □ 鋼 索 □ 披 覆 鋼 索 □ 鋼 線 □ 其 它 索 徑 (mm)<br />

頂 端 固 定 型 式 □ 夾 具 □ 鋼 索 夾 □ 調 節 器<br />

□ 纏 繞 式 □ 其 它<br />

底 端 固 定 型 式 □ 夾 具 □ 鋼 索 夾 □ 調 節 器<br />

□ 纏 繞 式 □ 其 它<br />

索 材<br />

□ 鋼 絞 索 □ 披 覆 鋼 絞 索<br />

□ 鋼 線 □ 其 它<br />

端 頭 固 定 型 式 □ 鋼 索 夾 □ 調 節 器 □ 橋 式 結 合<br />

□ 其 它<br />

防 鏽 方 式<br />

索 材<br />

□ 瀝 青 柏 油 □ 塗 漆 □HDPE □ 熱 浸 鍍 鋅<br />

□ 其 它<br />

□ 鋼 纜 □ 披 覆 鋼 纜 □ 鋼 線<br />

□ 其 它<br />

頂 端 固 定 型 式 □ 夾 具 □ 鋼 索 夾 □ 調 節 器 □ 纏 繞 式<br />

□ 其 它<br />

數 量 / 間 距<br />

索 徑 (mm)<br />

□ 變 斷 面 □ 均 勻 斷 面<br />

×<br />

□ 型 鋼 □ 原 木<br />

□ 其 它<br />

□ 型 鋼 □ 原 木<br />

□ 其 它<br />

□ 瀝 青 柏 油 □ 塗 漆 □HDPE<br />

□ 熱 浸 鍍 鋅 □ 其 它<br />

□ 活 動 式 □ 固 定 式 □ 其 它<br />

@<br />

□ 混 凝 土 座 □ 預 力 地 錨<br />

錨 碇 基 礎 型 式<br />

□ 其 它<br />

索 徑 (mm)<br />

數 量 / 間 距<br />

@<br />

-422-


底 端 固 定 型 式 □ 夾 具 □ 鋼 索 夾 □ 調 節 器<br />

□ 纏 繞 式 □ 其 它<br />

□ 護 欄 □ 木 構 □ 鋼 構 □ 鋼 索 □ 鋼 網 □ 繩 網 □ 鋼 線 □ 其 它<br />

設 計 資 料<br />

設 計 載 重 主 索 垂 度 橋 面 拱 度<br />

主 索 設 計 索 力 抗 風 索 設 計 索 力 垂 吊 索 設 計 索 力<br />

抗 風 支 索 設 計 索 力 設 計 風 力 設 計 地 震 力<br />

附 註<br />

四 、 吊 橋 檢 測 作 業<br />

吊 橋 常 用 檢 測 方 法 可 以 分 為 使 用 目 視 或 簡 單 量 測 工 具 的 一 般 檢 測 , 以 及 使 用 各 種 輔 助 儀 器 之 儀<br />

器 檢 測 , 必 要 時 亦 可 利 用 舉 高 車 將 檢 測 人 員 運 送 至 吊 索 或 橋 塔 等 高 處 位 置 進 行 近 距 離 的 檢 查 。 然 而 ,<br />

外 觀 的 檢 查 法 容 易 流 於 不 精 確 , 因 為 即 使 外 觀 完 好 也 有 機 會 遇 到 內 部 損 壞 的 狀 況 , 亦 即 外 觀 的 完 整<br />

並 無 法 保 證 吊 橋 各 構 件 皆 仍 能 發 揮 應 有 的 功 能 。<br />

4.1. 目 視 檢 測<br />

現 地 吊 橋 調 查 紀 錄 的 用 意 , 在 於 了 解 目 前 台 灣 已 有 吊 橋 之 現 況 。 但 除 了 基 本 資 料 調 查 外 , 對 於<br />

吊 橋 的 使 用 現 況 目 前 尚 無 一 套 評 估 判 斷 的 標 準 。 在 此 表 5 參 考 現 有 之 案 例 資 料 , 列 出 針 對 吊 橋 所 包<br />

含 的 各 重 要 構 件 所 對 應 可 能 發 生 之 破 壞 項 目 。<br />

表 5 各 構 件 之 目 視 檢 測 要 點<br />

( 一 ) 橋 塔 橋 塔 傾 斜 、 裂 縫 、 表 層 剝 落 、 鏽 蝕 、 基 礎 沈 陷 、 主 纜 於 鞍 座 是 否 有 不 正 常 滑 動 、 邊 坡 是 否 有 隱 含<br />

潛 在 問 題 ?<br />

( 二 ) 橋 面 板 破 損 、 傾 斜 、 生 苔 或 腐 爛 、 止 搖 索 損 壞 、 底 索 鬆 脫<br />

( 三 ) 主 索 斷 裂 、 局 部 鋼 纜 破 壞 、 表 面 鏽 蝕 、 錨 碇 基 礎 不 穩 、 固 定 端 鋼 索 夾 方 向 是 否 正 確 無 誤 、 錨 碇 位 置 是<br />

否 妨 礙 附 近 交 通 ?<br />

( 四 ) 垂 吊 索 表 面 鏽 蝕 、 斷 裂 或 鬆 脫 、 夾 具 破 壞 或 鏽 蝕 、 垂 吊 索 鋼 索 夾 方 向 是 否 正 確 無 誤 ?<br />

( 五 ) 抗 風 索 斜 拉 角 度 不 足 、 表 面 鏽 蝕 或 鬆 弛 、 斷 裂 、 錨 碇 鬆 動 、 鏽 蝕 、 鋼 索 夾 方 向 是 否 正 確 無 誤 ?<br />

( 六 ) 抗 風 支 索 表 面 鏽 蝕 、 斷 裂 或 鬆 脫 、 兩 端 之 夾 具 是 否 破 壞 或 產 生 鏽 蝕 、 鋼 索 夾 方 向 是 否 正 確 無 誤 ?<br />

( 七 ) 護 欄 金 屬 鏽 蝕 或 鬆 動 、 高 度 不 足 、 安 全 間 距 過 大 ?<br />

( 八 ) 整 體 性 安 全<br />

評 估<br />

是 否 搖 晃 過 於 劇 烈 、 是 否 有 清 楚 之 載 重 限 制 及 告 示 號 誌 ?<br />

若 有 危 險 疑 慮 之 吊 橋 , 是 否 有 封 橋 措 施 及 警 示 標 語 ?<br />

4.2 儀 器 檢 測<br />

吊 橋 在 使 用 一 段 時 間 之 後 主 索 可 能 會 出 現 下 垂 的 現 象 , 可 能 造 成 下 垂 的 因 素 有 : 主 索 腐 蝕 使 應<br />

力 增 加 、 溫 度 伸 長 、 鬆 弛 、 疲 乏 等 。 主 索 之 索 力 量 測 可 採 用 懸 垂 量 量 測 與 振 動 頻 率 量 測 等 方 法 來 間<br />

接 換 算 纜 索 索 力 的 大 小 。 懸 垂 量 的 變 化 可 據 以 推 斷 吊 索 張 力 的 變 化 , 但 此 法 量 測 成 敗 的 關 鍵 在 能 否<br />

使 用 高 精 度 的 測 量 儀 器 , 而 且 必 須 能 有 效 的 在 纜 索 動 態 振 動 的 狀 態 下 量 測 到 纜 索 之 平 衡 位 置 。<br />

4.2.1 三 維 雷 射 掃 描<br />

三 維 雷 射 掃 描 儀 是 一 種 能 在 短 時 間 內 快 速 獲 取 大 量 高 精 度 三 維 點 位 相 對 坐 標 的 儀 器 , 且 只 需 一<br />

個 儀 器 立 足 點 , 即 能 以 不 接 觸 被 測 物 的 方 式 快 速 獲 得 待 測 物 表 面 非 常 高 密 度 且 高 精 度 的 三 維 點 位 ,<br />

可 突 破 傳 統 測 量 受 限 的 方 法 , 能 精 確 、 大 量 、 快 速 且 連 續 地 蒐 集 資 料 , 對 於 分 析 結 果 更 具 有 真 正 的<br />

-423-


意 義 及 代 表 性 。 圖 2 為 利 用 三 維 雷 射 掃 描 完 整 的 將 碧 潭 吊 橋 之 形 狀 點 雲 資 料 化 , 量 測 結 果 顯 示 橋 長<br />

199.13 公 尺 、 橋 寬 3.5 公 尺 、 主 纜 索 之 懸 垂 量 為 7.54m, 藉 此 得 到 吊 橋 各 構 件 之 尺 寸 、 長 度 , 尤 其<br />

是 主 纜 索 之 懸 垂 量 及 橋 面 版 之 拱 度 , 可 保 存 作 為 日 後 重 建 、 整 修 之 重 要 依 據 。<br />

圖 2 三 維 雷 射 掃 描 之 案 例<br />

圖 3 以 速 度 計 量 測 纜 索 振 動 情 形<br />

圖 4 纜 索 之 振 動 歷 時 曲 線<br />

圖 5 纜 索 之 傅 氏 頻 譜<br />

表 6 纜 索 振 動 頻 率 與 索 力<br />

階 次 振 動 頻 率 (Hz) 索 力 (tf) 拉 應 力 (tf/cm 2 ) 階 次 振 動 頻 率 (Hz) 索 力 (tf) 拉 應 力 (tf/cm 2 )<br />

1 2.59 26.011 1.985 5 13.29 27.395 2.090<br />

2 5.25 26.719 2.039 6 16.26 28.477 2.173<br />

3 7.92 27.025 2.062 7 18.38 26.733 2.040<br />

4 10.62 27.333 2.086<br />

4.2.2 振 動 量 測 法<br />

振 動 測 量 法 則 是 目 前 量 測 索 力 最 普 遍 使 用 的 方 法 。 量 測 時 將 感 應 器 以 膠 帶 固 定 於 纜 索 上 ( 如 圖<br />

3), 並 記 錄 該 纜 索 之 振 動 歷 時 , 每 次 量 測 所 記 錄 之 時 間 長 度 應 以 便 利 於 後 續 之 傅 氏 頻 譜 分 析 為 主 。<br />

影 響 拉 索 拉 力 之 因 素 包 含 拉 索 固 有 頻 率 、 拉 索 撓 曲 剛 度 、 垂 跨 比 、 邊 界 約 束 等 , 並 採 用 弦 震 動 理 論<br />

[7]<br />

之 簡 算 公 式 進 行 量 測 資 料 之 分 析 :<br />

T<br />

2 2<br />

4wL<br />

f<br />

n<br />

= , (1)<br />

2<br />

n g<br />

其 中 :T 為 纜 索 拉 力 、 L 為 斜 拉 索 之 索 長 、w 為 單 位 索 長 的 質 量 、f n 為 第 n 階 振 動 頻 率 、g 為 重<br />

力 加 速 度 (m 2 /sec)。 圖 4 及 圖 5 分 別 為 分 析 案 例 之 振 動 歷 時 及 傅 氏 頻 譜 圖 , 由 圖 5 之 頻 譜 圖 中 可 清 楚<br />

補 捉 到 不 同 階 次 所 對 應 的 頻 率 值 。 表 6 則 為 換 算 之 索 力 大 小 , 發 現 在 不 同 的 階 次 所 換 算 出 來 的 索 力<br />

值 彼 此 相 當 吻 合 。<br />

-424-


4.2.3 吊 橋 扭 轉 行 為 之 分 析<br />

吊 橋 之 動 態 反 應 除 了 垂 直 振 動 與 水 平 搖 擺 之 外 , 橋 面 板 扭 轉 的 現 象 也 是 探 討 的 重 點 。 量 測 扭 轉<br />

的 方 式 可 利 用 在 橋 面 板 的 兩 側 擺 設 三 軸 向 加 速 度 計 , 由 其 各 方 向 之 加 速 度 資 料 進 行 相 位 角 頻 譜 分<br />

析 。 圖 6 左 的 A3 與 A4 為 橋 面 板 兩 側 之 兩 組 三 軸 向 加 速 度 計 , 圖 6 上 為 X 向 ( 行 人 方 向 ) 之 相 對 相 位<br />

圖 , 比 對 結 果 並 無 明 顯 之 相 位 差 ; 圖 6 中 為 Y 向 ( 水 平 方 向 ) 之 相 對 相 位 圖 , 顯 示 橋 面 板 兩 側 有 明 顯<br />

的 同 向 趨 勢 ; 圖 6 下 為 Z 向 ( 垂 直 方 向 ) 之 相 對 相 位 圖 , 發 現 約 在 0.8Hz 附 近 有 同 向 趨 勢 , 但 在 1.3Hz<br />

左 右 則 出 現 明 顯 的 相 位 角 差 異 , 反 應 在 該 頻 率 橋 面 板 的 兩 側 是 反 向 運 動 , 即 為 橋 面 板 之 扭 轉 行 為 。<br />

行 人 方 向 相 對 相 位 圖<br />

無 明 顯 趨 勢<br />

水 平 方 向 相 對 相 位 圖<br />

A1<br />

1/4<br />

同 向 趨 勢<br />

A2<br />

1/2<br />

垂 直 方 向 相 對 相 位 圖<br />

A3<br />

A4<br />

3/4<br />

同 向 趨 勢<br />

反 向 趨 勢<br />

圖 6 加 速 度 相 位 角 分 析 吊 橋 扭 轉<br />

4.2.4 其 它 調 查 方 法<br />

完 整 的 調 查 資 料 應 包 含 對 於 吊 橋 地 理 位 置 之 周 遭 環 境 , 對 此 可 使 用 無 人 載 具 (Unmanned aerial<br />

vehicle, UAV) 進 行 空 中 攝 影 取 得 數 位 影 像 , 再 依 不 同 成 果 之 需 求 進 行 影 像 處 理 、 接 圖 及 編 排 工 作 。<br />

此 外 , 為 使 吊 橋 之 影 像 紀 錄 能 有 更 真 實 的 效 果 , 本 計 畫 也 將 3D 攝 影 應 用 於 吊 橋 調 查 之 紀 錄 。<br />

五 、 吊 橋 管 理 系 統<br />

本 計 畫 經 由 文 化 大 學 詹 添 全 教 授 與 立 德 大 學 林 宏 明 教 授 等 人 之 協 助 , 目 前 已 完 成 如 圖 7 所 示 之<br />

61 座 吊 橋 之 調 查 , 並 將 成 果 建 置 於 吊 橋 管 理 系 統 中 。 該 管 理 系 統 包 括 「 基 本 資 料 建 置 」、「 基 本 資 料<br />

更 新 與 刪 除 」、「 基 本 資 料 閱 覽 與 列 印 」、「 檢 測 資 料 建 置 」、「 檢 測 資 料 更 新 與 刪 除 」、「 檢 測 資 料 閱 覽<br />

與 列 印 」 及 「 地 理 資 訊 系 統 整 合 」 等 七 個 模 組 ( 如 圖 8), 可 提 供 使 用 者 將 調 查 填 表 內 容 及 照 片 上 傳 至<br />

系 統 中 , 並 可 下 載 成 excel 與 word 格 式 的 成 果 列 表 , 直 接 輸 出 成 報 告 格 式 。 並 配 合 與 Google Earth<br />

的 連 結 , 可 提 供 使 用 者 方 便 迅 速 的 查 詢 瞭 解 各 吊 橋 之 地 理 位 置 , 對 於 吊 橋 之 管 理 單 位 對 於 其 所 管 轄<br />

之 吊 橋 數 量 、 地 點 、 結 構 型 式 、 健 康 現 況 等 資 訊 有 極 大 之 助 益 , 也 將 可 有 效 提 昇 吊 橋 之 安 全 維 護 作<br />

業 成 效 。<br />

-425-


圖 7 完 成 調 查 之 61 座 吊 橋 分 佈<br />

圖 8 吊 橋 管 理 系 統 之 網 頁<br />

六 、 結 論<br />

人 行 吊 橋 在 台 灣 的 交 通 發 展 史 上 具 有 舉 足 輕 重 的 地 位 , 不 論 是 跨 越 河 流 峽 谷 都 常 需 利 用 到 吊 橋<br />

的 便 利 性 , 且 近 年 來 吊 橋 的 觀 光 價 值 尤 其 備 受 重 視 , 常 作 為 地 區 觀 光 發 展 的 中 心 。 台 灣 地 區 吊 橋 之<br />

規 劃 、 設 計 、 建 造 、 維 護 , 於 過 去 幾 十 年 來 管 理 上 並 不 十 分 完 善 , 而 一 些 老 舊 的 吊 橋 在 長 期 腐 蝕 環<br />

境 及 風 雨 效 應 作 用 下 , 其 老 劣 化 的 現 象 對 於 人 行 安 全 上 的 問 題 是 著 實 不 可 忽 視 的 。 歷 年 來 台 灣 的 吊<br />

橋 管 理 在 各 地 各 行 其 政 , 尚 無 一 個 完 整 的 管 理 作 業 程 序 , 也 未 見 一 套 統 一 的 管 理 系 統 , 但 由 過 去 曾<br />

經 發 生 的 吊 橋 意 外 中 發 現 , 吊 橋 一 旦 發 生 事 故 時 容 易 帶 給 社 會 大 眾 不 安 , 使 吊 橋 的 安 全 性 及 可 靠 性<br />

受 到 質 疑 。 本 研 究 本 於 公 共 安 全 與 利 益 的 宗 旨 , 協 助 政 府 將 散 佈 在 全 島 的 吊 橋 開 始 予 以 普 查 , 建 置<br />

地 理 資 訊 管 理 平 台 , 蒐 集 基 本 資 料 及 照 片 , 並 開 始 規 劃 相 關 之 檢 測 方 法 與 評 估 維 修 策 略 , 其 中 亦 應<br />

用 適 當 之 先 進 檢 測 、 量 測 儀 器 與 方 法 , 以 使 能 獲 得 更 為 充 分 可 靠 的 資 訊 作 為 維 護 管 理 上 的 參 考 。<br />

参 考 文 獻<br />

[1] 林 金 田 . 台 灣 的 吊 橋 . 臺 灣 文 獻 館 , 2002 年 .<br />

[2] Latincsics, T. Design and Construction of the Pochuck Quagmire Bridge-A Suspension Timber Bridge. New Jersey<br />

Division of Parks and Forestry, 1998.<br />

[3] 王 如 鈺 , 王 雲 程 . 橋 樑 吊 索 講 義 . 鋼 橋 文 化 叢 書 : 2, 1996 年 .<br />

[4] NCHRP. LRFD Guide Specifications for the Design of Pedestrian Bridges, NCHRP, 20-07 Task 244, 2009.<br />

[5] French association of civil engineering, Technical guide Footbridges Assessment of vibrational behaviour of footbridges<br />

under pedestrian loading, Sétra, 2006.<br />

[6] 社 団 法 人 日 本 道 路 協 会 . 小 規 模 吊 橋 指 針 . 同 解 說 , 1996 年 .<br />

[7] 宋 一 帆 . 公 路 橋 梁 動 力 學 . 人 民 交 通 出 版 社 , 大 陸 . 1999 年 , 210-211.<br />

-426-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

索 杆 张 力 结 构 风 振 主 动 控 制 研 究<br />

1,2<br />

1<br />

2<br />

肖 南 苗 永 志 陈 华 鹏<br />

(1. 浙 江 大 学 建 筑 工 程 学 院 , 浙 江 杭 州 310058 2. 英 国 格 林 威 治 大 学 土 木 系 , 英 国 肯 特 ME4 4TB)<br />

摘 要 : 索 杆 张 力 结 构 由 于 质 量 轻 , 刚 度 柔 , 在 脉 动 风 荷 载 作 用 下 容 易 产 生 较 大 的 变 形 和 振<br />

动 。 为 了 实 现 对 该 类 结 构 风 致 振 动 的 主 动 控 制 , 本 文 考 虑 结 构 几 何 非 线 性 及 风 与 结 构 的 相 互 作<br />

用 , 基 于 瞬 时 最 优 理 论 推 导 了 张 力 结 构 风 振 主 动 控 制 方 程 , 并 通 过 Levy 型 索 穹 顶 算 例 进 行 主<br />

动 控 制 分 析 。 结 果 表 明 , 本 文 中 的 风 振 主 动 控 制 理 论 对 索 杆 张 力 结 构 具 有 很 好 地 适 用 性 , 在 较<br />

小 的 控 制 力 下 即 可 对 结 构 的 位 移 和 速 度 响 应 取 得 明 显 的 控 制 效 果 。 此 外 , 主 动 控 制 调 节 了 受 控<br />

结 构 在 风 荷 载 作 用 下 的 内 力 分 布 , 从 而 使 杆 件 轴 力 维 持 在 初 始 预 应 力 的 水 平 。<br />

关 键 词 : 索 杆 张 力 结 构 风 振 主 动 控 制 索 穹 顶<br />

ACTIVE VIBRATION CONTROL OF CABLE-STRUT TENSEGRITY STRUCTURES<br />

UNDER WIND EXCITATION<br />

Nan XIAO 1,2 , Yong-zhi MIAO 1 and Hua-peng CHEN 2<br />

1. College of Civil Engineering and Architecture, Zhejiang <strong>University</strong>, Hangzhou 310058, China<br />

2. School of Engineering, <strong>University</strong> of Greenwich at Medway, Chatham Maritime, Kent, ME4 4TB, UK<br />

Abstract: Cable-strut tensegriy structures tend to present significant deformation and vibration under dynamic<br />

wind loading. In order to eliminate or at least minimize the undesirable effects of this perturbation, the theory of<br />

active control is introduced. By considering geometric nonlinearity and the effect of wind-structure interaction,<br />

the active control approach is developed based on the instantaneous optimal control algorithm. In order to<br />

validate the effectiveness of the theory, studies on a Levy cable dome are performed. <strong>The</strong> results reveal that the<br />

active control theory presented in this paper is advantageous to tensegrity structures under dynamic wind<br />

excitation, and only small control force is required to reduce the structural responses. In addition, active control<br />

may modify the distribution of member forces in order to maintain the structure in the level of pre-tension.<br />

Keywords: Cable-strut tensegriy structures, wind-induced vibration, active control, cable dome.<br />

一 、 前 言<br />

索 杆 张 力 结 构 是 由 张 力 索 和 压 力 杆 为 基 本 构 成 单 元 、 通 过 预 应 力 提 供 结 构 刚 度 的 一 类 空 间 结 构<br />

体 系 [1] 。 该 类 结 构 能 够 最 大 限 度 地 利 用 自 身 的 材 料 特 性 , 以 尽 量 少 的 材 料 建 造 更 大 跨 度 的 空 间 。 虽<br />

然 如 此 , 由 于 其 结 构 的 质 量 轻 , 刚 度 柔 , 对 风 荷 载 十 分 敏 感 , 在 强 风 荷 载 作 用 下 , 容 易 发 生 较 大 的<br />

变 形 和 振 动 。 这 种 过 大 的 振 动 不 仅 可 能 引 起 结 构 表 面 膜 材 的 撕 裂 破 坏 , 还 可 能 影 响 结 构 的 稳 定 [2]-[4] 。<br />

目 前 , 索 杆 张 力 结 构 的 抗 风 问 题 并 没 有 得 到 很 好 地 解 决 , 由 风 荷 载 引 起 的 结 构 振 动 及 变 形 仍 然 是 结<br />

构 安 全 的 一 个 重 大 隐 患 。<br />

主 动 控 制 是 利 用 外 部 能 源 ( 计 算 机 控 制 系 统 或 智 能 材 料 ), 在 结 构 受 到 激 励 而 振 动 的 过 程 中 ,<br />

瞬 时 施 加 控 制 力 或 瞬 时 改 变 结 构 的 动 力 特 性 , 以 迅 速 控 制 或 衰 减 结 构 的 振 动 反 应 。 近 些 年 来 , 张 力<br />

结 构 的 主 动 控 制 研 究 引 起 了 广 泛 的 关 注 。Skelton [5] 等 人 的 研 究 表 明 , 对 张 力 结 构 的 主 动 控 制 仅 需 要<br />

[6]<br />

很 小 的 外 部 能 量 , 从 而 使 得 张 力 结 构 主 动 控 制 具 有 了 可 行 性 。Djouadi 等 基 于 有 限 单 元 法 对 一 个 由<br />

4 个 三 棱 柱 张 力 集 成 单 元 组 成 的 结 构 进 行 加 速 度 激 励 下 的 瞬 时 最 优 的 控 制 研 究 , 证 明 了 该 方 法 的 有<br />

效 性 。Kamada [7] 等 人 采 用 压 电 材 料 为 作 动 器 , 研 究 柔 性 结 构 在 随 机 荷 载 作 用 下 的 主 动 控 制 效 果 。Raja<br />

和 Narayanan [8] 基 于 H 2 和 H ∞ 的 最 优 控 制 理 论 研 究 某 包 含 2 个 单 元 的 张 力 集 成 结 构 在 随 机 荷 载 作 用 下<br />

的 反 应 , 并 且 比 较 了 两 种 控 制 理 论 的 不 同 。<br />

索 杆 张 力 结 构 在 强 风 作 用 下 容 易 进 入 几 何 非 线 性 状 态 , 因 而 该 类 结 构 的 风 振 主 动 控 制 应 采 用 非<br />

线 性 系 统 的 控 制 算 法 。J.N.Yang 等 在 文 献 [9] 中 指 出 经 典 线 性 最 优 控 制 理 论 由 于 没 有 考 虑 外 部 干 扰 ,<br />

因 而 不 是 真 正 的 最 优 , 并 提 出 了 新 的 最 优 控 制 算 法 : 瞬 时 最 优 控 制 法 。 该 瞬 时 最 优 控 制 法 在 每 个 时<br />

间 步 内 优 化 性 能 泛 函 , 由 于 将 结 构 位 移 按 振 型 分 解 , 因 而 仍 然 是 线 性 系 统 的 控 制 方 法 。 在 文 献 [10]<br />

中 Yang 等 推 导 了 地 面 加 速 度 荷 载 作 用 下 , 以 状 态 空 间 表 示 的 非 线 性 结 构 的 最 优 控 制 方 法 。Chang<br />

基 金 项 目 : 国 家 自 然 科 学 基 金 资 助 项 目 (50978228).<br />

作 者 简 介 : 肖 南 (1965—), 男 , 博 士 , 副 教 授 , 从 事 空 间 结 构 的 教 学 、 科 研 和 工 程 实 践 工 作 . E-mail:sholran@zju.edu.cn.<br />

-427-


和 Yang [11] 基 于 Yang 提 出 的 瞬 时 最 优 控 制 方 法 , 结 合 Newmark-β 法 , 给 出 了 一 种 新 的 控 制 方 法 , 该<br />

方 法 避 免 了 复 杂 Riccati 方 程 的 求 解 , 通 过 在 每 个 时 间 步 内 更 新 结 构 刚 度 或 阻 尼 矩 阵 , 可 实 现 非 线 性<br />

结 构 的 主 动 最 优 控 制 。<br />

各 国 学 者 在 大 跨 空 间 结 构 的 风 振 控 制 方 面 做 了 很 多 研 究 工 作 [12][13] , 但 针 对 索 杆 张 力 结 构 的 风 振<br />

控 制 研 究 目 前 还 不 多 见 。 本 文 基 于 可 考 虑 结 构 几 何 非 线 性 的 瞬 时 最 优 控 制 理 论 , 考 虑 风 与 结 构 的 相<br />

互 作 用 , 得 到 了 索 杆 张 力 结 构 的 风 振 主 动 控 制 方 程 , 并 采 用 一 Levy 型 索 穹 顶 算 例 进 行 主 动 控 制 效<br />

果 分 析 。<br />

二 、 基 于 瞬 时 最 优 的 风 振 主 动 控 制 理 论<br />

具 有 n 个 自 由 度 的 结 构 在 干 扰 P 及 控 制 力 U 作 用 下 ,t 时 刻 的 主 动 控 制 方 程 可 表 示 为<br />

Mx &&<br />

t<br />

+ Cx&<br />

t<br />

+ Kxt = DPt + BU<br />

t<br />

. (1)<br />

式 中 , M、 C、<br />

K —n×n 阶 的 质 量 、 阻 尼 、 刚 度 矩 阵 ;x 为 结 构 的 n 维 位 移 向 量 ;P 为 r 维 外 荷 载<br />

向 量 ,D 为 n×r 阶 干 扰 位 置 矩 阵 ;U 为 p 维 控 制 力 向 量 , 相 应 的 位 置 矩 阵 为 n×p 阶 阵 B。 考 虑 结 构<br />

几 何 非 线 性 时 , 刚 度 K 也 是 时 间 的 函 数 K t , 将 K t x t 写 成 F t 的 形 式 , 则 (1) 可 写 为 :<br />

Mx &&<br />

t<br />

+ Cx&<br />

t<br />

+ Ft = DPt + BU<br />

t. (2)<br />

其 中 , 系 统 的 位 移 和 速 度 都 是 独 立 变 量 。<br />

结 构 风 振 控 制 的 目 的 是 使 结 构 风 振 反 应 尽 量 最 小 , 同 时 又 使 控 制 力 不 至 于 过 大 , 为 此 取 系 统 的<br />

性 能 泛 函 为 :<br />

T T T<br />

J<br />

t<br />

= xQx<br />

t 1 t<br />

+ xQx & &<br />

t 2 t<br />

+ Ut RU<br />

t<br />

. (3)<br />

式 中 ,Q 1 、Q 2 和 R 均 为 权 矩 阵 ,Q 1 、Q 2 为 n×n 阶 半 正 定 矩 阵 ,R 为 p×p 阶 正 定 矩 阵 。<br />

引 入 Newmark 法 的 基 本 假 定 :<br />

xt = xt−Δ<br />

t<br />

+Δx . (4)<br />

x& t<br />

= (1 −a5)<br />

x& t−Δt − a &&<br />

6xt−Δt<br />

+ a4Δx. (5)<br />

&& xt = (1 −a3)<br />

&& xt−Δt − a2x&<br />

t−Δt<br />

+ a1Δx . (6)<br />

式 中 ,<br />

1 1 1 α α α<br />

a1 = ; a<br />

2 2<br />

= ; a3 = ; a4 = ; a5 = ; a6<br />

=Δt( −1)<br />

.<br />

β( Δt) βΔt 2β βΔt<br />

β 2β<br />

α 、 β 为 Newmark 法 的 参 数 , 一 般 取 α ≥ 0.5 , β ≥ 0.25 ( 0.5+ α), 本 文 中 取 α= 0.5 , β= 0.25 ,<br />

即 Newmark 法 退 化 为 平 均 加 速 度 法 , 是 目 前 广 泛 应 用 的 逐 步 积 分 法 。<br />

上 述 系 统 的 最 优 控 制 问 题 可 看 作 在 Newmark 法 基 本 假 定 为 约 束 条 件 下 , 求 性 能 泛 函 极 小<br />

值 。 引 入 Lagrange 乘 子 法 , 得 Hamilton 函 数 :<br />

1 T T T T T<br />

Ht = ( xQx<br />

t 1 t<br />

+ xQx & &<br />

t 2 t<br />

+ Ut RUt) + λ1 ( xt −xt−Δt −Δ x)<br />

+ λ2 ( x& t<br />

−(1 − a5) x& t−Δt + a6x&&<br />

t−Δt<br />

−a4Δx)<br />

2<br />

. (7)<br />

T<br />

+ λ3 (&& xt −(1 − a3) x&& t−Δt + a &<br />

2xt−Δt<br />

−a1Δx)<br />

为 了 消 除 Δx, 将 F t 线 性 化 表 示 为 如 下 形 式 [14] :<br />

Ft = KT( t−Δt)<br />

Δ x+<br />

F<br />

t−Δt. (8)<br />

此 处 ,K T 表 示 切 线 刚 度 矩 阵 , 将 (4)~(6)、(8) 代 入 (2) 式 , 可 得 :<br />

−1<br />

Δ x = K ( DP + BU + M( a x& + ( a − 1) x&& ) + C(( a − 1) x& + a && x ) + F ). (9)<br />

t t t 2 t−Δt 3 t−Δt 5 t−Δt 6 t−Δt t−Δt<br />

其 中 , Kt = a1M + a4 C + K<br />

T( t−Δt)<br />

。<br />

为 使 性 能 泛 函 (7) 式 取 极 小 值 , 有 :<br />

∂H t<br />

∂H t<br />

∂H t<br />

∂H<br />

t<br />

= = = = 0 . (10)<br />

T T T T<br />

∂xt ∂x&<br />

t<br />

∂&&<br />

xt ∂Ut<br />

将 (9) 代 入 (10), 得<br />

Qx + = 0<br />

1 t<br />

λ<br />

1<br />

. (11)<br />

Q x & + λ = 0 . (12)<br />

2 t 2<br />

-428-


由 式 (11)~(13) 可 求 得 闭 环 控 制 系 统 的 控 制 力 :<br />

将 (15) 式 代 入 (2) 式 , 并 记<br />

整 理 得<br />

λ = 0 3 (13)<br />

RU − B( a M + a C + K ) ( λ + a λ + aλ ) = 0. (14)<br />

t 1 4 T( t−Δt) 1 4 2 1 3<br />

−1<br />

T<br />

−T<br />

t<br />

=− t 1 t<br />

+ a &<br />

4 2 t<br />

U R B K ( Q x Q x ). (15)<br />

add<br />

t<br />

−1<br />

T<br />

−T<br />

4 t 2<br />

C = a B R B K Q . (16)<br />

add<br />

−1<br />

T<br />

−T<br />

Kt<br />

= B R B Kt<br />

Q<br />

1<br />

. (17)<br />

add<br />

add<br />

Mx && + ( C + C ) x&<br />

+ ( K + K ) x = D P . (18)<br />

t t t t t t t<br />

此 即 考 虑 结 构 几 何 非 线 性 时 的 主 动 控 制 方 程 , 其 中 K t 为 割 线 刚 度 矩 阵 ,C 采 用 Rayleigh 阻 尼 [15] :<br />

C = α[ M] + β[ K ]. (19)<br />

α 为 质 量 阻 尼 系 数 ,β 为 刚 度 阻 尼 系 数 。 这 两 个 阻 尼 系 数 可 通 过 振 型 阻 尼 比 计 算 得 到 。<br />

风 振 控 制 时 , 外 荷 载 P t 为 风 压 时 程 , 设 t 时 刻 风 速 向 量 为 v t ,v t 包 含 平 均 风 和 脉 动 风 。 由 于 张<br />

力 结 构 质 量 较 轻 , 刚 度 相 对 较 弱 , 在 风 荷 载 作 用 下 会 产 生 较 大 的 变 形 , 而 结 构 的 变 形 又 改 变 了 作 用<br />

在 其 上 的 风 荷 载 的 方 向 和 大 小 , 反 过 来 又 要 影 响 结 构 的 变 形 , 因 此 应 考 虑 风 与 结 构 的 相 互 作 用 [4] 。<br />

此 时 , 风 荷 载 可 表 示 为 :<br />

1 2 1<br />

2 2<br />

Pt = ρAμs( vt − x& t) = ρAμ s( vt − 2 v &<br />

t<br />

xt + x&<br />

t<br />

). (20)<br />

2 2<br />

式 中 ,ρ 为 空 气 质 量 密 度 ,A 为 风 荷 载 作 用 面 积 向 量 ,μ s 为 体 型 系 数 向 量 , 风 速 随 高 度 的 变 化 系 数<br />

已 在 平 均 风 中 考 虑 。<br />

由 于 &x 与 v<br />

t t 相 比 较 小 [16] , 略 去 与 之 相 关 的 二 阶 小 量 , 得<br />

1 2 1<br />

2 w<br />

Pt = ρAμsvt − ρAμsvt x&<br />

t<br />

= ρAμ svt −C &<br />

t<br />

xt. (21)<br />

2 2<br />

其 中 , C w<br />

t<br />

为 风 荷 载 引 起 的 附 加 阻 尼 矩 阵 , 对 n 个 自 由 度 的 系 统 而 言 可 写 成 一 n 阶 对 角 阵 :<br />

⎡ρA1μs<br />

1vt1<br />

0 L 0 ⎤<br />

⎢<br />

⎥<br />

w ⎢ 0 ρA2μs2vt2<br />

L 0<br />

C<br />

t<br />

= ⎥ . (22)<br />

⎢ M M O M ⎥<br />

⎢<br />

⎥<br />

⎢⎣<br />

0 0 L ρAnμ snvtn⎥⎦n × n<br />

式 中 ,A i 为 第 i 自 由 度 对 应 的 风 荷 载 作 用 面 积 , μ si 为 计 算 点 的 体 型 系 数 ,v ti 为 t 时 刻 第 i 个 自 由 度 对<br />

应 的 风 速 。 将 (21) 代 入 (18) 得<br />

add w add 1<br />

2<br />

Mx &&<br />

t<br />

+ ( C + Ct + Ct ) x&<br />

t<br />

+ ( Kt + Kt)<br />

xt = Dρμ sv<br />

. (23)<br />

t<br />

2<br />

此 即 考 虑 结 构 几 何 非 线 性 和 风 与 结 构 共 同 作 用 时 , 索 杆 张 力 结 构 的 主 动 控 制 方 程 。 记<br />

* add w<br />

Ct = C + Ct + C<br />

t<br />

. (24)<br />

* add<br />

Kt = Kt + K<br />

t<br />

. (25)<br />

* 1<br />

2<br />

Ft = Dρμ svt<br />

. (26)<br />

2<br />

则 (23) 式 可 简 写 为 :<br />

* * *<br />

Mx &&<br />

t<br />

+ Ct x&<br />

t<br />

+ Kt xt = F<br />

t<br />

. (27)<br />

求 解 此 非 线 性 的 动 力 方 程 即 可 得 到 受 控 结 构 的 响 应 时 程 及 控 制 力 时 程 , 本 文 对 此 方 程 的 求 解 采 用<br />

了 文 献 [17] 中 的 精 细 时 程 积 分 法 。<br />

三 、 算 例<br />

3.1 模 型 取 值<br />

-429-


如 图 1 所 示 一 Levy 型 索 穹 顶 , 直 径 40 m, 设 置 一 道 环 索 , 节 点 数 50, 约 束 节 点 数 16, 杆 件 数<br />

129, 其 中 压 杆 数 17, 拉 杆 数 112, 非 约 束 自 由 度 总 数 为 102。 节 点 1~16 为 支 座 节 点 , 约 束 三 个 方<br />

向 的 位 移 自 由 度 。 根 据 对 称 性 , 将 索 穹 顶 模 型 的 杆 件 类 型 分 为 7 组 , 如 图 2 所 示 , 其 中 , 第 1 组 、<br />

第 2 组 为 脊 索 , 第 3 组 、 第 4 组 为 斜 索 , 第 5 组 为 环 索 , 第 6 组 、 第 7 组 为 竖 杆 。 索 穹 顶 上 弦 节 点<br />

编 号 如 图 3 所 示 。<br />

[1]<br />

对 上 述 模 型 采 用 二 次 奇 异 值 分 解 法 求 得 整 体 可 行 自 应 力 模 态 分 布 如 表 1 所 示 (N 为 杆 件 分 组 号 ,<br />

T 为 自 应 力 模 态 )。<br />

图 1 Levy 型 索 穹 顶<br />

图 2 索 穹 顶 杆 件 分 组 信 息<br />

图 3 索 穹 顶 上 弦 节 点 编 号<br />

表 1 Levy 型 索 穹 顶 整 体 自 应 力 模 态 分 布<br />

N 1 2 3 4 5 6 7<br />

T 0.0293 0.0242 0.0142 0.04833 0.2273 -0.0452 -0.0136<br />

选 取 杆 件 截 面 , 并 按 自 应 力 模 态 分 布 施 加 预 应 力 , 如 表 2 所 示 , 其 中 钢 索 抗 拉 强 度 为 1860 MPa,<br />

弹 性 模 量 185 GPa, 压 杆 强 度 为 235 MPa, 弹 性 模 量 210 GPa。<br />

对 该 模 型 采 用 有 限 元 软 件 ANSYS 建 模 计 算 , 得 结 构 的 第 一 阶 和 第 二 阶 自 振 频 率 :ω 1 =4.7099Hz,<br />

ω 2 =8.7581Hz。 取 前 两 阶 振 型 阻 尼 比 ξ 1 =ξ 2 =0.02, 可 得 :α=0.1225,β=0.00297, 进 而 可 求 得 结 构 阻 尼<br />

矩 阵 。<br />

表 2 Levy 型 索 穹 顶 杆 件 信 息<br />

组 数 规 格 面 积 重 量 预 应 力 kN<br />

编 号<br />

mm 2 kg/m<br />

1 37Φ5 726 6.5 285<br />

2 37Φ5 726 6.5 235.4<br />

3 37Φ5 726 6.5 138.1<br />

4 55Φ5 1080 9.2 470.1<br />

5 283Φ5 5555 45.6 2210.9<br />

6 Φ133×5 2010 15.78 -439.7<br />

7 Φ133×5 2010 15.78 -132.3<br />

-430-


3.2 脉 动 风 压 时 程<br />

结 构 的 风 振 响 应 分 析 必 须 要 利 用 风 速 时 程 样 本 , 目 前 风 速 时 程 主 要 通 过 计 算 机 数 值 模 拟 得 到 。<br />

脉 动 风 一 般 看 作 各 态 历 经 的 零 均 值 高 斯 平 稳 随 机 过 程 , 因 此 可 用 随 机 过 程 的 模 拟 方 法 。 目 前 , 主 要<br />

是 基 于 三 角 函 数 加 权 叠 加 的 谐 波 叠 加 法 和 线 性 滤 波 法 。 本 文 利 用 线 性 滤 波 法 中 的 AR 自 回 归 模 型 模<br />

拟 索 穹 顶 各 风 荷 载 作 用 点 的 风 速 时 程 。 仅 考 虑 顺 风 向 风 荷 载 时 , 风 荷 载 作 用 点 在 上 弦 层 节 点 上 , 共<br />

有 17 个 :17、18、21、23、25、27、29、31、33、35、37、39、41、43、45、47、49。 设 该 索 穹<br />

顶 支 座 标 高 为 20m, 平 均 风 速 剖 面 采 用 对 数 律 , 水 平 阵 风 采 用 Kaimal 谱 。10 m 高 度 基 本 风 速 为 28.28<br />

m/s, 模 拟 时 间 步 长 0.1 s; 模 拟 风 速 时 程 时 间 总 长 200 s。 下 面 仅 列 出 17、18、35 号 节 点 的 风 速 时 程<br />

及 17 号 节 点 的 功 率 谱 对 比 结 果 , 如 图 4、5 所 示 ( 图 中 ,t 为 时 间 ,V 为 速 度 ,S 为 功 率 谱 密 度 ,f 为<br />

频 率 )。 为 了 便 于 比 较 , 将 18 和 35 点 的 脉 动 风 速 分 别 提 高 20 m/s 和 40 m/s。 由 图 5 可 以 看 出 模 拟 风<br />

速 时 程 的 自 功 率 谱 密 度 与 Kaimal 目 标 谱 较 好 地 吻 合 。<br />

V/(m/s)<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

35<br />

18<br />

17<br />

S /f<br />

10 3<br />

10 2<br />

10 1<br />

10 0<br />

10 -1<br />

10 -2<br />

10 -3<br />

模 拟 谱<br />

目 标 谱<br />

-10<br />

0 50 100 150 200<br />

t/s<br />

图 4 17、18、35 点 模 拟 风 速 时 程<br />

0.001 0.01 0.1 1 5<br />

f /Hz<br />

图 5 17 号 点 功 率 谱 密 度<br />

[18]<br />

得 到 各 风 荷 载 作 用 点 处 的 风 速 时 程 后 , 可 按 规 范 公 式 求 得 体 型 系 数 , 进 而 由 式 (21) 得 到 各<br />

点 与 屋 面 垂 直 方 向 的 风 压 时 程 , 将 风 压 时 程 按 x、y、z 三 个 方 向 分 解 可 得 结 构 各 自 由 度 对 应 的 风 压<br />

时 程 。<br />

3.3 主 动 控 制 参 数 取 值<br />

对 前 述 Levy 型 索 穹 顶 进 行 主 动 控 制 设 计 , 计 算 时 长 取 30 s。 结 构 控 制 力 可 以 施 加 在 拉 索 或 者 压<br />

杆 上 , 可 全 部 施 加 也 可 部 分 施 加 , 本 例 假 定 在 第 1~4 组 杆 件 , 即 全 部 脊 索 和 斜 索 施 加 控 制 力 。 权 矩<br />

阵 Q 1 、Q 2 和 R 分 别 取 为 :<br />

Q 1 =α 1 K,Q 2 =α 2 M,R=βI<br />

式 中 ,K 为 结 构 刚 度 矩 阵 ,M 为 结 构 质 量 矩 阵 ,I 为 p 阶 单 位 阵 ,p 为 控 制 力 个 数 ,α 1 ,α 2 ,β 均 为<br />

待 定 系 数 。 本 例 通 过 试 算 , 取 α 1 =α 2 =1000,β=1.0×10 -6 。<br />

四 、 结 果 分 析<br />

4.1 结 构 位 移 及 速 度 响 应 分 析<br />

图 6~9( 图 中 ,D 为 位 移 ) 给 出 了 索 穹 顶 17 号 节 点 的 水 平 x 向 和 竖 向 z 向 的 位 移 、 速 度 控 制 效<br />

果 。 可 以 看 出 , 索 穹 顶 在 风 荷 载 和 自 重 作 用 下 竖 向 的 位 移 和 速 度 响 应 数 值 较 大 ,z 向 的 响 应 应 作 为<br />

主 要 控 制 目 标 。 施 加 主 动 控 制 后 ,17 号 点 三 个 自 由 度 方 向 的 位 移 和 速 度 均 取 得 了 较 好 的 控 制 效 果 ,<br />

x 向 位 移 峰 值 降 低 了 60.98%, 振 幅 均 值 降 低 了 60.67%;x 向 速 度 峰 值 降 低 了 57.24%, 振 幅 均 值 降 低<br />

了 54.13%; 结 构 响 应 最 大 的 竖 向 位 移 和 速 度 控 制 效 果 更 为 明 显 ,z 向 位 移 峰 值 和 均 值 分 别 降 低 了<br />

98.50% 和 98.56%,z 向 速 度 降 幅 则 达 到 了 97.43% 和 97.65%。 即 受 控 结 构 17 号 节 点 z 向 位 移 及 速 度<br />

响 应 基 本 减 小 为 0。<br />

-431-


4 x 10-3<br />

3<br />

2<br />

1<br />

无 控<br />

受 控<br />

0.015<br />

0.01<br />

0.005<br />

无 控<br />

受 控<br />

D/m<br />

0<br />

-1<br />

-2<br />

V/(m/s)<br />

0<br />

-0.005<br />

-3<br />

-0.01<br />

-4<br />

0 10 20 30 -0.015<br />

0 10 20 30<br />

t /s<br />

图 6 17 号 节 点 x 向 位 移<br />

图 7 17 号 节 点 x 向 速 度<br />

t /s<br />

0.03<br />

0.02<br />

0.01<br />

无 控<br />

受 控<br />

0.04<br />

0.03<br />

V/(m/s)<br />

0<br />

-0.01<br />

D/m<br />

0.02<br />

0.01<br />

无 控<br />

受 控<br />

-0.02<br />

0<br />

-0.03<br />

0 10 20 30<br />

t /s<br />

图 9 17 号 节 点 z 向 速 度<br />

-0.01<br />

0 10 20 30<br />

t /s<br />

图 8 17 号 节 点 z 向 位 移<br />

考 查 其 他 节 点 的 控 制 效 果 , 提 取 所 有 节 点 受 控 前 后 的 x、y、z 三 向 位 移 峰 值 , 其 中 x、z 向 位 移<br />

峰 值 的 控 制 效 果 如 图 10~11 所 示 ( 图 中 ,n 为 节 点 号 )。<br />

2.5<br />

3 x 10-3<br />

无 控<br />

受 控<br />

0.03<br />

0.025<br />

无 控<br />

受 控<br />

D/m<br />

2<br />

0.02<br />

1.5<br />

D/m<br />

0.015<br />

1<br />

0.01<br />

0.5<br />

0.005<br />

0<br />

17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49<br />

n<br />

图 10 各 节 点 x 向 位 移 峰 值<br />

0<br />

17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49<br />

n<br />

图 11 各 节 点 z 向 位 移 峰 值<br />

对 所 有 节 点 而 言 , 水 平 向 位 移 和 竖 向 位 移 经 主 动 控 制 之 后 均 有 了 明 显 的 减 小 , 其 中 ,x 向 位 移 最<br />

大 点 35 号 节 点 的 峰 值 减 小 了 75.74%。 由 此 可 见 , 主 动 控 制 可 以 有 效 地 减 小 结 构 的 位 移 及 速 度 响 应 。<br />

-432-


4.2 主 动 控 制 力 分 析<br />

图 12 给 出 了 第 1~4 组 杆 件 对 应 控 制 力 的 平 均 值 ( 图 中 ,F C 为 控 制 力 ,m 为 杆 件 数 )。 可 以 看<br />

出 , 由 于 结 构 对 称 , 风 荷 载 也 基 本 对 称 , 同 一 组 杆 件 对 应 的 控 制 力 基 本 相 同 , 特 别 是 当 杆 件 控<br />

制 力 水 平 越 大 时 , 同 组 杆 件 不 同 杆 的 控 制 力 离 散 性 越 小 , 第 4 组 杆 件 各 杆 的 平 均 控 制 力 虽 然 出<br />

现 了 一 定 的 波 动 , 但 仍 保 持 在 同 一 水 平 。 此 外 , 在 不 考 虑 控 制 力 数 量 的 前 提 下 , 可 以 看 出 脊 索<br />

的 控 制 力 大 于 斜 索 , 内 圈 索 的 控 制 力 大 于 外 圈 索 。 图 13 给 出 了 第 一 组 杆 件 中 1 号 杆 ( 对 应 节<br />

点 17、18) 的 控 制 力 时 程 曲 线 , 控 制 力 峰 值 为 90.10 kN, 仅 为 杆 件 承 载 力 的 7.9%, 可 见 张 力<br />

结 构 只 需 很 小 的 控 制 力 即 可 达 到 较 好 的 控 制 效 果 , 这 与 文 献 [5] 中 的 结 论 相 同 。<br />

F C /N<br />

7 x 104<br />

6<br />

5<br />

4<br />

3<br />

第 一 组<br />

第 二 组<br />

第 三 组<br />

第 四 组<br />

F C /N<br />

10 x 104 1 号 杆<br />

9<br />

8<br />

7<br />

2<br />

6<br />

1<br />

0<br />

-1<br />

1 16<br />

m<br />

32<br />

图 12 第 1-4 组 杆 件 各 杆 平 均 控 制 力<br />

5<br />

4<br />

0 10<br />

t /s<br />

20 30<br />

图 13 1 号 杆 控 制 力 时 程<br />

4.3 拉 索 和 压 杆 内 力 分 析<br />

为 保 证 索 穹 顶 结 构 具 有 一 定 的 初 始 刚 度 , 在 荷 载 作 用 之 前 按 照 可 行 自 应 力 模 态 施 加 了 一 定<br />

水 平 的 预 应 力 , 该 预 应 力 应 该 保 证 结 构 在 荷 载 作 用 之 后 不 出 现 索 的 松 弛 , 并 保 证 一 定 的 稳 定 性 。<br />

由 于 对 索 穹 顶 而 言 , 每 组 杆 件 受 力 基 本 相 同 , 本 文 在 第 1~7 组 杆 件 中 , 每 组 取 一 根 杆 件 考 查 主<br />

动 控 制 过 程 中 各 杆 件 的 内 力 变 化 。 各 杆 控 制 前 后 内 力 峰 值 和 均 值 数 据 如 表 3 所 示 。`<br />

由 表 3 可 以 看 出 , 与 初 始 预 应 力 水 平 相 比 , 结 构 无 控 时 , 第 1、3、4 组 的 杆 件 轴 力 平 均 值<br />

分 别 增 大 了 11.23%、24.98%、3.74%, 第 2、5 组 的 杆 件 轴 力 平 均 值 分 别 降 低 了 10.45%、7.09%,<br />

而 第 6、7 组 竖 杆 的 轴 力 没 有 变 化 。 结 构 施 加 主 动 控 制 之 后 , 各 组 杆 件 轴 力 均 值 与 初 始 预 应 力<br />

分 布 相 差 无 几 , 可 见 主 动 控 制 对 轴 力 增 大 的 杆 件 起 到 了 减 小 轴 力 的 作 用 , 而 对 轴 力 减 小 的 杆 件<br />

