24 EACS2012 Expert System for the Preliminary Design of a Seismic Isolation Scheme for Bridges

EACS 2012 – 5 th European Conference on Structural Control Genoa, Italy – 18-20 June 2012
Paper No. # XXX
Expert System for the Preliminary Design of a
Seismic Isolation Scheme for Bridges
George MANOS 1, Stergios MITOULIS 2*
Laboratory of Experimental Strength of Materials, Dept. of Civil Engineering,
Aristotle University of Thessaloniki, Greece
ABSTRACT
Seismic design of isolated bridges, either new or existing ones that need retrofitting, involves
conceptual, preliminary and detailed design. However, preliminary design is a non-straightforward procedure
because of the sensitivity of bridge response on the initial decisions made by the designer, as well as on a
series of broader criteria such as serviceability, target performance level and cost-effectiveness of the various
design alternatives. Given the lack of design guidelines to ensure compliance with the above requirements, a
“trial and error” procedure is typically followed in the design office to decide on the most appropriate
isolation system; the latter typically based on engineering judgment to balance performance with cost. To
this end, the particular research effort aims to develop a decision-making system for the optimal preliminary
design of seismically isolated bridges. The proposed decision-making process is based on the current design
provisions of Eurocode 8, but is complemented by additional criteria set according to expert judgment and
laboratory testing, while using a combined cost/performance criterion. Software is also developed for the
implementation of the system. The paper concludes with the validation of the methodology and software
developed through more rigorous MDOF numerical analysis for the case of a real bridge.
Keywords: bridges, seismic isolation, preliminary design, software, experiment.
1 INTRODUCTION
The main objective of this research is to provide tools for helping design the isolation system
of relatively regular short to medium span bridge structures as highway overpasses with the
emphasis being on their earthquake performance and the inclusion of elastomeric bearings in their
design. This objective addresses both new and existing structures in an effort to upgrade their
earthquake performance [1] [2] and is motivated by the fact that during the last decade, an extensive
road network has been constructed in Northern Greece with more than 646 bridges built of a total of
40km length. These are structures of relatively small total lengths (i.e. L

University and c) the development of a an expert system which can link the commercially available
elastomeric bearings or the ones to be produced in the selection of an appropriate elastomeric
bearing scheme for typical R/C bridge overpasses in Greece (figure 1) [4].
L
Type 1: One span
Bearings on
abutments
L 1 L L 1
Type 3: Three spans
Bearings on piers and
abutments
L L L 1 L L L 1 1
Type 2: Two spans
Bearings on piers
and abutments
Type 4: Four spans
Bearings on piers and
abutments
Figure 1 - Overpasses with 1,2,3 and 4 spans and variable span lengths.
A series of tests were performed on unit elastomeric slice specimens of properties and
dimensions representative of the pilot bearings as described by current codes [5] [6] [7]. Summary
results are presented in this paper of the aforementioned ongoing experimental investigation along
with the measured mechanical characteristics under standard loading sequences. Numerical
simulations are also performed for validation purposes and an expert system is developed.
Summary information of this expert system is presented herein; it is aimed at selecting, during
preliminary design, of a set of ‘optimal’ bearings from a database of commercially available ones,
inclusive the foreseen pilot production. This selection process is based on the criteria set by the
Greek Guidelines that conform with the relevant clauses of the Eurocodes, for seismic isolation,
engineering judgment as well as on international standards. It is believed that the above described
scheme can contribute towards a better link between production that is based on local industries and
needs, experimentally-based optimization and final seismic design of bridges.
2 MECHANICAL CHARACTERISTICS OF RUBBER BEARINGS
A series of standard tests were performed at the Laboratory of Strength of Materials and
Structures of Aristotle University according to the [8]. Initially, these tests were used as
qualification tests for the materials used in the production; that is the elastomer, the steel plates and
the adhesion materials and processes. Next, in an effort to study the influence of certain parameters
in the mechanical characteristics of these elastomeric bearings, the vulcanization process was
investigated and in particular influence of the time duration in the mechanical characteristics of the
final product. Thus, a set of specimens were produced and tested with their vulcanization lasting
either 15minutes or 25 minute.
2.1 Shear cyclic tests of unit slices of elastomeric bearings
Shear cyclic tests of a specimens made of two unit slices were performed according to the [8].
Each unit slice included a layer of elastomer and two steel plates from a bearing with plan
dimensions 200mm by 200mm. The dimensions of each slice of elastomer were 200mm by 200mm
and 7.62mm thickness. Thus, the tested specimens were formed by two slices of elastomer and four
steel plates. Each steel plate had a thickness of 2.94mm and dimensions in plan of 200mm by
200mm. Figure 2 depicts the loading arrangement. The final specimen was of relatively large
dimensions as to be in plan a one to one representations of elastomeric bearings produced by the
2

