Proceedings of the First PRC-US

Workshop
Workshop on Seismic Analysis and
Design of Special Bridges

Edited by
Lichu Fan
State Key Laboratory for Disaster Reduction in Civil
Engineering
Tongji University
Shanghai, China
George Lee
Multidisciplinary Center for Earthquake Engineering
Research
University at Buffalo, State University of New York
Buffalo, New York

Technical Report MCEER-03-0004 UNIVE

JI R
July 15, 2003 G 1907
SI
TON

TY

This workshop was held October 8-10, 2002 and hosted by the State
Key Laboratory for Disaster Reduction in Civil Engineering, Tongji
University, Shanghai, China and was supported by the Federal Highway
ISSN 1520-295X Administration under contract number DTFH61-92-C-00106.
 NOTICE
This report was jointly prepared by Tongji University and the Multidisciplinary
Center for Earthquake Engineering (MCEER) as a result of research sponsored by
the Multidisciplinary Center for Earthquake Engineering Research (MCEER) through
a contract from the Federal Highway Administration. Neither MCEER, associates
of MCEER, its sponsors, Tongji University, nor any person acting on their behalf:

a. makes any warranty, express or implied, with respect to the use of

any information, apparatus, method, or process disclosed in this re-
port or that such use may not infringe upon privately owned rights;
or

b. assumes any liabilities of whatsoever kind with respect to the use of,
or the damage resulting from the use of, any information, apparatus,
method, or process disclosed in this report.

Any opinions, findings, and conclusions or recommendations expressed in this pub-

lication are those of the author(s) and do not necessarily reflect the views of MCEER
or the Federal Highway Administration.
 UNIVE
JI R
G 1907

SI
TON

TY
Proceedings of the
First PRC-US Workshop
on Seismic Analysis and Design of Special Bridges

Held at
State Key Laboratory for Disaster Reduction in Civil Engineering
Tongji University
October 8-10, 2002

Edited by
Li-Chu Fan1 and George C. Lee2

Publication Date: July 15, 2003

Technical Report MCEER-03-0004

Workshop Steering Committee

Lichu Fan, SLDRCE, Tongji University
Mengling Lou, Tongji University
Weigang Bao, PRC Ministry of Communication
George Lee, MCEER, University at Buffalo
Ian Buckle, University of Nevada, Reno
W. Phillip Yen, Federal Highway Administration

Local Organizing Committee

Lichu Fan, Shide Hu, Yan Xu and Suwen Chen, all of Tongji University, Shanghai, PRC

FHWA Contract Number DTFH61-98-00094

1 Professor, State Key Laboratory for Diasaster Reduction in Civil Engineering, Tongji Univer-
sity, Shanghai, PRC
2 Director, Multidisciplinary Center for Earthquake Engineering Research, University at
Buffalo, State University of New York

MULTIDISCIPLINARY CENTER FOR EARTHQUAKE ENGINEERING RESEARCH

University at Buffalo, State University of New York
Red Jacket Quadrangle, Buffalo, NY 14261
 Preface

This seismic analysis and design of special bridges (SADSB) workshop series is based on a
Memorandum of Understanding (MOU) between the Multidisciplinary Center for Earthquake
Engineering Research (MCEER), University at Buffalo, Buffalo, New York and the State Key
Laboratory for Disaster Reduction in Civil Engineering (SLDRCE), Tongji University, Shanghai
China. The MOU was signed by Professor George C. Lee of MCEER and Professor Lichu Fan of
SLDRCE on May 26 2001, and resulted from the PRC-US earthquake engineering and
earthquake disaster mitigation collaboration project. Four international workshops will be carried
out in China and US between 2002-2005, alternating locations each year.
The first workshop was held on October 8 - 10, 2002 at Tongji University in Shanghai. A
total of 35 participants, ten from the US 12 from China, and 11 observers attended this workshop.
Following the two-day meeting and discussion, a technical tour to the LuPu Bridge, a tied arch
bridge under construction over the Huangpu River in Shanghai, was arranged. The lead designer
of the energy dissipation system in the main span was Professor Shi-de Hu, one of the workshop
organizers.
Workshop themes included seismic design and retrofit of long span bridges, performance
based design, seismic safety evaluation, soil-pile-structure interaction and pseudo-dynamic and
hydrodynamic experimental study. These proceeding contain 22 papers covering a wide range of
research fields, including a discussion of seismicity in China.

iii
 Contents
Page

Introduction to the State Key Laboratory for Disaster Reduction in

Civil Engineering
Li-chu Fan 1

Introduction to the Seismicity of China

Luo Qi-feng & Li Shi-dong 3

Performance-Based Seismic Design of Highway Systems

Ian G. Buckle 15

Mitigate Earthquake Hazard & Risk for Highway Bridges Through Planning,
Design and Retrofitting
W. Philip Yen and James D. Cooper 27

Seismic Safety Evaluation of Large Scale Interchange System in Shanghai

Li-chu Fan, Jian-zhong Li, Shi-de Hu, Gui-ping Bi and Li-ying Nie 31

A New Approach to Analysis of Soil-Pile-Structure Interactions for

Long-Span Bridges
Seung-Il Nam & Jamshid Ghaboussi 43

Preliminary Study of Hydrodynamic Effects on Seismic Response of Bridges

Jun-jie Wang, Wei Lai, Ning-yong Zhang, Li-min Sun and Li-chu Fan 55

Observed Pile and Pipeline Performance in the Full-scale Lateral

Spread Experiment
Scott A. Ashford & Teerawut Juirnarongrit 69

Seismic Response of Railway Bridges Considering Track Restriction

Gui-ping Yan, Yan Huang and Guanyuan Zhao 81

Seismic Design and Retrofit Guidelines for Bridges in New Jersey,

A Low-To-Moderate Seismic Hazard Area
Harry Allen Capers, Jr 95

Generation of Spectral Compatible Non-stationary Ground Motions based on Phase

Difference Spectra
Qing-shan Yang & Hai-peng Jiang 107

The Seismic Retrofit of the Golden Gate Bridge

Charles Seim, P.E. 119

v
 Page

Response of Seismic Isolated Bridges using M-DOF Model and 2D Excitation

George C. Lee & Zach Liang 129

Seismic Conceptual Design for Bridge Tower of a Long-span Cable-stayed Bridge

Ai-jun Ye, Shi-de Hu & Li-chu Fan 141

Analytical Investigation of the Response of Lu-Pu Bridge with Added

Viscous Dampers
She-de Hu, Zhi-qiang Wang & Li-chu Fan 151

Seismic Performance and Retrofit of a 24-Span Freeway Bridge

M. Saiidi, A. Itani, Q. Yang & T. Isakovic 161

Nonlinear Seismic Response Analysis of Effects of Sliding and Pounding

of Urban Interchange Bridges with Rubber Bearings
Li-ying Nie, Jian-zhong Li & Li-chu Fan 173

Seismic Design and Analysis for Urban Viaducts with a Double Deck
Jian-zhong Li, Shi-de Hu and Li-chu Fan 183

A Pseudodynamic Test of an Urban Viaduct with a Double-Deck

Tian-bo Peng, Shi-de Hu, Jian-zhong Li and Li-chu Fan 193

Seismic Design and Retrofit Strategies of Cable-Supported Bridges:

An Overview of Current U.S. Practice
George C. Lee, John Sun and Chuck Seim 203

Seismic Response Analysis of Wuhu Yangtze River Bridge

Xi Zhu 215

A Study on Shear Strengthening of Reinforced Concrete Beams Using ±45°

Glass Fiber Reinforced Polymer
Jie Li, Yuan-de Xue and Wen-xiao Li 231

Research on Flexural Properties of Hybrid GFRP/CFRP Tube Confined

Concrete Beams
Hua Yuan, Yuan-de Xue and Wen-xiao Li 241

PRC-US Workshop Agenda 255

Participants 257

vi
 Introduction to the State Key Laboratory For Disaster
Reduction in Civil Engineering (SLDRCE)

Li-chu Fan1
The State Key Laboratory for Disaster Reduction in Civil Engineering (SLDRCE) was
established in 1988 at Tongji University, which is one of 81 state key laboratories led directly by
the State Commission of Education & Science, and also the only state key laboratory in the field
of civil engineering in China. The main objectives of the SLDRCE at Tongji University, as one
of the main research centers for disaster reduction in China, are to:
● Improve basic understanding of disaster phenomena through fundamental studies
● Establish effective measures for disaster reduction
● Provide opportunity for domestic applicants through opening the laboratory to
outside scientists
● Participate in international exchanges and collaboration with government and
non-government organizations, universities and research institutes
● Serve as consultants in disaster reduction
● Organize training activities

In addition, the SLDRCE will make a due contribution to the “International Decade for
Natural Disaster Reduction” organized by the United Nations.

The SLDRCE’s main facilities include:

● Shaking Table
● Boundary layer wind tunnel (TJ-1,TJ-2 and TJ-3)
● Strong Ground Motion Instrumentation

Research is performed in the following areas:

● Fundamental Study of Earthquake Engineering
● Applied Study of Earthquake Engineering
● Fundamental Study of Architectural Aerodynamics and Aeroelastic Phenomena
● Applied Study of Wind Engineering
However, in this presentation, special attention is paid to the research projects, further
development and study trends of the bridge seismic laboratory, which is a subarea of applied
study in earthquake engineering.
__________________
1 State
Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, PRC

1
 Introduction to the Seismicity of China

Luo Qi-feng and Li Shi-dong1

ABSTRACT

In this paper, China seismicity is introduced, which includes seismic regions in China,
seismic activity, fault model, attenuation model and its characteristic, and a seismic zoning
map of China, etc.

Keywords: seismic region in China, seismic activity, seismic zonation map of China,
fault model, attenuation model, estimation of ground motion

1
Research Institute of Structural Engineering and Disaster Reduction, Tongji University, Shanghai 200092

3
PREFACE

China is a large country where there were many large earthquakes occurred in the past
thousand years. In 1556 about 830,000 persons were killed in Guanzhong earthquake, Shaanxi
province. In 1976 more than 230,000 persons were killed in Tangshan earthquake, Hebei
province. In the past 50 years, there are 51 earthquakes with a magnitude larger than 7.0
occurred in China. Therefore how to predict earthquake and how to mitigate seismic loses are
the main purposes for scientists and engineers of China, who are studying and working on
seismology, geophysics, geology, earthquake engineering, civil engineering and sociality, etc.
Here the authors will introduce some researches on seismic region of China, seismic activity,
fault model, attenuation model and its characteristics, seismic zoning map of China, and
estimation of ground motion, etc.

MAIN SEISMIC REGIONS AND SEISMIC ACTIVITY

Seven main seismic regions

Based on the analysis of the seismic activity and seismo-tectonic environment, 7 large
seismic regions are divided in China. They are Northeast region, North China region, South
China region, Taiwan region, South China Sea region, Xinjiang region and Qinghai-Tibet
Plateau region (see Figure 1). Every region has its own characteristic of seismic activity. In
Taiwan region, Xinjiang region and Qinghai-Tibet Plateau region, the seismic intensities are
stronger and the frequencies of earthquake occurrence are higher. In North China region the
seismic intensity is stronger and the frequency of earthquake occurrence is high. The seismic
activity is medium in South China region, where the seismic intensity is between medium and
strong, while the frequency of earthquake occurrence lower. In Northeast China region and
South China Sea region, the seismic activities are not active, where the seismic intensities are
weaker and the frequencies of earthquake occurrence lower.
Taiwan seismic region is suited on the west boundary of Pacific plate and it belongs to
Circum-Pacific seismic zone. And South China Sea region is mainly influenced by Philippine
plate. Beside the earthquakes occurred in the two regions, most of others belong to intra-plat
earthquakes. Some seismologists in United States also take China into an intra-plate region
(Butler et al. 1979).
Also there are other methods to divide seismic regions. For example, Sichuan-Yunnan
seismic belt, which is in the central area of Qinghai-Tibet Plateau region, and Shanxi fault belt
in North China region compose one seismic belt, which is named Great Central Seismic Belt or
Great North-South Seismic Belt (see Figure 1).

4
 Figure 1 Distribution of 7 seismic regions and
revised seismic intensity zoning map of China (1990)

5
 Since 1900 there are many serious disaster earthquakes occurred in China, among
which there are 9 earthquakes with a magnitude larger than 8.0. All the epicenters of
earthquakes occurred in China with magnitude 3 from 1901 to 1950 are shown in
M ≥ 4
4
Figure 2.

Figure 2 Distribution of epicenters of earthquakes ( M 3 )

≥ 4
4
Occurring in China from 1901 to 1950

According to the data we’ve collected, in the past 50 years there are about 51
earthquakes with magnitude larger than 7.0 occurring in China. Their distribution in different
seismic regions is shown in Table 1. Figure 2 and Table 1 all show the characteristics of
seismicity in different seismic regions in China.

Table 1 Distribution of large earthquakes (M ≥ 7.0, Since 1950)

Northeast North Qinghai-Tibet South South

Region Xinjiang Taiwan Total
China China Plateau China China Sea
Times 3 6 4 18 20 0 0 51

From Table 1 we can see that there are many large earthquakes occurred in
Qinghai-Tibet Plateau region and Taiwan region. Because the population in Qinghai-Tibet
Plateau is very small and most of the earthquakes took place in the wild field, these earthquakes
did not cause serious disasters. For example, Kunlun Mountain Pass earthquake (Ms 8.1) on
14th Nov.2001 is only one great earthquake with magnitude larger than 8 occurred in Chinese
continent since 1951. Fortunately, its epicenter is in frigid zone, where there are no people. And
the Qinghai-Tibet railway, which runs across the near field of the earthquake, is still under
construction, so that the great earthquake did not cause serious disaster.
Because of the dense population in North China region, although there are only a few
earthquakes occurred there in the past 50 years, if one great earthquake occurs it will cause

6
serious disaster. The lesson embedded in our mind is the tragedy caused by the great Tangshan
earthquake (Ms7.8) on 28th July 1976.

Seismic occurrence period

There are more than 3800 years of historical seismic records in China; Seismic
activities can be analyzed from these records. Figure 3 shows the active periods and non-active
periods of seismic activity in different seismic regions (Figure 3 is a revised figure, reference to
Hu, 1988). Comparison of seismic activities shows that the active and non-active periods in
one region are almost the same, but the periods in different regions vary from each other. In
North China, South China, Taiwan and Qinghai-Tibet Plateau regions, the active period is
about 300~400 years. In Xinjiang region, it is about 100 years, while in Taiwan it is only about
decades. Figure 3 also implies that China is now in active seismic period.

Figure 3 Seismic active periods in main seismic regions

The magnitude-frequency relationship can be described by equation

lg N = a − bM . (1)

Where M is the earthquake magnitude, N is the occurring times of earthquakes whose

magnitudes are M. a and b are constants. Figure 4 shows the magnitude-frequency
relationships in the main seismic regions (Figure 4 is a revised figure, reference to Hu, 1988).
Comparing Figure 3 with Figure 4, we can know that if the active period is shorter in
one seismic region, the constants a and b are larger in its magnitude-frequency relationship.
For example, the values a and b in Taiwan region are larger than those in North China, while
the active periods in the two regions are opposite.

7
 Figure 4 Magnitude-frequency relationship in main seismic regions

Seismic fault model and source depth

Fault model. Table 2 shows percentage of different fault models and maximum
magnitude of earthquakes in some seismic regions (Table 2 is a revised table, reference to
Huan et.al. 1990 (got from the book edited by Hu, 1999)). In North China and Central
Qinghai-Tibet Plateau, most of the fault models are strike-slip faults and their percents are 90%
and 84%, and the maximum magnitudes are 7.8 and 7.9. But in Southeast China, the percent is
only 63% and the maximum magnitude is 7.3. Tangshan earthquake on 28th July1976 is a
typical strike-slip fault event, and Jiji earthquake of Ms7.4 on 21 Sept.1999 is a typical reverse
fault event.
Fault length. There are many relationships between fault length and magnitude.
Table 3 shows one reference relationship of them (Table 3 is a revised table, reference to Huan
et.al. 1991 (got from the book edited by Hu, 1999)). The fault length of Tangshan earthquake
Ms7.8 in 1976 is about 100km (Luo and Hu, 1997).
Seismic focus depth. Among the 51 earthquakes with magnitude larger than 7
mentioned above, there are 23 events from which we can collect seismic focus depth data.
Except for two earthquakes, which occurred in Northeast China and their focus depths are
595km and 570km, others are shallow focus earthquakes. Their focus depths range from 1 km
to 59 km and most of them are between 10km and 30km.

8
 Table 2 Percent of fault model & maximum magnitude
of earthquake in some seismic regions

Fault Model Strike-Slip Fault Reverse Fault Normal Fault

Regions % Mmax % Mmax % Mmax
Vicinage of
52 8.5 44 7.25 4 6.9
China
North China 90 7.8 7 4.5 3 7.1
Central of
Qinghai-Tibet 84 7.9 11 7.2 5 6.8
Plateau
Southeast
63 7.5 13 3.4 19 4.6
China
)
Table 3 relationships between fault length and magnitude

Magnitude 6 6.5 7.0 7.5 8.0

Fault length (km) About 20 30~40 50~60 80~100 100~200

SEISMIC ZONING MAP

There are four generations of seismic zoning maps in China. The third generation of
zoning map was compiled in 1990 and the fourth generation 2001. These two zoning maps are
compiled by adopting probabilistic method of the seismic hazard analysis.

Seismic intensity zoning map of China (1990)

It is the first time to compile seismic intensity map of China by using probabilistic
seismic hazard analysis method. Based on the analysis of the seismic activity and
seismo-tectonic environment, 7 seismic regions are firstly divided into 26 seismic provinces as
the statistic elements of the seismicity analysis. Then 733 seismic potential source areas are
divided in the seismic provinces. Considering the inhomogeneity of seismicity distribution in
both space and time, the seismic intensities with various probabilities of exceedance are
calculated at about 30,000 controlling points in the whole country. And then the seismic
intensity zoning map (scale with 1:4,000,000) is compiled with 10 percent of probability of
exceedance during a period of 50 years. Figure 1 is a revised seismic intensity zoning map. In
the original zoning map, the whole country is divided into 5 grades of zones, they are MMI<V,
VI, VII, VIII and >IX (CCSZMC, 1992). But in Figure 1 there are only 4 grades, they are
MMI<V, VI, VII and >VIII.

Seismic ground motion parameter zonation map of China (2001)

China Seismological Bureau published a new seismic zoning map in 2001 (scale with
1:4,000,000). There are four main characteristics in the new zoning map. Firstly, it is the first
seismic ground motion parameter zoning map of China. Peak acceleration and characteristic

9
period Tg ( Tg=2πV/A) of acceleration response spectrum are used to make zonation. Therefore
there are two zoning maps, one peak acceleration zoning map and another characteristic period
Tg of response spectrum zoning map. In the acceleration zoning map (Figure 5, Hu, 2002) the
whole country is divided into 7 grades of zones, they are A<0.05g, 0.05g, 0.10g, 0.15g, 0.20g,
0.30g and ≥0.40g.

Figure 5 Peak acceleration zonation map of China

Figure 6 Characteristic period Tg of response spectrum zonation map of China

In Tg zoning map (Figure 6, Hu, 2002) there are 3 grades, they are 0.30s, 0.35s and
0.40s. Secondly, the compile methodology considers the inhomogeneieou characteristics of
seismicity in China, which reflects the progress made on seismic activity in strong and weak
regions in the past 10 years. Thirdly, the two zoning maps are compiled for type II site soil
condition, and they accompany a table that contains some values such as Tg for type I and type
III site conditions. Furthermore, considering the uncertainties in some aspects and different
opinions of experts in seismology and geology, etc., several schemes are adopted to make
hazard analysis. Compared with the seismic intensity zoning map of China, the grades in the
new map are changed from 5 to 7. The increased grades are A=0.15g and 0.30g, which equal to
MMI=7.5 and 8.5.

10
ESTIMATION OF GROUND MOTIONS

Attenuation relationship of MMI in China

The ellipse intensity attenuation models of different rigions are adopted in seismic
intensity zonation (CCSZMC, 1992). In Eastern China, the attenuation model is as follows.

In long axis:
I = 6.046 + 1.480M − 2.081ln(R + 26) (2a)

In short axis:
I = 2.617 + 1.435M − 1.441ln(R + 7) (2b)

In Western of China, the attenuation model is as follows

In long axis:
I = 5.643 + 1.538M − 2.109 ln(R + 25) (3a)

In short axis:
I = 2.941 + 1.363 M − 1.494 ln( R + 7 ) (3b)

Where I is MMI, M surface magnitude and R epicentral distance in km.

Comparison of attenuation characteristics in North China and in US

Butler et al considered China to be an intra-plate region. They stated in their paper

(1979) “it is likely that the intensity fall off characteristics of northeast China are more akin to
those of the stable continental eastern United States than the basin and range, tectonic
environment of the west”.
Four recently developed attenuation models are calibrated by using a very limited
amount of strong motion data recorded in North China (Wong, et al. 2002). Their research
shows that the attenuation characteristics of ground motions in the North China are similar to
those in western US. The supporting evidence includes coda and S-wave Q factors, preliminary
results of kappa values, stress drop, shear wave velocity profile in the shallow earth crust, and
areas enclosed by the isoseismal of Modified Mercalli intensity V etc. From the comparisons
between results obtained by using different attenuation models it is recommended that the
Crouse and McGuire (1996) spectral attenuation model could be used as a potential attenuation
model for the North China (Wong, et al., 2002).

Probability-Consistent Scenario Earthquake and its application

Luo proposed a new definition of Probability-Consistent Scenario Earthquake (PCSE)

and an evaluation method of its magnitude M ( p 0 ) , focal distance R ( p 0 ) and orientation

ϕ ( p 0 ) (Luo, 1996, 2000a). The exceeded probability of ground motion y ( p 0 ) , which caused

11
by PCSE M ( p 0 ) , is equal to p 0 . p 0 is calculated from conventional probability hazard

analysis method. PCSE and improved empirical Green’s function method can be used to
estimate ground motions (Luo and Hu, 1990, 1997, Luo and Dan, 1994a, Luo et al., 1994b,
Luo, 2000b). HHT (Hilbert-Huang Transform, Huang et al.1998) method is used to analyze
strong ground motion (Luo and Shi, 2002). It is a useful tool to separate different frequency
components from ground motion records.

DISCUSSION AND CONCLUSION

From the introduction above we can come to the conclusion as follows. Firstly, China is
a large country where there were many large earthquakes occurred in the past thousand years,
and China is now in active seismic period. Secondly, the seismicity of China is inhomogeneous
both in space and time. Thirdly, the seismic zonation maps of China reflect research level of
Chinese scientists and engineers in Seismology, Geology and Earthquake Engineering. And it
can be applied to aseismic design, programs of national land use, the countermeasure for
earthquake disaster mitigation and protection. Fourthly, except for some earthquakes occurred
in Taiwan, South China Sea, Northeast regions are inter-plate earthquakes, most of them
occurred in China are intra-plate earthquakes. Fifthly, the attenuation characteristics of ground
motions in the North China are similar to those in western US. But the attenuation
characteristics of whole China should be studied more. Furthermore, PCSE, Empirical Green
function method and HHT can be used to analyze and estimate seismic ground motion.

ACKNOWLEDGMENT

This paper is written under the encouragement of Professor Fan Lichu of Tongji
University, academism of Academic of Engineering of China and Professor George, C. Lee,
director of MCEER, University of Buffalo, United States.

REFERENCES

Butler R, Stewart G S, and Kanamori H (1979), The July 27, 1976 Tangshan, China earthquake – a complex
sequence of interpolate events, Bull. Seism. Soc. Am., Vol.83, pp1178-1798.
CCSZMC (the compiling committee of seismic zoning map in China), (1992), Seismic intensity zoning map of
China (1990), Earthquake Research in China, Vol.8, No.4, pp1-11
Crouse C B and McGuire J W (1996), Site response studies for purpose of revising NEHRP seismic provisions,
Earthquake Spectra, Vol.12, No.3, pp 407-439
Hu, Y. (1988), Earthquake Engineering, Seismology Press, Beijing, pp114-115.
Hu, Y. (2002), Introduce of seismic ground motion zoning map, Advances in Modern Earthquake Engineering,
edited by Wang Y., Li A and Cui J., Press of East South University, Nanjing, pp1-7.
Hu. Y. (1999), Technical book for seismic safety analysis, edited by Y. Hu, Seismology Press, Beijing, 1999,
pp.115-116.

12
Huang, et al. 1998. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-station
time series analysis. Proc. R. Soc. A, 454, pp903-995
Luo, Q. and Y. Hu, (1990), An Improved Empirical Green's Function Method and Synthesis of Near-field
Accelerograms in Lulong Earthquake, Earthquake Engineering and Engineering Vibration, Vol.10, No.3,
1-13.
Luo, Q., Y. Hu and K. Dan, (1994a), Statistical-empirical Green’s function method for estimating near-field
earthquake ground motions (in Chinese with English abstract), Journal of Natural Disasters, Vol.3, No.3,
pp.1-10.
Luo, Q. and K. Dan, (1994b), Simulation of Accelerations in Epicentral Region for 1923 Kanto, Japan,
Earthquake (M7.9) (in Chinese with English abstract), Earthquake Engineering and Engineering Vibration,
Vol.14, No.4, 35-43.
Luo, Q., (1996), Probability-Consistent Scenario Earthquake and Its Determination (in Chinese with English
abstract), Earthquake Engineering and Engineering Vibration, Vol.16, No.3, 22-29.
Luo, Q. and Y. Hu, (1997), Synthesis of Accelerations of the 1976 Tangshan Earthquake Ms7.6 in Near- and
Far-Field by Using Semi-Empirical Method (in English), Acta Seismologica Sinica, Vol.19, No.3, 347-354.
Luo, Q., (2000a), Probability-consistent Scenario Earthquake and Its Application in Estimation of Ground
Motions, Proceedings/12th WCEE, Auckland, New Zealand, Paper ID.0010.
Luo, Q., (2000b), Estimation of ground motions affecting Shanghai by long distance earthquake, Advances in
Structural Dynamics, Proceedings of Conference on Advances in Structural Dynamics, Vol. I, pp.225-232,
Dec. 13-15, 2000, Hong Kong, China, edited by J. M. Ko and Y. L. Xu, Elsevier Science Ltd, Oxford.
Luo, Q. and C. Shi, (2002), HHT and its application in analysis of seismic wave spectra, Advances in Modern
Earthquake Engineering, edited by Wang Y., Li A and Cui J., Press of East South University, Nanjing,
pp.373-376.
Wong Y, Zhao J X and Luo Q (2002), Attenuation characteristics of ground motions in North China, Earthquake
Engineering and Engineering Vibration (English edition), Vol.1, No.2 (in printing).