起 到 了 增 大 轴 力 的 作 用 , 从 而 使 杆 件 轴 力 维 持 在 初 始 预 应 力 的 水 平 。<br />

组 号<br />

预 应<br />

表 3 控 制 前 后 各 杆 内 力 比 较 (kN)<br />

结 构 无 控<br />

结 构 受 控<br />

力 最 大 值 最 小 值 均 值 最 大 值 最 小 值 均 值<br />

1 285.0 337.1 303.08 317.1 292.8 279.1 285.6<br />

2 235.4 252.9 164.6 210.8 253.5 216.0 235.8<br />

3 138.1 211.6 128.4 172.6 153.6 124.8 140.0<br />

4 470.1 520.1 459.1 487.7 485.0 457.3 470.4<br />

5 2210.9 2141.8 2054.1 2103.3 2264.0 2144.2 2207.4<br />

6 -439.7 -439.7 -439.7 -439.7 -439.7 -439.7 -439.7<br />

7 -132.3 -132.3 -132.3 -132.3 -132.3 -132.3 -132.3<br />

图 14、15 给 出 了 第 2 组 ( 脊 索 ) 和 第 3 组 ( 斜 索 ) 的 轴 力 时 程 ( 图 中 ,F 为 杆 件 轴 力 )。<br />

可 以 看 出 , 主 动 控 制 降 低 了 第 3 组 杆 件 的 轴 力 而 增 大 了 第 2 组 杆 件 的 轴 力 。<br />

-433-


2.6 x 105<br />

2.4<br />

无 控<br />

受 控<br />

2.2 x 105<br />

2<br />

无 控<br />

受 控<br />

2.2<br />

F/N<br />

1.8<br />

F/N<br />

2<br />

1.6<br />

1.8<br />

1.4<br />

1.6<br />

0 10 20 30<br />

t / s<br />

图 14 第 2 组 杆 件 轴 力 时 程<br />

1.2<br />

0 10 20 30<br />

t/ s<br />

图 15 第 3 组 杆 件 轴 力 时 程<br />

五 、 结 语<br />

(1) 本 文 基 于 传 统 瞬 时 最 优 控 制 理 论 , 考 虑 了 风 与 结 构 的 相 互 作 用 , 得 到 了 索 杆 张 力 结<br />

构 的 风 振 主 动 控 制 方 程 。<br />

(2) 对 Levy 型 索 穹 顶 进 行 了 主 动 控 制 的 特 性 分 析 。 计 算 表 明 , 受 控 结 构 的 位 移 和 速 度 均<br />

有 明 显 的 降 低 , 控 制 效 果 明 显 ; 同 组 杆 件 控 制 力 基 本 相 同 , 且 较 小 的 控 制 力 即 可 取 得 很 好 的 控<br />

制 效 果 ; 主 动 控 制 调 节 了 杆 件 承 受 荷 载 后 的 内 力 分 布 , 从 而 使 杆 件 轴 力 维 持 在 初 始 预 应 力 的 水<br />

平 。<br />

(3) 本 文 中 假 定 控 制 力 施 加 在 全 部 脊 索 和 斜 索 上 , 这 可 能 导 致 过 大 的 控 制 成 本 。 对 控 制<br />

力 的 布 置 位 置 和 数 目 进 行 合 理 的 优 化 应 当 是 进 一 步 需 要 研 究 的 内 容 。<br />

参 考 文 献<br />

[1] 董 石 麟 , 罗 尧 治 , 赵 阳 等 . 新 型 空 间 结 构 分 析 、 设 计 与 施 工 , 北 京 : 人 民 交 通 出 版 社 , 2006.<br />

[2] Massimiliano, L. and Renato, V. V. et al. Dynamic behavior of a tensegrity system subjected to follower wind loading.<br />

Computers & Structures, 2003, 81: 2199-2207.<br />

[3] Massimiliano, L., Massimo, M. and Renato, V. V. et al. Nonlinear FE. analysis of Montreal Olympic stadium roof under<br />

natural loading conditions. Engineering Structures, 2009, 31: 16-31.<br />

[4] 沈 世 钊 , 武 岳 . 大 跨 度 张 拉 结 构 风 致 动 力 响 应 研 究 进 展 . 同 济 大 学 学 报 , 2002, 30(5): 533-538.<br />

[5] Skelton, R. E., Helton, J. W. and Adhikari, R. et al. Introduction to the mechanics of tensegrity structures, CRC Press,<br />

2002.<br />

[6] Djouadi, S., Motro, R. and Pons, J. C. et al. Active control of tensegriy systems. Journal of the Aerospace Engineering,<br />

1998, 11(2): 37-44.<br />

[7] Kamda, T., Fujita, T. and Hatayama, T. et al. Active vibration control of flexural-shear type frame structures with smart<br />

structures using piezoelectric actuators. Smart Materials and Structures, 1998, 7: 479-488.<br />

[8] Raja, M. G. and Narayanan, S. Active control of tensegrity structures under random excitation. Smart Materials and<br />

Structures, 2007, 16: 809-817.<br />

[9] Yang, J. N., Akbarpour, A. and Ghaemmaghami, P. New optimal control algorithms for structural control. Journal of<br />

Engineering Mechanics, 1987, 113(9): 1369-1386.<br />

[10] Yang, J. N., Li. Z. and Vongchavalitkul, S. Generalization of optimal control theory: linear and nonlinear control.<br />

Journal of Engineering Mechanics, 1994, 120(2): 266-283.<br />

[11] Chang, C. C. and Yang, H. T. Y. Instantaneous optimal control of building frames. Journal of Structural Engineering,<br />

1994, 120(4): 1307-1326.<br />

[12] 周 岱 , 郭 军 慧 . 空 间 结 构 风 振 控 制 系 统 的 神 经 网 络 时 滞 补 偿 . 空 间 结 构 , 2008, 14(2): 8-13.<br />

[13] 胡 继 军 , 黄 金 枝 , 李 春 祥 等 . 网 壳 -TMD 风 振 控 制 分 析 . 建 筑 结 构 学 报 , 2001, 22(3): 32-35.<br />

[14] Djouadi, S., Motro, R. and Pons, J. C. et al. Active control of tensegriy systems. Journal of the Aerospace Engineering,<br />

1998, 11(2): 37-44.<br />

[15] 爱 德 华 ·L· 威 尔 逊 著 , 北 京 金 土 木 软 件 技 术 公 司 、 中 国 建 筑 标 准 设 计 研 究 院 译 . 结 构 静 力 与 动 力 分 析 — 强 调 地 震<br />

工 程 学 的 物 理 方 法 . 北 京 : 中 国 建 筑 工 业 出 版 社 , 2006.<br />

[16] 张 华 , 单 键 . 薄 膜 结 构 随 机 风 场 模 拟 和 耦 合 风 振 响 应 分 析 . 工 程 力 学 , 2006, 23(10): 19-24.<br />

[17] 张 洵 安 , 姜 节 胜 . 结 构 非 线 性 动 力 方 程 的 精 细 积 分 算 法 , 2000, 17(4): 164-168.<br />

[18] 建 筑 结 构 荷 载 规 范 (GB 50009-2001)(2006). 北 京 : 中 国 建 筑 工 业 出 版 社 , 2006.<br />

-434-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

PREDICTION OF RESIDUAL SERVICE LIFE AND THROUGH-LIFE<br />

MAINTENANCE COSTS FOR HONG KONG PUBLIC RENTAL HOUSING<br />

ESTATES<br />

H. W. Pang 1 , C. O. Chan 2 and W. B Chan 3<br />

1 Estate Management Division,<br />

<strong>Hong</strong> <strong>Kong</strong> Housing Authority, <strong>Hong</strong> <strong>Kong</strong>, China. Email: hw.pang@housingauthority.gov.hk<br />

2 Estate Management Division,<br />

<strong>Hong</strong> <strong>Kong</strong> Housing Authority, <strong>Hong</strong> <strong>Kong</strong>, China. Email: co.chan@housingauthority.gov.hk<br />

3 Estate Management Division,<br />

<strong>Hong</strong> <strong>Kong</strong> Housing Authority, <strong>Hong</strong> <strong>Kong</strong>, China. Email: wb.chan@housingauthority.gov.hk<br />

ABSTRACT<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> Housing Authority (HKHA) is responsible for the construction and maintenance of the public<br />

rental housing estates in <strong>Hong</strong> <strong>Kong</strong>. As at 2011, there are over 1000 reinforced concrete public rental buildings,<br />

which are of 7 to 40 storeys with age ranging from newly built to over 50 years old. With ageing of the buildings,<br />

HKHA launched in 2005 the Comprehensive Structural Investigation Programme (CSIP) to assess the structural<br />

conditions of the older buildings and to formulate repair solutions to sustain them. <strong>The</strong> investigation generates a<br />

huge amount of data on the conditions and performance of these aged buildings, which include concrete strength,<br />

steel area loss and corrosion rates and others. This provides the necessary information for the prediction of the<br />

residual service life and the through-life maintenance costs of these aged buildings. This paper describes the<br />

methodology adopted for the prediction, which is developed based on the latest international standards and<br />

authoritative references. An example is also quoted for illustrative purposes.<br />

KEYWORDS<br />

Service life, repair cycle, corrosion rate, maintenance, structural investigation.<br />

INTRODUCTION<br />

<strong>The</strong> <strong>Hong</strong> <strong>Kong</strong> Housing Authority (HKHA) maintains some 700,000 public rental flats in over 1000 buildings,<br />

a stock which has gradually built up over the past half a century. <strong>The</strong> HKHA formulates building maintenance<br />

and redevelopment strategies for older estates based on assessment on whether these old buildings are<br />

cost-effective to keep, or whether it would be more cost-effective to demolish these buildings and build new<br />

ones. Efforts were made to gauge the residual service life and through-life maintenance costs of these aged<br />

buildings. This paper outlines the methodology adopted and gives illustration examples. <strong>The</strong> method is<br />

developed primarily for HKHA’s aged buildings only and appropriate modifications and adaptations are required<br />

for other applications.<br />

THE COMPREHENSIVE STRUCTURAL INVESTIGATION PROGRAMME (CSIP)<br />

Prediction of the residual service life and the through-life maintenance cost of an existing building require<br />

knowledge on the current conditions as well as the estimated rate of deterioration of it. As part of its<br />

maintenance strategy the HKHA carried out the CSIP on all her public rental buildings of age around 40 years<br />

since 2005. <strong>The</strong> purpose is to ascertain their structural conditions and to provide necessary information to enable<br />

the HKHA to decide if these aged buildings can be sustained for another 15 years or if demolition would have to<br />

be considered. <strong>The</strong> investigations have generated large amount information relating to the existing structural<br />

conditions of the buildings, such as carbonation depth, chloride content, core strength, steel area loss as well as<br />

corrosion rates. Inspection and testing are however mainly carried out in public area and limited number of<br />

vacant flats to minimize disturbance to tenants. Investigation is further affected by past repairs and presence of<br />

finishes and furniture. Regime of the investigation and test sampling are tailored made to suit the specific<br />

conditions of HKHA’s building stock and was reported elsewhere (Pang and Chan 2008) and (HKIE 2008). In<br />

essence, a representative sampling approach has been used in the investigation, which is further supported by<br />

-435-


adopting method of grouping and categorization of elements of similar properties or deterioration performance.<br />

INTERNATIONAL CODES AND PRECODES ON SERVICE LIFE<br />

<strong>The</strong>re has been much development in the concrete of service life design and planning in the last 15 years or so.<br />

<strong>The</strong>re has been a growing acceptance that safety levels of both new designs and existing structures to be<br />

mandated should be in reliability or probabilistic terms. Considerable work has been done towards codification<br />

of methods of assessment. <strong>The</strong> ISO 2394:1998 (ISO 1998) sets out the general principles of reliability-based<br />

design of structures as well as the target reliability levels for the different limit states designs. 2001 saw three<br />

heavy-weight publications on reliability-based assessment of existing structures. <strong>The</strong>se are the ISO 13822:<br />

Bases for Design of Structures – Assessment of Existing Structures (ISO 2001), and Probabilistic Assessment of<br />

Existing Structures (JCSS 2001) and Probabilistic Mode Code (JCSS 2001a) by the Joint Committee on<br />

Structural Safety (JCSS). Whilst (JCSS 2001a) details the general concepts and the framework of code<br />

calibration, (JCSS 2001) and (JCSS 2001a) provide further guidelines for assessment related to existing<br />

structure. <strong>The</strong> concepts of reliability and limited states, including the reversible serviceability limit states, the<br />

management and levels of reliability contained in these documents are filtered into the Eurocode EN 1990:2002<br />

Basis of Structural Design (BSI 2002). EN1990 explains the calibrations of the partial factors method against<br />

the target reliability levels. <strong>The</strong> fib Model Code for Service Life Design 2006 (fib 2006) and the ten-part ISO<br />

15686: Buildings and Constructed Assets- Service Life Planning (ISO 2000-2010) outline methods for planning,<br />

designing and predicting service life. <strong>The</strong> concept of reliability is embedded in the recently published fib Model<br />

Code 2010 (fib 2010).<br />

<strong>The</strong>se publications issued by international research bodies provide operational rules as well as the background,<br />

trends and design recommendations. <strong>The</strong> “model codes” provide framework for the code drafters to calibrate or<br />

produce standards and code of practices for use by practitioners. <strong>The</strong> relationship between these references can<br />

best be illustrated in “Figure 1” below.<br />

Fib Model Code for<br />

Fib Model<br />

Service Life Design: 2006<br />

Code 2010<br />

Model codes<br />

Codes and<br />

Standards<br />

Fib<br />

International federation<br />

for concrete structures<br />

Euro Codes<br />

ISO<br />

Other international<br />

organizations: CIB, RILEM,<br />

IABSE, JCSS, etc<br />

JCCS Probabilistic<br />

RILEM 2001 Probabilistic<br />

Model Code: 2001<br />

assessment of existing<br />

structures<br />

EN1990:2002<br />

Basis of<br />

structural<br />

design<br />

ISO 15686<br />

(10 parts, 2000-2010)<br />

Service Life Planning<br />

ISO 2394:1998<br />

general principles on<br />

reliability of<br />

structures<br />

ISO 13822: 2001<br />

bases for design of<br />

structures-assessment of<br />

existing structures<br />

Figure 1 <strong>The</strong> world knowledge map of service life<br />

Though principles are said to be applicable for the assessment of existing structures, direct application is<br />

however extremely difficult due to the specific conditions of HKHA buildings and the large amount of variables<br />

in the same building. Methodology developed for HKHA’s aged buildings has taken the advantage of<br />

availability of huge amount of corrosion-related field data collected from CSIP and is based on the above<br />

international models codes and standards with suitable adaptation.<br />

THE CONCEPTS<br />

Cost-effective of repairs as a serviceability limit state<br />

Concrete spalling in domestic buildings is normally an acceptable defect which can be repaired. It can thus be<br />

-436-


egarded as a serviceability limit state which is acceptable even when exceeded, and is reversible. <strong>The</strong> term<br />

“reversible limit state” is used in EN 1990: 2002 (BSI 2002). But if the number of times the need for concrete<br />

repair is too large, it could be unacceptable, as it might become not cost-effective to maintain the building and<br />

thus redevelopment has to be considered. To have a vigorous projection of concrete repair expenses of a<br />

building, it is necessary to have the understanding of the building’s current conditions, corrosion inducing agents,<br />

rate of deterioration as well as repair means and relevant costs for life-cycle cost analysis. Regarding the<br />

prediction of the repair extent and occurrence frequency, the mean values of the corrosion rate were adopted in<br />

the assessment. This is in line with Table E.2 (Target reliability index β – life-time, examples) in Clause E4.3 of<br />

Annex E in Ref. (ISO 1998) which suggests that for serviceability limit states which are reversible, the target<br />

reliability index can be taken as zero (β=0). Hence there will be no further assessment on the level of<br />

reliability of the occurrence of spalling as a serviceability limit state. But rather, life time cost of concrete<br />

repair in a predicted through-life performance, under a virtual inspection and maintenance trajectory is<br />

established to gauge its cost-effectiveness.<br />

Reliability levels of existing structures<br />

Proper regular inspection and maintenance are central to keeping the building in good serviceable condition.<br />

<strong>The</strong> levels of reliability related to structural safety can be achieved by protecting the structure from deterioration<br />

and through inspection and maintenance. That is the concept embraced by (BSI 2002).Clause 2.2 (reliability<br />

management) states,<br />

“..<strong>The</strong> level of reliability relating to structural resistance and serviceability can be achieved by suitable<br />

combination of (a) preventive and protective measures….design….(g) adequate inspection and maintenance….”<br />

With proper maintenance and because existing structures normally require a higher cost of achieving a higher<br />

reliability level, the target level of reliability for redesign of service life of existing structures usually should be<br />

lower.<br />

It can be argued that an inspection plan can be devised such that an inspection is done before a projected level of<br />

unacceptable reliability is reached, so that remedial actions keep the structure safe with adequate reliability. <strong>The</strong><br />

concept is illustrated in “Figure 2”.<br />

Reliability of structural safety<br />

building under<br />

natural deterioration<br />

building with proper<br />

inspection and maintenance<br />

Years in service<br />

Figure 2 Reliability management to buildings<br />

<strong>The</strong> above concepts of special consideration of reliability for various limit states for existing building structures<br />

are widely embraced in international codes and pre-codes, such as (fib 2010) (clause 3.3.3 tables 3.3-4 and<br />

3.3-5).<br />

Structural safety as the ultimate limit state<br />

Limit state design with partial factors derived by historical and empirical methods are calibrated by probabilistic<br />

methods to ensure the required reliability levels are generally achieved. <strong>The</strong> major parameters involved in a<br />

structural durability assessment of a building include loss of steel area (SAL), rate of further loss and strength<br />

-437-


eserve. Each SAL value is that figure associated with the most corroded section of the bars exposed in a small<br />

open-up of typically 300mm by 150mm. Determination of representative values for these parameters for a<br />

tenant-occupied building is however extremely difficult. Not only sampling and testing are greatly affected by<br />

the presence of tenants and furniture, these parameters themselves are very complex variable, which vary<br />

between buildings, between elements and even within element. Strength reserve determination is even more<br />

complex bearing in mind of the large number of elements, many loading combinations and re-distribution of<br />

loadings.<br />

<strong>The</strong> rate of deterioration of an existing building and hence its residual service life are related to a multitude of<br />

factors, from the original design standards, the construction workmanship to subsequent level of maintenance as<br />

well as other internal and external environmental effects. <strong>The</strong> factors will act interactively to give complicated<br />

multiplying effects.<br />

All these call for a simplified approach. Strength reserve is established generally being mindful that proactive<br />

inspection and intervention focused on the vulnerable local elements may be a better option to keep the building<br />

at an acceptable reliability level. Prediction of building residual life which is related to structural safety,<br />

demands a higher reliability in comparison with repair cycle projection. Mean values of SAL and corrosion<br />

rate are not used in the assessment as a significant probability of target reliability level may not be achieved.<br />

Instead, 90 percentile values of the parameters are adopted as a prudent approach.<br />

METHODOLGY<br />

Through-life maintenance cost<br />

Spalling repair for HKHA aged buildings has long been attributing a considerable proportion of the overall<br />

building maintenance expenses. <strong>The</strong> average yearly repair cost for different concrete elements or the same<br />

element type but at different locations varies greatly, depending very much on the present condition and its rate<br />

of deterioration. Reinforcement corrosion rates so far measured in 7 HKHA older estates for elements under<br />

different usages show great variances. <strong>The</strong> toilet slabs which are subject to cyclic wet and dry exposure due to<br />

showering water were found to be undergoing much acute corrosion whilst readings taken at constantly dry<br />

living room slabs and walls are generally mild. As shown in Table 1, the overall mean values for toilet, balcony,<br />

living room slabs and walls are respectively 0.0454, 0.0201, 0.0023 and 0.0022 mm depth/year. On estate basis,<br />

the mean and 90 percentile values for all four types of elements are also shown in Table 1. <strong>The</strong> values used in<br />

the assessments, i.e. mean values for slabs and 90 percentile values for walls are then gauged against the values<br />

of similar exposure conditions specified in EN206 (ECS 2000). Table 2 shows the corresponding corrosion rates<br />

for various exposure classes from CONTECVET Report 2001 (CONTECVET 2001).<br />

Table 1 Corrosion rates measured in 7 old estates<br />

Corrosion rates [mm/year]<br />

Toilet slab Balcony slab Living room slab Wall<br />

Mean 0.0207 – 0.0725 0.0086 – 0.0298 0.0012 – 0.0044 0.0006 – 0.0053<br />

90 Percentile 0.0510 – 0.2570 0.0119 – 0.0760 0.0019 – 0.0138 0.0018 – 0.0150<br />

Overall mean 0.0454 0.0201 0.0023 0.0022<br />

Table 2 Corrosion figures in CONTECVET Report 2001<br />

CONTECVET Report 2001<br />

Exposure Classes<br />

(Chloride initiated corrosion)<br />

Icorr [μA/cm 2 ]<br />

D1 Moderate humidity 0.1 - 0.2<br />

D2 Wet - rarely - dry 0.1 - 0.5<br />

D3 Cyclic wet - dry 0.5 - 5<br />

S1 Airborne sea water 0.5 - 5<br />

S2 Submerged 0.1 - 1.0<br />

S3 Tidal zone 1 - 10<br />

Note 1: CONTECVET 2001 (Section C5) also suggests applying a pitting factor (α) of 5 to 10 to arrive at the<br />

depth of pitting penetration.<br />

A comparison plot of measured corrosion rates against those in (CONTECVET 2001) is given in Figure 2. It is<br />

-438-


observed that the measured values closely aligned with the figures quoted in (CONTECVET 2001) of similar<br />

exposure classes.<br />

Figure 3 Corrosion rates measured in HKHA estates<br />

Corrosion rate has much effect on the repair frequency, and the repair method used relates closely to the extent<br />

of deterioration. Slab categorization is therefore necessary for these buildings for life-cycle repair cost analysis.<br />

A virtual repair cycle taking into account of the above has been modeled as outlined below, for through-life<br />

repair cost estimation for slabs in an aged estate:<br />

• Determine the corrosion rates and present SAL for toilet, balcony and living slabs based on the corrosion<br />

current measurements and opening-up.<br />

• Categorize the slabs into 4 grades according to the SAL so measured. Grade A is the best with SAL less<br />

than 8% and Grade D be the worst with SAL exceeding 30%, and estimate the relative proportion of slabs<br />

falling within slab grade.<br />

• Formulate repair solution for each grade of concrete deterioration and determine the relevant repair market<br />

unit rates.<br />

• Calculate the time taken by corrosion of each slab type (toilet, balcony and living room) propagating<br />

through each specified SAL range.<br />

• Determine the different starting points in the typical virtual repair cycle because of different present SAL<br />

of each slab grade. It is assumed that the slab after complete re-hauling repair, a new repair cycle then<br />

restarts again from zero SAL.<br />

• Calculate the repair costs of each slab category along period of each deterioration grade and then sum up<br />

all for over a specified time period, say 50 years, to obtain the estimated total slab repair cost for a typical<br />

flat in the period.<br />

<strong>The</strong> above concept is illustrated in “Figure 4”. Each slab may go through cycles of repairs of 4 repair modes (1)<br />

no repair, (2) patch repairs, (3) extensive patch repairs and (4) large area spalling repair with bar replacement.<br />

<strong>The</strong> cycles would depend on the choice of repair strategy.<br />

-439-


Balcony Slab<br />

<br />

Repair Cost ( $ / slab )<br />

<br />

<br />

Toilet Slab<br />

<br />

<br />

<br />

<br />

<br />

<br />

(year)<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

(year)<br />

– mode of repair (2)<br />

Residual service life prediction<br />

Figure 4 Illustrative diagram of repair cycles<br />

If concrete repair is not so excessive as to be less than cost-effective, prediction of the building residual service<br />

life can be made based on wall assessment. Using the corrosion rates measured in walls of these estates, the<br />

likely residual period of time before the steel corrosion reach a state that inadequate reserve is available can then<br />

be predicted at 90% probability. As discussed above, it is a prudent approach to adopt the values at 90%<br />

probability for both the SAL and corrosion rate for strength assessment. <strong>The</strong> residual service life calculation<br />

model is outlined as follows:<br />

• Plot the Cumulative Frequency Curve of the latest measured SAL for vertical elements.<br />

• Plot the steel corrosion rate of vertical elements based on representative corrosion rate measurements and<br />

determine the 90 percentile value.<br />

• Project the Cumulative Frequency Curves of SAL at different ages based on the 90 percentile steel<br />

corrosion rate.<br />

• Determine the strength reserve of the vertical elements, in terms of allowable SAL, based on loadings and<br />

capacity of the elements.<br />

• Predict the residual life of the estate at 90% probability based on the above.<br />

AN ILLUSTRATIVE EXAMPLE<br />

To illustrate the above prediction models, the details of the assessment carried out in one of the aged estates,<br />

Estate E, was given below. This estate comprises buildings with age approaching 40 years. In this exercise, the<br />

corrosion rate adopted is the mean or 90 percentile value of corresponding measurements taken in the group of<br />

estates of similar conditions.<br />

Based on the measured corrosion rates, the repair cycle for various categories of slab elements for Estate E is<br />

established and integrated over the coming years, say 50 years, to arrive at a total life time cost.<br />

As for the prediction of the residual service life of Estate E, the cumulative frequency plots of the SAL<br />

measured at walls of the Estate at various residual life were shown in “Figure 5” below. Based on the allowable<br />

strength reserve, it is estimated that Estate E can enjoy a residual life of more than 59 years under proper<br />

building maintenance. It can be argued that with intermediate in-depth inspections and focused interventions,<br />

further extension of the residual service life is achievable.<br />

-440-


_<br />

Cumulative Frequency of Samples (%)<br />

100%<br />

90%<br />

80%<br />

70%<br />

60%<br />

50%<br />

40%<br />

30%<br />

20%<br />

10%<br />

0%<br />

after 20 years<br />

after 30 years<br />

after 40 years<br />

after 50 years<br />

after 60 years<br />

at present (CSIP)<br />

after 10 years<br />

Estate “E”,<br />

with steel<br />

reserve of<br />

30%, can<br />

enjoy a period<br />

of 59 years<br />

before steel<br />

corrosion will<br />

affect the<br />

global safety.<br />

0 5 10 15 20 25 30 35 40<br />

Steel Area Loss (SAL) %<br />

Figure 5 Cumulative Frequency Curves of SAL at different ages of Estate E<br />

CONCLUSIONS<br />

Prediction models for residual service life and through life maintenance costs tailored made to suit HKHA’s<br />

specific aged building conditions have been developed based on latest international standards and authoritative<br />

reference. <strong>The</strong>y have been successfully applied to some aged HKHA estates. However, significant variations in<br />

conditions and performance were observed between estates, buildings and structural elements, which are<br />

attributable to difference in area usage, level of maintenance and inherent quality of the elements. <strong>The</strong><br />

predictions therefore must be crude ones. Such unique conditions also point to the need of well-structured and<br />

regular inspection and repair strategy to ensure that the buildings can be sustained for their residual service life.<br />

REFERENCES<br />

BSI (2002). BS EN 1990:2002 Eurocode – Basis of structural design, British Standards Institute.<br />

CONTECVET (2001). A validated users manual for assessing the residual service life of concrete structure,<br />

GEOCISA and Torroja Institute, Spain.<br />

ECS (2000). EN 206:2000 Concrete, Specification, Performance, Production and Conformity, European<br />

Committee for Standardization.<br />

Fib (2006). Model Code for Service Life Design, International Federation for Structural Concrete.<br />

Fib (2010). Model Code 2010, International Federation for Structural Concrete.<br />

HKIE (2008). Article on “Comprehensive Structural Investigation of Public Housing”, <strong>Hong</strong> <strong>Kong</strong> Institution<br />

of Engineers.<br />

ISO (1998). ISO 2394:1998 General principles on reliability for structures, International Standards<br />

Organization.<br />

ISO (2000-2010). ISO 15686 Buildings and Constructed Assets- Service Life Planning, International Standards<br />

Organization.<br />

ISO (2001). ISO 13822:2001 Bases for design of structures – Assessment of existing structures, International<br />

Standards Organization.<br />

JCSS (2001). Probabilistic assessment of existing structures, Joint Committee on Structural Safety.<br />

JCSS (2001a). Probabilistic model code Part 1 – Basis of design, Joint Committee on Structural Safety.<br />

Pang H.W., and Chan C.O. (2008). “Structural Investigation of <strong>Hong</strong> <strong>Kong</strong>’s ageing public rental buildings -<br />

Creating and Renewing Urban Structures”, Proceedings of 17th Congress of IABSE, Chicago 2008, 168-<br />

169.<br />

-441-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

超 高 层 建 筑 两 种 结 构 体 系 的 技 术 与 经 济 比 较<br />

周 红 梅 叶 甲 淳 谢 忠 良<br />

( 浙 江 省 建 筑 设 计 研 究 院 , 浙 江 杭 州 310006)<br />

摘 要 : 通 过 对 高 度 为 118 米 、150 米 、177 米 的 三 个 工 程 实 例 , 分 别 采 用 钢 筋 混 凝 土 框 架 - 核<br />

心 筒 结 构 体 系 和 混 合 框 架 - 钢 筋 混 凝 土 核 心 筒 混 合 结 构 体 系 进 行 设 计 和 计 算 , 对 两 种 结 构 体 系 进 行<br />

了 技 术 和 经 济 两 方 面 的 对 比 和 分 析 , 并 探 讨 了 两 种 结 构 体 系 随 高 度 增 加 显 现 出 的 规 律 和 经 验 。<br />

关 键 词 : 钢 筋 混 凝 土 框 架 - 核 心 筒 混 合 框 架 - 钢 筋 混 凝 土 核 心 筒 钢 管 混 凝 土 柱 双<br />

重 抗 侧 力 体 系 对 比<br />

TECHNOLOGICAL AND ECONOMIC COMPARISON OF TWO SUPER HIGH-RISE<br />

STRCUTURAL SYSTEMS<br />

<strong>Hong</strong>mei Zhou, Jiachun Ye and Zhongliang Xie<br />

Zhejiang Prov. Institute of Architectural Design & Research, Hangzhou 310006, China<br />

Abstract: <strong>The</strong> paper provides us with three construction projects in the height of 118, 150 and 177 meters<br />

which are designed and calculated on the basis of RC frame-cored tube structure and Mixed frame-RC<br />

cored tube structure. Comparison and analysis in the aspects of technology and economy to those two<br />

structures have been made, meanwhile the arising regular manners with the rising structural heights have<br />

been discussed.<br />

Keywords: RC frame-core, mixed frame-RC core, concrete-filled steel tubular column, dual horizontal force<br />

resisting system, comparison.<br />

一 、 引 言<br />

在 我 国 100 米 ~200 米 的 超 高 层 建 筑 中 , 钢 筋 混 凝 土 框 架 - 核 心 筒 结 构 体 系 应 用 较 为 广 泛 。 近 年<br />

来 , 混 合 框 架 - 钢 筋 混 凝 土 核 心 筒 混 合 结 构 体 系 的 应 用 逐 渐 增 多 , 混 合 框 架 或 采 用 钢 骨 混 凝 土 柱 -<br />

钢 梁 , 或 采 用 钢 管 混 凝 土 柱 - 钢 梁 等 , 核 心 筒 为 钢 筋 混 凝 土 实 腹 筒 。 这 两 种 结 构 体 系 各 有 怎 样 的 优<br />

势 和 缺 点 ? 结 构 性 能 有 怎 样 的 异 同 ? 两 种 体 系 的 材 料 用 量 相 差 多 少 ? 综 合 的 经 济 效 益 如 何 ? 广 大<br />

设 计 人 员 和 建 设 单 位 在 结 构 方 案 选 择 时 常 常 会 面 临 这 些 问 题 。<br />

本 文 选 择 了 屋 面 高 度 分 别 为 118 米 、150 米 、177 米 的 三 个 具 有 较 强 代 表 性 的 工 程 实 例 , 对 三 幢<br />

建 筑 分 别 采 用 钢 筋 混 凝 土 框 架 - 核 心 筒 结 构 体 系 ( 本 文 以 下 均 简 称 “ 混 凝 土 结 构 体 系 ”), 和 混 合 框<br />

架 - 钢 筋 混 凝 土 核 心 筒 结 构 体 系 ( 本 文 以 下 均 简 称 “ 混 合 结 构 体 系 ”, 混 合 框 架 采 用 钢 管 混 凝 土 柱<br />

- 钢 梁 ) 进 行 设 计 和 计 算 分 析 , 既 对 每 个 工 程 的 两 种 不 同 结 构 体 系 进 行 分 析 和 对 比 , 也 对 每 种 结 构<br />

体 系 随 高 度 增 加 显 现 出 的 规 律 进 行 探 讨 , 并 力 求 从 技 术 和 经 济 两 方 面 对 类 似 超 高 层 建 筑 的 两 种 结 构<br />

体 系 进 行 一 些 分 析 。<br />

二 、 工 程 实 例 概 况<br />

本 文 选 取 的 工 程 实 例 均 为 地 处 杭 州 的 已 建 和 在 建 工 程 , 三 个 项 目 分 别 为 杭 州 坤 和 中 心 、 中 华 航<br />

空 大 厦 和 亚 包 大 厦 , 项 目 的 概 况 、 结 构 平 面 布 置 和 效 果 图 如 下 :<br />

-442-


三 个 项 目 基 本 概 况 对 比 表<br />

主 要<br />

核 心 筒<br />

建 筑<br />

建 筑 核 心 筒<br />

屋 面<br />

外 边 与 外 框<br />

工 程 名 称 总 高 度<br />

层 数 外 包 尺 寸 ( 米 外 包 尺 寸 ( 米<br />

高 度<br />

柱 距 离 ( 米 )<br />

( 米 )<br />

× 米 ) × 米 )<br />

( 米 )<br />

标 准<br />

层 高<br />

( 米 )<br />

标 准 层<br />

面 积<br />

( 平 方<br />

米 )<br />

杭 州 坤 和 中 心 128 118 32 39.9×39.9 11.8×20.5 10.35 3.45 1500<br />

中 华 航 空 大 厦 155 150 38 45.6×45.6 18.5×18.5 10.5 3.8 1500<br />

亚 包 大 厦 180 177 46 42.6×42.6 19.5×19.5 10.5 3.95 1818<br />

KZ1<br />

杭 州 坤 和 中 心<br />

坤 和 中 心 结 构 平 面 布 置<br />

杭 州 坤 和 中 心<br />

KZ1<br />

中 华 航 空 大 厦<br />

中 华 航 空 大 厦 结 构 平 面 布 置<br />

-443-


KZ1<br />

亚 包 大 厦<br />

亚 包 大 厦 结 构 平 面 布 置<br />

杭 州 设 计 基 本 地 震 加 速 度 值 为 0.05g, 抗 震 设 防 基 本 烈 度 6 度 , 设 计 地 震 分 组 为 第 一 组 。 三 个<br />

工 程 场 地 类 别 均 为 Ⅲ 类 , 均 属 标 准 设 防 类 ( 丙 类 ) 建 筑 。100 年 重 现 期 的 基 本 风 压 取 值 0.50KN/m 2 ,<br />

阻 尼 比 取 值 0.04。 杭 州 坤 和 中 心 和 中 华 航 空 大 厦 属 于 A 级 高 度 , 亚 包 大 厦 属 于 B 级 高 度 高 层 建 筑 。<br />

本 文 分 析 的 混 合 结 构 体 系 中 , 杭 州 坤 和 中 心 采 用 方 钢 管 混 凝 土 柱 , 中 华 航 空 大 厦 和 亚 包 大 厦 采<br />

用 圆 钢 管 混 凝 土 柱 ; 楼 面 梁 均 采 用 H 型 钢 梁 或 箱 型 钢 梁 , 楼 板 采 用 压 型 钢 板 组 合 楼 板 或 钢 筋 桁 架 组<br />

合 楼 板 。 钢 框 架 梁 与 外 围 钢 管 混 凝 土 柱 刚 接 , 与 钢 筋 混 凝 土 核 心 筒 之 间 的 连 接 , 中 华 航 空 大 厦 全 部<br />

为 铰 接 , 杭 州 坤 和 中 心 和 亚 包 大 厦 在 核 心 筒 角 部 采 用 刚 接 , 其 余 部 位 采 用 铰 接 。 为 了 有 利 于 刚 性 连<br />

接 , 在 核 心 筒 角 部 设 置 了 型 钢 。<br />

钢 材 均 采 用 Q345B 钢 , 混 凝 土 强 度 等 级 为 C60~C30。<br />

三 、 结 构 计 算 分 析 及 结 构 性 能 对 比<br />

本 文 选 用 了 SATWE 和 MIDAS 两 种 分 析 软 件 对 三 幢 建 筑 的 两 种 结 构 体 系 进 行 了 整 体 分 析 计 算 。<br />

采 用 考 虑 扭 转 耦 联 振 动 影 响 的 振 型 分 解 反 应 谱 法 , 对 结 构 作 了 多 遇 地 震 作 用 下 的 整 体 弹 性 分 析 , 计<br />

算 结 果 详 见 下 表 。 表 中 无 括 号 的 数 值 为 SATWE 软 件 计 算 结 果 , 括 号 内 的 数 值 为 MIDAS 软 件 计 算<br />

结 果 。 表 后 分 析 采 用 的 数 据 为 SATWE 软 件 的 计 算 结 果 。<br />

主 要 技 术 指 标 对 比 表 之 一<br />

工 程 名 称 杭 州 坤 和 中 心 中 华 航 空 大 厦 亚 包 大 厦<br />

结 构 体 系<br />

钢 筋 砼 框 架 混 合 框 架 - 钢 筋 砼 框 架 混 合 框 架 钢 筋 砼 框 架 混 合 框 架<br />

- 核 心 筒 核 心 筒 - 核 心 筒 - 核 心 筒 - 核 心 筒 - 核 心 筒<br />

上 部 结 构 总 质 量 (t)<br />

77140 61330 93792 75884 124341 100328<br />

(77242) (61791) (92997) (75163) (122844) (101921)<br />

上 部 结 构 面 积 (m 2 ) 48700 55600 76000<br />

单 位 面 积 质 量 分 布 (kg/m 2 )<br />

1584 1259 1687 1365 1636 1320<br />

(1586) (1269) (1672) (1352) (1616) (1341)<br />

总 基 底 墙 柱<br />

最 大 组 合 轴 力 (KN)<br />

1064330 878814 1309318 1103856 1608866 1311482<br />

KZ1 柱 底 最 大 组 合<br />

轴 力 (KN)<br />

29096 20564 34015 25085 36291 27487<br />

-444-


风 作 用 下 基 底 总<br />

X 向<br />

7442<br />

(7127)<br />

8151<br />

(7268)<br />

10328<br />

(10183)<br />

11245<br />

(10223)<br />

14544<br />

(14416)<br />

15804<br />

(14470)<br />

剪 力 (kN)<br />

Y 向<br />

7442<br />

(7127)<br />

8151<br />

(7261)<br />

10062<br />

(9803)<br />

10971<br />

(9830)<br />

14379<br />

(14215)<br />

15633<br />

(14224)<br />

风 作 用 下 基 底 弯<br />

矩 (kN-m)<br />

X 向<br />

Y 向<br />

530992<br />

(474785)<br />

530992<br />

(476568)<br />

586617<br />

(542029)<br />

586617<br />

(540748)<br />

892491<br />

(911504)<br />

888965<br />

(882864)<br />

981327<br />

(922170)<br />

977703<br />

(888773)<br />

1465370<br />

(1500821)<br />

1467263<br />

(1495524)<br />

1607488<br />

(1516405)<br />

1609861<br />

(1496513)<br />

水 平 地 震 作 用 下<br />

X 向<br />

6323<br />

(6392)<br />

5281<br />

(5285)<br />

6709<br />

(6583)<br />

5650<br />

(5676)<br />

8949<br />

(8897)<br />

7513<br />

(7519)<br />

基 底 总 剪 力 (kN)<br />

Y 向<br />

5992<br />

(6037)<br />

4998<br />

(4974)<br />

8010<br />

(7946)<br />

6786<br />

(6734)<br />

9045<br />

(8986)<br />

7681<br />

(7663)<br />

水 平 地 震 作 用 下 X 向<br />

基 底 弯 矩<br />

(kN-m) Y 向<br />

T1<br />

前 3 阶 振 型<br />

T2<br />

周 期 (s)<br />

T3<br />

X 向<br />

风 载 作 用 下 最 大<br />

层 间 位 移 角<br />

Y 向<br />

X 向<br />

地 震 作 用 下 最 大<br />

层 间 位 移 角<br />

Y 向<br />

框 架 承 担 的 地 震 X 向<br />

剪 力 比<br />

( 底 层 ) Y 向<br />

407129<br />

(414370)<br />

306689<br />

(333945)<br />

541738<br />

(552553)<br />

439296<br />

(462999)<br />

846014<br />

(863000)<br />

694074<br />

(717423)<br />

404500 311614 580296 482697 858154 720778<br />

(413235) (336730) (582304) (488901) (876022) (741136)<br />

3.40 3.61 4.19 4.70 4.01 4.45<br />

(3.26) (3.52) (4.01) (4.36) (4.11) (4.45)<br />

3.17 3.52 3.20 3.52 3.79 3.92<br />

(3.07) (3.42) (3.04) (3.27) (3.86) (3.93)<br />

2.82 2.70 2.45 2.33 2.79 2.75<br />

(2.55) (2.47) (2.24) (2.06) (2.63) (2.60)<br />

1/2343 1/1096 1/1398 1/ 803 1/1706 1/1063<br />

(1/2537) (1/1426) (1/1565) (1/1008) (1/1607) (1/1110)<br />

1/2109 1/1246 1/2233 1/1309 1/1847 1/1323<br />

(1/2289) (1/1575) (1/2605) (1/1702) (1/1765) (1/1397)<br />

1/2497 1/1845 1/2037 1/1593 1/2579 1/2176<br />

(1/2786) (1/2138) (1/2308) (1/1868) (1/2614) (1/2208)<br />

1/2335 1/1979 1/2878 1/2310 1/2700 1/2424<br />

(1/2661) (1/2278) (1/3454) (1/2766) (1/2772) (1/2605)<br />

10.57% 5.93% 13.67% 12.88% 15.70% 8.90%<br />

(7.95%) (4.78%) (10.7%) (9.61%) (17.6%) (10.2%)<br />

10.12% 5.27% 8.45% 7.39% 13.75% 7.37%<br />

(8.38%) (4.19%) (6.92%) (6.11%) (15.1%) (8.21%)<br />

框 架 承 担 的 地 震<br />

倾 覆 力 矩 比 ( 底<br />

层 )<br />

X 向<br />

Y 向<br />

23.18%<br />

22.03%<br />

11.12%<br />

10.14%<br />

21.19%<br />

12.72%<br />

14.86%<br />

8.64%<br />

23.79%<br />

19.06%<br />

15.28%<br />

13.21%<br />

顶 点 加 速 度 (m/s 2 )<br />

顺 风 向 0.041 0.050<br />

横 风 向 0.112 0.134<br />

从 以 上 计 算 结 果 , 分 析 比 较 如 下 :<br />

横 向 比 较 : 混 合 结 构 体 系 对 比 混 凝 土 结 构 体 系 具 有 以 下 特 征 :<br />

1. 总 质 量 减 少<br />

杭 州 坤 和 中 心 减 少 20%, 中 华 航 空 大 厦 19%, 亚 包 大 厦 19%( 以 下 均 按 此 项 目 排 序 进 行 比 较 )。<br />

-445-


混 合 体 系 单 位 面 积 质 量 约 1200~1400kg/m 2 , 混 凝 土 结 构 体 系 为 1500~1700kg/m 2 。<br />

2. 墙 柱 最 大 组 合 轴 力 减 小 , 且 外 围 框 架 减 幅 大 于 核 心 筒<br />

总 基 底 墙 柱 最 大 组 合 轴 力 减 小 幅 度 分 别 为 : 17%,16%,18%;<br />

外 围 框 架 柱 基 底 最 大 组 合 轴 力 减 小 幅 度 分 别 为 29%,26%,24%。<br />

3. 结 构 抗 侧 刚 度 减 小 , 第 一 、 二 周 期 明 显 加 大<br />

第 一 周 期 分 别 增 加 0.21s,0.51s,0.44s; 第 二 周 期 分 别 增 加 0.35s,0.32s,0.13s。<br />

4. 由 于 楼 层 质 量 减 轻 , 地 震 效 应 相 应 减 小<br />

地 震 荷 载 作 用 下 基 底 总 剪 力 :X 方 向 均 减 小 16%,Y 方 向 减 小 15~17%;<br />

基 底 倾 覆 弯 矩 :X 方 向 分 别 减 小 25%、19%、18%;,Y 方 向 分 别 减 小 23%、17%、16%。<br />

5. 刚 度 减 小 , 结 构 自 振 周 期 增 大 , 因 此 风 荷 载 效 应 增 大<br />

风 荷 载 作 用 下 , X 和 Y 方 向 基 底 剪 力 增 大 的 百 分 比 分 别 为 : 9.5%,9.0% 和 8.7%;<br />

基 底 倾 覆 弯 矩 增 加 10.5%,10.0% 和 9.7%。<br />

6. 风 载 作 用 下 的 层 间 位 移 角 增 幅 大 于 地 震 作 用 下 的 增 幅<br />

结 构 刚 度 变 柔 , 风 效 应 增 大 , 风 载 作 用 下 , 层 间 位 移 角 增 加 39%~114%;<br />

地 震 荷 载 作 用 下 , 层 间 位 移 角 增 加 10%~35%。<br />

7. 框 架 层 剪 力 分 担 率 减 小<br />

地 震 荷 载 作 用 下 , 钢 筋 混 凝 土 框 架 分 配 到 的 层 剪 力 为 9%~35%, 混 合 框 架 分 配 到 的 层 剪 力 只<br />

有 3.5%~25%, 两 种 形 式 框 架 部 分 分 配 到 的 剪 力 百 分 比 曲 线 对 比 如 下 图 。<br />

31<br />

36<br />

41<br />

楼 层<br />

26<br />

21<br />

16<br />

11<br />

6<br />

坤 和 钢 筋 砼 框 架 ‐ 核 心 筒 x 方 向<br />

坤 和 钢 筋 砼 框 架 ‐ 核 心 筒 y 方 向<br />

坤 和 混 合 框 架 ‐ 核 心 筒 x 方 向<br />

坤 和 混 合 框 架 ‐ 核 心 筒 y 方 向<br />

楼 层<br />

31<br />

26<br />

21<br />

16<br />

11<br />

6<br />

中 航 钢 筋 砼 框 架 ‐ 核 心 筒 x 方 向<br />

中 航 钢 筋 砼 框 架 ‐ 核 心 筒 y 方 向<br />

中 航 混 合 框 架 ‐ 核 心 筒 x 方 向<br />

中 航 混 合 框 架 ‐ 核 心 筒 y 方 向<br />

楼 层<br />

36<br />

31<br />

26<br />

21<br />

16<br />

11<br />

6<br />

亚 包 钢 筋 砼 框 架 ‐ 核 心 筒 x 方 向<br />

亚 包 钢 筋 砼 框 架 ‐ 核 心 筒 y 方 向<br />

亚 包 混 合 框 架 ‐ 核 心 筒 x 方 向<br />

亚 包 混 合 框 架 ‐ 核 心 筒 y 方 向<br />

1<br />

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1<br />

柱 剪 力 所 占 百 分 比<br />

1<br />

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1<br />

柱 剪 力 所 占 百 分 比<br />

1<br />

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1<br />

柱 剪 力 所 占 百 分 比<br />

上 图 虚 线 所 示 , 混 合 结 构 体 系 的 框 架 柱 分 配 的 层 剪 力 大 多 达 不 到 20%, 下 部 楼 层 甚 至 达 不 到<br />

10%, 绝 大 部 分 剪 力 由 核 心 筒 承 担 。<br />

8. 两 种 结 构 体 系 起 控 制 作 用 的 水 平 荷 载 均 是 风 荷 载<br />

纵 向 比 较 : 每 种 结 构 形 式 的 三 个 不 同 高 度 建 筑 从 低 到 高 分 别 为 杭 州 坤 和 中 心 (118 米 ), 中 华 航<br />

空 大 厦 (150 米 ), 亚 包 大 厦 (177 米 ), 随 高 度 不 同 , 结 构 性 能 显 示 以 下 比 例 关 系 :<br />

-446-


主 要 技 术 指 标 对 比 表 之 二<br />

结 构 体 系 混 凝 土 结 构 体 系 混 合 结 构 体 系<br />

总 高 度 之 比 1 : 1.27 : 1.50 1 : 1.27 : 1.50<br />

总 层 数 之 比 1 : 1.19 : 1.48 1 : 1.19 : 1.48<br />

总 质 量 之 比 1 : 1.22 : 1.61 1 : 1.24 : 1.63<br />

总 基 底 最 大 组 合 轴 力 之 比 1 : 1.23 : 1.51 1 : 1.26 : 1.49<br />

风 作 用 下 基 底<br />

总 剪 力 之 比<br />

风 作 用 下 基 底<br />

总 弯 矩 之 比<br />

地 震 作 用 下 基<br />

底 总 剪 力 之 比<br />

地 震 作 用 下 基<br />

底 总 弯 矩 之 比<br />

X 向 1 : 1.38 : 1.95 1 : 1.38 : 1.94<br />

Y 向 1 : 1.35 : 1.93 1 : 1.35 : 1.92<br />

X 向 1 : 1.68 : 2.76 1 : 1.67 : 2.74<br />

Y 向 1 : 1.67 : 2.76 1 : 1.67 : 2.74<br />

X 向 1 : 1.06 : 1.42 1 : 1.07 : 1.42<br />

Y 向 1 : 1.34 : 1.51 1 : 1.40 : 1.54<br />

X 向 1 : 1.33 : 2.07 1 : 1.43 : 2.26<br />

Y 向 1 : 1.43 : 2.11 1 : 1.55 : 2.31<br />

从 上 表 的 计 算 结 果 分 析 , 建 筑 高 度 的 增 加 , 在 两 种 结 构 体 系 上 所 产 生 的 效 应 较 为 相 似 :<br />

1. 总 质 量 和 基 底 组 合 轴 力 近 似 线 性 增 加 。<br />

2. 风 载 效 应 随 高 度 增 大 , 且 风 载 剪 力 效 应 增 幅 大 于 高 度 增 幅 ; 弯 矩 效 应 增 幅 大 于 高 度 增 幅 的 平 方 。<br />

3. 风 荷 载 效 应 的 增 幅 , 两 种 结 构 体 系 基 本 相 同 。<br />

4. 地 震 荷 载 效 应 的 增 幅 , 混 合 结 构 体 系 大 于 混 凝 土 结 构 体 系 。<br />

四 、 基 础 设 计 及 对 比<br />

三 个 工 程 均 设 三 层 地 下 室 , 基 础 形 式 均 采 用 桩 筏 基 础 , 其 中 杭 州 坤 和 中 心 采 用 桩 径 1000 的 钻<br />

孔 灌 注 桩 , 持 力 层 为 中 风 化 泥 质 粉 砂 岩 , 有 效 桩 长 44~57 米 , 单 桩 承 载 力 特 征 值 为 5350KN, 每 桩<br />

的 造 价 约 32000 元 ( 以 桩 长 50 米 计 ); 中 华 航 空 大 厦 及 亚 包 大 厦 均 为 桩 径 850 的 钻 孔 灌 注 桩 ( 压 力<br />

注 浆 ), 持 力 层 为 圆 砾 层 , 有 效 桩 长 约 为 30 米 , 单 桩 承 载 力 特 征 值 分 别 为 4850KN 及 4500KN, 每<br />

桩 的 造 价 约 23500 元 。 基 础 对 比 情 况 见 下 表 :<br />

基 础 对 比 表<br />

工 程 名 称 杭 州 坤 和 中 心 中 华 航 空 大 厦 亚 包 大 厦<br />

结 构 体 系 混 凝 土 结 构 混 合 结 构 混 凝 土 结 构 混 合 结 构 混 凝 土 结 构 混 合 结 构<br />

高 层 范 围 总 桩 数 ( 根 ) 160 132 216 182 286 234<br />

高 层 部 分 底 板 厚 (m) 2.4 2.2 2.6 2.4 2.8 2.5<br />

由 上 表 比 较 可 知 , 当 采 用 混 合 结 构 体 系 时 , 总 桩 数 比 混 凝 土 结 构 体 系 减 少 约 16~18%, 三 幢 建<br />

