See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/290075661

Seismic Design of Bridges for Prevention of Girder Pounding

Article in Electronic Journal of Structural Engineering · March 2008

DOI: 10.56748/ejse.9301

CITATIONS READS

15 945

2 authors, including:

Hong Hao
Curtin University
915 PUBLICATIONS 30,938 CITATIONS

SEE PROFILE

All content following this page was uploaded by Hong Hao on 18 July 2016.

The user has requested enhancement of the downloaded file.

 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

Seismic Design of Bridges for Prevention of Girder Pounding

H. Hao1, N. Chouw2
The University of Western Australia, Australia1
The University of Auckland, New Zealand, Email: [email protected]

ABSTRACT: In design of conventional bridges the gap between bridge spans is usually only a few centime-
tres. For such an expansion joint poundings between adjacent bridge girders during strong earthquake shaking
are usually unavoidable. Pounding often causes damages to girders. In extreme situations it may push one of
the bridge decks off the support. In this work a new design philosophy using a modular expansion joint (MEJ)
is introduced. So far MEJs have been used mainly to cope with large thermal expansion and contraction of
long bridges. For proper design of bridges to avoid the consequence of poundings under strong earthquakes
not only a minimum total gap but also the maximum opening movement of the joint are essential. In this study
the simultaneous effect of varying vibration properties of adjacent bridge spans, spatially varying ground exci-
tations and soil-structure interaction on the total closing and opening movements of a MEJ, required to elimi-
nate possible pounding and to ensure the join in perfect working order, is estimated, and the main influence
factors are discussed.

Keywords: Modular expansion joint, spatial variation, soil-structure interaction, separation distance, pounding
prevention

1 INTRODUCTION by the different footing properties and non-uniform

ground conditions as well as by the interaction be-
tween buildings and subsoil, which may induce out-
Pounding can cause substantial damage to neigh- of-phase responses between adjacent buildings. In
bouring structures when they are sufficiently close to the case of bridge structures, besides soil-structure
each other. Despite research effort in the past dec- interaction (SSI), inevitable spatially non-uniform
ades and recommendations provided in almost all ground excitation at the neighbouring bridge pier
seismic design codes, pounding damage of bridge supports is another factor that may produce out-of-
girders has still been observed in many recent major phase responses of adjacent bridge spans. In such a
earthquakes, e.g. during the 1994 Northridge earth- case the current design recommendations can cause
quake (Moehle et al. 1995), the Kobe earthquake an adverse effect (Chouw & Hao, 2008). While a
(Park et al. 1995), the Chi-Chi earthquake (JSCE minimum distance between buildings is a possible
1999) and the Yogyakarta earthquake (Elnashai et al. measure for avoiding pounding damage, in the case
2007). Current design regulations, e.g. CALTRANS of bridge structures a large gap between adjacent
(1999), AASHTO (1998) and JRA (2004), recom- girders will strongly hinder the passage of traffic. An
mend that adjacent structures should have a suffi- adjustment of the fundamental frequencies of the ad-
cient separation distance and the same or at least jacent bridge structures may not be a sufficient or
similar fundamental vibration period to ensure their suitable approach to reduce out-of-phase responses,
in-phase vibrations. The idea is that this in-phase because bridge structures will more likely experience
overall vibration will then prevent the adjacent struc- spatially non-uniform ground excitation. To over-
tures from colliding. This recommendation, how- come this difficulty, in this work a new design phi-
ever, is made under the assumption that the struc- losophy is introduced.
tures experience the same ground excitation, and
their behaviour is determined only by the structural
properties. In the case of adjacent buildings an as-
sumption of same ground excitation is justifiable.
However, their dynamic behaviour can be affected

133
 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

2 CAUSE OF RELATIVE MOVEMENT When adjacent bridge segments have different

BETWEEN BRIDGE GIRDERS slenderness, even if they have the same fixed-base
fundamental frequency and experience the same
There are a number of sources that induce relative ground excitation and the subsoil at the bridge piers
movement between bridge girders. They include: is uniform with the same properties, relative re-
- Unequal vibration properties of adjacent bridge sponse between the bridge segments will occur. This
structures is because both bridge structures will interact differ-
- Spatial variation of ground motions ently with the supporting soil. Consequently, each
- Different soil-structure interaction because of va- structure responds differently and contributes to their
rying soil properties at bridge supports and be- relative movements. The influence of the soil on the
cause of different footings and slenderness ratios period T%i of the system (bridge segment including
of adjacent bridge piers subsoil) can be easily seen from Equation (1).