13
 Performance-Based Seismic Design of Highway Systems

Ian G. Buckle1

ABSTRACT

In recent years, a major review of performance criteria for bridges has been undertaken in the
United States and a move towards performance-based, multi-level seismic design of bridges has
begun. In a parallel exercise, a risk-based methodology has been developed for assessing the
performance of highway systems taking into account the seismic fragility of bridges and their
interconnectivity. These efforts have opened the door to performance-based seismic design of
highway systems, in which system-level performance criteria, such as maximum permissible traffic
delay times, are targeted for highway systems immediately following earthquakes of varying size.
This paper explores the feasibility of such a design approach and potential applications for resource
allocation and emergency planning.

BACKGROUND

The seismic performance of highway systems in recent earthquakes has been less than
satisfactory. In the last decade, numerous highways have been closed due to earthquakes in California,
Costa Rica, Japan, Turkey and Taiwan, and although life-safety was generally preserved, public
frustration with closures and restricted access has been widespread.
Just as with many other lifeline and infrastructure systems, highways are rarely designed for
seismic loads and there are no known codes or specifications for the seismic design of highway
systems. Instead most of the progress that has been made towards reducing the vulnerability of these
systems has been directed towards the performance of bridges, essential components of most highway
systems. But despite the widespread use of seismic bridge codes and specifications, many of the
highway closures in recent earthquakes have been due to bridge damage and collapse.
Historically, the United States and many other countries have used a single-level earthquake to
seismically design bridges and other structures. This earthquake, usually called the design earthquake,
is intended to represent the largest earthquake that could reasonably be expected to occur during the
life of the bridge. Inherent in such a statement is the notion of ‘uniform risk’ since the design level is
intended to be an earthquake with the same probability of exceedance from one region to another,
rather than using the maximum historical event for each region, which may have a very low
probability of occurrence.
The Standard Specification for Highway Bridges in the United States (AASHTO 1992) adopted
this uniform risk approach following the 1989 Loma Prieta earthquake, and uses a level of hazard that
has a 10% probability of exceedance in a 50-year exposure period. This corresponds to an event with
a return period of about 500 years (actually 475 years). In more recent years the exposure period has
been adjusted to 75 years, corresponding to the assumed life of a normal highway bridge (AASHTO
1994). The probability of exceedance was then raised to 15%, so as to maintain, approximately, the
same return period (500 years).
At the same time as adopting this uniform risk approach, a corresponding set of performance
expectations were included in the philosophy of the AASHTO specifications (AASHTO 1992).
These are given in Art. 1.1 of the specification and summarized below:
• Small to moderate earthquakes should be resisted within the elastic range, without significant
damage.
• Realistic seismic ground motion intensities and forces are used in the design procedures.

1
Professor, Department of Civil Engineering, MS 258, University of Nevada, Reno, 89557, USA

15
 • Exposure to shaking from large earthquakes should not cause collapse of all or part of the
bridge. Where possible, damage that does occur should be readily detectable and accessible
for inspection and repair.
A set of basic concepts for seismic design was derived from this philosophy (Art. 1.3, AASHTO
1992), and these are summarized below:
• Hazard to life to be minimized.
• Bridges may suffer damage but have a low probability of collapse.
• Function of essential bridges to be maintained.
• Ground motions used in design should have a low probability of being exceeded in the normal
lifetime of the bridge.
Characterized by a lack of specificity, these criteria were nevertheless a significant advance over the
then prevailing requirements for seismic design.
By contrast, little has been achieved by way of assessing the performance of an inventory of
bridges interconnected by a network of roads, and subjected to the same earthquake. Nor have other
components of highway systems (retaining walls, slopes, tunnels, culverts and the like) been
systematically studied and their contribution to system vulnerability determined. Applications of
seismic risk assessment procedures to water supply systems and other utilities have been developed,
but until very recently their application to highway systems had not been attempted.
Bridges are critical components of highway systems and with few exceptions are designed to
single-level performance criteria. Furthermore, the overall impact of bridge vulnerability on the
performance of complete highway systems is not generally known, due to lack of data on the other
components of highway systems and a credible methodology for performing such an analysis.
Improving the performance of bridges, and the systems of which they are part, is urgently required.
Performance-based design and seismic risk assessment procedures appear to offer a way forward.

THE CASE FOR PERFORMANCE-BASED DESIGN OF BRIDGES

The assumption is made in single-level design (and retrofit) that if performance at the design
event is satisfactory, it will be satisfactory at all other levels, both smaller and larger. Such an
assumption is generally not true, as seen in recent earthquakes in California, Costa Rica, Japan,
Turkey and Taiwan. It would be true for smaller events if elastic performance was required at the
design event, and it may also be true for larger events, if the design event was sufficiently large and a
generous degree of conservatism used in the design. But under the design event, inelastic performance
(damage) is explicitly intended (in most bridges), and provided life safety is preserved, the
consequential restrictions on access are considered to be tolerable.
However, these restrictions become unacceptable, if they were to occur on a more frequent basis
such as during a smaller earthquake. Since this is a nonlinear problem, assurances regarding
performance during smaller earthquakes cannot be obtained simply by scaling performance at the
design event and thus explicit design (or at least a design check) should be made at this level, to gain
this assurance.
Similarly, performance during a larger event cannot be estimated by scaling upwards and relying
on reserve strength. Without explicit quantification, this approach is unreliable because it is based on
engineering judgment and an experience database that is thin and largely unverified, especially in the
central and eastern United States (CEUS).
The argument is thus made, that to avoid adverse performance, such as seen in Loma Prieta,
Northridge, Kobe and Taiwan, explicit consideration of bridge performance during at least two levels
of earthquake (and perhaps more) should be undertaken. Furthermore, the expected level of
performance during these earthquakes should be stated with a greater level of specificity than has
been the case in the past, and assurances given that these performance levels will be met. This
argument leads to the consideration of performance-based engineering for the seismic design and
retrofit of bridges.
Performance based engineering (PBE) has been defined as consisting of the selection of design
criteria, structural systems (layout, proportioning and detailing), and the assurance and control of
construction quality and long-term maintenance, such that at specified levels of ground motions, and

16
with defined levels of reliability, the structure will not be damaged beyond certain limiting states or
other usefulness limits. (SEAOC 1995). This definition has been paraphrased from that developed for
buildings in the SEAOC Vision 2000 Project where PBE was explored and its potential for improving
the seismic performance of new buildings was clearly demonstrated.
Application of the design phase of PBE requires several fundamental issues be addressed. These
include:
• Selecting the ground motions (hazard levels) and corresponding damage states (performance
objectives)
• Developing analytical methods for the verification of damage states and performance
objectives.
The first of these bullets requires that ground motions be known with a degree of confidence (i.e.
the 500-yr seismic design coefficient for a given bridge in a given site class, is known within
acceptable limits), and that realistic and meaningful objectives can be defined. The second bullet
requires a level of sophistication in analysis that can be implemented with ease and reliability.
Further, the relationship between damage states and performance objectives (such as crack width to
lane closures) must be not known with a degree of certainty.

HAZARD LEVELS AND PERFORMANCE OBJECTIVES FOR BRIDGES

Factors to be considered when selecting hazard levels and setting performance objectives include:
• How many earthquake levels should be used? Ideally it should be many, but in practice two or
three levels should sufficient to assure that the desired range of performance is achieved.
These might be small, moderate and large if three events are favored, or small and large if
only two events are considered. In the latter case, they might also be referred to as frequent
and rare events.
• How many different kinds of bridges should be considered? It is unreasonable to expect that
all bridges should have the same performance criteria for the same earthquake. More
important bridges for example, might be expected to perform to a higher level than less
important bridges. Temporary bridges and those under construction might also have specific
criteria. Setting aside these special cases, two or three categories should be again be
sufficient, and these might be based solely on importance, although it might be preferable to
use expected performance level as the differentiating parameter.
• How should these performance requirements be specified? It is not a simple matter to measure
performance and therefore to be able to specify it. One measure might be the number of days
a bridge is closed for repair following an earthquake, or has restricted access (lane reduction
or weight reduction or both). Another measure might be the extent of damage as given by
residual displacements or offsets, crack widths, extent of spalled concrete and exposed rebar,
number of misaligned or unseated bearings, settlement of approach fills, distress to expansion
joints and vehicle barriers, and the like. Neither measure is particularly satisfactory and in
practice both are used to complement each other. In this case both a performance level (PL)
and a damage level (DL) is used to set the performance criteria.

If dual events are considered (rather than three levels) and two bridge types identified, the above
performance criteria may be formatted in a 2 x 2 matrix with the rows assigned to the earthquake level
and the columns to bridge type. Elements within the matrix are the required performance and damage
levels. Table I shows such a performance criteria matrix.
Four performance levels and four damage levels are shown in Table I corresponding to two
earthquake levels and two bridge types. If more hazard levels and/or more bridge types are to be
considered, the number of performance and damage levels (PL, DL) would, in principle, increase. But
in practice duplication among the PLs and DLs is common and the number of separate and distinct
levels may not even be as many as shown in Table I.

17
 TABLE I. PERFORMANCE CRITERIA MATRIX FOR HIGHWAY BRIDGES.

BRIDGE TYPE 1 BRIDGE TYPE 2

EARTHQUAKE
(e.g. Standard Bridges) (e.g. Important Bridges)

PL1 PL2
Frequent Earthquake
DL1 DL2

Rare Earthquake PL3 PL4

DL3 DL4

where PL1 through PL4 is Performance Level 1 through 4

and DL1 through DL4 is Damage Level 1 through 4.

Caltrans Experience

Immediately following the 1989 Loma Prieta earthquake, the California Department of
Transportation (Caltrans) moved towards dual-level performance-based design. Endorsed by the
Caltrans Advisory Board and an independent review by the Applied Technology Council, these
criteria were customized to the seismic hazard in California and Caltrans prevailing practice. Two
bridge classes were identified (Ordinary and Important) and the rare and frequent earthquakes
identified as the safety evaluation earthquake (SEE) and the functional evaluation earthquake (FEE)
respectively.
The safety evaluation earthquake is determined deterministically by the California Division of
Mines and Geology (CDMG) for each site and is identified as the maximum considered earthquake
(MCE). In some circumstances this rare earthquake is determined probabilistically, using a 1000 –
2000 year return period. The functional evaluation event (the frequent earthquake) is also a
probabilistic event with about a 40% PE in the expected life of the bridge (about a 200-year event).
An important bridge is one satisfying any of the following:
• required to provide secondary life safety
• would create a major economic impact if closed for restoration of functionality, or
• designated in a local emergency response plan as critical.
An ordinary bridge is any bridge not classified as important.
It is noted that for important bridges, the same performance level is required for both earthquakes,
but greater damage is tolerated for the rare earthquake (SEE) than for the frequent event (FEE). For
ordinary bridges, less stringent performance is required for the SEE than for the FEE and a greater
level of damage is also tolerated. There is a similar differential between the performance and damage
levels for ordinary and important bridges.

Applications in California

As noted above, Caltrans has been using explicit dual level designs for major bridges for almost a
decade, and implicitly for ordinary bridges for about the same period of time. Two examples, where
the above criteria have been applied, are the new I-80 crossing over the Carquinez Straits in the North
Bay (40 km north of San Francisco), and the new East Bay structure between San Francisco and
Oakland. Both bridges are major lifeline structures and of critical importance to the region. As a
consequence both are classified as important and fall into the right hand column of the performance
matrices, where stringent performance is required even for rare events.
For the FEE, it is intended that there be no loss of capacity in the Carquinez Bridge (725m main
span), that damage that does occur be minimal and not require immediate repair, and that there be no
permanent structure displacements. For the SEE, there again be no loss in capacity, damage that does
occur be repairable without disruption, and no permanent structure displacements.

18
 The new East Bay crossing of the San Francisco Oakland Bay Bridge, which will replace the truss
spans damaged in the 1989 Loma Prieta earthquake, comprises four separate structures totaling about
3 km in length. The main structure is a single-tower, self-anchored suspension bridge with a 385 m
main span. The FEE for this bridge is a 450-year event, and full service is required almost
immediately after this event with only minimal damage. Thus essentially elastic performance is
required which implies only minor cracking, no apparent permanent deformations and no damage to
the expansion joints. The SEE has a return period of 1500 years and again full service is required
almost immediately after this earthquake. Any damage that does occur must be repairable with
minimal impact on functionality. Such damage must be limited to reinforcement yield, spalling of
concrete cover, and minor yielding of structural steel. Despite the relatively high FEE, the design has
been governed by the SEE event. It is noted that for wind, the 100-year speed has been used for
service load design and that a 10,000-year speed was used to determine the critical flutter velocity
threshold.
Much of the above discussion also applies to the retrofit of existing structures. As a third example,
the criteria proposed for the retrofit of Golden Gate Bridge in San Francisco are described. Instead of
two-level criteria as above, three earthquake levels are defined and performance criteria developed for
each. The three levels are: (1) frequent but small earthquakes, (2) moderate earthquakes, and (3) the
maximum credible earthquake.
Performance criteria are as follows. For the frequent but small events, there should be no loss of
capacity, only minimal damage and then only to non-essential elements; no elements requiring
immediate repair, and no impact on margins of safety. For moderate earthquakes, there should be no
loss of capacity, no disruption to regular function, and repairable without disruption. Damage that
does occur will be small, and although residual stresses and deformations will be tolerated, they will
be confined to non-critical members. For the maximum credible earthquake, the bridge will not be
closed for more that 48 hours. Nevertheless emergency vehicles must have immediate access; public
transportation has access within ‘days’ and the structure must be fully operational within ‘weeks’.

VERIFICATION OF PERFORMANCE OBJECTIVES FOR BRIDGES

Traditional methods of seismic design are force-based using either modified forces from elastic
models (including R-factor methods) or nonlinear capacity and demand analyses (including pushover
methods). Some of these methods are listed below in order of increasing rigor and complexity
(MCEER 2001a):
• Capacity spectrum method, in which demand and capacity evaluation are combined in a single
procedure. Method is restricted to very regular structures, which can be modeled as single degree-
of-freedom systems; is the basis of the AASHTO guide specification for isolated bridges.
• Elastic response spectrum methods, in which demands are calculated from response spectrum
analysis using elastic spectra and single- or multi-mode techniques depending on the complexity
of the structure. R-factors are used to obtain design forces based on assumed capacity of the
structure for inelastic action. Design displacements are set equal to elastic displacements.
• Nonlinear static displacement capacity verification methods (pushover analysis), in which the
displacement capacities of individual bridge substructures are determined from lateral load-
displacement analyses taking into account the nonlinear behavior of their components.
• Nonlinear dynamic analysis methods, in which force and displacement demands are found from
step-by-step time-history analyses using ground motion records and taking into account the
nonlinear behavior of various bridge components.

Although the development of capacity-spectrum and capacity-verification methods have greatly

improved the analyst’s ability to directly address various damage states, and by implication, various
performance objectives, they are essentially force-based and appear to be less powerful than the newer
displacement-based methods which use nonlinear displacement spectra rather than acceleration
spectra to characterize the earthquake loads.
In these latter methods, displacements and deformations are calculated directly and forces follow
from the displacements. Since many damage states and performance objectives are in fact

19
displacement states, a displacement-based method allows these damage states to be targeted directly.
For regular bridges, that may be modeled as single-degree-of-freedom systems, the method is
straightforward and may be summarized in the following nine steps (after Priestley 2000):
Step 1: Develop an equivalent single-degree-of-freedom model of the bridge.
Step 2: Select the damage state (performance objective) for the earthquake under consideration.
Step 3: Determine the displacement of the bridge, ∆max, that is consistent with the selected damage
state, using a nonlinear static analysis of the structure (a pushover analysis) or similar.
Step 4: Calculate the corresponding displacement ductility demand, µ
Step 5: Calculate the corresponding equivalent viscous damping ratio, ζ 
Step 6: Using ∆max and ζ, obtain the required effective period, Teff, from the displacement spectrum
for the earthquake under consideration.
Step 7: From the effective period, calculate the required effective stiffness, Keff and hence the base
shear, Vb, from Vb = Keff * ∆max.
Step 8: Check actual Keff against required Keff from (7) and redesign substructure accordingly. Repeat
from (3) as necessary.
Step 9: Select new damage state for a new earthquake and repeat from (2).

The main advantage of the method is the transparent manner in which the damage state is used to
influence the design and the inclusion of nonlinear static analysis in a logical manner. Disadvantages
include relating damage states to performance objectives, the need for displacement spectra, and the
complications that arise for bridges that cannot be modeled as single-degree-of-freedom systems. But
it is noted that, apart from the need for displacement spectra, these disadvantages also apply to force-
based methods.

Damage States and Performance Objectives

In this paper, the term damage state has been used to describe, say, the first yield in the vertical
reinforcement of a bridge column, or the transverse offset of a girder, or the formation of a sufficient
number of fully developed plastic hinges to form a collapse mechanism. On the other hand
performance objective has been used to describe the serviceability of a bridge after an earthquake
usually expressed in such terms as ‘no collapse but closed to traffic except emergency vehicles’, or
‘unrestricted access after 24 hours’, or ‘no interruption to bridge function at any time’.
The relationship between damage states and performance objectives is ill-defined at best and this
is one reason for specifying both in Table I. But they are not independent of each other and
inconsistencies can arise if their interdependence is not realized. Attempts to link one with the other
have been made based on experience and engineering judgment. For example in the proposed
revisions to the AASHTO LRFD specifications for seismic design, (MCEER 2001a),
recommendations are made for limiting values of structure displacements and plastic hinge rotations
as they might affect serviceability. As noted in the commentary to these recommendations (MCEER
2001b), these values are based on consensus opinions of workshop participants and are subject to
review. Some typical values, based primarily on California experience, are shown in Table II.

SEISMIC RISK ASSESSMENT PROCEDURE FOR HIGHWAY SYSTEMS

General Description

An outline of a seismic risk assessment (SRA) procedure for a highway system is shown in Figure
1, where it is seen to involve four main steps. These are: (1) initialization of the SRA; (2)
development of system SRA results for each scenario earthquake and simulation specified under Step
1; (3) incrementation of the simulations and the scenario earthquakes and repeat of Step 2; and (4)
aggregation of the SRA results for all earthquakes and simulations.
This SRA procedure has several desirable features. First, it may be carried out within a
geographical information systems (GIS) framework, which enhances data management, improves the
efficiency of the analysis, and enables the immediate display of analysis results. Second, if the GIS

20
 TABLE II. PROPOSED BRIDGE DAMAGE STATES FOR SELECTED PARAMETERS
ACCORDING TO REQUIRED PERFORMANCE OBJECTIVE (MCEER 2001a)

PERFORMANCE OBJECTIVE

PARAMETER Bridge Type 1 Bridge Type 2

Life Safety Only Fully Operational
Significant Disruption Immediate Access

Column plastic hinge rotation 0.035 rad or by analysis 0.01 rad

Vertical offset in girders 0.2 m 0.03 m

Horizontal offset in deck shoulder width 0.1 m

Longitudinal joint opening or

not more than seat width 0.1 m
1.0 m whichever is less

database is modular, the addition of improved data, procedures, and models as they are developed
from future research and development efforts is facilitated. Third, the procedure enables the effects of
uncertainties in the earthquake characterization, hazard models, and vulnerability models to be
considered, and has the capability of developing aggregate SRA results that could be either
deterministic or probabilistic, depending on user needs.
Four modules comprise the SRA procedure as shown in Figure 1. These are noted below:
• system module: network inventory, traffic data, origin-destination zones, trip tables, traffic
management, network analysis models
• hazards module: seismic zones, topography, local soils, ground motion attenuation, geologic
hazard models, model uncertainties
• component module: structural data, repair costs, repair procedures, traffic states, loss models,
fragility models, model uncertainties
• economic module: economic sectors (locations, productivity, damageability), stakeholder
impacts, economic models

Detailed descriptions of these modules are given by Werner et al (2000).

Demonstration of Seismic Vulnerability Assessment

The above SRA methodology has been applied to the highway system in the City of Memphis
(Figure 2) in conjunction with currently available data and models, to demonstrate the application of
the procedure and give one example of the type of results that can be obtained.

System

The City of Memphis is located in the southwestern corner of Tennessee, just east of the Mississippi
River and just north of the Tennessee-Mississippi border (Figure 2). Because of its proximity to the
New Madrid seismic zone, the potential seismic risks to the Memphis area are well recognized and
have been studied extensively. The highway system evaluated under this demonstration exercise
includes the beltway of interstate highways that surrounds the city, the two crossing of the Mississippi
River (at Interstate Highways 40 and 55), major roadways within the beltway, and highways just
outside of the beltway that extend to important transportation, residential and commercial centers to
the south, east, and north. The system contains a total of 286 bridges.

21
 (a) Outline of four-step procedure

(b) Modules comprising GIS database

Figure 1. Seismic Risk Analysis Procedure for Highway Systems (Werner et al 2000).

Assumptions

As noted earlier, this demonstration SRA is based on currently available data and models only.
Because the data and models are very preliminary at this time, it has been necessary to make certain
simplifying assumptions in this assessment. These include:
• a scenario earthquake, with a moment magnitude = 5.5, and epicentral distance between 35
and 50 km to the closest and furthest points of the Memphis highway system

22
 • traffic flow and volume data, roadway capacities and O-D zones as provided by the Memphis
and Shelby County Office of Planning and Development (OPD); traffic flow data were based
on OPD’s 1988 traffic forecasting model
• ATC-25 loss models for conventional highway bridges that differentiate between simple span
bridges and those with continuous girders but do not consider the influence of other structural
attributes on bridge performance (ATC 1991)
• simplified functionality models for estimating closure impacts and restoration times for
simple span and continuous girder bridges; post-earthquake traffic management was not
considered.
• MINUTP traffic forecasting models for calculating O-D times pre- and post- earthquake

Figure 2. Highway Network for City of Memphis.

Results

Detailed results of this study are given by Werner et al (2000). In this short paper only the overall
travel times and distances are presented to illustrate the effect of network redundancy on system
response. Implications of these results are further discussed by Buckle et al (1998).

Overview of Seismic Vulnerability Assessment Procedure.

The results of this SRA illustrate the potential impact of earthquake damage on traffic flows in the
Memphis area highway system. The analysis consisted of two parts. First, the PGAs estimated for
the scenario earthquake were applied to the fragility models after each earthquake in order to estimate
the state of the system at times of three days and six months after each earthquake (in terms of the
number of available lanes along each roadway in the system). Then, the effects of any reductions in
the available lanes (due to earthquake damage) on traffic flows throughout the system were estimated
by using the MINUTP transportation forecasting software, together with a regional traffic capacity
and flow data base developed at the Memphis and Shelby County OPD. From this, travel times and
distances throughout the system after each earthquake were compared to pre-earthquake travel times
and distances (in which all travel times and distances are average values for a 24 hour period).
Overall travel time and distance for the entire system may be compared, which is computed as the
sum of the travel times and distances respectively between all origin-destination (O-D) zones in the

23
system. This set of comparisons provides an approximate measure of the impacts of the earthquake
on overall system performance. Further, a breakdown of these total travel times and distances may be
compared for particular key O-D zones. These latter comparisons indicate the spatial distribution of
the earthquake impacts through out the system and also show how travel to, and from, critical O-D
zones are impacted by the earthquake damage to the highway system. Only the former comparisons
are presented below.

Overall System Travel Times.

Table III shows that, as a result of the estimated bridge damage due to the assumed earthquake,
overall system travel times three days after the earthquake are nearly 34 percent larger than the pre-
earthquake values. Six months after the earthquake, bridge repairs during that time have reduced the
overall system travel time; however it is still nearly 20 percent larger than the pre-earthquake value.

TABLE III. EFFECTS OF EARTHQUAKE ON TOTAL SYSTEM TRAVEL

TIMES AND DISTANCES.

Pre- Value @ Increase over Value @ Increase over

PARAMETER earthquake T=3 days pre- T=6 days pre-
value earthquake earthquake
value value

Total vehicle hours

traveled in 24-hr 3.73 x 105 4.99 x 105 33.8% 4.46 x 104 19.6%
period

Total travel
distance in 24-hr 15.5 x 106 15.6 x 106 small 15.6 x 106 small
period (miles)

Overall System Travel Distances.

Table III also shows that overall system travel distances are not sensitive to the estimated bridge
damage due to the earthquake, despite the fact that the total number of trips estimated over a 24-hour
period by MINUTP (solely on the basis of demographics) was nearly the same for the pre-earthquake
system and for the scenario earthquake. This lack of change of travel distances, despite significant
increases in travel times, is due to the availability of more direct, but less time-efficient, routes that
are taken after the earthquake. For example, if faster but less direct routes along interstate highways
and beltways, that would ordinarily be used, are closed because of bridge damage, slower but more
direct routes along city streets with no damaged bridges would be used instead.

THE CASE FOR PERFORMANCE-BASED DESIGN OF HIGHWAY SYSTEMS

The methodology described above has opened the door to implementing performance-based
seismic design for highway systems. As for highway bridges, the goal of such a design approach is to
satisfy certain specified performance criteria following earthquakes of different sizes, but in this case,
the objectives are set for a highway network or subset thereof. Following the approach for bridges,
successful application will require two issues to be addressed:
• Establishment of realistic and meaningful performance objectives at various hazard levels, and
• Verification of the performance objectives.

Performance objectives for highway systems might simply be related to changes in total system
travel times (Table III) for emergency traffic should a small, medium or large earthquake occur in the
region. More stringent criteria might be imposed for small and more frequent earthquakes, than for the
large and rare events. Alternatively, performance might be measured by system restoration time,

24
which is the time required to restore a system back to full capacity (or some fraction thereof)
following an earthquake. For small earthquakes this might less than a day, but for larger events
restoration times might be measured in months. Table IV presents a possible set of criteria based on
maximum acceptable restoration times, using two sets of times corresponding to 80 and 100%
restoration respectively.