筑 总 桩 数 分 别 减 少 28 根 ,34 根 和 52 根 , 桩 基 造 价 分 别 节 约 90 万 元 ,80 万 元 和 122 万 元 。 基 础 底<br />

板 厚 度 减 小 0.2~0.3 米 , 可 减 少 造 价 约 15~30 万 元 。<br />

-447-


五 、 结 构 构 件 设 计 及 对 比<br />

主 要 结 构 构 件 尺 寸 对 比 表<br />

工 程 名 称 杭 州 坤 和 中 心 中 华 航 空 大 厦 亚 包 大 厦<br />

结 构 体 系 混 凝 土 结 构 混 合 结 构 混 凝 土 结 构 混 合 结 构 混 凝 土 结 构 混 合 结 构<br />

最 大 柱 截 面<br />

(mm)<br />

梁<br />

(mm)<br />

核 心 筒 外 墙 厚<br />

(mm)<br />

核 心 筒 内 墙 厚<br />

(mm)<br />

600×600×30<br />

Φ800×36<br />

1150×1150<br />

Φ750×28 1400×1400<br />

450×600×30 Φ1300<br />

Φ700×30<br />

700×1150<br />

( 圆 钢 管 砼 ) 1100×1100<br />

( 方 钢 管 砼 )<br />

( 圆 钢 管 砼 )<br />

500×600<br />

500×650 箱 形 400×450 500×700 H550×300 700×750<br />

( 箱 形 梁 )<br />

300×650 H 型 450×200 300×600 H450×200 400×700<br />

H600×250<br />

300~450 300~500 300~600 300~600 400~600 400~600<br />

250~300 250~350 200~300 200~300 200~250 200~250<br />

通 过 对 三 幢 建 筑 的 构 件 设 计 , 总 结 了 几 点 体 会 :<br />

1. 采 用 钢 管 混 凝 土 柱 可 大 幅 减 小 柱 的 断 面 , 增 加 有 效 使 用 面 积<br />

以 KZ1 为 例 , 坤 和 中 心 减 少 约 73%, 中 华 航 空 减 少 约 75%, 亚 包 大 厦 减 少 约 74%; 柱 截<br />

面 减 小 使 坤 和 中 心 单 层 有 效 面 积 增 加 约 16m 2 , 中 华 航 空 增 加 16m 2 , 亚 包 大 厦 增 加 约 24m 2 。<br />

2. 钢 梁 高 度 加 上 楼 板 高 度 后 , 梁 高 降 低 的 优 势 不 明 显<br />

类 似 的 超 高 层 建 筑 , 核 心 筒 外 边 与 外 框 柱 距 离 大 约 为 10.5 米 , 钢 梁 与 混 凝 土 梁 相 比 , 梁 高 可 适<br />

当 降 低 , 钢 梁 高 度 加 上 楼 板 高 度 比 混 凝 土 梁 高 减 少 大 约 40~90mm。 钢 梁 的 梁 高 优 势 对 于 跨 度 较 大<br />

的 建 筑 方 能 更 好 体 现 。<br />

3. 部 分 设 备 管 线 从 钢 梁 腹 板 的 开 洞 处 穿 越 , 可 以 减 小 吊 顶 高 度 , 有 效 增 加 楼 层 净 高<br />

钢 梁 腹 板 易 于 开 洞 , 可 使 部 分 水 电 管 线 从 腹 板 穿 过 。 做 好 综 合 管 线 设 计 , 在 保 证 楼 层 净 高 不 变<br />

的 情 况 下 , 若 每 层 设 备 管 线 占 用 高 度 减 少 100mm, 则 有 可 能 增 加 出 一 个 楼 层 , 这 会 产 生 很 大 的 经 济<br />

效 益 !<br />

4. 采 用 钢 管 混 凝 土 柱 可 更 好 地 满 足 建 筑 立 面 和 功 能 需 求<br />

以 坤 和 中 心 为 例 , 由 于 建 筑 立 面 要 求 纤 细 的 密 柱 效 果 , 若 采 用 混 凝 土 柱 宽 度 需 要 700, 加 上 两<br />

侧 干 挂 石 材 的 厚 度 , 完 成 后 柱 子 总 宽 度 近 900, 完 全 达 不 到 建 筑 要 求 的 效 果 , 改 用 钢 管 混 凝 土 柱 后<br />

建 筑 外 包 宽 度 减 小 为 700, 解 决 了 建 筑 与 结 构 间 的 矛 盾 , 从 平 面 布 置 图 中 可 看 到 两 者 的 不 同 效 果 。<br />

5. 按 刚 度 分 配 框 架 的 层 剪 力 分 担 率 较 小 时 , 应 加 强 核 心 筒 抗 震 性 能<br />

混 合 结 构 外 围 框 架 刚 度 较 小 , 有 必 要 加 强 混 凝 土 核 心 筒 的 侧 向 承 载 能 力 , 以 保 证 核 心 筒 的 抗 震<br />

性 能 。 为 提 高 核 心 筒 的 延 性 , 坤 和 中 心 采 取 了 以 下 措 施 : 在 核 心 筒 角 部 设 角 柱 , 并 内 设 型 钢 ( 也 便<br />

于 刚 接 ), 加 强 角 部 配 筋 , 以 加 强 筒 体 角 部 延 性 和 受 力 性 能 , 同 时 加 强 了 核 心 筒 的 抗 震 构 造 ; 核 心<br />

筒 部 分 墙 体 加 厚 , 以 控 制 墙 体 剪 应 力 ; 在 楼 层 标 高 处 均 设 钢 筋 混 凝 土 暗 梁 ; 连 梁 内 配 置 交 叉 暗 撑 。<br />

6. 钢 梁 应 考 虑 与 楼 板 形 成 组 合 梁 , 以 减 少 钢 梁 高 度 及 用 钢 量<br />

组 合 梁 可 提 高 梁 的 受 力 和 变 形 性 能 , 有 效 减 少 钢 梁 用 钢 量 。 注 意 框 架 梁 的 上 部 负 弯 矩 不 应 考 虑<br />

梁 板 组 合 作 用 , 可 在 梁 柱 ( 墙 ) 节 点 处 加 腋 , 增 强 支 座 的 抗 弯 能 力 , 提 高 钢 梁 整 体 的 受 力 性 能 。<br />

组 合 梁 中 钢 梁 上 翼 缘 可 适 当 减 小 以 充 分 发 挥 组 合 梁 的 作 用 , 降 低 用 钢 量 。<br />

-448-


六 、 上 部 结 构 材 料 用 量 及 经 济 性 对 比<br />

上 部 结 构 材 料 用 量 及 经 济 性 对 比 表<br />

工 程 名 称 杭 州 坤 和 中 心 中 华 航 空 大 厦 亚 包 大 厦<br />

结 构 体 系 混 凝 土 结 构 混 合 结 构 混 凝 土 结 构 混 合 结 构 混 凝 土 结 构 混 合 结 构<br />

钢 筋<br />

混 凝<br />

土 部<br />

分<br />

钢 结<br />

构 部<br />

分<br />

柱 1936 952 2213 1089 5523 2354<br />

混 凝 土<br />

墙 5887 5930 8254 8254 13358 13358<br />

用 量<br />

梁 6004 126 6644 364 9407 882<br />

(m 3 )<br />

板 5197 5197 5789 5789 8285 8285<br />

总 计 19024 12205 22900 15496 36573 24879<br />

钢 筋 混 凝 土 造 价 ( 万<br />

元 )<br />

2854 1831 3435 2324 5486 3732<br />

单 位 面 积<br />

混 凝 土 用 量 (m 3 )<br />

0.39 0.25 0.41 0.28 0.48 0.33<br />

单 位 面 积 混 凝 土<br />

结 构 造 价 ( 元 /m 2 )<br />

586 376 618 415 722 491<br />

柱 917 1009 1719<br />

型 钢<br />

梁 2117 2482 3323<br />

用 量 (T)<br />

总 计 3034 3491 5042<br />

型 钢 造 价 ( 万 元 ) 2579 2967 4286<br />

单 位 面 积 型 钢<br />

62.3 62.8 66.3<br />

用 量 ( kg/m 2 )<br />

单 位 面 积 型 钢<br />

造 价 ( 元 /m 2 )<br />

防 火 涂 料 造 价<br />

( 万 元 )<br />

530 534 564<br />

244 278 380<br />

总 造 总 造 价 ( 万 元 ) 2854 4654 3435 5569 5486 8398<br />

价 单 位 面 积 造 价 586 956 618 1002 722 1105<br />

注 : 钢 筋 混 凝 土 结 构 造 价 按 0.15 万 元 / m 3 计 算<br />

钢 结 构 造 价 按 0.85 万 元 / T 计 算<br />

防 火 涂 料 按 50 元 / m 2 计 算<br />

从 上 表 的 计 算 结 果 可 以 看 出 :<br />

1. 单 纯 就 上 部 建 筑 的 结 构 部 分 ( 仅 地 下 室 以 上 梁 柱 板 墙 ) 造 价 而 言 , 混 合 结 构 体 系 相 比 混 凝 土 结<br />

构 体 系 增 加 较 多<br />

坤 和 中 心 造 价 增 加 1800 万 元 , 增 幅 为 63%; 中 华 航 空 增 加 2134 万 元 , 增 幅 62%; 亚 包 大 厦 增<br />

加 2912 万 元 , 增 幅 53%。<br />

2. 混 合 结 构 与 混 凝 土 结 构 相 比 , 可 大 幅 度 节 约 混 凝 土 材 料 用 量<br />

混 凝 土 用 量 节 约 的 幅 度 , 坤 和 中 心 为 36%, 中 华 航 空 为 32%, 亚 包 大 厦 为 32%。<br />

3. 建 筑 高 度 增 加 , 材 料 用 量 也 增 加 , 且 两 者 呈 非 线 性 关 系<br />

以 混 凝 土 结 构 为 例 , 从 118m 到 150m, 高 度 的 增 量 为 27%, 混 凝 土 材 料 用 量 的 增 量 为 20%; 从<br />

150m 到 177m, 高 度 的 增 量 为 18%, 材 料 用 量 的 增 量 为 60% 。 混 合 结 构 , 型 钢 用 量 的 增 量 分 别<br />

为 15% 和 66%。<br />

4. 混 合 结 构 中 , 钢 筋 混 凝 土 造 价 占 总 造 价 的 百 分 比 随 高 度 增 加 而 增 加<br />

-449-


三 幢 建 筑 钢 筋 混 凝 土 造 价 占 总 造 价 的 百 分 比 分 别 为 :39%,42%,44%。<br />

5. 随 着 建 筑 高 度 增 加 , 地 下 室 以 上 结 构 部 分 造 价 的 增 幅 , 混 合 结 构 小 于 混 凝 土 结 构<br />

三 幢 建 筑 上 部 结 构 部 分 总 造 价 之 比 分 别 为 : 混 凝 土 结 构 1:1.20:1.92; 混 合 结 构 1:1.20:1.80。<br />

七 、 综 合 技 术 和 经 济 分 析<br />

技 术 分 析 总 结 :<br />

1. 对 于 100~200 米 的 超 高 层 建 筑 , 两 种 结 构 体 系 均 适 用 。 混 合 结 构 体 系 与 混 凝 土 结 构 体 系 相<br />

比 , 总 质 量 减 小 , 风 载 效 应 增 大 , 地 震 效 应 减 小 , 柱 截 面 大 幅 减 小 , 尽 管 外 围 框 架 的 刚 度 相 对 减 弱 ,<br />

但 位 移 和 内 力 控 制 都 能 满 足 要 求 。<br />

2. 外 围 框 架 部 分 的 刚 度 , 能 否 起 到 二 道 防 线 作 用 , 是 这 两 种 结 构 形 式 设 计 的 核 心 问 题<br />

混 合 结 构 外 围 框 架 的 刚 度 小 于 混 凝 土 结 构 , 外 围 框 架 的 刚 度 及 框 架 层 剪 力 分 担 率 的 设 计 应 具 体<br />

斟 酌 。 要 保 证 外 框 架 起 到 二 道 防 线 作 用 , 外 框 架 应 具 有 适 当 刚 度 ; 对 框 架 层 剪 力 分 担 率 不 满 足 要 求<br />

的 楼 层 , 应 按 规 范 要 求 乘 以 地 震 剪 力 调 整 系 数 。<br />

3. 混 合 结 构 体 系 应 更 加 重 视 和 加 强 核 心 筒 的 设 计<br />

两 种 结 构 形 式 内 部 核 心 筒 的 尺 度 、 刚 度 几 乎 一 样 , 变 化 的 仅 是 外 围 框 架 部 分 。 混 合 框 架 比 混 凝<br />

土 框 架 刚 度 小 , 分 担 的 地 震 剪 力 很 小 , 整 个 建 筑 的 抗 震 性 能 很 大 程 度 取 决 于 核 心 筒 , 保 证 核 心 筒 的<br />

延 性 十 分 重 要 。 当 框 架 按 刚 度 计 算 分 配 的 最 大 楼 层 地 震 剪 力 小 于 结 构 总 地 震 剪 力 的 10% 时 , 核 心 筒<br />

承 担 的 地 震 作 用 应 加 大 , 甚 至 让 筒 体 具 有 承 担 100% 的 地 震 剪 力 的 能 力 , 而 且 将 核 心 筒 抗 震 等 级 提<br />

高 一 级 。<br />

4. 与 钢 筋 混 凝 土 柱 相 比 , 钢 管 混 凝 土 柱 具 有 明 显 优 势<br />

a), 柱 截 面 面 积 大 幅 减 少 约 70%~75%, 有 效 使 用 面 积 增 加 ; b), 钢 管 取 代 了 混 凝 土 柱 中 的<br />

纵 筋 和 箍 筋 , 工 厂 制 作 取 代 了 现 场 绑 扎 ;c), 钢 管 自 身 作 为 抗 侧 压 模 板 , 施 工 不 需 要 支 模 , 也 不 受<br />

混 凝 土 养 护 时 间 限 制 ;d), 钢 管 约 束 了 混 凝 土 的 横 向 变 形 , 使 混 凝 土 强 度 得 以 提 高 ;e), 钢 管 混 凝<br />

土 柱 的 延 性 好 ;f), 混 凝 土 的 填 充 限 制 了 钢 管 的 局 部 屈 曲 , 使 钢 材 强 度 更 好 地 发 挥 , 也 使 钢 管 的 防<br />

火 性 能 提 高 。<br />

钢 管 混 凝 土 应 用 中 尚 存 在 一 些 问 题 , 如 管 内 混 凝 土 在 硬 化 过 程 中 的 收 缩 导 致 管 壁 与 混 凝 土<br />

粘 结 不 紧 密 、 长 期 混 凝 土 徐 变 对 钢 管 与 混 凝 土 整 体 性 的 影 响 等 问 题 尚 需 进 一 步 研 究 。<br />

5. 混 合 结 构 工 厂 化 程 度 高 , 施 工 周 期 短 , 更 为 环 保<br />

采 取 交 错 施 工 的 方 案 , 混 凝 土 核 心 筒 的 施 工 比 钢 框 架 的 安 装 提 前 4~5 层 , 达 到 一 定 强 度 后 的<br />

核 心 筒 可 以 作 为 钢 框 架 施 工 时 的 稳 定 支 承 , 使 周 边 框 架 易 于 施 工 ; 梁 、 柱 的 施 工 省 去 了 钢 筋 混 凝 土<br />

绑 扎 钢 筋 、 支 模 拆 模 、 等 候 养 护 等 过 程 ; 楼 板 采 用 压 型 钢 板 做 模 板 , 上 浇 钢 筋 混 凝 土 楼 板 , 可 多 层<br />

同 时 操 作 , 不 必 像 混 凝 土 结 构 必 须 等 混 凝 土 达 到 一 定 强 度 后 才 能 进 行 下 一 层 施 工 , 可 有 效 加 快 施 工<br />

速 度 。<br />

6. 混 合 结 构 造 价 较 高 , 防 腐 与 防 火 需 要 后 期 维 护<br />

经 济 分 析 总 结 :<br />

混 合 结 构 体 系 与 混 凝 土 结 构 体 系 相 比 , 综 合 经 济 比 较 如 下 :<br />

1. 地 下 室 以 上 的 结 构 造 价 增 加 约 53%~63%, 桩 基 基 础 造 价 节 省 约 16~18%, 综 合 后 结 构 造 价<br />

的 增 加 值 如 下 :( 折 合 面 积 仅 计 地 下 室 以 上 的 面 积 , 若 计 入 包 括 地 下 室 在 内 的 整 个 工 程 总 面 积 , 则<br />

单 位 面 积 增 加 的 造 价 相 应 降 低 )<br />

坤 和 中 心 :1690( 万 元 ), 折 合 单 位 面 积 增 加 347 元 /m 2 。<br />

中 华 航 空 :2034( 万 元 ), 折 合 单 位 面 积 增 加 366 元 /m 2 。<br />

亚 包 大 厦 :2760( 万 元 ), 折 合 单 位 面 积 增 加 363 元 /m 2 。<br />

2. 上 部 结 构 造 价 仅 占 整 个 工 程 总 造 价 的 小 部 分 , 假 设 整 个 建 筑 的 单 位 面 积 总 造 价 为 3500 元<br />

-450-


m 2 , 三 个 项 目 采 用 混 合 结 构 所 增 加 的 造 价 分 别 为 建 筑 总 造 价 的 9.9%,10.5%,10.4%。 所 以 采 用 混<br />

合 结 构 增 加 的 造 价 占 整 个 工 程 总 造 价 的 比 例 约 为 10%( 钢 结 构 按 0.85 万 元 / 吨 计 算 ), 若 钢 结 构 按 1<br />

万 元 / 吨 计 算 , 增 加 的 造 价 大 约 占 工 程 总 造 价 的 12%。<br />

3. 钢 梁 的 应 用 使 结 构 和 设 备 所 占 高 度 减 少 , 在 满 足 净 高 要 求 下 降 低 层 高 , 由 此 可 使 建 筑 总 层<br />

数 得 以 增 加 一 ~ 二 层 。 以 杭 州 坤 和 中 心 为 例 , 在 保 证 楼 层 净 高 不 变 的 前 提 下 , 标 准 层 层 高 由 原 来 3.6<br />

米 降 低 为 3.45 米 , 每 层 减 少 0.15 米 ,24 个 标 准 层 累 计 减 少 3.6 米 , 因 此 整 个 建 筑 增 加 了 一 层 , 经<br />

济 效 益 可 观 。<br />

4. 良 好 的 施 工 组 织 可 使 混 合 结 构 建 造 工 期 缩 短 , 产 生 相 应 经 济 效 益 。<br />

5. 有 效 使 用 面 积 增 加 , 下 部 楼 层 有 效 面 积 可 增 加 1% 以 上 , 为 使 用 者 带 来 实 实 在 在 的 利 益 。<br />

八 、 后 语<br />

两 种 结 构 选 型 应 结 合 工 程 具 体 条 件 、 工 期 、 造 价 、 建 筑 效 果 等 因 素 综 合 考 虑 。<br />

混 合 结 构 在 超 高 层 建 筑 中 有 一 定 的 综 合 优 势 , 因 而 发 展 趋 势 较 好 , 目 前 制 约 混 合 结 构 发 展 的 主<br />

要 因 素 有 :1, 造 价 较 高 ;2, 相 关 研 究 和 规 范 编 制 滞 后 ;3, 设 计 人 员 更 熟 悉 传 统 混 凝 土 结 构 的 设<br />

计 ;4, 对 两 种 结 构 形 式 的 工 程 对 比 经 验 有 限 。<br />

希 望 本 文 能 为 超 高 层 结 构 选 型 提 供 借 鉴 和 帮 助 。 本 文 仅 讨 论 了 抗 震 设 防 烈 度 六 度 的 情 况 , 以 后<br />

将 另 文 讨 论 类 似 100 米 ~200 米 超 高 层 建 筑 在 更 高 烈 度 区 的 情 况 比 较 。<br />

参 考 文 献<br />

[1] 徐 培 福 等 . 复 杂 高 层 建 筑 结 构 设 计 .<br />

[2] 高 立 人 等 . 高 层 建 筑 结 构 概 念 设 计 .<br />

[3] 建 筑 抗 震 设 计 规 范 GB 50011-2010.<br />

[4] 高 层 建 筑 钢 - 混 凝 土 混 合 结 构 设 计 规 程 , CECS 230: 2008.<br />

-451-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

WIND-INDUCED STRUCTURAL DYNAMIC RESPONSE ANALYSIS FOR A<br />

LARGE STADIUM CABLE-NET ROOF<br />

W. J. CHEN 1 *, T. REN 2 , Y.L.HE 3<br />

1 * Professor, Space Structures Research Centre, Shanghai Jiao Tong <strong>University</strong> Shanghai 200030, P. R. China<br />

http://www.ssrc.cn, http:// naoce.sjtu.edu.cn, cwj@sjtu.edu.cn<br />

2 Graduate Student, Space Structures Research Centre, Shanghai Jiao Tong <strong>University</strong><br />

3 Associate Professor, Space Structures Research Centre, Shanghai Jiao Tong <strong>University</strong><br />

ABSTRACT<br />

Jaber Al-Ahmad International Stadium adopted a novel lightweight cable-net roof system, single layer saddle<br />

radial-ring cable-net. This paper investigated the structural performance thoroughly, the pre-stress level and<br />

stiffness, static loading and deformation with respect to asymmetry wind simulated through CFD. And the<br />

dynamic fluctuation of tension and deflection were evaluated with developed algorithm. Finally, the lifting<br />

process was simulated with the developed method, which adopt generalized inverse and consider rigid free<br />

movement and elastic deformation.<br />

KEYWORDS<br />

Single layer saddle radial-ring cable-net, dynamic response, lifting simulation.<br />

INTRODUCTION<br />

<strong>The</strong> cable has high strength, small size and lightweight. It is the key structural member for the modern large<br />

span/space structure. This makes the structural member moment free with pre-tension introduction to develop<br />

the overall form and stiffness tendency and require high accuracy manufacture and installation.<br />

<strong>The</strong> cable-net is one of the ways to realize large span with high mast. <strong>The</strong> pioneer was ILEK, lightweight<br />

structure institute. <strong>The</strong> classical representative is Munchen Olympic stadium in 1972 (shown in Figure.1) and<br />

Germany Pavilion in 1967 EXPO[1]. With development of membrane structure in 1960’s, the tensioned<br />

cable-truss appeared, jointed with fabric. <strong>The</strong> conception of flat wheel with spokes was created and evolutes.<br />

<strong>The</strong> centre changed into a tension ring and out compression ring. <strong>The</strong> tie cables were added. <strong>The</strong> plan can be<br />

oval, quasi-square, un-full polar symmetry. Some time there has only single outer compression ring, two centre<br />

tension-ring, like inverse wheel. <strong>The</strong> typical projects have Rome Olympic stadium in 1990, Gottlieb Daimler<br />

stadium in 1993 (shown in Figue.2), Outdoor stadium in Kuala Lumpur Malaysia 1998[1], Olympic stadium La<br />

Cartuja Seville Spain 1999[2], Fushan century lotus stadium in Fushan China 2006, and Baan stadium in<br />

Shenzhen China due finished 2010.<br />

Jaber Al-Ahmad International Stadium in Kuwait 2004 adopted a novel lightweight cable-net roof system, single<br />

layer saddle radial-ring cable-net, finished in end of 2007. This paper investigated comprehensively the<br />

structural performance of this stadium to understand the general feature of this kind of structure and to develop<br />

the involved structural design and numerical simulation method.<br />

Figure.1 Munchen Olympic stadium<br />

Figure.2 Gottlieb Daimler stadium<br />

-452-


STADIUM ROOF BRIEF<br />

Jaber Al-Ahmad International Stadium adopted a single layer saddle radial-ring cable-net covered with<br />

membranes. <strong>The</strong> upper edge of the stand undulates greatly due to optimization of the viewing distance and at the<br />

same time provide requisite for a saddle-shaped double-curved surface. <strong>The</strong> plan is almost circular with main<br />

axes 280mx260m. <strong>The</strong> roof opening over the field is formed by the cable tension ring, which has an affinity to<br />

the outer compression ring. By contrast, the oval membrane roof is aligned to the spectator stands [2].<br />

a) photo in construction b) single layer cable net roof<br />

Figure.3 Jaber Al-Ahmad International Stadium<br />

As shown in Figure.3, the overall plan is 282m 256m elliptic with openning 109m 119m elliptic. <strong>The</strong> roof<br />

consists of 55 radial cables (RAC) and 10 ring cables (RIC), a PTFE fabric tensegrity module covers a cable<br />

gird with four lower stabilized cables and a flying strut, over three hundred modules with various sizes cover<br />

whole roof. This looks like white wavelet over the sand [3].<br />

As the main structural performance of cable net concerned, only cable nets was thus taken into the analysis<br />

model without outer undulation compression ring, huge circular hollow section steel beam, and coverage of<br />

membrane. <strong>The</strong>y are assumed boundary and dead load. Table.1 shows the cable size. <strong>The</strong> modulus set 190GPa.<br />

Location<br />

Table.1 <strong>The</strong> size of cables<br />

RIC (Numbering from opening to outer ring)<br />

RAC<br />

1 2 3 4 5~10<br />

size 8-(5×187) 5×223 5×211 5×199 5×187 5×199<br />

PRE-STRESS AND STATIC PERFORMANCE<br />

Pre-stress has two functions to the flexible cable nets, the first is to form the shape, the second is to develop and<br />

maintain stiffness. Rational geometry of negative gauss curvature and introduction of pre-stress are the basic<br />

requisite to achieve high performance of the cable net roof.<br />

Figue.4 shows cable net layout, the size gives in Table.1. <strong>The</strong> initial location of the cable net was set according<br />

to architect design. <strong>The</strong> initial pre-stress were optimized with respect to the deflection requirement of centre<br />

tension ring. <strong>The</strong> cable net can be assumed all un-compression pre-stressed link assembly, which is proved<br />

one static redundancy and one kinematic mechanism. Only one cable set specific pre-stress, the tensions of other<br />

cables can be easily calculated from the called force-finding method which is based on the linear adjustment<br />

theory and force density method. It was a try-on process and finally the optimized pre-stress were got in<br />

engineering sense. Table.2 gives the initial pre-stress of the cable nets.<br />

-453-


(8)<br />

(10)<br />

(12) B<br />

(6)<br />

(4)<br />

(2)<br />

A<br />

C<br />

D<br />

Figue.4 Cable net layout<br />

Table.2 the initial pre-stress of the cable nets<br />

Location<br />

RIC<br />

1 2 3 4 5 6<br />

Pre- stress( 103kN ) 51.05~54.60 21.57~22.65 19.64~20.83 17.42~18.39 17.30~18.16 16.89~17.61<br />

RIC<br />

Location<br />

RAC<br />

7 8 9 10<br />

Pre- stress( 103kN ) 16.15~16.72 15.04~15.49 14.91~15.29 14.59~14.91 5.91~26.64<br />

Deflexion (m)<br />

0.40<br />

0.35<br />

0.30<br />

0.25<br />

0.20<br />

0.15<br />

0.10<br />

0.05<br />

0.00<br />

Load step<br />

1<br />

2<br />

3<br />

4<br />

5<br />

Link A<br />

1 2 3 4 5 6 7 8 9 10 11<br />

Node<br />

Figure.5 Deflection of cable nets<br />

Tension (kN)<br />

51110<br />

51100<br />

1<br />

51090<br />

51080<br />

51070<br />

51060<br />

51050<br />

1/5 2/5 3/5 4/5 1<br />

Load Step<br />

Figure.6 Tension of cable nets<br />

In the case of uniform loading condition as dead load plus snow, 0.55kN/m2, and nonlinear static analysis was<br />

carried out with load increment of five steps. Figure.5 shows the deflection of cable link-A. <strong>The</strong> maximum is<br />

about 0.4m, less 1/250 cover span, and the deflection increase approximated to linear with load add, but increase<br />

fairly small from outer to opening unlike cantilever behaviour. Figure.6 shows the inner ring cable tension with<br />

respect to loading, which increase un-strong nonlinearly.<br />

WIND-INDUCED DYNAMIC PERFORMANCE<br />

<strong>The</strong> light structure is proved sensitive to wind action. <strong>The</strong> basic dynamic structural performance needs<br />

evaluating in structural design. <strong>The</strong> static nonlinear analysis results indicated the pre-stressed cable net does not<br />

behaviour strong nonlinearly. <strong>The</strong> frequency domain method was employed to calculate the response. <strong>The</strong><br />

modal analysis was thus performed firstly.<br />

-454-


a) 1st modal shape<br />

Figure.7 Modal shape of the cable net<br />

b) 10th modal shape<br />

Figure.7 shows the modal shape of cable net. <strong>The</strong> first modal has frequency 1.49Hz, this shows fairly high<br />

stiffness “unlike flexible”. And the 10th modal has 4.172Hz. <strong>The</strong> overall modal is much dense.<br />

In the case of complex wind, the equivalent static load is always used for design, and the dynamic response used<br />

for advanced performance evaluation. <strong>The</strong> equivalent static load is calculated from the reference wind pressure,<br />

terrain roughness factor, topography factor and Form Pressure Coefficients (FPC). Due to free form, FPC is<br />

usually got by wind tunnel. CFX was herein employed to get FPC with various wind inward direction. Figure.8a<br />

shows the FPC with zero wind inward, and Figure.8b shows the facet FPC over the roof. <strong>The</strong> overall roof is<br />

subjected to asymmetry wind.<br />

a) FPC over the roof b) Faceted FPC over the roof<br />

Figure.8 Form pressure coefficients of the cable net roof<br />

<strong>The</strong> frequency domain method is proved simple and effective to calculate the dynamic response for the linear or<br />

weak nonlinear structure. <strong>The</strong> modal compensation method was developed here. All normal modal considered is<br />

inefficient due to great node, the cut-off error would happen without reasonable number of modal taken into<br />

account. <strong>The</strong> modal compensation method firstly consider specific number of modal, together with the<br />

compensation modal which is calculated on the total strain energy. <strong>The</strong> nodal displacement wind fluctuation<br />

factor is defined<br />

where,<br />

wind,<br />

−<br />

u<br />

i+ μσ ⋅<br />

ui<br />

⋅sign( ui)<br />

Ui<br />

βU<br />

= = (1)<br />

i<br />

−<br />

−<br />

u<br />

u<br />

u i is the displacement of i-node subjected to average wind,<br />

U is the sum, μ is the peak factor, assumes 2.5.<br />

i<br />

i<br />

−<br />

i<br />

σ<br />

u i<br />

is the root-square of fluctuate<br />

-455-


Table.3 the nodal displacement wind fluctuation factor<br />

Nodal No. 5M 5M+MC 10M 10M+MC 30M 30M+MC 500M<br />

1 1.723 1.914 1.943 2.469 1.947 2.462 2.589<br />

2 1.534 1.705 1.738 2.013 1.744 2.019 2.117<br />

4 1.353 1.503 1.547 1.637 1.556 1.662 1.732<br />

6 1.254 1.393 1.457 1.466 1.470 1.521 1.568<br />

8 1.183 1.315 1.401 1.362 1.433 1.422 1.462<br />

10 1.143 1.270 1.369 1.307 1.429 1.386 1.414<br />

Table.3 gives the nodal displacement wind fluctuation factor for one RAC partial node at inward roof. 5M<br />

means five modals. 5M+MC means five modals plus compensation modal. It indicate that the reasonable result<br />

can be obtained in 30M+MC, there has little difference compared with 500M. And the results also clarify that<br />

the cable net behaves significant dynamic flutter. Similarly, the fluctuation factor of cable tension can also be<br />

defined and calculated, in contrast, the results indicate the change is too small with respect to fairly big initial<br />

pre-stress.<br />

LIFTING PROCESS SIMULATION<br />

<strong>The</strong> cable nets are generally assembled into a whole roof on the ground, then lift the boundary node with<br />

auxiliary cable and jack to the stand mount location, finally stretch the cables to introduce pre-tension. <strong>The</strong> cable<br />

has large rigid movement and elastic deformation during this process, this pose great change to simulate.<br />

Step-20<br />

Step-15<br />

Step-10<br />

Step-5 状<br />

Step-1 状<br />

Step-0 放<br />

Figure 9. Lifting simulation of cable net roof<br />

<strong>The</strong> catenary flexible element was simulated as cable without pre-stress. And the generalized inverse was used<br />

to solve singular equation to get ordinary solution, plus the equilibrium solution. <strong>The</strong> developed algorithm was<br />

coded to simulate the construction. Figure.9 shows the cable net form variation, and clarifies the introduction of<br />

the pre-stress.<br />

CONCLUSION<br />

This paper investigated the structural performance thoroughly of a new single-layer cable-net roof to interpret<br />

general nature of this system. And introduced the analysis method as pre-stress optimization, wind-induced<br />

dynamic response and lifting process simulation algorithm.<br />

REFERENCES<br />

[1] Kazuo Ishii (1999). Membrane designs and structures in the world, Shinkenchiku-sha.<br />

[2] Schlaich Bergermann und Partner (2004). Light structures, Prestel, DAM.<br />

[3] Ren, T. (2008). Structural analysis and construction simulation research for the radial-ring single-layer<br />

cable-net stadium roof, Shanghai Jiao Tong <strong>University</strong>.<br />

-456-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

转 体 施 工 的 中 承 式 拱 梁 组 合 体 系 桥 的 受 力 性 能 分 析 研 究<br />

1<br />

项 贻 强<br />

1,2<br />

陈 建 伟<br />

1<br />

汪 劲 丰<br />

1,3<br />

杨 万 里<br />

1<br />

程 坤<br />

(1 浙 江 大 学 土 木 工 程 系 , 杭 州 310058;<br />

2 宁 波 市 城 建 设 计 研 究 院 , 宁 波 315012;<br />

3 交 通 运 输 部 科 学 研 究 院 , 北 京 100029)<br />

摘 要 : 为 最 大 限 度 地 减 少 跨 线 桥 梁 或 跨 运 河 航 道 桥 对 交 通 的 影 响 , 采 用 转 体 施 工 桥 梁 方 案 具 有<br />

较 大 的 优 越 性 。 本 文 对 一 座 跨 运 河 的 桥 梁 , 提 出 并 介 绍 了 采 用 转 体 施 工 的 中 承 式 梁 拱 组 合 体 系 方 案<br />

及 桥 型 布 置 特 点 , 并 对 其 静 动 力 特 性 进 行 了 分 析 和 试 验 研 究 , 结 果 表 明 采 用 转 体 施 工 的 中 承 式 梁 拱<br />

组 合 体 系 桥 具 有 施 工 期 间 交 通 干 扰 少 、 混 凝 土 浇 筑 方 便 、 造 价 低 和 结 构 刚 度 大 、 空 间 协 同 工 作 性 能<br />

好 等 优 点 , 并 为 静 动 力 试 验 研 究 所 证 实 。<br />

关 键 词 : 桥 梁 工 程 转 体 施 工 中 承 式 拱 梁 组 合 体 系 桥 静 动 力 性 能 空 间 分 析 试 验 研 究<br />

ANALYSIS OF MECHANICAL BEHAVIOR OF HALF-THROUGH ARCH BEAM<br />

COMBINATION BRIDGE WITH SWING CONSTRUCTION<br />

Y. Q. Xiang 1 , J. W. Chen 1,2 , J. F. Wang 1 , W. L. Yang 1,3 and K. Cheng 1<br />

1 Department of Civil Engineering, Zhejiang <strong>University</strong>, Hangzhou 310058,China<br />

2 Institute of Ningbo Urban and Building Design, Ningbo 315012,China<br />

3 China Academy of Transportation Science, Beijing 100029,China<br />

Abstract: In order to reduce traffic interfering under construction of crossing highway or busy canal, the bridge<br />

with swing construction has a lot of superiorities. Taking as background of a bridge crossing the canal, project<br />

and layout features of half-through combination bridge of arch beam with swing construction are presented.<br />

Static and dynamic behavior of the bridge is analyzed. <strong>The</strong> field test of the bridge is completed. <strong>The</strong> results<br />

show the half-through combination bridge of arch beam with swing construction has advantages of rare traffic<br />

interfering under construction, convenient casting of concrete, low cost, large stiffness of structure, good spatial<br />

harmony work behavior. <strong>The</strong>se behaviors have been approved by static and dynamic test.<br />

Keywords: Bridge engineering, swing construction, half-through combination bridge of arch beam, static and<br />

dynamic behavior, spatial analysis, experimental study.<br />

一 、 引 言<br />

在 通 航 频 繁 的 运 河 上 或 跨 线 桥 , 为 使 桥 梁 的 施 工 不 影 响 河 道 通 航 或 桥 下 通 车 , 且 尽 可 能 的 降 低<br />

桥 面 标 高 , 往 往 采 用 一 跨 过 河 , 同 时 在 桥 位 两 侧 有 场 地 的 条 件 下 , 有 时 会 优 先 考 虑 转 体 施 工 方 案 ,<br />

关 于 桥 梁 转 体 施 工 方 法 , 文 献 [1] 对 该 施 工 方 法 在 我 国 桥 梁 的 应 用 与 发 展 作 了 概 述 和 总 结 , 文 献 [2]<br />

针 对 高 速 公 路 , 为 最 大 限 度 减 少 对 高 速 公 路 正 常 交 通 的 影 响 , 对 某 些 跨 线 桥 梁 采 用 转 体 施 工 桥 梁 , 具<br />

作 者 简 介 : 项 贻 强 (1959-), 浙 江 杭 州 人 , 教 授 、 博 士 生 导 师 , 主 要 从 事 桥 梁 的 设 计 理 论 、 健 康 监 测 与 养 护 管 理 、 悬 浮 隧 道 、 古 桥 保 护 的<br />

研 究 。e-mail: Xiangyiq@zju.edu.cn<br />

-457-


有 较 大 的 优 越 性 , 从 现 有 已 经 发 表 的 文 章 来 看 , 主 要 是 针 对 如 何 实 施 转 体 施 工 工 艺 , 而 对 这 类 桥 梁<br />

的 静 动 力 特 性 分 析 和 试 验 研 究 的 论 述 较 为 少 见 。 本 文 结 合 一 座 跨 运 河 的 中 承 式 梁 拱 组 合 体 系 桥 梁 ,<br />

提 出 并 介 绍 了 其 采 用 转 体 施 工 的 方 案 及 桥 型 布 置 特 点 , 同 时 对 其 静 动 力 特 性 进 行 了 分 析 和 试 验 研<br />

究 , 并 给 出 相 关 的 结 论 。<br />

二 、 桥 梁 布 置 特 点<br />

某 桥 主 跨 采 用 三 跨 30+80+30m 的 中 承 式 梁 拱 组 合 体 系 桥 梁 , 边 孔 梁 段 设 平 衡 重 , 将 引 桥 边 梁 压<br />

在 主 桥 边 跨 梁 段 牛 腿 上 , 以 确 保 恒 载 作 用 下 副 墩 不 出 现 负 反 力 。 该 桥 设 计 荷 载 为 汽 —20 级 、 挂 —100。<br />

主 桥 承 重 构 件 为 拱 及 加 劲 梁 , 设 二 榀 分 列 , 桥 面 宽 度 为 1.0( 加 劲 梁 )+0.25( 防 撞 栏 杆 )+11.5( 行 车<br />

道 )+0.25( 防 撞 栏 杆 )+1.0( 加 劲 梁 ), 肋 中 距 为 12.8m。 梁 底 标 高 10.98m, 桥 中 心 标 高 12.28m, 拱<br />

顶 标 高 25.346m( 上 缘 )。 主 跨 拱 轴 线 采 用 二 次 抛 物 线 , 矢 跨 比 1/f=1/4。 拱 肋 尺 寸 b×h=100×180cm,<br />

桥 面 以 下 为 矩 形 , 其 余 为 工 字 形 。 纵 向 加 劲 梁 尺 寸 为 b×h=100×144cm 带 臂 0.25m 的 矩 形 梁 。 纵 向<br />

加 劲 梁 间 每 隔 6m 设 置 一 道 横 梁 , 横 梁 上 铺 设 桥 面 板 厚 为 30cm, 现 浇 接 头 成 连 续 板 , 横 梁 为 预 制 肋<br />

与 桥 面 板 的 现 浇 接 头 形 成 组 合 截 面 , 跨 中 高 度 h=130cm, 主 桥 桥 面 铺 装 为 厚 6~12cm 的 30 号 钢 纤<br />

维 混 凝 土 , 主 拱 肋 间 在 桥 面 上 设 6 道 上 风 撑 , 桥 面 以 下 设 下 K 型 风 撑 , 以 保 证 转 体 施 工 时 的 稳 定 性 。<br />

拱 与 加 劲 梁 在 桥 面 以 下 设 立 柱 , 桥 面 上 设 吊 杆 , 短 立 柱 与 主 拱 肋 下 端 做 成 假 铰 连 接 , 即 立 柱 下<br />

端 与 主 拱 只 有 柱 钢 筋 伸 入 连 接 , 混 凝 土 用 油 毛 毡 隔 离 层 隔 开 。 吊 杆 用 6Φ17 钢 丝 组 成 的 厂 制 成 品 索 ,<br />

冷 铸 锚 、 钢 丝 标 准 强 度 为 1670MPa, 并 在 拱 肋 处 设 张 拉 端 , 纵 梁 端 为 固 定 端 。 纵 向 加 劲 梁 采 用 符 合<br />

ASTM416-90a(270k) 标 准 的 7Φ15.24 的 高 强 度 低 松 驰 钢 铰 线 , 标 准 强 度 为 R b y =1860MPa, 设 计 张 拉<br />

控 制 应 力 采 用 σ k =0.75R b y =1395MPa。<br />

Ⅶ<br />

Ⅵ<br />

Ⅶ<br />

Ⅷ<br />

Ⅰ<br />

Ⅴ<br />

Ⅹ<br />

Ⅵ<br />

防 撞 栏 杆<br />

Ⅷ<br />

Ⅺ Ⅻ<br />

Ⅱ<br />

Ⅰ<br />

Ⅱ<br />

Ⅲ<br />

Ⅳ<br />

Ⅳ<br />

Ⅴ<br />

Ⅹ<br />

Ⅺ Ⅻ<br />

Ⅸ<br />

Ⅲ<br />

Ⅸ<br />

图 1 桥 梁 总 体 布 置 图 及 各 分 析 测 试 截 面 ( 单 位 :cm)<br />

三 、 转 体 施 工 方 案<br />

为 保 证 桥 梁 的 施 工 不 影 响 河 道 通 航 , 且 尽 可 能 的 降 低 桥 面 标 高 , 本 桥 设 计 提 出 采 用 转 体 施 工 方<br />

案 , 如 图 示 2。<br />

3.1. 转 体 施 工 特 点<br />

上 锚 箱<br />

2*9?15.24<br />

张 拉 2600KN<br />

2*9?15.24<br />

张 拉 2142KN<br />

9?15.24<br />

张 拉 950KN<br />

9?15.24<br />

张 拉 1220KN<br />

加 劲 梁<br />

上 支 承 面<br />

扎 花 锚 固 , 锚 固 长<br />

度 不 小 于 120cm<br />

图 2 桥 梁 的 转 体 施 工 方 案 ( 单 位 :cm)<br />

扎 花 锚 固 , 锚 固 长<br />

度 不 小 于 120cm<br />

图 3 空 间 有 限 元 网 格 模 型<br />

-458-


1) 所 需 的 机 具 设 备 少 , 工 艺 简 单 , 操 作 安 全 。 在 相 同 条 件 下 , 转 体 法 与 传 统 的 悬 吊 拼 装 或 悬<br />

浇 施 工 法 、 桁 架 伸 臂 法 、 搭 架 法 相 比 , 经 济 效 益 和 社 会 效 益 十 分 显 著 。<br />

2) 转 体 法 能 较 好 地 克 服 在 高 山 峡 谷 、 水 深 流 急 或 通 航 繁 忙 的 河 道 上 架 设 大 跨 度 构 造 物 的 困 难 ,<br />

尤 其 是 对 修 建 处 于 交 通 运 输 繁 忙 的 城 市 立 交 桥 和 铁 路 跨 线 桥 , 更 具 有 施 工 期 间 交 通 干 扰 少 , 混 凝 土<br />

浇 筑 方 便 等 优 点 。<br />

3.2. 施 工 顺 序<br />

跨 运 河 的 中 承 式 梁 拱 组 合 体 系 桥 梁 进 行 转 体 施 工 的 顺 序 为 :<br />

1) 主 桥 的 基 础 以 及 墩 台 的 施 工 ;<br />

2) 在 两 岸 边 搭 设 临 时 支 架 , 浇 筑 劲 性 骨 架 , 包 括 拱 肋 、 下 横 梁 、 下 风 撑 、 加 劲 梁 及 上 横 梁 连<br />

成 整 体 ;<br />

3) 在 主 墩 上 方 加 劲 梁 上 设 置 独 立 桅 杆 , 安 装 主 桥 岸 跨 桥 面 板 , 并 进 行 配 重 ( 配 重 加 在 端 部 三<br />

根 横 梁 上 );<br />

4) 安 装 主 桥 中 跨 吊 杆 , 张 拉 吊 杆 索 力 ;<br />

5) 双 向 张 拉 桅 杆 扣 索 索 力 , 使 拱 架 与 支 架 脱 离 。 拆 除 岸 边 支 架 ;<br />

6) 利 用 千 斤 顶 顶 推 上 盘 , 两 岸 同 时 平 衡 转 体 就 位 , 拱 肋 和 加 劲 梁 接 头 临 时 连 接 ;<br />

7) 调 整 中 跨 产 面 横 截 面 各 项 标 高 , 浇 筑 拱 肋 接 头 混 凝 土 ;<br />

8) 浇 筑 加 劲 梁 接 头 混 凝 土 , 张 拉 吊 杆 索 力 , 张 拉 加 劲 梁 部 分 预 应 力 筋 并 灌 浆 ;<br />

9) 拆 除 独 立 桅 杆 及 扣 索 , 卸 掉 边 跨 平 衡 重 , 张 拉 加 劲 梁 部 分 预 应 力 筋 并 灌 浆 ;<br />

10) 安 装 主 桥 中 跨 桥 面 板 , 张 拉 吊 杆 索 力 , 张 拉 加 劲 梁 部 分 预 应 力 筋 并 灌 浆 ;<br />

11) 完 成 桥 面 铺 装 及 防 撞 栏 杆 。<br />

四 、 结 构 受 力 性 能 分 析<br />

4.1. 静 力 性 能 的 分 析 计 算<br />

由 于 该 桥 结 构 复 杂 , 空 间 效 应 明 显 , 为 准 确 分 析 其 静 动 力 性 能 , 这 里 采 用 大 型 通 用 程 序<br />

ANSYS9.0 进 行 空 间 有 限 元 分 析 建 模 , 见 图 3。 该 模 型 有 空 间 节 点 6788 个 , 划 分 各 类 单 元 5826 个 。<br />

其 中 桥 面 板 用 SHELL181 单 元 模 拟 , 有 2898 个 单 元 ; 墩 台 用 SOLID45 单 元 模 拟 , 有 个 1760 单 元 ;<br />

吊 杆 用 LINK10 单 元 模 拟 , 有 20 个 单 元 ; 拱 肋 , 加 劲 梁 , 横 梁 , 立 柱 , 风 撑 以 及 两 边 跨 平 衡 重 都 用<br />

BEAM188 单 元 模 拟 , 有 1158 个 单 元 , 两 主 墩 设 铰 支 座 , 两 边 跨 各 设 五 个 竖 向 支 座 。 同 时 采 用 杠 杆<br />

法 考 虑 其 荷 载 的 横 向 分 布 将 空 间 两 片 拱 简 化 为 平 面 一 片 拱 肋 后 用 平 面 理 论 进 行 各 种 内 力 的 计 算 和<br />

分 析 , 以 便 于 空 间 分 析 结 果 比 较 , 具 体 模 型 见 图 4。<br />

4.2. 动 力 特 性 分 析<br />

图 4 平 面 有 限 元 模 型<br />

空 间 有 限 元 分 析 网 格 模 型 同 静 力 分 析 模 型 , 前 四 阶 振 型 计 算 结 果 见 图 5。<br />

-459-


a) 第 一 阶 拱 肋 正 对 称 竖 弯 (0.68) b) 第 二 阶 拱 梁 反 对 称 竖 弯 (1.45)<br />

c) 第 三 阶 拱 肋 反 对 称 横 扭 (1.74) d) 第 四 阶 拱 梁 正 对 称 竖 弯 (2.95)<br />

图 5 空 间 分 析 的 前 四 阶 振 型 图<br />

五 、 荷 载 试 验 研 究<br />

为 了 验 证 理 论 分 析 的 正 确 性 和 转 体 施 工 桥 梁 的 实 际 受 力 性 能 , 对 该 桥 进 行 了 静 动 载 试 验 研 究 。<br />

5.1. 静 载 试 验<br />

选 取 如 图 1 所 示 的 Ⅰ-Ⅰ~Ⅻ-Ⅻ 十 二 个 截 面 作 为 主 要 的 测 试 截 面 , 其 中 Ⅰ-Ⅰ~Ⅷ-Ⅷ 八 个 截 面 为 拱<br />

肋 的 主 要 测 试 截 面 , 在 各 截 面 的 上 下 缘 均 布 贴 应 变 片 ( 除 一 个 主 跨 四 分 点 截 面 ), 在 Ⅴ-Ⅴ、Ⅵ-Ⅵ、<br />

Ⅶ-Ⅶ、Ⅹ-Ⅹ、Ⅺ-Ⅺ、Ⅻ-Ⅻ 截 面 的 上 缘 装 置 “ 索 佳 ”2" 级 全 站 仪 十 字 丝 , 用 架 设 在 能 通 视 测 点 的<br />

桥 头 的 全 站 仪 观 测 。 在 Ⅳ-Ⅳ、Ⅲ-Ⅲ 截 面 的 内 侧 装 置 位 移 计 测 拱 脚 处 的 水 平 位 移 ; 选 取 Ⅰ-Ⅰ 截 面 右<br />

侧 3 米 处 的 横 梁 作 为 试 验 研 究 对 象 , 在 横 梁 的 中 跨 和 四 分 点 布 贴 应 变 片 、 装 置 百 分 表 分 别 测 读 各 截<br />

面 的 应 力 和 挠 度 。<br />

试 验 荷 载 系 根 据 设 计 荷 载 采 用 汽 车 在 桥 上 直 接 加 载 的 方 式 , 试 验 的 单 辆 车 总 重 300KN, 其 中 前<br />

轴 重 60KN, 中 后 轴 重 240KN, 前 中 轴 距 330cm, 中 后 轴 距 135cm。 为 便 于 与 空 间 理 论 方 法 相 比 较 ,<br />

表 1 给 出 了 按 实 际 加 载 载 位 排 列 和 平 面 理 论 计 算 的 试 验 荷 载 横 向 分 布 系 数 。<br />

表 1 试 验 荷 载 横 向 分 布 系 数 表<br />

布 载 方 式 1 列 偏 载 2 列 偏 载 3 列 偏 载<br />

荷 载 横 向 分 布 系 数 0.835 1.435 1.795<br />

5.2. 动 载 试 验<br />

在 拱 肋 测 试 截 面 的 相 关 部 位 安 装 若 干 高 灵 敏 度 加 速 度 传 感 器 , 由 智 能 动 态 测 试 分 析 系 统 测 试 其<br />

在 环 境 脉 动 荷 载 下 和 不 同 车 速 的 移 动 荷 载 下 , 拱 肋 跨 中 和 四 分 点 截 面 关 键 点 的 振 动 响 应 和 结 构 的 自<br />

振 频 率 。<br />

六 、 试 验 结 果 及 与 理 论 计 算 值 的 比 较<br />

-460-


6.1. 挠 度<br />

选 取 典 型 加 截 工 况 下 截 面 的 挠 度 实 测 值 与 挠 度 理 论 值 的 比 较 见 表 2,Ⅲ-Ⅲ 和 Ⅳ-Ⅳ 为 拱 脚 的<br />

水 平 位 移 , 其 中 东 拱 Ⅳ-Ⅳ 截 面 由 于 位 移 小 , 没 有 测 出 水 平 位 移 。 其 中 空 间 理 论 计 算 为 ANSYS 计<br />

算 结 果 。 其 中 各 截 面 挠 度 测 点 分 别 位 于 该 截 面 加 劲 梁 底 、 拱 肋 或 横 梁 截 面 的 中 部 。<br />

从 表 2 可 以 看 出 : 空 间 有 限 元 理 论 的 分 析 结 果 与 实 测 挠 度 值 吻 合 较 好 , 而 平 面 分 析 结 果 与 实 测<br />