The response of a structure is as good as the as- k k h2

T%i = Ti 1 + i (1 + i x i ) (1)
sumption of the excitation. In the case of earthquake ki x ki φ
loading the ground excitation of the structure is de-
termined not only by the geological properties along where Ti is the period of the bridge segment with an
the path of wave propagation from the source to the assumed fixed base (without the influence of sup-
structural site and the properties of the local site, but porting soil). ki , ki x , ki φ are the bending stiffness of
also by the stiffness ratio between the structural foot- the bridge structure, and the static soil stiffness for
ing and the supporting ground. The embedment of horizontal and rocking movements of the assumed
the footing and the interface between footing and rigid bridge footing, respectively.
subsoil can also alter the characteristics of the in- In Equation (1) the influence of the vertical soil
coming waves and consequently the ground excita- stiffness is assumed to be negligible. In reality the
tion of the structure. soil stiffness is frequency dependent which makes
The inertial forces activated by the excitation the influence of subsoil even more complex (see e.g.
cause vibrations of the whole structure. If the struc- Sieffert & Cevaer 1992). From this equation and
ture is low, the structural footing tends to move lat- Figure 1 it is obvious that even if both bridge seg-
erally from side to side. This movement initiates ments have the same fixed-base fundamental period
pressure waves at both sides of the footing and at the ( T1 = T2 ) and the same supporting ground ( k1x = k2 x
same time shear waves from the bottom interface of and k1φ = k2φ ), the different structural heights (h1 ≠ h2)
the footing. If the structure is tall, the footing rocks will cause different system periods, and conse-
because of the bending movement of the whole quently, each bridge segment will respond differ-
structure. This footing movement activates mainly ently, even if the ground excitation is the same.
pressure waves at both edges of the footing and
shear waves at both sides of the footing. Since vi-
brating footing is a wave source, different movement
of the footing means different wave propagation, and h1
consequently different dynamic soil stiffness and ra- h2
diation damping will be initiated.
In the case of long-extended structures, like cs1, ν1 , β1
bridges, the soil plays another significant role. Since
cs2, ν2, β2
seismic waves need time to travel from one bridge
pier footing to another pier support, the ground exci-
tation at the adjacent support experiences a time de- Figure 1 Soil-structure system.
lay. Additionally, the soil properties between bridge
supports are heterogeneous and non-uniform; there- In reality the local site at adjacent bridge pier
fore ground motions at the supports are incoherent. supports is normally not the same. Hence, non-
Consequently, the response of structures to spatially uniform soil properties contribute additionally to the
varying ground excitations is not the same as that to relative response between the bridge structures.
commonly-assumed uniform ground motions. In the Because of the nature of the structures to span a
case of bridge structures the spatial variation of the valley or river, often the subsoil is soft. The profile
ground excitations causes different relative response of soft supporting sediments can additionally am-
between the adjoining bridge segments as those un- plify the incoming seismic waves and consequently
der the assumption that both bridge segments ex- the ground excitation of the bridge supports. Figure
perience the same ground motions. 1 shows a possible complex system. The adjacent
bridge structures have different slenderness, and the

134
 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

supporting subsoil has different profiles and proper- 3 NEW DESIGN PHILOSOPHY
ties: shear wave velocity cs, Poisson’s ratio ν and
material damping β. Several causes of structural Recently, to cope with large thermal expansion and
poundings are described, e.g. in the reference contraction of long bridges more and more modular
(Chouw 2005). expansion joints (MEJ) are used. Figure 2 shows two
The consequence of relative girder movements for segments of a bridge with a MEJ. The upper figure
adjacent bridge structures is damage to the bridge displays the cross section of the joint in the longitu-
decks or even collapse. When closing relative dinal direction of the bridge. The bridge segments
movements are larger than the gap between adjacent are connected by edge beams at both girder ends and
girders, pounding between the girders will take by middle beams. Support beams and rubber bear-
place. Depending on the magnitude of the relative ings transfer the traffic loading from the joint to the
girder movement the damage can result in a bridge adjoining bridge girders. To ensure watertightness of
that is out of function. the joint, free and moveable rubber sealing is in-
When opening relative movements are larger than stalled between the beams. The bearings ensure that
the seat length at the expansion joint unseating and a the beams move uniformly. Details of MEJ can be
subsequent collapse of the bridge deck can happen. found, e.g. in the work by Dexter et al. (2002).
To prevent the consequence of large opening relative The authors propose to apply the ability of MEJ
movements between bridge girders several measures to mitigate the pounding problems due to large rela-
have been developed and applied, e.g. stoppers and tive movement between bridge girders. Up to now,
restrainers at the expansion joints (see e.g. Yashinsky the suitability of MEJ to mitigate pounding damages
& Karshenas 2003). It is, however, so far not possi- of girders under strong earthquakes is unknown.
ble to prevent girder pounding under strong earth- So far investigations of MEJ have been focused
quakes, because in order to ensure a serviceability of mainly on traffic-induced noise (e.g. Ravshanovich et
the bridge the gap at the expansion joints should be al., 2007) and long-term MEJ fatigue behaviour due
not more than a few centimetres. This small gap is, to repeated vehicle loading and continuous opening
however, usually not sufficient to prevent adjacent and closing movements of the MEJ beams.
bridge decks from colliding into each other. In this
work a new design approach is introduced that en-
able large closing movement at the expansion joint
without causing any pounding and at the same time
the bridge serviceability is ensured.