System performance can be verified using the risk assessment methodology described above and
illustrated in Figure1. In this way more intelligent allocation of resources can be made with respect to
either seismic retrofitting of highway structures, or the deployment of emergency response measures.
For example, bridges might be retrofitted in order of their impact on overall system performance, such
as the total travel time for emergency vehicles, or the time required to restore 80% of the network
capacity. It may be found that retrofitting 10% of the deficient bridges in an inventory may be all that
is necessary to get a system back to 80% of its pre-earthquake performance. Such a result could have
a profound effect on the allocation of resources to bridge retrofit programs.

TABLE IV. PERFORMANCE CRITERIA MATRIX FOR HIGHWAY SYSTEMS

BASED ON RESTORATION TIMES (ILLUSTRATIVE PURPOSES ONLY)

HIGHWAY SYSTEM TYPE 1 HIGHWAY SYSTEM TYPE 2

EARTHQUAKE Standard Operating Essential Operating
Requirements Requirements

Frequent Earthquake T80 < 2 days T80 < 1 day

(FEE) T100 < 7 days T100 < 1 day

Rare Earthquake T80 < 30 days T80 < 7 days

(SEE) T100 < 90 days T100 < 30 days

NOTES:
1. A ‘highway system’ may be a subset of a larger highway network, subdivided according
to operational requirements.

2. Two classes of operating requirements are defined: standard and essential. ‘Essential’
requirements are more rigorous than ‘standard’ requirements.

3. System performance is measured by time required to restore network to given percentage

of traffic capacity before earthquake.

4. T80 and T100 are times required to restore system to 80% and 100% of capacity before
earthquake, respectively.

As with the performance-based design of bridges, consequential issues arise when considering
application to highway systems. For example, the uncertainty in the ground motion needs to be
reduced and the relationship between component damage states (e.g. bridge column crack widths) and
overall system performance (e.g. travel times to emergency care facilities) needs to be better
understood. Nevertheless the above tools show great promise and deserve further study.

CONCLUSIONS

A major review of performance criteria for bridges has been undertaken in the United States and a
move towards performance-based, multi-level seismic design of bridges has begun. In a parallel
exercise, a risk-based methodology has been developed for assessing the performance of highway
systems taking into account the seismic fragility of bridges and their interconnectivity. These efforts
have opened the door to performance-based seismic design of highway systems. The goal of such an
approach is to satisfy specified performance levels for highway systems immediately following
earthquakes of different size. Such criteria might be minimum delay times for emergency traffic for a

25
small, medium or large earthquake. More stringent criteria might be imposed for smaller and more
frequent earthquakes, than for larger and rare events. Alternatively, performance might be measured
by system restoration time, which is the time required following an earthquake to restore the system
back to full capacity (or some fraction thereof). For small earthquakes this might less than a day; for
large events restoration times might be measured in months.
It is concluded that there are significant benefits to be gained by combining the progress made in
performance-based bridge design with risk-based assessment of highway systems. Smarter use of
scarce sources in seismic retrofitting can be expected, together with more efficient deployment of
emergency services following a damaging earthquake.

ACKNOWLEDGEMENTS

The development of a performance-based seismic design methodology for bridges in the United
States has been the work of many people over the last decade. Principally funded by the California
Department of Transportation (Caltrans), the American Association of State Highway and
Transportation Officials (AASHTO), and the Federal Highway Administration (FHWA), the author is
grateful for this support and for the opportunity to contribute to this effort.
The development of a seismic risk assessment procedure for highway systems has been the
achievement of Stuart Werner at Seismic Systems and Engineering Consultants, and has also been
funded by the Federal Highway Administration. The assistance of the City of Memphis and the State
of Tennessee with the analysis of Shelby County is gratefully acknowledged.

REFERENCES

AASHTO (1992), Standard specifications for highway bridges, Fifteenth Edition, American Association of
State Highway and Transportation Officials, Washington DC.

AASHTO (1994), LRFD bridge design specifications, First Edition, American Association of State Highway
and Transportation Officials, Washington DC.

ATC (1991), Seismic vulnerability and impact of disruption of lifelines in the contiguous United States, Report
ATC-25, Applied Technology Council, Redwood City, CA.

Buckle, I.G., Friedland, I.M. and Werner, S.D. (1998), Seismic retrofitting of highway systems, Proc Asia-
Pacific Workshop Seismic Design & Retrofit of Structures, National Center for Research on Earthquake
Engineering, Chinese Taipei.

MCEER (2001a), Recommended LRFD guidelines for the seismic design of highway bridges, Part I:
Specifications, MCEER Technical Report, Buffalo, NY.

MCEER (2001b), Recommended LRFD guidelines for the seismic design of highway bridges, Part II:
Commentary and Appendices, MCEER Technical Report, Buffalo, NY.

Priestley M.J.N. (2000), Performance based seismic design, Bull NZ Soc Earthquake Eng., Vol 33,
No 3, pp325-346.

SEAOC (1995), Performance based seismic engineering of buildings, Structural Engineers Association of
California, Vol 1, Sacramento, CA.

USGS (1996), National seismic hazard maps, http://geohazards.cr.usgs.gov/eqint/cgi-bin/find-ll-l.cgi.

Werner, S.D., Taylor, C.E., Moore III, J.E., Walton, J.S., and Cho, S. (2000), A risk-based methodology for
assessing the seismic performance of highway systems, Technical Report MCEER-00-0014, Multidisciplinary
Center for Earthquake Engineering Research, University at Buffalo, Buffalo NY.

26
 Mitigate Earthquake Hazard & Risk for Highway Bridges
Through Planning, Design and Retrofitting
W. Phillip Yen1 and James D. Cooper2

EXTENDED ABSTRACT

An earthquake hazard is an inevitable natural hazard with the potential for large numbers
of fatalities and injuries, major property and infrastructure damage and serious disruption of
every day life. Our transportation systems, which connect nations, states and cities, are our
lifelines delivering daily needs such as food, water, and intra-city communication. Among these
systems, highway bridges are the most vulnerable to seismic damages. However, earthquake
losses may be reduced to a minimum through an integrated assessment and planning system, and
through better techniques and specifications for designing earthquake-resistant bridges and
highways. This paper summarizes seismic mitigation measures for bridges and highways
through planning, design and retrofitting.

The planning, recognized as part of “Risk Management”, is a process of deciding which

hazards at what scale should be managed, determining what should be done for a hazard, and in
what priority. This planning, called “Seismic Risk Assessment (SRA)”, evaluates earthquake
risks, and has a systemwide approach. It provides the methods, models, data needs, and
procedures for conducting a system-wide analysis of scenarios for earthquake impacts on a
highway system or sub-network, e.g. primary and secondary highways within a city, county or
other geographic region. This assessment methodology describes expected damage to the
highway elements and the resulting effect on vehicular travel between various origins and
destinations within the network. Developed to serve as a decision support (guide) tool for bridge
owners such as State highway departments. It includes establishment of appropriate levels of
seismic design, prioritizing the retrofit of existing bridges and development of post-earthquake
response plans. SRA contains procedures that provide a basis for addressing these seismic
performance issues and incorporate data and methodology pertaining to engineering issues
(structural, geotechnical and transportation), repair and reconstruction factors, system network
and risk analysis, and socio-economic effects from damage to the system. It also provides a
mechanism for estimating system-wide direct losses (i.e., costs for repair of damaged
components) and indirect losses due to reduced traffic flow and/or increased travel time
(economic impacts). A trail demonstration of the SRA procedure was performed to calibrate and
validate this methodology. This exercise was conducted on the highway system in Memphis,
Tennessee. The Memphis was selected because of its proximity to the New Madrid seismic
zone. The potential seismic risks to the Memphis area from this fault are well recognized and
have been studied extensively.

1
Research Structural Engineer, Office of Infrastructure, R&D Federal Highway Administration, 6300 Georgetown
Pike, McLean, VA 20121
2
Office of Bridge Technology, Federal Highway Administration, 400 7th Street NW, Washington, D.C. 20590

27
 Design is the first step for equipping bridges to resist earthquakes. Good design details
have saved many bridges from collapsing from unseating or shear failure. Design methods
evolved over time and produce details that directly affect bridge performance under earthquake
and other natural hazard loading. Design methods are steadily improved based on experience
with destructive earthquakes and advanced seismic research. The current seismic design
specification, adopted as a standard in 1992 by AASHTO, was primarily developed by US
highway agencies, including FHWA and CALTRANS. Realistic seismic provisions first entered
this code after the 1971 San Fernando earthquake. The fundamental design objective of the
current seismic specifications is to prevent collapse in large earthquakes. In small to moderate
earthquakes, the intent of the code is to resist these loads within the elastic range without
significant damage to structural components. In large earthquakes, no span or part of a span
should collapse. However, the AASHTO specs consider some damage acceptable in these
circumstances, provided it is limited to flexural hinging in pier columns and that it occurs above
ground in regions that are visible and accessible for inspection and repair. The design
earthquake is a single level event with a 475-year return period. Design forces are calculated
from an elastic analysis of the bridge using response spectra approximating the design quake. As
the result of an effort by the National Cooperative Highway Research Program and FHWA, a
recommended new seismic design specification was completed in March, 2001. This
recommended code contains significant changes in the design approach and criteria to reflect
lessons learned from recent earthquakes and research studies. A dual-level design method has
now been introduced in the recommended design specification. Bridge design objectives are
categorized in two levels of seismic performance. They are “Life Safety” and “Operational”.
The greatest advantage of the dual-level approach is that it addresses safety and functional
performance directly and separately to better assure that performance goals are met. The new
code will include Nonlinear Static Displacement Capacity Verification, (Pushover Analysis) in
the design procedure. This analysis, appearing for the first time in the design specifications, is a
displacement-based approach for analyzing dynamic response. The objective is to determine the
displacement at which the earthquake-resisting elements achieve their inelastic deformation
capacity. Damage states are defined by local deformation limits, such as plastic hinge rotation,
footing settlement or lift, and abutment displacement. Displacement may be limited by loss of
capacity such as degradation of strength under large inelastic deformation or ρ − ∆ effects.

Retrofitting is an urgent process to upgrade those bridges constructed (usually prior to

1971), with little or no consideration given to seismic forces. These structures are very
vulnerable to earthquake strikes, and need to be retrofitted based on site seismicity and structural
types. FHWA’s Seismic Retrofitting Guidelines for Highway Bridges was first issued in 1983
and was followed in 1987 by Seismic Design and Retrofitting for Highway Bridges. In 1995,
FHWA updated these manuals with more current knowledge and practical technology. This
revision, reflecting recent changes in retrofitting philosophy and performance criteria, was
adopted by AASHTO shortly after 1995. The fundamental retrofitting philosophy is to prevent
unacceptable damage. In general, unacceptable damage includes: 1) Serious injury or loss of
life, 2) Collapse of all or part of the bridge, and 3) Loss of use of a vital transportation route.
Seismic retrofitting is a planning and design solution for mitigating earthquake hazards to
existing bridges. Because not all bridges can be retrofitted simultaneously, the most critical
bridges should be retrofitted first. The selection of bridges for retrofitting requires an overall
evaluation of the structure and highway system including seismic risk, economics and societal

28
impacts. Thus, the process for the retrofitting bridges involves the assessment of multiple
variables and requires considerable judgment. The retrofitting process is divided into three
stages 1) Preliminary Screening, 2) Detailed Evaluation, and 3) Design of Retrofitting Measures.
This process ensures a significant improvement in seismic capacity of these older structures that
are at risk from earthquakes.

29
 Seismic Safety Evaluation of Large Scale Interchange
System in Shanghai

Lichu Fan1, Jianzhong Li1, Shide Hu1, Guiping Bi1 and Liying Nie1

ABSTRACT

The report for the seismic evaluation and retrofit of a large scale interchange system in
Shanghai Xin-zhuang has been completed by State Key Laboratory for Disaster Reduction of
Civil Engineering (SLDRCE) in Tongji University. In this paper, the important problems of the
seismic evaluation procedure and the analysis model were investigated. The suggestions for local
models of plate girder bridges and continuous girder bridges in a large scale interchange system
were given. Additional studies were conducted to determine pounding effects at the structure
interfaces in bridges.

__________________
1. State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, PRC

31
INTRODUCTION

Xin-zhuang interchange system is the largest transportation engineering in Shanghai, and

it is also the largest interchange system in Asia at present as shown in Figure 1. The interchange
site is between the outer ring road’s station at the No.1 Subway and the Xin-zhuang’s station.
The site locates approximately at 1km south west of the Xin-zhuang Town, and is the start-point
of the Hu, Hang, Yong’s freeway. Xin-zhuang interchange system includes four trunk highways
with six entrances and exits. The four layers interchange system with 20 directional bifurcated
girders to make a complex cloverleaf interchange. The whole system has 11.1km-long bridges
with total plan area of more than 84 thousand square meters. The highest construction at the
interchange system is 21 meters. In the original seismic design, according to the China Code for
Seismic Design of Highway Engineering (JTJ 041-89)[1], the seismic intensity 7 is considered
for the interchange system.

STRUCTURAL CHARACTERISTICS

The superstructure of bridges in Xin-zhuang interchange system includes four kinds of

girders:
(1) The concrete hollow plate girders with the depth ranging from 0.95m to 1.05m;
(2) The posttensioned box girders with uniform depth ranging form 1.1m to 1.6m;
(3) The irregular box girders with varied width;
(4) The steel and concrete composite girders.
The heights of columns of the piers range from 2m to 20m. The girders are directly
supported on the teflon bearings and the elastomeric bearings without connection between the
bearings and girders or cap beams. The teflon bearings are located at expansion joints. For
simple girder bridges, the inverse T type cap beams are used as shown in Figure 2. Most piers
have a specified concrete compressive strength of 30 Mpa with longitudinal steel
reinforcement ratio ranging from 1.4% to 2.5%. The volumetric ratio of lateral steel
reinforcement provided inside the plastic is about 0.3%.
In the bridges in the interchange system, the shear keys and seismic concrete block
structures (shown in Figure 3) were widely used to prevent a girder fall.

SEISMIC SAFETY EVALUATION PROCEDURE

According to the characteristics of Xin-zhuang interchange system, seismic evaluation

procedures for the overall bridge have evolved as outlined below:
(1) Earthquake safety assessment at site. This includes site-specific information on the
expected ground motion that consist of a basic responses spectrum shape, or shapes, with peak
ground acceleration related to annual probability, and artificial generated accelerograms that
closely match the responded spectrum.

32
 Figure 1. Xin-zhuang interchange system

Simple girder Simple girder

Inverse T type
cap beam
Simple girder

Figure 2. Inverse T type cap beam

Figure 3. Seismic concrete block

33
 (2) Testing of material behavior and investigating of structural component at site. In order to
determine the difference between the design and actual bridges, the concrete and reinforcement
strength of actual bridges in Xin-zhuang interchange system were tested and were compared with
the design values.
(3) Two level dynamic analysis. Linear Level-1 with global model of entire interchange
system and linear local models was first performed using three-dimensional elements. The global
model focuses on the overall behavior and includes all structural components. The global model
analysis is an important first step in the initial assessment of the seismic vulnerabilities of a
structure. Such analysis can provide initial indication of "hot spots" to plan an evaluation strategy,
categorize members by their demand/capacity ratio, and envelope peak response quantities to a
design spectrum. The linear local analysis models were applied to investigate the difference
between linear dynamic analysis of global and local models to get simplified model methods.
Nonlinear level-2 analysis is defined as nonlinear dynamic time history analysis with
considerations of geometrical nonlinearity, nonlinear boundary conditions, other inelastic
element (for example, bearings)
and inelastic members. Level-2 analysis is mainly applied to local models. The local models
emphasize the localized behavior, especially complex inelastic and nonlinear behavior.
(4) Determination of capacity for piers, bearings and connection elements. Based on
material properties for concrete and reinforcement, nominal flexural strength, shear strength of
structural components have been determined in accordance with specified code formula.
(5) Capacity/demand ratio analyses. Based on the results of nonlinear dynamic time history
analysis, capacity/demand ratio analyses were carried out. Based on the results of capacity/demand
ratio analyses, the suggestion plan for retrofit is given.
The flow chart outlining the step in the seismic evaluation procedures for a large scale
interchange system is given in Figure 4.

ANALYSIS MODELS

Global Model

The global model focuses on the overall behavior and includes all structural components in
interchange system as shown in Figure 5. The superstructure is modeled by three-dimensional linear
elastic beam-column elements placed at the geometric centroid of the cross section. At end of each
continuous beam or simple beam, the three-node frame ends are connected by rigid elements that
extend transversely from the centerline of the superstructure (detail A in Figure 5). For models of
the straight bridge, the rigid elements extend perpendicular to the centerline. For skewed models,
the rigid elements extend at skew angle. Three-dimensional linear elastic beam- column elements
are also used to model the column and cap beam for each of the piers in the bridges in the global
model. Each elastomeric bearing in bridges is modeled by a linear spring element as shown in

34
detail A in Figure 5. Total of 6310 three-dimensional linear elastic beam-elements and 2104 linear
spring elements were used in the global model.

Earthquake Safety Testing of Material Investigating Structural

Assessment at Site Behavior Components at Site

Global Analysis Model to Represent Overall Behavior

Local Model to Represent Localized Behavior

Emphasizing Complex Inelastic and Nonlinear Behavior

Level 1 AnalysisLinear Elastic Dynamic

Estimate Bearing and Analysis Estimate Flexure
Connection Element Level 2 AnalysisNonlinear Dynamic and Shear Capacity
Capacity Time History Analysis of Piers

Overall Seismic Evaluation for

Interchange System

Suggestion Plan for Retrofit

Figure 4. Seismic evaluation procedures

Detail A

3-1 Line

6-1 Line

Figure 5. Global model

35
Nonlinear Elements and Nonlinear Local Models

Based on the analysis results of global model and characteristics of the Xin-zhuang
interchange system, the linear and nonlinear analysis local models were established. The detail
description of linear local analysis models was presented in the report of the seismic evaluation
of Xin-zhuang interchange system [2]. Here, the typical nonlinear elements and nonlinear local
models are discussed as following:

Pier columns

The inelastic three-dimensional beam-column element with a fiber model of the cross
section [3] was used to model each column of the piers in the bridges. Figure 6 shows the fibers
of the section. Each fiber has a specified stress-strain relationship, which can be specified to
represent unconfined concrete, confined concrete, and longitudinal steel reinforcement. The
distribution of inelastic deformation and forces is simply by specifying cross section slices along
the length of the element. The fiber model approach provides versatile modeling of bi–axial
moment-axial force interaction with distributed inelastic hinges and can represent the loss of
stiffness caused by concrete cracking, yielding of reinforcing steel, and stain hardening. In this
study, Mander’s model for confined concrete and unconfined concrete [4] were used to represent
the stress-strain behavior of concrete as is shown in Figure7.

Bearing

As mentioned above, the girders in the bridges in the Xin-zhuang interchange system are
directly supported on the teflon bearings and the elastomeric bearings without connection
between the bearings and girders or cap beams. The horizontal sliding behavior of interface
between the bearings and girders or cap beams is presented by nonlinear spring elements with
bilinear model shown in Figure 8. The value of initial stiffness k 0 for the bilinear model is
determined by shear stiffness of a bearing. The frictional force, F f , at a sliding interface, may
be described by following equation (1).

f cc Confined
Fiber (x,y)
f c'
Unconfined

ε sp ε cu

Fig.6 Column elements with fibers Fig.7 Stress-strain behavior of concrete

36
 F
x
Ff
k0 k0

M x
c

..
ug

(a) (b)

Figure 8. (a) Nonlinear spring element; (b) Bilinear mode

Fy = µ d R (1)

in which R= the vertical reaction force of a bearing; µ d =sliding fraction coefficient of interface.

Pounding effects

Based on the characteristics of Xin-zhuang interchange system, the following three kinds
of collisions may occur during earthquakes:
(1) Collisions between the simple girders and inverse T type cap beams;
(2) Pounding of adjacent girder segments at expansion joints for the continuous girder;
(3) Collisions between the girders and the seismic concrete block structures;
The collision is modeled by a nonlinear spring element with gas as shown in Figure 9.
The nonlinear spring element with gap becomes active when the relative displacement between
adjacent structures is smaller than the initial gap D0 . The initial stiffness k 0 and post-yield
stiffness k1 of the spring are used to represent the elastic and plastic behavior of pounding
structures.
Force

k1
Gap k Qy
J
I k0

c
Displacement
D0 Dy

Figure 9. Nonlinear spring element with gap

37
Nonlinear local models

As mentioned above, in the bridges in the Xin-zhuage interchange system, there are three
kinds of bridge, simple girder bridges, curve continuous girder bridges and bifurcated girder
bridges. Here the typical nonlinear local models for a simple girder bridge and a curve
continuous girder bridge are given as following.
The 3-1 Line in Xin-zhuang interchange is a 36 span bridge with simply supported
concrete hollow plate superstructure. Pier10 to pier 12 are two column bents, and other piers are
single column bents. The column heights vary considerably over the bridge from 1m to 15m. The
6-1 Line includes simply supported girder and continuously supported girder bridges. The part of
continuous bridge is a 13 span bridge with box girder superstructure and single column pier. The
expansion joints located at top of pier 5, pier 10, pier 14 and pier 18.
The nonlinear local analysis models for simple the girder bridge in 3-1 Line and the
continuous girder bridge in the 6-1 Line are presented in Figure 10 and Figure 11 respectively.
The superstructure and the cap beam are modeled by three-dimensional linear elastic beam-
column elements. A nonlinear spring element with gap is used to model impact (Detail A in
Figure 10 and Figure 11). To be able to capture impact caused by in-plane rotation of the
superstructure, three nonlinear springs are used at each end of rigid element. The nonlinear
spring element with bilinear was used to model horizontal sliding behavior of interface between
the bearings and girders or cap beams. For the interface between the teflon bearing and girder
(cap beam), sliding fraction coefficient 0.02 is suggested; for the interface between the
elastomeric bearing and girder (cap beam), 0.15 is suggested. Each column of the piers in the
bridge is modeled by inelastic three-dimensional beam-column element with a fiber model.

36#

30#

24#

18#
D e ta il A
12#

Detail A
6#
Nonlinear Spring
Element with Gap Rigid Frame
Elastic Beam End Element 0#
Element

Seismic Concrete
Block Element
Bearing Element Cap Beam
Column Element
Element Bearing Element
with Fiber

Figure 10. Nonlinear local model of simple girder bridge in 3-1 line

38
 Detail A

12# 11# 10#

13# 9#
14#
15# 8#
16# 7#
17# 6#
18# 5#
Detail A
Nonlinear Spring Rigid Frame
Element with Gap End Element
Elastic Beam
Element

Bearing Element
Cap Beam
Column Element Element Bearing Element
with Fiber

Figure 11. Nonlinear local model of continuous girder bridge in 6-1 line

EARTHQUAKE LOADING

According to earthquake safety assessment at site, site-specific acceleration coefficient

(shown in Figure 12) and typical site-specific time histories of input acceleration for a 10% and
2% probability of exceedance in 50 years are provided. The input acceleration for a 10%
probability of exceedance in 50 years was used to represent design earthquake and the input
acceleration for a 2% probability was used to represent severe earthquake in the seismic safety
evaluation of the Xin-zhuang interchange system.

0.5
Acceleration coefficient

2%
0.4 10%

0.3

0.2

0.1

0.0
0 1 2 3 4 5
Period T(s)

Figure 12. Site- specific acceleration coefficient

39
ANALYSIS RESULTS

The computer program ANSYS [5] was used to perform linear spectral analysis for the
global and linear local model. For the nonlinear dynamic time history analysis, the specific
computer program developed by State Key Laboratory for Disaster Reduction of Civil
Engineering in Tongji University was adopted.
To assess the response, the maximum relative displacements between the girder and the
pier, column displacement, column curvature ductilities, column bent moments, column shear
forces and the impact forces for seismic concrete block were calculated with the site-specific
input acceleration of a 10% and 2% probability, respectively. The detail analysis results are
presented in the report of the seismic evaluation of Xin-zhuang interchange system. Parts of
import results are discussed as following:

Under Design Earthquake

The pier columns in the interchange system work generally in the elastic range and the
shear capacity of the columns is adequate to resist the design earthquake. Because the girders in
the bridges are directly supported on the elastomeric bearings without connection between the
bearings and girders or cap beams, the horizontal sliding between the elastomeric bearings and
girders or cap beams in simple girder bridges occur. The collisions between the girders and the
seismic concrete block structures induce a large impact force in simple girder bridges. The
maximum values of impact force between the girders and the seismic concrete block structures in
the 3-1 Line are about 9320kN.

Under Severe Earthquake

Parts of pier columns enter to plastic work range and maximum demand/capacity ratio of
curvature ductility is about 5.293. The shear capacity of the columns is adequate to resist the
severe earthquake according to China Seismic Code for Urban Bridges [6]. The horizontal
sliding between the elastomeric bearings and girders or cap beams in simple girder bridges and
parts of continuous occur. The typical sliding displacement history between the elastomeric
bearing and girder at pier 35 in 3-1 Line is shown in Figure 13.
The collisions at concrete block structures, expansion joints and inverse T cap beams
induce a large impact force, and may cause considerable damage or even lead to collapse of
colliding structures under severe earthquake. The typical pounding time history between the
girder and the seismic concrete block at pier 2 and moment time history at column bottom of
pier2 in 3-1 line are presented in Figure 14 and Figure 15. Figure 16 and Figure 17 show the
typical pounding time histories between the girder end and inverse T cap beam at pier 35 and
moment time history at column bottom of pier35 in 3-1 Line.