same process; that is employing identical unit slices and building it up at the desired height with the
appropriate number of such unit slices [9]. A dynamic actuator of considerable displacement and
force capability was utilized to introduce a series of cyclic shear strain imposed loading sequences
to the specimens (see Figure 2). Initially, the series of tests did not exceed a maximum strain level
of 100%. Next a series of similar tests introduced maximum shear strain levels larger than 100% up
to the failure of the specimen that appeared in the form of unbonding of the elastomer from the steel
plating.
Apart from test specimens with the above dimensions, additional tests were also performed that
corresponded to cylindrical elastomeric bearings having as plan dimensions of each slice of
elastomer a diameter of 250mm with 7.62mm thickness. These specimens were produced with the
same process described before. Next, typical tests results are presented in brief. Throughout all the
tests the applied load producing the shear strains was monitored together with the corresponding
displacements of the specimen that were utilized to deduce the applied shear stress and shear strain
levels to the specimen. At the same time the applied vertical load, normal to the slices of neoprene,
was recorded and checked for any significant variations; the objective in this case being to keep the
vertical load almost constant at the range of 2.0 to 2.5 Mpa.
Figure 2 - Loading arrangement of a unit slice.
2.2 Shear cyclic tests with strain amplitude lower than 100%
Cyclic tests were performed for the following combinations: 3 to 11 cycles for each test, with
temperature 23 o C and cyclic loading varying with frequency 0.2Hz. The shear strain amplitude was
varied from 5% to 75% in the following steps: 5% (0.38mm), 10% (0.76mm), 25% (1.91mm), 50%
(3.81mm), and 75% (5.72mm). The corresponding results for the initial shear tests are depicted in
Figure 3. An increase in the shear stiffness was observed when the cyclic shear strain became larger
than 75%. Subsequent tests that followed the initial tests with shear strain amplitudes varying again
in the range from 5% to 75% exhibited an increase in the shear stiffness when they are compared
with the results of the initial shear tests (Figure 4). Again, the specimen exhibited a stable
performance throughout the increasing shear strain amplitude from 5% to 75%.
Additional tests were also performed with the same specimens whereby the studied variable
this time was the frequency of imposing the shear strain, keeping the maximum target strain
amplitude constant and equal to 75%. The corresponding results are illustrated in Figure 5. In this
case the specimen’s performance was examined for loading frequencies equal to 0.1Hz, 0.5Hz and
1.0Hz. As can be seen from this Figure no significant variation in the performance of the specimen
could be observed from the obtained response whereby the loading frequency was varied from
0.1Hz to 1.00Hz
3