Figure 2 Bridge structures with subsoil and modular expansion joint.

135
 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

The most significant requirement for proper de- The bridge structures with their footings and the
sign of a MEJ to cope with strong earthquake in- subsoil are described by finite elements and bound-
duced relative movement between bridge girders is ary elements, respectively.
the minimum total gap between MEJ beams to pre- Because in the 2D analysis an exact shape func-
vent pounding. To ensure a perfectly working condi- tion (continuous-mass model) is used, only one finite
tion the joint must also be able to open without tear- element is necessary for each bridge pier and each
ing off the seals. Since the MEJ system ensures a girder. To couple the footing with the subsoil, 8
uniform movement between the beams, in the inves- boundary elements are used to model the supporting
tigation the influence of the rubber sealing is consid- soil at each footing-soil interface. The algorithm for
ered to be negligible. Instead, the investigation fo- the calculation of the girder responses with non-
cuses on the most significant influence factors linear soil-structure interaction is described in the
identified in previous studies (Chouw & Hao reference (Chouw & Hao 2008).
2008): For simplicity the multiple piers of the left and
- Characteristics of the spatially varying ground right bridge segments are described as a single pier,
excitation: coherency loss and wave apparent ve- and the distance between these left and right mod-
locities eled piers is assumed to be 100 m. It should be noted
- Ratio of the fundamental frequencies of the adja- that to obtain a precise result the influence of multi-
cent bridge structures ple piers of each bridge segment and different
- Interaction between bridge structures and subsoil ground excitations of these piers should be consid-
- Combined effect of these factors ered.
The considered left and right bridge structures in
Figure 2 have the heights of 12.2 m and 18.3 m, re-
spectively. To focus on the influence factors it is as-
sumed that both structures have very similar fixed-
base fundamental frequencies with a ratio fII/fI of
0.99. The soil is assumed to be a half space with a
shear wave velocity cs of 100 m/s, a density ρ of
2000 kg/m3 and a Poisson’s ratio ν of 0.33.

Figure 3 Simulated ground motions with different wave apparent velocities ca.

136
 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

Simplifying the multiple piers into one actually In this study three degrees of coherency loss are
overestimates the ground motion spatial variation ef- considered, which represent weakly, intermediately
fect as multiple piers tend to average out such effect and highly correlated spatial ground motions at the
(Hao 1997). Although the simplified model is used, two bridge pier supports.
this study is a clear advancement compared to other For each parameter of the considered ground mo-
studies, e.g. that performed by DesRoches & Mutku- tion twenty sets of spatially non-uniform ground mo-
kumar (2002), where the influence of the pier, sub- tions are simulated. In total 100 sets of ground mo-
soil and spatial variation of the ground excitation is tions are generated. Details of the ground motion
not considered at all. Figure 3 shows the influence of simulation are given in (Chouw & Hao, 2008).
wave apparent velocity ca on the spatial variation of
the ground motions at the two distant bridge pier lo-
cations. 4 DESIGN PARAMETERS
The ground motions are simulated based on a
near-source ground motion model introduced by
Ambraseys & Douglas (2003). The dominant fre-
quencies of the simulated ground motions range be- 4.1 Closing relative movements
tween 2.5 Hz and 12.5 Hz with the peak ground ac- Figure 4 displays the combined influence of SSI, the
celeration of 3 m/s2. The considered wave apparent wave apparent velocity ca and the coherency loss on
velocities ca are 200 m/s, 500 m/s and 1000 m/s. the mean values of the minimum total MEJ gap re-
With increasing wave velocity the delay of the quired to avoid girder pounding. The total gap gc is
ground motions at the right bridge pier support de- the sum of each gap between the MEJ beams (see
creases as the occurrence of the peak motions ag1 and Figure 2).
ag2 shows. However, the spatial variation of the
ground motions is not only characterized by the time
delay but also by the coherency loss.