挠 度 值 差 异 较 大 , 说 明 该 桥 的 空 间 受 力 特 征 明 显 , 如 果 用 平 面 分 析 结 果 去 评 定 测 试 结 果 , 则 可 能 出<br />

现 认 为 桥 梁 的 变 形 不 满 足 受 力 的 要 求 , 造 成 误 判 , 这 一 点 引 引 起 桥 梁 工 程 师 的 注 意 。<br />

加 载<br />

截 面<br />

表 2 典 型 工 况 最 大 一 级 荷 载 下 控 制 截 面 实 测 挠 度 与 理 论 值 的 比 较<br />

测 试 测 点 试 验 总 挠 空 间 理 论 计 算<br />

平 面 理 论 计 算<br />

位 置 截 面 度 St(mm) 挠 度 Ss1(mm) St/Ss1 挠 度 Ss2(mm) St/Ss2<br />

边 跨 跨 中 东 拱<br />

I-I<br />

( 偏 载 ) 西 拱<br />

中 跨 跨 中<br />

东 拱<br />

Ⅶ-Ⅶ<br />

( 偏 载 )<br />

西 拱<br />

I-I 2.66 3.98 0.67 9.33 0.29<br />

II-II 1.45 2.55 0.57 5.79 0.23<br />

I-I 2.06 3.06 0.67 4.98 0.41<br />

II-II 1.23 2.15 0.57 3.17 0.34<br />

Ⅳ-Ⅳ 0 0.14 — 0.01 —<br />

Ⅵ-Ⅵ 0.1 0.14 0.71 0.17 0.59<br />

Ⅶ-Ⅶ 3.7 4.25 0.87 3.6 1.03<br />

Ⅳ-Ⅳ 0.01 0.02 0.50 0.003 3.33<br />

Ⅵ-Ⅵ 0.09 0.14 0.64 0.07 1.29<br />

Ⅶ-Ⅶ 1.9 2.03 0.93 1.42 1.06<br />

6.2. 应 变<br />

典 型 加 载 工 况 截 面 的 应 力 实 测 值 及 与 应 力 理 论 值 的 比 较 见 表 3, 空 间 理 论 计 算 为 ANSYS 计 算<br />

结 果 。 测 点 编 号 为 :1 表 示 顶 缘 东 侧 ,2 表 示 顶 缘 西 侧 ,3 表 示 底 缘 东 侧 ,4 表 示 底 缘 西 侧 。 所 有 表<br />

中 拉 应 力 为 正 , 压 应 力 为 负 。<br />

表 3 典 型 工 况 最 大 一 级 荷 载 下 控 制 截 面 实 测 应 力 、 理 论 应 力 的 比 较<br />

加 截<br />

截 面<br />

测 试 截 面 及<br />

位 置<br />

测 点 位 置<br />

及 编 号<br />

实 测<br />

应 力<br />

空 间 理 论 计 算<br />

理 论 实 测<br />

应 力 理 论<br />

平 面 理 论 计 算<br />

理 论 实 测<br />

应 力 理 论<br />

加 劲 梁<br />

1 -0.35 -0.39 0.89 -2.91 0.12<br />

边 跨<br />

东<br />

顶 缘 2 - -0.27 - -2.91 -<br />

加 劲 梁 3 0.67 0.84 0.80 2.98 0.23<br />

跨 中<br />

Ⅰ—Ⅰ<br />

Ⅰ—Ⅰ<br />

底 缘 4 0.85 0.95 0.89 2.98 0.29<br />

加 劲 梁 1 - -0.53 - 1.78 -<br />

( 偏 载 )<br />

西<br />

顶 缘 2 -0.31 -0.43 0.74 1.78 -0.18<br />

加 劲 梁 3 0.68 0.95 0.72 -1.83 -0.38<br />

底 缘 4 0.95 1.05 0.90 -1.83 -0.52<br />

中 跨<br />

Ⅶ-Ⅶ<br />

拱 肋 顶<br />

1 -0.95 -1.94 0.49 -1.86 0.51<br />

跨 中<br />

Ⅶ-Ⅶ<br />

东<br />

缘 2 -0.95 -1.94 0.49 -1.86 0.51<br />

拱 肋 底 3 0.73 0.77 0.94 0.79 0.92<br />

( 偏 载 )<br />

缘 4 0.74 0.77 0.96 0.79 0.93<br />

-461-


西<br />

东<br />

西<br />

拱 肋 顶 1 -0.95 -0.95 0.99 -0.70 1.36<br />

缘 2 -0.94 -0.95 0.98 -0.70 1.35<br />

拱 肋 底 3 0.31 0.40 0.77 0.23 1.35<br />

缘 4 0.31 0.40 0.78 0.23 1.37<br />

加 劲 梁 1 -1.21 -1.42 0.86 -2.51 0.48<br />

顶 缘 2 -1.03 -1.26 0.82 -2.51 0.41<br />

加 劲 梁 3 1.03 2.05 0.50 2.93 0.35<br />

底 缘 4 - 2.21 - 2.93 -<br />

加 劲 梁 1 -0.29 -0.59 0.49 -1.21 0.24<br />

顶 缘 2 -0.32 -0.60 0.54 -1.21 0.27<br />

加 劲 梁 3 - 1.08 - 1.35 -<br />

底 缘 4 1.00 1.08 0.93 1.35 0.74<br />

6.3. 动 力 性 能<br />

环 境 脉 动 下 的 实 测 时 间 波 形 和 频 谱 分 析 图 见 图 7, 典 型 跑 车 荷 载 下 的 实 测 时 间 波 形 和 频 谱 分 析<br />

图 分 别 见 图 8。 试 验 孔 跨 前 四 阶 固 有 频 率 的 理 论 值 和 实 测 值 的 比 较 见 表 4, 通 过 对 车 辆 在 5、10、20、<br />

30、40km/h 跑 车 试 验 及 跳 车 试 验 , 测 试 其 频 率 均 为 0.79 Hz。 通 过 现 场 实 测 , 结 果 表 明 理 论 与 实 测<br />

的 前 几 阶 固 有 频 率 吻 合 很 好 。<br />

表 4 前 四 阶 固 有 频 率 的 理 论 值 和 实 测 值 的 比 较<br />

频 率 (Hz) 第 一 阶 第 二 阶 第 三 阶 第 四 阶<br />

理 论 值 0.68 1.45 1.74 2.95<br />

实 测 值 0.79 1.37 1.77 2.94<br />

a) 时 间 波 形 图 b) 频 谱 图<br />

图 7 环 境 脉 动 下 实 测 时 间 波 形 及 频 谱 分 析 图<br />

a) 时 间 波 形 图 b) 频 谱 图<br />

图 8 30km/h 跑 车 实 测 时 间 波 形 及 频 谱 分 析 图<br />

-462-


七 、 结 论<br />

通 过 上 述 研 究 , 可 以 得 出 如 下 结 论 :<br />

采 用 转 体 施 工 的 中 承 式 梁 拱 组 合 体 系 桥 具 有 施 工 期 间 交 通 干 扰 少 , 混 凝 土 浇 筑 方 便 、 造 价 低 和<br />

结 构 刚 度 大 、 空 间 协 调 工 作 性 能 好 等 优 点 , 并 为 静 动 力 试 验 研 究 所 证 实 。<br />

与 平 面 分 析 结 果 相 比 较 , 空 间 有 限 元 理 论 的 分 析 结 果 与 实 测 挠 度 值 吻 合 较 好 , 说 明 该 桥 的 受 力<br />

空 间 特 征 明 显 。 实 测 挠 度 相 对 桥 跨 比 为 L/2162, 远 小 于 《 桥 规 》 规 定 的 L/800。 拱 肋 、 加 劲 梁 及 横<br />

梁 各 测 试 截 面 混 凝 土 测 点 的 实 测 应 力 基 本 小 于 理 论 计 算 应 力 , 关 键 应 力 测 试 断 面 实 测 应 力 值 与 空 间<br />

理 论 计 算 值 基 本 都 很 接 近 。<br />

由 动 载 试 验 分 析 得 到 的 前 几 阶 固 有 频 率 与 理 论 分 析 结 果 吻 合 很 好 , 且 实 测 频 率 大 于 理 论 计 算<br />

值 , 表 明 该 桥 梁 的 实 际 刚 度 较 设 计 预 期 的 略 大 , 桥 梁 的 刚 度 达 到 设 计 和 规 范 的 变 形 要 求 , 且 有 一 定 的<br />

安 全 储 备 。<br />

参 考 文 献<br />

[1] 陈 宝 春 , 孙 潮 . 桥 梁 转 体 施 工 方 法 在 我 国 的 应 用 与 发 展 . 公 路 交 通 科 技 . 2001, 18(2):24-28<br />

[2] 刘 志 宏 . 跨 越 高 速 公 路 转 体 桥 梁 设 计 与 施 工 控 制 . 北 方 交 通 . 2006, 12:39-42<br />

[3] 交 通 部 颁 . 公 路 桥 涵 设 计 规 范 . 北 京 : 人 民 交 通 出 版 社 .1989 年<br />

[4] Xiang Y. Q. et al. Study on Static and Dynamic Behavior of T-shape Rigid Frame Bridge Reinforced by External Tendons.<br />

INTERNATIONAL CONFERENCE ICACS 2003, 2003,9:17-19<br />

[5] 项 贻 强 , 杨 万 里 , 范 永 根 等 . 拱 索 体 系 加 固 刚 架 拱 桥 的 荷 载 横 向 分 布 . 中 国 公 路 学 报 , 2007,20(4):91-95<br />

-463-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

FINITE ELEMENT STUDY OF INELASTIC LOCAL WEB BUCKLING CAPACITY<br />

OF COPED STEEL I-BEAM<br />

Y. Qin 1 , C. C. Lam 2 , V. P. Iu 3 , and K. P. Kou 4<br />

1,2,3,4 Department of Civil and Environmental Engineering,<br />

<strong>University</strong> of Macau, Macau SAR, China.<br />

2 Email: fstccl@umac.mo<br />

ABSTRACT<br />

In steel construction, beams often have to be coped at the flange to provide clearance for the framing beams or<br />

to maintain the main beam and the secondary beam at the same level. When the beams are coped, the local web<br />

buckling capacity of the beams at the coped region may be reduced. Depending on the cope details, the local<br />

web buckling capacity can be classified as elastic and inelastic web buckling. For coped beam with relatively<br />

short coped length and depth, part or the whole section of the web may yield before local web buckling occurs.<br />

In this paper, non-linear finite element analyze, which included both material and geometric nonlinearities, of<br />

the inelastic local web buckling capacity of coped steel I-beam is presented. <strong>The</strong> effects of different parameters,<br />

such as (1) web slenderness (d/t w ), (2) cope depth to beam depth ratio (d c /D), (3) cope length to reduced web<br />

depth ratio (c/h o ) and (4) initial imperfection of web section, to the web buckling capacity of coped steel I-beam<br />

were studied. <strong>The</strong> current finite element results showed that local web buckling capacity decreased with<br />

increased initial imperfection. Meanwhile, for beams with the smallest beam depth to web thickness ratio, both<br />

shear yielding capacity and elastic web buckling capacity gave conservative prediction when c/h o is larger than<br />

0.5. On the other hand, for beams with largest beam depth to web thickness ratio, the predicted elastic web<br />

buckling capacity was closed to the finite element results for c/h o larger than 0.75. However, the prediction of<br />

both shear yielding capacity and elastic web buckling capacity gave conservative prediction when c/h o is smaller<br />

than 0.75 for most of the cases in this study.<br />

KEYWORDS<br />

Cope beam; finite element method; inelastic analysis; local web buckling; steel structures.<br />

INTRODUCTION<br />

In order to make the beam flanges at the same elevation, in practical steel construction, beam flanges often have to<br />

be cut away from the major part to provide enough clearance for supports. <strong>The</strong> cope can be at the top (as shown in<br />

Figure 1), the bottom, or both flanges in combination. As a result, due to the coped flanges, the corresponding<br />

lateral-torsional buckling capacity and the local web buckling (LWB) capacity of the beam may be reduced. For<br />

the beams with lateral support, block shear failure of the connection at beam end and local web buckling were two<br />

potential failure modes. For thin webs, failure could occur by local web buckling at the coped region. <strong>The</strong> global<br />

lateral-torsional buckling behavior and the elastic local web buckling strength of coped steel-I beam were studied<br />

both experimentally and analytically by Cheng et al. (1984). Further studies of the design and behavior of coped I<br />

beams were also conducted by some researchers (Cheng and Yura 1986; Lam et al. 2000; Yam et al. 2000; Johan<br />

Maljaars et al. 2002; Yam et al. 2007). Based on the experimental and numerical investigations of the local web<br />

buckling strength of coped section, Yam et al. (2003) proposed a modified plate buckling equation that considered<br />

the shear buckling phenomenon for predicting the elastic local web buckling capacity of coped steel I-beam. It is<br />

proposed that the critical elastic web buckling reaction (R cr ) can be predicted by Eqs. (1) to (3).<br />

k<br />

s<br />

b<br />

0<br />

,<br />

⎛ h ⎞<br />

= a⎜<br />

⎟<br />

⎝ c ⎠<br />

R<br />

cr<br />

=τ t D − d )<br />

(1)<br />

cr<br />

w ( c<br />

2<br />

2<br />

π E ⎛ tw<br />

⎞<br />

τ = k<br />

2<br />

12( 1 )<br />

⎜<br />

⎟<br />

(2)<br />

cr s<br />

−υ<br />

⎝ h0<br />

⎠<br />

2<br />

d<br />

c<br />

a = 1.38<br />

−1.79<br />

,<br />

⎛ dc ⎞ ⎛ dc<br />

⎞<br />

b = 3.64⎜<br />

⎟ − 3.36⎜<br />

⎟ + 1. 55<br />

(3)<br />

D ⎝ D ⎠ ⎝ D ⎠<br />

-464-


where c = cope length,D = beam depth, d c<br />

= cope depth, R cr<br />

= critical reaction, τ cr<br />

= critical shear stress, t w<br />

=<br />

web thickness, k s<br />

= shear buckling coefficient, h 0<br />

= height of web of T section.<br />

<strong>The</strong> above equations are suitable for coped beams with relatively long coped length and coped depth since<br />

elastic web buckling may occur. However, for coped beams with relatively short coped length and depth, local<br />

web buckling may occur inelastically. <strong>The</strong>refore, the above elastic web buckling equations might give<br />

non-conservative prediction for coped beams with inelastic web buckling failure.<br />

Figure 1 Top flange coped I beam<br />

In 2004, ten full-scale coped beam tests were conducted by Zhong et al. to investigate the potential local failure<br />

mode of coped steel I-beam with relatively short coped length and depth and it was shown that inelastic local<br />

web buckling is one of the potential failure modes. In order to further study the local web buckling capacity of<br />

coped steel I-beam with relatively short coped length and depth, finite element analysis was carried out to<br />

analyze those test specimens with local web buckling failure. Subsequently, by using the validated finite element<br />

model, parametric studies which included (1) web slenderness (d/t w ), (2) cope depth to beam depth ratio (d c /D), (3)<br />

cope length to reduced web depth ratio (c/h o ) and (4) initial imperfection of web section, were conducted to obtain<br />

the local web buckling capacity of coped steel I-beams.<br />

FINITE ELEMENT MODEL FOR COPED BEAMS<br />

General<br />

Finite element method is employed to present the numerical analyses of local web buckling strength of coped<br />

steel I-beams. All finite element (FE) models were modeled and analyzed using the finite element software<br />

ABAQUS 6.7 (2007). <strong>The</strong> validity of the FE models were examined by comparing the numerical results of load<br />

deflection curve of the models with those results obtained from the tests of Zhong et al. (2004). Detail test<br />

procedures and results can be obtained from Zhong et al. (2004). In this paper, five finite element models were<br />

established according to the dimension and materials details of the test specimens (A1, A2, B1, B2, and D1) of<br />

Zhong et al. (2004). Detail dimensions of the specimens are shown in Table 1.<br />

Table 1 <strong>The</strong> details of dimension of the test specimens of Zhong et al. (2004)<br />

Test Web Flange Beam Flange Connected Connection Cope Cope Weld<br />

No. Thickness Thickness Depth Width Length Position Length Depth Size<br />

t w (mm) T (mm) D (mm) B (mm) a (mm) b (mm) p (mm) c (mm) d c (mm) s (mm)<br />

A1 6.8 11.1 404.2 140.9 50 160 20 100 33 8.9<br />

A2 6.8 11.1 404.2 140.9 70 140 20 120 31 9.4<br />

B1 6.8 11.1 404.2 140.9 50 120 20 100 30 10<br />

B2 6.8 11.1 404.2 140.9 90 110 20 130 30 9.9<br />

D1 9.2 14.2 456.3 189.1 90 120 20 150 30 11.8<br />

c<br />

b'<br />

p<br />

b<br />

dc<br />

D<br />

tw<br />

s<br />

s<br />

a<br />

B<br />

T<br />

-465-


Boundary Simulation<br />

Three-dimensional shell elements S4R, were selected to simulate the whole beam, including beam web, flanges,<br />

and stiffeners and the clip angle connection was modeled by using brick element (C3D8R in ABAQUS). As<br />

lateral bracings were provided in the test, lateral displacement (U 3 =0) was prevented in the finite element<br />

models at the positions of bracing as shown in Figure 2. Suitable boundary conditions were provided to the<br />

finite element models to simulate the simply supported condition and point load was applied to simulate the<br />

loading condition. Bilinear spring elements were used to simulate the bolts between the connection of beam and<br />

the flange of supporting column. <strong>The</strong> tensile and compressive stiffness of the spring was set as 200 N/mm at<br />

first 0.5 mm in order to account for the slip behaviour. Afterwards, the stiffness was changed to a relative large<br />

value of 40000 N/mm per spring. Those spring properties were obtained by comparing the load versus<br />

deflection curve of the numerical results and test results. In addition, at the edge of bolt hole, freedom of the<br />

outer surface nodes was constrained to be zero on vertical and out of plane direction (U 2 =0,U 3 =0). To simulate<br />

the welding near the clip angles and beam web, multi-point constraints (MPCs) were introduced. It allows<br />

constraints to be imposed between different degrees of freedom of the model. MPC type BEAM provides a rigid<br />

beam between two nodes to constrain the displacement and rotation at the first node to the displacement and<br />

rotation at the second node. Two series of MPC BEAM were used in the models to link the corresponding nodes.<br />

For the nodes associated as Group 1, it is needed to constrain all nodes at the clip angle edge to beam web<br />

directly at the weld root. In order to account for the weld size, the nodes associated as Group 2 paralleled to the<br />

angle edge was linked from a distance of angle thickness to the beam web at the weld leg to be consistent with<br />

the measured weld sizes. Typical finite element mesh and boundary conditions were shown in Figure 2.<br />

Nodes directly connected<br />

to beam web (Group 1)<br />

Using MPC element to<br />

connect to the beam web<br />

(Group 2)<br />

Material Property<br />

Figure 2 Typical finite element mesh and boundary conditions<br />

<strong>The</strong> isotropic elastic-plastic material properties with the von Mises yield criterion were used for the FE analyses<br />

to account for the effect of the material nonlinearities. As the input values of ABAQUS, the stress and strain are<br />

p<br />

required as true stress ( σ ) and true plastic stain (<br />

true<br />

ε ). <strong>The</strong>refore, the nominal stress (<br />

true<br />

σ ) and nominal<br />

nom<br />

strain ( ε ) getting from the tension coupons tests of Zhong et al. (2004), were converted to true stress (<br />

nom<br />

σ )<br />

true<br />

p<br />

and true plastic stain ( ε<br />

true<br />

), using Eqs. (4) and (5). It is emphasized that, the nominal stress ( σ ) and nominal<br />

nom<br />

strain ( ε nom<br />

) were taken from the static stress-strain curves from the coupon tests.<br />

σ<br />

true<br />

= σ<br />

nom( 1+<br />

ε<br />

nom)<br />

(4)<br />

p<br />

⎛ σ<br />

true ⎞<br />

ε<br />

true<br />

= ln( 1+<br />

ε<br />

nom<br />

) − ⎜ ⎟<br />

(5)<br />

⎝ E ⎠<br />

-466-


A post-ultimate stiffness of 200 MPa was assumed, and the material true plastic strain was extrapolated linearly<br />

to 1.0. Typical true stress versus true plastic strain curve is shown in Figure 3.<br />

Figure 3 Typical true stress vs. true strain curve<br />

Test Specimens Finite Element Analysis Procedure, Results and Discussion<br />

<strong>The</strong> numerical analysis included two steps in this study. <strong>The</strong> first step was the eigenvalue buckling analysis,<br />

which is generally used to obtain the elastic critical load and the corresponding buckling shape. <strong>The</strong> second<br />

analysis step was the non-linear analysis which considered initial imperfection of the specimens based on the<br />

first buckling shape obtained from the elastic analysis. <strong>The</strong> maximum initial lateral imperfection of each models<br />

were obtained by comparing the finite element results of load versus vertical deflection curves to the test results.<br />

It is found that by assigning the maximum initial lateral imperfection of 0.7, 1.7, 4.0, 3.0, and 3.0 mm,<br />

respectively for specimens A1, A2, B1, B2, and D1, the load versus vertical deflection curves obtained from the<br />

finite element analysis compared well with the test results. <strong>The</strong> comparison of FE results and test results of the<br />

load versus vertical deflection curves of two typical specimens (A2 and D1) are shown in Figure 4 and the<br />

comparison of the ultimate load obtained from FEA and test are shown in Table 2. It is shown that the mean test<br />

beam capacity to predicted ratio is 1.016 and the mean reaction to predicted ratio is 0.988. Meanwhile, the beam<br />

end reaction verses lateral deflection of cope region obtained from FE analysis was shown in Figure 5 for<br />

specimens A2 and D1. It is observed that significant lateral deflection occurred when the ultimate load was<br />

reached.<br />

<strong>The</strong> von Mises stress contours and the equivalent plastic strain contours of model A2 at maximum load level are<br />

shown in Figure 6(a) and (b). As it is shown in those figures, significant yielding and plastic deformation is<br />

observed at coped region and the region underneath the clip angles. Typical deformed shapes of test specimen<br />

A2 is shown in Figure 6(c) in which local web buckling failure is observed at the coped region.<br />

As it is shown that the current FE model and analysis procedures gave reasonable results when compared with<br />

the test results, parametric studies were carried out by using the validated FE model and the results are discussed<br />

in the following sections.<br />

Test No.<br />

Test Ultimate<br />

Load P test<br />

(kN)<br />

Table 2 Comparison of FEA results with test results<br />

FEA<br />

Ultimate FEA Ultimate<br />

Load Reaction R FEA<br />

P FEA (kN)<br />

(kN)<br />

Test Ultimate<br />

Reaction R test<br />

(kN)<br />

P test /P FEA<br />

Ratio<br />

R test /R FEA<br />

Ratio<br />

Maximum<br />

Imperfection<br />

value<br />

(mm)<br />

A1 495.5 395.2 505.0 412.0 0.98 0.96 0.7<br />

A2 550.1 437.4 537.0 443.6 1.02 0.99 1.7<br />

B1 500.3 394.0 492.0 405.1 1.02 0.97 4.0<br />

B2 490.2 390.3 481.0 396.1 1.02 0.99 3.0<br />

D1 811.0 623.0 781.0 605.0 1.04 1.03 3.0<br />

-467-


A pply Load, P (kN)<br />

900<br />

800<br />

700<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

FE results D1<br />

Test results D1<br />

0<br />

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1<br />

Deflection, δ (mm)<br />

Test results A2<br />

Figure 4 Load vs. deflection curves<br />

FE results A2<br />

A2 Test<br />

imperfection=1.7<br />

D1 Test<br />

imperfection=3<br />

Reaction, R (kN)<br />

700<br />

600<br />

500<br />

400<br />

300<br />

200<br />

100<br />

0<br />

R<br />

δ (Lateral)<br />

D1<br />

0 1 2 3 4 5 6 7 8 9 1<br />

Lateral deformation, δ (mm)<br />

Figure 5 Reaction vs. Lateral deformation<br />

A2<br />

Local<br />

buckling<br />

web<br />

(a) Stress contours (b) Equivalent plastic strain contours (c) Deformed shapes of the connection<br />

Figure 6 Typical character of the connection<br />

PARAMETRIC STUDY AND RESULTS<br />

In order to study the effect of different parameters to the capacity of the web, parametric study was conducted<br />

based on the validated FE models discussed in previous section. <strong>The</strong> parameters chosen in this study were (1)<br />

initial imperfection of web section, (2) ratio of coped length to reduced beam depth (c/h 0 ), (3) the ratio of cope<br />

depth to beam depth (d c /D) and (4) web slenderness (d/t w ). For a beam with large web slenderness, the critical<br />

elastic buckling load may dominate the resistance, while the plastic capacity may dominate the resistance for<br />

beams with small web slenderness. For the purpose of investigating the inelastic local web buckling and to<br />

ensure the results to be within the application in practice, the web slenderness was chosen from 27.4 to 57.1.<br />

<strong>The</strong> proposed web slenderness covered a wide range of sections according to the British Standard (2000). In this<br />

study, four universal beam sections (UB203×133×30, UB610×305×179, UB356×171×51, and UB406×140×39)<br />

according to the British Standard (2001) were chosen. <strong>The</strong> span length of simulated beams was assumed to be<br />

ten times of the beam depth. <strong>The</strong> load position from the coped end of beam was approximately 1.5 times the<br />

beam depth.<br />

<strong>The</strong> coped depth to beam depth ratio (d c /D) was set as 0.05, 0.1 and 0.15 for all models. Based on a given beam<br />

depth and d c /D ratio, the corresponding reduced beam depth at coped (h 0 ) for each beam sections can be<br />

obtained. <strong>The</strong> corresponding coped length (c) for each beam sections was selected in order to achieve the coped<br />

length to reduced beam depth ratio (c/h 0 ) to vary from 0.1 to 1.0. Based on the above parameters requirement,<br />

the corresponding coped end geometry of each beam sections was obtained. All the variables of parametric<br />

study are presented in Table 3.<br />

-468-


Table 3 Summary of the variables of parametric study<br />

Parameters UB203×133×30 UB610×305×179 UB356×171×51 UB406×140×39<br />

D (mm) 206.8 617.5 355.6 397.3<br />

T (mm) 9.6 23.6 11.5 8.6<br />

B (mm) 133.8 307 171.5 141.8<br />

depth between fillets, d (mm) 172.4 537.3 312.2 359.7<br />

web thickness, t w (mm) 6.3 14.1 7.3 6.3<br />

d/t w 27.4 38.1 42.8 57.1<br />

d c /D 0.05, 0.1, 0.15<br />

cope depth, d c (mm) 17 21 31 31 62 93 18 36 53 20 40 60<br />

h 0 =D-d c (mm) 190 186 176 587 556 525 338 320 302 377 358 338<br />

c/h 0 0.1, 0.25, 0.5, 0.75, 1.0<br />

19 19 18 59 56 52 34 32 30 38 36 34<br />

48 47 44 147 139 131 84 80 76 94 89 84<br />

cope length, c (mm) 95 93 88 293 278 262 169 160 151 189 179 169<br />

143 140 132 440 417 394 253 240 227 283 268 253<br />

190 186 176 587 556 525 338 320 302 377 358 338<br />

In the parametric study, boundary conditions and material properties were the same as mentioned in the previous<br />

finite element analysis. <strong>The</strong> weld contact of end plate and beam web was simulated by rigidly MPC beam<br />

constrains (ABAQUS 6.7, 2007). For the FE models used in the parametric study, simply welded end plate was<br />

used for the connection between the beam end and the supported column. With this modification, the effect of<br />

clip angle to the buckling capacity of the web was eliminated. <strong>The</strong> different schematic connections of the FE<br />

models for test specimens and the model for parametric study were illustrated in Figure 7.<br />

Effects of Imperfection<br />

Figure 7 Schematic of the connection of the finite element models<br />

For the FEA of the test specimens, it is shown that the FE results compared well with the test results with the<br />

maximum web imperfection value varied from 0.7mm to 4.0mm. In order to investigate the effect of maximum<br />

web imperfection to the capacity of coped I-beam, a series of analysis were conducted for UB203 models with a<br />

fixed values of c/h 0 =0.5, d c /D=0.15 and d/t w =27.4. Initial maximum web imperfections values were selected to<br />

be 0.7, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, and 4.0 mm and the corresponding maximum capacity of the coped beam was<br />

obtained from the FEA. <strong>The</strong> load versus deflection curve of this series of models was shown in Figure 8(a). It<br />

was found that, the effect of initial imperfection to the load versus vertical deflection behavior in the early stage<br />

is very limited; however, it does affect the ultimate capacity of coped beams. <strong>The</strong> ultimate capacities versus<br />

initial maximum web imperfection values were shown in Figure 8(b). In general, as the imperfection increases,<br />

the ultimate capacity of the coped beam deceases. It is found that the ultimate capacity decreases significantly<br />

by 12.35% with the maximum web imperfection value increases from 0.7mm to 2.0mm. However, when the<br />

maximum web imperfection value increased from 2.5mm to 4.0mm, the decrease of ultimate capacity becomes<br />

less and the percentage reduction is about 4.65%. Since the decrease of ultimate capacity is not significant when<br />

the maximum web imperfection value is larger than 2.5mm, this maximum web imperfection value was used in<br />

the following FE analysis to obtain the effect of other parameters to the ultimate capacity of coped beam.<br />

-469-


(a)<br />

(b)<br />

Figure 8 Load vs. deflection curves with varied imperfection<br />

(UB203, c/h 0 =0.5, d c /D=0.15, d/t w =27.4 for all models)<br />

Effects of Ratio of Cope Length to Reduced Beam Depth (c/h 0 )<br />

In order to investigate the effects of the ratio of cope length to reduced beam depth (c/h 0 ) on local buckling<br />

capacity of coped beams, models with c/h 0 varied from 0.1 to 1.0 were analyzed. <strong>The</strong> proposed c/h 0 is achieved<br />

by designing suitable coped length and coped depth dimension for each proposed beam sections. Details of the<br />

dimension of coped length and reduced coped depth for each beam sections are shown in Table 3. <strong>The</strong>se three<br />

series of analysis were developed on UB356 with 2.5mm initial maximum web imperfection and slenderness<br />

ratio equals to 42.8 and the results were shown in Figure 9. In general, the ultimate reaction of coped beams was<br />

decreased by increasing the ratio of cope length to reduced beam depth for all FE models. It is found that the<br />

ultimate reaction of coped beam was significantly increased when the ratio of cope length to reduced beam<br />

depth was decreased in those models. In other words, the instability of coped beams would increase by<br />

increasing the ratio of cope length to reduced beam depth.<br />

Additionally, it can be found from Figure 9 that the slope of curves (the curves about the effects of the ratio of<br />

cope length to reduced beam depth) was not linear. <strong>The</strong> capacity decreases more significantly when c/h 0 is less<br />

than 0.5. When c/h 0 varied from 0.1 to 0.5, the ultimate reaction of cope beam decreased from 484kN to 359kN,<br />

(25.8% decreased) for the series of beams with d c /D=0.05; and the percentage decrease of the other two series of<br />

beams with d c /D=0.1 and d c /D=0.15 were 29.1% and 29%, respectively. When c/h 0 varied from 0.5 to 1.0, the<br />

ultimate capacity dropped from 271kN to 212kN (21.8% decreased) for the series of beams with d c /D=0.15.<br />

Effects of the Ratio of Cope Depth to Beam Depth (d c /D)<br />

<strong>The</strong> effect of cope depth to beam depth ratio to the capacity of coped end was investigated. Three different ratio<br />

of cope depth to beam depth (d c /D = 0.05, 0.1 and 0.15) was designed for each beam sections. Details of the<br />

dimension of coped depth for each beam sections are shown in Table 3. <strong>The</strong> analysis for this parameter was<br />

carried out on five different cope length to reduced beam depth ratios (c/h 0 ) and the ratios were designed as 0.1,<br />

0.25, 0.5, 0.75 and 1.0. It is observed from Figure 10 that, when the ratio of cope depth to beam depth increased,<br />

the ultimate reaction decreased, however, the variation was not significant. <strong>The</strong> ultimate reaction decreased by<br />

1%, 11.1%, 15%, 10.7%, and 6.6% for the five specified c/h 0 value, with a mean value equals to 8.9%.<br />

Figure 9 Effects of the ratio of cope length to reduced beam depth<br />

Figure 10 Effects of the ratio of cope depth to beam depth<br />

-470-


Effects of Web Slenderness (d/t w )<br />

All the results of the ultimate reaction obtained from the FE analysis were normalized by the shear yielding<br />

capacity and the results were shown in Figure 11. <strong>The</strong> corresponding shear yielding reaction is predicted by the<br />

following equation.<br />

3<br />

R<br />

vy<br />

= × σ<br />

y<br />

× h0<br />

× t<br />

w<br />

3<br />

where<br />

R = shear yield reaction,<br />

vy<br />

σ = yield stress,<br />

y<br />

h = reduced beam depth,<br />

0<br />

w<br />

(5)<br />

t = web thickness. Meanwhile,<br />

the elastic web buckling capacity of the coped beam according to the prediction from Eqs. (1) to (3) was shown<br />

in Figure 11 as well for comparison.<br />

It can be seen from the figures that the elastic web buckling capacity is more likely to occur when the web<br />

slenderness (d/t w ) is large. It was found that, for d/t w =57.1, when c/h 0 >1.0, the reaction of model UB406 was<br />

larger than the critical reaction of coped beam. With a fixed value of c/h 0 (i.e.c/h 0 =1.0), the ratio of the ultimate<br />

reaction of FE model to the shear yielding reaction was 0.42, for d/t w =42.8 and d c /D=0.1. Meanwhile the ratio of<br />

the critical reaction to the shear yielding reaction was 0.53, which is 20.7% higher than the former one. For the<br />

case of d/t w =38.1 and d c /D=0.1, the ratio of the ultimate reaction of FE model to the shear yielding reaction was<br />

0.54, and the ratio of the critical reaction to the shear yielding reaction was 0.68, which is 21% higher than the<br />

former one. For series of d/t w =27.4, when the ratio of cope length to reduced section depth was 1.0, the ultimate<br />

reaction obtained from FEA was less than the shear yielding reaction significantly. <strong>The</strong>refore, it is shown that<br />

both shear yielding capacity and elastic web buckling capacity gave non-conservative prediction for most cases.<br />

As a result, prediction of the inelastic web buckling capacity of coped steel I-beam is necessary in order to<br />

provide conservative prediction.<br />

(a) Comparison of ELWB with FEA results of UB203<br />

series<br />

(b) Comparison of ELWB with FEA results of UB610<br />

series<br />

(c) Comparison of ELWB with FEA results of UB356 (d) Comparison of ELWB with FEA results of UB406<br />

series<br />

series<br />

Figure 11 Comparison of ELWB with FEA results of all series<br />

-471-


A summary of all FE results of parametric study is shown in Figure 12 in which the ultimate reaction (R u ) from<br />

the FEA was divided by the shear area ( As = tw<br />

× h ) and the average shear stress (<br />

0<br />

F<br />

ave<br />

= Ru<br />

/ A ) was<br />

s<br />

normalized by the shear yielding stress ( F = σ / 1/ 3 ; where σ y is the tensile yield stress). <strong>The</strong> parameters<br />

y<br />

y<br />

which are the ratio of coped length to reduced beam depth (c/h 0 ), ratio of cope depth to beam depth (d c /D) and<br />

web slenderness (d/t w ) are shown in the figure as well. <strong>The</strong> results showed that fully shear yielding of web did<br />

not occur for most cases except for those sections with small d/t w and c/h o values. Meanwhile, for section with<br />

the largest d/t w value, the ultimate capacity of coped region is always less than the fully shear yielding capacity<br />

of web. <strong>The</strong>refore, the prediction of fully shear yielding capacity of web might not give conservative results for<br />

coped steel I-beam when inelastic buckling of web occurred.<br />

CONCLUSIONS<br />

Figure 12 FE results of normalized shear stress of coped section<br />

More than 60 nonlinear FE models of coped beam were analyzed in this paper to study the inelastic local web<br />

buckling capacity of coped steel I-beam. Both material and geometric nonlinearities were included in the finite<br />

element analyzes with suitable boundary conditions assigned to the models. <strong>The</strong> results obtained from the FE<br />

analysis were calibrated and compared to the test results from Zhong et al. (2004). <strong>The</strong>n, parametric studies<br />

were conducted based on the validated FE models. <strong>The</strong> parameters considered in this study included (1)<br />

geometry initial imperfection of web, (2) ratio of coped length to reduced beam depth (c/h 0 ), (3) ratio of cope<br />

depth to beam depth (d c /D), and (4) web slenderness (d/t w ).<br />

<strong>The</strong> FE results showed that the local web buckling capacity decreased when initial imperfection increased, as<br />

well as the ratio of coped length to reduced beam depth increased. Increase of the ratio of cope depth to beam<br />

depth also lead to the decrease of cope I-beam capacity, however, the reduction is not significant. <strong>The</strong> nonlinear<br />

finite element results were compared to the shear yielding capacity and the elastic local web buckling capacity<br />

and it is shown that both shear yielding capacity and elastic local web buckling capacity gave non-conservative<br />

prediction for most cases. As a result, prediction of the inelastic web buckling capacity of coped steel I-beam is<br />

necessary in order to provide conservative predictions. As the FE models in this paper provided a basis for the<br />

prediction of local web buckling capacity of coped steel I-beams, further work is going on which includes the<br />

preparation of analytical solutions for predicting the inelastic local web buckling strength of coped steel I<br />

beams.<br />

ACKNOWLEDGMENTS<br />

<strong>The</strong> authors would like to extend their gratitude to the Research Committee of the <strong>University</strong> of Macau in<br />

providing financial support to this project. <strong>The</strong> assistance of the technical staff, Mr. Tou Ka Man, of the Structures<br />

and Construction Materials Laboratory of <strong>University</strong> of Macau is acknowledged.<br />

-472-


REFERENCES<br />

ABAQUS/Standard User’s Manual Version 6.3, (2002), Hibbitt, Karlsson & Sorensen,Inc.<br />

Cheng, J.J.R., and Yura, J.A. (1986). “Local web buckling of coped beams”, Journal of Structural Engineering,<br />

ASCE, 112(10), 2314-2331.<br />

Cheng, J.J.R., Yura, J.A., and Johnson, C.P. (1984). “Design and behavior of coped beams”, Phil M. Ferguson<br />

Structural Engineering Laboratory Rep. No. 84-1, Univ. of Texas at Austin, Austine, Tex.<br />

Johan Maljaars, Jan W.B. Stark, Henri M.G.M. Steenbergen (2002). “Numerical simulation of lateral-torsional<br />

buckling of coped girders”, 15th ASCE Engineering Mechanics Conference, NY<br />

Lam, C.C., Yam, M.C.H., Iu, V.P., and Cheng, J.J.R. (2000). “Design for lateral torsional buckling of coped<br />

I-beams.’’ Journal of Constructional Steel Research, 54(3), 423-443.<br />

Michael C.H. Yam, Angus C. C. Lam, and V. P. Iu (2003). “Local web buckling strength of coped steel I<br />

beams.” Journal of Structural Engineering, 129(1).<br />

Michael C.H. Yam, Y.C. Zhong, Angus C.C. Lam (2007). “An investigation of the block shear strength of coped<br />

beams with a welded clip angle connection - Part I: experimental study”, Journal of Constructional Steel<br />

Research 63, 96-115.<br />

Michael C.H. Yam, Y.C. Zhong, Angus C.C. Lam (2007). “An investigation of the block shear strength of coped<br />

beams with a welded clip angle connection - Part II: Numerical study”, Journal of Constructional Steel<br />

Research 63, 116-134<br />

Steelwork Design Guide to BS 5950: Part 1: 1990, Volume 1 Section properties member capacities, (2001), <strong>The</strong><br />

Steel Construction Institute, UK.<br />

Zhong, Y.C.J., Yam, C.H.M., Lam, C.C.A., and Iu, V.P. (2004). “Experimental investigation of block shear of<br />

coped beams with welded clip angles connection”, Proceedings of 5th Structural Specialty Conference of<br />

the Canadian Society for Civil Engineering, Saskatoon, Saskatchewan, Canada.<br />

-473-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

焊 接 球 对 网 壳 稳 定 性 影 响 的 精 细 化 分 析<br />

1<br />

2<br />

2<br />

武 芳 丁 茂 强 顾 磊<br />

3<br />

傅 学 怡<br />

3<br />

陈 贤 川<br />

(1. 深 圳 华 森 建 筑 与 工 程 设 计 顾 问 有 限 公 司 , 广 东 深 圳 ,518054;<br />

2. 哈 尔 滨 工 业 大 学 深 圳 研 究 生 院 , 广 东 深 圳 518055;3. 深 圳 大 学 , 广 东 深 圳 518060)<br />

摘 要 : 焊 接 空 心 球 节 点 在 网 壳 结 构 中 应 用 广 泛 , 目 前 单 层 网 壳 均 采 用 刚 接 于 结 点 的 空 间 梁 单 元<br />

进 行 分 析 , 未 能 考 虑 焊 接 球 节 点 的 影 响 。 本 文 采 用 ANSYS 软 件 中 的 Shell181 单 元 模 拟 钢 管 杆 件 和<br />

空 心 球 节 点 , 建 立 了 一 个 K6(2) 型 单 层 球 面 网 壳 的 精 细 化 有 限 元 模 型 , 进 行 考 虑 几 何 非 线 性 和 材 料<br />

弹 塑 性 的 双 重 非 线 性 分 析 , 并 通 过 精 细 化 分 析 与 传 统 的 梁 单 元 模 型 分 析 的 对 比 , 得 到 了 单 层 球 面 网<br />

壳 极 限 破 坏 时 杆 件 和 节 点 塑 性 应 力 分 布 、 细 部 破 坏 模 式 和 极 限 荷 载 的 精 细 化 分 析 结 果 。 本 文 还 进 行<br />

了 参 数 分 析 , 讨 论 了 焊 接 球 节 点 壁 厚 、 直 径 对 网 壳 极 限 荷 载 的 影 响 规 律 , 对 焊 接 空 心 球 节 点 单 层 网<br />

壳 的 有 限 元 分 析 和 工 程 设 计 提 出 了 建 议 。<br />

关 键 词 : 精 细 化 有 限 元 分 析 焊 接 空 心 球 节 点 单 层 球 面 网 壳 稳 定 性<br />

A REFINED ULTIMATE BEARING CAPACITY ANALYSIS ON SINGLE LAYER<br />

LATTICED DOMES WITH WELDED HOLLOW SPHERICAL JOINTS<br />

Wu Fang 1 Ding Maoqiang 2 Gu Lei 2 Fu Xueyi 3 Chen Xianchuan 3<br />

1<br />

Shenzhen Hua Sen Architecture & Engineering Design Consultants Ltd., Shenzhen 518054, China;<br />

2 shenzhen Graduate School, Harbin Institute of Technology, Shenzhen 518055, China;<br />

3 Shenzhen <strong>University</strong>, Shenzhen 518060, China<br />

Abstract: <strong>The</strong> Welded hollow spherical joints are widely used in latticed shells. Calculate model of space beam<br />

elements jointing rigidly at nodes is adopted in FEM analysis at present, the influence of welded hollow<br />

spherical joints is ignored. In this paper, shell181 element of ANSYS software was chosen to simulate steel<br />

tubes and spherical joints, a refined FEM structure model of a Kiewitt-6-2 latticed dome was established,<br />

nonlinear analysis considering geometrical nonlinear and material elasticity-plasticity was carried out. By<br />

comparing refine analysis with traditional beam element analysis, refined results on plastic stress distribution,<br />

detailed failure mode and limit load was concluded. With a good deal of parameters comparison, impacting<br />

factors such as thickness and diameter of welded hollow spherical joints was discussed, the influence<br />

characteristic was revealed. <strong>The</strong>n, some suggestions for FEM analysis and design of latticed dome with welded<br />

hollow spherical joints were put forward.<br />

Keywords: Refined finite element analysis, welded hollow spherical joint, single layer latticed dome, stability.<br />

一 、 前 言<br />

杆 焊 接 空 心 球 节 点 在 网 壳 结 构 中 应 用 较 广 泛 , 目 前 单 层 网 壳 均 采 用 刚 接 于 结 点 的 空 间 梁 单 元 进<br />

行 分 析 , 未 能 考 虑 焊 接 球 节 点 的 影 响 ; 而 焊 接 球 节 点 的 研 究 主 要 集 中 在 单 一 节 点 的 强 度 和 承 载 力 ,<br />

未 能 考 虑 整 体 结 构 中 球 节 点 的 性 能 。 文 献 [1] 基 于 几 何 非 线 性 的 海 量 参 数 研 究 总 结 了 网 壳 结 构 稳 定 破<br />

坏 的 全 过 程 规 律 , 提 炼 了 可 供 工 程 技 术 人 员 使 用 的 单 层 网 壳 稳 定 承 载 力 的 计 算 公 式 , 并 已 纳 入 我 国<br />

基 金 项 目 : 国 家 自 然 科 学 基 金 资 助 项 目 (90715012,50878065)、 哈 尔 滨 工 业 大 学 科 研 创 新 基 金 (HIT.NSRIF.2009133) 资 助 项 目<br />

作 者 简 介 : 武 芳 (1981-), 女 , 硕 士 , 结 构 工 程 师 , 从 事 建 筑 结 构 设 计 工 作 。<br />

-474-


规 程 [7]。 文 献 [2]、[3] 则 进 一 步 考 虑 了 材 料 弹 塑 性 、 半 刚 性 节 点 对 网 壳 稳 定 性 的 影 响 。 文 献 [4]、[5]、<br />

[6] 针 对 焊 接 球 节 点 进 行 了 数 值 模 拟 和 试 验 研 究 , 提 出 了 球 节 点 刚 度 的 计 算 方 法 和 节 点 设 计 公 式 。 至<br />

于 在 网 壳 整 体 结 构 中 考 虑 球 节 点 影 响 的 研 究 尚 不 多 见 。 本 文 采 用 ANSYS 软 件 中 的 Shell181 单 元 模<br />

拟 钢 管 杆 件 和 空 心 球 节 点 , 建 立 了 一 个 K6(2) 型 单 层 球 面 网 壳 的 精 细 化 有 限 元 模 型 , 进 行 考 虑 几 何<br />

非 线 性 和 材 料 弹 塑 性 的 双 重 非 线 性 分 析 , 并 通 过 精 细 化 分 析 与 传 统 的 梁 单 元 模 型 分 析 的 对 比 , 得 到<br />

了 单 层 球 面 网 壳 极 限 破 坏 时 杆 件 和 节 点 塑 性 应 力 分 布 、 细 部 破 坏 模 式 和 极 限 荷 载 的 精 细 化 分 析 结<br />

果 。<br />

二 、 精 细 化 有 限 元 模 型 的 初 步 分 析<br />

2.1. 分 析 模 型<br />

为 实 现 在 模 型 中 考 虑 焊 接 空 心 球 , 数 值 模 拟 时 选 用 ANSYS 中 的 Shell181 单 元 , 它 适 于 分 析 薄<br />

至 中 等 厚 度 的 壳 形 结 构 , 每 个 节 点 具 有 6 个 自 由 度 的 4 节 点 单 元 , 非 常 适 用 于 线 性 、 大 转 角 或 大 应<br />

变 非 线 性 的 计 算 。 钢 材 选 用 Q235, 采 用 双 折 线 硬 化 本 构 关 系 , 考 虑 应 力 应 变 关 系 强 化 1%, 不 考 虑<br />

包 辛 格 效 应 。<br />

选 取 K6 单 层 球 面 网 壳 为 研 究 对 象 , 鉴 于 精 细 化 分 析 时 计 算 量 限 制 , 网 壳 径 向 分 割 频 数 为 2, 即<br />

为 K6(2) 网 壳 , 跨 度 2m、 矢 高 0.4m, 如 图 1 所 示 。 模 型 荷 载 施 加 方 式 为 作 用 于 网 壳 球 节 点 的 均 布 荷<br />

载 , 周 边 采 用 固 定 刚 接 支 座 。<br />

(a) 结 构 整 体 模 型<br />

(b) 球 节 点 与 杆 件 细 部<br />

图 1 K6(2) 单 层 球 面 网 壳 精 细 化 模 型<br />

2.2. 精 细 化 分 析 与 传 统 分 析 对 比<br />

采 用 ANSYS 中 的 空 间 梁 单 元 Beam189 构 造 分 析 模 型 , 杆 件 截 面 Ф18×2。 分 几 何 非 线 性 ( 弹 性<br />

大 变 形 ) 和 双 重 非 线 性 ( 弹 塑 性 大 变 形 ) 两 种 情 况 , 提 取 了 网 壳 的 全 过 程 曲 线 ( 图 2)。<br />

采 用 壳 单 元 构 造 相 同 的 K6(2) 网 壳 模 型 , 球 节 点 直 径 D=60mm、 壁 厚 t=3mm, 进 行 双 重 非 线 性<br />

分 析 提 取 其 全 过 程 曲 线 ( 图 3), 与 空 间 梁 单 元 模 型 对 比 , 可 看 出 :<br />

图 2 梁 单 元 模 型 的 荷 载 - 位 移 曲 线<br />

图 3 精 细 化 模 型 与 梁 单 元 模 型 荷 载 - 位 移 曲 线<br />

-475-


与 考 虑 双 重 非 线 性 的 空 间 梁 单 元 网 壳 模 型 极 限 承 载 力 24.2kN 相 比 , 考 虑 空 心 球 节 点 的 壳 单 元 精<br />

细 化 网 壳 模 型 的 极 限 承 载 力 增 加 到 28.1kN, 可 见 球 节 点 的 加 入 使 得 结 构 的 整 体 稳 定 性 得 到 了 提 高 。<br />

当 网 壳 的 节 点 设 计 正 确 时 , 梁 单 元 模 型 和 考 虑 空 心 球 节 点 的 精 细 化 模 型 分 析 的 结 果 基 本 一 致 ,<br />

梁 单 元 模 型 在 材 料 进 入 塑 性 后 很 快 到 达 极 限 荷 载 。 而 精 细 化 模 型 前 期 与 梁 单 元 模 型 几 乎 重 合 , 超 越<br />

梁 单 元 模 型 的 极 限 荷 载 后 , 荷 载 - 位 移 曲 线 有 一 个 转 折 , 揭 示 出 材 料 进 入 塑 性 后 结 构 刚 度 的 下 降 , 弹<br />

塑 性 经 过 一 定 发 展 后 到 达 极 限 荷 载 。<br />

三 、 球 节 点 壁 厚 对 网 壳 的 影 响<br />

模 型 钢 管 选 为 Ф18×2, 焊 接 球 节 点 直 径 为 60mm, 球 节 点 壁 厚 选 取 如 表 1 所 示 , 采 用 弧 长 法 对<br />

网 壳 加 载 过 程 进 行 跟 踪 , 不 同 节 点 壁 厚 的 精 细 化 模 型 和 空 间 梁 单 元 模 型 的 荷 载 - 位 移 曲 线 如 图 4 所<br />