40
 0 .0 1

0 .0 0

Displacement (m)
- 0 .0 1

- 0 .0 2

- 0 .0 3

- 0 .0 4

- 0 .0 5

0 10 20 30 40 50
T im e (s )

Figure 13. The typical sliding displacement history between

the elastomeric bearing and girder

10000
2500
Pounding force (KN)

8000

Moment (KN.m)
2000 6000

1500 4000

2000
1000
0
500
-2000
0
0 10 20 30 40 50 0 10 20 30 40 50

Time(s) Time(s)

Figure 14. Pounding time histories at Figure 15. Moment time histories at
the concrete block column bottom

2000
Pounding force (KN)

8000
0
Moment (KN)

6000
-2000

4000
-4000

2000 -6000

0 -8000
0 10 20 30 40 50 0 10 20 30 40 50
Time(s) Time(s)

Figure 16. Pounding time histories at Figure 17. Moment time histories at
inverse T cap beam column bottom

41
Retrofit Suggestions

Based on the analysis results of the Xin-zhuang interchange system, following

suggestions are given for retrofit:
(1) To insure that sliding between the elastomeric bearings and girders or cap beams does
not occur under design and severe earthquakes, the bolts should be applied to connect elastomeric
bearings with girder or cap beam. The links shall be adequate to resist the design and severe
earthquakes.
(2) To reduce pounding effects in the interchange system, the rubber layer should be attach
on the inner side of seismic concrete block. A rubber layer on the interface of pounding can reduce
the pounding time, than reduce pounding forces.

CONCLUSIONS

Based on the characteristics of the large scale interchange system, the seismic evaluation
procedures have been investigated. Three kinds of analysis models, global model, linear local
and nonlinear local models are established, and two level dynamic analysis, linear spectral
analysis and nonlinear dynamic time history analysis were carried out for Xin-zhuang
interchange system.
The analysis results were investigated. The results show that the horizontal sliding
between the elastomeric bearings and girders or cap beams occur under design earthquake and
severe earthquakes. The collisions at concrete block structures, expansion joints and inverse T
cap beams induce a large impact force, and may cause considerable damage or even lead to
collapse of colliding structures under severe earthquake. Based on the analysis results,
suggestions are given for retrofit.

REFERENCES

Transportation Ministry of P.R.C. 1989. “Code for Seismic Design of Highway Engineering (JTJ 041-89)”
Lichu Fan, Jianzhong Li, Shide Hu, Guiping Bi and Liying Nie. 2001. “Seismic Safety Evaluation of xin-Zhuang
interchange system,” Stare Key Laboratory for Disaster Reduction of Civil Engineering in Tongji University.
Enrico Spacone, Filip. C. Filippou and Fabio F. Taucer. 1996. “Fiber Beam-Column Model for Nonlinear Analysis
of R/C Frames: Part 1. Formulation,” Earthquake Engineering and Structure Dynamics.25 (4):711-725.
J. B. Mander, M. J. N. Priestley and R. Park. 1988. “Theoretical Stress-Stain Model for Confined Concrete,”ASCE
Journal of Structure Engineering. 114 (8):1804-1826.
ANANS, Inc.1997. “ ANASYS Workbook Release 5.4,” Third Edition. SAS, IP Inc.
Stare Key Laboratory for Disaster Reduction of Civil Engineering. “China Seismic Code for Urban Bridges (Draft
2001),’’ Shanghai, P.R.C.

42
 A New Approach to Analysis of Soil-Pile-Structure
Interactions for Long-Span Bridges

Seung-Il Nam1 and Jamshid Ghaboussi2

ABSTRACT

Dynamic soil-structure interaction problems are a major part of the seismic analysis and
evaluation of the long-span bridges. The accurate modeling of major bridges with multiple spans
for seismic analysis requires that the whole bridge and its foundations be modeled, usually using
finite element method. This is specially so when nonlinear behavior is present in both the
superstructure and the substructure. However, finite element methods for bridges having
multiple spans crossing extended area are difficult to be implemented because of the enormous
amount of computations. As a result, simplified models have been proposed for modeling of the
foundations of the long span bridges.
The effectiveness of conventional Winkler-type foundation models become questionable
for the case of a large number of closely spaced piles, which often form the foundations of long-
span bridges. A new approach for developing a discrete parameter models using genetic
algorithms is introduced in this paper. This method gives the simplified Winkler-type foundation
models with finite-element-method-like accuracy regardless the complexity of the problem.
Simple preliminary results are presented. It is shown that the Winkler model constructed from
genetic algorithms gives satisfactory result for seismic analysis. Due to the flexibility of the
method, the model can be modified and even more simplified without severe loss of accuracy for
any given purpose.

__________________
1. Doctoral student, Department of Civil and Environmental Engineering, University of Illinois at Urbana-
Champaign, Urbana, IL 61801, USA
2. Professor, Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign,
Urbana, IL 61801, USA

43
INTRODUCTION

Considerable amount of research has been devoted to the seismic analysis of structures,
including the soil-structure interaction effects. For seismic analysis of the combined system of
the superstructure, its foundations and surrounding soil medium, various numerical techniques
are currently at hand. Most of these methods have been developed for seismic analysis of
buildings. Considerably less research effort has gone into developing methods for seismic
analysis of major bridges, which are inherently different, and behave differently under seismic
loads. In fact most of the methods for seismic analysis of major bridges are the extensions of the
methods of analysis that have been developed for buildings.
Soil-structure interactive effects are important in buildings and they can be included in
seismic analysis almost routinely. However, soil-structure interaction effects are far more
important in major bridges and they are more difficult to model with reasonable accuracy. In
fact, the very excitations that the major bridge superstructures experience are the direct result of
the foundation-soil-superstructure interaction. Current methods consisting of springs and
dashpots that are used in practice do not adequately account for the soil-structure interaction
effects in major bridges. Far more research is needed to properly account for these effects.
For the design and retrofit of large structural systems having strong nonlinearity in both
structure and soil, such as long-span bridges subjected to strong ground motions, it is imperative
to perform three-dimensional nonlinear finite element analysis in time domain in order to extract
transient and nonlinear characteristics of dynamic behavior.
Despite the extensive research on this subject, proper modeling of soil-structure
interactions in numerical analysis of soil-pile-structure systems is one of the most challenging
parts in the seismic analysis of long-span bridges. For finite element analysis of long-span
bridges, besides the difficulties in representing absorbing boundaries, extremely large number of
elements required to model the soil medium is still frustrating on most pc’s and microcomputers.
Therefore, representing soil-foundation-structure interactions with satisfactory accuracy and with
reasonable computational cost is the key task in numerical analysis of long-span bridges.
As alternatives to the finite element methods, boundary element methods using Green’s
function are often carried out to reduce the computational cost and to properly represent
infinitely large soil medium. Several hybrid models have been introduced (Guin and Banerjee,
1998, Pavlatos and Beskos, 1994) where finite element method is used for structural components
and then boundary element method is used for soil medium. However, boundary element method
is not suitable for describing the nonlinear behavior of soil. Another alternative method widely
accepted is the Winkler-type foundation models. These models easily allow for the nonlinear
behavior of soil and significantly reduce computational effort compared to finite element
methods (El Naggar and Novak 1666, Nogami et al. 1990). Wang et al. preformed a comparison
of various spring-dashpot models and they pointed out that radiation damping and nonlinear p-y
curves have significant effect on the response of the systems. Generally, the effectiveness of the
Winkler models is mainly controlled by the determination of nonlinear p-y curves of the spring
and the representation of radiation energy and the pile-to-pile interaction. Many long-span
bridges have foundations with a large number of closely spaced piles in complex arrangements.
Although the Winkler models are easily derived for simple single pile systems, the effectiveness
of the model becomes questionable for increasingly complex arrangement of pile groups.

44
 In this paper, a new method is introduced to develop Winkler-type foundation models for
pile foundations. The method is expected to properly construct Winkler models regardless of the
complexity of the geometry of the pile group and to produce the seismic response as accurately
as the response of the finite element model. The use of genetic algorithms play an important role
in the proposed method.

PROPOSED METHOD FOR NUMERICAL MODELING OF PILE FOUNDATIONS

In conventional Winkler models used in pile foundations, the effect of soil is represented
by two elements, the near-field element and the far-field element. Each element consists of
masses, springs and dashpot. Various methods have been developed to determine the values and
the variation of the masses, springs and dashpots. Near-field elements, representing the effect of
soil around the piles, generally are modeled as nonlinear springs, whose behavior is to be
determined by empirical formula. On the other hand, far-field elements consist of linear springs
and dashpots and they represent radiation of scattering waves from the structure. As far as far-
field behavior is concerned, there are analytical solutions in frequency-domain for homogeneous
linear halfspace. In the time domain, it is possible to closely represent the radiation of energy in
a certain range of frequencies with the proper choice of spring and dashpot constants (Nogami et
al. 1990). However, in reality the situation is far more complex. The ground conditions may
consist of layered deposits, or even more complex geological formations. For a soil deposit
composed of different layers or more complex geological formations applying analytical
solutions that cannot be adequately verified are unlikely to model the radiation effects correctly
and to produce reliable results.
Most methods for modeling the pile-soil interaction are developed for single piles. An
additional complicating factor is the pile-to-pile interaction in pile groups. The behavior of a pile
in a pile group can be very different than the behavior a single pile. Consequently, the elements
in the Winkler model should account for these differences. The method proposed in this paper
accounts for the pile group effects.
In the method proposed here, a conventional two elements Winkler model, where each
element has a mass, a spring and a dashpot, are adopted to account for both nonlinear soil
reaction around the pile and the effect of radiation damping. However, it has to be noted that
different models, with varying degrees of complexity can be used for this purpose. In the
proposed model we account for the pile group effects by placing another set of springs and
dashpots between the piles to represent pile-to-pile interactions.
The most important part of any pile foundation model is how to determine the values and
variation of the various springs and dashpots of the model. At the outset the discrete Winkler
parameters, spring and dashpot constants, etc., are treated as unknowns. A reference data set for
the same structure-foundation-soil system is needed to determine these parameters. In the
proposed method we first perform a detailed nonlinear dynamic finite element analysis of the
soil-pile-foundation system. This finite element model would normally contain the foundation,
the bridge pier above the foundation, and the pile group and a portion of the soil mass
surrounding the pile group, as shown in Figure 1. This finite element model is the subject to the
ground shaking and the response of the system is computed. The computed response of the finite
element model is then used as the reference data set.

45
 The problem of determination of the discrete Winkler parameters, spring and dashpot
constants from this reference data set is a classical problem of system identification and
parameter determination. This is accomplished by applying an optimization method. We have
chosen to use genetic algorithm for solving the optimization problem. The discrete parameters
of the system are the unknowns variables in the genetic algorithm. The values of the discrete
parameters evolve during the generations in genetic algorithm in such a way that the results of
the discrete Winkler type model matches the results of the finite element analysis. The main
steps of this process are schematically illustrated in Figure1.
Genetic algorithms are employed to carry out the optimization process. Genetic
algorithms are powerful tools for searching for the optimal solutions. Unlike the conventional
calculus based optimization methods, genetic algorithms search solutions in a global manner
where the derivatives of the object function are difficult or even impossible to find. In the
example presented later in this paper, displacement time histories of the finite element system at
certain locations in the pier and the piles are picked as reference data and they are compared to
those of the discrete parameter system. The parameters keep evolving such that the difference
between displacements of two systems becomes smaller and smaller. There is no way to find the
explicit search directions for the optimal values of the parameters with the displacement
difference. As discussed later, after evolutions, it is expected that the discrete parameter systems
are capable of replacing finite elements around the pile.

Figure 1. Schematic procedure for the proposed method

46
CASE STUDY

An I-57 bridge at Cairo across Mississippi River is examined as a case study. The bridge
consists of approach spans and a main span (see Figure 2). The soil deposit under the bridge is
about 160m thick and consists of 5 layers of sand and clay. In order to examine the dynamic
characteristics of the bridge, a three-dimensional finite element models was constructed using
SAP2000. Figure 3 shows the finite element model of the superstructure, along with three
models of the foundation system. The modal characteristics of the bridge showed that there are
two longitudinal modes around the period of 2.15sec, and 47 modes above the period of 1.0sec.
It is also noteworthy that it has a significant number of modes associated with the torsion of
approach spans. The torsional modes of the approach spans may play an important role in the
seismic behavior of the bridge.

The foundations of approach spans consist of a large number of group piles, while the
main spans have caissons. For example, a typical pier in the approach spans of the bridge has
over 80 piles under the 20m×7m pier base cap. The closely spaced pile groups make the
application of conventional Winkler modes nearly impossible. Arbitrary reduction factor can be
introduced to Winkler springs or a large number of piles can be replaced with equivalent 2 by 2
pile group. However, the effectiveness of those methods has not been verified for seismic base
excitations. The method proposed here is capable of producing the simple discrete parameter
model for any complex arrangement of the pile groups.

Figure 2. Elevation view of Cairo Bridge

47
 (a)

(a)

(b)

(c)

Figure 3. Finite element models using different foundation models: a) fixed base, b) 6 by 6
stiffness matrix, c) nonlinear springs at pile nodes.

48
49
Figure 4. Displacements at the bent cap of pier 1 (after Kornkasem, 2001). Case 2A: 6 by 6 stiffness matrix without pile interaction,
Case 2B: 6 by 6 stiffness matrix with pile interaction, Case 6A: nonlinear springs at pile loads.
 Artificially generated acceleration time histories developed for the use in the New Madrid
seismic zone (Hwang 1998) were applied at the bedrock to obtain ground motions. As a
preliminary study, three types of models were used in time-domain analysis. In model, A, the
base of each pier is fixed and the model is completely linear, while in model B the base is still
fixed but nonlinear behavior in bearings and expansion joints were taken into account. In model
C, in addition to the nonlinear behavior in the superstructure, spring elements were attached at
the base of each pier to account for the foundation impedance. Absolute values of maximum
displacements along the centerline of the deck were given for each model. The results show that
there exist significant differences between longitudinal and transverse displacements from model
A and C or B and C as shown in see Figure 4. This implies that soil-structure interactions play a
more important role in those displacements that in the vertical displacements. Transverse
displacements are observed to be so high that they are capable of causing severe damage to
bearings. Significant torsion is also present between shear centers and mass centers of the bridge
decks and truss sections

6.00E-03

u5 (DP)
4.00E-03 u7 (DP)
Displacement (m)

u5 (FE)
2.00E-03 u7 (FE)

0.00E+00
0.000 0.200 0.400 0.600 0.800 1.000 1.200 1.400 1.600 1.800 2.000
-2.00E-03

-4.00E-03
(a)
-6.00E-03

3.00E-03

2.00E-03
Displacement (m)

1.00E-03

0.00E+00
0.000 0.200 0.400 0.600 0.800 1.000 1.200 1.400 1.600 1.800 2.000
-1.00E-03

-2.00E-03

-3.00E-03
(b)
-4.00E-03

Figure 5. Displacements at selective nodes. a) F(t) = step loading, b) F(t) = banded white noise

50
 The seismic behavior of this bridge has been thoroughly examined with various
foundation models (Kornkasem 2001). Cases for a fixed base, linear and nonlinear 6 by 6
stiffness matrices at the base of the bridge pier and nonlinear springs attached at pile nodes are
analyzed using SAP2000, as shown in figure 4. Pile-to-pile interaction effects have been taken
into account for 6 by 6 stiffness matrices by introducing the static interaction factor. It was
pointed out that this interaction factor tends to overestimate stiffness reductions. It was also
shown that different foundation models have a significant affect on the dynamic response and the
natural periods of the bridge.

PRELIMINARY STUDY WITH THE PROPOSED METHOD

The 2 by 2 pile group has been analyzed with the method described above. Three-
dimensional finite element mesh including the bridge pier, pier base, piles and soil medium were
constructed. The same pier and piles were constructed for the discrete parameter model, where
nonlinear springs were used in near-field elements in order for the response to resemble
nonlinear finite element analysis. Two sets of loading time histories, step loading and banded
white noise, were applied at the top of the column and the displacement time histories were used
to determine the spring and dashpot constants. The frequency content of the banded white noise
should span the sufficiently large range to obtain the discrete parameter system that works in any
possible loading condition. After the evolution of genetic algorithms, the displacements of the
discrete parameter model are shown to be similar to those of the FE model, as shown in figure 5.
In the figure, u4, u5 are nodes of the frame above ground and u6 and u7 are nodes of the pile
below the ground. To verify the performance of the model, different superstructure geometry
with a different loading were analyzed. It can be noted that the Winkler model determined by
genetic algorithms closely matches the response of finite elements model, except the high
frequency response of the pile.

CONCLUDING REMARKS

As shown in the case study, the soil-structure interaction effects plays an important role
and their modeling is one of the most important factors in the seismic analysis of the long-span
bridges. The modeling of the whole superstructure and the foundation system, specially the pile
group foundation, is very important in computing realistic seismic response of major bridges, as
demonstrated in this paper. Moreover, we have also shown the importance of detailed modeling
of the foundation and the soil-foundation interaction. More realistic and detailed modeling of the
pile foundation lead to the response of the superstructure that is very different than when the
simplified models are used.

51
 A new approach to construct computationally efficient, yet relatively accurate discrete
parameter model has been proposed for pile foundations. In this approach the parameters of
discrete Winkler type models are determined by matching the response of a detailed finite
element model that includes the foundation and the pile group. We have proposed using genetic
algorithm for the purpose of parameter identification. Genetic algorithm is a powerful tool for
this purpose and it is highly flexible. This research is still in progress and we have only
presented the results of a preliminary study. A simple 2 by 2 pile group is examined as a
preliminary study and it is shown that the proposed method produces the Winkler-type model
that performs as well as the nonlinear three-dimensional finite element analysis.
Having the flexibility to modify and simplify the model and not to have to depend upon
analytical solutions or empirical formula, the proposed method has advantages over the
conventional methods for computing Winkler foundation models for the case of extremely
complex arrangement of piles. The future research should be focused on the actual complex pile
groups and the verification of the capability of the method to produce simplified and accurate
models. The method requires carrying out finite element analysis every time a discrete parameter
model is constructed. It is possible to make design curves for constructing Winkler models
through extensive parameter study using this method.

ACKNOWLEDGEMENTS

The research reported in this paper was funded by National Research Foundation
through a grant from the Mid-America Earthquake Center. The first author was supported by
this grant. This support is gratefully acknowledged. The opinions expressed in this paper are
solely those of the authors.

REFERENCES
Cai, Y. X., Gould, P. L. and Desai, C. S. 1995. “Investigation of 3-D nonlinear seismic performance of pile-
supported structures”, Performance of Deep Foundations under Seismic Loading, ASCE Geotech. Special Pub.
No. 51, 17-31.

Elassaly, M., Ghali, A. and Elbadry, M. M. 1995. “Influence of soil conditions on the seismic behaviour of two
cable-stayed bridges”, Canadian J. Civ. Eng., 22, 1021-1040.

El Naggar, M. H. and Novak, M. 1996. “Nonlinear analysis for dynamic lateral pile response.”, Soil Dynamic and
Earthquake Engineering, 15, 233-244.

Finn, W. D. L., Thavaraj, T. and Wu, G. 1996. “Seismic Analysis of pile foundations: State-of-art”, Proc. 11th
WCEE, paper no. 2073.

Finn, W. D. L. and Wu, G. 1996. “Nonlinear seismic analysis of pile foundations”, Proc. 11th WCEE, paper no.
1414.

Goldberg, D. E. 1989. Genetic Algorithms in Search, Optimization, and Machine Learning, Addison-Wesley.

Guin, J. and Banerjee, P. K. 1998. “Coupled soil-pile-structure interaction analysis under seismic excitation”, J. of
Struct. Eng., ASCE, 124, no 4, 434-444.

52
Hwang, H. M. 1998. Artificial acceleration time histories, Center for Earthquake Research and Information,
University of Memphis.

Kornkasem, W. et al. 1999. “Seismic analysis of a truss-arch bridge across the Mississippi River”, Presented for
publication on the 5th U.S. Nat. Conf. Lifeline Earthquake Eng.

Kornkasem, W. 2001. Seismic Behavior of Pile-Supported Bridges, Ph. D. Thesis, University of Illinois at Urbana-
Champaign.

Nam, S.-I. et al. 2000. “ Analysis of soil-structure Interaction of major river-crossing bridges.”, Proceedings of
12WCEE. Paper no. 0496.

Nogami, T., Otani, J., Konagai, K. and Chen, H.-L. 1992. “Nonlinear soil-pile interaction model for dynamic lateral
motion”, J. of Geotech. Eng., ASCE, 118, no 1, 89-106.

Pavlatos, G. D. and Beskos, D. E. 1994. “Dynamic elastoplastic analysis by BEM/FEM.”, Engineering Analysis by
Boundary Elements, 14. 51-63.

Wang, S. et. al. 1998. “Nonlinear seismic soil-pile structure interaction.” Earthquake Spectra, EERI, 14, no 2, 377-
396.

53
 Preliminary Study of Hydrodynamic Effects on Seismic
Response of Bridges

Jun-jie Wang1, Wei Lai1, Ning-yong Zhang1, Li-min Sun and Li-chu Fan1

ABSTRACT

The planning background of sea strait bridges remind engineers paying close attention to
the aseismic design of bridges in deep water. In this paper a preliminary study of the water
dynamic effects on seismic responses of bridges in deep water has been carried out. The Morison
Equation is used to model water dynamic pressure under seismic action, but only the inertia term
of Morison Equation is considered. Two types of bridges are analyzed, the first is long span
bridge with GBS, and the second is short span bridge with group pile foundation. The maximum
water depth is about 280m for the first type of bridge, and the water depth for the second bridge
is about 30m. The results show that the seismic hydrodynamic pressure may be important for
correctly predicting the responses of bridges in deep water under seismic action in some cases, it
may changes the seismic response by 10% to 100%.

_____________
1
State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, PRC

55
INTRODUCTION

Sea straits and sea bays dotted in China have been great obstacles of economy development
for regions in a country, such as Hainan Province in China, and archipelago of Zhoushan in China.
Recently, a planning was lunched by the government of China (in 1996) to construct a series of sea
strait(or sea bay) crossing bridges(or tunnels), one of them is shown in Figure 1. In fact, there are
many sea-crossing bridges in planning worldwide. Figure 2 is an example.
To cross these sea bays or sea straits, one of the solutions is to construct bridges, or part
of a crossing link consists of bridges. These bridges stand in sea with water depth up to
80-100m, the water might have important effects on seismic response of the bridges. However,
engineers have a lack of knowledge about such a problem.
The prediction of seismic response of bridges in deep water requires special
consideration due to fluid-structure interaction. In 1965 Hisao Goto and Kenzo[1] Toki[1],
Kotsubo[2] investigated the seismic response of bridge pier in water. From 1973-1980, a series of
research work had been finished by Liaw, Chopra, Rea, Byrd, and Nilrat[3-6]. The above pioneer
research work demonstrated the importance of the fluid-structure interaction in the earthquake
response behaviour of tower surrounded by water. The experimental research had also been done
by Rea et al[5] and Byrd[6]. The experimental results are used to verify the analytical theory and
to find main factors for hydrodynamic forces for tank-type structures.
In 1980, Nilrat[7] proposed a finite element method for determining the hydrodynamic
pressure distribution on rigid axisymmetric bodies oscillating in a fluid. The results of the finite
element analysis are compared with experimental results showing good comparisons in most
cases. In 1986 Tanaka and Hudspeth[8] developed employed an eigenfunction solution to
calculate the hydrodynamic forces under sinusoidal horizontal ground motion, and the results
from eigenfunction solution was verified by experimental data from a squatty circular column
and a slender circular column.

Figure 1. The Bohai Bay. Figure 2. Honshu-Hokkaido Strait in Japan.

56
 So far the research about the importance of hydrodynamic forces to structures in water
mainly focuses on three types of structures, marine tanks, intake-outlet towers[3-5,11-13], offshore
platforms[14,16,17]. However, few papers contribute to the bridges in deep water, eg. sea strait
bridges.
As a preliminary investigation, the Morision equation is employed to consider the water
effects on the seismic response of bridges in this paper, and two types of bridges are analyzed,
the first is long span bridge, and the second is bridge with group piles foundations.

DYNAMIC EQUATION

Seismic action will produce relative motion between the bridge pier and sea water. This
relative motion acts on the bridge pier in two manner, the first is the inertia action because of the
change of the water, and the second is the viscosity action of the water. For a slender column (eg.
pile) with small diameter, the water dynamic effects on the pier can be expressed by Morision
equation:

1
F = ρVu
&& + (C M − 1) ρV (u
&& − &x& − &x& g ) + C D ρA P [(u& − x& − x& g ) | (u& − x& − x& g ) |] (1)
2

where ρ is the density of the water; u& 、 u

&& are the absolute velocity and acceleration the water

respectively; x 、 x& and &x& are the relative displacement, velocity and acceleration respectively;
&x& g is the seismic ground motion; C M is the inertia factor, and C D is the viscosity factor. In this

study, C D is equal to 2，and C M is equal to 2.0.

In this study, it assumed that the water is still, i.e., u& = u
&& = 0 , therefore equation (1) can
be rewritten as,

1
F = −(C M − 1) ρV (&x& + &x& g ) − C D ρA P [(x& + x& g ) | ( x& + x& g ) |] (2)
2

then one can obtain the motion equation of a bridge under seismic action with the water effects,

1
M&x& + Cx& + Kx = −M&x& g − (C M − 1) ρV (&x& + &x& g ) − C D ρA P [( x& + x& g ) | ( x& + x& g ) |] (3)
2

The third term in the right hand is nonlinear one, by adopting the linearization technique[5] to this
term, one can obtain the following linearized equation,

57
 1 8
M&x& + Cx& + Kx = −M&x& g − (C M − 1) ρV (&x& + &x& g ) − C D ρA Pσ x& + x& g ( x& + x& g ) (4)
2 π
Then,
[M + M w ]&x& + [C + C w ]x& + Kx = −[M + M w ]&x& g − C w x& g (5)

1 8
M w = (C M − 1) ρV, C w = C D ρA P σ x& + x& g (6)
2 π

where σ x& + x& g is the variance response of absolute accelerations of the bridge. σ x& + x& g is a unknown

vector, the computation of sub-matrix of concerns the unknown σ x& + x& g , an iterative procedure is

needed.
If the effect of water viscosity is neglected, the motion equation of a bridge under the
seismic action can be simplified,

[M + M w ]&x& + Cx& + Kx = −[M + M w ]&x& g (7)

In this simple case, the effects of the water is equivalent to an additional mass matrix.
The additional mass matrix changes the dynamic characteristics of the analyzed bridge, and then
changes the seismic responses of the analyzed bridge. In this paper only the inertia effect is taken
into account.