Figure 4 Minimum required total gap g of a MEJ.

Figure 5 Dependence of the minimum gap gc required on the frequency ratio fII/fI, the apparent wave velocity ca
of the spatially varying ground motions and SSI.

137
 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

If fixed-base structures and uniform ground exci- In the higher ratio range, fII/fI above 1.15, the in-
tation are assumed, the minimum gap gc is only 0.59 fluence of the wave apparent velocity ca is dominant.
cm. This is to be expected, because both structures As expected, the minimum total MEJ gap reduces
with assumed fixed base have very similar funda- with increasing wave speed.
mental frequencies (fII/fI = 0.99). In the range of lower frequency ratios there is no
If the effect of the subsoil is considered, the re- clear tendency of the influence of the wave ap-
quired gap is not similar as one might expect. Even parent velocity. Higher wave speed does not neces-
though the frequencies of the two structures without sarily cause smaller required gap. This is because
considering SSI are similar and both structures ex- spatially varying ground motions induce both quasi-
perience the same ground excitation, owing to their static and dynamic responses. When a structure is
different structural slenderness both bridge structures relatively stiff, quasi-static response dominates the
interact with their ground differently. The unequal responses. That is why when the frequency ratio is
SSI effect causes relative movements, and conse- larger than 1.15, which is achieved by increasing the
quently a much larger minimum required total MEJ stiffness of the right bridge structure while that of
gap gc of 10.42 cm. In Figure 4 it is indicated as a the left bridge structure remains unchanged, increas-
horizontal solid line. ing wave apparent velocity reduces the relative re-
The results show that an assumption of uniform sponses. On the other hand, if the structure is rela-
ground excitation clearly underestimates the mini- tively flexible, dynamic response dominates, which
mum required total gap of a MEJ to avoid pounding, is very much influenced by the dynamic properties of
especially when the structures are assumed to be the bridge and the frequency content of the ground
fixed at their base. In the case of highly correlated motions. These observations indicated that the min-
spatially varying ground motions the minimum re- imum gap is not only significantly affected by the
quired gap does not decrease with higher wave ap- frequency ratio of two adjacent structures, but also
parent velocity ca. In the case of the wave apparent by the absolute vibration frequencies of the two
velocity ca of 500 m/s the minimum total gap also structures.
does not decrease with less coherency loss (e.g. the A comparison of the results with and without SSI
highly correlated case) as one would expect. These shows that an additional effect of SSI further in-
results reveal that the minimum total MEJ gap can- creases the minimum gap that a MEJ must have to
not be related to a single influence factor, because ensure that pounding will not take place. This result
the combined influence of these factors is dominant. clearly shows the significance of the combined in-
Another influence factor, not considered so far, is fluence of SSI, the spatial variation of the ground
the ratio fII/fI of the fixed-base fundamental frequen- excitations and the ratio of the fundamental fre-
cies of the adjacent bridge structures. The involve- quency of the adjacent bridge structures.
ment of this factor in causing relative girder re-
sponses is displayed in Figure 5.
To enable a clear interpretation of the considered 4.2 Opening relative movements
factors the effect of the structural slenderness is ne- To ensure a perfect function of MEJ each of the
glected. It is assumed that both adjacent structures seals between the MEJ beams should not be over-
have the same height of 9 m. Hence, the relative stretched during the opening relative movement be-
girder movement due to unequal interaction between tween the bridge girders. In order to achieve this
the structures and soil is not considered. As a refer- goal MEJ must be designed with sufficient number
ence the case of structures with assumed fixed-base of centre beams so that it can cope with the largest
is displayed on the left side of Figure 5. opening relative movement without causing any
The results clearly show that the recommendation damage to the covering seals.
of current design regulations to avoid relative girder Figure 6 shows the combined influence of SSI
movement by designing structures with similar or and the spatial variation of the ground excitations on
equal fundamental frequencies is not adequate, when the mean values of the largest expected opening
spatially non-uniform ground excitation does occur. movement go between the bridge girders. If a uni-
In fact the fixed-base frequency ratio cannot be used form ground excitation is assumed, theoretically al-
as the only design parameter. most no opening relative movement will take place,
At the frequency ratio fII/fI = 1.0 the minimum to- because both adjacent structures have nearly the
tal gap does not have the smallest value, and this same fundamental frequencies (fII/fI = 0.99). How-
value is definitely not equal to zero. In the investi- ever, because the adjacent structures have different
gation it is assumed that the spatially varying ground heights (Figure 2), the subsoil has not the same in-
motions are highly correlated. In both cases, with fluence on each of the structures as it can be easily
and without SSI, the influence of the frequency ratio estimated from Equation (1). Consequently, the un-
fII/fI is obvious. equal soil-structure interaction causes opening rela-