示 。 分 析 得 出 :<br />

(1) 当 节 点 壁 厚 在 2~3mm 范 围 内 ,<br />

即 节 点 壁 厚 与 杆 件 壁 厚 之 比 为 1~1.5 倍<br />

表 1 焊 接 空 心 球 节 点 壁 厚 t<br />

时 , 由 于 球 壁 厚 的 增 加 使 节 点 的 刚 度 迅<br />

速 提 增 强 , 网 壳 的 极 限 荷 载 相 应 地 有 较<br />

大 幅 度 提 高 。 当 节 点 壁 厚 在 3~5mm 范<br />

围 内 , 即 节 点 壁 厚 与 杆 件 壁 厚 之 比 为<br />

1.5~2.5 倍 变 化 时 , 虽 然 节 点 刚 度 继 续 增<br />

加 , 但 球 节 点 的 刚 度 已 接 近 网 壳 对 节 点<br />

刚 度 的 最 大 要 求 , 因 而 极 限 荷 载 变 化 很<br />

小 。<br />

(2) 文 献 [7] 中 规 定 空 心 球 壁 厚 与 钢<br />

管 最 大 壁 厚 的 比 值 宜 选 用 1.5~2.0, 由<br />

精 细 化 模 型 的 计 算 分 析 可 知 : 球 节 点 壁<br />

厚 取 1.5 倍 杆 件 壁 厚 以 上 时 , 可 按 刚 接 节<br />

[7]<br />

点 考 虑 , 证 明 网 壳 规 程 的 构 造 规 定 正 确<br />

合 理 。<br />

(3) 传 统 梁 单 元 模 型 的 极 限 荷 载 为<br />

24.2kN。 而 精 细 化 模 型 当 节 点 壁 厚 t 达 到<br />

3mm 以 后 , 网 壳 的 极 限 荷 载 保 持 为<br />

球 节 点 壁 厚 t (mm)<br />

球 壁 厚 与 管 壁 厚 比 值<br />

2<br />

1<br />

2.4<br />

1.2<br />

3<br />

1.5<br />

4<br />

2<br />

5<br />

2.5<br />

27.3kN 以 上 , 可 见 采 用 壳 单 元 的 精 细 化 图 4 不 同 球 节 点 壁 厚 的 网 壳 荷 载 - 位 移 曲 线<br />

数 值 分 析 方 法 能 反 映 出 球 节 点 使 网 壳 极<br />

限 荷 载 提 高 。<br />

0<br />

1.079<br />

2.157<br />

3.236<br />

4.314<br />

5.393<br />

6.472<br />

7.55<br />

8.629<br />

9.707<br />

0<br />

.507347 1.015 1.522<br />

2.029<br />

2.537<br />

3.044<br />

3.551<br />

4.059<br />

4.566<br />

0<br />

.394885 .78977 1.185<br />

1.58<br />

1.974<br />

2.369<br />

2.764<br />

3.159<br />

3.554<br />

(a) 球 壁 厚 t=2mm (b) 球 壁 厚 t=3mm (c) 球 壁 厚 t=5mm<br />

图 5 不 同 球 节 点 壁 厚 的 网 壳 位 移 图<br />

-476-


为 探 究 球 节 点 壁 厚 对 网 壳 稳 定 性 影 响 的 机 理 , 提 取 了 不 同 节 点 壁 厚 的 网 壳 应 力 、 塑 性 应 变 等 分<br />

布 云 图 进 行 分 析 。 网 壳 在 极 限 荷 载 时 对 应 的 位 移 图 如 图 5 所 示 , 可 以 看 出 : 节 点 壁 厚 t=2mm 时 , 网<br />

壳 的 最 大 位 移 发 生 在 顶 节 点 球 ; 节 点 壁 厚 t=3mm 时 , 顶 节 点 位 移 仍 然 最 大 , 但 相 对 t=2mm, 节 点<br />

相 连 杆 件 已 经 发 生 较 大 位 移 ; 壁 厚 t=5mm 时 网 壳 节 点 处 的 位 移 变 小 , 杆 件 局 部 区 域 出 现 最 大 位 移 。<br />

可 见 随 着 网 壳 节 点 壁 厚 增 加 , 网 壳 的 最 大 变 形 逐 步 由 球 节 点 向 杆 件 部 位 转 移 。<br />

考 察 不 同 球 节 点 壁 厚 的 网 壳 达 到 极 限 荷 载 时 的 顶 节 点 塑 性 分 布 云 图 ( 图 6), 可 看 出 : 当 球 节<br />

点 壁 厚 t=3mm 以 后 节 点 的 刚 度 相 对 于 杆 件 迅 速 提 高 , 虽 然 在 顶 节 点 处 仍 发 现 有 部 分 区 域 进 入 塑 性 ,<br />

但 相 对 于 t=2mm 时 的 节 点 塑 性 区 域 已 经 减 小 , 当 球 节 点 壁 厚 t=5mm 时 球 节 点 基 本 无 塑 性 分 布 。<br />

0<br />

.015437 .030874 .046311 .061748 .077186 .092623 .10806 .123497 0<br />

.1389 .003182 .006364 .009546 .012728 .01591 .019092 .022274 .025456 0<br />

.028 .226E-03 .452E-03 .678E-03 .904E-03 .001356 .001808<br />

.00113<br />

.001582 .002035<br />

(a) 球 壁 厚 t=2mm (b) 球 壁 厚 t=3mm (c) 球 壁 厚 t=5mm<br />

图 6 不 同 球 节 点 壁 厚 细 部 塑 性 云 图<br />

塑 性 区 域 的 分 布 可 以 很 好 地 解 释 图 4 中 球 节 点 壁 厚 从 2mm 增 加 到 5mm 过 程 中 , 荷 载 - 位 移 曲 线<br />

出 现 的 差 异 : 球 节 点 壁 厚 t=2mm, 与 杆 件 壁 厚 相 等 , 塑 性 区 域 在 球 节 点 内 发 展 , 因 此 荷 载 - 位 移 曲 线<br />

是 一 条 光 滑 的 弹 塑 性 曲 线 ; 球 节 点 壁 厚 t=2.4mm、3mm, 塑 性 区 域 先 由 杆 端 发 展 到 节 点 , 此 时 荷 载 -<br />

位 移 曲 线 出 现 转 折 , 待 节 点 区 塑 性 发 展 后 到 达 极 限 荷 载 ; 球 节 点 壁 厚 t=4mm、5mm, 塑 性 区 域 仅 在<br />

杆 件 中 发 展 并 到 达 极 限 荷 载 , 属 杆 件 弹 塑 性 失 稳 , 承 载 力 迅 速 下 降 , 因 而 荷 载 - 位 移 曲 线 与 传 统 的 梁<br />

单 元 模 型 形 式 一 致 , 但 考 虑 节 点 的 作 用 , 极 限 荷 载 有 所 提 高 。<br />

四 、 球 节 点 直 径 对 网 壳 的 影 响<br />

球 节 点 壁 厚 统 一 取 4mm, 直 径 选 取 如 表 2 所 示 , 采 用 弧 长 法 对 网 壳 加 载 过 程 进 行 跟 踪 , 获 取 不<br />

同 节 点 直 径 时 的 荷 载 位 移 曲 线 如 图 7 所 示 。 分 析 得 出 :<br />

(1) 球 节 点 直 径 由 55mm 增 大 到 80mm 过 程 中 网 壳 的 荷 载 - 位 移 曲 线 几 乎 完 全 重 合 , 网 壳 的 失<br />

稳 形 式 基 本 一 致 。<br />

(2) 由 于 球 节 点 的 壁 厚 保 持 不 变 , 随 着 直 径 的 增 加 球 节 点 自 身 的 刚 度 下 降 , 而 节 点 直 径 的 增<br />

大 对 网 壳 节 点 与 杆 件 几 何 比 例 的 影 响 甚 微 , 因 而 网 壳 的 极 限 荷 载 略 有 下 降 。<br />

(3) 球 节 点 直 径 为 90mm 时 网 壳 的 荷 载 - 位 移 曲 线 略 有 变 化 , 参 照 位 移 图 ( 图 8) 可 发 现 网 壳<br />

球 节 点 直 径 为 70mm 时 , 网 壳 的 失 稳 形 式 为 顶 节 点 一 侧 的 3 根 杆 件 大 变 形 引 起 的 非 对 称 失 稳 ; 球 节<br />

点 直 径 增 大 到 90mm 后 , 网 壳 的 失 稳 形 式 变 为 顶 节 点 周 围 所 有 6 根 杆 件 对 称 大 变 形 引 起 的 环 状 失 稳 ,<br />

由 于 失 稳 模 态 的 变 化 , 网 壳 的 极 限 承 载 力 没 有 继 续 下 降 , 反 而 略 有 提 高 。<br />

总 之 , 焊 接 球 直 径 通 常 由 交 汇 杆 件 便 于 焊 接 的 几 何 构 造 确 定 , 取 值 不 会 过 大 或 过 小 , 因 此 对 网<br />

壳 稳 定 性 影 响 不 大 , 但 会 影 响 局 部 的 失 稳 形 态 。<br />

表 2 焊 接 空 心 球 节 点 直 径 D<br />

球 节 点 直 径 D(mm) 55 60 65 70 80 90<br />

球 直 径 与 球 壁 厚 比 值 13.75 15 15.5 17.5 20 22.5<br />

-477-


图 7 不 同 球 节 点 直 径 的 网 壳 荷 载 - 位 移 曲 线<br />

(a) 球 直 径 D=70mm<br />

图 8 不 同 球 节 点 直 径 的 网 壳 位 移 图<br />

(b) 球 直 径 D=90mm<br />

五 、 结 论<br />

本 文 采 用 Shell181 单 元 建 立 了 考 虑 空 心 球 节 点 的 精 细 化 网 壳 有 限 元 模 型 , 分 析 发 现 当 球 节 点 设<br />

计 合 理 、 强 节 点 弱 杆 件 的 条 件 下 , 传 统 梁 单 元 和 壳 单 元 精 细 化 分 析 具 有 一 致 性 , 传 统 梁 单 元 模 型 具<br />

有 可 靠 的 精 度 , 完 全 满 足 工 程 分 析 和 设 计 的 需 求 。 但 精 细 化 模 型 更 能 反 映 出 焊 接 球 节 点 的 影 响 , 并<br />

可 揭 示 出 网 壳 稳 定 破 坏 的 弹 塑 性 受 力 机 理 。 本 文 主 要 结 论 如 下 :<br />

(1) 网 壳 规 程 [7] 中 规 定 空 心 球 壁 厚 与 钢 管 最 大 壁 厚 的 比 值 宜 选 用 1.5~2.0 正 确 合 理 , 球 节 点<br />

壁 厚 取 1.5 倍 杆 件 壁 厚 以 上 时 , 可 按 刚 接 节 点 考 虑 , 精 细 化 分 析 的 极 限 荷 载 略 高 于 传 统 梁 单 元 模 型<br />

的 极 限 荷 载 。<br />

(2) 若 球 节 点 设 计 合 理 , 网 壳 破 坏 时 塑 性 区 先 由 杆 端 开 始 , 逐 渐 向 球 体 扩 散 , 塑 性 有 所 发 展<br />

后 到 达 弹 塑 性 极 限 荷 载 。<br />

(3) 若 球 节 点 不 满 足 壁 厚 构 造 要 求 , 则 刚 接 于 节 点 的 梁 单 元 计 算 模 型 将 高 估 网 壳 的 极 限 荷 载<br />

和 结 构 刚 度 。<br />

(4) 球 节 点 直 径 的 变 化 对 网 壳 稳 定 性 影 响 不 大 , 但 会 影 响 局 部 的 失 稳 形 态 。<br />

参 考 文 献<br />

[1] 沈 世 钊 , 陈 昕 . 网 壳 结 构 稳 定 性 , 北 京 : 科 学 出 版 社 , 1999, 102-109.<br />

[2] 曹 正 罡 , 范 峰 , 沈 世 钊 . 单 层 球 面 网 壳 的 弹 塑 性 稳 定 性 . 土 木 工 程 学 报 , 2006, 39(10): 6-10.<br />

[3] 范 峰 , 马 会 环 , 沈 世 钊 . 半 刚 性 型 螺 栓 球 节 点 单 层 K8 型 网 壳 弹 塑 性 稳 定 分 析 . 土 木 工 程 学 报 , 2009, 42(2): 45-52.<br />

-478-


[4] 王 星 , 董 石 麟 , 完 海 鹰 . 焊 接 球 节 点 刚 度 的 有 限 元 分 析 . 浙 江 大 学 学 报 ( 工 学 版 ), 2000, 34(1): 77-82.<br />

[5] 董 石 麟 , 唐 海 军 , 赵 阳 , 傅 学 怡 , 顾 磊 . 轴 力 和 弯 矩 共 同 作 用 下 焊 接 空 心 球 节 点 承 载 力 研 究 与 实 用 计 算 方 法 , 2005,<br />

38(1): 21-30.<br />

[6] 韩 庆 华 , 潘 延 东 , 刘 锡 良 . 焊 接 空 心 球 节 点 的 拉 压 极 限 承 载 力 分 析 . 土 木 工 程 学 报 , 2003, 36(10): 1-6.<br />

[7] 中 华 人 民 共 和 国 行 业 标 准 , JGJ 61─2003 网 壳 结 构 技 术 规 程 , 北 京 : 中 国 建 筑 工 业 出 版 社 , 2003.<br />

[8] 丁 茂 强 . 单 层 球 面 网 壳 极 限 荷 载 的 精 细 化 有 限 元 分 析 , 哈 尔 滨 工 业 大 学 深 圳 研 究 生 院 , 2010, 空 间 结 构 , 2000, 6(1):<br />

14-21.<br />

-479-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

錨 錠 系 統 對 於 短 鋼 纜 振 態 頻 率 識 別 之 影 響<br />

陳 建 州 吳 文 華 劉 郡 晏<br />

( 國 立 雲 林 科 技 大 學 營 建 工 程 系 , 斗 六 , 雲 林 )<br />

摘 要 : 為 建 立 斜 張 鋼 纜 振 態 頻 率 與 其 索 力 的 關 係 式 , 以 往 研 究 均 假 設 鋼 纜 為 一 維 構 件 , 兩 端<br />

錨 錠 裝 置 常 視 為 單 點 固 接 、 鉸 接 或 同 時 組 合 旋 轉 彈 簧 ; 若 進 一 步 考 量 鋼 纜 上 其 他 橡 膠 束 制 裝 置 的 影<br />

響 , 則 可 加 入 側 向 線 性 彈 簧 模 擬 或 另 外 決 定 有 效 振 動 長 度 。 換 句 話 說 , 即 是 將 錨 錠 裝 置 簡 化 為 僅 提<br />

供 束 制 的 邊 界 條 件 。 對 於 較 長 之 斜 張 鋼 纜 而 言 , 此 種 分 析 方 法 已 確 認 可 精 準 掌 握 鋼 纜 的 振 動 特 性 。<br />

但 對 常 見 於 矮 塔 斜 張 橋 的 較 短 鋼 纜 來 說 , 錨 錠 裝 置 相 對 在 整 體 鋼 纜 系 統 中 扮 演 更 為 顯 著 的 角 色 , 所<br />

以 可 能 必 須 將 其 納 為 分 析 之 次 系 統 而 非 僅 邊 界 條 件 , 如 此 方 足 以 精 確 反 應 鋼 纜 的 實 際 振 動 行 為 。 透<br />

過 根 據 愛 蘭 矮 塔 斜 張 橋 之 鋼 纜 微 振 訊 號 所 進 行 的 振 態 頻 率 識 別 , 同 時 經 由 有 限 元 素 模 型 交 叉 分 析 比<br />

對 , 本 研 究 驗 證 較 短 鋼 纜 的 振 態 頻 率 確 實 會 受 到 錨 錠 系 統 振 動 的 干 擾 而 造 成 識 別 上 的 混 淆 , 並 且 提<br />

出 選 用 鋼 纜 頻 率 進 行 索 力 計 算 時 應 須 注 意 的 要 點 。<br />

關 鍵 詞 : 短 鋼 纜 錨 錠 系 統 振 態 頻 率 識 別 混 淆<br />

EFFECT OF ANCHORAGE SYSTEM IN IDENTIFYING MODAL FREQUENCIES OF<br />

SHORT STAY CABLES<br />

C.-C. Chen, 1 W.-H. Wu 1 and C.-Y. Liu 1<br />

1 Dept. of Construction Eng., National Yunlin <strong>University</strong> of Science and Technology, Touliu, Yunlin 640, Taiwan<br />

Abstract: To establish the relationship between the modal frequencies of a stay cable and its axial force, the<br />

cable is conventionally assumed to be a one-dimensional member and the anchorage devices at its both ends are<br />

usually regarded as a single point boundary with either a fixed, hinged, or rotationally sprung constraint. If the<br />

other centralized rubber devices installed on cables are also considered, either certain linear springs in the<br />

transverse direction can be added in the model or the effective cable vibration length can be more delicately<br />

determined. In other words, the anchorage devices in the above analysis are simplified as certain boundary<br />

conditions to provide constraints. This simplification has been verified to be effective in accurately<br />

imitating the vibration behaviour of longer stay cables. For shorter cables commonly adopted in extradosed<br />

bridges, however, the anchorage devices play a relatively more prominent role in the whole cable system.<br />

<strong>The</strong>refore, it is necessary to further include the anchorage devices as sub-systems in this case, instead of<br />

just taking them as boundary conditions, such that the vibration behaviour of cable can be precisely<br />

reflected. According to the modal frequency identification based on the ambient vibration measurements<br />

for the cables of Ai-Lan Extradosed Bridge and the cross examination with finite element analysis, this<br />

study confirms that the modal frequencies of short cables are actually interfered by the anchorage systems<br />

and can cause serious problems in identification due to peak ambiguity. Certain guidelines in screening the<br />

cable frequencies for determining the cable force are also proposed in this paper.<br />

Keywords: Short cable, anchorage system, modal frequency, peak ambiguity.<br />

-480-


一 、 前 言<br />

鋼 纜 為 斜 張 橋 的 主 要 受 力 構 件 , 斜 張 橋 體 的 內 力 狀 態 與 鋼 纜 索 力 的 大 小 及 配 置 息 息 相 關 。 當 橋<br />

體 受 異 常 荷 載 作 用 或 受 損 時 , 索 力 常 會 隨 之 改 變 , 因 此 長 期 以 來 監 測 索 力 的 變 化 成 為 進 行 斜 張 橋 安<br />

全 評 估 的 重 要 工 作 。 微 振 法 為 求 取 索 力 常 用 的 方 法 之 一 , 主 要 是 經 由 量 測 的 微 振 訊 號 識 別 鋼 纜 的 振<br />

態 頻 率 , 再 依 鋼 纜 振 態 頻 率 與 索 力 之 關 係 計 算 索 力 值 的 大 小 。 為 求 取 準 確 的 索 力 值 , 經 由 微 振 訊 號<br />

識 別 出 鋼 纜 正 確 振 態 頻 率 為 首 要 條 件 。 由 於 鋼 纜 為 一 細 長 比 相 當 大 的 構 件 , 雖 然 微 振 訊 號 僅 來 自 於<br />

環 境 外 力 的 影 響 , 振 態 頻 率 的 識 別 卻 非 特 別 困 難 的 作 業 , 一 般 常 直 接 經 由 微 振 反 應 之 傅 利 葉 振 幅 譜<br />

進 行 主 要 尖 峰 頻 率 判 讀 。 由 於 所 使 用 為 鋼 纜 反 應 之 傅 利 葉 振 幅 譜 , 其 振 幅 尖 峰 處 亦 可 能 來 自 外 力 頻<br />

率 之 集 中 而 與 系 統 無 關 , 故 為 避 免 判 讀 可 能 發 生 的 錯 誤 , 常 輔 以 鋼 纜 振 態 頻 率 間 的 特 定 關 係 來 協 助<br />

精 準 判 讀 。 鋼 纜 各 振 態 頻 率 的 大 小 與 相 鄰 間 距 的 關 係 除 與 索 力 值 有 關 外 , 亦 與 重 力 垂 度 、 撓 曲 剛 度<br />

及 邊 界 條 件 等 有 關 , 此 一 課 題 在 以 往 研 究 中 已 多 有 探 討 。 相 關 研 究 均 假 設 鋼 纜 為 一 維 構 件 , 兩 端 錨<br />

錠 裝 置 常 視 為 單 點 固 接 、 鉸 接 或 同 時 組 合 旋 轉 彈 簧 ; 若 進 一 步 考 量 鋼 纜 上 其 他 橡 膠 束 制 裝 置 ( 或 阻<br />

尼 器 ) 的 影 響 , 則 可 加 入 側 向 線 性 彈 簧 模 擬 或 另 外 決 定 有 效 振 動 長 度 。 換 句 話 說 , 即 是 將 錨 錠 裝 置<br />

簡 化 為 僅 提 供 束 制 的 邊 界 條 件 , 再 以 理 論 推 導 或 有 限 元 素 分 析 求 解 索 力 與 振 態 頻 率 間 的 關 係 以 及 各<br />

振 態 頻 率 間 的 關 係 [1-2]。<br />

對 於 較 長 之 斜 張 鋼 纜 而 言 , 上 述 分 析 方 法 已 確 認 可 精 準 掌 握 鋼 纜 的 振 動 特 性 , 若 配 合 微 振 訊 號<br />

之 量 測 便 可 識 別 出 準 確 的 鋼 纜 振 態 頻 率 , 並 進 一 步 算 出 正 確 的 索 力 值 。 由 以 往 長 期 累 積 的 鋼 纜 微 振<br />

訊 號 量 測 識 別 資 料 可 知 , 傳 統 斜 張 橋 之 鋼 纜 若 假 設 為 兩 端 受 不 同 邊 界 條 件 束 制 的 單 純 一 維 構 件 進 行<br />

分 析 , 通 常 均 可 求 得 與 實 測 振 態 頻 率 相 當 一 致 的 分 析 頻 率 。 同 時 鋼 纜 各 振 態 頻 率 間 的 關 係 可 概 分 為<br />

三 類 , 若 鋼 纜 振 動 為 弦 模 式 , 則 其 各 振 態 頻 率 呈 等 差 級 數 的 關 係 ; 若 為 拉 力 梁 模 式 , 則 各 相 鄰 振 態<br />

頻 率 的 差 值 隨 振 態 之 升 高 而 增 大 ; 若 鋼 纜 受 重 力 垂 度 的 影 響 , 則 基 本 振 態 之 頻 率 將 特 別 變 大 。 上 述<br />

各 振 態 頻 率 關 係 均 可 由 傳 統 斜 張 橋 鋼 纜 微 振 訊 號 的 傅 利 葉 振 幅 譜 得 到 證 實 。 除 鋼 纜 的 振 態 頻 率 外 ,<br />

振 幅 譜 中 亦 常 可 看 出 橋 體 的 某 些 振 態 頻 率 , 這 些 橋 體 振 態 頻 率 主 要 座 落 於 低 頻 區 域 , 常 與 鋼 纜 頻 率<br />

同 時 存 在 而 不 相 互 干 擾 [3]。<br />

近 年 來 , 為 因 應 主 跨 200 公 尺 左 右 造 型 橋 的 設 計 需 求 , 結 構 系 統 介 於 傳 統 斜 張 橋 與 與 變 梁 深 預<br />

力 混 凝 土 橋 之 矮 塔 斜 張 橋 的 建 造 數 量 已 有 快 速 增 多 的 趨 勢 , 尤 其 是 在 台 灣 、 中 國 大 陸 及 日 本 等 三 地 。<br />

相 對 於 傳 統 斜 張 橋 , 矮 塔 斜 張 橋 的 塔 高 與 跨 徑 之 比 值 較 小 , 鋼 纜 長 度 也 因 此 較 短 且 傾 角 較 小 , 而 振<br />

態 頻 率 則 相 對 較 大 。 但 是 近 來 對 於 矮 塔 斜 張 橋 部 分 短 鋼 纜 所 進 行 的 實 測 顯 示 , 前 述 一 維 構 件 分 析 模<br />

型 卻 無 法 解 釋 實 測 所 得 之 鋼 纜 振 態 頻 率 , 即 所 識 別 振 態 頻 率 並 非 屬 於 弦 模 式 或 拉 力 梁 模 式 , 亦 不 似<br />

重 力 垂 度 影 響 所 造 成 的 改 變 。 對 此 類 常 見 於 矮 塔 斜 張 橋 的 較 短 鋼 纜 來 說 , 錨 錠 裝 置 相 對 在 整 體 鋼 纜<br />

系 統 中 扮 演 更 為 顯 著 的 角 色 , 所 以 可 能 必 須 將 其 納 為 分 析 之 次 系 統 而 非 僅 邊 界 條 件 , 如 此 方 足 以 精<br />

確 反 應 鋼 纜 的 實 際 振 動 行 為 。<br />

本 研 究 針 對 位 於 台 灣 中 部 的 愛 蘭 矮 塔 斜 張 橋 鋼 纜 進 行 微 振 量 測 , 根 據 這 些 微 振 訊 號 進 行 振 態 頻<br />

率 識 別 的 詳 細 分 析 發 現 , 其 對 應 傅 利 葉 振 幅 譜 在 高 於 某 一 特 定 的 頻 率 之 後 顯 得 特 別 複 雜 , 因 此 也 相<br />

對 極 易 干 擾 這 個 高 頻 範 圍 內 鋼 纜 振 態 頻 率 的 識 別 。 更 明 確 地 說 , 振 幅 譜 中 存 在 兩 組 明 顯 的 尖 峰 頻 率 ,<br />

其 中 一 組 接 近 鋼 纜 的 振 態 頻 率 , 但 各 頻 率 間 的 關 係 卻 又 難 以 完 全 透 過 前 述 的 單 一 系 統 一 維 構 件 分 析<br />

模 型 來 解 釋 , 而 另 一 組 尖 峰 頻 率 亦 明 顯 並 非 橋 體 的 振 態 頻 率 。 若 再 進 一 步 檢 視 這 兩 組 尖 峰 頻 率 , 其<br />

關 係 像 是 其 中 一 至 數 個 鋼 纜 振 態 頻 率 偏 離 單 一 系 統 模 型 所 分 析 的 頻 率 值 , 同 時 另 一 組 頻 率 又 與 其 相<br />

互 干 擾 對 應 。 此 行 為 類 似 結 構 控 制 中 主 結 構 系 統 受 到 調 諧 質 量 阻 尼 器 (tuned mass damper) 次 系 統 的<br />

影 響 , 所 以 綜 合 研 判 非 鋼 纜 振 態 頻 率 的 另 一 組 尖 峰 頻 率 極 可 能 是 來 自 於 錨 錠 裝 置 所 形 成 次 系 統 之 貢<br />

獻 。 因 此 , 本 研 究 依 錨 錠 裝 置 的 組 成 進 行 有 限 元 素 模 擬 , 將 錨 錠 裝 置 與 鋼 纜 構 件 視 為 兩 個 獨 立 系 統 ,<br />

其 間 以 線 性 彈 簧 元 素 進 行 連 結 , 此 一 線 性 彈 簧 元 素 主 要 用 於 模 擬 鋼 纜 束 制 機 制 ( 阻 尼 器 ) 的 行 為 。 經<br />

由 這 項 有 限 元 素 分 析 的 結 果 , 可 清 楚 證 實 分 屬 鋼 纜 構 件 與 錨 錠 裝 置 等 兩 組 振 態 頻 率 的 存 在 , 同 時 其<br />

中 部 分 鋼 纜 振 態 確 因 調 諧 作 用 而 偏 離 原 有 的 頻 率 值 。 本 文 首 先 以 單 一 系 統 一 維 構 件 模 型 解 析 說 明 鋼<br />

-481-


纜 的 振 動 行 為 ; 其 次 利 用 集 鹿 斜 張 橋 與 愛 蘭 矮 塔 斜 張 橋 之 鋼 纜 的 識 別 頻 率 對 照 說 明 長 、 短 鋼 纜 之 間<br />

振 動 特 性 的 差 異 ; 最 後 則 以 有 限 元 素 模 型 結 果 分 析 錨 錠 次 系 統 對 鋼 纜 主 系 統 振 態 頻 率 的 影 響 。<br />

二 、 鋼 纜 側 向 振 動 行 為<br />

鋼 纜 斜 張 鋼 纜 的 設 計 斷 面 組 成 除 主 拉 鋼 材 外 , 一 般 常 包 含 保 護 套 管 與 灌 漿 材 , 鋼 材 斷 面 的 大 小<br />

與 鋼 纜 的 設 計 力 量 有 關 , 一 般 鋼 纜 斷 面 直 徑 約 在 10~30 公 分 之 間 。 鋼 纜 設 計 長 度 則 與 斜 張 橋 的 設 計<br />

跨 徑 規 模 有 關 , 以 現 代 化 的 斜 張 橋 而 言 , 最 短 十 幾 公 尺 , 最 長 則 達 數 百 公 尺 。 由 組 成 構 造 可 知 , 鋼<br />

纜 為 一 細 長 比 相 當 大 的 拉 力 構 件 , 若 將 鋼 纜 視 為 單 系 統 的 一 維 受 拉 力 構 件 , 鋼 纜 面 內 側 向 振 動 之 運<br />

動 方 程 式 可 表 示 如 下 :<br />

4<br />

2<br />

EI ∂ v ∂ v ∂ v d y<br />

− T + m = h<br />

4 2<br />

2<br />

∂x<br />

∂ x ∂x<br />

d<br />

2 x<br />

2<br />

2<br />

(1)<br />

其 中 x = 鋼 纜 軸 向 座 標 、y(x) = 鋼 纜 面 內 垂 直 x 軸 方 向 受 鋼 纜 自 重 作 用 之 變 形 、v(x,t) = 相 對 於 鋼<br />

纜 自 重 變 形 狀 態 之 側 向 振 動 位 移 、EI = 鋼 纜 撓 曲 剛 度 、T = 鋼 纜 索 力 值 、m = 鋼 纜 單 位 長 度 質 量 、<br />

h = 鋼 纜 索 力 因 振 動 所 引 起 的 改 變 量 。(1) 式 中 第 一 項 為 側 向 撓 曲 剛 度 對 鋼 纜 振 動 的 影 響 , 第 二 項 為<br />

索 力 對 鋼 纜 振 動 的 影 響 , 而 等 號 右 側 則 為 重 力 垂 度 與 索 力 改 變 量 對 於 振 動 的 影 響 。 假 設 重 力 垂 度 及<br />

撓 曲 勁 度 皆 可 忽 略 的 情 況 下 , 此 即 鋼 纜 符 合 弦 理 論 (string theory) 模 式 , 則 可 理 論 解 出 索 力 與 各 振 態<br />

頻 率 關 係 的 解 析 式 如 下 :<br />

2<br />

2 ⎛ f ⎞<br />

4ml<br />

⎟<br />

⎠<br />

n f<br />

T = ⎜ 或 n T<br />

= (2)<br />

2<br />

⎝ n n 4ml<br />

其 中 f n = 鋼 纜 第 n 振 態 頻 率 、l = 鋼 纜 振 動 長 度 。 由 (2) 式 之 數 學 式 可 知 鋼 纜 各 振 態 頻 率 值 成 等 差<br />

級 數 , 亦 即 由 鋼 纜 振 動 訊 號 之 傅 利 葉 振 幅 譜 所 識 別 的 各 振 態 識 別 頻 率 將 呈 等 間 距 關 係 。 若 須 進 一 步<br />

考 慮 撓 曲 剛 度 的 影 響 , 則 鋼 纜 面 內 側 向 振 動 行 為 較 接 近 拉 力 梁 模 式 。 假 設 兩 端 邊 界 條 件 為 鉸 接 , 可<br />

理 論 解 出 索 力 與 各 振 態 頻 率 關 係 的 解 析 式 如 下 :<br />

2 2 2<br />

2⎛<br />

fn ⎞ EIn π<br />

T 4ml<br />

⎜ ⎟ −<br />

2<br />

= 或<br />

⎝ n ⎠ l<br />

f n<br />

n<br />

2<br />

2<br />

T EIn π<br />

= +<br />

(3)<br />

2<br />

4<br />

4ml<br />

4ml<br />

由 (3) 式 可 知 鋼 纜 相 鄰 振 態 頻 率 間 距 振 態 之 升 高 而 增 大 。 對 於 兩 端 邊 界 條 件 為 固 接 之 鋼 纜 , 雖 無 法<br />

直 接 根 據 (3) 式 精 確 描 述 索 力 與 振 態 頻 率 之 關 係 , 但 仍 可 用 其 說 明 相 鄰 振 態 頻 率 間 距 逐 漸 變 大 的 關<br />

係 。 若 鋼 纜 須 考 慮 重 力 垂 度 的 影 響 , 目 前 仍 無 法 以 理 論 解 得 索 力 與 振 態 頻 率 的 解 析 式 , 但 可 推 導 出<br />

偶 數 振 態 幾 乎 不 受 重 力 垂 度 的 影 響 , 而 各 奇 數 振 態 中 則 以 基 本 振 態 所 受 影 響 最 為 顯 著 。<br />

綜 合 前 述 可 知 , 若 將 鋼 纜 視 為 單 系 統 一 維 構 件 模 型 , 則 由 其 微 振 訊 號 之 傅 利 葉 振 幅 譜 所 識 別 的<br />

各 鋼 纜 振 態 頻 率 約 可 概 分 為 三 種 模 式 :(a) 鋼 纜 不 受 重 力 垂 度 及 撓 曲 剛 度 的 影 響 , 各 振 態 頻 率 呈 等<br />

間 距 分 佈 , 常 見 於 中 鋼 纜 ;(b) 鋼 纜 不 受 重 力 垂 度 的 影 響 , 但 受 撓 曲 剛 度 的 影 響 , 振 態 頻 率 間 距 逐<br />

漸 變 大 , 常 見 於 短 纜 纜 ;(c) 鋼 纜 不 受 撓 曲 剛 度 的 影 響 , 但 受 重 力 垂 度 的 影 響 , 基 本 振 態 頻 率 會 被<br />

放 大 , 常 見 於 長 鋼 纜 。<br />

三 、 鋼 纜 微 振 量 測 與 振 態 頻 率 識 別<br />

為 解 說 錨 錠 系 統 對 鋼 纜 系 統 振 態 頻 率 的 影 響 , 本 研 究 除 針 對 愛 蘭 橋 斜 張 鋼 纜 進 行 微 振 訊 號 量 測<br />

外 , 亦 量 測 集 鹿 橋 的 斜 張 鋼 纜 , 前 者 採 矮 塔 斜 張 橋 之 結 構 配 置 , 後 者 則 為 傳 統 斜 張 橋 。 愛 蘭 橋 爲 一<br />

對 稱 雙 塔 矮 塔 斜 張 橋 , 主 梁 全 長 300m, 跨 徑 配 置 為 80+140+80m , 橋 塔 較 橋 面 高 出 20.25 m, 鋼 纜<br />

沿 主 梁 中 心 線 採 單 面 豎 琴 形 配 置 , 如 圖 1 所 示 。 愛 蘭 橋 所 有 鋼 纜 均 由 31 根 15.2mm 之 七 股 鋼 鉸 線 平<br />

行 組 成 , 兩 端 以 OVM250 錨 錠 裝 置 與 橋 塔 及 主 梁 結 合 , 每 座 橋 塔 兩 側 均 配 置 9 對 鋼 纜 , 共 計 配 置<br />

72 根 鋼 纜 , 鋼 纜 長 度 由 31.26m~64.63m, 鋼 纜 的 編 號 亦 如 圖 1 所 示 。 集 鹿 斜 張 橋 採 單 塔 對 稱 配 置 ,<br />

-482-


圖 1 愛 蘭 矮 塔 斜 張 橋 配 置 圖<br />

圖 2 集 鹿 斜 張 橋 配 置 圖<br />

主 梁 全 長 240m, 跨 徑 配 置 為 120m+120m, 橋 塔 由 橋 面 至 塔 頂 高 58m, 鋼 纜 沿 主 梁 中 心 線 採 單 面 混<br />

合 扇 形 配 置 , 如 圖 2 所 示 。 集 鹿 斜 張 橋 所 有 鋼 纜 均 由 12.7mm 七 股 鋼 鉸 線 平 行 組 成 , 依 組 成 根 數 差<br />

異 可 分 為 37 根 、43 根 及 55 根 三 種 型 式 , 鋼 纜 兩 端 以 APS 錨 錠 裝 置 與 橋 塔 及 主 梁 連 結 , 橋 塔 兩 側<br />

均 配 置 17 對 鋼 纜 , 共 計 配 置 68 根 鋼 纜 , 鋼 纜 長 度 由 28.31m~125.22m, 鋼 纜 的 編 號 亦 如 圖 2 所 示 。<br />

鋼 纜 振 態 頻 率 識 別 之 分 析 數 據 來 自 於 面 內 垂 直 鋼 纜 向 的 量 測 微 振 速 度 訊 號 , 量 測 儀 器 為 日 本 東<br />

京 測 振 VSE-15D 伺 服 型 速 度 計 , 取 樣 頻 率 為 200Hz, 每 次 量 測 時 間 為 300 秒 , 每 筆 共 計 60000 點 速<br />

度 歷 時 資 料 。 鋼 纜 振 態 頻 率 主 要 是 以 微 振 速 度 訊 號 的 傅 利 葉 振 幅 譜 之 尖 峰 值 進 行 識 別 , 再 輔 以 前 節<br />

所 述 振 態 頻 率 之 可 能 關 係 進 行 確 認 。 利 用 傳 統 的 傅 利 葉 分 析 , 可 將 一 個 時 間 域 的 函 數 經 由 傅 利 葉 轉<br />

換 (Fourier transform) 呈 現 為 頻 率 域 中 的 複 數 函 數 。 而 此 複 數 函 數 在 不 同 頻 率 下 之 振 幅 大 小 , 即 直 接<br />

代 表 原 函 數 在 各 個 頻 率 的 組 成 權 重 , 一 般 以 其 直 接 對 頻 率 作 圖 來 檢 視 原 函 數 之 頻 率 內 容 , 即 為 所 謂<br />

的 傅 利 葉 振 幅 譜 (Fourier amplitude spectrum)。 對 於 一 個 廣 義 的 結 構 系 統 來 說 , 其 輸 出 反 應 在 頻 率 域<br />

為 轉 換 函 數 與 輸 入 外 力 傅 利 葉 轉 換 之 乘 積 , 所 以 結 構 輸 出 量 測 之 傅 利 葉 振 幅 譜 必 然 反 映 轉 換 函 數 在<br />

系 統 振 態 頻 率 處 所 產 生 的 振 幅 尖 峰 值 。 就 頻 率 識 別 而 言 , 因 鋼 纜 各 振 態 頻 率 均 成 近 似 等 差 數 列 的 簡<br />

單 數 學 關 係 , 所 以 經 由 傅 利 葉 振 幅 譜 的 主 要 規 則 尖 峰 值 進 行 判 讀 , 一 般 識 別 作 業 的 進 行 並 無 任 何 困<br />

難 。<br />

本 文 首 先 取 集 鹿 斜 張 橋 長 、 中 、 短 三 根 鋼 纜 微 振 訊 號 的 傅 利 葉 振 幅 譜 說 明 傳 統 斜 張 橋 之 鋼 纜 振<br />

態 頻 率 分 佈 , 以 及 可 能 同 時 存 在 的 橋 體 振 態 。 長 、 中 、 短 三 根 鋼 纜 的 編 號 依 序 為 L34、L17、L01,<br />

分 別 位 於 一 扇 鋼 纜 的 兩 側 與 中 央 , 如 圖 2 紅 色 三 角 標 所 示 。 圖 3 至 圖 5 依 序 為 L34、L17、L01 三 根<br />

鋼 纜 之 傅 利 葉 振 幅 譜 , 均 顯 示 出 清 晰 的 尖 峰 值 。 由 圖 3 可 明 確 識 別 出 R34 長 鋼 纜 之 前 10 個 振 態 頻<br />

率 , 數 值 依 序 為 0.91Hz、1.80 Hz、2.70 Hz、3.60 Hz、4.50 Hz、……。 從 其 振 態 頻 率 值 的 等 間 距 變<br />

化 趨 勢 可 知 ,R34 長 鋼 纜 的 振 動 行 為 屬 於 弦 模 式 。 由 圖 4 亦 可 明 確 識 別 出 L17 中 鋼 纜 之 前 10 個 振 態<br />

頻 率 , 數 值 依 序 為 1.67 Hz、3.35 Hz、5.03 Hz、6.72 Hz、8.41 Hz、……。 根 據 其 振 態 頻 率 間 距 緩 慢<br />

逐 漸 增 大 的 趨 勢 可 知 ,L17 中 鋼 纜 的 振 動 行 為 雖 近 似 弦 模 式 , 但 已 些 微 受 撓 曲 剛 度 的 影 響 。 相 對 於<br />

L34 長 鋼 纜 與 L17 中 鋼 纜 , 圖 5 顯 示 L01 短 鋼 纜 各 振 態 頻 率 之 尖 峰 處 在 傅 利 葉 振 幅 譜 中 較 不 明 顯 ,<br />

綜 合 研 判 其 振 態 頻 率 依 序 為 4.51 Hz、9.15 Hz、13.76 Hz。 由 其 振 態 頻 率 間 距 明 顯 增 大 的 趨 勢 可 知 ,<br />

L01 短 鋼 纜 的 振 動 行 為 受 到 撓 曲 剛 度 相 當 程 度 的 影 響 , 振 動 模 式 接 近 拉 力 梁 。 綜 合 比 對 三 根 鋼 纜 之<br />

振 態 頻 率 識 別 結 果 可 推 論 , 集 鹿 斜 張 橋 所 有 鋼 纜 受 重 力 垂 度 的 影 響 均 相 當 小 , 短 鋼 纜 的 振 動 行 為 則<br />

受 到 撓 曲 剛 度 的 影 響 。 除 此 之 外 , 其 振 態 頻 率 識 別 結 果 亦 說 明 單 系 統 一 維 構 件 分 析 模 型 可 充 分 解 釋<br />

所 有 鋼 纜 的 振 動 行 為 。 除 鋼 纜 振 態 頻 率 外 , 三 根 鋼 纜 之 振 幅 譜 內 亦 可 同 時 識 別 出 其 他 的 系 統 頻 率 ,<br />

交 互 比 對 可 知 部 分 系 統 頻 率 共 同 存 在 於 三 根 鋼 纜 之 振 幅 譜 , 如 0.61 Hz、1.53 Hz, 再 與 橋 面 板 垂 直<br />

向 微 振 量 測 之 傅 利 葉 振 幅 譜 比 對 可 知 , 此 兩 組 數 值 為 橋 體 的 振 態 頻 率 。 事 實 上 , 這 種 橋 體 振 態 頻 率<br />

常 見 於 鋼 纜 微 振 量 測 之 頻 率 識 別 , 一 般 位 於 低 頻 區 域 , 其 與 鋼 纜 振 態 頻 率 並 不 相 互 干 擾 。 前 述 鋼 纜<br />

振 態 與 橋 體 振 態 之 識 別 頻 率 於 鋼 纜 微 振 訊 號 振 幅 譜 的 分 佈 狀 態 不 僅 可 見 於 集 鹿 斜 張 橋 , 亦 常 見 於 其<br />

他 傳 統 斜 張 橋 , 說 明 傳 統 斜 張 橋 鋼 纜 的 振 動 行 為 應 均 可 以 單 系 統 一 維 構 件 分 析 模 型 進 行 模 擬 分 析 。<br />

-483-


圖 3 集 鹿 橋 L33 鋼 纜 之 傅 利 葉 振 幅 譜<br />

圖 4 集 鹿 橋 L17 鋼 纜 之 傅 利 葉 振 幅 譜<br />

圖 5 集 鹿 橋 L01 鋼 纜 之 傅 利 葉 振 幅 譜<br />

圖 6 愛 蘭 橋 E9 鋼 纜 之 傅 利 葉 振 幅 譜<br />

圖 7 愛 蘭 橋 E5 鋼 纜 之 傅 利 葉 振 幅 譜<br />

圖 8 愛 蘭 橋 E1 鋼 纜 之 傅 利 葉 振 幅 譜<br />

對 比 於 集 鹿 斜 張 橋 鋼 纜 的 振 幅 譜 , 愛 蘭 橋 鋼 纜 的 振 幅 譜 雖 亦 顯 現 清 晰 尖 峰 值 , 但 部 分 鋼 纜 的 振<br />

態 頻 率 分 佈 卻 無 法 以 前 述 頻 率 關 係 進 行 說 明 。 本 文 首 先 取 三 根 具 代 表 性 鋼 纜 說 明 愛 蘭 橋 鋼 纜 振 態 頻<br />

率 獨 特 之 處 , 依 序 為 該 橋 的 長 、 中 、 短 三 根 鋼 纜 , 編 號 為 E9、E5、E1, 所 在 位 置 如 圖 1 紅 色 三 角<br />

標 所 示 。 圖 6 至 圖 8 依 序 繪 出 三 根 鋼 纜 微 振 訊 號 之 傅 利 葉 振 幅 譜 , 比 對 三 個 振 幅 譜 可 明 顯 發 現 位 於<br />

低 頻 區 域 的 四 個 橋 體 振 態 頻 率 , 其 依 序 為 0.86Hz、1.41 Hz、2.43 Hz、3.59(3.64) Hz。 濾 除 橋 體 振<br />

-484-


態 頻 率 後 , 由 圖 6 可 清 楚 識 別 出 E9 長 鋼 纜 前 三 個 振 態 頻 率 , 依 序 為 2.32Hz、4.64Hz、6.93Hz, 由<br />

頻 率 等 間 距 的 變 化 趨 勢 可 知 ,E9 長 鋼 纜 之 振 動 行 為 應 為 弦 模 式 。 除 橋 體 振 態 頻 率 外 , 圖 7 可 清 楚 識<br />

別 出 E5 中 鋼 纜 前 兩 個 振 態 頻 率 , 依 序 為 3.26 Hz、6.50 Hz, 呈 近 似 等 間 距 變 化 關 係 , 研 判 屬 於 弦 振<br />

動 模 式 之 振 態 頻 率 ; 隨 後 較 清 晰 頻 率 為 10.01Hz, 因 無 近 似 等 間 距 變 化 關 係 , 初 判 應 非 該 鋼 纜 的 振<br />

態 頻 率 。 圖 8 為 E1 短 鋼 纜 之 振 幅 譜 , 濾 除 橋 體 振 態 頻 率 後 , 可 看 出 一 個 清 晰 頻 率 5.43Hz 與 兩 組 成<br />

對 頻 率 (10.43Hz、11.43Hz)、(16.07Hz、16.71Hz), 但 無 法 確 認 何 者 為 鋼 纜 振 態 頻 率 。 為 找 出 鋼 纜<br />

振 態 頻 率 , 首 先 將 前 述 三 組 頻 率 配 對 成 兩 組 較 可 能 的 鋼 纜 振 態 頻 率 排 列 , 分 別 為 5.43Hz、10.43Hz、<br />

16.07Hz 及 5.43Hz、11.45Hz、16.71Hz, 從 這 兩 組 配 對 的 頻 率 變 化 趨 勢 來 看 , 頻 率 間 的 關 係 既 非 等<br />

間 距 之 弦 模 式 , 也 非 間 距 逐 漸 增 加 之 拉 力 梁 模 式 , 亦 無 因 重 力 垂 度 影 響 造 成 基 本 振 態 頻 率 放 大 的 狀<br />

況 。 若 視 鋼 纜 為 單 系 統 的 一 維 構 件 , 這 兩 組 頻 率 均 不 符 合 我 們 所 要 尋 找 的 鋼 纜 振 態 頻 率 。 再 仔 細 檢<br />

視 圖 8 振 幅 譜 兩 組 成 對 頻 率 的 尖 錐 形 狀 可 研 判 其 應 為 系 統 頻 率 , 而 非 如 一 般 外 力 集 中 頻 率 之 單 一 脈<br />

衝 形 狀 ; 但 其 同 時 亦 非 橋 體 振 態 頻 率 , 因 在 其 它 鋼 纜 的 振 幅 譜 並 無 顯 現 相 同 的 尖 峰 頻 率 。 若 假 設 第<br />

一 個 清 晰 頻 率 5.43Hz 為 E1 短 鋼 纜 的 基 本 振 態 頻 率 , 並 以 等 間 距 關 係 概 估 第 二 與 第 三 振 態 的 頻 率 ,<br />

其 數 值 依 序 為 10.86Hz、16.29Hz, 分 別 落 在 兩 組 成 對 頻 率 之 間 。 因 兩 組 成 對 頻 率 為 系 統 頻 率 , 且 位<br />

於 振 態 頻 率 估 算 值 的 兩 側 , 故 研 判 成 對 頻 率 之 一 可 能 為 鋼 纜 振 態 頻 率 , 偏 離 估 算 位 置 的 原 因 應 是 來<br />

自 於 相 對 應 另 一 組 系 統 頻 率 的 影 響 , 此 行 為 類 似 一 個 主 系 統 受 到 次 系 統 調 諧 影 響 所 產 生 的 頻 率 耦 合<br />

現 象 。<br />

為 檢 視 此 一 耦 合 現 象 是 否 亦 發 生 於 其 它 鋼 纜 , 首 先 再 次 詳 細 檢 視 E5 與 E9 兩 根 鋼 纜 之 振 幅 譜 ,<br />

重 點 在 於 高 頻 區 域 。 為 利 於 說 明 , 將 大 於 7Hz 區 域 切 割 出 來 , 並 將 振 幅 的 尺 度 放 大 , 如 圖 9 至 10<br />

所 示 。 由 圖 9 可 發 現 前 述 E5 中 鋼 纜 的 一 個 識 別 頻 率 10.01Hz 與 另 一 識 別 頻 率 9.17Hz 成 對 出 現 , 除<br />

此 之 外 , 尚 有 一 組 較 高 頻 率 (15.97Hz、17.45Hz) 也 成 對 出 現 。 類 似 情 形 亦 可 見 於 E9 長 鋼 纜 的 振 幅 譜 ,<br />

圖 10 顯 示 高 頻 區 域 可 找 出 一 組 成 對 頻 率 (8.83Hz、9.45Hz)。 綜 合 三 根 鋼 纜 振 幅 譜 之 識 別 頻 率 可 歸 納<br />

出 一 個 現 象 , 不 論 鋼 纜 的 長 短 , 在 高 頻 區 域 ( 約 大 於 8Hz) 均 可 找 出 成 對 頻 率 。 除 了 前 述 三 根 鋼 纜 外 ,<br />

本 研 究 針 對 同 一 扇 其 它 鋼 纜 的 振 幅 譜 進 行 詳 細 搜 尋 , 主 要 對 象 為 成 對 系 統 頻 率 。 表 1 所 示 為 各 鋼 纜<br />

不 含 橋 體 振 態 頻 率 的 其 他 識 別 頻 率 , 成 對 出 現 的 系 統 頻 率 以 分 隔 線 標 示 在 估 計 對 應 的 鋼 纜 振 態 階 數<br />

位 置 。 表 列 資 料 共 計 有 18 根 鋼 纜 的 前 四 振 態 頻 率 值 , 資 料 顯 示 超 過 一 半 以 上 鋼 纜 的 振 幅 譜 可 找 出<br />

成 對 系 統 頻 率 ,E1 鋼 纜 有 兩 組 成 對 系 統 頻 率 , 成 對 系 統 頻 率 均 發 生 在 基 本 振 態 之 後 , 惟 所 對 應 振 態<br />

階 數 於 各 鋼 纜 並 不 相 同 。 若 以 基 本 振 態 頻 率 為 基 礎 , 假 設 等 間 距 的 關 係 概 估 成 對 系 統 頻 率 所 對 應 振<br />

態 頻 率 值 , 則 其 數 值 約 落 在 兩 個 區 域 , 依 序 是 9.0Hz 至 11.6Hz 與 15.6Hz 至 16.3Hz 之 間 。 綜 合 這 些<br />

結 果 可 知 , 成 對 系 統 頻 率 的 出 現 與 鋼 纜 的 位 置 、 長 度 及 振 態 階 數 並 無 明 確 的 關 係 , 但 似 乎 主 要 與 振<br />

態 頻 率 數 值 有 關 , 此 再 次 說 明 成 對 頻 率 應 是 來 自 於 鋼 纜 系 統 與 另 一 個 次 系 統 振 動 的 耦 合 效 應 。 同 時<br />

圖 9 愛 蘭 橋 E5 鋼 纜 之 局 部 放 大 傅 利 葉 振 幅 譜<br />

圖 10 愛 蘭 橋 E9 鋼 纜 之 局 部 放 大 傅 利 葉 振 幅 譜<br />

-485-


表 1 愛 蘭 橋 P1G2 各 鋼 纜 振 態 頻 率<br />

Cable<br />

Modal frequencies (Hz)<br />

f1<br />

f2<br />

f3<br />

f4<br />

E1 5.43 10.43 11.43 16.07 16.71 .<br />

E2 4.54 8.60 9.42 13.71 18.64<br />

E3 3.88 7.72 11.71 .<br />

E4 3.53 . 10.59 .<br />

E5 3.26 6.50 9.17 10.10 13.20<br />

E6 2.90 5.75 8.61 10.58 11.79<br />

E7 2.63 5.26 7.90 10.50<br />

E8 2.41 4.86 7.25 9.39 10.28<br />

E9 2.32 4.64 6.93 8.83 9.45<br />

E10 5.46 9.54 11.23 . .<br />

E11 4.46 7.82 9.13 13.42 18.04<br />

E12 7.83 11.64 14.54 16.65<br />

E13 3.51 6.96 10.47 .<br />

E14 3.11 6.29 9.14 10.02 12.57<br />

E15 2.88 5.74 8.58 11.32<br />

E16 2.60 5.21 7.77 10.04 11.10<br />

E17 2.39 4.80 7.16 9.52<br />

E18 2.27 4.54 6.77 9.15<br />

因 為 此 一 現 象 出 現 在 所 有 的 鋼 纜 中 , 由 鋼 纜 的 組 成 研 判 此 一 次 系 統 應 是 來 自 鋼 纜 兩 端 的 錨 錠 裝 置 。<br />