WATER EFFECTS ON LONG SPAN BRIDGE

In general, the superstructure of a sea strait bridge take the usual configuration as the
long span bridges which go across a river. But a sea strait bridge featured by its type of
foundation, an example is shown in Figure 3.
In this paper as an example the planning bridge for Honshu-Hokkaido[18] strait in Japan is
analyzed, and the effects of water on the seismic responses long span bridges is investigated.
To verify the effects of the water on the seismic responses of a long span bridge over a
sea strait, two structural models are used in computation: (1) the pylon only; (2) the plane bridge
model. Six cases of the water depth are considered in the analyses, which are 0, 30, 60, 90, 120
and 150m, and the depth of 150m is the natural situation.
The comparison between the free vibrational frequencies without water and with water of 150m
are shown in Figure 4. It can be seen that the hydrodynamic pressure largely changes the he free
vibrational frequencies of the bridge, therefore it can be expected that the hydrodynamic
pressure would largely changes the seismic response of bridge.

58
 The Cross

The Tower

Plane

Transverse Longitudinal
Figure 3. The Planning Bridge Solution For Honshu-Hokkaido Strait in Japan.
Define the ratio:

RSP − RSP0
Rw = × 100 (8)
RSP0

where RSP is the seismic response for case with water, and RSP0 is the seismic response
without water. The results of R w are shown in Figure5 and 6.
4

3 without water
with water ( 150m )
Frequency (Hz)

0
0 5 10 15 20
modal number
Figure 4. Frequencies of the Analyzed Bridge.

59
 600 600
water depth (m) water depth (m)
500 30 500 30
60 60
400 90 400 90
120 120
300 150 300 150
H(m)

H(m)
200 200

100 100

0 0
-40 -20 0 20 40 60 80 100 120 140 -40 -20 0 20 40 60 80 100 120 140
Rw (%) Rw (%)

Bending Moment Shear Force

Figure 5. The Water Effects on Seismic response For Tower Model.

600 600
water depth (m) water depth (m)
500 500 30
30
60 60
400 90 400 90
120 120
300 150 150
H(m)

300
H(m)

200 200

100 100

0 0
-20 0 20 40 60 80 100 120 -20 0 20 40 60 80 100 120 140
Rw (%) Rw (%)
Bending Moment Shear Force
Figure 6. The Water Effects on Seismic response For Plane Bridge model.
600 600
Longitudinal
500 500 Longitudinal
Transverse
Transverse

400 400

300 300
H(m)
H(m)

200 200

100 100

0 0
-20 0 20 40 60 80 100 120 -40 -20 0 20 40 60 80 100 120 140
R w(%) R w(%)

Bending Moment Shear Force

Figure 7. Responses in Both Longitudinal and Transverse Direction For Tower Model.

60
 600 600
Longitudinal Longitudinal
500 Transverse 500
Transverse

400 400

300 300

H(m)
H(m)

200 200

100 100

0 0
0 20 40 60 80 100 -20 0 20 40 60 80 100 120 140
R w (%) R w (%)

Bending Moment Shear Force

Figure 8. Responses in Both Longitudinal and Transverse Direction For Plane Bridge Model.

At first, with the increase of water depth, the seismic response of the tower simply
increase, reach its maximum value at water depth of 120m, but get less at water depth of 150m.
In fact the dynamic effect of water is equvalent a additional mass matrix, which change the
inertia of the tower, so the dynamic characteristic of free vibration of the tower changes. The
effect of this change on the seismic response of the tower depends on the characteristics of the
input seismic ground motion. In principle, the tower will obtain its maximum response if the
dominate frequency of input seismic ground motion coincides with controlling frequency of free
vibration of the tower. For the tower, the water depth that will give maximum response is 120m.
Therefore it is not always true that deeper water gives larger seismic response of the tower.
In general, the effect of the water is really important for the correct predict of the seismic
response of the tower, and the net increase may reach 100%. The effect of the water on the
seismic response of the tower changes with the height of the tower, Rw get its maximum values

at the bottom or at some height of the tower, it depends on which response is considered. This
shows the complexity of the water effect on the seismic response of the tower.
The results from the tower model and the results from plane bridge model reach the same
conclusions.
The Figure 7 and 8 show the comparison of water effect on the seismic response in
transverse direction and the longitudinal direction. The same conclusion can be reached for seismic
response of the tower in the transverse direction.
Two examples of the time histories of seismic responses for tower model are shown in
Figure 9 and 10.

61
 4 6
8.0x10 4x10

Seismic Shear Force Response at the Section I

Without Water Without Water

Seismic Moment Response at the Section I

6.0x10
4
With Water 3x10
6 With Water

4 6
4.0x10 2x10

4 6
2.0x10 1x10

0.0 0

4 6
-2.0x10 -1x10

4 6
-4.0x10 -2x10

4 6
-6.0x10 -3x10

4 6
-8.0x10 -4x10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time (sec) Time (sec)

Figure 9. Seismic Responses at the Bottom of the Tower for Tower Model.

4 6
Seismic Shear Force Response at the Section III

1.8x10 2.5x10
Without Water

Seismic Moment Response at the Section III

4 6
1.5x10 With Water 2.0x10
4 6
1.2x10 1.5x10
3 6
9.0x10 1.0x10
3 5
6.0x10 5.0x10

3.0x10
3
0.0
5
0.0 -5.0x10
3 6
-3.0x10 -1.0x10
3 6
-6.0x10 -1.5x10
3 6
-9.0x10 -2.0x10
4 6 Without Water
-1.2x10 -2.5x10
With Water
4 6
-1.5x10 -3.0x10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time (sec) Time (sec)

Figure 10. Seismic Responses at the Height of Free Water Surface for Tower Model.

WATER EFFECTS ON BRIDGES WITH GROUP PILES

In China, group piles foundations are widely used in bridge construction, from short span
bridges to long span bridges, an example is shown in Figure 9. In this section, Pingtan Bridge[19] is
taken as an example to investigate water dynamic effects on the seismic response of bridges with
group pile foundations.
Five cases are considered in computation : (1) without water and without the mass of
superstructure; (2) with water below the pile-cap and without the mass of superstructure; (3) with
water above the pile-cap and without the mass of superstructure; (4) without water and with the
mass of superstructure; (5) with water above pile-cap and with the mass of superstructure.
The effect of hydrodynamic pressure on the free vibrational frequencies is shown in
Figure 10; the seismic responses are shown in Figure11-16.

62
 It can be seen from Figure 10 that the added mass makes important changes of the free
vibrational frequencies of mode 6, 7, 8(the values of frequencies are about 25-35rad/s, i.e.,
4.0-5.5Hz). That means the seismic response might be largely changed by the added mass if the
dominate spectral area of the seismic excitation is at this frequency phase.

Figure 11. The Pingtan Bridge Over Pingtan Sea Strait.

It can be observed from Figure 11-18 that the effect of the added mass(hydrodynamic
pressure) on the seismic response of the bridge depends on several factors: (1) the total mass of the
bridge; (2) the added mass; (3) the response that is concerned. To the Pingtan Bridge, the effect of
the hydrodynamic pressure on the seismic response the bridge show the following features:
the added mass enlarges the seismic responses at the first part of time, and then reduce the
responses, either for the response displacements, or for the internal force responses.
The effect of added mass on the response of piles is larger than that of the pier since the
added mass is attached to the piles.
The effect of seismic hydrodynamic pressure of the seismic responses of the bridge is
significantly different if or not the pile cap is below the free surface of the water.

63
 35 35

without water without water

30 with water below cap 30 with water below cap
with water above cap
25 25
Frequency (rad/s)

Frequency (rad/s)
20 20

15 15

10 10

5 5

0 0
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Mode Number Mode Number

Figure 12. Frequencies of Free Vibration of the Analyzed Bridge.

0.4 0.4

0.3 0.3
Disp. Response at the Top of the Pier

Disp. Response at the Top of the Pier

0.2 0.2

0.1 0.1

0.0 0.0

-0.1 -0.1

-0.2 -0.2

-0.3 -0.3
Without Water Without Water
-0.4 With Water Above the Cap -0.4 With Water Above the Cap

-0.5 -0.5
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Time (sec) Time (sec)

Figure 13. Seismic Displacement Response At the Top of The Pier.

0.15 without water 0.10
with water below the cap without water
Disp. Response at the Mid of the Pile

with water above the cap with water above the cap
Disp. Response at the Mid of the Pile

0.10

0.05

0.05

0.00 0.00

-0.05

-0.05

-0.10

-0.15 -0.10
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Time (sec) Time (sec)

Figure 14. Seismic Displacement Response At the Mid of The Pile.

64
 Bending Moment Response at the Bottom of the Pile(KN-m)

Bending Moment Response at the Bottom of the Pile(KN-m)

Without Water
4
With Water Below the Cap 4
1.5x10 With Water Above the Cap 1.2x10

1.0x10
4
Without Water
4 With Water Above the Cap
1.0x10 8.0x10
3

3
6.0x10
3
5.0x10 3
4.0x10
3
2.0x10
0.0
0.0
3
3 -2.0x10
-5.0x10
3
-4.0x10
3
-1.0x10
4 -6.0x10
3
-8.0x10
4 4
-1.5x10 -1.0x10
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Time (sec) Time (sec)

Without the Mass of Superstructure With the Mass of Superstructure

Figure 15. Seismic Bending Moment Response At the Bottom of The Pile.
Without Water

Shear Force Response at the Bottom of the Pile(KN)

Shear Force Response at the Bottom of the Pile(KN)

With Water Below the Cap

900 With Water Above the Cap 600

400
600

200
300

0
0
-200

-300
-400

-600
-600 Without Water
With Water Above the Cap
-900 -800
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Time (sec) Time (sec)

Without the Mass of Superstructure With the Mass of Superstructure

Figure 16. Seismic Shear Force Response At the Bottom of The Pile.
Bending Moment Response at the Bottom of the Pier(KN-m)

Bending Moment Response at the Bottom of the Pier(KN-m)

Without Water
6.0x10
3 With Water Below the Cap 2.0x10
4

With Water Above the Cap

4
3 1.5x10
4.0x10
4
1.0x10
3
2.0x10
3
5.0x10

0.0
0.0

3
-2.0x10
3 -5.0x10

4
-1.0x10
3
-4.0x10
4
-1.5x10
3
-6.0x10 4
-2.0x10 Without Water
3
With Water Above the Cap
4
-8.0x10 -2.5x10
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Time (sec) Time (sec)

Without the Mass of Superstructure With the Mass of Superstructure

Figure 17. Seismic Bending Moment Response At the Bottom of The Pier.

65
 Shear Force Response at the Bottom of the Pier(KN)
Shear Force Response at the Bottom of the Pier(KN)
Without Water 800
400
With Water Below the Cap
With Water Above the Cap 600
300

400
200

200
100
0
0
-200
-100
-400

-200
-600

-300 -800 Without Water

With Water Above the Cap
-400 -1000
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Time (sec) Time (sec)

Without the Mass of Superstructure With the Mass of Superstructure

Figure 18. Seismic Shear Force Response At the Bottom of The Pier.

CONCLUSIONS

The planning background of sea strait bridges remind engineers paying close attention to the
aseismic design of bridges in deep water. But so far engineers have a lack of knowledge about this
problem. In this paper a preliminary study of the water dynamic effects on seismic responses of
bridges in deep water has been carried out, and the following conclusions have been reached.
The hydrodynamic pressure may have important effect on the seismic response of the
bridges in deep water, research in details on this effect should be a primary part for
aseismic designs of sea strait bridges.
The effect of hydrodynamic pressure on the seismic response of the bridge depends on
several factors, the total mass of the bridge, the added mass, the response that is concerned.
The hydrodynamic pressure may changes the seismic responses of Honshu-Hokkaido
Strait Bridge by 30%-100%, and more than 15% for Pingtan Bridge.
Finally, it should be mentioned that the calculated results presented in this paper are
preliminary, it assumed that the Morison equation is suitable for the cases studied, and the
hydrodynamic damping is neglected. Because the importance of hydrodynamic effects on the
response of bridges suggested by the numerical analysis in this paper, to verify the theoretical
method and the conclusions from the numerical computation, a shaking-table test is under
preparation and is going to be finished in by the end of 2002.

ACKNOWLEDGEMENT

This research is supported by the National Science Foundation of China at granted No.
40072088.

66
REFERENCES

H. Goto and K. Toki. 1965. “Vibration Characteristics and Aseismic Design of Submerged Bridge Piers,” Proc. Third
World Conf. On Earthq. Eng, Auckland, Welligton, New Zealand.
S.Kotsubo. 1965. “Seismic force effect on submerged bridge piers with elliptic cross-section,” Proc. Third World Conf.
On Earthq. Eng., Auckland, Welligton, New Zealand.
C.Y. Liaw and A.K. Chopra. 1973. “Earthquake response of axisymmetric tower structures surrounded by water,”
Report No. UCB/EERC 73/25, EERC, University of Califirnia, Berkeley, CA.
C.Y. Liaw and A.K. Chopra. 1974. “Dynamics of towers surrounded by water,” Earthquake Engineering & Structural
Dynamics, Vol.3, No.1, 33-49.
D.Rea, C.Y. Liaw and A.K. Chopra. 1975. “Dynamic properties of San Bernardino intake tower,” Report No.
UCB/EERC 75/05, EERC, University of Califirnia, Berkeley, CA.
R.C.Byrd. 1978. “A laboratory study of the fluid-structure interaction of submerged tanks and caissons in earthquakes,”
Report No. UCB/EERC 78/08, EERC, University of Califirnia, Berkeley, CA.
F.Nilrat. 1980. “Hydrodynamic pressure and added mass for axismmetric bodies,” Report No. UCB/EERC 80/12,
EERC, University of Califirnia, Berkeley, CA.
Tanaka Y. And Hudspeth R.T. 1986. “Earthq. Response of Circular Cylindricl Structures in Water,” 5th International
Symp. on Offshore Mech. And Arctic Eng., Tokyo, Japan.
Anthony N. Williams. 1986. “Earthq. Response of Submerged Circular Cylinders,” Ocean Eng., 13(6).
Tanaka Y. And Hudspeth R.T. 1988. “Restoring forces on vertical circular cylinders forced by earthquakes,” Earthquake
Engineering & Structural Dynamics, Vol.16, No.1, pages 99-119.
Goyal, A.; Chopra, A. K. 1989. “Earthquake response spectrum analysis of intake-outlet towers,” Journal of Engineering
Mechanics, 115, 7, July, pages 1413-1433.
Goyal, A.; Chopra, A. K. 1989. “Simplified evaluation of added hydrodynamic mass for intake towers,” Journal of
Engineering Mechanics, 115, 7, July, pages 1393-1412.1`
Chopra, A. K.; Goyal, A. 1991. “Simplified earthquake analysis of intake-outlet towers,” Journal of Structural
Engineering, 117, 3, Mar, pages 767-788.
El Naggar, M. H.; Novak, M. 1996. “Influence of foundation nonlinearity on offshore towers response,” Journal of
Geotechnical Engineering, 122, 9, Sept., pages 717-724.
Spyrakos, C. C.; Xu, C. 1997. “Soil-structure-water interaction of intake-outlet towers allowed to uplift,” Soil Dynamics
and Earthquake Engineering, 16, 2, Feb, pages 151-159.
Sun, K.; Nogami, T. 1991. “Earthquake induced hydrodynamic pressure on axisymmetric offshore structures,”
Earthquake Engineering & Structural Dynamics, 20, 5, May, pages 429-440.
Yamada, Y.; et al. 1989. “Seismic response of offshore structures in random seas,” Earthquake Engineering & Structural
Dynamics, 18, 7, Oct, pages 965-981.
Bridge and Offshore Engineering Association, Sea Strait Crossing, Vol.13, 47-53, 1998.
Fujian Highway Design Institute, The Pre-feasibility Study of Pingtan Bridge, 1999.

67
 Observed Pile and Pipeline Performance in the Full-Scale
Lateral Spread Experiment
Scott A. Ashford1and Teerawut Juirnarongrit2

ABSTRACT

Two full-scale experiments using controlled blasting were conducted in the Port of Tokachi on
Hokkaido Island, Japan, to assess the performance of lifeline facilities subjected to lateral spreading.
Lifeline specimens in this study included a single pile, a 4-pile group, a 9-pile group, two natural gas
pipelines, and one electrical conduit. All of them were extensively instrumented with strain gauges to
measure the distribution of moment during lateral spreading. This allowed to compute the loading
condition, as well as to conduct the damage and performance assessments on the lifeline facilities. Other
instrumentation including pore pressure transducers, GPS units, and slope inclinometers, were also
installed to measure the degree of liquefaction as well as the movements of soil and lifelines. This paper
presents the test results and provides some discussions on the performance of piles and pipelines observed
from the experiments.

Scott A. Ashford, Associate Professor, Department of Structural Engrg., Univ. of California, San Diego, La Jolla,
CA 92093-0085
Teerawut Juirnarongrit, Graduate Student Researcher, Department of Structural Engrg., Univ. of California, San
Diego, La Jolla, CA 92093-0085.

69
INTRODUCTION

Lateral spreading, which usually refers to global displacements of gently sloping ground due to
liquefaction, is one of the primary earthquake hazards. In past earthquakes, lateral spreading has caused
considerable damage to civil infrastructure including port facilities, buildings, bridges, and utilities. Good
examples are the damage of quay walls and buildings in the 1995 Kobe earthquake; the damage of pile
foundations in the 1964 Niigata earthquake; the damage of over 250 bridges and numerous embankments
along the Alaskan Railroad and Highway during the 1964 earthquake; the damages of numerous water
and gas lines in the 1906 earthquake; and the significant damage in the San Francisco area in 1989 Loma
Prieta earthquake (Bartlett and Youd 1992b; Seed 1987; Youd and Hoose 1976; Bardet and Kapuskar
1993; Clough et al. 1994; and O’ Rourke and Pease 1992). Therefore, it is extremely essential to
understand the behavior of soil as well as structures during lateral spreading in order to improve the
current design method for structures and lifeline utilities to prevent the catastrophic failure for future
earthquakes. Meanwhile, most lateral spreading research to date has focused on small-scale centrifuge
studies (e.g. Abdoun et al. 1996), limited area 1-g shake table tests (e.g. Tokida et al. 1993), or case
histories (e.g. Hamada and O’Rourke 1992; O’Rourke 1996). In addition, some full-scale has been
carried out to study the behavior of deep foundations in sand liquefied by controlled blasting (e.g.
Ashford et al. 2000), but these tests do not account for the global translations of the lateral spreading soil
mass. In light of this, the full-scale instrumented lifeline components in controlled lateral spreading tests
were carried out in order to understand the performance of lifelines and be able to implement the test
results in engineering practice. The test results will be a valuable source of data for further development
of the empirical methods and/or complex numeral models to use to design lifeline facilities subjected to
lateral spreading.
Two full-scale experiments using controlled blasting were conducted in November and December
2001 in the Port of Tokachi on Hokkaido Island, Japan, to study the performance of lifeline facilities
subjected to lateral spreading. This research project was the joint collaboration between the University of
California San Diego (UCSD) and several Japanese organizations. This overall research effort was lead
by the Port and Airport Research Institute (PARI). The primary objective of the test was to assess the
performance of quay walls subjected to lateral spreading using controlled blasting. One quay wall was of
traditional design and new seismic design criteria was applied to the other. Since the test area was so
large, it enabled researchers to include additional experiments in the zone of liquefaction and lateral
spreading without interfering with the primary objective of the quay wall test. The University of
California, San Diego, together with Waseda University (WU) collaborated with other Japanese
researchers to install the lifeline specimens in the zone of lateral spreading through the PEER Lifelines
Program with support from Caltrans, Pacific Gas & Electric and the California Energy Commission.
In all, UCSD installed 6 test specimens. The pile specimens in the experiment program consisted
of a single pile, a 4-pile group, and a 9-pile group. In addition, two natural gas pipelines and one
electrical conduit were installed. The objectives of this study is to conduct damage and performance
assessments of those lifelines subjected to lateral spreading, as well as to evaluate loading conditions on
the structures during lateral spreading.

SITE CHARACTERIZATION

The test site was a recent man-made land that was completed just a few years ago. The land was built
by hydraulically placing fill without any ground improvement; therefore, the soil was very loose and
highly susceptible to liquefaction.
A subsurface soil exploration program was carried out in many areas throughout the test site to
characterize the soil condition. Generally, the soil condition consisted of 7.5 m of hydraulic fill underlain
by 1 m of medium dense sand overlying a very dense gravel layer as presented in Figure 1. The hydraulic

70
fill was comprised of a 4-m layer of very loose silty sand with uncorrected SPT-N values ranging from 1
to 5. This was underlain by a 3.5-m layer of very soft lean to fat clay with sand. Uncorrected SPT blow
counts ranged from 0 to 2 blows per foot in this layer. The water table was approximately 1 m below the
ground surface. Figure 2 presents the grain size distribution of the hydraulic fill plotted together with the
Japanese standard curves for liquefaction potential evaluation. The first 4 m of the soil fell into a zone of
highly susceptible to liquefaction. Below this layer, fine contents increased with depth. Only a thin layer
of soil at depths between 7.0 and 7.5 m was not liquefiable. Based on the results of grain size analysis
and the soil strength characteristic, the soil at the test site was highly susceptible to liquefaction, and
therefore appropriate for conducting the full-scale lateral spreading test.

SITE DESCRIPTION AND TEST SETUP

The UCSD experiments were located in a zone of the unapplied seismic design quay wall where
the large global translation of the soil was expected. A layout of the test site for the first experiment is
shown in Figure 3. The test site was approximately 25 m wide by 100 m long. The front face was
bordered by a water way. The water elevation was approximately +2.00 m on the test day. The sheet pile
quay wall was driven to the elevation of -8.00 m and was anchored by the tied rods which were fixed to
H-piles to prevent the movement of the quay wall. The quay wall retained approximately 7.5 m of
hydraulic fill. The ground surface started to gently elevate upwards at 25.2 m away from the quay wall
with the embankment slope of 4%. The test site was surrounded by the sheet piles to tip elevations
between -5.00 and -8.00 m.
The UCSD pile specimens were located 19.0 m away from the quay wall. The pile specimens
consisted of a single pile, a 4-pile group, and a 9-pile group. A group of free head single piles of WU
were also located in this region. The pile diameters were 318 mm with wall thickness of 10.5 mm, and a
nominal length of 11.5 m. The yield strength of these steel pipe piles was 400 MPa.

~ 1.0 m
Gravel Sand Silt Clay
100
Very Loose Silty SAND ~ 4.0 m
SPT ~ 1 to 5 Susceptible
to Liquefaction
80
Hydraulic
Highly Susceptible
Fill
Percent Finer (%)

to Liquefaction
60
Very Soft CLAY
SPT ~ 0 to 2 ~ 3.5 m
Depth from Design
40 Ground Surface(El. +3.00)
0.93~1.23m
2.93~3.38m
Medium Dense SAND 3.93~4.38m
~ 1.0 m 20
SPT~20 4.93~5.38m
5.93~6.23m
6.93~7.23m

Very Dense GRAVEL Natural Soil 0

SPT>50 100 10 1 0.1 0.01 0.001
Grain Size (mm)

Figure 1. Typical Soil Profile of Test Site. Figure 2. Grain Size Distribution of Soil at Test
Site.

71
 El -0.50

El -4.50

El -8.00

Figure 3. Site Layout of 1st Lateral Spreading Experiment.

In addition, two natural gas pipelines and one electrical conduit were installed. The gas pipeline
consisted of a 500 mm diameter pipe with wall thickness of 6 mm and yield strength of 400 MPa. The
electrical conduit consisted of a 268 mm diameter with wall thickness of 6 mm and yield strength of 400
MPa. Both pipelines were about 25 m long and located across the test sites at 30 m and 32.2 m away
from the quay wall. The bottoms of both pipelines were installed at the elevation of +1.75 m. The other
gas pipeline was 22 m long and installed parallel to the direction of the flow. The center of the pipeline
was 1 m below the ground surface along its entire length.
Due to the success in using the controlled blast to induce liquefaction of the soil in several tests in
Japan as well as the full-scale lateral load tests at Treasure Island (Ashford et al. 2000), the same
technique was implemented to liquefy the soil at the test site, and thus induce lateral spreading. The blast
holes were spaced at 6.0 m on centers in a regular grid pattern. The charges were installed at depths of
3.5 m and 7.5 m below the ground surface. The amount of charges varied from 2 kg nearby the pile
specimens to 3-5 kg at other areas. It was done this way so as to prevent damage to a large number of
instruments installed in the vicinity of pile specimens. The first experiment was carried out on November
13, 2001. The sequence of the blasting started from the back corner of the embankment and then
continued to the next holes of the same rows and proceeded successively towards the quay wall. This
was followed by the detonation of the secondary blast holes around the perimeter of the test site. The
purpose of these explosives was to loosen the soil in the vicinity of the sheet pile to allow unrestricted
flow of the soil in such region. Approximately 20 seconds after the completion of the secondary blasting,
the additional explosives were used to break the tie rods of the quay wall and allowed the quay wall to
move freely to create additional movement of the soil within the test area.

72
 The second lateral spreading test was performed with an attempt to induce additional ground
deformations and further evaluate the performance of lifeline facilities subjected to a higher level of soil
deformation. The test site for the second lateral spreading test was significantly modified from the first
one as presented in Figure 4. The test site was approximately 30 m wide by 40 m long. The quay wall
and sheet piles surrounding the test site were removed to allow the soil to move freely. The waterway was
excavated on one end of the test site to an elevation of -1.00 m with the slope of 1:2 and then filled with
water to an elevation of +2.00 m. The ground surface was leveled for a distance of 7.5 m away from the
edge of the waterway and then started to rise up with an embankment slope of 6% over a distance of 18.0
m. The blast holes were spaced at 6.0 m on centers in a regular grid pattern. Charges were installed at
depths of 4.0 m and 8.0 m below the design ground surface (El +3.00m). The amount of charges varied
from 2 kg to 4 kg. Two additional rows of blast holes were drilled. One was located on the steep slope
adjacent to the waterway with the amount of explosives ranging from 1 to 3 kg. The purpose of these
explosives was to loosen the soil at the slope toe prior to the primary blasting sequence such that the
embankment soil behind it had a high potential to move freely with larger deformation once the primary
blasting initiated. The other was located between the pipelines and piles as denoted as blast holes No. 7
to No. 9. Three kilograms of explosives were installed at El. -3.00 m.
The weather condition for the second lateral spreading experiment was poor as presented in
Figure 5 due to a heavy snowfall with a snow thickness of about 0.50 m and a new record of wind speed
of 100 kph on the test day. The ground was frozen throughout the test site which would likely impede the
global translation of the soil mass. In an attempt to mitigate this, jackhammers were used to break up the
frozen ground in the vicinity of test specimens to depths of approximately 20 to 30 cm below the ground
surface as presented in Figure 6. The second test was carried out on December 14, 2001. The explosives
on the steep slopes were detonated initially from S1 to S5. Approximately 15 second later, the primary
sequence of the blasting was started. The primary blast began at blast hole No.1 on the back of the
embankment. Then, the blasting proceeded to the next holes of the same rows, and then continued to the
next row towards the waterway (i.e., from No.1 to No.17).