shear stress (Mpa)
Slice Bearing 7.5mm thick
Test of varying shear 0.6
strain amplitude
5%, 10%, 25%,
0.3
50% 75%
0
-1.2 -0.6 0 0.6 1.2
-0.3
-0.6
shear strain
0.2Hz (Test 002)
Figure 3 - Initial shear cyclic test results for frequency 0.2Hz and shear strain amplitude
from 5% to 75%.
shear stress (Mpa)
Slice Bearing 7.5mm thick
0.6
Test 001 preceded
test 001a
0.3
0
-1.2 -0.6 0 0.6 1.2
-0.3
-0.6
shear strain
0.2Hz (Test 001)
0.2Hz (Test 001a)
Figure 4 - Subsequent shear cyclic test results for frequency 0.2Hz and shear strain amplitude
from 5% to 75%.
shear stress (Mpa)
Slice Bearing 7.5mm thick 200mm x 200mm
Tests with varying
frequency 0.1Hz,
0.5Hz and 1.0Hz.
Shear strain amplitude
75%
0.4
0.2
0
-1 -0.5 0 0.5 1
-0.2
-0.4
shear strain
(Test 003)
0.1Hz
0.5Hz
1.0Hz
Figure 5 - Shear cyclic test results for shear strain amplitude 75% and frequencies 0.1Hz,
0.5Hz and 1.0Hz.
An additional specimen of the same geometry and produced by the same process was tested
by the loading arrangement shown in Figure 6. This time, apart from imposing the shear strain
levels of continuously increasing amplitude, the specimen was placed under a constant compressive
stress field normal to the horizontal plane of the elastomer. This stress field corresponded to an
equivalent compressive stress equal to 2.4Mpa. The frequency of the cyclic load was equal to 0.5Hz
and the shear strain amplitude was continuously increasing from 10% to 75%. The summary results
of this test are depicted in Figure 7. By comparing the results of Figure 4 with those of Figure 7 an
increase in the stiffness and a decrease in the equivalent damping ratio is evident when the
specimen is subjected to the previously described compressive stress field of 2.4Mpa equivalent
uniform stress normal to the elastomer. However, this observation should not be generalized; as was
shown from the measurements of another investigation [10] further increase in this compressive
stress field has the opposite effect that is a decrease in the shear stiffness and an increase in the
equivalent damping ratio.
4

Figure 6 - Simultaneous application of a compressive stress field.
Shear stress (Mpa)
Test slice 200mm x 200mm (7.1mm thick) 0.5Hz
2.4Mpa Normal stress
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
Shear strain
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8
shear strain 10%, 20%, 30%
shear strain 30%, 50%, 75%
Figure 7 - Shear cyclic test results for frequency 0.5Hz and shear strain amplitude from 5%
to 75%.
2.3 Shear cyclic tests with strain amplitude higher than 100%
The previously described loading sequences were repeated again with unit slice specimens
being loaded with shear strain amplitudes higher than 100% up to the failure of the specimen. Two
distinct loading arrangements were again adopted. First, the shear strains were introduced without
the application of compressive load normal to the horizontal plane of the elastomer (Figure 8),
whereas in the second case a vertical load was applied and kept constant producing an equivalent
uniform compressive stress normal to the plane of the elastomer (Figure 6) in the range of 2.0 to 2.5
Mpa.
Figure 8 depicts the failing mode of the specimen during this loading process, without the
presence of the compressive stress field. Figure 9 depicts the results of this test with large shear
strains without the application of compressive normal stress, whereas Figures 8 and 10 show the
resulting unbonding of the elastomer from the steel plating at the end of this loading sequence. The
load-unload behavior of this specimen is shown in Figure 9 for levels of shear strain starting from
100% and gradually increasing to 275%. It can be observed that for shear strain levels lower than
200%, which corresponds to Eurocode’s [5] maximum allowable shear strain, the specimen’s
behavior remains stable even for this demanding test that corresponds to an elastomeric bearing that
does not have the beneficial stabilizing effect of the compressive stress field normal to the slices of
the elastomer within the bearing. The maximum capacity reached for this specimen was equal to
1.7Mpa; this occurred for a maximum strain level equal to 275%. The load-unload behavior of the
specimen with the simultaneous application of continuously increasing shear strains and an imposed
compressive stress field normal to the elastomer equal to 2.4Mpa is shown in Figure 11; Figure 12
depicts the failure mode. The drops of stress that can be observed in this plot are due to the
relaxation of the elastomer during small duration stops in the loading sequence. The levels of shear
strain reached 230% during the first loading cyclic; then they were gradually increased to 300%,
whereby the maximum bearing capacity equal to 2.0Mpa was reached for this specimen.
5