以 往 因 傳 統 斜 張 橋 的 鋼 纜 長 度 相 當 大 , 錨 錠 裝 置 的 長 度 與 質 量 相 對 極 其 有 限 , 故 常 被 視 為 單 點 的 邊<br />

界 束 制 條 件 。 但 對 於 愛 蘭 橋 而 言 , 其 鋼 纜 長 度 相 對 較 短 , 特 別 是 短 鋼 纜 部 分 ; 同 時 雪 上 加 霜 的 是 ,<br />

其 所 採 用 之 錨 錠 裝 置 內 有 填 充 水 泥 漿 體 而 增 加 不 少 重 量 。 故 整 組 錨 錠 裝 置 若 以 單 系 統 之 一 維 構 件 簡<br />

化 模 型 模 擬 可 能 無 法 反 應 真 實 狀 況 , 應 依 其 組 成 構 造 進 行 系 統 模 擬 , 將 其 視 為 另 一 個 次 系 統 連 接 於<br />

鋼 纜 的 兩 端 。 為 驗 證 此 一 研 判 結 果 , 本 研 究 隨 後 針 愛 蘭 橋 的 鋼 纜 與 兩 端 錨 錠 裝 置 進 行 有 限 元 素 分 析 。<br />

四 、 鋼 纜 與 錨 錠 裝 置 有 限 元 素 分 析<br />

愛 蘭 橋 所 有 鋼 纜 均 由 31 根 15.2mm 之 七 股 鋼 鉸 線 平 行 組 成 , 外 套 高 密 度 聚 乙 稀 套 管 , 內 部 灌<br />

注 聚 氨 酯 漿 材 , 如 圖 11 所 示 。 鋼 纜 兩 端 以 OVM250 錨 錠 裝 置 與 橋 塔 及 主 梁 結 合 ,OVM250 屬 於<br />

HYDIN 型 式 錨 錠 裝 置 , 主 要 是 以 夾 片 傳 遞 鋼 鉸 線 應 力 , 鋼 鉸 線 外 套 分 離 管 (separation tube), 並 以<br />

阻 尼 器 (damper) 束 制 於 外 導 管 (guide pipe) 前 端 , 如 圖 12 所 示 。 分 離 管 與 保 護 套 管 (transition pipe)<br />

之 間 灌 注 無 收 縮 水 泥 砂 漿 , 如 圖 13 所 示 。 本 研 究 以 E9 鋼 纜 為 例 進 行 有 限 元 素 分 析 ,E9 鋼 纜 錨 錠 端<br />

之 間 全 長 為 31.26m, 錨 錠 端 至 阻 尼 器 距 離 於 橋 面 端 長 4.00m, 於 橋 塔 端 長 2.57m, 鋼 纜 單 位 長 度 重<br />

40.2kgf/m。<br />

為 探 討 鋼 纜 振 態 頻 率 因 兩 端 錨 錠 裝 置 模 擬 機 制 的 差 異 所 造 成 的 影 響 , 本 研 究 利 用 SAP2000 程 式<br />

建 構 分 析 模 型 , 如 圖 14 所 示 ; 以 梁 元 素 分 別 模 擬 鋼 纜 與 錨 錠 裝 置 兩 構 件 , 兩 者 之 間 以 線 性 彈 簧 元<br />

-486-


素 連 接 , 假 設 兩 構 件 主 要 連 接 點 是 經 由 阻 尼 器 。 模 擬 分 析 乃 是 經 由 調 整 彈 簧 元 素 之 勁 度 值 觀 察 鋼 纜<br />

系 統 與 錨 錠 系 統 振 態 頻 率 的 改 變 。 當 彈 簧 勁 度 值 極 小 時 , 兩 系 統 近 似 相 互 獨 立 ; 而 當 彈 簧 勁 度 值 逐<br />

次 增 大 時 , 以 錨 錠 系 統 振 動 為 主 的 振 態 頻 率 會 變 大 , 同 時 接 近 錨 錠 系 統 振 態 頻 率 的 鋼 纜 系 統 振 態 頻<br />

率 也 會 受 影 響 。 表 2 所 列 為 彈 簧 勁 度 值 由 2.5*10 4 N/m 遞 增 至 2.5*10 5 N/m 之 各 振 態 頻 率 值 , 黑 色 字<br />

體 為 以 鋼 纜 系 統 振 動 為 主 的 振 態 頻 率 值 , 紅 色 字 體 表 示 以 錨 錠 系 統 振 動 為 主 的 振 態 頻 率 值 , 而 黃 色<br />

底 則 為 主 要 受 干 擾 的 鋼 纜 振 態 。 觀 察 表 列 的 鋼 纜 系 統 第 二 振 態 頻 率 值 之 變 化 趨 勢 可 發 現 , 其 數 值 隨<br />

彈 簧 勁 度 值 漸 增 而 些 微 變 大 (11.00Hz→11.11Hz), 但 於 彈 簧 勁 度 值 增 至 4.75*10 4 N/m 時 , 其 數 值 突<br />

然 變 小 至 10.83Hz, 此 時 錨 錠 系 統 的 振 態 頻 率 值 已 接 近 鋼 纜 系 統 第 二 振 態 頻 率 初 始 值 11.00Hz; 隨<br />

後 , 其 數 值 再 緩 慢 增 大 回 至 初 始 值 。 當 彈 簧 勁 度 值 增 至 2.5*10 5 N/m 時 , 第 二 振 態 頻 率 值 已 脫 離 錨 錠<br />

系 統 的 影 響 , 但 第 三 振 態 頻 率 值 則 逐 漸 受 其 影 響 。 此 一 分 析 結 果 說 明 鋼 纜 系 統 振 態 頻 率 值 可 能 會 受<br />

錨 錠 系 統 的 干 擾 , 亦 證 實 前 節 針 對 成 對 識 別 頻 率 成 因 的 研 判 。 此 一 有 限 元 素 分 析 仍 存 在 兩 個 問 題 須<br />

進 一 步 研 析 , 一 是 成 對 頻 率 識 別 值 與 分 析 值 並 不 一 致 , 其 次 是 並 無 兩 組 成 對 頻 率 同 時 發 生 , 研 判 應<br />

是 與 錨 錠 系 統 的 模 擬 有 關 。<br />

圖 11 鋼 纜 斷 面 組 成 圖<br />

圖 12 OVM 錨 錠 裝 置 示 意 圖<br />

圖 13 錨 頭 灌 漿<br />

圖 14 SAP2000 鋼 纜 模 型 示 意 圖<br />

-487-


表 2 SAP2000 模 擬 分 析 之 各 振 態 頻 率 值<br />

Mode<br />

Modal Frequency (Hz) for different connecting spring stiffness k (N/m)<br />

0 2.5*10 4 4.5*10 4 4.75*10 4 5*10 4 7.5*10 4 10 5 2.5*10 5<br />

1 5.49 5.49 5.49 5.49 5.49 5.50 5.50 5.50<br />

2 11.00 10.03 10.66 10.72 10.77 10.94 10.96 10.98<br />

3 16.57 10.04 10.75 10.83 10.91 11.69 12.40 15.30<br />

4 22.23 11.03 11.11 11.14 11.18 11.80 12.51 15.83<br />

5 28.00 16.62 16.63 16.63 16.64 16.66 16.70 17.25<br />

6 33.90 22.30 22.32 22.32 22.33 22.35 22.38 22.58<br />

7 39.96 28.12 28.14 28.14 28.14 28.17 28.19 28.36<br />

8 46.21 34.10 34.12 34.12 34.12 34.15 34.17 34.33<br />

9 52.66 40.26 40.28 40.28 40.28 40.30 40.33 40.47<br />

10 59.34 46.63 46.65 46.65 46.65 46.67 46.69 46.82<br />

五 、 結 論<br />

透 過 根 據 愛 蘭 矮 塔 斜 張 橋 之 鋼 纜 微 振 訊 號 所 進 行 的 振 態 頻 率 識 別 , 同 時 經 由 有 限 元 素 模 型 交 叉<br />

分 析 比 對 , 本 研 究 驗 證 較 短 鋼 纜 的 振 態 頻 率 確 實 會 受 到 錨 錠 系 統 振 動 的 干 擾 而 造 成 識 別 上 的 混 淆 。<br />

同 時 受 影 響 的 振 態 頻 率 與 鋼 纜 的 位 置 、 長 度 及 振 態 階 數 無 關 , 主 要 乃 為 錨 錠 系 統 的 振 態 頻 率 所 控<br />

制 。 事 實 上 , 此 一 頻 率 干 擾 現 象 可 能 亦 存 在 於 傳 統 斜 張 橋 的 鋼 纜 , 以 往 未 被 發 現 應 是 錨 錠 系 統 的 振<br />

態 頻 率 一 般 均 相 當 大 , 發 生 干 擾 的 鋼 纜 振 態 階 數 較 高 , 因 此 頻 率 識 別 作 業 時 常 被 忽 略 。 但 對 於 如 矮<br />

塔 斜 張 橋 的 較 短 斜 張 鋼 纜 , 由 於 其 振 態 頻 率 較 易 受 錨 錠 系 統 振 動 的 影 響 , 因 此 在 選 用 振 態 頻 率 進 行<br />

索 力 計 算 時 須 特 別 小 心 交 叉 判 讀 。<br />

参 考 文 献<br />

[1] 李 政 寬 , 陳 建 州 , 周 智 傑 , 張 國 鎮 . 斜 張 鋼 纜 微 振 訊 號 有 限 元 素 法 解 析 . 中 國 土 木 水 利 工 程 學 刊 ,2006, 18(2):<br />

279-288.<br />

[2] 陳 建 州 , 李 政 寬 , 周 智 傑 , 張 國 鎮 . 斜 張 鋼 纜 索 力 分 析 之 參 數 研 究 . 中 國 土 木 水 利 工 程 學 刊 , 2006, 18(4):<br />

337-345.<br />

[3] 陳 建 州 , 吳 文 華 , 廖 釗 銨 . 以 鋼 纜 微 振 訊 號 進 行 鋼 纜 與 橋 體 之 振 態 頻 率 識 別 . 中 國 土 木 水 利 工 程 學 刊 , 2008,<br />

20(4): 415-427.<br />

-488-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

体 外 预 应 力 加 固 T 梁 桥 的 抗 剪 性 能 分 析<br />

项 贻 强 朱 小 辉 楼 铁 炯 吴 强 强<br />

( 浙 江 大 学 土 木 工 程 系 , 浙 江 杭 州 310058)<br />

摘 要 : 以 某 座 钢 筋 混 凝 土 T 梁 桥 采 用 体 外 预 应 力 进 行 加 固 改 造 和 提 高 荷 载 等 级 为 研 究 背 景 , 采<br />

用 通 用 有 限 元 软 件 ABAQUS, 考 虑 几 何 、 材 料 、 边 界 条 件 等 多 重 非 线 性 , 建 立 有 限 元 模 型 , 分 析 加<br />

固 前 后 T 梁 桥 的 抗 剪 性 能 , 评 价 加 固 效 果 。 分 析 结 果 表 明 : 采 用 体 外 预 应 力 加 固 , 提 高 了 T 梁 桥 支<br />

点 附 近 梁 截 面 的 抗 剪 性 能 , 而 对 跨 中 附 近 梁 截 面 的 抗 剪 性 能 提 高 并 不 明 显 。<br />

关 键 词 : 桥 梁 工 程 钢 筋 混 凝 土 T 梁 桥 体 外 预 应 力 加 固 非 线 性 分 析 抗 剪 性 能<br />

ANALYSIS OF SHEAR PERFORMANCE OF CONCRETE T-BEAM BRIDGE<br />

STRENGTHENED BY EXTERNAL PRESTRESSED TENDON<br />

Xiang Yi-qiang 2<br />

Zhu Xiao-hui Lou Tie-jiong and Wu Qiang-qiang<br />

Department of Civil Engineering, Zhejiang <strong>University</strong>, Hangzhou 310058, China<br />

Abstract: Taking a typical reinforced concrete T-beam bridge as background, which was strengthened by<br />

external prestressed tendon for upgrading traffic load, the nonlinear finite element models of the external<br />

prestressed concrete bridge is established by using finite element analytical software ABAQUS. Considering<br />

effects of the material and geometry nonlinearity, boundary conditions in this model, the shear performance of<br />

T-beam bridge has been analyzed under dead load and live load before and after strengthening the bridge to<br />

evaluated strengthening effectiveness of the bridge. <strong>The</strong> results show that the shear performance in the beam<br />

sections nearby supporting point is improved, but the one in the beam sections nearby mid-section is not<br />

significantly improved.<br />

Keywords: Bridge engineering, Reinforced concrete T-beam bridge, External prestressed force, Strengthening,<br />

Nonlinear analysis, Shear performance.<br />

一 、 引 言<br />

体 外 预 应 力 加 固 技 术 已 广 泛 运 用 于 桥 梁 加 固 工 程 中 。 工 程 实 践 表 明 , 体 外 预 应 力 加 固 技 术 具 有<br />

如 下 优 点 [1] : 在 自 重 很 小 的 情 况 下 , 能 较 大 幅 度 地 改 善 和 调 整 原 结 构 的 受 力 情 况 , 提 高 承 重 结 构 的<br />

刚 度 、 抗 裂 性 ; 所 需 设 备 简 单 , 人 力 投 入 少 , 经 济 效 益 明 显 ; 对 结 构 营 运 影 响 较 小 , 对 原 结 构 损 伤<br />

较 小 。 目 前 对 体 外 预 应 力 混 凝 土 桥 梁 抗 剪 性 能 的 深 入 研 究 总 的 来 说 并 不 多 。 同 济 大 学 李 国 平 教 授 等<br />

[2]<br />

做 了 一 系 列 体 外 预 应 力 混 凝 土 模 型 梁 试 验 , 研 究 其 剪 切 性 能 和 抗 剪 承 载 力 [3] ; 对 于 体 外 预 应 力 加<br />

固 桥 的 抗 剪 性 能 研 究 , 不 少 学 者 是 采 用 平 面 杆 系 做 线 弹 性 分 析 [4] , 而 绝 大 多 数 的 加 固 方 案 设 计 则 是 按<br />

照 规 范 公 式 来 近 似 计 算 的 [5] 。 本 文 以 某 座 钢 筋 混 凝 土 T 梁 桥 采 用 体 外 预 应 力 进 行 加 固 改 造 和 提 高 荷<br />

2<br />

作 者 简 介 : 项 贻 强 (1959-), 男 , 浙 江 杭 州 人 , 教 授 、 博 士 生 导 师 , 主 要 从 事 桥 梁 的 设 计 理 论 、 健 康 监 测 与 养 护 管 理 、 悬 浮 隧 道 、 古 桥<br />

保 护 的 研 究 。e-mail: xiangyiq@zju.edu.cn<br />

-489-


载 等 级 为 研 究 背 景 , 采 用 通 用 有 限 元 软 件 ABAQUS, 考 虑 几 何 、 材 料 、 边 界 条 件 等 多 重 非 线 性 , 建<br />

立 有 限 元 模 型 , 分 析 加 固 前 后 T 梁 桥 的 抗 剪 性 能 , 评 价 加 固 效 果 。<br />

[5]<br />

二 、 工 程 背 景<br />

某 钢 筋 混 凝 土 T 型 简 支 梁 桥 , 桥 面 宽 度 为 9m( 机 动 车 道 7m, 人 行 道 两 边 各 1m), 上 部 结 构 由<br />

5 片 跨 径 14.1m 的 钢 筋 混 凝 土 T 梁 组 成 。 桥 面 铺 装 为 钢 筋 混 凝 土 (7.5、11、7.5cm) 和 3cm 的 沥 青<br />

混 凝 土 面 层 。 该 桥 原 设 计 等 级 为 汽 -13、 拖 -60。 旧 桥 混 凝 土 设 计 标 号 为 25 号 , 普 通 钢 筋 屈 服 强<br />

度 340MPa。T 梁 典 型 截 面 尺 寸 和 配 筋 如 图 1。<br />

图 1 截 面 尺 寸 及 配 筋 ( 单 位 : cm ) 图 2 体 外 预 应 力 筋 布 置 图 ( 单 位 : cm )<br />

由 于 原 有 公 路 的 技 术 标 准 低 , 加 之 目 前 交 通 量 的 增 加 和 汽 车 载 重 的 增 加 , 上 述 旧 桥 已 不 能 满 足<br />

承 载 力 要 求 的 。 考 虑 到 资 金 和 材 料 资 源 及 时 间 的 限 制 , 也 不 可 能 全 部 拆 除 并 新 建 , 只 能 考 虑 投 资 较<br />

少 、 工 期 时 间 短 且 能 提 高 承 载 力 的 桥 梁 加 固 技 术 予 以 加 固 。<br />

下 部 结 构 墩 柱 及 盖 梁 是 在 原 桥 位 上 游 侧 重 新 设 计 建 造 的 , 荷 载 等 级 : 汽 -20、 挂 -100。 采 用<br />

体 外 预 应 力 钢 筋 加 固 改 造 技 术 , 确 为 一 种 简 单 易 行 且 能 与 新 建 下 部 结 构 荷 载 相 适 应 的 有 效 方 法 。<br />

本 桥 体 外 预 应 力 筋 采 用 粗 钢 筋 。 体 外 索 水 平 筋 取 为 2φ28, 斜 筋 取 为 2φ32, 均 为 冷 拉 Ⅲ 级 钢 ( 单<br />

控 ), 其 标 准 强 度 为 530MPa。 张 拉 控 制 应 力 为 σ k =0.8R b y=424MPa, 有 效 预 应 力 经 计 算 为 317MPa。<br />

体 外 预 应 力 筋 至 T 梁 底 的 距 离 C=10cm。 两 垫 板 中 心 之 间 的 水 平 距 离 为 938cm, 上 锚 固 点 至 垫 板 中<br />

心 的 水 平 距 离 为 205cm, 滑 块 与 梁 底 之 间 的 摩 擦 系 数 为 0.11。<br />

三 、 有 限 元 模 型<br />

考 虑 到 建 立 整 桥 有 限 元 模 型 进 行 多 重 非 线 性 计 算 时 , 单 元 数 目 将 非 常 大 , 普 通 PC 电 脑 一 般 难<br />

以 完 成 这 样 庞 大 的 计 算 量 , 因 此 本 文 引 入 横 向 分 布 系 数 , 将 多 梁 的 计 算 转 化 成 只 对 单 片 最 不 利 的 梁<br />

进 行 受 力 分 析 , 以 减 小 计 算 量 。<br />

考 虑 材 料 非 线 性 。 混 凝 土 应 力 - 应 变 采 用 Hognestad 应 力 - 应 变 曲 线 , 普 通 钢 筋 采 用 理 想 弹 塑 性<br />

模 型 , 体 外 预 应 力 筋 采 用 线 性 强 化 模 型 。<br />

采 用 分 离 式 钢 筋 混 凝 土 模 型 , 通 过 “ 拉 伸 硬 化 ” 考 虑 钢 筋 和 混 凝 土 的 相 互 作 用 , 钢 筋 单 元 嵌 入<br />

到 混 凝 土 单 元 中 。 裂 缝 行 为 采 用 弥 散 裂 缝 模 型 模 拟 。<br />

体 外 预 应 力 钢 筋 与 转 向 块 之 间 的 接 触 用 库 伦 摩 擦 来 模 拟 , 摩 擦 系 数 按 文 献 [1] 取 0.11, 考 虑 边 界<br />

条 件 的 非 线 性 。<br />

为 减 小 工 作 量 , 根 据 结 构 和 荷 载 以 及 边 界 条 件 关 于 梁 轴 线 正 对 称 的 特 点 , 只 建 半 片 梁 , 混 凝 土<br />

部 分 离 散 为 9966 个 8 节 点 六 面 体 线 性 减 缩 积 分 单 元 (C3D8R), 普 通 钢 筋 采 用 线 性 空 间 桁 架 单 元<br />

(T3D2), 离 散 为 2970 个 单 元 , 体 外 预 应 力 粗 钢 筋 采 用 空 间 线 性 梁 单 元 (B31) 模 拟 。 有 限 元 模 型<br />

如 图 3。<br />

-490-


图 3 半 片 梁 的 有 限 元 模 型<br />

四 、 分 析 的 步 骤 及 荷 载 作 用 方 式<br />

ABAQUS/Standard 应 用 Newton-Raphson 算 法 获 得 非 线 性 问 题 的 解 答 。 对 于 加 固 前 的 普 通 钢 筋<br />

混 凝 土 结 构 , 可 以 将 分 析 过 程 分 成 两 个 分 析 步 , 第 一 个 分 析 步 施 加 恒 载 , 第 二 个 分 析 步 施 加 活 载 ;<br />

对 于 加 固 后 的 体 外 预 应 力 结 构 , 将 分 析 过 程 分 成 三 个 分 析 步 , 第 一 个 分 析 步 施 加 恒 载 , 第 二 个 分 析<br />

步 施 加 预 应 力 , 第 三 个 分 析 步 施 加 活 载 。<br />

本 文 所 采 用 的 荷 载 组 合 为 : 恒 载 + 挂 —100 级 荷 载 。 挂 车 荷 载 采 用 横 向 分 布 系 数 进 行 折 减 , 对 最<br />

不 利 梁 进 行 计 算 分 析 。 挂 车 荷 载 的 纵 桥 向 布 置 位 置 见 图 4。<br />

五 、 分 析 结 果<br />

图 4 挂 车 荷 载 沿 桥 纵 向 布 置 位 置 ( 单 位 : mm )<br />

为 了 在 图 中 表 述 方 便 , 加 固 前 的 普 通 钢 筋 混 凝 土 桥 简 称 为 RC(Reinforced concrete) 桥 , 而 加<br />

固 后 的 体 外 预 应 力 混 凝 土 桥 简 称 为 EPC(Externally prestressed concrete) 桥 。<br />

主 梁 控 制 截 面 A 和 B 位 置 如 图 5 所 示 , 控 制 截 面 上 控 制 点 位 置 见 图 6 所 示 ,A、B 截 面 上 的 控<br />

制 点 分 别 记 作 a、b 点 。 分 析 结 果 中 规 定 拉 应 力 为 正 号 , 压 应 力 为 负 号 。<br />

有 限 元 分 析 结 果 分 别 如 图 7~14 所 示 。<br />

A B<br />

815 785 12200<br />

13800<br />

图 5 主 梁 控 制 截 面 A和 B位 置 ( 单 位 : mm )<br />

图 6 控 制 截 面 上 控 制 点 位 置 ( 单 位 : mm )<br />

-491-


图 7 RC 桥 在 恒 载 和 挂 车 荷 载 共 同 作 用 下<br />

最 大 主 轴 塑 性 应 变 分 布 云 图<br />

图 8 EPC 桥 在 恒 载 和 挂 车 荷 载 共 同 作 用 下 最<br />

大 主 轴 塑 性 应 变 分 布 云 图<br />

1000<br />

RC 桥<br />

EPC 桥<br />

1000<br />

RC 桥<br />

EPC 桥<br />

800<br />

800<br />

梁 高 ( mm)<br />

600<br />

400<br />

梁 高 ( mm)<br />

600<br />

400<br />

200<br />

200<br />

0<br />

0.0 0.2 0.4 0.6 0.8 1.0 1.2<br />

第 一 主 应 力 ( MPa)<br />

0<br />

0.0 0.5 1.0 1.5 2.0<br />

第 一 主 应 力 (MPa)<br />

图 9 恒 载 作 用 下 主 梁 控 制 截 面 A 第 一 主 应<br />

力 沿 梁 高 变 化 状 况<br />

图 10 挂 车 与 恒 载 共 同 作 用 下 主 梁 控 制 截 面<br />

A 第 一 主 应 力 沿 梁 高 变 化 状 况<br />

1000<br />

RC 桥<br />

EPC 桥<br />

1000<br />

RC 桥<br />

EPC 桥<br />

800<br />

800<br />

梁 高 ( mm)<br />

600<br />

400<br />

梁 高 ( mm)<br />

600<br />

400<br />

200<br />

200<br />

0<br />

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0<br />

第 一 主 应 力 ( MPa)<br />

图 11 恒 载 作 用 下 主 梁 控 制 截 面 B 第 一 主<br />

应 力 沿 梁 高 变 化 状 况<br />

0<br />

-0.5 0.0 0.5 1.0 1.5 2.0 2.5<br />

第 一 主 应 力 ( MPa)<br />

图 12 挂 车 与 恒 载 共 同 作 用 下 主 梁 控 制 截 面<br />

B 第 一 主 应 力 沿 梁 高 变 化 状 况<br />

2.2<br />

2.0<br />

RC 桥<br />

EPC 桥<br />

2.2<br />

2.0<br />

RC 桥<br />

EPC 桥<br />

1.8<br />

1.8<br />

第 一 主 应 力 ( MPa)<br />

1.6<br />

1.4<br />

1.2<br />

1.0<br />

0.8<br />

0.6<br />

0.4<br />

第 一 主 应 力 ( MPa)<br />

1.6<br />

1.4<br />

1.2<br />

1.0<br />

0.8<br />

0.6<br />

0.4<br />

0.2<br />

0.2<br />

0.0<br />

0.0<br />

0.0 0.2 0.4 0.6 0.8 1.0<br />

挂 车 荷 载 系 数 λ<br />

图 13 挂 车 与 恒 载 共 同 作 用 下 控 制 点 a 第<br />

一 主 应 力 随 挂 车 荷 载 系 数 λ 变 化 状 况<br />

0.0 0.2 0.4 0.6 0.8 1.0<br />

挂 车 荷 载 系 数 λ<br />

图 14 挂 车 与 恒 载 共 同 作 用 下 控 制 点 b 第 一<br />

主 应 力 随 挂 车 荷 载 系 数 λ 变 化 状 况<br />

-492-


加 固 前 的 钢 筋 混 凝 土 T 梁 桥 , 在 挂 车 与 恒 载 的 共 同 作 用 下 , 支 座 附 近 截 面 出 现 主 轴 方 向 的 塑 性<br />

应 变 , 说 明 这 些 截 面 上 某 些 点 主 拉 应 力 超 过 混 凝 土 的 开 裂 应 力 而 产 生 裂 缝 。 这 些 开 裂 点 的 分 布 与 梁<br />

轴 线 大 致 呈 45° 角 。 另 外 , 跨 中 截 面 附 近 的 两 侧 混 凝 土 也 因 主 拉 应 力 较 大 而 开 裂 。<br />

采 用 体 外 预 应 力 筋 加 固 后 的 T 梁 桥 , 在 挂 车 与 恒 载 的 共 同 作 用 下 , 三 个 转 向 块 附 近 的 混 凝 土 由<br />

于 局 部 承 压 和 应 力 集 中 , 也 将 产 生 较 大 的 主 拉 应 力 ; 另 外 , 从 两 侧 转 向 块 至 跨 中 截 面 两 侧 混 凝 土 也<br />

将 产 生 较 大 的 主 拉 应 力 , 因 此 必 须 通 过 构 造 措 施 减 少 应 力 集 中 , 分 布 荷 载 , 同 时 增 加 抗 剪 和 混 凝 土<br />

的 抗 拉 , 如 粘 贴 纤 维 材 料 等 。<br />

恒 载 作 用 下 , 控 制 截 面 A 越 靠 近 底 缘 混 凝 土 主 拉 应 力 越 大 , 最 大 值 达 到 1.1MPa; 采 用 体 外 预<br />

应 力 筋 加 固 后 , 截 面 A 的 混 凝 土 主 拉 应 力 分 布 呈 上 下 缘 小 、 中 间 大 的 趋 势 , 最 大 值 达 0.44MPa, 但<br />

都 没 有 达 到 开 裂 应 力 。<br />

挂 车 与 恒 载 的 共 同 作 用 下 , 控 制 截 面 A 在 加 固 前 , 沿 高 度 的 大 部 分 混 凝 土 的 主 拉 应 力 将 达 到 或<br />

超 过 开 裂 应 力 ; 加 固 之 后 沿 梁 高 由 上 至 下 混 凝 土 主 拉 应 力 先 是 减 小 , 然 后 增 大 , 底 缘 处 最 大 值 为<br />

1.7MPa, 但 并 未 开 裂 。<br />

恒 载 作 用 下 , 加 固 前 的 钢 筋 混 凝 土 T 梁 桥 , 控 制 截 面 B 的 混 凝 土 主 拉 应 力 沿 梁 高 由 上 至 下 逐 渐<br />

增 大 , 最 大 值 出 现 在 底 缘 , 为 1.72MPa; 加 固 后 , 混 凝 土 主 拉 应 力 沿 梁 高 分 布 也 基 本 呈 上 下 小 、 中<br />

间 大 的 形 状 , 最 大 值 约 为 0.5MPa。<br />

挂 车 与 恒 载 的 共 同 作 用 下 , 控 制 截 面 B 在 加 固 前 , 大 多 数 点 的 主 拉 应 力 达 到 开 裂 应 力 而 开 裂 ;<br />

加 固 之 后 沿 梁 高 由 上 至 下 混 凝 土 主 拉 应 力 先 是 减 小 , 然 后 增 大 , 底 缘 处 达 到 最 大 值 为 1.8MPa, 距 底<br />

缘 0.6~0.8m 区 域 的 混 凝 土 主 拉 应 力 很 小 , 甚 至 出 现 双 向 受 压 的 情 况 。<br />

挂 车 与 恒 载 的 共 同 作 用 下 , 随 着 挂 车 荷 载 的 增 加 , 加 固 前 控 制 点 a 处 的 混 凝 土 主 拉 应 力 呈 增 大<br />

趋 势 , 最 终 达 到 开 裂 应 力 ; 而 采 用 有 转 向 块 的 体 外 预 应 力 筋 加 固 后 , 控 制 点 a 处 的 混 凝 土 主 拉 应 力<br />

却 随 荷 载 增 加 而 减 小 , 之 后 又 有 所 增 加 。<br />

挂 车 与 恒 载 的 共 同 作 用 下 , 随 着 挂 车 荷 载 的 增 加 , 加 固 前 控 制 点 b 处 的 混 凝 土 主 拉 应 力 呈 增 大<br />

趋 势 , 最 终 达 到 开 裂 应 力 ; 而 采 用 有 转 向 块 的 体 外 预 应 力 筋 加 固 后 , 控 制 点 b 处 的 主 拉 应 力 却 随 荷<br />

载 增 加 而 减 小 , 最 终 小 于 零 , 说 明 此 时 该 点 附 近 的 混 凝 土 有 双 向 受 压 的 趋 势 。<br />

六 、 控 制 截 面 第 一 主 应 力 沿 梁 高 分 布 分 析<br />

从 图 9~ 图 14 可 以 看 出 , 控 制 截 面 上 处 于 中 性 轴 附 近 的 点 的 第 一 主 应 力 , 随 挂 车 荷 载 的 施 加 ,<br />

反 而 呈 减 小 的 趋 势 , 似 乎 和 普 通 钢 筋 混 凝 土 以 及 有 粘 结 预 应 力 结 构 的 情 况 不 同 , 分 析 结 果 是 否 正 确 ,<br />

是 何 原 因 引 起 的 值 得 研 究 。 为 此 , 作 者 进 行 了 相 关 的 对 比 分 析 计 算 。<br />

为 便 于 比 较 , 分 别 分 析 了 体 外 预 应 力 梁 有 3 个 转 向 块 、 体 外 预 应 力 梁 只 有 跨 中 1 个 转 向 块 、 体<br />

外 预 应 力 梁 无 转 向 块 等 3 种 结 构 在 恒 载 加 挂 车 荷 载 下 的 抗 剪 行 为 , 取 控 制 截 面 A 上 的 第 一 主 应 力 。<br />

为 了 减 小 工 作 量 , 模 型 采 用 弹 性 材 料 。 分 析 结 果 如 图 15~ 图 17 所 示 。<br />

通 过 对 比 分 析 可 以 看 出 , 如 果 没 有 转 向 块 或 者 仅 在 跨 中 有 一 个 转 向 块 的 结 构 , 控 制 截 面 上 中 性<br />

轴 附 近 点 的 第 一 主 应 力 , 随 着 挂 车 荷 载 的 施 加 , 确 实 是 呈 增 加 趋 势 的 , 符 合 一 般 简 支 梁 受 剪 时 随 着<br />

剪 切 荷 载 增 加 , 梁 中 性 轴 附 近 点 主 拉 应 力 也 随 之 增 加 的 规 律 。 表 明 在 体 外 预 应 力 进 行 加 固 桥 梁 时 ,<br />

其 内 力 分 布 还 与 两 侧 转 向 块 的 约 束 和 体 外 预 应 力 的 设 置 有 关 。<br />

-493-


1000<br />

恒 载<br />

恒 载 + 挂 车<br />

1000<br />

恒 载<br />

恒 载 + 挂 车<br />

800<br />

800<br />

梁 高 ( mm)<br />

600<br />

400<br />

梁 高 ( mm)<br />

600<br />

400<br />

200<br />

200<br />

0<br />

0.0 0.5 1.0 1.5 2.0<br />

第 一 主 应 力 (MPa)<br />

图 15 3 个 转 向 块 的 体 外 预 应 力 梁 截 面 A 第<br />

一 主 应 力 沿 梁 高 变 化 状 况<br />

1000<br />

0<br />

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5<br />

第 一 主 应 力 (MPa)<br />

图 16 跨 中 有 1 个 转 向 块 的 体 外 预 应 力 梁 截<br />

恒 载<br />

恒 载 + 挂 车<br />

面 A 第 一 主 应 力 沿 梁 高 变 化 状 况<br />

800<br />

梁 高 ( mm)<br />

600<br />

400<br />

200<br />

0<br />

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5<br />

第 一 主 应 力 (MPa)<br />

图 17 无 转 向 块 的 体 外 预 应 力 梁 截 面 A 第 一 主 应 力 沿 梁 高 变 化 状 况<br />

七 、 结 语<br />

1) 采 用 体 外 预 应 力 筋 加 固 简 支 梁 后 , 在 靠 近 梁 支 座 截 面 处 的 主 拉 应 力 得 到 明 显 减 小 , 有 效 地<br />

减 少 了 该 区 域 混 凝 土 斜 裂 缝 的 开 展 。 但 是 转 向 块 的 出 现 , 又 使 得 设 置 转 向 块 处 梁 周 围 混 凝 土 的 应 力<br />

分 布 比 较 复 杂 , 出 现 了 较 大 的 主 拉 应 力 , 同 样 会 引 起 混 凝 土 的 开 裂 , 应 引 起 加 固 设 计 的 注 意 。<br />

2) 采 用 所 提 出 的 体 外 预 应 力 筋 布 置 方 式 对 钢 筋 混 凝 土 T 梁 桥 进 行 加 固 , 对 支 点 附 近 截 面 的 梁<br />

抗 剪 性 能 有 显 著 提 高 , 但 是 这 种 加 固 方 法 并 不 能 有 效 控 制 梁 跨 中 截 面 附 近 由 于 主 拉 应 力 过 大 而 引 起<br />

的 混 凝 土 斜 裂 缝 , 因 此 , 对 梁 跨 中 截 面 附 近 抗 剪 性 能 的 提 高 并 不 理 想 , 建 议 对 该 处 的 混 凝 土 裂 缝 采<br />

用 其 他 加 固 手 段 进 行 加 固 。<br />

3) 采 用 体 外 预 应 力 筋 加 固 混 凝 土 梁 后 , 其 支 座 附 近 梁 截 面 中 性 轴 附 近 点 的 第 一 主 应 力 随 着 挂<br />

车 荷 载 的 施 加 呈 减 小 趋 势 , 通 过 分 析 , 该 内 力 与 转 向 块 的 布 置 和 约 束 有 关 。<br />

参 考 文 献<br />

[1] 康 叶 铭 . 体 外 预 应 力 加 固 机 理 及 应 用 [D]. 南 京 : 河 海 大 学 硕 士 学 位 论 文 , 2006.<br />

[2] 李 国 平 . 体 外 预 应 力 混 凝 土 简 支 梁 剪 切 性 能 试 验 研 究 [J]. 土 木 工 程 学 报 , 2007, 40(2): 58-63.<br />

[3] 李 国 平 , 沈 殷 . 体 外 预 应 力 混 凝 土 简 支 梁 抗 剪 承 载 力 计 算 方 法 [J]. 土 木 工 程 学 报 , 2007, 40(2): 64-69.<br />

[4] 盛 丽 娟 . T 型 刚 构 桥 的 体 外 预 应 力 加 固 技 术 研 究 [D]. 哈 尔 滨 : 东 北 林 业 大 学 硕 士 学 位 论 文 , 2006.<br />

[5] 倪 建 . 大 刘 坡 桥 体 外 预 应 力 的 加 固 方 案 设 计 [J]. 天 津 市 政 工 程 , 2001, 13(3): 13-16.<br />

[6] 石 亦 平 , 周 玉 蓉 . ABAQUS 有 限 元 分 析 实 例 详 解 [M]. 北 京 : 机 械 工 业 出 版 社 , 2006.6.<br />

[7] 庄 茁 等 . ABAQUS 非 线 性 有 限 元 分 析 与 实 例 [M]. 浙 江 : 科 学 出 版 社 , 2005.<br />

-494-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

IN-PLANE PURE SHEAR PANEL TEST OF SELF-CONSOLIDATING HIGH<br />

PERFORMANCE FIBER REINFORCED CONCRETE (SCHPFRC)<br />

W.-C. Liao<br />

Department of Civil Engineering, National Taiwan <strong>University</strong>, Taiwan<br />

ABSTRACT<br />

Self-consolidating high performance fiber reinforced concrete (SCHPFRC) combines the self-consolidating<br />

property of self-consolidating concrete (SCC) in their fresh state, with the strain-hardening and multiple<br />

cracking characteristics of high performance fiber reinforced cement composites (HPFRCC) in their hardened<br />

state. This paper presents test results of in-plane pure shear test of a panel made of SCHPFRC with 1.5% fiber<br />

volume fraction. <strong>The</strong> test was conducted using the Panel Tester Machine at the <strong>University</strong> of Toronto, which was<br />

developed by Vecchio. <strong>The</strong> shear stress-shear strain response and principal tensile stress-principal tensile strain<br />

response of SCHPFRC are further calculated based on the constitutive concrete model and the measurement of<br />

LVDTs, Zurich gauges, and reinforcement strain gauges. <strong>The</strong> results show SCHPFRC exhibits higher shear<br />

strain capacity and tensile strain hardening compared to conventional concrete.<br />

KEYWORDS<br />

HPFRC, SCC, fiber reinforced concrete, strain hardening, in-plane pure shear.<br />

INTRODUCTION<br />

Self-consolidating high performance fiber reinforced concrete (SCHPFRC) combines the self-consolidating<br />

property of self-consolidating concrete (SCC) with the strain-hardening and multiple cracking characteristics of<br />

high performance fiber reinforced cement composites (HPFRCC) as seen in Figure 1 (Naaman 2003). Several<br />

SCHPFRC mix designs have been developed for different strength demands from 35 to 105 MPa (Liao et al.<br />

2010). SCHPFRC is being addressed as part of a project for the U.S. Network for Earthquake Engineering<br />

Simulation (NEES) with the objective to develop concrete based SC-HPFRCC mixtures that can be easily<br />

manufactured and delivered by ready-mix trucks for use on the job site, with particular application in seismic<br />

resistant structures. In this paper, in-plane pure shear panel test results, including shear stress-shear strain<br />

response and principal tensile stress-principal tensile strain response of SCHPFRC are presented.<br />

s cc<br />

s<br />

pc<br />

e cc<br />

s<br />

pc<br />

s cc<br />

e cc<br />

e pc d 0<br />

Figure 1 Stress-strain response of conventional FRC and HPFRCC (Naaman, 2003)<br />

-495-


EXPERIMENTAL PROGRAMS<br />

Materials and Mix Proportion<br />

<strong>The</strong> cementitious materials used in this study were ASTM Type III Portland cement and class C fly ash. <strong>The</strong><br />

coarse aggregate had a maximum size of 12.7 mm and consisted of solid crushed limestone from a local source,<br />

with a density of about 2.70 g/cm 3 . <strong>The</strong> fine aggregate was #16 flint silica sand (ASTM 50-70).<br />

polycarboxylate-based superplasticizer (SP) was used to achieve desired workability. In addition to the SP, a<br />

viscosity modifying admixture (VMA) was also used to enhance the viscosity and avoid fiber segregation. This<br />

mixture has 1.5% volume fraction of hooked steel fiber with circular cross-section was used with tensile<br />

strength of 2300MPa and aspect ratio of 79 (diameter= 0.38mm and length= 30mm). Average 28-day<br />

compressive strength based on 100×200 mm cylinders is approximately 40 MPa. Details of matrix composition<br />

are given in Tables 1.<br />

Cement<br />

Table 1 Relative composition of SCHPFRC mixture by weight and compressive strength<br />

Fly<br />

Ash<br />

Sand<br />

Coarse<br />

Aggregate<br />

Superplasticizer<br />

f ¢<br />

Water VMA Steel Fiber c<br />

(MPa)<br />

1.0 0.875 2.2 1.2 0.005 0.8 0.038 0.315 40<br />

Specimen Geometry and Test Setup<br />

A 890×890×70 mm concrete panel was fabricated for the shear panel test. Figure 2 illustrates the geometry and<br />

reinforcement layout of the panel. In order to provide an adequate post-cracking resistance of the panel, forty D6<br />

deformed wires were provided for the x-direction reinforcement, giving a total reinforcement area of 1543 mm²,<br />

which equated to a reinforcement ratio of 2.47%.<br />

5/16" Threaded Rod<br />

D6 Bars<br />

ρ x= 2.47%<br />

Pure Shear<br />

Pure Shear<br />

70<br />

70<br />

εy<br />

εh<br />

εv<br />

εx<br />

890<br />

890<br />

Pure Shear<br />

Pure Shear<br />

Y<br />

X<br />

Unit: mm<br />

Figure 2 Dimension, reinforcement layout and LVDT setup of panel<br />

This in-plane pure shear panel test was conducted by using the Panel Tester Machine developed by Vecchio<br />

(1979, 1982) at the <strong>University</strong> of Toronto. This machine and experimental setup were designed to apply various<br />

in-plane loading conditions to a concrete panel. <strong>The</strong> in-plane shear forces are generated by the horizontal and<br />

vertical jacks on each opposing side of the panel. <strong>The</strong>re are adjustable links on the frame to prevent out-of-plane<br />

displacement and keep the panel aligned. Deformations (strains) of the panel were obtained continuously from<br />

the LVDTs and strain gauges throughout the duration of each test. Additional data were also obtained from<br />

Zurich gauge readings that were taken at each load stage to verify the accuracy of LVDT. <strong>The</strong>se data were<br />

subsequently analysed to investigate the response characteristics of the concrete panels under in-plane pure<br />

shear loading.<br />

RESULTS AND DISCUSSIONS<br />

<strong>The</strong> shear stress-shear strain response of SCHPFRC panel is plotted in Figure 3. <strong>The</strong> behavior kept linear up to<br />

first cracking at the shear stress of 1.5 MPa. <strong>The</strong> panel failure occurred at a maximum shear stress of 4.3MPa<br />

-496-


and a corresponding shear strain of 6.20×10 -3 . At the onset of panel failure, the applied load gradually declined<br />

and the major cracks opened up slowly. <strong>The</strong> failure of the panel was dictated by an aggregate interlock failure as<br />

there was no indication of concrete crushing or reinforcement rupture.<br />

Figure 3 <strong>The</strong> shear stress - shear strain response of SCHPFRC panel<br />

<strong>The</strong> comparison of crack pattern of SCHPFRC and control panels (Susetyo 2009) at failure is shown in Figure 4.<br />

It is noted that the control panel (f’ c = 66MPa), which was made of conventional concrete, has higher<br />

reinforcement ratios (x-direction: reinforcement area of 2061 mm², ρ x = 3.31%; y-direction: reinforcement area<br />

of 260 mm²,ρ y = 0.42% ) compared to the SCHPFRC panel. It was also observed that the failure of the<br />

SCHPFRC panel was caused by the inability of the concrete to transmit the load across the cracks (coarse<br />

aggregate interlock failure) instead of the yielding or fracture of the transverse reinforcement in the conventional<br />

concrete panel. <strong>The</strong> cracks in SCHPFRC panel were significantly fine and well distributed across the face of the<br />

panel, with crack spacing smaller than those observed in the control panel.<br />

Figure 4 Crack patterns at failure of SCHPFRC (left) and conventional concrete (right) panels<br />

<strong>The</strong> principal tensile stress-strain relationships of SCHPFRC and conventional concrete are conducted based on<br />

Mohr’s circle method and shown in Figure 5. Due to higher compressive strength of conventional concrete, the<br />

-497-


maximum principal tensile stress of conventional concrete is higher than that of SCHPFRC as expected.<br />

However, in contrast to strain softening behavior of conventional concrete, SCHPFRC exhibits tensile strain<br />

hardening until failure occurs at which the principal tensile strain of 4.1×10 -3 . This can be deemed that the<br />

strain hardening behavior can be achieved with addition of fibers and that fiber addition can significantly<br />

improve the tensile carrying capacity of the concrete.<br />

Figure 5 <strong>The</strong> principal tensile stress-strain relationships of SCHPFRC and conventional concrete<br />

CONCLUSIONS<br />

SCHPFRC combines the self-consolidating property of SCC with the strain-hardening and multiple cracking<br />

characteristics of HPFRCC. This paper presents in-plane pure shear panel test results of SCHPFRC, including<br />

shear stress- shear strain responses and principal tensile stress-strain behavior. Noticeable tensile strain<br />

hardening can be observed in SCHPFRC. In terms of failure mode and crack pattern, SCHPFRC showed fine<br />

and well distributed cracks with much smaller crack spacing than those observed in the conventional concrete<br />

panel. <strong>The</strong> test results also demonstrated the applicability of SCHPFRC mixtures to further applications on shear<br />

walls and panels.<br />

ACKNOWLEDGMENTS<br />

<strong>The</strong> research described herein was sponsored by the National Science Foundation under Grants Nos. CMS<br />

0530383 and CMS 0754505, and the <strong>University</strong> of Michigan and the <strong>University</strong> of Toronto. <strong>The</strong>ir support is<br />

gratefully acknowledged. <strong>The</strong> opinions expressed in this paper are those of the authors and do not necessarily<br />

reflect the views of the sponsor.<br />

REFERENCES<br />

Liao, W.-C., Chao, S.-H. and Naaman, A.E. Experience with Self-Consolidating High Performance Fiber<br />

Reinforced Mortar and Concrete. ACI Journal, Special Publication No. 247, Fiber Reinforced<br />

Self-Consolidating Concrete - Research and Applications, 2010, 79-94<br />

Naaman, A.E. Strain Hardening and Deflection Hardening Fiber Reinforced Cement Composites. High<br />

Performance Fiber Reinforced Cement Composites 4, Proceedings of the Fourth International RILEM<br />

Workshop, 2003, 95-113.<br />

Susetyo J. Fiber Reinforcement for Shrinkage Crack Control in Prestressed, Precast Segmental Bridges, Ph.D.<br />

<strong>The</strong>sis, Department of Civil Engineering, <strong>University</strong> of Toronto, 2009, 532.<br />

Vecchio, F.J. Shear Rig Design. M.Eng. <strong>The</strong>sis, Department of Civil Engineering, <strong>University</strong> of Toronto, 1979,<br />

246.<br />

Vecchio, F.J. and Collins, M.P. <strong>The</strong> Response of Reinforced Concrete to In-Plane Shear and Normal Stresses.<br />

Ph.D. <strong>The</strong>sis, Department of Civil Engineering, <strong>University</strong> of Toronto, 1982, 332.<br />

-498-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

山 西 体 育 中 心 体 育 场 结 构 设 计<br />

傅 学 怡 朱 勇 军 周 颖 杨 想 兵 成 永 平 王 涛 张 晶<br />

( 中 建 ( 北 京 ) 国 际 设 计 顾 问 有 限 公 司 , 北 京 100013)<br />

摘 要 : 山 西 体 育 中 心 体 育 场 为 甲 级 大 型 体 育 建 筑 , 近 六 万 人 座 , 看 台 及 附 属 用 房 采 用 框 架 剪 力<br />

墙 结 构 体 系 , 钢 结 构 罩 棚 采 用 三 角 形 管 桁 架 结 构 。 罩 棚 最 大 悬 挑 长 度 52m, 混 凝 土 结 构 外 缘 边 长 约<br />

944m, 不 设 永 久 变 形 缝 , 属 超 长 结 构 , 设 计 中 考 虑 地 基 的 有 限 约 束 刚 度 , 结 合 工 程 的 施 工 过 程 、 全<br />

年 气 候 统 计 资 料 并 考 虑 混 凝 土 的 收 缩 徐 变 效 应 , 合 理 进 行 温 度 分 析 。 同 时 进 行 总 装 模 型 的 重 力 、 风 、<br />

地 震 作 用 分 析 , 设 计 节 点 构 造 , 保 证 结 构 设 计 安 全 合 理 经 济 。<br />

关 键 词 : 悬 挑 罩 棚 , 超 长 结 构 , 温 度 分 析<br />

THE STRUCTURAL DESIGN OF SHANXI SPORTS CENTER STADIUM<br />

X. Y. Fu Y. J. Zhu Y. Zhou X. B. Yang Y. P. Cheng T. Wang J. Zhang<br />

China Construction Design International (Beijing), Beijing 100013, China<br />

Abstract: <strong>The</strong> stadium of Shanxi sports center is a grade large stadium, nearly 60,000 people can be<br />

accommodated. Reinforced Concrete Frame-shear Wall Structure is adopted in the structural design of lower<br />

structure, and Triangle Tubular Truss in the upper cantilevered canopy. <strong>The</strong> maximum length of cantilevered<br />

canopy is 52 meters and the outer edge of concrete structures is 944 meters, the stadium is a super long structure<br />

without permanent deformation crack. Temperature analysis is performed in design considering the limited<br />

stiffness of the foundation's constraint, construction process, weather statistics, and the effect of shrinkage and<br />

creep of concrete. Gravity and wind analysis, seismic analysis and finite element analysis of nodes are<br />

performed to ensure the safety and economy of the structure.<br />

Keywords: Cantilevered canopy, Super long structure, Temperature analysis.<br />

一 、 前 言<br />

山 西 体 育 中 心 位 于 山 西 省 太 原 市 滨 河 南 路 西 侧 , 新 晋 祠 路 东 侧 , 占 地 1238 亩 。 包 括 主 体 育 场 、<br />

体 育 馆 、 自 行 车 馆 、 游 泳 馆 、 综 合 训 练 馆 以 及 公 寓 和 运 动 员 训 练 基 地 , 并 预 留 体 育 培 训 中 心 发 展 用 地 。<br />

将 满 足 国 内 综 合 运 动 会 和 国 际 单 项 比 赛 的 要 求 , 并 充 分 考 虑 赛 后 运 营 的 灵 活 性 , 同 时 满 足 日 常 训 练 、<br />

群 众 健 身 和 商 业 文 化 活 动 的 需 要 。<br />

主 体 育 场 的 形 象 以 山 西 民 俗 特 色 的 “ 大 红 灯 笼 ”、“ 大 鼓 ” 及 剪 纸 为 创 意 元 素 , 使 民 族 特 色 、 地<br />

域 特 征 与 时 代 精 神 在 其 中 得 到 有 机 的 融 合 和 贴 切 的 体 现 , 建 筑 效 果 图 见 图 1。<br />

本 工 程 为 一 复 杂 的 大 型 甲 级 综 合 体 育 场 , 总 建 筑 面 积 约 90,066 m2, 观 众 座 位 57597 座 , 上 部 为<br />

钢 结 构 悬 挑 罩 棚 , 采 用 三 角 形 空 间 管 桁 架 悬 挑 结 构 , 下 部 为 混 凝 土 看 台 及 各 功 能 用 房 , 采 用 框 架 剪<br />

力 墙 结 构 , 整 体 结 构 总 装 分 析 三 维 模 型 见 图 2。<br />

作 者 简 介 : 傅 学 怡 (1945-), 男 , 江 苏 南 京 人 , 教 授 , 博 导 , 结 构 设 计 大 师 , 中 建 国 际 设 计 顾 问 有 限 公 司 总 工 程 师 。<br />