INSTRUMENTATION

Piles and pipelines were extensively instrumented with electrical strain gauges. The strain gauges
of pile specimens were located at 0.6 m intervals on both upstream and downstream sides of the piles to
measure the bending moment along the length of the pile. A series of tiltmeters at various depths were
also installed on one pile of each foundation system to use as backup data for strain gauges.
Unfortunately, all of them were damaged during the pile installation. The 75x40x5 steel channels with
yield strengths of 400 MPa were welded to the steel pipe piles to protect the strain gauges from damage
during the pile installation. The strain gauges of the gas pipeline were spaced between 1.0 m and 3.0 m
along the top and the side of the pipelines to measure the bending moment along the pipelines in both
vertical and horizontal directions, respectively.
Apart from the strain gauges, other instrumentation was also installed to capture behaviors of soil
and lifelines in more details. These include pore pressure transducers, soil pressure cells, string-activated
linear potentiometers, accelerometers, slope inclinometer casings, and Global Positioning System (GPS)
units. A layout of instrumentation for the first experiment is presented in Figure 7. The instrumentation
for the second experiment was essentially the same as the first test; therefore, it is not shown in this paper.

73
 S1
1:2
14 13 6

A S2 7 A'
15 12 5 1

S3 8

16 11 4 2

S4 9
1:2
17 10 3

S5 Hinge point
Figure 5. Poor Weather Condition during 2nd
Lateral Spreading Experiment.
16.0 m
17.0 m [email protected] m = 18.0 m
20.0 m
23.5 m
31.2 m
33.2 m
41.5 m

3.0 m 7.5 m 18.0 m

EL +3.00 1.5 m 6%
EL +3.00
EL +2.00

EL -1.00 EL -1.00

EL -3.00

EL -5.00
o
30

Scale (m)
Section A-A'
0 10 20

Figure 4. Site Layout of 2nd Lateral Spreading Figure 6. Breaking up Frozen Ground Surface
Experiment. Using Jack Hammer.

, ,G G

PPT-9B G

1C
G G
G G
G

G
G

Inclinometer Casing G GPS Unit

Pore Pressure Transducer Tiltmeter
Potentiometer Accelerometer

Figure 7. Instrumentation Plan for 1st Lateral Spreading Test.

74
TEST RESULTS

Excess Pore Water Pressure

An example of excess pore water pressure ratio time-history nearby the 9-pile group at depth of 2
m below the ground surface is presented in Figure 8. The excess pore water pressure ratios built up
immediately after the blast though this transducer was located about 50 m away from the first blast hole.
The rate of increase in pore water pressure became more rapid as the blast moved closer to the transducers.
The increase in pore water pressure ratios proceeded to reach the maximum values at approximately 30
seconds. Fluctuation of pore pressure ratios was obvious as the blasting occurred in the vicinity of the
transducer location. The results show that the soil in the vicinity of the 9-pile group was liquefied with
the maximum excess pore pressure ratios exceeding 100%. The ratios dropped to about 80% after the
blast stopped, then proceeded to dissipate with time. The evidence of increase in excess pore water
pressure ratio at times of about 40s and 86s was due to the effect of the secondary blasting of tied rods,
respectively. The characteristics of excess pore water pressure ratios in other locations were basically the
same as the one presented herein. The excess pore water pressure ratios throughout the entire test site
exceeded 70%. Some of them were slightly over 100%. Sand boil was observed following the blasting
as presented in Figure 9 confirming that the liquefaction had occurred.
The excess pore pressure ratios in the 2nd test appeared to be much less than those measured
during the first test with values ranging between 30% and 80%. No sand boil was observed in the 2nd test.
Two possible reasons can be explained regarding the lower excess pore pressure ratios. First, the soil was
less susceptible to liquefaction because some settlement took place after the first experiment and caused
the soil to become denser. Second, the frozen ground decreased the liquefaction potential due to terrible
weather in the second test.

150

PPT-9B
Depth 2.00 m

100
Ru (%)

50

0
0 20 40 60 80 100
Time (s)

Figure 8. Excess Pore Pressure Ratio vs. Time Figure 9. Sand Boil after the 1st Experiment.
nearby 9-Pile Group.

75
Deformations of Ground and Lifelines

The GPS units were used to monitor the movements of both ground and lifeline facilities during
lateral spreading. The measurements were conducted by a research team from the California Department
of Transportation, Caltrans (Turner 2002). An example of time history of soil movements on the
downstream side of the gas pipeline (denoted as unit 1C) in longitudinal, transverse, and vertical
directions is presented in Figure 10a. The movements of GPS units were observed at about 10 seconds
after blasting initiated. As the blasting moved closer to the GPS location, more movements in all
directions were observed. The lateral movements between 10 seconds and 27 seconds were due to not
only the liquefaction-induced lateral spreading but also the dynamic forces generated by the blasting.
With the blasting past the location of GPS units (at about 27 seconds), the effect of dynamic forces from
the blasting was not important as indicated by the insignificant movements in transverse and vertical
directions. The longitudinal movement observed after 27 seconds was therefore primarily due to
liquefaction-induced lateral spreading. Figure 10b presents the displacement path of GPS unit in the
horizontal plane of the gravel surface showing that the horizontal movement mainly occurred in the
longitudinal direction towards the quay wall.
The vector displacements in the horizontal plane throughout the test site for the first test are
presented in Figure 11a. The largest horizontal displacement was about 43 cm occurring at the pile head
of the WU pile. The UCSD single pile moved only 32 cm, which was significantly less than the WU piles.
This was likely due to the fact that the WU piles were shorter in length and the pile tips were located just
above the dense layer; while the UCSD pile was penetrated about 3.5 meters into the dense soils. The
WU piles were therefore likely behaved as rigid piles, in which the rotation and movement at the pile tip
were expected. In contrast, the UCSD pile acted as a flexible pile where the rotation and the movement at
the pile tip was insignificant. As a result, the displacement at the pile head of the UCSD single pile was
less than those of the WU piles. The 4-pile group and the 9-pile group moved 21cm and 18 cm,
respectively. The data from the GPS units in the vicinity of the pipelines show that the movements of the
gas pipeline and electrical conduit were similar with a magnitude of about 38 cm. The average of soil
movement was about 35 to 40 cm.
As presented in Figure 11b, the horizontal movements occurred in the second test were generally
lower than those occurred in the first test, especially in the vicinity of the pipelines. This was mainly due
to the weather condition that decreased the liquefaction potential and thus impeded the global translation.
The movement of the gas pipeline was about 50% of that occurred in the first test. The average soil
movement in the second test ranged from 10 cm to 23 cm. One GPS unit installed between two pile caps
showed the soil movements as high as 45 cm. However, 10 cm of 45 cm attributed to the movement of
slope toe due to the effect of initial blasting along the slope edge. The movement of pile groups in the 2nd
test ranged from 16 cm to 18 cm, slightly less than that in the first test. The movements of WU and
UCSD single piles at the ground surface were 39 cm and 28 cm, respectively.

76
 0.60
Relative Displacement (m) (a) Displacement Time-History GPS-1C

0.40
(a) 1st Experiment (b) 2 nd Experiment
Longitudinal
0.20 Transverse
Vertical

0.00

-0.20
0 20 40 60
Time (sec)
0.60
(b) Displacement Path Downstream
0.40
Transverse (m)

0.20
Initial Position
0.00
0 0.5 1.0
-0.20
Vector Displacement Scale (m)
-0.40

-0.60
-0.20 0.00 0.20 0.40 0.60
Longitudinal (m)

Figure 10. GPS Data (a) Displacement Time- Figure 11. Vector Displacements from GPS
History, and (b) Displacement Path Data for 1st and 2nd Experiments
(after Turner 2002). (after Turner 2002).

Moment Distribution

Moment distribution along the length of the single pile at the end of the test is presented in
Figure 12. The test results indicate that the moments at depths between 0m and 4m were insignificant.
One possible explanation for this phenomenon is that after the soil was liquefied, it comes to behave like
a viscous fluid material, being able to flow around the pile without significant force acting on the pile.
The soil resistance began to increase with depth for the next 3.5 m where a very soft clay layer existed.
The maximum moment occurred in a dense soil layer at a depth of about 9 m below the ground surface.
The pile was yielded after the second test.
Figure 13 presents the moment distribution of pile No.5 in the 9-pile group. The shape of moment
profile from the experiment agreed well with a typical analysis of a pile with fixed head condition
showing that the results from the test were reasonable and appropriate for further analysis to estimate the
loading distribution of liquefiable soil on the pile. Figure 14 presents the moment profile of each pile in
the group after the first experiment. The moment distribution of all piles in the 9-pile group was more or
less similar, except for pile No. 2 and No.4 where the moments were smaller than the others. This is likely
due to the fact that both piles were shorter in length, and had a smaller degree of fixity into the dense soil
layer, resulting in a smaller moment in the piles. It is noted that pile No. 2 and No.4 reached refusal
during the pile installation, likely due to the presence of a boulder at that particular depth. The similarity
of moment distribution of each pile in the group indicates that a shadowing effect was unimportant in
liquefied soil. This conclusion was similar to that of a recent research on the behavior of pile group in
liquefied soil conducted at Treasure Island (Ashford and Rollins 2002).

77
 After the second test, all piles in the 4-pile group and the 9-pile group remained elastic with the
maximum moment below 70% and 60% of yield moment, respectively. No structural damage was
observed on piles to pile cap connections on both pile groups though both pile groups experienced the
total movements of nearly 40 cm. In addition, based on the strain gauge data of pipelines (not presented
in this paper), all pipelines performed comparatively well without any yielding.

Back Front
Instrumented Piles
6 5 4

Moment after 1st Test Flow Direction

Moment after 2nd Test 3 2 1

Moment (kN-m) Moment from 1st Test

-600 -400 -200 0 200 400 600 Additional Moment from 2nd Test
El +3.000
0 Combined Moment (1st Test+2nd Test)

El +2.000 9-Pile Group

Moment (kN-m)
-400 -200 0 200 400

Yield Moment
V. Loose El +3.000
SAND
2 0
Pile Cap

El -0.800 El +2.000
V. Loose
4 SAND
2
Depth (m)

9 V. Soft
CLAY El -0.800
6 4

Depth (m)
El -4.300 6 5 4 V. Soft
CLAY
Med. Dense
SAND 8 6
El -5.300

El -4.300
V. Dense Med. Dense
GRAVEL SAND 8
El -5.300
10
V. Dense
GRAVEL
10
12

12

Figure 12. Moment along Single Pile. Figure 13. Moment along Pile No.5 of 9-Pile
Group.

78
 Moment (kN-m) Moment (kN-m)
-200 -100 0 100 200 -200 -100 0 100 200
0 0

2 2

4 4

Depth (m)
Depth (m)
6 6

8 8

10 10
No.1 No.4
No.2 No.5
No.3 No.6

12 12

Figure 14. Moment Distribution of Each Pile in the 9-Pile Group (Data Extracted from 1st Experiment).

CONCLUSIONS

Based on the results obtained from two full-scale experiments, the following conclusions can be
obtained:
1. Controlled blasting successfully liquefied the soil and induced lateral spreading.
2. The excess pore water pressure ratios exceeded 70% for the first experiment. The degree of
liquefaction in the second experiment was much lower than the first one with excess pore pressure
ratios ranging between 30% and 80%, likely due to the weather condition.
3. The average soil movements of the first experiment were about 35 cm to 40 cm, while about 9 cm to
22 cm of ground movements were observed in the second test.
4. The total movements of the single pile, 4-pile group, and 9-pile group were 58 cm, 39 cm, and 34 cm,
respectively.
5. The total movements at the middle of the gas pipeline and electrical conduit were about 54 cm.
6. Shadowing effect of the 9-pile group in liquefied soil was not observed.
7. The single pile yielded at the end of the second experiment. Both 4-pile and 9-pile groups performed
well during both experiments. Piles remained in elastic range with maximum moments of less than
70% of yield moment. No damaged was observed on piles to pile cap connection.
8. All pipelines also performed very well without any yielding.

79
ACKNOWLEDGEMENT

The authors wish to express their sincere gratitude to the sponsors of the Performance of Lifelines
Subjected to Lateral Spreading project, PEER Lifelines Program with support from Caltrans, Pacific Gas
& Electric and the California Energy Commission.
The assistance from Dr. Takehiro Sugano of Port and Airport Research Institute (PARI) and
Professor Masanori Hamada from Waseda University(WU), the original leaders of this project, in
providing a great opportunity for us to participate in this test, as well as effectively managing the project
is gratefully acknowledged. Loren Turner, Cliff Roblee and Tom Shantz from Caltrans’ Division of New
Technology and Research are gratefully acknowledged for their help in monitoring the movements of soil
and lifelines using GPS equipments.
We would like to thank Andrea Martinez and Jeremy Allen, the undergraduate students, for their
help in analyzing the data and preparing the figures in this paper.

REFERENCES

Abdoun, T. Dobry, R., and O-Rourke, T. D., 1997. “Centrifuge and numerical modeling of soil-pile interaction
during earthquake induced soil liquefaction and lateral spreading,” Observation and Modeling in Numerical
Analysis and Model Tests in Dynamic Soil-Structure Interaction Problems, Geotechnical Special
Publication No. 64, ASCE, New York, pp.64-70.
Ashford, S. A., Rollins, K. M., 2002. TILT: Treasure Island Liquefaction Test Final Report, Report No. SSRP-
2001/17, Department of Structural Engineering, UCSD.
Ashford, S. A., Rollins, K. M., Bradford, S. C., Weaver, T. J., and Baez, J. I., 2000. “Liquefaction mitigation using
stone columns around deep foundations: fill-scale test results,” Soil Mechanics 2000, Transportation
Research Record No. 1736, TRB, Washington D.C., pp.110-118.
Bardet, J. P., and Kapusker, M., 1993. “Liquefaction sand boils in San Francisco during 1989 Loma Prieta
earthquake.” J. Geotech. Engrg., ASCE, Vol. 119, No.3, March, pp. 543-562.
Bartlett, S. F. and Youd, T. L., 1992b. “Empirical analysis of horizontal ground displacement generated by
liquefaction-induced lateral spreads.” Tech. Rep. NCEER-92-0021, National Center for Earthquake
Engineering Research, Buffalo, N.Y., M. Hamada and T.D. O’Rourke (eds.), August 17.
Clough, G. W., Martin, J. R., II, and Chameau, J. L., 1994. “The geotechnical aspects.” Practical Lessons from the
Loma Prieta Earthquake, National Research Council, National Academy Press, Washington, D.C., pp. 29-
63.
Hamada, M., and O’Rourke, T., Editors., 1992. Case Studies of Liquefaction and Lifeline Performance During Past
Earthquakes, Report No. NCEER-92-0001, Vol.1, National Center for Earthquake Engineering Research,
Buffalo, N.Y.
O’Rouke, T. D., 1996. “Lessons learned for lifeline engineering from major urban earthquakes,” Proceedings,
Eleventh World Conference on Earthquake Engineering, Elsevier Science Ltd., 18 p.
O’Rouke, M. J., and Pease, J. W., 1992. “Large ground deformations and their effects on lifeline facilities: 1989
Loma Prieta earthquake.” Case Studies of Liquefaction and Lifeline Performance During Past Earthquakes,
Volume 2: United States Case Studies, Tech. Rep. NCEER-92-0002, M. Hamada and T.D. O’Rourke (eds.),
February 17, 85 pages.
Seed, H. B., 1987. “Design problems in soil liquefaction.” J. Geotech. Engrg., ASCE, Vol. 113, No. 8, August, pp.
827-845.
Tokida, K., Iwasaki, H., Matsumoto, H., and Hamasa, T., 1993. “Liquefaction potential and drag force acting on
piles in flowing soils,” Soil Dynamic and Earthquake Engineering, Computational Mechanics, South
Hampton, England, pp. 349-364.
Turner, L. L., 2002. Measurements of Lateral Spread Using a Real Time Kinematic Global Positioning System,
California Department of Transportation Division of New Technology &Research, April 2002, 94 pages..
Youd, T. L., and Hoose, S. N., 1976. “Liquefaction during 1906 San Francisco Earthquake.” J. Geotech. Engrg.
Div., ASCE, Vol. 102, No. GT5, May, pp. 425-439.

80
 Seismic Response of Railway Bridges

Considering TRACK Restriction

Guiping Yan1, Yan Huang1 and Guanyuan Zhao1

ABSTRACT

A simplified model is introduced to describe the dynamic interaction between track

and bridge spans. With the model a track-beam model is established and the nonlinear
dynamic FEM program is developed to analyze the effects of track restriction to the seismic
response of railway bridges. Through the numerical analysis, it is shown that the existence of
track is helpful for the piers with lower stiffness to resist earthquake, otherwise for the stiffer
piers it may be harmful. By considering the variant track resistance, number of bridge spans,
and pier stiffness distribution, the effects of track on the seismic response of railway bridges
are discussed.

_____________
1.
College of Civil Engineering and Architecture, Northern Jiaotong University, Beijing, 100044, PRC

81
INTRODUCTION

Bridges play an important role in the railway transportation system. With the use of
long rail track, a much stronger connection between bridge spans is provided. How to
consider the effects of long rail track on earthquake properties of bridges is still a problem to
be studied. The existence of stronger track restriction can cause the variation of the bridge
dynamic characteristics, change the internal force distribution, and therefore affect the
behavior of bridge during earthquakes. Some research works show that the existence of track
can provide an additional restriction to transfer the seismic loads to the roadbed. Thus, the
structural elements of the bridge can be partly relieved from the demand to carry all of the
seismic loads[1]-[3]. Maragakis has measured the frequencies and damping properties of the
Strawberry Park Railway Bridge, and investigated the effects of the rails on the dynamic
response of the bridge[1]. The experimental results shows that with the existence of track, the
vibration transferred to the adjacent roadbed is obvious. When the rails are cut, the
fundamental frequencies of bridges in different directions are decreased. The full-scale
experiments by Luo on the simply supported railway bridges also show that the track has the
considerable restriction on railway bridges along the longitudinal direction, but they had few
function in the transverse direction[3].
So far, no systematical research works have been done about the effects of track
restriction on the seismic response of railway bridges. The effects of track are usually
neglected in the seismic analysis of railway bridges; therefore the influence of the track on
the seismic response of bridges can not be exactly evaluated. CWRs (Continuously Welded
Rails) have been used on high-speed railway bridges in many countries. In China, CWRs are
also used on some long bridges. With the wide use of CWRs on railway bridges, the effects
of track on the seismic response of railway bridges should be further studied.
In this paper, a simplified method of describing the interaction between rails and
bridge is developed, in which the beam, rails and roadbed are regarded as a hybrid
beam-track element. The beam-track element is consisted of segments of beam-track element.
The elasto-plastic relation in the roadbed for each segment is determined according to the real
states. The DOF reduction method is used to form the stiffness matrix of the hybrid
beam-track element. Based on this, the earthquake response is solved.

ANALYSIS MODEL

Basic Assumption

Rail Roadbed

Girder
200m 200m
Pier
1 2 3 5 6
4

Figure 1 Structural model

82
 M

K2
My 1

K1

θy θ
Figure 2 Bilinear model of piers

For the structural system shown in Figure 1, the motion equation of the structure can
be expressed as:
M&x& + Cx& + Kx = MI {&x&g } (1)

where M , K , and C are the mass, stiffness, and damping matrixes, &x&, x&, x

and &x&g are the vectors of acceleration, velocity, displacement, and ground acceleration,

respectively.
Rails and beams are modeled as linear beam elements and the nonlinear properties are
considered about the interaction between track and beams. The non-linear characteristics of
piers are defined as the bilinear institute relation as shown in Figure 2. The tracks outside the
bridge will share the deformation and longitudinal forces. In the analysis, the length of the
rails outside the bridge is selected as 200m.

Nonlinear Properties of Piers

Following conditions are assumed in the analysis. (1) The pier bottom is fixed to the
foundation and the interaction between the foundation and soil is ignored. (2) The relationship
between the moment ( m ) at pier bottom and the rotation angle at pier top ( θ ) is given by:
3EI
m= ⋅θ (2)
l
where EI /l is the linear stiffness of piers. The yielding moment is defined as the moment that the
outside reinforcements of pier section begin to yield. The pier stiffness after yield ( k 2 ) is

assumed as 10% of the initial one ( k1 ) .

83
Interaction Between Track and Beams

Under an intense earthquake, the relative displacements between beam and track will
occur. Because of the roadbed resistance, the longitudinal forces will be applied to the rails and
beams. The relationship between relative deformation and longitudinal forces are shown in
Figure 3.

Force endured by rails 0

Tension
Dis. of rails
Press
Dis. of bridge
Force endured by bridge y

(a) Longitudinal forces (b) Deformation

Figure 3 Deformation and longitudinal forces in the rails and bridge

In order to simulate the interaction correctly, an elasto-plastic model is used to
represent the force-displacement relationship of the roadbed, as shown in Figure 4[4][5].
A series of nonlinear springs are used to model the function of track resistance. Only
when the space between two adjacent spring elements is small enough the model can have
enough accuracy. Therefore the beam, rails, and roadbed are divided into many small
segments and nonlinear spring elements. To simplify the data preparation and input works, all
sub-elements within one span are combined to form a hybrid beam-track element. The
stiffness matrix and its nonlinear characteristics of the hybrid element is calculate by a
subroutine and formed by DOF reduced technique.

r(KN/m)

u(mm)
2mm

Figure 4 Model of roadbed

The mesh of beam-track element is shown in Fig5, where I, J, K, L are the nodes of the
hybrid element. The interaction between the roadbed and the rail is modeled in a similar way.

84
 raRail
1 3 n-1 J
I
2 4 n
K L
Girder

Figure 5 Beam-track element

DEDUCTION OF THE STIFFNESS MATRIX OF THE BEAM-TRACK ELEMENT

Distributed Resistance of Roadbed at Linear Stage

U1 Rail U2

F1 F2
Roadbed resistance ku
U3 U4
F3
F4
Beam

Figure 6 The beam-track element

In Figure 6, u1 , u2 , F1 , F2 are the displacements and forces at rail ends, u3 , u4 , F3 , F4 are

the displacements and forces at beam ends, respectively. ur and ub are the displacements

along rail and beam. u 0 is the displacement corresponding to the position where the roadbed

begin to yield. When − u0 < u r − ub < u0 , resistance of roadbed is a linear function of

u= u r − u b and distributed resistance in roadbed is ku . Let Young’s module and section area

for rail are E and A for beam are Eb and Ab respectively. Assuming that the resistance

has no effects on the deformation of beam, the displacement of beam is linear one, that is
x
ub = u3 + (u 4 − u3 ) .
L
The differential equation of the rail is:

d 2u r
AE − ku = 0
dx 2
(3)

k
Assume λ2 =
EA

85
The solution of the differential equation is
u 4 − u3
u r = c1e λx + c2 e −λx + x + u3 (4)
L

According to the boundary conditions, with the linear assumption of roadbed resistance the
element stiffness matrix can be given as:

 λγ − 2λ β 2λL − α 
 EA α EA
α
EA
Lα
EA
Lα 
 F1    u1 
   EA − 2λ EA
λγ
EA
2λL − α
EA
β
 u 
 F2   α α Lα Lα  2 
F  =  β 2λL − α Eb Ab β E A 2λL − α  u 3 
 3   EA EA − EA − b b − EA  
F  Lα Lα L Lα L Lα   u 
 4   2λL − α β Eb Ab 2λL − α Eb Ab β  4 
 EA EA − − EA − EA 
 Lα Lα L Lα L Lα  (5)

α = e λL − e − λL
 λL − λL
where γ = e + e
 β = (1 − λL)e λL − (1 + λL)e −λL


Resistance of Roadbed at Nonlinear Stage

When the relative displacement between rail and beam u r − ub is greater than u0 , the
maximum resistance of roadbed is m , and the equilibrium differential equation of the rail is
d 2 ur d 2 ur
AE = m . When u r − ub is less than − u 0 , AE = −m .
dx 2 dx 2
Assuming that u r − ub is greater than u0 , we have

d 2u r m
2
= (6)
dx AE
The solution of the differential equation is
1 m 2
ur = ( ) x + c1 x + c2 (7)
2 EA

According to the boundary conditions, the element stiffness matrix can be developed
can be given as when the resistance of the roadbed is linear:

86
  1 
 F1 − 2 mL   
EA EA
−
   L L 
1
 F − mL   EA EA  u1 
 2 2
−
  L  u 
L
 = E b Ab E A
 2 
 F3 + 1 mL   − b b  u 3 
 2   L L  u 
   E A E b Ab   4 
1  − b b 
 F4 + mL   L L 
 2  (8)

DOF Reduction

To reduce the element input work, the DOFS of each beam and rail segment should be
eliminated and only the DOFS of beam and rail nodes at span ends are left.
To establish the system equation, the relation of stiffness matrix and nodal
displacement vector can be written as:

 K bb K bi  ab   Pb 
K   =   （9）
 ib K ii  ai   Pi 

Where a b and ai are displacement vectors at boundary and inner nodes respectively.