138
 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

tive movement go of 10.4 cm, even if both structures To focus on the combined influence of SSI, the
experience the same ground excitation. In Figure 6 spatial variation of the ground motions and the ratio
this result is displayed as a solid horizontal line. It of the fundamental frequencies of the adjacent struc-
clearly shows the significance of the unequal soil- tures, it is assumed that both bridge structures have
structure interaction effect. the same height of 9 m and the ground motions are
If the influence of the wave apparent velocity is highly correlated.
considered as well, the largest opening relative gird- Figure 7 shows this combined effect on the larg-
er movement increases to 12.5 cm in the case of ca = est opening relative movement that a MEJ should
500 m/s. Similar to the total gap that a MEJ requires expect. If SSI effect is neglected, the apparent veloc-
to avoid pounding, the relative movement between ity of the spatially varying ground motions is domi-
the girders does not increase proportionally with the nant when the frequency ratio fII/fI is larger than 1.
wave speed which indicates that other influence is Even though the opening relative movement in the
also strongly involved. higher frequency ratio range does decrease with the
The results in the case of a constant wave velocity apparent velocity, similar to the total gap required to
ca = 500 m/s confirms the involvement of many fac- avoid girder pounding, the smallest opening relative
tors in the development of the relative movement be- movement does not occur when both adjacent struc-
tween the girders. In contrast to the expectation the tures have the same fundamental frequency (fII/fI =
largest opening relative movement go of 12.5 cm oc- 1).
curs, when the spatially varying ground excitations An additional consideration of SSI alters the re-
have the highest correlation. Weakly correlated sult significantly. In general it can be observed that
ground motions do not cause the largest opening rel- SSI amplifies in almost all cases the opening relative
ative girder movement, but the value of 11.75 cm, movement.
while the intermediately correlated ground motions
produce a smaller opening relative movement of
10.72 cm.

Figure 6 Minimum opening movability go of a MEJ.

Figure 7 Dependence of the minimum required opening movability go on the frequency ratio fII/fI, the apparent wave velocity
ca of the spatially varying ground motions and SSI.