-499-


图 1 山 西 体 育 中 心 体 育 场 建 筑 效 果 图<br />

图 2 体 育 场 整 体 结 构 三 维 模 型 图<br />

二 、 结 构 设 计<br />

2.1. 基 础 设 计<br />

根 据 山 西 省 勘 察 设 计 研 究 院 《 山 西 体 育 中 心 岩 土 工 程 勘 察 报 告 》( 详 勘 ,2008 年 12 月 ), 拟 建<br />

场 地 表 第 四 系 全 新 统 冲 洪 积 层 。 以 下 土 层 依 次 为 : 粉 质 粘 土 、 粉 土 、 粉 细 砂 、 粉 质 粘 土 。 其 中 第 3<br />

层 粉 土 、 粉 细 砂 为 中 等 液 化 土 层 。<br />

本 工 程 荷 载 差 异 较 大 , 根 据 分 析 结 果 , 并 结 合 当 地 类 似 场 地 情 况 的 设 计 、 施 工 经 验 , 浅 层 地 基<br />

承 载 力 相 对 较 低 , 且 液 化 土 层 需 进 行 消 除 液 化 处 理 , 采 用 天 然 地 基 基 础 方 案 难 以 满 足 要 求 。 经 综 合<br />

比 较 分 析 , 本 工 程 基 础 形 式 采 用 桩 基 础 , 桩 型 采 用 钻 孔 灌 注 桩 。 根 据 试 桩 结 果 和 经 验 公 式 计 算 , 桩<br />

径 分 别 采 用 Φ600,Φ800 桩 ,Φ600 桩 以 第 5 层 粉 土 作 为 桩 端 持 力 层 , 桩 长 24 米 , 单 桩 极 限 承 载<br />

力 4000KN, 考 虑 液 化 折 减 后 竖 向 承 载 力 特 征 值 1500KN;Φ800 桩 以 第 7 层 细 中 砂 作 为 桩 端 持 力 层 ,<br />

桩 长 40 米 , 单 桩 竖 向 极 限 承 载 力 7350KN, 考 虑 液 化 折 减 后 竖 向 承 载 力 特 征 值 3200KN。 桩 混 凝 土<br />

强 度 等 级 C40, 并 采 取 有 关 措 施 进 行 了 耐 久 性 设 计 , 配 筋 率 0.4~0.6%。<br />

2.2. 下 部 混 凝 土 结 构<br />

下 部 钢 筋 混 凝 土 主 体 结 构 高 度 34.85 米 , 外 轮 廓 为 直 径 300.6 米 的 圆 形 。 看 台 外 轮 廓 为 近 椭 圆<br />

形 , 南 北 长 约 267 米 , 东 西 宽 约 245 米 , 比 赛 场 地 内 南 北 长 约 195 米 , 东 西 宽 约 138 米 , 看 台 最 高<br />

点 标 高 约 为 34.85 米 。 混 凝 土 结 构 总 共 地 上 五 层 , 其 中 二 层 为 局 部 夹 层 , 竖 向 构 件 混 凝 土 强 度 均 为<br />

C50, 梁 板 混 凝 土 强 度 等 级 均 为 C30。 采 用 框 架 剪 力 墙 结 构 体 系 , 上 部 支 承 悬 挑 钢 结 构 罩 棚 。 利 用<br />

建 筑 楼 电 梯 间 及 隔 墙 的 位 置 , 均 匀 对 称 布 置 混 凝 土 剪 力 墙 及 筒 体 , 见 图 3、 图 4。<br />

图 3 混 凝 土 结 构 模 型 三 维 视 图 图 4 混 凝 土 筒 体 及 剪 力 墙 平 面 布 置 ( 红 色 )<br />

混 凝 土 结 构 最 外 缘 周 长 约 944 米 , 不 设 永 久 变 形 缝 , 设 计 考 虑 地 基 的 有 限 约 束 刚 度 , 对 混 凝 土<br />

结 构 进 行 整 体 温 差 分 析 。 结 合 工 程 的 施 工 过 程 、 全 年 气 候 统 计 资 料 并 考 虑 混 凝 土 的 收 缩 徐 变 效 应 ,<br />

-500-


对 温 度 变 化 效 应 进 行 分 析 。 并 进 一 步 采 取 施 工 构 造 措 施 减 小 混 凝 土 收 缩 及 温 度 应 力 不 利 影 响 , 留 设<br />

后 浇 带 , 混 凝 土 低 温 入 模 ; 加 强 混 凝 土 养 护 、 覆 盖 ; 降 低 水 泥 用 量 , 减 小 水 灰 比 等 。<br />

2.3. 上 部 钢 结 构<br />

本 工 程 上 部 钢 结 构 为 钢 结 构 罩 棚 , 呈 弧 形 状 , 东 西 高 南 北 低 , 覆 盖 看 台 观 众 席 。 罩 棚 最 宽 处 约<br />

68 米 , 最 窄 处 约 49m, 罩 棚 外 边 缘 南 北 向 最 长 处 约 293 米 , 东 西 向 最 宽 约 275 米 。 整 个 罩 棚 外 立 面<br />

落 于 混 凝 土 结 构 上 , 屋 面 与 墙 体 浑 为 一 体 , 采 用 三 角 形 管 桁 架 空 间 悬 挑 结 构 。<br />

主 结 构 布 置 : 屋 面 钢 结 构 罩 棚 由 径 向 主 桁 架 和 环 向 次 桁 架 组 成 , 采 用 三 角 形 空 间 管 桁 架 悬 挑 结<br />

构 。 整 个 罩 棚 中 间 高 , 两 端 低 , 高 差 9.3m, 罩 棚 最 高 点 离 地 面 约 55.45m。 其 中 , 主 桁 架 最 大 悬 挑<br />

长 度 约 为 52 米 , 桁 架 根 部 最 大 高 度 7.8m, 最 小 悬 挑 长 度 约 为 34.7 米 , 根 部 高 度 5m, 每 榀 桁 架 内 支<br />

座 间 距 约 26 米 , 结 构 轴 测 及 立 面 图 见 图 5~6。 主 桁 架 之 间 布 置 环 向 次 桁 架 , 次 桁 架 为 片 状 管 桁 架 结<br />

构 , 高 度 2 米 , 间 距 约 6 米 ~9 米 。 在 主 桁 架 端 部 、 中 间 根 部 布 置 环 向 桁 架 , 如 图 7 所 示 , 内 环 为 梯<br />

型 桁 架 , 中 环 为 矩 形 桁 架 , 外 环 为 三 角 形 桁 架 。 结 构 墙 面 杆 件 布 置 配 合 建 筑 立 面 造 型 , 采 用 正 交 及<br />

斜 交 两 种 单 层 墙 面 结 构 , 两 种 形 式 之 间 采 用 平 面 桁 架 过 渡 , 并 在 外 支 座 处 设 置 三 角 形 边 桁 架 , 见 图<br />

8。<br />

图 5 罩 棚 结 构 轴 侧 图<br />

(a) 南 北 立 面<br />

(b) 东 西 立 面<br />

图 6 钢 结 构 立 面 视 图<br />

-501-


-0.150<br />

观 众 平 台<br />

i=5%<br />

6.900 6.900<br />

i=1%<br />

i=1%<br />

±0.000<br />

15.300<br />

11.100<br />

6.900<br />

i=3.3%<br />

i=6%<br />

±0.000<br />

6.950<br />

-1.200<br />

i=5%<br />

0.750<br />

图 7 罩 棚 结 构 平 面 布 置 图<br />

图 8 钢 结 构 墙 面 结 构 布 置 图<br />

支 座 : 罩 棚 内 支 座 采 用 四 角 锥 分 叉 钢 管 柱 , 落 在 混 凝 土 看 台 混 凝 土 框 架 柱 上 , 罩 棚 外 侧 主 桁 架<br />

上 下 弦 分 别 落 于 混 凝 土 柱 墩 上 , 并 通 过 混 凝 土 框 架 柱 连 接 于 基 础 , 见 结 构 剖 面 图 9 及 图 10 所 示 。<br />

支 座 节 点 设 计 采 用 铰 接 、 刚 接 双 控 。<br />

51.200<br />

48.000<br />

图 9 结 构 典 型 剖 面 图<br />

图 10 最 大 悬 挑 处 一 榀 主 桁 架 布 置 图<br />

支 撑 体 系 : 为 了 增 强 结 构 的 整 体 稳 定 性 和 侧 向 抗 扭 刚 度 , 在 罩 棚 桁 架 上 弦 平 面 设 置 了 8 组 水 平<br />

支 撑 , 并 向 下 延 伸 至 墙 面 , 增 强 整 体 作 用 , 另 外 设 置 3 组 环 向 支 撑 , 如 图 11, 并 在 次 桁 架 间 布 置 水<br />

平 系 杆 保 证 弦 杆 的 受 压 稳 定 性 。 同 时 , 结 构 立 面 交 叉 杆 件 的 布 置 很 好 的 增 强 了 结 构 的 整 体 抗 扭 刚 度 。<br />

图 11 罩 棚 水 平 支 撑 布 置 ( 红 色 杆 件 )<br />

-502-


典 型 节 点 : 在 钢 结 构 罩 棚 的 前 端 , 如 图 12 所 示 , 在 竖 向 荷 载 作 用 下 , 最 前 端 环 向 杆 轴 向 受 压 ,<br />

且 随 杆 件 刚 度 的 增 大 , 其 轴 向 力 也 随 之 增 大 , 导 致 该 杆 件 的 强 度 及 稳 定 均 难 以 满 足 要 求 , 因 此 , 采<br />

取 构 造 措 施 释 放 该 杆 件 轴 向 力 , 并 结 合 马 道 结 构 布 置 , 将 相 邻 一 榀 平 面 次 环 桁 架 改 为 空 间 四 边 形 环<br />

桁 架 , 提 高 了 受 压 环 自 身 的 稳 定 性 , 且 对 主 桁 架 前 端 竖 向 振 型 周 期 及 变 形 影 响 不 大 。<br />

图 12 钢 结 构 罩 棚 前 端 关 键 节 点 做 法 示 意<br />

罩 棚 钢 结 构 理 论 用 钢 量 5970t, 按 其 覆 盖 面 积 41786m2 计 142.9kg/m2, 按 展 开 面 积 64969m2 计<br />

91.9kg/m2, 与 目 前 国 内 同 等 规 模 其 他 体 育 场 罩 棚 相 比 , 结 构 经 济 指 标 较 好 。<br />

三 、 荷 载 作 用<br />

3.1. 重 力 荷 载<br />

(1) 混 凝 土 结 构 : 附 加 恒 载 、 活 载 根 据 建 筑 楼 ( 屋 ) 面 做 法 及 功 能 取 值 。<br />

(2) 钢 结 构 : 附 加 恒 载 : 金 属 屋 面 + 檩 条 + 天 沟 、 防 水 等 :0.50kN/m2; 阳 光 板 + 檩 条 等 :0.4kN/m2;<br />

建 筑 铝 方 管 装 饰 : 根 据 材 料 自 重 计 算 附 加 , 约 0.4KN/m; 悬 挂 荷 载 : 根 据 设 备 专 业 提 供 的 设 备 重 量<br />

及 布 置 位 置 及 马 道 的 平 面 布 置 输 入 。<br />

活 荷 载 :0.3 KN/ m2<br />

基 本 雪 压 :0.4 kN/m2(100 年 )<br />

(3) 结 构 自 重 由 程 序 自 动 计 算 。<br />

3.2. 风 荷 载<br />

地 面 类 型 B 类 。 按 荷 载 规 范 取 值 , 钢 结 构 计 算 基 本 风 压 按 100 年 重 现 期 取 值 0.45 KN/ m2, 混<br />

凝 土 按 50 年 重 现 期 取 值 0.4kN/m2; 风 压 高 度 系 数 μz:1.72( 钢 结 构 最 高 处 离 地 面 55.5m),1.49( 看<br />

台 最 高 处 离 地 面 34.9m); 风 振 系 数 βz: 混 凝 土 结 构 取 1.0, 钢 结 构 悬 臂 罩 棚 取 2.0 ; 风 载 体 型 系 数<br />

μs: 下 部 混 凝 土 结 构 体 型 系 数 µS=1.3, 钢 结 构 取 值 体 型 系 数 见 下 图 13:<br />

图 13 钢 结 构 罩 棚 体 型 系 数 取 值<br />

建 研 科 技 股 份 有 限 公 司 对 本 工 程 进 行 了 刚 性 模 型 风 洞 试 验 研 究 , 并 计 算 风 振 系 数 , 给 出 了 36<br />

个 风 向 角 下 试 验 结 果 , 考 虑 结 构 的 对 称 性 , 设 计 时 取 β=60°、100°、150°、190°、230° 五 种 角 度 下<br />

的 试 验 结 果 进 行 验 算 。<br />

3.3. 温 度 作 用<br />

-503-


随 主 体 结 构 生 成 , 逐 层 输 入 随 时 间 变 化 的 温 差 , 并 考 虑 混 凝 土 徐 变 收 缩 效 应 (CEB-FIP MODEL<br />

CODE 1990) 及 桩 基 础 对 上 部 结 构 的 有 限 刚 度 约 束 (《 建 筑 桩 基 技 术 规 范 》JGJ94-2008)。<br />

(1) 施 工 模 拟<br />

结 构 从 8 月 份 开 始 施 工 , 第 1、2 个 月 完 成 第 一 层 混 凝 土 结 构 , 第 3 个 月 完 成 第 二 、 三 层 结 构 ,<br />

第 4 个 月 完 成 第 四 、 五 层 结 构 , 然 后 封 闭 一 到 五 层 后 浇 带 , 第 5~7 个 月 完 成 钢 结 构 施 工 , 并 进 入 装<br />

修 阶 段 。 主 体 结 构 逐 层 生 成 , 后 浇 带 逐 层 生 成 , 逐 层 施 加 温 差 , 同 时 随 季 节 变 化 改 变 温 差 , 后 浇 带<br />

选 择 在 低 温 月 合 拢 , 可 以 在 主 体 混 凝 土 完 工 后 一 次 浇 筑 。<br />

(2) 温 差 计 算<br />

由 于 结 构 局 部 温 差 ( 构 件 内 外 表 面 温 差 ) 影 响 较 小 , 计 算 仅 考 虑 整 体 温 差 , 混 凝 土 合 拢 温 差 取 月<br />

平 均 气 温 , 整 体 温 差 = 月 最 高 ( 最 低 ) 气 温 - 合 拢 温 度 。<br />

3.4. 地 震 作 用<br />

地 震 作 用 重 现 期 50 年 , 特 征 周 期 0.45s, 抗 震 设 防 烈 度 8 度 , 设 计 地 震 分 组 为 第 一 组 ,Ⅲ 类 场<br />

地 , 地 震 影 响 系 数 最 大 值 :αmax=0.16( 小 震 )、0.46( 中 震 )、0.90( 大 震 ); 小 震 水 平 峰 值 加 速 度 70gal,<br />

大 震 水 平 峰 值 加 速 度 394gal。 三 向 地 震 作 用 效 应 组 合 系 数 1:0.85:0.65; 下 部 混 凝 土 结 构 阻 尼 比 :0.05<br />

( 单 独 分 析 )0.04( 总 装 分 析 ); 上 部 钢 结 构 阻 尼 比 :0.02( 单 独 、 总 装 分 析 ); 考 虑 5% 单 向 偶 然<br />

偏 心 。<br />

四 、 结 构 性 能 化 设 计 目 标<br />

(1) 小 震 弹 性 : 所 有 构 件 、 支 座 均 处 于 弹 性 。<br />

(2) 中 震 弹 性 ( 基 本 弹 性 ): 中 震 时 下 部 混 凝 土 结 构 竖 向 构 件 处 于 基 本 弹 性 ; 上 部 钢 结 构 构 件 、<br />

支 座 、 节 点 弹 性 。<br />

(3) 大 震 不 倒 塌 : 大 震 时 层 间 位 移 角 满 足 规 范 要 求 , 钢 构 件 产 生 少 量 第 一 、 二 阶 段 塑 性 铰 ;<br />

混 凝 土 结 构 出 现 少 量 少 量 第 一 、 二 阶 塑 性 铰 。<br />

(4) 静 力 荷 载 作 用 下 ( 包 括 风 荷 载 ), 钢 结 构 稳 定 性 满 足 要 求 。<br />

五 、 整 体 结 构 性 能<br />

对 钢 结 构 、 混 凝 土 总 装 结 构 进 行 了 重 力 、 风 荷 载 分 析 , 地 震 反 应 谱 及 时 程 分 析 、 钢 结 构 屈 曲 稳<br />

定 分 析 、 温 度 分 析 、 静 动 力 弹 塑 性 分 析 以 及 节 点 有 限 元 分 析 。<br />

5.1. 模 态 分 析<br />

阶 数<br />

表 1 模 态 分 析 质 量 参 与 系 数<br />

质 量 参 与 系 数<br />

周 期 (s)<br />

Ux Uy Uz Rx Ry Rz<br />

1 0.955 0.003 0.000 0.000 0.000 0.001 0.000<br />

2 0.952 0.000 0.000 0.002 0.000 0.000 0.000<br />

4 0.913 0.000 0.005 0.000 0.000 0.000 0.000<br />

12 0.590 0.000 0.000 0.000 0.000 0.000 0.004<br />

13 0.567 0.000 0.000 0.000 0.000 0.000 0.031<br />

18 0.429 0.293 0.000 0.000 0.000 0.037 0.000<br />

21 0.418 0.000 0.190 0.000 0.034 0.000 0.000<br />

27 0.345 0.000 0.001 0.000 0.000 0.000 0.247<br />

Σ1~50 0.916 0.918 0.788 0.083 0.057 0.910<br />

-504-


钢 结 构 第 一 阶 振 型 为 局 部 x 向 水 平 振 动 , 周 期 0.955s, 第 二 阶 为 东 西 侧 罩 棚 竖 向 振 型 , 周 期<br />

0.952s, 频 率 1.05Hz>1Hz, 满 足 结 构 竖 向 频 率 控 制 要 求 , 第 四 阶 为 y 向 水 平 振 动 , 扭 转 主 振 型 为 第<br />

12 、 13 阶 , 其 中 第 13 阶 的 整 体 扭 转 耦 合 了 部 分 下 部 混 凝 土 看 台 的 质 量 ,TT =0.567s ,<br />

TT/T1=0.567/0.955=0.594。 钢 结 构 振 型 频 率 密 集 , 且 存 在 很 多 高 阶 局 部 振 型 。 第 18、19 阶 振 型 为 混<br />

凝 土 看 台 的 x 向 平 动 振 型 , 第 19 阶 振 型 为 看 台 的 45 度 方 向 斜 向 平 动 振 型 , 第 21 阶 为 看 台 的 y 向<br />

平 动 振 型 , 第 27 阶 为 整 体 看 台 的 扭 转 ,T T =0.345s,T T /T 1 =0.345/0.429=0.804


4000<br />

3000<br />

杆 件 数<br />

2000<br />

1000<br />

0<br />

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0<br />

应 力 比<br />

图 14 钢 结 构 构 件 应 力 水 平 分 布<br />

图 15 非 线 性 屈 曲 分 析 荷 载 步 — 竖 向 基 底 总<br />

反 力 曲 线 (kN)<br />

5.6. 温 度 作 用 分 析<br />

通 过 温 度 分 析 , 大 部 分 径 向 梁 内 无 须 增 加 温 度 筋 , 局 部 内 力 较 高 者 按 计 算 结 果 配 置 温 度 筋 , 但<br />

不 超 过 总 截 面 配 筋 面 积 的 15%, 环 向 梁 额 外 增 加 温 度 筋 约 为 总 截 面 配 筋 面 积 的 5~20%。 外 环 2~3 排<br />

框 架 柱 配 筋 率 有 所 增 长 , 考 虑 温 度 作 用 不 与 中 震 作 用 效 应 组 合 , 剪 力 墙 考 虑 温 度 作 用 分 析 后 基 本 不<br />

用 额 外 增 加 配 筋 。<br />

在 温 度 荷 载 作 用 下 , 环 向 结 构 受 力 较 长 条 结 构 受 力 有 所 不 同 。 本 工 程 椭 圆 环 的 几 何 结 构 使 得 水<br />

平 构 件 在 温 度 作 用 下 互 相 挤 压 , 处 在 整 个 环 带 薄 弱 部 位 ( 南 北 向 ) 梁 板 柱 内 温 度 应 力 均 较 高 , 若 整<br />

个 环 带 水 平 刚 度 均 匀 , 则 对 温 度 应 力 的 集 中 将 会 有 较 好 的 改 善 。 同 时 这 种 “ 挤 压 ” 也 会 对 梁 板 内 温<br />

度 应 力 有 一 定 的 抵 消 , 因 此 环 状 结 构 水 平 构 件 在 环 向 的 温 度 应 力 小 于 同 长 度 的 条 状 结 构 。 环 状 结 构<br />

水 平 刚 度 薄 弱 部 位 的 竖 向 构 件 内 将 产 生 较 大 的 双 向 弯 矩 , 设 计 中 需 引 起 注 意 。<br />

5.7. 静 、 动 力 弹 塑 性 分 析<br />

通 过 进 行 静 、 动 力 弹 塑 性 分 析 可 知 , 结 构 在 多 遇 地 震 作 用 下 完 全 处 于 弹 性 状 态 , 罕 遇 地 震 作 用<br />

下 有 少 量 塑 性 铰 出 现 , 但 整 体 结 构 的 刚 度 并 不 会 很 快 削 弱 , 结 构 内 力 重 分 布 , 整 体 结 构 的 承 载 力 并<br />

没 有 明 显 的 降 低 。 支 撑 钢 结 构 的 柱 , 悬 挑 、 大 跨 及 看 台 后 部 斜 梁 是 整 个 结 构 中 比 较 薄 弱 的 部 位 , 设<br />

计 中 予 以 加 强 。<br />

5.8. 节 点 有 限 元 分 析<br />

采 用 ANSYS 有 限 元 分 析 软 件 ,solid45、solid92 单 元 , 选 取 悬 挑 跨 度 最 大 的 一 榀 框 架 , 选 取 其<br />

中 受 力 最 大 的 主 要 构 件 的 节 点 进 行 分 析 , 做 到 构 造 简 单 、 传 力 直 接 、 施 工 方 便 、 安 全 可 靠 , 达 到 强<br />

节 点 弱 构 件 。 采 取 加 设 肋 板 等 构 造 措 施 , 保 证 节 点 区 承 载 力 。 结 构 上 、 下 支 座 节 点 有 限 元 分 析 结 果<br />

见 图 16, 可 以 看 到 在 最 大 设 计 荷 载 下 , 节 点 区 应 力 除 个 别 点 因 应 力 集 中 稍 高 外 , 大 部 分 区 域 应 力 都<br />

很 低 , 节 点 基 本 保 持 弹 性 , 可 以 保 证 结 构 的 正 常 安 全 工 作 。<br />

-506-


(a) 上 支 座 节 点<br />

(b) 下 支 座 节 点<br />

图 16 节 点 有 限 元 分 析 Von—Mises 应 力 云 图 (MPa)<br />

六 、 结 论<br />

山 西 体 育 中 心 体 育 场 规 模 较 大 , 罩 棚 悬 挑 较 长 , 且 地 处 8 度 区 , 场 地 属 抗 震 不 利 地 段 , 地 震 作<br />

用 组 合 为 主 要 控 制 工 况 , 通 过 进 行 反 应 谱 分 析 、 地 震 波 时 程 分 析 、 静 动 力 弹 塑 性 分 析 , 对 结 构 抗 震<br />

薄 弱 部 位 进 行 加 强 , 看 台 斜 梁 、 钢 结 构 支 承 柱 、 混 凝 土 首 层 剪 力 墙 、 钢 结 构 支 座 附 近 杆 件 等 。 对 结<br />

构 整 体 进 行 考 虑 施 工 全 过 程 的 温 度 作 用 分 析 , 在 温 度 应 力 集 中 区 域 适 当 配 置 温 度 筋 , 满 足 结 构 承 载<br />

力 需 要 并 控 制 裂 缝 宽 度 。 进 行 节 点 有 限 元 分 析 , 通 过 在 节 点 区 域 设 置 加 劲 肋 等 构 造 措 施 , 保 证 强 节<br />

点 弱 构 件 , 提 高 结 构 延 性 。 整 体 结 构 形 式 简 洁 , 传 力 明 确 , 经 济 指 标 较 好 。<br />

参 考 文 献<br />

[1] 徐 培 福 , 傅 学 怡 , 王 翠 坤 , 肖 从 真 编 著 . 复 杂 高 层 建 筑 结 构 设 计 . 北 京 : 中 国 建 筑 工 业 出 版 社 , 2005.<br />

[2] 中 建 国 际 ( 深 圳 ) 设 计 顾 问 有 限 公 司 . 山 西 体 育 中 心 体 育 场 结 构 超 限 设 计 可 行 性 论 证 报 告 , 2009.<br />

-507-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

深 圳 火 车 北 站 站 房 下 部 结 构 设 计<br />

孟 美 莉 吴 兵 傅 学 怡 陈 朝 晖 冯 叶 文 邵 建 伟<br />

( 深 圳 大 学 建 筑 设 计 研 究 院 , 深 圳 ,518060;)<br />

摘 要 : 深 圳 火 车 北 站 为 京 广 铁 路 上 省 级 重 大 型 枢 纽 车 站 之 一 , 站 房 结 构 由 下 部 结 构 和 上 部 屋<br />

盖 钢 结 构 组 成 , 由 于 地 铁 5 号 线 、 平 南 铁 路 和 市 政 道 路 从 站 房 下 方 穿 越 , 站 房 下 部 结 构 不 仅 支 撑 着<br />

上 部 屋 面 钢 结 构 , 同 时 兼 做 城 市 轻 轨 的 支 承 结 构 , 结 构 设 计 复 杂 。 本 文 结 合 工 程 项 目 背 景 , 全 面 介<br />

绍 了 该 工 程 下 部 结 构 设 计 要 点 , 包 括 基 础 结 构 设 计 、 城 市 轻 轨 支 承 结 构 设 计 、 大 跨 楼 盖 舒 适 度 计 算<br />

控 制 、 组 合 梁 弹 性 屈 曲 稳 定 计 算 分 析 、 蜂 窝 梁 结 构 设 计 等 , 计 算 分 析 及 结 构 设 计 方 法 可 供 类 似 工<br />

程 参 考 。<br />

关 键 词 : 复 杂 交 通 枢 纽 城 市 轻 轨 高 架 超 长 结 构 楼 盖 舒 适 度 组 合 梁 屈 曲 分 析<br />

蜂 窝 梁<br />

DESIGN ON THE SUB-STRUCTURE OF SHENZHEN NORTH TRAIN STATION<br />

Meng Meili,Wu Bing,Fu Xueyi,Chen Zhaohui,Feng Yewen,Shao Jianwei,<br />

<strong>The</strong> Institute of Architecture Design & Research Shenzhen <strong>University</strong>, Shenzhen 518060<br />

Abstract: Shenzhen north train station, as one of the major hub stations in the Beijing-Guangzhou railway<br />

system, is consisted of sub-structure and super-steel structure. <strong>The</strong> urban 5 th light railway, Pinnan raiway and the<br />

other municipal road go through the train station. <strong>The</strong> sub-structure supports the super-steel structure, and at the<br />

same time supports the urban light raiway structure. Sub-structure is complex. Combined the project<br />

background, this paper does introduce the key technical issues in the sub-structural design process, such as<br />

foundation design, structure design on the supporting system of urban light raiway, the analysis and control on<br />

floor comfortability, elastic buckling analysis of composite beam, structure design of honeycomb beam etc.<br />

which can be as reference for similar projects.<br />

Keywords: Transport hub, Urban light rail, Super-long structure, floor comfortability analysis, buckling<br />

analysis.<br />

一 、 工 程 概 况<br />

深 圳 火 车 北 站 总 建 筑 面 积 18 万 m2。 高 43m, 其 中 站 房 由 上 部 屋 面 钢 结 构 和 下 部 结 构 组 成 , 如<br />

图 1。 站 房 下 部 结 构 作 为 口 岸 进 出 站 及 候 车 空 间 , 东 西 垂 直 股 道 方 向 为 340m, 南 北 顺 股 道 方 向 长 度<br />

为 207m, 面 积 约 为 7 万 平 米 。 主 体 结 构 采 用 钢 管 混 凝 土 柱 与 钢 - 混 凝 土 楼 板 组 合 梁 框 架 结 构 体 系 ,<br />

局 部 设 钢 管 混 凝 土 空 心 柱 。 标 准 柱 距 43m、27m, 采 用 Φ1400、1500、1600mm 厚 40、50mm 的 钢<br />

管 混 凝 土 柱 ; 高 架 地 铁 4、6 号 线 支 承 柱 2500×4000 钢 管 混 凝 土 空 心 柱 , 在 高 架 层 楼 盖 以 上 上 分 叉<br />

形 成 Y 型 矩 形 钢 管 混 凝 土 空 心 柱 , 支 承 地 铁 4、6 号 线 及 站 房 屋 面 钢 结 构 。 下 部 结 构 平 面 示 意 如 下<br />

图 2 所 示 。<br />

基 金 项 目 : 国 家 自 然 科 学 基 金 资 助 项 目 (50778113)<br />

作 者 简 介 : 孟 美 莉 (1970-), 女 , 山 西 人 , 高 级 工 程 师 , 一 级 注 册 结 构 工 程 师 , 从 事 高 层 超 高 层 、 大 跨 空 间 结 构 的 科 研 设 计 工 作 。<br />

-508-


上 部 屋 面 钢 结 构<br />

下 部 结 构<br />

图 1 深 圳 北 站 站 房 南 北 剖 面 示 意<br />

二 、 基 础 结 构 设 计<br />

本 工 程 基 础 形 式 采 用 扩 底 钻 孔 灌 注 桩 基 础 , 局 部 采 用 墩 基 础 或 嵌 弱 风 化 花 岗 岩 天 然 基 础 。<br />

本 项 目 为 复 杂 交 通 枢 纽 工 程 , 地 铁 5 号 线 、 平 南 铁 路 沿 东 西 方 向 地 下 穿 越 站 房 , 而 市 政 新 区 大<br />

道 在 其 下 方 沿 南 北 向 穿 越 。 站 房 基 础 设 计 与 受 上 述 多 项 相 关 地 下 工 程 施 工 的 约 束 , 从 而 使 得 基 础 结<br />

构 设 计 变 得 很 复 杂 。 地 铁 5 号 线 、 平 南 铁 路 及 新 区 大 道 与 站 房 基 础 平 面 关 系 如 下 图 3 所 示 。<br />

由 于 站 房 与 地 铁 5 号 线 、 平 南 铁 路 及 新 区 大 道 采 用 不 同 的 结 构 体 系 , 基 础 型 式 与 沉 降 形 态 完 全<br />

不 同 , 因 此 , 需 采 用 合 理 技 术 措 施 消 除 站 房 基 础 结 构 与 其 它 交 通 线 轨 工 程 的 相 互 干 扰 。<br />

地 铁 5 号 线 结 构<br />

褥 垫 层 , 范 围 为<br />

地 梁 宽 + 每 侧 1m<br />

承 台 上 设 褥 垫 层 , 范 围 为<br />

承 台 宽 + 每 侧 1m<br />

图 2 站 房 下 部 结 构 图<br />

图 3 基 础 与 各 交 通 枢 纽 平 面 关 系 图 4 基 础 与 地 铁 5 号 线 结 构 剖 面 图<br />

设 计 时 首 先 确 定 的 原 则 是 站 房 与 多 条 相 关 枢 纽 基 础 之 间 相 互 脱 离 , 各 自 独 立 设 置 基 础 , 站 房 受<br />

枢 纽 影 响 部 分 的 基 础 承 台 标 高 降 至 与 各 交 通 枢 纽 基 础 相 平 或 之 下 , 各 交 通 枢 纽 与 站 房 柱 之 间 设 褥 垫<br />

层 过 渡 隔 离 , 消 除 站 房 柱 对 各 交 通 枢 纽 不 均 匀 沉 降 的 影 响 , 如 图 4。 褥 垫 层 可 采 用 粗 砂 + 残 积 层 亚 粘<br />

土 均 匀 搅 拌 夯 实 。 同 时 , 站 房 桩 基 础 设 计 时 考 虑 地 铁 、 新 区 大 道 及 其 上 部 回 填 土 的 部 分 重 量 。<br />

三 、 城 市 轻 轨 4、6 号 线 支 承 结 构<br />

3.1 城 市 轻 轨 4、6 号 线<br />

深 圳 市 城 市 轻 轨 4、6 号 线 沿 南 北 方 向 高 架 穿 越 站 房 , 如 图 5 示 , 图 中 通 过 柱 墩 支 承 于 18.95m<br />

平 台 的 阴 影 部 分 即 为 4、6 号 线 结 构 。 该 部 分 结 构 作 为 4、6 号 线 的 站 台 、 站 厅 层 , 主 要 由 列 车 行 车<br />

梁 结 构 组 成 , 由 北 京 城 建 院 设 计 , 结 构 剖 面 见 图 6.<br />

-509-


图 5 地 铁 4、6 号 线<br />

图 6 4、6 号 线 结 构 剖 面 图<br />

3.2 城 市 轻 轨 4、6 号 线 荷 载 及 荷 载 组 合<br />

4、6 号 线 结 构 传 给 下 部 支 承 结 构 荷 载 主 要 有 :(1)4、6 号 线 站 台 、 站 厅 层 结 构 恒 载 、 活 载 ;(2)<br />

双 线 列 车 梁 及 列 车 : 恒 载 5850kN, 活 载 1400 kN( 每 个 支 座 作 用 力 ); 单 线 列 车 梁 及 列 车 : 恒 载 4500<br />

kN, 活 载 1400 kN( 每 个 支 座 作 用 力 );(3) 列 车 轨 道 伸 缩 力 220 kN; 断 轨 力 918kN; 牵 引 力 144kN;<br />

摇 摆 力 96kN<br />

当 有 地 震 力 组 合 时 , 上 述 全 部 荷 载 均 作 为 活 荷 载 考 虑 , 而 无 地 震 作 用 参 与 的 承 载 力 极 限 状 态 设<br />

计 组 合 有 两 种 :1.35 恒 载 +1.4 活 载 +1.4 伸 缩 力 ( 或 断 轨 力 )+1.4 牵 引 力 +1.4 摇 摆 力 。<br />

3.3 城 市 轻 轨 4、6 号 线 支 承 结 构 设 计<br />

深 圳 北 站 结 构 最 大 的 一 个 特 点 就 是 城 市 轻 轨 4、6 从 车 站 建 筑 中 穿 过 , 城 市 轻 轨 的 桥 梁 结 构 和<br />

车 站 建 筑 结 构 合 二 为 一 。 由 于 轻 轨 列 车 荷 载 较 大 , 且 列 车 刹 车 停 站 、 启 动 和 高 速 通 过 时 对 整 个 站 房<br />

结 构 、 对 候 车 室 振 动 和 舒 适 度 都 将 产 生 较 大 的 影 响 。 结 构 布 置 受 其 影 响 较 大 , 从 而 也 使 得 结 构 设 计<br />

变 得 很 复 杂 。4、6 支 承 结 构 示 意 及 布 置 图 如 下 :<br />

图 7 4、6 号 线 站 厅 ( 支 承 ) 层 结 构 布 置<br />

图 8 支 承 结 构 垂 直 股 道 方 向 主 要 剖 面 A-A<br />

支 承 柱 较 为 复 杂 , 共 16 根 。 结 合 建 筑 造 型 首 层 采 用 八 边 形 空 心 钢 管 混 凝 土 柱 , 截 面<br />

2500×4000mm, 柱 在 二 层 以 上 分 叉 为 截 面 2500×2000mm 的 两 七 边 形 柱 , 形 成 Y 型 柱 , 如 图 7 示 。<br />

基 于 以 下 考 虑 , 4、6 号 线 Y 型 支 承 柱 采 用 2500×4000mm 的 钢 管 混 凝 土 空 心 柱 : 1 4、6 号 线<br />

集 中 质 量 大 , 支 承 结 构 需 相 应 较 大 的 刚 度 ; 2 列 车 通 过 可 能 会 引 起 站 房 结 构 产 生 较 大 的 振 动 , 有<br />

必 要 加 强 列 车 支 承 结 构 的 刚 度 ; 3 Y 型 柱 上 支 承 4 道 列 车 轨 道 梁 , 列 车 行 驶 的 不 均 匀 性 会 引 起 Y 型<br />

柱 两 上 撑 点 受 力 不 均 、 变 形 不 均 ,Y 型 柱 垂 直 股 道 方 向 会 产 生 较 大 的 弯 矩 ,Y 型 柱 垂 直 股 道 方 向 边<br />

长 适 当 增 大 为 4m; 4 4、6 号 线 轨 道 梁 支 承 系 统 要 求 下 部 柱 在 顺 股 道 方 向 尺 寸 必 须 大 于 2.5m; 5 空<br />

心 可 节 省 混 凝 土 , 避 免 大 体 积 混 凝 土 效 应 , 同 时 可 基 本 保 持 柱 抗 弯 刚 度 。<br />

由 于 Y 型 柱 作 为 4、6 号 线 桥 墩 的 支 承 结 构 , 除 需 按 普 通 的 房 屋 建 筑 进 行 设 计 外 , 还 需 满 足 《 地<br />

铁 设 计 规 范 》(GB50157-2003) 的 相 关 要 求 。 根 据 该 规 范 第 9.1.7 款 、 表 9.1.7, 桥 墩 纵 向 水 平 线 刚 度<br />

须 满 足 : 桥 梁 跨 度 20~30m 时 , 最 小 水 平 线 刚 度 为 320KN/cm。 同 时 根 据 9.1.8 款 , 横 向 水 平 刚 度<br />

限 值 可 取 为 320x5/4=400 KN/cm。 考 虑 到 Y 型 柱 除 作 为 上 部 地 铁 4、6 号 线 桥 墩 的 支 承 外 , 尚 需 支<br />

承 站 房 屋 盖 钢 结 构 , 故 Y 型 柱 水 平 线 刚 度 控 制 值 适 当 提 高 为 ≥1000KN/cm。<br />

同 时 , 由 于 4、6 号 线 采 用 无 缝 轻 轨 , 根 据 《 地 铁 设 计 规 范 》, 需 严 格 控 制 结 构 不 均 匀 沉 降 , 其 总<br />

沉 降 量 与 相 邻 柱 墩 沉 降 差 , 不 应 超 过 容 许 值 : 柱 总 沉 降 量 50mm, 相 邻 墩 柱 沉 降 量 之 差 20mm。 经<br />

-510-


验 算 4、6 号 线 重 力 荷 载 标 准 值 加 列 车 重 作 用 下 ( 考 虑 了 列 车 行 驶 的 不 均 匀 性 ), 柱 端 最 大 沉 降 量 为<br />

4.6mm, 相 邻 墩 柱 沉 降 差 为 2.8mm, 满 足 要 求 。<br />

四 、 楼 盖 舒 适 度 结 构 设 计<br />

楼 盖 的 振 动 , 一 般 由 人 的 行 走 、 运 动 或 机 械 车 辆 设 备 运 行 等 产 生 , 有 关 楼 盖 的 振 动 对 人 的 生 活<br />

工 作 舒 适 度 的 影 响 , 已 经 越 来 越 引 起 人 们 的 关 注 , 对 于 北 站 这 种 人 群 密 集 、 大 跨 楼 盖 的 公 共 建 筑 ,<br />

楼 盖 舒 适 度 的 控 制 显 得 尤 为 重 要 。<br />

深 圳 北 站 站 房 舒 适 度 分 析 主 要 包 括 有 两 方 面 : 人 行 走 、 跳 跃 等 引 起 楼 盖 振 动 对 人 们 舒 适 度 的 影<br />

响 ; 高 架 轻 轨 列 车 振 动 对 楼 盖 舒 适 度 的 影 响 。 其 中 第 2 项 另 文 详 述 。<br />

对 于 钢 梁 组 合 楼 盖 , 其 舒 适 度 控 制 往 往 以 楼 盖 振 动 峰 值 加 速 度 和 楼 盖 竖 向 频 率 控 制 。 楼 盖 振 动<br />

峰 值 加 速 度 可 参 考 ATC40 控 制 , 该 部 分 内 容 另 文 , 而 在 楼 盖 竖 向 频 率 控 制 方 面 , 一 般 有 : 对 于 轻 钢<br />

楼 盖 结 构 , 竖 向 自 振 频 率 需 >8HZ, 混 凝 土 楼 盖 , 竖 向 频 率 需 >3HZ。 本 小 节 仅 从 控 制 站 房 下 部 楼 盖<br />

竖 向 频 率 方 面 出 发 , 介 绍 有 关 舒 适 度 的 控 制 的 简 化 、 可 行 的 计 算 分 析 方 法 。<br />

深 圳 北 站 站 房 下 部 楼 盖 采 用 钢 梁 - 混 凝 土 楼 板 体 系 , 平 面 布 置 见 图 2, 标 准 单 元 主 次 梁 截 面 如 下<br />

图 9 所 示 , 次 梁 间 距 6.75m, 钢 梁 以 上 钢 筋 混 凝 土 板 厚 度 为 180mm, 上 迭 浇 150mm 厚 细 石 混 凝 土 ,<br />

次 梁 跨 度 43m, 短 向 主 梁 跨 度 27m, 主 次 梁 截 面 尺 寸 由 刚 度 和 强 度 条 件 计 算 结 果 分 别 为<br />

H1400(900)x2300x50x70、H700x2300x40x50, 材 料 为 Q345B, 混 凝 土 强 度 等 级 C30, 板 上 荷 载 ( 含 面<br />

层 、 隔 墙 )6.25kN/m 2 , 活 荷 载 3.5kN/m 2 。<br />

图 9 标 准 单 元 梁 板 布 置 图<br />

4.1 楼 盖 刚 度 简 化 计 算<br />

本 工 程 钢 梁 高 2300mm, 板 位 于 离 梁 中 性 轴 较 远 的 翼 缘 上 方 , 对 钢 梁 刚 度 贡 献 较 大 , 采 用 传 统<br />

方 法 把 梁 、 板 单 独 刚 度 相 加 不 足 以 反 映 组 合 楼 盖 真 实 刚 度 , 以 下 楼 盖 刚 度 简 化 计 算 清 晰 地 表 明 板 对<br />

组 合 楼 盖 刚 度 贡 献 。<br />

取 标 准 长 向 梁 (H700x2300x40x50)、 混 凝 土 板 h=180mm+150 mm, 梁 跨 度 43m。 美 国 ( 加 拿 大 )<br />

钢 结 构 协 会 标 准 中 楼 盖 系 统 自 振 频 率 是 统 一 按 照 简 支 梁 计 算 , 同 时 考 虑 到 连 续 梁 , 控 制 楼 盖 结 构 竖<br />

向 自 振 频 率 实 际 上 就 是 控 制 下 图 所 示 振 型 的 自 振 频 率 , 故 本 工 程 简 化 计 算 采 用 简 支 梁 模 型 。<br />

图 10 连 续 梁 振 动 振 型<br />

简 化 手 算 和 有 限 元 程 序 (SAP2000) 计 算 相 结 合 , 所 采 用 的 计 算 模 型 及 假 定 有 : 模 型 1: 手 算<br />

-511-


4<br />

模 型 1, 梁 刚 度 和 板 刚 度 单 纯 相 加 , 挠 度 Δ= Cm<br />

5 ql /384EI<br />

(mm), C<br />

m<br />

为 梁 连 续 性 影 响 系 数 , 简 支<br />

及 等 跨 连 续 梁 计 及 邻 跨 反 向 振 动 取 1; 竖 向 频 率 f<br />

n<br />

= 18/ Δ (Hz); 模 型 2: 手 算 模 型 2, 按 组 合 截<br />

面 计 算 刚 度 ; 模 型 3: 电 算 模 型 , 梁 用 杆 元 模 拟 , 板 用 壳 元 模 拟 , 杆 + 壳 模 型 ; 模 型 4: 电 算 模 型 ,<br />

梁 板 均 用 壳 元 模 拟 ; 模 型 5: 电 算 模 型 , 梁 、 板 均 用 SOLID 实 体 单 元 模 拟 。<br />

表 1 自 振 频 率 计 算 结 果 比 较<br />

模 型 1 模 型 2 模 型 3 模 型 4 模 型 5<br />

有 效 重 力 荷 载 作 用 下 楼 盖 梁 跨 中 挠 Δ(mm) 70.7 32 71 32 30.6<br />

楼 盖 竖 向 频 率 f (Hz) 2.14 3.2 2.13 3.15 3.25<br />

n<br />

由 上 表 可 以 看 出 , 模 型 1 和 模 型 3 计 算 假 定 较 为 接 近 , 得 到 的 计 算 结 果 也 较 为 接 近 , 由 于 均 未<br />

考 虑 板 在 组 合 截 面 中 位 置 对 截 面 刚 度 贡 献 , 得 到 的 挠 度 偏 大 ; 模 型 2 和 4、5 假 设 接 近 , 得 到 的 结<br />

果 比 较 接 近 , 其 中 又 以 全 SOLID 模 型 梁 、 板 协 调 变 形 合 理 , 得 到 的 结 果 可 认 为 是 精 确 的 有 限 元 解 。<br />

4.2 整 体 结 构 楼 盖 竖 向 振 动 频 率 计 算 分 析<br />

根 据 楼 盖 振 动 分 析 和 控 制 的 原 理 , 本 工 程 楼 盖 竖 向 频 率 整 体 模 型 计 算 时 考 虑 以 下 4 个 方 面 ;(1)<br />

考 虑 整 浇 混 凝 土 楼 板 与 钢 梁 共 同 受 力 变 形 , 调 整 钢 梁 刚 度 以 反 映 板 的 实 际 刚 度 贡 献 ;(2) 采 用 全<br />

弹 性 楼 盖 多 自 由 度 振 动 模 型 , 计 及 连 续 性 影 响 采 用 相 邻 跨 反 向 运 动 对 应 竖 向 振 动 振 型 作 为 第 一 自 振<br />

频 率 ;(3) 考 虑 动 力 材 料 弹 性 模 量 提 高 1.2 倍 ;(4) 考 虑 有 效 活 荷 载 , 对 正 常 使 用 状 态 活 荷 载 予<br />

以 折 减 。 基 于 以 上 4 方 面 , 采 用 ETABS 电 算 模 型 , 计 算 得 到 的 楼 盖 竖 向 第 一 振 动 频 率 如 图 11.<br />

由 于 本 工 程 楼 盖 结 构 自 重 较 大 , 楼 盖 阻 抗 大 , 频 率 控 制 为 主 控 指 标 , 当 其 在 控 制 范 围 内 时 , 峰<br />

值 加 速 度 控 制 易 满 足 。 本 工 程 另 外 进 行 的 人 行 走 、 跳 跃 峰 值 加 速 度 计 算 分 析 结 果 也 证 明 了 这 点 。<br />

图 11 楼 盖 竖 向 第 一 振 型 ( 频 率 3.27HZ)<br />

五 、 组 合 梁 受 弯 屈 曲 稳 定 计 算 分 析<br />

站 房 楼 盖 采 用 H 型 钢 梁 — 混 凝 土 楼 板 组 合 结 构 形 式 , 南 北 顺 股 道 方 向 主 梁 标 准 跨 为 27m, 东 西<br />

方 向 主 梁 标 准 跨 43m, 东 西 向 每 跨 设 3 道 次 梁 , 次 梁 间 距 为 6.75m, 梁 高 h=2300mm, 板 厚 180mm,<br />

C30 混 凝 土 , 组 合 楼 盖 结 构 布 置 见 图 2。<br />

组 合 钢 梁 上 翼 缘 受 混 凝 土 楼 板 约 束 , 基 本 上 没 有 侧 向 位 移 , 其 稳 定 性 分 析 应 与 《 钢 结 构 设 计 规<br />

范 》(GB50017) 规 定 的 自 由 梁 弯 扭 失 稳 不 同 。 在 连 续 组 合 梁 支 座 负 弯 矩 区 段 下 翼 缘 受 压 , 容 易 发 生<br />

侧 扭 屈 曲 。 对 于 组 合 钢 梁 侧 扭 屈 曲 的 计 算 , 我 国 钢 结 构 设 计 规 范 采 用 限 制 绕 弱 轴 的 长 细 比 的 方 法 ( 参<br />

见 规 范 9.3.2 条 )。 这 种 方 法 未 能 考 虑 楼 板 作 用 , 不 适 用 于 本 工 程 组 合 钢 梁 稳 定 性 分 析 。<br />

此 外 , 在 连 续 组 合 梁 的 负 弯 矩 区 , 钢 梁 不 仅 承 受 较 大 的 剪 力 , 同 时 弯 曲 应 力 、 局 部 压 力 也 都 较<br />

大 。 腹 板 处 于 弯 、 剪 、 压 作 用 下 的 复 杂 应 力 状 态 , 局 部 稳 定 性 问 题 也 要 考 虑 。<br />

本 工 程 组 合 钢 梁 关 于 抗 弯 承 载 力 验 算 采 用 我 国 《 钢 结 构 设 计 规 范 》(GB50017) 和 欧 洲 规 范 组 合<br />

结 构 设 计 规 范 EuroCode4 双 控 、 侧 扭 屈 曲 和 局 部 稳 定 性 的 分 析 采 用 EC4。 同 时 , 利 用 SAP2000 通 用<br />

有 限 元 程 序 对 组 合 钢 梁 进 行 了 线 性 屈 曲 分 析 。 以 下 重 点 介 绍 关 于 组 合 梁 的 屈 曲 稳 定 有 限 元 计 算 分<br />

析 。<br />

-512-


采 用 SAP2000, 对 各 种 截 面 的 钢 梁 进 行 屈 曲 分 析 , 不 考 虑 初 始 缺 陷 , 梁 板 均 采 用 全 壳 单 元 模 拟 ,<br />

通 过 共 用 节 点 实 现 梁 板 之 间 的 连 接 , 不 考 虑 梁 板 之 间 的 滑 移 , 结 合 工 程 连 续 组 合 钢 梁 采 用 两 种 不 同<br />

的 约 束 条 件 : 两 端 固 支 —— 模 拟 中 间 梁 ; 一 端 固 支 、 一 端 简 支 —— 模 拟 边 跨 梁 , 为 了 较 为 真 实 的 体<br />

现 整 浇 混 凝 土 板 刚 度 对 钢 梁 的 约 束 作 用 , 取 3 跨 板 进 行 计 算 分 析 , 计 算 模 型 如 下 :<br />

图 12 计 算 模 型 示 意 ( 楼 板 不 显 示 )<br />

图 13 剖 面 示 意<br />

图 14 两 端 固 支 模 型 第 一 屈 曲 模 态<br />

图 15 一 端 固 支 、 一 端 简 支 第 一 屈 曲 模 态<br />

( 支 座 负 弯 矩 区 腹 板 、 下 翼 缘 屈 曲 ) ( 梁 发 生 弯 扭 屈 曲 )<br />

( 对 应 的 屈 曲 因 子 4.88) ( 对 应 的 屈 曲 因 子 4.64)<br />

有 限 元 屈 曲 稳 定 计 算 分 析 表 明 , 重 力 荷 载 标 准 值 作 用 下 , 一 端 固 支 、 一 端 简 支 梁 最 小 屈 曲 因 子<br />

为 4.64, 也 即 总 重 力 荷 载 标 准 值 的 安 全 系 数 4.64。 说 明 现 设 计 采 用 的 截 面 能 够 保 证 钢 梁 在 组 合 工 况<br />

最 大 值 M( 约 为 标 准 值 的 1.8 倍 ) 作 用 下 钢 梁 不 会 产 生 侧 扭 屈 曲 失 稳 , 梁 具 有 足 够 的 稳 定 承 载 力 。<br />

设 计 时 在 梁 端 1/10 梁 跨 处 设 置 构 造 隅 撑 , 如 图 16, 进 一 步 加 强 了 钢 梁 的 整 体 稳 定 性 。<br />

图 16 隅 撑 连 接 节 点<br />

六 、 蜂 窝 梁 结 构 设 计<br />

组 合 梁 梁 高 控 制 为 2300mm。 为 解 决 设 备 管 线 穿 越 及 节 省 用 钢 量 , 在 长 向 组 合 梁 腹 板 设 置 均 匀<br />