From equation (9):

−1
ai = K ii ( Pi − K ib ab ) （10）

Substitute equation (10) into equation (9), we have:

−1 −1
( K bb − K bi K ii K ib )ab = Pb − K bi K ii Pi （11）

* *
or simply: K bb ab = Pb （12）

* −1
where K bb = K bb − K bi K ii K ib

* −1
Pb = Pb − K bi K ii Pi （13）

If ai is a vector with k element, equation (12) can be got by using Gauss-Jordan

elimination k times to equation (9). That is to get the result:

87
  K bb * 0 ab   Pb * 
 *   =  *  （14）
 K ib I  ai   Pi 
* *
where K bb , Pb are the stiffness matrix and load vector of the substructure after

* *
eliminating, and K ib , Pi are the corresponding matrix related to inner nodes.
* −1
Pi = K ii Pi
* −1
（15）
K ib = K ii K ib

Verification of the Program

With the results given above, the nonlinear dynamic program is developed based on
Wilson-θ method. The following 3-span bridge with the same piers is selected to check the
reliability of the program. The height, section size, and reinforcement ratio of the piers are
10m, 1.75m×1.75m and 1%, for the spans, the span length and section area are 32m, 2.32m2,
and the roadbed resistance factor is 10KN/m. By inputting the El Centro ground motion, the
result given by the program is compared with that given by ANSYS program. In the analysis
later, the discrete nonlinear spring models are used. The pier displacements from both analysis
are shown in Figure 7. It gives the good consistent.

1.5
Ansys results
1 Programe results
0.5
Dis.（cm）

Time（s）
0
-0.5 0 2 4 6 8 10 12

-1
-1.5
-2

Figure 7 Comparison of results

88
EFFECTS OF TRACK RESTRICTION ON THE SEISMIC RESPONSE OF
BRIDGES WITH THE SAME PIERS

The effects of track on the seismic response of a railway bridge is also affected by
other factors, such as the characteristics of piers, the roadbed conditions, and the span
numbers, etc. Therefore the discussion will be given according to different cases.
A simply supported railway bridges with piers of equal heights and section are used to
seismic response analysis. The bridges have the pre-reinforced concrete T-shaped
superstructures with the reinforced concrete substructures. The span length is 32m.
The longitudinal roadbed resistance coefficient is 10KN/m. In following table, the
main parameters of the piers are listed for various cases.

TABLE I PARAMETERS OF PIERS

Heights (m) 6 10 14 18 22 26
Section size 1.5×1.5 1.75×1.75 2.0×2.0 2.25×2.25 2.5×2.5 2.75×2.75
（m×m）
Stiffness 96 39 25 19 16 14
（MN/m）
Yielding moment 6020 9760 14790 21300 29490 39560
(KN-m)

Under the excitation of EL Centro ground motion record, the displacements of the left
pier tops with and without track restriction are shown in Figure 8.

8 14
Displacement (cm）

Displacement(cm)

12 Intensity Ⅸ without rails

6 without rails
10
with rails with rails
4 8
Intensity Ⅷ 6
2 4
2
0 0
0 20 40 60 80 100 120
0 50 100 150
Pier stiffness（MN/m） Pier stiffness（MN/m）

(a) Displacements for intensity Ⅷ (b) Displacements for intensity Ⅸ

Figure 8 Seismic response of bridges

It can be seen that the displacements of the pier tops are reduced when the rails are
taken into account. Fore this bridge the displacements are reduced by 10%~50%. The degree
of reduction relates to the structure, the site, and the seismic parameters.

89
Influence of Roadbed Resistance

According to the real track resistance, 3 resistance coefficients of roadbed, i.e. 10kN/m,
15kN/m, 20kN/m are chosen, to study how the change of the roadbed resistance affects the
seismic response of bridge. Two ground motion records, the EL Centro S00E and Los
Angeles N00W are used in the analysis. With different roadbed resistance, the displacements
of the pier top under intensity Ⅷ and Ⅸ are shown in Figure 9 (a) and Figure 9 (b),
respectively.

4.5 9
4 8

Displacment （cm）
Intensity Ⅷ Los Angeles Intensity Ⅸ Elcentro
Displacement (cm)

3.5 7
Elcentro Los Angeles
3 6
2.5 5
2 4
1.5 3
1 2
0.5 1
0 0
0 5 10 15 20 25 0 5 10 15 20 25
Roadbed resistence(KN/m)
Roadbed resistence（KN/m）

(a) Displacements for intensity Ⅷ (b) Displacements for intensity Ⅸ

Figure 9 Influence of roadbed resistance on the seismic response of bridge

The results show that with the increase of the roadbed resistance, the responses of the
bridge are reduced, but for different earthquake intensity the reduction have different
features.

Influence of Span Numbers

By considering different span number of bridge, the effects of the track restriction are
further discussed. With the same conditions and excitations the seismic response of bridge for
various span number are given in Figure 10 (a) and Figure 10 (b).
(a) Displacements for intensity Ⅷ (b) Displacements for intensity Ⅸ
5 7
Elcentro
6
Displacement （cm）

4 Los Angeles
Displacement

5
3
（cm）

4
2 3
1 2 Elcentro
Intensity Ⅷ
1 Intensity Ⅸ Los Angeles
0 0
0 5Number of bridge
10 spans 15 20 0 5 10 15 20
Number of bridge spans

Figure 10 Influence of span number

90
 It can be seen that with the increase of the span number, the seismic response of the
pier in the middle increases, but it is not larger than the response of the pier without
considering the track restriction.

EFFECTS OF TRACK RESTRICTION ON SEISMIC RESPONSE OF BRIDGES

WITH DIFFERENT PIERS

A 7-span railway bridge is taken as the example. The height of No.1 pier on the left side
is 6m and the heights of rest piers are 26m. The roadbed resistance coefficient is 10KN/m. By
inputting the EL Centro ground motion, the displacements of pier No.1 with and without
considering track restriction are shown in Figure 11.

1.5
Cosidering rails
1 Not considering rails
0.5
Dis(cm)

0
-0.5 0 1 2 3 4 5 6 7 8Time(s) 9
-1
-1.5
-2

Figure 11 Displacements of pier No. 1

It is can be seen that for stiffer pier the response is obvious higher than that without the
track restriction. At the same time, other piers show lower responses when track restriction is
considered. That is to say, if the track effect is neglected the safe result can be given for more
flexible piers and the not safe one can be given for stiffer piers.
8
Considering rails,same piers
6 Considering rails,different piers
4 Not considering rails
Dis.（cm）

2
0
-2 0 1 2 3 4 5 6 7 8 9

-4 Time（s）
-6

Figure 12 Displacements of pier No. 2

For other flexible piers, the effects of track shoe the same features as that discussed
above. In Figure 12 the response of pier No.2 (adjacent to the stiffer pier) is given.

Influence of Roadbed Resistance

With the same 7-span simply supported railway bridge and 3 different roadbed resistant
coefficients 10kN/m, 15kN/m, 20kN/m, the seismic responses are compared under the action of
EL Centro S00E, Imperial Valley, and Los Angeles N00Wground motion records. The results of
pier No.1 and pier No.2 are shown in Figure 13.

91
 7
3 Elcentro
6
2.5 Imperial Valley

Dis.（cm）
Los angeles
Dis.(cm）
2
4
1.5 3
1 2
Elcentro
0.5 Impreial Valley 1
Los angeles
0 0
0 10 20 30 0 10 20 30
Roadbed resistence(KN/m） Roadbed resistence(KN/m）
（a）Displacements of pier No.1 （b）Displacements of pier No.2

Figure 13 Influence of roadbed resistance

With the increase of roadbed resistance, the response of pier No.1 become obvious
bigger for the EL Centro earthquake. But the response for other two earthquake waves keep
at the same level approximately. Displacement of pier No.2 is reduced as roadbed resistance
increase, which shows track restriction to be helpful to flexible pier adjacent to a stiffer pier
with increase of roadbed resistance.

Influence of Span Number

By considering different span number, the effects of track restriction are discussed. The
seismic response of the pier No.1 and Pier No.2 are given in Figure 14.

5
2
Elcentro
4
Dis.（cm）

Imperial Valley
1.5
Dis.（cm）

Los Angeles 3
1 2
Elcentro

0.5 1 Imperial Valley

Los Angeles
0
0
0 5 10
0 5 10 Number of spans
Number of spans
（a）Displacements of pier No.1 （b）Displacements of pier No.2

Figure 14 Influence of span number

The displacement of pier No.1 increase as number of span increase. The track
restriction is not beneficial to this pier as the span number increase. The displacement of pier
No.2 becomes bigger with more spans.

CONCLUSIONS

The effects of track restriction on the seismic response of railway bridges are analyzed
above. From the discussion, following conclusion can be drawn:
(1) For most flexible pier, the tract restriction can reduce the seismic response of
bridges. This function is more obvious with the increase of roadbed resistance.
(2) With the even distribution of pier stiffness, the response of middle pier become

92
larger with the increase of span number, but it would not be larger than that without the track
effect.
(3) For different pier stiffness, track restriction has different effects for stiffer and
flexible piers. Especially for stiffer piers, the effect may be negative and should be take into
account.

REFERENCES

E.Maragakis, et al, Full-scale Resonance Tests of a Railway Bridge, Structures Congress - Proceedings 1, 1996
Duane E.Otter, et al, Seismic Performance of Railway Bridges, Railway Track & Structure, 1996.5
Luo Xuehai, et al, A Simplified Method for the Dynamic Analysis of Railway Piers, Earthquake Engineering
and Engineering Vibration, 1992.6
Jiang Jinzhou, Additional Longitudinal Forces in Continuously Welded Rails and Their Transmission on
Railway Bridges, China Railway Science, 1998,6
Guang Zhongyan, et al, Long Welded Railway Tracks, China Railway Press, 1995.7

93
 Seismic Design And Retrofit Guidelines For Bridges
In New Jersey,
A Low-To-Moderate Seismic Hazard Area

Harry Allen Capers, Jr., PE

ABSTRACT

This paper provides a description of the methodology prescribed for the design and retrofit of
bridges for the State of New Jersey, which is located in a low-to-moderate Seismic Hazard Area
of the United States. Considering designing structures to meet published seismic design and
retrofit requirements of the AASHTO bridge codes has long been a concern of many bridge
engineers in the central and eastern states of the United States. To date, most US standards have
been based on experiences gained from events occurring in the western states and on codes
written in Japan, New Zealand and Europe. Bridge owners outside of the high seismic zones of
the United States have either experienced or perceive an increase in the design effort and
construction costs associated with meeting these requirements. While no bridge owner will
knowingly design a bridge that would put the traveling public at risk, they must also attempt to
tailor the requirements of the AASHTO Specification and FHWA to meet the resources available
to them and a level of risk acceptable to the jurisdiction in which they work. This paper outlines
the approach that New Jersey has taken to address seismic design and retrofit of our structures to
meet the level of capital investment it can make at a level of risk that is acceptable to the state by
providing modifications to Adopted AASHTO standards. The paper assumes that the reader has
a basic understanding of the AASHTO Bridge Design Specifications.

_____________
Harry A. Capers, Jr., PE, Manager, Structural Engineering (State Bridge Engineer), New Jersey Department of
Transportation, PO Box 615, Trenton, New Jersey, 08625-0615, USA

95
INTRODUCTION

This paper provides a description of the methodology prescribed for the design and retrofit of
bridges for the State of New Jersey, which is located in a low-to-moderate Seismic Hazard Area
of the United States. The methodology described herein was first prepared and released in the
New Jersey Department of Transportation’s (NJDOT) Design Manual for Bridges and
Structures, Third Edition, 1998 for use by all designers engaged by agencies within the state to
provide highway structure designs in state and federally funded projects. At this time, New
Jersey was using the American Association of State Highway and Transportation Officials
(AASHTO) Standard Specifications for Highway Bridges, 16th edition with certain state directed
modifications as its design specification. Since that time, NJDOT has adopted the AASHTO
LRFD Bridge Design Specifications, 2nd Edition with certain state directed modifications, as it’s
current design specification. The material presented herein is based on that specification.
New Jersey is located in the northeastern portion of the United States, a region that has not in
recorded history experienced damaging earthquakes. While not uncommon in the region,
earthquakes that occur are usually low intensity events as can be in Table 1.

Intensity Magnitude
Location Date
MM Scale Richter Scale
Newark September 1, 1895 VI 5.00
Asbury Park June 1, 1927 VII 5.00
Trenton January 24, 1933 V 4.00
Central NJ August 22, 1922 V 4.00
Salem County November 14, 1939 V 4.00
West-Central NJ March 23, 1957 VI 5.00
NJ-PA border December 27, 1961 V 4.00
Southern NJ December 10, 1968 V 4.00
Table 1
As one can see, experience with earthquakes in this part of the country is much different than
that of western or central areas of the United States where most of the earthquakes have occurred
in this country. Consequently until recently the AASHTO Seismic Design codes were largely
based on these experiences that focused on larger events and not the magnitudes or risks that
could be expected in lower seismic zones. The effect of this has been that many owners were
experiencing cost increases in their designs as much as 10% due to seismic design requirements
when applying the AASHTO Design specifications without modification.
As can be seen from the above table, for most in New Jersey earthquakes are less than a
reality. Further this being typical, seismic design of bridges has always been a difficult issue for
bridge owners to accept as a necessary consideration when designing or rebuilding bridges
within their jurisdiction especially as the potential exists to draw down as much as 10% of the
capital dollars available to them in attempting to meet a low risk requirement of the code. As
such, in New Jersey, we have assessed what we feel to be the owner’s risks and modified the
AASHTO LRFD Bridge Design Specifications, 2nd Edition to, in our opinion, to more
realistically represent our state’s needs.

96
SEISMIC DESIGN OF HIGHWAY BRIDGES IN NEW JERSEY

Our current design manual, Design Manual for Bridges and Structures, Third Edition, 1998
directs that seismic design of new highway structures shall follow the requirements of Division
1-A of the AASHTO Standard Specifications for Highway Bridges, 16th edition. In our new
manual, we will be directing that the Seismic design of new highway structures shall follow the
requirements of Subsection 3.10 of the AASHTO LRFD Bridge Design Specifications. For the
designer, this is consistent in that in general seismic design provisions contained in the AASHTO
LRFD Specification were based on the code contained in the 16th Edition of the Standard
Specification. In view of the fact that our new Design Manual will be issued shortly, this paper
will reference the final draft of our LRFD version, Design Manual for Bridges and Structures,
Fourth Edition.
Our modifications to the AASHTO Specification are intended to apply to bridge spans not
greater than 500 feet and to superstructures of slab bridges, steel girders, concrete girders, box
girders or truss bridges as does the AASHTO Specification itself. We require that any structure
not meeting these requirements be designed using the results of a site-specific seismic analysis
and on occasion, have also required site-specific analysis on unusual structure configurations.
Under either code, the first considerations of an owner is to determine for the Standard
Specification which seismic performance category or for the LRFD code which seismic
performance zone is to be used for the design of the structure. To make this determination, two
characteristics of the bridge must be known. The first is the importance category of the structure.
Under the Standard specification this became a non-issue as classifying the bridge as either an
essential bridge or other did not change the seismic performance category. However, this
classification becomes more significant in the LRFD specification as it segregates post event use
of the structure into critical, meaning immediate use by all traffic or designing for no damage,
and essential, use by emergency vehicles only which allows accepting some damage to the
structure. Both specifications only require no collapse of the superstructure for the lowest
importance category. Because of the significant redundancy in the roadway system of the State
it was decided that classifying our bridges as essential would provide an acceptable level of risk.
The second characteristic is the ground acceleration coefficient at the location of the bridge.
This is found by referencing contour maps of horizontal acceleration prepared for the whole
country by the US Geological Survey. For the State of New Jersey ranges from a high of .18 just
to the west of the New York metropolitan area to a low of .075 at the very southern end of the
state are provided in the USGS maps. For the sake of uniformity of design, to determine the
seismic performance category or seismic performance zone throughout the state we have
specified that the minimum seismic performance shall be Zone 2 for the entire State of New
Jersey. The guidance provided in Subsection 3.10.9.3 of the AASHTO LRFD Bridge Design
Specifications should be referred to for clarification on the use of this designation.
The acceleration coefficient for horizontal force effects for use in the design and retrofit of
bridge structures is specified in our guidance to designers. It is given on a county-to-county
basis as shown in Table 2.

97
 Acceleration Coefficient County
A = 0.10 Atlantic, Cape May, Cumberland, Salem
A = 0.15 Burlington, Camden, Gloucester, Monmouth, Ocean
Bergen, Essex, Hudson, Hunterton, Mercer,
Middlesex, Morris, Passaic, Somerset, Sussex,
A = 0.18 Union, Warren
Table 2
For a bridge structure that is located on the border between two counties with different
acceleration coefficients, the larger value shall be used. We also direct that vertical components
of acceleration shall be neglected.

METHODS OF ANALYSIS

With respect to specific guidance that is provided to designers we direct that the magnitude
of seismic forces be determined by considering several factors. These include dead weight of the
structure, ground motion (acceleration coefficient), type of soil, fundamental period of vibration,
and the Importance classification of the bridge. Other specific guidance provided includes the
following items

Single Span Bridges

Due to the higher relative stiffness of abutments when compared to piers, with a single span,
the ability to resist earthquakes is increased. Accordingly, for single span design, no formal
analysis is required for seismic forces. Designers are directed to consider minimum force
requirements as provided in the AASHTO LRFD Bridge Design Specifications for connection of
the superstructure to the substructure and to satisfy minimum seat width requirements.
The abutments shall be designed for the effects of static earth pressure and the additional
seismic induced earth pressure forces, using the Mononobe-Okabe method. This method is an
extension of Coulomb’s method for analyzing soil pressure on retaining walls. With the use of
this method, we direct that the backfill be assumed to be unsaturated so that liquefaction effects
are negligible, the backfill is assumed cohesion-less and that seismically induced active and
passive pressures is considered.

Single Span Bridges with Integral Abutments

NJDOT directs that the abutments be designed for seismic forces from the superstructure in
addition to the static earth pressure and seismic induced forces using the Mononoke-Okabe
method. Both active and passive pressures are to be considered.

Multi-Span Regular Bridges

The Uniform load (Equivalent static load) method or the single mode spectral method is
acceptable methods of design. The Uniform load method will require hand calculations, while

98
the single mode spectral method will require the use of computer software. Also, in using the
uniform load method the specific guidance provided by the AASHTO LRFD Bridge Design
Specifications are to be followed.

Multi-Span Irregular Bridges

For such bridges, a multi-mode spectral analysis method is required for the analysis of
substructures.

Load Combinations

We specify that the AASHTO LRFD Extreme Event-I load combinations be applicable for a
seismic analysis. As directed in the AASHTO LRFD Bridge Design Specifications, Dead Load
and Live Load forces are to be combined with the forces from a single or multi-mode analysis as
follows:
100% of longitudinal seismic forces + 30% of transverse seismic forces
100% of transverse seismic forces + 30% of longitudinal seismic forces

Site Coefficients and Site Effects

Due to large variations in the values of site coefficients, soil profiles will be based on soil
composition at the specific bridge site. A geotechnical investigation is to be performed at all
proposed bridge locations to determine the soil conditions, whether cohesive or cohesionless, the
type of rock, sand, gravel, and stiff clay, soft clay or silt.

Liquefaction

The potential for soil liquefaction and liquefaction related ground instability is to be
investigated at relevant locations along proposed project alignments. Effects of settlement of
footings, loss in bearing capacity and increased lateral earth pressures is to be considered in the
design of abutments, walls and footings.

Seismic Slope Instability and Landslide

The potential for seismic induced slope movements and landslides along the proposed
alignment must be investigated for all projects. Mitigation measures are to be incorporated in the
design of abutments, walls and footings.

Response Modification Factors

The LRFD Bridge Design Specifications recognize that it is uneconomical to design a bridge
to resist large earthquakes elastically and therefore are assumed to deform inelastically once they
exceed their design level. According to the provisions of the AASHTO LRFD Specifications,
Response Modification Factors or R-factors shall be used to reduce the moments and forces due
to the ability of a member to develop a plastic hinge. Moments and forces for member and
connection designs are to be computed by dividing the forces and moments obtained from load

99
combinations by the appropriate Response Modification Factors. This is the design level of the
member.

Abutment Analysis

For all multi-span bridges, regular, irregular or with integral abutments, earthquake forces
from the superstructure shall be considered. Abutments need to be analyzed for seismic forces
from the superstructure, in addition to the static earth pressure and seismic induced forces using
the Mononobe-Okabe method. Both active and passive pressures shall be considered.

Miscellaneous Guidance

Seismic ductility design at locations where plastic hinges will form shall be accounted for on
all new structures. Seismic effect considerations are not required for buried structures or culvert
structures. In addition to the AASHTO LRFD Bridge Design Specifications, we also recognize
the FHWA Geotechnical Engineering Circular No. 3 titled, “Design Guidance: Geotechnical
Earthquake Engineering for Highways” as a reference on seismic design.

Alternate Design Criteria

As an alternative to the use of the AASHTO LRFD Specifications, NCHRP Report 472,
“Comprehensive Specification for the Seismic Design of Bridges” may be used. Designers may
submit a request to the Manager, Bureau of Structural Engineering for the use of the NCHRP
Report. A comparison of the effects on the design of a project between the two documents
should be made to validate the request.
This specification was the result of an AASHTO sponsored National Cooperative Highway
Research Project initiated in 1998 to develop a state of the art seismic specification compatible
with the AASHTO LRFD Bridge Design Specification. The proposed specification was
intended to replace the current specifications by incorporating experiences gained around the
world in all areas of seismic design in a nationally acceptable specification. As was previously
stated the existing provisions are based on data that is now considered 10 to 20 years out of date.
The project was completed by a joint venture of the Applied Technology Council and
MCEER last spring and was considered by the AASHTO Subcommittee on Bridges and
Structures at it’s annual meeting this past May. Apparently due to concerns raised over the new
USGS seismic hazard mapping the new specifications were not adopted to replace the existing
provisions at this meeting.
The specification is available for use and does provide certain benefits to bridge owners in
low to moderate seismic risk areas by focusing on proper detailing for seismic events rather than
analysis. In the two trail designs performed for the project by New Jersey designers, significant
reductions in seismic forces were noted. Other benefits include treatment of new seismic load
resistant systems, improved soils information, it forces the designer to identify the seismic load
path and it provides a state of the art, performance based method of designing bridges for seismic
loading.

100
SEISMIC RETROFIT OF EXISTING HIGHWAY BRIDGES IN NEW JERSEY

An even greater challenge to us over time has been how to address existing bridges in our
inventory that were designed to less than today’s standards. New Jersey has approximately 6500
highway carrying bridges within their jurisdictions of which all but about 1250 are over 25 years
old. Needless to say, seismic loading was not considered in their design.
Further, in that resistance to seismic loads has not been considered at all in their design,
retrofitting these structures would be a very difficult and costly task to undertake. To minimize
the impact to our available budget New Jersey has chosen to address this issue, as it affects our
existing inventory, as our structures are programmed for rehabilitation or replacement.
Typically, preventing collapse of the superstructure is targeted.
NJDOT specifies that the seismic retrofit design of existing highway structures shall follow
the guidelines of the FHWA publication titled “Seismic Retrofitting Manual for Highway
Bridges” currently numbered as, FHWA-RD-94-052, May 1995. We also specify that highway
structures shall be retrofitted for a Zone 2 earthquake, which is consistent with our criteria for
new bridges.
However, prior to proceeding with retrofit design, we require a Seismic Retrofit Report to be
prepared to provide a determination as to a bridge structure’s eligibility for a seismic retrofit. A
flow chart to provide guidance in determining if a bridge structure qualifies as a seismic retrofit
candidate is included herein as Figure 1. The results of the analysis, performed in accordance
with .the flow chart, shall be provided in the Seismic Retrofit Report.
In preparing the Seismic Retrofit Report, the following guidance shall be followed. Initially,
seismic retrofitting of a bridge structure shall only be considered under the following conditions:
a. The planned work will involve widening of a deck by more than 30% of its deck area; or,
b. The planned work will involve an entire deck replacement; or,
c. The planned work will involve superstructure rehabilitation or replacement, major
abutment or pier repairs to bearing seat areas or bearing repairs or replacement.

101
 Figure 1 - Additional Analysis Required for
Existing Bridges Found in Planned Projects
Bridge work is planned

Will
Does work
deck be widened
involve entire
by more than 30% of No No
deck replacement or
existing
greater?
area?

Yes
Yes

Is the Extent and level of retrofit to be approved

Structure a special by the Manager of Structural Engineering
type such as an arch, with concurrence by the FHWA on Full-
Yes
movable span, or on or Oversight Federal Projects and SHPO
eligible for historic consultation for structures that are on or eligible
register? for the historic register.

No

Is Structure
feasible to retrofit to No
Category B ?

Yes

Analyze the structure and determine retrofit

strategy and cost. Include cost for
rehabilitating structural deficiencies, scour
etc.

Is retrofit
strategy cost > 30% Should the structure be
Yes No
increase over original replaced ?
proposed work?

No Yes

Continue with planned work and Replace No Retrofit

include retrofit for Seismic, Scour, the needed. Continue with
and deficiencies Structure planned work

102
 The Report should also include a study of the proposed project to determine if retrofitting a
bridge is a cost-effective measure. At a minimum, we direct that two things be done. First, an
investigation to determine the extent of retrofitting which may be required must be performed
and documented. Second, prior to making a detailed evaluation of the seismic capacity of the
bridge structure, the relationship of the bridge structure to other bridge structures on neighboring
routes that may also be damaged during an earthquake, be considered. Consider two bridge
structures that have similar functions, such as bridge structures A and B as detailed in Figure 2.
Assume that bridge structure B is currently in a project and being considered for seismic retrofit.
It is possible, that retrofitting bridge structure A would be more economical or that bridge
structure A is more seismically adequate than bridge structure B. Accordingly, even though
bridge structure A is not in the project scope and bridge structure B is, it would be more rational
to retrofit bridge structure A than bridge structure B. If so, not seismically retrofitting the bridge
structure may be justified.

Br id g e A Br id g e B

Figure 2
Parallel Bridges

Seismic Retrofit of Bearings

Several methods of seismic retrofit are outlined for bearings and expansion joints within the
FHWA Retrofit Manual that is referenced above. Of these methods NJDOT design guidance
suggests the following for consideration in order of preference. If applicable, a recommendation
as to the proposed treatment of a bridge structure should be included in the Seismic Retrofit
Report.
a. Modify existing bearings to resist seismic loads or to prevent toppling of existing
bearings by installing longitudinal displacement stoppers.
b. Use of longitudinal joint restraints as outlined in the FHWA Retrofit Manual.
c. Bearing replacement with those type bearings identified in the NJDOT Design Manual
for Bridges and Structures. If conventional steel and elastomeric bearings, are proposed
to remain, typical modifications to these bearings to withstand the specified design
earthquake loadings would include the following:

Modifications To Steel Bearings

a. Increase size, number or embedment of anchor bolts.

b. Increase the outer diameter of the pinhead.
c. Increase the width of the expansion rocker.

