13.06.2014 Views

Nonlinear Finite Element Analysis of Concrete Structures

Nonlinear Finite Element Analysis of Concrete Structures

Nonlinear Finite Element Analysis of Concrete Structures

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

- 36 -<br />

affecting the descending curve in the post-failure region. Eq.<br />

(3) is a four-parameter expression determined by the parameters<br />

a , e , E., and D, and it infers that the initial slope is E.,<br />

and that there is a zero slope at failure, where (a,e) = (- c ,<br />

-e ) satisfies the equation. The parameter D determines the postfailure<br />

behaviour, and even though there are some indications <strong>of</strong><br />

this behaviour, e.g. Karsan and Jirsa U969),the precise form <strong>of</strong><br />

this part <strong>of</strong> the curve is unknown and is in fact, not obtained<br />

by a standard uniaxial compressive test. Therefore, the actual<br />

value <strong>of</strong> D is simply chosen so that a convenient post-failure<br />

curve results. However, there are certain limitations to D, if<br />

eq. (3) is to reflect: (1) an increasing function without inflexion<br />

points before failure; (2) a decreasing function with at<br />

most one inflexion point after failure; (3) a residual strength<br />

equal to zero after sufficiently large strain. To achieve these<br />

features A > 4/3 must hold, and the parameter D is subject to<br />

the following restrictions<br />

(1-35A) 2 < D _< l+A(A-2) when A < 2;<br />

0 _< D 4/3 is in practice not a restriction, and, in<br />

fact, eq. (3) provides a very flexible procedure to simulate the<br />

uniaxial stress-strain curve. For instance, the proposal <strong>of</strong> Saenz<br />

(1964) follows when D = 1, the Hognestad parabola (1951) follows<br />

when A = 2 and D = 0, and the suggestion <strong>of</strong> Desayi and Krishnan<br />

(1964) follows when A = 2 and D = 1. In addition, different postfailure<br />

behaviours can be simulated by means <strong>of</strong> the parameter D<br />

and this affects only the behaviour before failure insignificantly.<br />

This is shown in fig. 3, where A = 2 is assumed and where<br />

the limits <strong>of</strong> n are given by zero and unity.<br />

Using simple algebra, eq. (3) can be solved to obtain the actual<br />

secant value E<br />

s<br />

<strong>of</strong> Young's modulus. The expression for E<br />

s<br />

con-<br />

tains the actual stress in terms <strong>of</strong> the ratio - 0/0 . For unic<br />

axial compressive loading £ = - a/a<br />

for E<br />

holds, and the expression<br />

can therefore be generalized to triaxial compressive load-

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!