布 置 的 大 菱 形 孔 洞 , 采 用 图 17 的 切 割 组 合 的 方 法 实 现 , 由 于 梁 中 部 腹 板 形 成 的 孔 洞 为 正 六 变 形 ,<br />

像 峰 窝 状 而 得 名 。 因 为 有 六 边 形 孔 洞 的 存 在 , 所 以 必 须 要 验 算 腹 板 与 翼 缘 相 交 处 的 弯 曲 强 度 以 及 腹<br />

板 在 中 间 部 位 相 焊 处 的 抗 剪 强 度 , 此 外 , 尚 需 考 虑 孔 边 的 应 力 集 中 和 开 孔 对 梁 刚 度 的 削 弱 。<br />

图 17 组 合 梁 腹 板 切 割 组 合 示 意<br />

-513-


标 准 梁 跨 43m, 每 跨 共 设 13 个 孔 , 孔 离 梁 端 距 离 在 1.5h(h 为 梁 高 ) 以 上 。 采 用 通 用 有 限 元<br />

SAP2000 对 其 进 行 了 有 限 元 分 析 , 开 孔 工 字 钢 梁 、 混 凝 土 板 均 采 用 壳 单 元 模 拟 , 计 算 模 型 及 开 孔 位<br />

置 见 图 12、18。<br />

图 18 长 向 次 梁 开 孔 布 置 示 意 图<br />

2<br />

图 19 重 力 荷 载 标 准 值 作 用 下 梁 正 应 力 ( N / mm )( 图 中 数 字 为 该 孔 截 面 最 大 正 应 力 )<br />

2<br />

图 20 重 力 荷 载 标 准 值 作 用 下 梁 剪 应 力 ( N / mm )( 图 中 数 字 为 该 孔 截 面 最 大 剪 应 力 )<br />

图 21 重 力 荷 载 标 准 值 作 用 下 长 向 次 梁 变 形<br />

计 算 结 果 见 图 19~21, 可 以 看 到 , 组 合 梁 开 孔 引 起 一 定 的 应 力 集 中 现 象 , 重 力 荷 载 标 准 值 作 用<br />

2<br />

下 , 长 向 次 梁 的 最 大 正 应 力 峰 值 σ 约 为 94 N / mm , 且 应 力 不 丰 满 , 满 足 规 范 要 求 ; 开 孔 削 弱<br />

max<br />

了 梁 腹 板 , 对 梁 的 整 体 刚 度 有 一 定 的 影 响 , 梁 跨 中 最 大 挠 度 为 32.1mm(32.1/43000=1/1087) 比 梁 开<br />

孔 前 挠 度 值 23.86mm 增 大 40%, 但 仍 小 于 规 范 要 求 的 限 值 。<br />

七 、 结 语<br />

(1) 深 圳 火 车 北 站 工 程 体 量 巨 大 , 空 间 关 系 复 杂 , 多 条 交 通 枢 纽 纵 横 交 错 穿 越 站 房 下 部 , 结<br />

构 设 计 远 远 复 杂 于 普 通 的 空 间 结 构 。<br />

(2) 多 条 交 通 枢 纽 下 穿 站 房 , 为 了 消 除 各 基 础 之 间 的 相 互 影 响 , 设 置 褥 垫 层 过 渡 隔 离 是 一 种<br />

简 单 而 有 效 的 措 施 。<br />

(3) 下 部 结 构 同 时 作 为 上 部 屋 盖 钢 结 构 和 城 市 轻 轨 的 支 承 体 系 , 除 按 普 通 公 建 进 行 结 构 设 计<br />

外 , 尚 应 按 《 地 铁 设 计 规 范 》 等 相 关 规 范 进 行 复 核 计 算 , 以 确 保 结 构 的 安 全 。<br />

(4) 针 对 站 房 下 部 结 构 的 大 跨 楼 盖 , 展 开 了 楼 盖 舒 适 度 、 组 合 梁 屈 曲 稳 定 、 蜂 窝 梁 有 限 元 等<br />

的 详 尽 计 算 分 析 , 其 计 算 方 法 可 供 其 他 类 似 工 程 参 考 。<br />

参 考 文 献<br />

[1] 中 铁 第 四 勘 察 设 计 研 究 院 有 限 公 司 、 深 圳 大 学 建 筑 设 计 研 究 院 . 深 圳 北 站 超 长 结 构 专 家 论 证 审 查 报 告 , 2008.<br />

[2] 中 铁 第 四 勘 察 设 计 研 究 院 有 限 公 司 、 深 圳 大 学 建 筑 设 计 研 究 院 . 深 圳 北 站 抗 震 超 限 专 项 审 查 报 告 , 2009.<br />

[3] 徐 培 福 , 傅 学 怡 , 王 翠 坤 , 肖 从 真 . 复 杂 高 层 建 筑 结 构 设 计 . 中 国 建 筑 工 业 出 版 社 , 2005.<br />

[4] ATC Design Guide 1, Minimizing Flloor Vibration, Applied Technology Council, 1999<br />

-514-


<strong>The</strong> 5th Cross-strait Conference on Structural and Geotechnical Engineering (SGE-5)<br />

<strong>Hong</strong> <strong>Kong</strong>, China, 13-15 July 2011<br />

應 用 現 地 微 振 量 測 進 行 斜 張 橋 之 沖 刷 評 估<br />

吳 文 華 陳 建 州 石 峰 王 勝 威<br />

( 國 立 雲 林 科 技 大 學 營 建 工 程 系 , 斗 六 , 雲 林 )<br />

摘 要 : 橋 墩 與 基 礎 之 沖 刷 監 測 近 年 來 成 為 工 程 界 與 學 界 特 別 重 視 的 問 題 , 不 過 相 關 研 究 在 精<br />

確 度 和 穩 定 性 上 一 直 無 法 獲 得 重 大 突 破 。 本 研 究 選 定 高 屏 溪 斜 張 橋 為 對 象 , 目 標 在 發 展 出 一 套 根 據<br />

現 地 橋 體 微 振 量 測 進 行 有 效 沖 刷 評 估 以 至 長 期 監 測 的 方 法 。 首 先 針 對 主 梁 橋 面 、 箱 型 梁 內 部 及 橋 墩<br />

頂 進 行 各 振 動 方 向 之 微 振 量 測 , 而 後 據 此 識 別 主 梁 垂 直 、 水 流 、 車 行 、 扭 轉 向 及 局 部 橋 墩 車 行 向 之<br />

各 個 實 測 振 態 頻 率 。 接 著 以 SAP2000 軟 體 建 立 斜 張 橋 結 構 體 之 有 限 元 素 模 型 , 並 配 合 原 始 設 計 之 各<br />

項 參 數 進 行 振 態 頻 率 分 析 。 結 合 前 述 兩 項 結 果 , 本 研 究 進 一 步 交 叉 分 析 出 有 限 元 素 模 型 中 最 符 合 現<br />

地 實 況 之 各 項 邊 界 支 承 條 件 。 再 來 則 改 變 模 型 中 橋 墩 基 礎 之 土 層 深 度 進 行 參 數 分 析 , 由 此 決 定 出 對<br />

於 沖 刷 效 應 最 為 敏 感 的 幾 個 關 鍵 振 態 頻 率 。 最 後 以 所 得 關 鍵 振 態 頻 率 為 基 準 設 計 目 標 誤 差 函 數 , 然<br />

後 改 變 有 限 元 素 模 型 中 橋 墩 基 礎 之 土 層 深 度 與 勁 度 參 數 進 行 最 佳 化 分 析 , 由 此 便 足 以 有 效 評 估 出 最<br />

可 能 的 洪 水 沖 刷 現 況 。<br />

關 鍵 詞 : 沖 刷 評 估 斜 張 橋 微 振 量 測 有 限 元 素 模 型<br />

SCOURING EVALUATION OF CABLE-STAYED BRIDGES BASED ON AMBIENT<br />

VIBRATION MEASUREMENTS<br />

W.-H. Wu, 1 C.-C. Chen, 1 F. Shi 1 and S.-W. Wang 1<br />

1 Dept. of Construction Eng., National Yunlin <strong>University</strong> of Science and Technology, Touliu, Yunlin 640, Taiwan<br />

Abstract: Scouring monitoring of bridge piers and foundations has recently attracted a considerable<br />

attention in engineering practice and academics. Nevertheless, the stability and accuracy of its related<br />

research works have not been able to reach a satisfactory level. This study chooses Kao-Ping-His<br />

Cable-stayed Bridge as the research target and is aimed to develop an efficient scouring evaluation method<br />

based on the ambient vibration measurements of bridge structures. <strong>The</strong> ambient vibration measurements in<br />

different directions are first conducted on the girder, inside the girder and on the pier top to identify various<br />

actual modal frequencies of girder and local pier. <strong>The</strong> finite element model of the cable-stayed bridge is<br />

then constructed with SAP2000 to perform the modal frequency analysis with the original design<br />

parameters. Combining the above results, this study further determines the best boundary support<br />

conditions for the finite element model to fit the identified modal frequencies. Based on this globally best<br />

fitted model, the parameter analysis with the alteration of soil depth supporting the pier foundation is<br />

subsequently performed to reveal the critical bridge frequencies most sensitive to the scouring effect.<br />

Finally, an objective error function considering all the critical bridge frequencies is defined such that the<br />

optimal analysis can be carried out to effectively investigate the current scouring status from different<br />

possible values for soil depth and stiffness.<br />

Keywords: Scouring evaluation, cable-stayed bridge, ambient vibration measurement, finite element model.<br />

-515-


一 、 前 言<br />

近 年 來 因 地 球 環 境 的 急 劇 變 遷 使 全 球 天 然 災 害 逐 年 增 多 , 導 致 橋 梁 嚴 重 損 壞 甚 至 斷 裂 的 悲 劇 一<br />

再 重 演 。 以 台 灣 為 例 , 這 幾 年 最 令 人 無 法 釋 懷 的 斷 橋 案 例 包 括 2000 年 的 高 屏 大 橋 、2008 年 的 后 豐<br />

大 橋 及 2009 年 的 雙 園 大 橋 等 。 特 別 是 新 近 2009 年 8 月 莫 拉 克 颱 風 來 襲 , 中 南 部 地 區 於 當 年 8 月 7<br />

日 至 8 月 10 日 四 天 期 間 共 計 落 下 1000 至 3000 公 釐 的 雨 量 , 約 為 年 平 均 雨 量 的 四 成 至 八 成 。 此 超<br />

大 豪 雨 在 山 區 積 聚 後 經 由 各 河 流 洶 湧 暴 洩 , 引 發 坡 地 崩 塌 、 土 石 流 沖 擊 、 河 床 淤 積 及 洪 水 溢 流 等 嚴<br />

重 問 題 , 從 而 造 成 各 行 水 區 橋 梁 之 沖 毀 與 斷 裂 。 統 計 全 台 損 毀 橋 梁 超 過 110 座 , 其 災 情 慘 重 為 歷 年<br />

之 最 , 血 淚 斑 斑 地 彰 顯 出 研 究 洪 水 沖 刷 對 橋 梁 安 全 影 響 並 提 出 適 當 對 策 之 重 要 性 與 急 迫 性 。<br />

橋 梁 之 安 全 除 了 繫 於 分 析 與 設 計 時 對 其 動 力 行 為 與 破 壞 機 制 的 精 確 掌 握 外 , 正 式 使 用 後 之 有 效<br />

監 測 對 於 避 免 人 員 損 傷 及 橋 體 損 壞 其 實 可 能 扮 演 更 為 關 鍵 的 角 色 。 因 此 , 橋 墩 與 基 礎 之 沖 刷 監 測 近<br />

年 來 成 為 工 程 界 與 學 界 特 別 重 視 的 問 題 , 分 別 有 學 者 嘗 試 時 間 域 反 射 儀 [1-2] [3-4]<br />

、 雷 達 及 光 纖 布 拉 格<br />

[5-6]<br />

光 柵 (FBG) 感 測 器 等 不 同 手 段 。 不 過 相 關 研 究 在 精 確 度 和 穩 定 性 上 一 直 無 法 獲 得 重 大 突 破 , 主 要<br />

障 礙 在 於 直 接 量 測 水 流 及 土 石 之 困 難 度 。 最 近 國 內 外 開 始 有 學 者 嘗 試 從 間 接 分 析 橋 體 結 構 的 振 動 訊<br />

號 著 手 [7-10] , 希 望 由 此 得 以 明 確 判 斷 不 同 的 橋 墩 沖 刷 與 基 礎 掏 空 程 度 , 從 而 建 立 更 為 簡 便 有 效 的 監<br />

測 機 制 。 但 是 這 個 研 究 方 向 也 因 為 橋 梁 沖 刷 監 測 牽 涉 到 整 個 橋 體 、 土 壤 以 至 於 流 水 的 複 雜 互 制 系 統 ,<br />

所 以 極 不 易 僅 從 橋 體 振 動 訊 號 便 明 確 量 化 沖 刷 影 響 而 遭 遇 到 瓶 頸 。 在 這 個 脈 絡 之 下 , 本 研 究 選 定 台<br />

灣 跨 徑 最 長 且 具 有 高 度 沖 刷 或 淤 積 潛 在 危 機 的 高 屏 溪 斜 張 橋 為 對 象 , 目 標 在 發 展 出 一 套 根 據 現 地 橋<br />

體 微 振 量 測 進 行 有 效 沖 刷 評 估 以 至 長 期 監 測 的 方 法 。<br />

二 、 高 屏 溪 斜 張 橋 概 述<br />

高 屏 溪 斜 張 橋 全 長 2617 公 尺 , 位 於 高 雄 縣 大 樹 鄉 與 屏 東 縣 九 如 鄉 的 交 接 處 , 跨 越 南 台 灣 重 要<br />

河 川 高 屏 溪 。 全 橋 依 跨 徑 配 置 不 同 共 分 為 六 個 單 元 , 其 中 南 端 第 四 、 五 、 六 單 元 跨 徑 配 置 以 45.3 公<br />

尺 為 主 的 等 斷 面 預 力 混 凝 土 箱 型 梁 橋 , 採 支 撐 先 進 工 法 施 工 , 緊 鄰 第 二 、 三 單 元 跨 徑 配 置 分 別 以 120<br />

公 尺 及 80 公 尺 為 主 的 變 斷 面 預 力 混 凝 土 箱 型 梁 橋 , 採 懸 臂 施 工 法 施 工 , 最 北 端 則 設 計 一 座 全 世 界<br />

第 二 長 之 單 橋 塔 非 對 稱 混 合 式 斜 張 橋 , 如 圖 1 所 示 。 橋 下 高 屏 溪 流 域 寬 達 2 公 里 , 主 河 道 位 於 斜 張<br />

橋 主 跨 的 下 方 , 冬 天 河 道 窄 小 , 水 量 不 多 且 流 速 緩 , 夏 季 期 間 則 常 溪 水 氾 濫 , 尤 其 是 颱 風 來 臨 , 洪<br />

水 常 淹 沒 整 個 高 屏 溪 流 域 , 湍 急 洪 水 已 造 成 高 屏 溪 大 橋 上 下 游 多 座 橋 梁 損 毀 , 如 雙 園 大 橋 、 高 屏 大<br />

橋 及 里 港 大 橋 等 。<br />

高 屏 溪 斜 張 主 橋 橋 面 寬 34.5 公 尺 , 南 北 兩 向 共 6 車 道 , 全 長 510 公 尺 , 主 跨 330 公 尺 為 全 焊 接<br />

鋼 構 箱 型 梁 , 跨 越 河 川 主 河 道 , 邊 跨 180 公 尺 為 預 力 混 凝 土 箱 型 梁 , 跨 越 省 道 台 21 線 , 示 意 圖 可<br />

見 圖 2。 鋼 筋 混 凝 土 橋 塔 高 183.5 公 尺 , 採 造 型 優 美 及 結 構 穩 定 度 高 之 倒 Y 型 設 計 , 兩 隻 塔 腳 由 地<br />

梁 相 連 接 , 並 座 落 於 37 公 尺 深 之 隔 牆 箱 室 連 續 壁 式 基 礎 上 ( 稱 P1 塔 腳 )。 斜 拉 索 系 統 沿 橋 梁 中 心 成<br />

單 面 混 合 扇 形 配 置 , 於 橋 塔 兩 側 各 安 裝 14 組 , 兩 端 分 別 錨 錠 於 塔 柱 及 橋 面 版 中 央 。 主 梁 主 要 由 斜<br />

拉 索 系 統 懸 吊 支 撐 , 但 中 央 亦 座 落 於 橋 塔 橫 梁 , 兩 端 則 分 別 由 橋 台 ( 稱 A1 橋 台 ) 與 邊 墩 ( 稱 P2 橋 墩 )<br />

支 撐 , 皆 設 計 拉 力 連 桿 , 以 抵 抗 上 揚 力 。 橋 台 座 落 於 里 嶺 台 地 , 亦 由 連 續 壁 式 基 礎 支 撐 。 邊 墩 則 緊<br />

臨 主 河 道 一 側 , 採 28 公 尺 長 樁 基 礎 設 計 。<br />

圖 1 高 屏 溪 斜 張 橋 全 景<br />

圖 2 高 屏 溪 斜 張 橋 示 意 圖<br />

-516-


高 屏 溪 引 橋 共 包 含 五 個 橋 梁 單 元 , 均 為 預 力 混 凝 土 箱 型 梁 橋 , 皆 採 樁 基 礎 設 計 , 長 度 依 承 載 力<br />

設 計 約 在 20 公 尺 至 35 公 尺 之 間 。 除 非 洪 水 期 間 , 其 下 方 河 川 流 域 並 無 水 流 , 常 被 用 於 耕 作 與 養 殖 。<br />

颱 風 來 襲 時 , 洪 水 則 常 漫 及 此 區 域 , 並 沖 刷 橋 墩 基 礎 。 由 於 高 屏 溪 斜 張 橋 的 基 礎 較 深 , 雖 亦 因 洪 水<br />

作 用 發 生 嚴 重 沖 刷 掏 空 , 但 橋 梁 並 未 發 生 任 何 損 壞 , 但 不 可 否 認 的 是 沖 刷 掏 空 已 降 低 基 礎 的 承 載 力 。<br />

考 量 全 球 氣 候 異 常 加 劇 , 未 來 強 烈 颱 風 來 襲 將 更 加 頻 繁 , 風 災 所 引 起 洪 水 沖 刷 作 用 亦 將 日 益 嚴 重 ,<br />

高 屏 溪 斜 張 主 橋 是 否 可 抵 抗 風 災 洪 水 經 常 性 的 衝 擊 應 乃 是 一 個 必 須 嚴 肅 面 對 的 問 題 。<br />

三 、 現 地 微 振 量 測 與 振 態 頻 率 識 別<br />

本 研 究 團 隊 為 精 確 掌 握 高 屏 溪 斜 張 橋 之 現 況 與 動 力 特 性 , 於 2010 年 內 分 階 段 對 其 進 行 數 次 現<br />

地 微 振 量 測 , 範 圍 涵 括 主 跨 與 側 跨 之 主 梁 橋 面 、 鋼 構 箱 型 梁 內 部 、P2 橋 墩 頂 、P1 塔 腳 橫 梁 及 斜 張<br />

鋼 纜 。 使 用 儀 器 為 日 本 東 京 測 振 VSE-15D 伺 服 型 速 度 計 與 SPC51 主 機 , 設 定 取 樣 頻 率 200Hz, 每<br />

次 量 測 時 間 300 秒 , 每 筆 共 計 60000 點 速 度 歷 時 資 料 。<br />

以 下 首 先 針 對 與 本 文 相 關 的 主 梁 橋 面 及 箱 型 梁 內 部 之 量 測 配 置 加 以 說 明 , 詳 如 圖 3 和 圖 4 所 示 ,<br />

其 中 各 個 感 測 器 之 位 置 編 號 定 義 為 :<br />

首 碼 英 文 字 母 ~ M: 主 跨 橋 面 ;S: 側 跨 橋 面 ;I: 箱 型 梁 內 部<br />

數 字 碼 ~ 距 離 橋 塔 原 點 處 之 距 離 (m)<br />

附 屬 標 示 (-) ~ 箱 型 梁 內 水 流 側 向 之 不 同 位 置<br />

我 們 在 主 梁 橋 面 的 量 測 包 括 垂 直 (vertical or gravity direction)、 水 平 (horizontal or river direction) 與 車 行<br />

(axial or driving direction) 等 三 個 不 同 的 振 動 方 向 , 其 目 的 在 識 別 斜 張 橋 體 主 要 呈 現 於 這 些 方 向 的 各<br />

振 態 參 數 ; 而 箱 型 梁 內 部 的 量 測 則 只 鎖 定 垂 直 向 , 其 目 的 在 識 別 斜 張 橋 體 以 扭 轉 (torsional) 向 為 主 之<br />

各 振 態 參 數 。 值 得 一 提 的 是 , 由 於 感 測 器 數 目 的 限 制 , 我 們 進 行 主 梁 橋 面 各 方 向 的 量 測 均 分 為 (M280,<br />

M230, M180, M130, M80) 及 (M80, S50, S100, S150) 兩 階 段 進 行 , 但 以 M80 為 共 同 參 考 點 。 另 外 對 於<br />

箱 型 梁 內 部 的 量 測 , 我 們 分 別 選 定 主 跨 距 橋 塔 276m 及 165m 兩 個 截 斷 面 內 部 進 行 , 並 且 以 圖 4 所<br />

示 的 兩 點 位 置 同 步 成 對 量 測 。<br />

斜 張 橋 體 振 態 頻 率 主 要 是 以 微 振 速 度 訊 號 的 傅 利 葉 振 幅 譜 之 尖 峰 值 進 行 識 別 。 利 用 傳 統 的 傅 利<br />

葉 分 析 , 可 將 一 個 時 間 域 的 函 數 經 由 傅 利 葉 轉 換 (Fourier transform) 呈 現 為 頻 率 域 中 的 複 數 函 數 。 而<br />

此 複 數 函 數 在 不 同 頻 率 下 之 振 幅 大 小 , 即 直 接 代 表 原 函 數 在 各 個 頻 率 的 組 成 權 重 , 一 般 以 其 直 接 對<br />

頻 率 作 圖 來 檢 視 原 函 數 之 頻 率 內 容 , 即 為 所 謂 的 傅 利 葉 振 幅 譜 (Fourier amplitude spectrum)。 在 外 部<br />

擾 動 作 用 下 , 一 個 結 構 系 統 之 位 移 、 速 度 或 加 速 度 歷 時 反 應 因 為 均 經 由 系 統 之 轉 換 函 數 而 在 各 振 態<br />

頻 率 處 特 別 放 大 組 成 權 重 , 所 以 通 常 透 過 其 傅 利 葉 振 幅 譜 便 可 簡 易 判 斷 各 振 態 頻 率 的 可 能 位 置 。 不<br />

過 必 須 注 意 的 是 , 在 外 部 擾 動 頻 率 內 涵 未 知 的 情 形 下 , 結 構 反 應 歷 時 的 傅 利 葉 振 幅 譜 之 峰 值 所 在 頻<br />

率 , 除 了 可 能 對 應 系 統 各 振 態 頻 率 外 , 也 無 法 排 除 乃 是 外 力 集 中 於 少 數 特 定 頻 率 而 產 生 。 首 先 以 圖<br />

5 所 示 橋 面 M80 感 測 器 於 垂 直 向 量 測 之 歷 時 與 傅 利 葉 振 幅 譜 為 例 , 明 顯 可 以 看 出 基 本 上 不 難 直 接 由<br />

圖 3 主 梁 橋 面 之 感 測 器 配 置<br />

圖 4 箱 型 梁 內 部 之 感 測 器 配 置<br />

-517-


圖 5 M80 感 測 器 於 垂 直 向 量 測 之 歷 時 與 傅 利 葉 振 幅 譜<br />

傅 利 葉 振 幅 譜 明 確 判 別 橋 體 各 個 振 態 頻 率 。 為 降 低 因 外 力 或 環 境 因 素 影 響 所 造 成 的 振 態 頻 率 識 別 誤<br />

[11]<br />

差 , 本 研 究 亦 採 用 頻 率 域 平 滑 化 技 巧 以 識 別 出 更 精 確 有 效 的 橋 體 各 振 態 頻 率 , 如 此 方 能 提 供 有 效<br />

的 振 態 頻 率 變 化 基 礎 以 作 為 後 續 沖 刷 影 響 評 估 之 用 。<br />

進 一 步 整 理 各 點 量 測 之 傅 利 葉 振 幅 譜 , 可 以 發 現 不 少 振 態 頻 率 在 不 同 振 動 方 向 重 複 出 現 。 換 句<br />

話 說 , 這 些 振 態 極 可 能 為 不 同 方 向 的 耦 合 振 態 , 如 此 將 大 幅 增 加 判 讀 上 的 困 難 , 不 過 此 亦 為 複 雜 之<br />

斜 張 橋 結 構 所 可 預 期 之 特 性 。 為 後 續 研 究 之 清 楚 對 應 起 見 , 本 文 經 過 交 叉 比 對 將 斜 張 橋 體 在 各 振 動<br />

方 向 所 能 釐 清 識 別 的 前 幾 個 主 要 貢 獻 振 態 頻 率 , 將 其 依 不 同 位 置 與 振 動 方 向 列 於 表 1 至 表 4。 各 個<br />

不 同 振 動 方 向 的 振 態 以 英 文 字 母 編 碼 , 如 BV 表 示 垂 直 撓 曲 振 態 、BH 代 表 水 平 撓 曲 振 態 、A 指 車 行<br />

軸 向 振 態 、T 表 示 扭 轉 向 振 態 ; 至 於 英 文 字 母 後 之 數 字 碼 則 用 以 區 別 該 振 動 方 向 之 振 態 順 序 。 比 較<br />

表 中 各 項 結 果 可 以 發 現 , 不 同 量 測 位 置 所 得 之 各 振 態 頻 率 基 本 上 相 當 穩 定 一 致 , 僅 主 側 跨 測 點 因 分<br />

兩 階 段 不 同 時 間 進 行 量 測 而 可 能 受 小 幅 溫 度 變 化 影 響 造 成 識 別 頻 率 的 極 微 差 異 。 另 外 值 得 注 意 的<br />

是 , 主 梁 的 前 兩 個 水 平 撓 曲 振 態 頻 率 及 車 行 向 第 二 振 態 頻 率 均 難 以 由 側 跨 橋 面 測 點 的 量 測 識 別 出<br />

來 。 至 於 根 據 箱 型 梁 內 成 對 垂 直 向 量 測 所 識 別 的 主 梁 扭 轉 向 振 態 頻 率 , 在 靠 近 P2 橋 墩 的 276m 處 截<br />

斷 面 較 容 易 看 出 第 二 振 態 的 貢 獻 , 而 位 於 P2 橋 墩 與 橋 塔 中 點 的 165m 處 截 斷 面 則 相 對 利 於 第 一 振 態<br />

的 識 別 。 本 研 究 下 一 階 段 之 分 析 即 以 表 1 至 表 4 所 整 理 的 十 個 主 梁 各 方 向 實 測 振 態 頻 率 為 趨 近 目 標 ,<br />

建 構 出 斜 張 橋 有 限 元 素 模 型 中 最 符 合 現 地 實 況 之 各 項 邊 界 支 承 條 件 。<br />

表 1 由 高 屏 溪 斜 張 橋 主 梁 橋 面 垂 直 向 量 測 所 識 別 出 之 各 垂 直 撓 曲 振 態 頻 率<br />

Mode<br />

Identified Modal Frequency (Hz)<br />

M280 M230 M180 M130 M80 S50 S100 S150<br />

BV1 0.273 0.273 0.273 0.273 0.273; 0.270 0.270 0.270 0.270<br />

BV2 0.527 0.527 . 0.527 0.527; 0.523 0.523 0.523 0.523<br />

BV3 0.867 . 0.867 . 0.867; 0.863 0.863 0.863 0.863<br />

表 2 由 高 屏 溪 斜 張 橋 主 梁 橋 面 水 平 向 量 測 所 識 別 出 之 各 水 平 撓 曲 振 態 頻 率<br />

Mode<br />

Identified Modal Frequency (Hz)<br />

M280 M230 M180 M130 M80 S50 S100 S150<br />

BH1 0.980 . 0.980 . 0.980;0.980 . . .<br />

BH2 1.373 1.373 1.373 1.373 1.373;1.377 . . .<br />

BH3 2.197 2.193 2.193 2.200 2.197;2.193 . 2.193 2.193<br />

-518-


表 3 由 高 屏 溪 斜 張 橋 主 梁 橋 面 車 行 向 量 測 所 識 別 出 之 各 車 行 向 振 態 頻 率<br />

Mode<br />

Identified Modal Frequency (Hz)<br />

M280 M230 M180 M130 M80 S50 S100 S150<br />

A1 2.420 2.420 2.420 2.420 2.420;2.430 . 2.427 2.427<br />

A2 2.837 2.837 2.837 2.837 2.837;2.837 . . .<br />

表 4 由 高 屏 溪 斜 張 橋 箱 型 梁 內 部 垂 直 向 成 對 量 測 所 識 別 出 之 各 扭 轉 向 振 態 頻 率<br />

Mode<br />

Identified Modal Frequency (Hz)<br />

(G276-2) − (G276-1) (G165-2) − (G165-1)<br />

T1 . 0.740<br />

T2 1.477 .<br />

四 、 有 限 元 素 分 析 與 邊 界 支 承 條 件 調 整<br />

斜 張 橋 之 結 構 組 成 較 為 複 雜 , 進 行 有 限 元 素 分 析 時 必 須 將 包 括 橋 塔 與 主 梁 的 上 部 結 構 、 斜 張 鋼<br />

纜 、 包 含 基 礎 與 土 層 的 下 部 結 構 以 至 各 項 邊 界 支 承 條 件 等 精 確 有 效 地 模 擬 , 若 干 輸 入 參 數 細 微 的 變<br />

化 即 可 能 明 顯 影 響 分 析 結 果 。 對 於 高 屏 溪 斜 張 橋 本 研 究 採 用 SAP2000 程 式 建 構 有 限 元 素 模 型 , 同 時<br />

先 將 P2 橋 墩 簡 化 為 邊 界 支 承 以 與 現 地 實 測 之 各 振 態 頻 率 比 對 來 進 行 模 態 更 新 , 如 圖 6 所 示 。 橋 塔 、<br />

鋼 纜 、 主 跨 混 凝 土 箱 型 梁 、 邊 跨 鋼 箱 型 梁 及 橋 塔 沉 箱 基 礎 等 結 構 元 件 均 以 SPACE FRAME ( 空 間 剛<br />

架 ) 元 素 模 擬 , 同 時 每 個 構 件 不 同 的 斷 面 與 材 料 性 質 亦 都 參 照 中 華 顧 問 工 程 司 之 橋 梁 設 計 資 料 與 竣 工<br />

圖 進 行 細 部 輸 入 。 至 於 相 對 複 雜 的 土 層 與 主 梁 邊 界 支 承 條 件 等 兩 項 關 鍵 模 擬 因 素 , 我 們 則 在 以 下 詳<br />

細 說 明 。<br />

首 先 關 於 基 礎 周 邊 土 壤 部 份 的 模 擬 , 在 工 程 實 務 上 為 求 簡 化 常 採 用 僅 需 根 據 鑽 探 資 料 便 能 決 定<br />

的 等 效 彈 簧 。 基 於 此 , 本 研 究 乃 根 據 赤 井 ─ 高 橋 公 式 :<br />

0.<br />

37<br />

K = 502N<br />

(1)<br />

圖 6 高 屏 溪 斜 張 橋 之 SAP2000 有 限 元 素 分 析 模 型 ( 簡 化 P2 橋 墩 為 邊 界 支 承 )<br />

及 福 岡 ─ 宇 都 公 式 :<br />

-519-


0.<br />

406<br />

K = 691N<br />

(2)<br />

兩 者 的 平 均 來 計 算 水 平 向 的 地 盤 反 力 係 數 。 必 須 注 意 的 是 ,(1) 式 與 (2) 式 中 地 盤 反 力 K 的 單 位 為<br />

tf/m 3 , 而 N 則 代 表 土 壤 之 標 準 貫 入 度 。 求 出 平 均 地 盤 反 力 係 數 後 , 乘 上 每 一 段 基 礎 元 素 之 作 用 面 積<br />

便 可 出 其 所 對 應 的 土 壤 彈 簧 係 數 , 本 研 究 對 於 水 流 以 及 車 行 兩 水 平 向 的 等 效 土 壤 彈 簧 都 以 此 方 式 來<br />

處 理 。<br />

接 著 有 關 主 梁 的 邊 界 支 承 條 件 , 可 分 為 其 與 A1 橋 台 、P1 橋 塔 和 P2 橋 墩 等 三 部 份 來 討 論 。 主<br />

梁 在 A1 橋 台 的 現 地 支 承 系 統 由 多 種 支 承 機 制 組 成 , 其 組 合 構 件 包 含 拉 力 連 桿 、 人 造 橡 膠 支 承 墊 、<br />

剪 力 榫 與 縱 向 水 平 鋼 纜 等 四 部 分 ; 其 中 兩 條 拉 力 連 桿 用 以 束 制 向 上 位 移 , 三 個 人 造 橡 膠 支 承 墊 承 受<br />

垂 直 壓 力 , 兩 個 剪 力 榫 束 制 水 平 橫 向 位 移 , 十 二 根 縱 向 水 平 鋼 纜 束 制 水 平 縱 向 位 移 。 考 慮 這 些 支 承<br />

配 置 , 本 研 究 對 於 主 梁 在 A1 橋 台 處 之 支 承 條 件 首 先 假 設 為 一 個 單 向 鉸 接 端 , 亦 即 束 制 住 UX、UY、<br />

UZ、RX 及 RZ 等 五 個 方 向 。 至 於 主 梁 在 P2 橋 墩 的 現 地 支 承 系 統 則 包 含 盤 式 支 承 、 拉 力 連 桿 與 剪 力<br />

榫 等 三 部 分 ; 兩 個 盤 式 支 承 近 似 橡 膠 支 承 墊 用 來 承 受 垂 直 壓 力 。 根 據 這 些 支 承 配 置 , 本 研 究 對 於 主<br />

梁 在 P2 橋 墩 處 之 支 承 條 件 首 先 假 設 為 一 個 單 向 滾 接 端 , 亦 即 束 制 住 UY、UZ、RX 及 RZ 等 四 個 方<br />

向 。 另 外 , 主 梁 與 P1 橋 塔 之 橫 梁 介 面 本 研 究 僅 設 定 垂 直 向 UZ 之 束 制 , 而 主 梁 與 P1 橋 塔 之 斜 柱 介<br />

面 則 設 定 水 流 向 UY 之 束 制 , 以 上 關 於 有 限 元 素 分 析 的 各 項 主 梁 支 承 條 件 原 始 設 定 均 列 於 表 5。<br />

經 由 上 述 的 斜 張 橋 有 限 元 素 模 型 建 構 , 接 著 我 們 便 可 應 用 SAP2000 軟 體 進 行 振 態 分 析 , 從 而 完<br />

整 求 得 高 屏 溪 斜 張 橋 不 同 振 動 方 向 的 各 個 振 態 頻 率 與 其 對 應 振 形 。 透 過 全 橋 振 形 的 繪 出 , 足 以 清 楚<br />

判 別 出 各 個 振 態 的 主 要 振 動 方 向 , 如 圖 7 與 圖 8 所 示 即 分 別 為 垂 直 撓 曲 第 一 振 態 (BV1) 和 水 平 撓 曲<br />

第 一 振 態 (BH1) 的 有 限 元 素 分 析 振 形 。 藉 助 分 析 振 形 的 協 同 判 斷 , 我 們 可 以 進 一 步 針 對 表 1 至 表 4<br />

所 列 的 十 個 主 梁 各 方 向 實 測 振 態 頻 率 找 出 其 對 應 之 分 析 振 態 頻 率 並 進 行 比 較 , 結 果 整 理 於 表 6。 從<br />

表 5 高 屏 溪 斜 張 橋 有 限 元 素 分 析 模 型 之 主 梁 各 項 邊 界 支 承 條 件 調 整<br />

Support Boundary<br />

Constraint<br />

Original<br />

Modified<br />

UX Yes 10 7 t/m<br />

UY Yes Yes<br />

A1 Abutment<br />

UZ Yes Yes<br />

RX Yes Yes<br />

RY No 10 7 t-m<br />

RZ Yes Yes<br />

UX No No<br />

UY Yes Yes<br />

P1 Pylon<br />

UZ No No<br />

( 主 梁 與 橋 塔 斜 柱 )<br />

RX No Yes<br />

RY No No<br />

RZ No No<br />

UX No No<br />

UY Yes Yes<br />

P2 Pier<br />

UZ Yes Yes<br />

RX Yes Yes<br />

RY No No<br />

RZ Yes Yes<br />

-520-


圖 7 垂 直 撓 曲 第 一 振 態 (BV1) 振 形<br />

圖 8 水 平 撓 曲 第 一 振 態 (BH1) 振 形<br />

表 6 主 梁 邊 界 支 承 條 件 調 整 前 後 之 各 振 態 頻 率 比 較<br />

Modal Frequency (Hz)<br />

Mode<br />

Identified from<br />

Original FE Model<br />

Modified FE Model<br />

Measurements<br />

(Error Percentage)<br />

(Error Percentage)<br />

BV1 0.273<br />

BV2 0.527<br />

BV3 0.863<br />

BH1 0.980<br />

BH2 1.373<br />

BH3 2.193<br />

A1 2.420<br />

A2 2.837<br />

T1 0.740<br />

T2 1.477<br />

0.241<br />

(11.72%)<br />

0.524<br />

(0.57%)<br />

0.827<br />

(4.15%)<br />

0.971<br />

(0.92%)<br />

1.365<br />

(0.58%)<br />

2.246<br />

(−2.42%)<br />

2.432<br />

(−0.48%)<br />

2.848<br />

(−0.39%)<br />

0.696<br />

(5.95%)<br />

1.398<br />

(5.35%)<br />

0.273<br />

(0.01%)<br />

0.524<br />

(0.57%)<br />

0.844<br />

(2.26%)<br />

0.971<br />

(0.92%)<br />

1.365<br />

(0.58%)<br />

2.246<br />

(−2.42%)<br />

2.416<br />

(0.17%)<br />

2.842<br />

(−0.18%)<br />

0.742<br />

(−0.27%)<br />

1.474<br />

(0.20%)<br />

表 6 的 比 較 可 以 看 出 , 基 本 上 實 測 與 分 析 之 各 振 態 頻 率 誤 差 不 算 太 大 , 但 以 扭 轉 向 振 態 頻 率 差 異 較<br />

大 而 均 在 6% 上 下 , 同 時 垂 直 撓 曲 第 一 振 態 更 有 達 11.7% 的 誤 差 。 考 慮 改 變 材 料 參 數 並 無 法 解 決 如 此<br />

各 振 態 頻 率 不 均 勻 差 異 的 問 題 , 本 研 究 判 斷 應 由 主 梁 的 邊 界 支 承 條 件 進 行 調 整 。 參 考 現 地 實 際 支 承<br />

狀 況 及 各 振 態 頻 率 之 誤 差 方 向 , 我 們 針 對 主 梁 與 A1 橋 台 及 P1 橋 塔 橫 梁 之 束 制 條 件 修 正 , 其 設 定 亦<br />

列 於 表 5, 其 中 修 改 的 束 制 條 件 以 粗 黑 斜 體 標 出 。 仔 細 地 說 , 我 們 一 方 面 經 由 加 入 主 梁 與 P1 橋 塔 橫<br />

-521-


梁 之 RX 向 束 制 來 提 高 扭 轉 向 各 振 態 頻 率 ; 另 外 更 同 時 調 整 主 梁 與 A1 橋 台 的 介 面 , 藉 著 鬆 解 UX<br />

向 束 制 為 線 性 彈 簧 來 降 低 車 行 軸 向 各 振 態 頻 率 , 加 入 RY 向 之 旋 轉 彈 簧 以 增 大 垂 直 撓 曲 向 各 振 態 頻<br />

率 。 表 5 中 所 列 出 之 線 性 彈 簧 係 數 10 7 tf/m 與 旋 轉 彈 簧 係 數 10 7 tf-m, 均 為 多 次 試 誤 後 之 最 佳 値 。 綜<br />

合 言 之 , 主 梁 各 邊 界 支 承 條 件 調 整 後 , 實 測 與 分 析 之 各 振 態 頻 率 誤 差 在 各 振 動 方 向 皆 十 分 均 勻 。 以<br />

表 5 檢 視 的 十 個 振 態 頻 率 來 說 , 幾 乎 都 低 於 1%, 僅 BV3 和 BH3 兩 個 振 態 頻 率 之 誤 差 略 超 過 2%,<br />

可 謂 已 達 到 相 當 精 確 可 靠 的 有 限 元 素 模 型 標 準 。<br />

五 、 受 沖 刷 影 響 之 敏 感 振 態 頻 率 分 析<br />

在 以 前 節 討 論 之 有 限 元 素 模 型 確 認 主 梁 各 主 要 振 態 頻 率 均 可 得 到 精 確 模 擬 後 , 為 進 一 步 評 估 斜<br />

張 橋 現 地 橋 墩 的 沖 刷 情 形 , 本 研 究 依 照 結 構 實 際 狀 況 將 P2 橋 墩 也 納 入 有 限 元 素 模 型 , 如 圖 9 所 示 。<br />

特 別 值 得 一 提 , 此 一 新 的 有 限 元 素 模 型 在 P2 墩 頂 與 主 梁 的 介 面 維 持 表 5 所 示 之 UY、UZ、RX 與<br />

RZ 等 四 個 方 向 的 束 制 , 至 於 墩 底 群 樁 基 礎 與 周 邊 土 壤 的 模 擬 亦 完 全 仿 照 前 節 橋 塔 沉 箱 基 礎 與 周 邊<br />

土 層 之 模 擬 。 接 著 將 擴 充 P2 橋 墩 的 有 限 元 素 模 型 同 樣 進 行 振 態 分 析 , 可 以 發 現 除 了 新 增 如 圖 10 及<br />

圖 11 所 示 之 局 部 橋 墩 垂 直 撓 曲 振 態 外 , 其 他 主 梁 不 同 振 動 方 向 的 各 個 振 態 頻 率 與 對 應 振 形 基 本 上 皆<br />

不 見 改 變 , 由 此 亦 可 反 證 我 們 在 前 節 中 簡 化 P2 橋 墩 為 邊 界 支 承 來 分 析 各 個 主 梁 振 態 的 有 效 性 。<br />

根 據 新 建 立 的 完 整 有 限 元 素 模 型 , 本 研 究 進 而 改 變 橋 墩 樁 基 礎 之 土 層 深 度 來 進 行 參 數 分 析 , 希<br />

望 由 此 決 定 出 對 於 沖 刷 效 應 最 為 敏 感 的 關 鍵 振 態 頻 率 。 我 們 將 竣 工 圖 所 對 應 之 樁 基 土 壤 高 度 定 為 0m<br />

高 程 , 而 後 陸 續 探 討 往 下 沖 刷 2m、4m、6m、8m 以 至 10m 等 情 形 , 這 些 不 同 沖 刷 深 度 下 的 各 方 向<br />

分 析 振 態 頻 率 整 理 於 表 7。 表 7 中 先 簡 單 列 出 主 梁 在 垂 直 撓 曲 、 水 平 撓 曲 、 車 行 軸 向 與 扭 轉 向 第 一<br />

振 態 的 頻 率 , 可 以 清 楚 看 出 這 些 主 梁 的 振 態 頻 率 完 全 不 受 P2 橋 墩 沖 刷 深 度 的 影 響 。 另 外 , 表 7 更<br />

同 時 列 出 如 圖 10 及 圖 11 所 示 之 局 部 橋 墩 垂 直 撓 曲 前 兩 振 態 (PIER-BV1, PIER-BV2) 頻 率 , 此 二 者 明<br />

圖 9 高 屏 溪 斜 張 橋 之 SAP2000 有 限 元 素 分 析 模 型 ( 納 入 P2 橋 墩 )<br />

圖 10 局 部 橋 墩 垂 直 撓 曲 第 一 振 態 振 形<br />

圖 11 局 部 橋 墩 垂 直 撓 曲 第 二 振 態 振 形<br />

-522-


表 7 在 P2 橋 墩 不 同 沖 刷 深 度 下 各 振 態 頻 率 之 比 較<br />

Mode<br />

Modal Frequency for Different Scouring Depth of Pier (Hz)<br />

0m 2m 4m 6m 8m 10m<br />

BV1 0.273 0.273 0.273 0.273 0.273 0.273<br />

BH1 0.960 0.960 0.960 0.960 0.960 0.960<br />

A1 2.416 2.416 2.416 2.416 2.416 2.416<br />

T1 0.741 0.741 0.741 0.741 0.741 0.741<br />

PIER-BV1 1.209 1.193 1.171 1.138 1.089 1.089<br />

PIER-BV2 3.369 2.985 2.605 2.280 2.001 1.779<br />

顯 隨 沖 刷 深 度 之 增 加 而 降 低 其 值 , 尤 以 第 二 振 態 特 別 敏 感 。 因 此 , 本 研 究 決 定 選 取 這 兩 個 關 鍵 的 振<br />

態 頻 率 作 為 後 續 評 估 沖 刷 現 況 的 基 準 。<br />

六 、 沖 刷 現 況 評 估<br />

根 據 前 節 的 結 果 , 本 研 究 先 定 義 振 態 頻 率 之 平 均 相 對 誤 差 作 為 基 準 設 計 目 標 函 數 如 下 :<br />

E =<br />

1 n ⎛<br />

∑ ⎜<br />

n i=<br />

1<br />

⎝<br />

fi<br />

- fˆ<br />

i<br />

fˆ<br />

i<br />

⎞<br />

⎟<br />

⎠<br />

2<br />

(3)<br />

(3) 式 中 f i<br />

代 表 有 限 元 素 模 型 之 分 析 振 態 頻 率 , 而 fˆ<br />

i 則 表 示 實 測 之 振 態 頻 率 , 同 時 本 研 究 取 PIER-BV1<br />

與 PIER-BV2 兩 個 關 鍵 振 態 頻 率 進 行 比 對 , 亦 即 n = 2。 然 後 改 變 有 限 元 素 模 型 中 橋 墩 基 礎 之 土 層 深<br />

度 (1m、3m、5m、7m、9m) 與 勁 度 (100%、110%、120%、130%、140%) 進 行 參 數 分 析 , 並 配 合 對 高<br />

屏 溪 斜 張 橋 P2 墩 頂 現 地 實 測 所 識 別 之 PIER-BV1 與 PIER-BV2 兩 個 振 態 頻 率 代 入 (3) 式 計 算 , 其 結<br />

果 整 理 於 表 8。 由 表 中 最 小 平 均 相 對 誤 差 0.015 足 以 評 估 判 斷 , 最 可 能 的 沖 刷 現 況 應 為 5m 左 右 的 沖<br />

刷 深 度 , 同 時 土 壤 勁 度 大 致 約 為 (2) 及 (3) 兩 式 平 均 所 估 算 的 120%。<br />

表 8 不 同 橋 墩 土 壤 勁 度 與 沖 刷 深 度 下 之 振 態 頻 率 平 均 相 對 誤 差<br />

Scouring<br />

Depth (m)<br />

Weighted Soil Stiffness (%)<br />

100% 110% 120% 130% 140%<br />

1m 0.128 0.150 0.184 0.210 0.235<br />

3m 0.047 0.064 0.097 0.120 0.142<br />

5m 0.037 0.032 0.015 0.035 0.055<br />

7m 0.108 0.099 0.066 0.049 0.041<br />

9m 0.186 0.181 0.147 0.130 0.114<br />

七 、 結 論<br />

本 文 首 先 針 對 主 梁 橋 面 、 箱 型 梁 內 部 及 橋 墩 頂 進 行 各 振 動 方 向 之 微 振 量 測 , 而 後 據 此 識 別 主 梁<br />

-523-


垂 直 、 水 流 、 車 行 、 扭 轉 向 及 局 部 橋 墩 車 行 向 之 各 個 實 測 振 態 頻 率 。 接 著 以 SAP2000 軟 體 建 立 斜 張<br />

橋 結 構 體 之 有 限 元 素 模 型 , 並 配 合 原 始 設 計 之 各 項 參 數 進 行 振 態 頻 率 分 析 。 結 合 前 述 兩 項 結 果 , 本<br />

研 究 進 一 步 以 趨 近 這 些 實 測 振 態 頻 率 作 為 標 的 , 交 叉 分 析 出 斜 張 橋 有 限 元 素 模 型 中 最 符 合 現 地 實 況<br />

之 各 項 邊 界 支 承 條 件 。 再 來 則 改 變 有 限 元 素 模 型 中 橋 墩 基 礎 之 土 層 深 度 進 行 參 數 分 析 , 由 此 決 定 出<br />

對 於 沖 刷 效 應 最 為 敏 感 的 幾 個 關 鍵 振 態 頻 率 。 最 後 以 所 得 關 鍵 振 態 頻 率 為 基 準 設 計 目 標 誤 差 函 數 ,<br />

然 後 改 變 有 限 元 素 模 型 中 橋 墩 基 礎 之 土 層 深 度 與 勁 度 參 數 進 行 最 佳 化 分 析 , 由 此 便 足 以 有 效 評 估 出<br />

最 可 能 的 洪 水 沖 刷 現 況 。 由 此 一 方 面 期 能 開 拓 透 過 橋 體 訊 號 間 接 進 行 沖 刷 監 測 的 可 行 性 , 同 時 更 寄<br />

望 從 而 彰 顯 斜 張 橋 利 於 監 測 以 降 低 洪 水 沖 刷 危 害 的 優 越 性 。<br />

参 考 文 献<br />

[1] Yankielun, N. E. and Zabilansky, L. Laboratory investigation of time-domain reflectometry system for monitoring bridge<br />

scour. Journal of Hydraulic Engineering, 1999, 125(12): 1279-1284.<br />

[2] Yankielun, N. E. and Zabilansky, L. J. Bridge Scour Detection and Monitoring Apparatus Using Time Domain<br />

Reflectometry (TDR). United States Patent, No. 6100700.<br />

[3] Park, I., Lee,J. and Cho, W. Assessment of Bridge Scour and Riverbed Variation by a Ground Penetrating Radar.<br />

Proceedings of the Third IEEE Conference on Sensors, Vienna, Austria, 2004, 1: 411-414.<br />

[4] Vieux, B. and Vieux, J. Integrated Radar and Hydrologic Modeling for a Bridge Scour Monitoring System. Proceedings<br />

of the Fifth European Conference on Radar in Meteorology and Hydrology, Helsinki, Finland. 2008<br />

[5] Lin, Y. B., Chang, K. C. and La, J. S. Applications of Optical Fiber Sensor on Local Scour Monitoring. Proceedings of<br />

the Third IEEE Conference on Sensors, Vienna, Austria, 2004, 2: 832-835.<br />

[6] Lin, Y. B., Chen, J. C., Chang, K. C. and Lai, J. S. Real-time Monitoring of Local Scour by Using Fiber Bragg Grating<br />

Sensors. Smart Materials and Structures, 2005, 14(4): 664-670.<br />

[7] Zhang, R. R., King, Olson, L. and Xu, Y.-L. Dynamic Response of the Trinity River Relief Bridge to Controlled Pile<br />

Damage Modeling and Experimental Data Analysis Comparing Fourier and Hilbert-Huang Techniques. Journal of Sound<br />

and Vibration, 2005, 285(4-5): 1049-1070.<br />

[8] Dey, S. and Barbhuiya, A. K. Velocity and Turbulence in a Scour Hole at a Vertical-wall Abutment. Flow Measurement<br />

and Instrumentation, 2006, 17(1): 13-21.<br />

[9] 李 維 峰 , 張 嘉 峰 , 梅 興 泰 . 以 振 動 量 測 檢 驗 橋 基 沖 刷 , 營 建 知 訊 期 刊 ,2008, 304: 54-65.<br />

[10] 柯 永 彥 , 張 為 光 , 陳 正 興 . 結 構 基 礎 損 傷 評 估 技 術 發 展 (I), 國 家 地 震 工 程 研 究 中 心 研 究 報 告 , 編 號 :<br />

NCREE-10-004, 2010.<br />

[11] Samizo, M., Watanabe, S., Fuchiwaki, A. and Sugiyama, T. Evaluation of the Structural Integrity of Bridge Pier<br />

Foundations Using Microtremors in Flood Conditions. Quarterly Report of the Railway Technical Research Institute, 2007,<br />

48(3): 153-157.<br />

-524-

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!