139
 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

While in the higher frequency ratio range above - In the lower frequency ratio range the combined
1.5, an apparent velocity ca higher than 500 m/s does effect of the ground motion spatial variation, SSI
not further reduce the opening relative movement and the vibration frequencies of the adjacent
which remains almost constant around 7.6 cm. At structures governs the minimum required gap.
lower frequency ratios the difference in the funda- - In almost all cases SSI causes a larger total gap.
mental frequencies of the adjacent bridge structures
has significant influence on the development of the
largest opening relative movement that one should 6 ACKNOWLEDGEMENT
consider in the design of MEJ. The result with SSI
effect also shows that current design recommenda- The authors thank the editors of the Australian Earth-
tion by using only the frequency ratio fII/fI as the de- quake Engineering Society conference proceedings
sign parameter is clearly insufficient. for the invitation to extend the manuscript for this
For the design of MEJ both values, the minimum journal. The authors also would like to thank the two
total gap required and the largest opening relative anonymous reviewers for their constructive com-
movement, are essential. The number of the middle ments of the conference contribution as well as an-
beams of MEJ should be chosen so that the total other two anonymous reviewers of this article.
available gaps between MEJ beams and the total
available seals covering the gaps can cope with the
largest opening and closing relative movement be- 7 REFERENCES
tween the bridge girders.
Ambraseys NN. and Douglas J. (2003). Near-field horizontal
and vertical earthquake ground motions. J Soil Dynamics
Earthquake Engineering, Vol. 23, pp 1-8.
5 CONCLUSIONS American Association of State Highway and Transportation
(AASHTO) (1998). Load and resistance factor design spe-
A new design philosophy for preventing bridge gird- cifications for highway bridges. Washington DC.
ers from pounding due to strong earthquake is intro- CALTRANS Seismic Design Criteria (1999). Design manual-
duced. In contrast to the design of a conventional version 1.1. California Department of Transportation, Sac-
bridge expansion joint with only a few-centimetre ramento, California.
Chouw N. (2005). Earthquake-induced pounding: causes and
gap, in the new design, modular expansion joints effect on structures, In: Hao H, Lok T.S. and Lu G.X.
(MEJs) are installed so that the adjacent bridge gird- (eds.): Proceedings of the int. conf. on Shock and Impact
ers can have a large relative movement without caus- Loads on Structures, 7-9 December 2005, Perth, Australia,
ing any pounding, and consequently damage to the pp 15-25
girders. The most significant specification is the Chouw N. and Hao H. (2008). Significance of SSI and non-
minimum total gap of the joint. The MEJs should uniform near-fault ground motions in bridge response I: Ef-
fect on response with conventional expansion joint, Engi-
then be designed so that the total MEJ gap can cope neering Structures, Vol.30, pp 141-153.
with the largest expected closing relative movement. DesRoches R and Muthukumar S. (2002). Effect of pounding
The other significant specification is the largest and restrainers on seismic response of multiple-frame
opening relative movement that the MEJs can ex- bridges. ASCE J Structural Engineering, Vol. 128(7), pp
pect. They must be designed so that they can cope 860-869.
with these movements without causing any damages Dexter RJ., Mutziger MJ., Osberg CB. (2002). Performance
testing for modular bridge joint systems, National coopera-
to the seals between the MEJ beams. tive highway research program, Report 467, 73 p.
In this work the influence of the spatially varying Elnashai AS, Kim SJ, Yun GJ, Sidarta D. (2007). The Yogya-
ground motions, SSI and their combined effect are karta Earthquake on May 27, 2006, Mid-America Earth-
discussed. quake Center, Report No. 07-02, 57 p.
The investigation shows that: Hao H. (1997). Torsional response of building structure to spa-
tial random ground excitationns, International Journal of
- The recommendation of current design regula- Engineering Structures, Vol. 19(2), pp 105-112.
tions to adjust the fundamental frequencies of Japan Road Association (JRA) (2004). Specifications for high-
adjacent bridge structures does not necessarily way bridges –Part V seismic design, 5th ed., 406 p. (in Jap-
produce the smallest minimum total gap that a anese)
MEJ must have when the ground motions are not Japan Society of Civil Engineers (JSCE) (1999). The 1999 Ji-Ji
uniform and the soil is soft. earthquake, Taiwan –investigation into damage to civil en-
gineering structure, In: Hamada M, Nakamura S, Ohsumi
- When the frequency ratio of adjacent bridge T, Megro K, Wang E (eds.), 160 p.
spans is larger than 1.15, the effect of wave ap- Moehle J, Fenves G, Mayes R, Priestley N, Seible F, Uang C-
parent velocity dominates the relative response. M, et al. (1995). Highway bridges and traffic management.
The minimum required gap decreases, as ex- Earthquake Spectra, Vol. 11(S2), pp 287-372.
pected, with higher wave speed. Park R, Billings IJ, Clifton GC, Cousins J, Filiatrault A,
Jennings DN et al. (1995). The Hyogo-ken Nanbu earth-

140
 EJSE Special Issue:
Earthquake Engineering in the low and moderate seismic regions of Southeast Asia and Australia (2008

quake of 17 January 1995. Report of the NZNSEE recon-

naissance team. Bulletin of New Zealand National Society
for Earthquake Engineering, Vol. 28(1), 98 p.
Ravshanovich K.A., Yamaguchi H., Matsumoto Y., Tomida N.
and Uno S. (2007). Mechanism of noise generation from a
modular expansion joint under vehicle passage, Engineer-
ing Structures, Vol. 29, pp 2206-2218.
Sieffert J-G and Cevaer F. (1992). Handbook of impedance
functions, Quest Editions, Presses Academiques, Nantes,
France, 174 p.
Yashinsky M and Karshenas M.J. (2003). Fundamentals of
seismic protection for bridges, Earthquake Engineering Re-
search Institute, Monograph MNO-9, 184 p.

141
View publication stats