Figure 8 - Loading the unit slice specimen without the presence of
the compressive stress field.
It can be observed that for shear strain levels higher than 300% the specimen’s bearing
capacity degrades; however this degradation does not progress rapidly; for shear strain levels 350%
the specimen exhibits a bearing capacity almost equal to the maximum bearing capacity of the
specimen. From the comparison of the performance of the specimens with and without the
compressive stress field (Figures 11 and 9) it can be seen that the most severe test is that without
the normal compressive stress field.
3 AN EXPERT SYSTEM FOR THE PRELIMINARY DESIGN OF BRIDGE SEISMIC
ISOLATION
The process to select the cost-effective seismic isolation system is often time consuming, as it
commonly leads to iterative calculations and numerical analyses [3] and several design checks
against target code-based criteria concerning both the maximum bearing strain and the overall
performance of the bridge structure [5]. This was the reason that led to the development of an
expert system which aimed to become a design tool for Bridge Engineers in their preliminary stages
of selecting an appropriate and cost-effective scheme of rubber bearings that could be applied in
existing or new short to medium span regular bridges with typical structural forms and geometry. In
general, it can be claimed that no comprehensive procedure has been presented to this date for the
optimal (that is, cost-effective), preliminary design of seismically isolated highway bridges.
shear stress
(Mpa)
Slice Bearing 7.5mm thick
2
200mm x 200mm
1.6
1.2
0.8
0.4
0
0.2Hz 001 after
Tension Static 1
Tension Static 2
0 1 2 3
shear strain
Figure 9 - Large shear strains test without compressive normal stress.
6

3.1 Assumptions and limitations of the software
The rigid deck model used in the software for the preliminary analysis of the bridge, is
applicable when the total mass of the piers is less than 20% of the total bridge mass, as prescribed
by Eurocode 8-Part 2. Moreover, the software assumes that bridges under design are level and
straight; therefore structures with small curvature in plan or small longitudinal inclination can be
approximated by this assumption. The developed software is not limited by the choice of the deck
cross section. The user can employ a default vertical deck load value equal to 200 KN/m,
considering that the used combination of loads includes earthquake loading. Otherwise, the user
should input an appropriate vertical deck load value if the particular deck does not correspond to the
default value. The software can be used in all isolated bridges with low damping elastomeric
bearings with an effective damping not larger than 6% [5]. However, the validity of the software for
high damping rubber bearings (HDRBs) is currently investigated. The software cannot be used
when monolithic pier-deck or abutment-deck connections are combined with seismic isolation in a
bridge. This later structural scheme represents a design alternative implemented for long cast-in-situ
bridge decks or in irregular bridges, which have short piers and are usually protected from the
deck’s movements through isolation bearings.
As already mentioned, the software is developed for the preliminary design of the isolation
for bridges having up to four spans. This limitation is not related to the seismic action, but mostly to
the induced in-service movements of the deck. Serviceability movements become significant for
bridges with total lengths greater than 150 meters according to the limits imposed by various
transportation agencies as described by [11]. It was, consequently, decided that the simplified
software should be restricted to overpasses and typical small and intermediate span continuous deck
bridges, having a total length of approximately up to 150 meters.
The bearings are assumed to be located at the top of the mid-span piers or/and at the
abutments supporting the bridge deck. At first, all the structural parts supporting the elastomeric
bearings, such as the piers with their foundations and the abutments, are considered to have a much
higher stiffness in the horizontal direction than the stiffness provided by the elastomeric bearings.
Consequently, the software at this stage utilizes only the horizontal stiffness properties of the
elastomeric bearings and ignores influences arising from the stiffness characteristics of their
supporting structural parts (piers etc). It follows that this assumption presents a limitation in
applying this software to bridges with a flexible pier-foundation system, that is, to bridges with
slender or tall piers and/or with flexible foundations. To partially confront with this limitation an
additional investigation was performed aiming to study the effect of the pier’s flexibility. This was
done by examining parametrically the preliminary design of the seismic isolation system for typical
bridge-structures having 5 different types of piers, as shown in Figure 13.
Figure 10 - The unbonding of the elastomer from the steel plating
at the end of this loading sequence.
7

Shear stress (Mpa)
Test slice 200mm x 200mm (7.1mm thick) Static
0.5
2.4Mpa Normal stress
0
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
-0.5
Failure Test
shear strain 400%
shear strain 230%
Shear strain
Figure 11 - Large shear strains test with compressive normal stress.
The influence of twelve different types and dimensions of piers were investigated having
three different heights that conform to the required clearance for the bridge deck. This extra study
was based on the maximum deck displacement response results obtained by either ignoring or
including the flexibility of these piers. The obtained results demonstrate that the flexibility of the
piers becomes significant for piers with relatively small cross sectional dimension (circular, square
or rectangular sections) having a height greater than 15 m. This study also showed that the seismic
isolation system of the bridge is being adversely affected by the increasing flexibility of the piers.
This is due to the increase in the fundamental period of the bridge-elastomeric bearings-piers
dynamic system and the resulting increase of the total maximum deck displacement response. The
first translational mode, in the longitudinal direction, remains the predominant mode of response for
such an isolated structural system, even with relatively flexible piers. The results of this analysis
demonstrate that for the mid-span flexible piers the increased deck displacement demand is partially
provided by the displacement of the top of these piers; thus, the displacement demand of the
elastomeric bearing at these locations is smaller than the maximum deck displacement demand.
On the contrary, this increased deck displacement demand must be met at the abutments
solely by the elastomeric bearings at this location; this becomes critical for the design of the total
seismic isolation system. It was found, that the influence of the flexibility of the piers examined in
this way can result in almost 15% increase in the seismic displacement of the deck. When
comparing two identical seismically isolated bridges having as the only difference the flexibility of
the piers, both employing the same number of elastomeric bearings with the same dimensions in
plan, the increase in the deck displacement demand, due to the piers flexibility, will result in a 15%
corresponding increase in the height of the elastomer of the bearings for the structure with the
flexible piers. This height increase of the bearings will be dictated by the compliance to the safety
criterion incorporated in the code provisions, as clearly stated by Equations 3 and 5, outlined in
sections 3.2 and 3.3. It must be pointed out here that this performed study was based on the
possibility of multiple bearings per support; however, the software makes the assumption that, all
the bearings used for the support of the deck, either on top of the piers or at the abutments, have the
same geometry and are of the same number at each support. This is believed to be a reasonable
assumption for bridge structures up to 150m long, as, in this case, it is not useful to differentiate the
geometry of the bearings per support.
The software developed so far is also based on the assumption that the isolation system is
active only along the longitudinal direction. The response of the bridge in the transverse direction
was assumed to be restrained by seismic links, which join the deck with the piers’ heads. This
assumption is deemed reasonable, because most bridges with isolated decks use such seismic links
in the transverse direction ([5] Clause 6.6.3). For these structures the in-service movements,
because of creep or shrinkage [12] and prestressing [13], are negligible. Therefore, the software in
its current version does not deal with the simultaneous longitudinal and transverse deck earthquake
displacement response. The current extra study was extended to examine the case whereby the
bridge deck is not restrained by seismic links in the transverse direction. Three different deck
sections were examined for this purpose, as shown in Figure 14, for bridges with total lengths up to
150m The results of this study show that in this case the maximum deck displacement response for
8
-1
-1.5
-2
-2.5
-3

simultaneous longitudinal and transverse earthquake input motion is well approximated by the
SRSS summation of the longitudinal d Ed,X and the transverse d Ed,Y maximum deck displacement
response found separately along these two directions as expressed by the following equation:
d = d d
Υ
2 2
Ed Ed,X Ed,
+ (1)
Figure 12 - Failing mode of the unit slice specimen without the presence
of the compressive stress field.
Circular:
D
D=1m
D=1.5m
D=2m
Square:
a
Dimensions
1.0x1.0m
1.3x1.3m
1.8x1.8m
Hollow:
Multi
column:
Dεξ.=3.0m
Dεσ.=2.0m
Πάχος t=0.5m
Long.
Direction
Rectangular:
Long.
Direction b
h
Dimensions
1.0x5.0m
1.5x5.0m
1m
3m
Figure 13 - Different types of piers for typical short to medium span bridges.
3m
Figure 14 - Different types of decks for typical short to medium span bridges.
9

It is also shown that, due to the flexibility of the bridge system in the transverse direction, the
maximum deck response in the transverse direction d Ed,Y is up to 12% greater than the
corresponding transverse direction deck response d Ed,Y when the bridge system stiffness in the
transverse direction are quite stiff. As argued before, this is due to the increase in the fundamental
period of the bridge-elastomeric bearings-piers dynamic system. It is also shown that the in-plane
deformability of the deck in the transverse direction has negligible effect. Thus, for the transverse
direction the flexibility of the piers will also result in up to 12% increase deck displacement when
compared to the deck response when the piers at the transverse direction are quite stiff. These
influences: (a) Of the flexibility of the piers in the longitudinal and transverse direction and (b) Of
the simultaneous seismic input on both directions, are currently accommodated in a revised version
of the initial software for the preliminary design of the isolating bridge system.
3.2 Description of the software
Within the framework described in the introduction and depending on (a) the period of
construction and the corresponding seismic code used, (b) the pier type and cross-section, (c) the
deck type and (d) the pier-deck connection, the selected bridge overpasses have been classified to
four different categories based on the total number of spans (1 to 4 spans) that corresponds to
bridges up to say 150m. Furthermore, an electronic database has been developed specifically for
overpasses and commercially available and experimentally tested bearings that can be used in
design and construction. Numerous geometrical (rubber and steel plate thickness, height, width and
type) and material (shear modulus, maximum strain) data, have been archived for each (circular or
rectangular) bearing.
This developed expert system is capable to perform the selection of seismic isolation devices
among the ones stored in the inventory. In particular, an interactive and user-friendly environment
has been developed that elaborates the designer to define the structural configuration of the
overpass bridge and its corresponding seismic exposure. The later is based on Eurocode 8 Part 2.
First, the user defines (Figure 15 and Figure 16) the type of overpass (in terms of number of spans,
span and total length and locations of bearing-type pier-deck connections). Next, the number of
bearings at each support, the preferred bearing shape and the parameters that define seismic loading
(soil classification, seismic zone, importance category), are defined. Then, the system automatically
passes all bearings available in the database through a sequence of checks in order to filter out those
that do not comply with the design criteria set. This filtering process, illustrated in Figure 16, is
essentially broken down into two major checks, in order to ensure that the bearings satisfy
particular criteria related to their maximum compression and seismic deformation according to the
aforementioned guidelines. The two main relationships that control this performance are:
d
Ed
ε
s,d
= ≤ 2
(2)
t
t
ε
t,d
=ε
s,d
+ε
c,d
+εa,d ≤ 7 /1.15 = 6.09
(3)
where ε s,d the calculated maximum shear strain for the elastomeric bearing scheme due to the
design horizontal earthquake and ε c,d the calculated maximum shear strain for all the design vertical
actions, that corresponds to permanent loading plus 20% of the variable loads [14].
10

Figure 15 - Typical bridge configurations supported by the software and user’s interface.
Figure 16 - Program flow for evaluating the bearing performance (indicative).
11

3.3 Combined cost-performance criteria for the selection of bridge isolation systems
Once a set of bearings is found to meet the above criteria, the system automatically rates and
ranks them by using an ‘optimal performance’ (OP) criterion which aims at quantifying the relative
safety that provides each bearing given its actual cost. This factor is expressed in the form of the
following equation:
OP
(i)
SC
(i)
= (4)
CC
(i)
where SC (i) is a factor indicating the degree of compliance to the safety criterion incorporated
in the code provisions as clearly stated by Equation 3. The value for this safety criterion is derived
for each eligible bearing i through the following formula as:
SC
ε
s,max(i)
(i)
= (5)
εs,d(i)
ε s,max(i) is in turn, the allowable (by these code provisions) maximum shear strain and ε s,d(i) is
the predicted shear strain due to the total design seismic displacement. When this safety criterion
has a value just larger than one (1.0) the compliance to this safety criterion is only just satisfied
according to the code safety requirements, whereas when this value is larger than one (1.0) this
compliance to the code safety requirements has a corresponding margin.
CC
FC
(i)
(i)
= (6)
minFC(i)
where FC (i) is the corresponding final item cost of each bearing i expressed in Euro currency
(€) that is approximated as:
FC (i) = sc + Vb (i) ·c v (in €) (7)
that is, as a function of the bearing volume Vb(i), the cost per unit volume cv and the standard
structural cost (sc) related to the bearing volume.
4 CONCLUSIONS
From the test results with frequencies ranging from 0.2Hz to 1.0Hz it was demonstrated that
the stable performance of the tested unit elastomeric slice specimens was not affected by the
frequency of the imposed cyclic shear strain loading.
The presence of a relatively mild compressive field is beneficial as it extends the stable
behaviour of the tested specimens for shear strain amplitude up to 300%. Moreover, for shear strain
levels higher than 300% the specimen’s bearing capacity degrades; however this degradation does
not progress rapidly.
An expert system was developed aimed to become a design tool for professionals in their
preliminary stages of selecting an appropriate and cost-effective scheme of elastomeric bearings.
12

One of the main criteria for selecting a scheme of elastomeric bearings is that it has to meet
the requirements posed by the Eurocode, especially those safeguarding the satisfactory seismic
behaviour.
The flexibility of the piers in both bridge directions and the flexibility of the deck in the
transverse bridge response typically increase the periods of the bridge and hence the displacements
and the requirements of the seismic isolation system in terms of bearings’ total thickness of the
elastomer. More specifically the piers’ flexibility led to an up to 15 and 10% increase in the
longitudinal and transverse seismic movement, of the deck and to the same increase in the bearings’
required total height of the elastomer.
Acknowledgements
The project is funded by the General Secretariat of Research and Technology of the Greek
Ministry of Development; its support is here gratefully acknowledged.
REFERENCES
[1] Dolce, M. and Marnetto, R. 1998. SMA devices for seismic isolation of bridges: design and
experime-ntal behaviour. U.S.-Italy W. on Seismic Protective Systems for Bridges, MCEER.
pp. 313-333.
[2] Priestley, M.J.N., Seible, F. and Calvi, G.M. 1996. Seismic design and retrofit of bridges. John
Wiley and Sons, New York.
[3] Naeim, F. and Kelly, J.M. 1999. Design of seismic isolated structures, From theory to practice.
John Wiley and Sons, Inc.
[4] Calvi, G.M., Pavese, A., 1998. Optimal design of isolated bridges and isolation systems for
existing bridges. In Proc. of U.S.-Italy Workshop on Seismic Protective Systems for Bridges.
MCEER. pp. 407-429.
[5] Eurocode 8 2004. Design of structures for earthquake resistance - Part 2: Bridges, DRAFT No
3. European Committee for Standardization.
[6] EN 1337-3 (European Standard) 2005. Structural bearings - Part 3: Elastomeric bearings.
European Committee for Standardization.
[7] EN 15129 2009. Anti-seismic devices. European Committee for Standardization.
[8] International Standard ISO 22762-1 2005. Elastomeric seismic protection isolators, Part1: Test
methods.
[9] Abe, M., Yoshida, J., Fujino, Y. 2004. Multi-axial Behaviors of Laminated Rubber Bearings
and their Modeling. I: Experimental Study. ASCE Journal of Structural Engineering. 130(8):
1119-1132.
[10] Ryan K.L., Kelly J.M., Chopra A.K. 2004. Experimental observation of axial load effects in
isolation bearings. 13 th World Conference on Earthquake Engineering, Canada. Paper No.
1707.
[11] Dicleli, M., Albhaisi, S.M., 2004. Estimation of Length Limits for Integral Bridges Built on
Clay. ASCE Journal of Bridge Engineering. 9(6): 572-581.
13

[12] Arockiasamy, M., Sivakumar, M., 2005. Design Implications of Creep and Shrinkage in
Integral Abutment Bridges. ACI Special Publication. 227(5):85-106.
[13] PCI, Precast, prestressed concrete bridges, the high performance solution, Comprehensive
Bridge Design Manual, Accessed online: 1 Dec 2008 (Available via DIALOG.
ww.pci.org/publications/bridge).
[14] Eurocode 1 2003. Basis of design and actions on structures - Part 2: Traffic loads on
bridges. European Committee for Standardization.
14