103
 d. Increase the top and bottom dimension of the pintle detail for increased movement.

Modifications to Elastomeric Bearings

a. Secure bearing against horizontal and vertical movement.

b. Modify the plan area and/or thickness of the elastomeric bearing to reduce
seismic forces to the substructure.
The methods outlined above are recommended procedures and are not intended to restrict the
ingenuity and creativity of the Design Engineer. Each bridge is different; therefore, the owner
will approve retrofit procedures on a project-to-project basis.
If it is found through a seismic analysis that the substructure is in need of seismic retrofit, it
will probably be economically advantageous to study bearing replacement as part of a retrofit.
In evaluating a bridge structure’s history, a significant traffic count should warrant an
increase to the Seismic Load Extreme Event load combination. As such, the 0.50 live load
factor, listed in the AASHTO LRFD Bridge Design Specifications as an Extreme Event II load
combination, shall be combined with the Extreme Event I Earthquake load factor. A Designer
should use his engineering judgment in assessing the traffic count in applying the increase.

CONCLUSION

Considering designing structures to meet published seismic design and retrofit requirements
of the AASHTO bridge codes has long been a concern of many bridge engineers in the central
and eastern states of the United States. To date, most US standards have been based on
experiences gained from events occurring in the western states and on codes written in Japan,
New Zealand and Europe. Bridge owners outside of the high seismic zones of the United States
have either experienced or perceive an increase in the design effort and construction costs
associated with meeting these requirements. While no bridge owner will knowingly design a
bridge that would put the traveling public at risk, they must also attempt to tailor the
requirements of the AASHTO Specification and FHWA to meet the financial resources available
to them and a level of risk acceptable to the jurisdiction in which they work. In this paper, I have
outlined the approach that New Jersey has taken to address seismic design and retrofit of our
structures to meet the level of capital investment we can make at a level of risk that is acceptable
to the state.
The AASHTO Subcommittee on Bridges and Structures continues to work to find a
reasonable solution to this issue through the issuance of a nationally acceptable seismic design
specification. The recent work completed by the ATC/MCEER joint venture published as
NCHRP Report 472 promises to provide the solution. Hopefully, the remaining concerns
expressed by the states at the last annual meeting will be addressed shortly and a state of the art
specification addressing all geographic areas of the country will be adopted.

104
ACKNOWLEDGEMENT
A special thanks to Jack Mansfield, PE Manager, Geotechnical Engineering, NJDOT, Jose
Lopez, Project Engineer, Shirish Patel, Principal Engineer, and Sharon Harrison, Administrative
Assistant, all of Structural Engineering, NJDOT for their tremendous support in editing and
formatting this paper. Finally, thanks to Dr. Mohiuddin (Ali) Khan, PE of STV, Inc. for his
tremendous assistance in developing our modifications described herein.

REFERENCES

AASHTO 2002, “AASHTO LRFD Bridge Design Specifications”, Second Edition, American Association
of State Highway and Transportation Officials, Inc., 444 North Capital Street, NW, Suite 249, Washington, DC,
20001

AASHTO 2002, “Standard Specifications for Highway Bridges”, Sixteenth Edition, American Association
of State Highway and Transportation Officials, Inc., 444 North Capital Street, NW, Suite 249, Washington, DC,
20001

ATC/MCEER Joint Venture, 2002 “NCHRP Report 472 – Comprehensive Specification for the Seismic
Design of Bridges”, Transportation Research Board, National Research Council, 2101 Constitution Avenue, NW,
Washington, DC 20418

Dombroski, Daniel R., Jr., 1977 “Earthquakes in New Jersey”, New Jersey Department of Environmental
Protection, Bureau of Geology and Topography, PO Box 2809, Trenton, NJ 08625-2809

FHWA 1995, Report No. FHWA-RD-94-052 “Seismic Retrofit Manual for Highway Bridges”, Buckle, Ian
G. and Friedland, Ian M (Editors), Office of Engineering and Highway Operations, R&D, Federal Highway
Administration, 6300 Georgetown Pike, McLean, Va 22101-2296

NJDOT 2002, Final Draft of “Bridges and Structures Design Manual-LRFD”, Fourth Edition, New Jersey
Department of Transportation, Bureau of Structural Engineering, PO Box 615, Trenton, NJ 08625-0615

NJDOT 1998, “Bridges and Structures Design Manual”, Third Edition, New Jersey Department of
Transportation, Bureau of Structural Engineering, PO Box 615, Trenton, NJ 08625-0615

105
 Generation of Response Spectrum Compatible
Non-stationary Ground Motions Based on Phase Difference
Spectra

Qing-shan Yang1 Hai-peng Jiang1

ABSTRACT

A numerical method is presented for computing artificial earthquake records consistent

with a specified target response spectrum or power spectral density. The phase angle spectrum in
the proposed algorithm is not an independent random variable distributed in [0, 2π ]
uniformly, but rather is generated on the basis of the statistical characteristics and the
distribution law of phase different spectrum, which controls the nonstationarity of the generated
ground motions. The efficiency of the algorithm and the accuracy of the fitting process are
substantially improved by applying the component-by-component modulating approach. The
proposed procedures are validated by examples.

_____________
1
College of Civil Engineering and Architecture, Northern Jiaotong University, Beijing, 100044, PRC

107
INTRODUCTION

Response spectrum has been applied in many seismic codes over the world and used as
the target spectrum in generating the artificial ground motions which has been studied by a
number of authors through various methods (Kaul, 1978a; Preumont, 1984; Ghosh, 1993).
For artificial ground motion, nonstationarity in time and frequency domains is one of the
most concerns (Deodatis, 1996; Sabetta, 1996; Shrikhande, 1996). After systematic studies
(Yang, 200; Zhao, 1992; Nigam, 1982; Ohsaki, 1979), it is concluded that it is the phase
difference spectrum, which mainly controls the nonstationarity of the artificial ground motions.
This conclusion is applied in this proposed algorithm. The design response spectrum specified in
Code for Seismic Design of Building of China (GB11-89, China) is applied in the presented
paper. To improve the matching efficiency, the component-by-component modulating technique
is applied, in which the modulating ratios of amplitude spectra for different components are
different according to their contribution to the responses.

GENERATION ALGORITHMS

Among various procedures for ground motion simulation, the method relying on
trigonometric approximations (Shinozuka, 1967; Mignolet, 1996) is generally applied and
written as:
n
a (t ) = ∑ Ak cos(ωt + φ k ) (1)
k =1

Eq.1 may be rewritten in the discrete Fourier transformation as:

N −1 N −1
( )
a (t ) = a (m∆t ) = Re[∑ F (ω k )e iω k m ] = Re[∑ F (ω k ) e −iϕ k e iω k m ] ∆t = Td / N (2)
k =0 k =0

where Td is the time length of the generated time history; φ k is the phase angle spectrum; N
is the sample number and should be an exponential function of 2 to use Fast Fourier
Transformation technique; and | F (ω k ) | is the two-sided Fourier amplitude spectrum, which
may be derived from the one-sided Fourier amplitude spectrum | F1 (ω k ) | :

 F (ω ) k = 0, N 2
F (ω k ) =  1 k (3)
 F1 (ω k ) / 2 k = 1,2, L , N 2 − 1

and one-sided Fourier amplitude spectrum | F1 (ω k ) | may be determined from energy

considerations as
| F1 (ω k ) |= [4G (ω k )∆ω ]1 2 (4)

So, the two-sided Fourier amplitude spectrum may be denoted directly

108
 [4G (ω k )∆ω ]1 2 k = 0, N 2
F (ω k ) =  (5)
 [G (ω k )∆ω ]
12
k = 1,2, L, N 2 − 1

where G (ω k ) is the one-sided power spectral density. To establish the relation between the
Fourier amplitude spectrum and the response spectrum, the approximate relation between
response spectrum and power spectrum was developed (Kaul, 1978b):

ξ   −π 
G (ω ) = [ S aT (ω )] 2 − ln  ln(1 − r )  (6)
πω  ωTd 
in which Td is the time length of the recording, r is exceeding probability, equals 5~10%
usually, ξ is the damping ratio of the linear oscillator which is used to calculate the response
spectrum S aT (ω ) .

Parameters φ k for ground motion are usually selected to be independent random

variables uniformly distributed in [0, 2π ] (Kaul, 1978a; Ohsaki, 1979; Preumont, 1984). While
after systematic studies, it is concluded that the phase angle should be obtained according to the
following procedures (Yang, 2001; Zhao, 1992; Nigam, 1982; Ohsaki, 1979):
φ k +1 ( f ) = φ k ( f ) + ∆φ k ( f ) k φ k ∈ [0, 2π ] k = 0, 1, L n − 1 (7)
∆φ k ( f ) ∈ [−2π , 0]
∆φ k ( f ) = ∆φ k ( f ) + ε k k = 0, 1, L n − 1 (8)
and any value in [0, 2π ] may be specified as the initial phase angle φ 0 . In Eq.7 f = ω / 2π is
the linear frequency; ∆φ k ( f ) is the phase difference spectrum, ∆φ k ( f ) the mean function of
phase difference spectrum and ε k the fluctuating part of the phase difference spectrum about the
mean function. The empirical relation of ∆φ k ( f ) to its influence factors is:

log10 ∆ϕ ( f ) = a1 ( f ) + a 2 ( f ) ⋅ M + a 3 ( f ) log 10 ( R + R0 ) (9)

and ∆ϕ ( f ) = − ∆ϕ ( f ) < 0 (10)

where M is the earthquake magnitude; R (Km) is the epicenter distance; R0 =15 is a

constant. a1 ( f ), a 2 ( f ), a 3 ( f ) are regressive coefficients as functions of frequency and their
values for bedrock site are listed in Yang (2001) and Zhao (1992).
ε k is the fluctuating part of the phase difference spectrum about the mean function and
can be obtained from:

ε k = −(ε bk − 2π ) and ε bk = e ε ck
(11)

and ε ck is a random variable satisfying the normal distribution law with mean value mc and

109
standard deviation σ c which can be calculated from:
1 σ σ
mc = ln 2π − ln(1 + ( ε ) 2 ) σ c = ln(1 + ( ε ) 2 ) (12)
2 2π 2π
and σ ε is the standard deviation of random variable ε k and may be obtained from:
log10 σ ε = d1 + d 2 ⋅ M + d 3 log10 ( R + R0 ) (13)
where d 1 , d 2 , d 3 are coefficients; the other symbols are the same as that in Eq.9. d1 = −1.124 ,
d 2 = 0.089 , d 3 = 0.316 for bedrock site (Yang, 2001; Zhao, 1992).
Substitute Eqs.5~7 into IFFT (Eq.2), the ground motion compatible to response spectrum
S aT (ω ) may be obtained. While, as Eq.6 is an approximate relation, the obtained ground motion
may be not compatible to target response spectrum accurately enough, it is necessary to
modulate the obtained ground motion to improve the accuracy. Let the target response spectrum
be noted as S aT (ω ) , the response spectrum of the obtained ground motion as S a (ω ) , if S a (ω )
is not close enough to S aT (ω ) , at any fitting point ω = ω i , i.e.,
R(ω i ) −1 > ε ( ε is the tolerant error) (14)
where
R (ω i ) = S aT (ω i ) S a (ω i ) (15)

then iterative matching is needed and the fitting technique generally used to modulate the
Fourier amplitude spectrum (Shaw, 1975, Kaul,1978a) is:
| F I +1 (ω k ) |=| F I (ω k ) | ⋅R(ω i ) (16)
I is the iteration number and ω k is the frequencies near the fitting point ω i .
and substitute the modulated Fourier amplitude spectrum into IFFT (Eq.2), then obtained the
renewed ground motion. The convergence criteria should be satisfied following the above
iteration.

COMPONENT-BY-COMPONENT MODULATING TECHNIQUE

Though the above fitting technique is applied broadly, what should be noted is its
convergence speed is not fast, and sometimes the needed accuracy was not able to be reached at
some fitting frequencies. Many works has been done on this problem, and we believe
component-by -component modulating technique (Hu, 1986) is a good idea, which will be
illustrated by the following example.
It is assumed that there is an oscillator with damping ratio ξ =0.05 and natural circular
frequency ω =10.466 rad/s. Its absolute acceleration response x(t ) under the S00E component
of EL Centro Wave a(t ) (Figure1) may be obtained by numerical integration (Figure2). Its
maximum absolute value, i.e., the response spectrum S a is | x(t ) | max = 8.33m / s 2 which occurs
at 2.22s, noted this time as t max (Figure2).

110
 4 2
a max =3.41m /s
2

a(m/s )
2
0

-2

0 5 10 15 20 25 30 35 40
t[sec]

Figure1 S00E component of the time history recorded at EL Centro station in the
Imperial Valley earthquake, 1940
10 2
|x(t)|max=8.33m/s
8
tmax=2.22s
6
4
Accl[m/s ]
2

2
0
-2
-4
-6
-8
0 5 10 15 20 25 30 35
t [s]

Figure 2 Absolute acceleration response of the EL Centro wave

Meanwhile the S00E component of EL Centro wave may be decomposed to be a real Fourier
series:
n
a (t ) = ∑ Ak cos(ω k t + ϕ k )
k =1
n
(17)
= ∑ a k (t )
k =1

The absolute response of the oscillator under each

component of the EL Centro Wave a k (t ) , x k (t ) , 4
k=10.466/∆ω =135
may also be obtained by numerical integration (part 3
xk(tmax)
of them shown in Figure3). 2
2.5 tmax=2.22秒 1
xk(tmax)

2.0
0
1.5
1.0 -1
0.5 x232(tmax)=0.92 -2
xk(t)

0.0
-0.5 -3
x641(tmax)=-1.06 k=641
-1.0 k=232 -4
-1.5
-2.0
-5
0 500 1000 1500 2000
-2.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 component number (k)
t [s]

Figure 3 Components of the absolute acceleration Figure 4 Absolute acceleration response values of
each component at t = t max

111
 S a =| x(t ) | max =| x(t = t max ) |
It is apparent that: =| ∑ x k (t = t max ) | (18)
k

Response of the oscillator and the distribution of x k (t = t max ) (k = 1,2,3....n) is shown in

Figure4.
Figure4 and 3 indicate that though the total response x(t ) gets its maximum absolute value
at t = t max , the component response x k (t ) at t = t max may not be its maximum value, even its
sign may be also different with that of x(t max ) . We may group the response components x k (t )
based on their sign at t = t max as:
Group A: the sign of x k (t max ) is same as that of x(t max )
Group B: the sign of x k (t max ) is contrary to that of x(t max )
and apply different modulating strategy to them:
If S a =| x(t max ) | is less than the target response spectrum S aT (ω ) , the amplitude spectrum
F (ω k ) corresponding to Group A will be increased, and that corresponding to Group B
decreased. If S a =| x(t max ) | is bigger than S aT (ω ) , the modulating direction is contrary to the
previous. This modulating technique, named as component-by-component fitting technique, will
generally be more efficient than Eq.16 and may be formulated as:

For Group A: | F I +1 (ω k ) |=| F I (ω k ) | ⋅R(ω i ) (19a)

For Group B: | F I +1
(ω k ) |=| F (ω k ) | / R(ω i )
I
(19b)
and unified as:
sign[ xk ( t max ) x ( t max )]
F I +1 (ω k ) = FkI (ω k ) ⋅ ( R(ω k )) (20)

It is should be noted that the total response is practically contributed by a few components
with frequencies ω k close to the natural frequency of the oscillator ω i , named as effective
components (Figure4), and only the effective components need to be modulated for each
oscillator with different natural frequencies.

GENERATION OF RESPONSE-SPECTRUM-COMPATIBLE GROUND MOTIONS

Based on the ground motion generation procedures (Sect.2) and the modulating
technique (Sect.3), the procedures of generating fully nonstationary and
response-spectrum-compatible ground motions may be summarized as:
1. Determine the possible earthquake magnitude M, epicenter distance R(Km), and target
response spectrum, including parameters, such as the characteristic period of the site, the
maximum value of the response spectrum.
2. Obtain the power spectral density G (ω ) from the target response spectrum S aT (ω ) at the
fitting points ω = ω i ( i = 1,2, K j , j is the number of the points) based on Eq.6.
3. Obtain the amplitude spectrum of complex Fourier transformation from power spectral

112
 density at the fitting points ω i ( i = 1,2, K j ):

F (ω i ) = [G (ω i )∆ω ]1 2 ∆ω = 2π / Td (21)

then obtain the amplitude spectrum at each calculating frequency by linear interpolation:
| F (ω k ) |= 0 ω k < ω i =1 (22a)

| F (ω i +1 ) | − | F (ω i ) |
| F (ω k ) |=| F (ω i ) | + (ω k − ω i ) (22b)
ω i +1 − ω i
ω i ≤ ω k = k∆ω ≤ ω i +1 i = 1,2,K j − 1 k = 1,2, L , N 2 − 1

| F (ω j ) | − | F (ω j −1 ) |
| F (ω N 2 ) |= 2(| F (ω j −1 ) | + (ω N 2 − ω j −1 )) (22c)
ω j − ω j −1

Where N is the sample number which should be the exponential function of 2 for using Fast
Fourier Transformation.
4. Obtain the phase different spectrum based on the field condition, magnitude and epicenter
distance from Eqs.8~13.
5. Specify the initial phase angle and obtain the phase spectrum following Eq.7.
6. Obtain the complex Fourier Spectrum
F (ω k ) = F (ω k ) e − iϕ (ω k ) k = 0,1,2 K N 2 (23a)
F (ω k ) = F * (ω N − k ) k = N 2 + 1, K N − 1; (23b)
where * denotes conjugate complex.
7. Obtain the initial ground motion from Inverse Fast Fourier Transformation (Eq.2).
8. Check the fitting accuracy and modulate the ground motion. Calculating the absolute
acceleration response x(t ) of the oscillators with natural frequencies ω = ω i ( i = 1,2, K j )
under the generated ground motion and determine the absolute maximum value | x(t ) | max , its
occurring time t max , the algebra value of x(t max ) , then check the accuracy.
If R(ω i ) −1 ≤ ε for each i = 1,2, K j ( ε is the specified tolerant error), it is indicated
that the generated ground motion satisfies the fitting accuracy, and stop the iteration.
Otherwise, modulate the generated ground motion following the procedures:
i). Calculate absolute acceleration response x k (t max ) of each oscillator with frequency
ω = ω i ( i = 1,2,K J , J is the number of the fitting points where the generated ground
motion do not satisfied the accuracy) under each component of the generated ground motion
Ak (ω k t + ϕ k ) at t max .
ii). Determine the contribution of each component,
β k = x k (t max ) x(t max ) (24)
β k > ε 1 ( ε 1 is the specified contribution level, for example, ε 1 =0.1) indicates this
component gives an important contribution to the total response, i.e., it is effective
component, and should be modulated.
iii). Modulate the amplitude spectrum of every effective component for each controlling

113
 frequency following Eq.23 and re-substitute the modulated amplitude spectrum into Eq.2. to
obtain the renewed ground motion. Checking the fitting accuracy till the tolerant error is reached
at each fitting point.

EXAMPLES

It is assumed the soil of the researched field is bedrock, i.e., site category I; the
earthquake intensity is VII; the earthquake risk analysis shows the maximum magnitude of the
possible earthquake is 6.5 and the epicenter distance is 150km. It is needed to generate the
possible earthquake ground motions compatible to the response spectrum specified by the Code
for Seismic Design of Buildings of China (GBJ 11-89) with the above far-earthquake risk
parameters.

0.9
(Tg/T) ∗αmax
0.8
0.8

Response spectrum [m/s ]

2
Response spectrum [m/s ]
2

O fitting points
0.6
0.6
linear function linear function
0.4
0.4

0.1568
0.2 0.2

linear function

0.0 0.0
0.01 0.1 Tg=0.25s 1 3 0.01 0.04 0.1 Tg=0.25 1 3 10
period [s] Period [s]

Figure 5 Target response spectrum Figure 6 Modified response spectrum

1. Parameters of the target response spectrum. The response spectrum curve specified by Code
for Seismic Design of Buildings of China (GBJ 11-89) is shown in Figure5 where the
characteristic period is 0.25s for site category I and far-earthquake, the maximum value of
response spectrum S amax is 0.08g for intensity VII (g is the gravity acceleration). The
response spectrum curve may be divided into three parts according to the natural period, from
T =0s to T =0.1s, the response spectrum varies from 0.45 S amax =0.036g to S amax =0.08g;
from T =0.1s to T = Tg =0.25s, it maintains its maximum value S amax =0.08g; from
T = Tg =0.25s to T =3s, it varies following function S a = (Tg T ) 0.9 S amax , but if
(Tg T ) 0.9 < 0.2 , S a = 0.2 S amax = 0.016 g .
2. Some pre-specified parameters of the target ground motions. Specify the sample number
N=212=4096 and the time step as 0.02s, then the time length of the ground motion is
Td = 80.92s; the frequency step ∆ω = 2π Td =0.07765 and the highest frequency in Fourier

114
 2π N
transform is ω N 2 = ⋅ = 50π , i.e., the frequency scope of the Fourier spectrum is
N∆t 2
[0, 50π ], the corresponding period scope is [0.04s, ∞] . The definition domain of the target
response spectrum should also be [0.04 s, ∞] in determining the Fourier amplitude
spectrum from Eqs.5 and 6. The period domain of response spectrum in Code for Seismic
Design of Buildings of China (GBJ 11-89) should be extended to be that the response
spectrum is zero when the period is bigger than 20s and the response spectrum for periods
within [3s, 20 s ] are interpolated linearly between S a (3) =0.016g and S a (20) =0. Based
on these parameters, the modified response spectrum is shown as Figure6.
3. Specify the fitting points. Within period T ∈ [0.04, 0.08] , period step ∆T = 0.005 , the
number of fitting points j1 =8; ∆T = 0.009 , j 2 =10 in T ∈ [0.139,0.229] ; ∆T = 0.002 ,
j3 =24 and j 4 =25 in T ∈ [0.082, 0.130] and T ∈ [0.232, 0.28] , respectively, where the
spectrum function changes from one to another; ∆T = 0.2 , j5 =14 in T ∈ [0.48, 3.08] . The
total number of the fitting points is j = j1 + j 2 + j3 + j 4 + j5 =81 (Figure 6).
4. Obtain the power spectrum at these fitting points based on Eq.6 and show it in Figure7.
5. Calculate the complex Fourier amplitude spectrum following Eqs.21 and 22, and present it in
Figure8.
-5
5.0x10 -3
2.0x10

-5
4.0x10
-3
F(ωk) [m/s]

1.5x10
G(ω)[m /s ]
3
2

-5
3.0x10

-3
1.0x10
-5
2.0x10

-4
1.0x10
-5 5.0x10

0.0 0.0
0 20 40 60 80 100 120 140
0 20 40 60 80 100 120 140 160
ω [rad]
ω [rad]

Figure 7 Initial power spectral density Figure 8 Initial amplitude spectrum

6.28

4.71
φ( f ) (0 - 2π)

3.14

1.57

0.00
0 2 4 6 8 10 12 14 16 18 20 22 24
f [Hz]
Figure 9 Generated phase angle spectrum for φ 0 = 0
6. Obtain the phase spectrum (Figure9) following the procedures presented in Eqs.7-13.
7. Operate IFFT to the obtained amplitude spectrum (Figure8) and phase spectrum (Figure9) to
generate the ground motion (Figure10).

115
 0.5
|accl|max=0

Response spectrum [m/s ]

0.08

2
0.4

0.04
0.3
accl [m/s ]
2

0.00
0.2

-0.04
0.1

-0.08
0.0
0.04 0.1 1 3
0 10 20 30 Period [s]

Figure 10 Initial ground motion Figure 11 Response spectrum of the initial ground
ti
Checking the response spectrum of the generated ground motion at the fitting points (Figure11).
8. Modulate the Fourier amplitude and the generated ground motion following the procedures
presented in Sect.4. As there is a big difference between Figure11 and the target spectrum
(Figure5), iteration calculation is needed. After a few times (no more than six) iteration, the
ground motion satisfying the tolerant error (5% in this example) is obtained (Figure12). The
comparison of the response spectrum of the generated ground motion to the target response
spectrum is presented in Figure13.

2
|accl|max=0.231m/s

0.2

0.1
accl [m/s ]
2

0.0

-0.1

-0.2

0 10 20 30 40 50 60 70 80
t [s]

Figure 12 Generated ground motion

116
 0.8
0.8

Response spectrum [m/s ]

2
Response spectrum [m/s ]
2

0.6
0.6

0.4 0.4

0.2 Target response spectrum 0.2 Target response spectrum

Fitted response spectrum Fitted response spectrum

0.0 0.0
0.04 0.1 1 3 0.04 0.1 1 3
Period [s] period [s]

Figure 13 Comparison of the fitted and the target Figure 14 Comparison of the fitted and the
response spectrum on the fitting points points target response spectrum on other

The fitting accuracy of the generated ground motion (Figure12) about another set of
fitting points is also investigated and it is shown that tolerant error may be not satisfied at a few
points but the error is not big, just about 7% (Figure14). It is believed that this error may satisfy
the engineering applications, higher accuracy can be reached by increasing the iteration steps if it
is needed. The further comparison of the fitting efficiency between Eq.20 and Eq.16 verifies that
Eq.23 is more efficient than Eq.16 as the generating procedure in the above example will be
divergent if Eq.16 is applied.

CONCLUSIONS

A new integrated algorithm is presented for the generation of artificial records which
comply with any target response spectrum and/or target power spectral density requirement and
possess the time and frequency nonstationarity typical of the real earthquake recordings. The
algorithm proposed in this work, in contrast to most of previous publications, applies the
component-by-component fitting technique which makes the response spectrum fitting
procedure more efficient and with high convergence speed.

117
ACKNOWLEDGEMENTS

This study was supported by the National Science Foundation of China under the Grant
No. 59708007 and the Teaching and Research Award Fund for Outstanding Young Teachers in
Higher Education Institute of MOE, PRC. The authors wish to thank Prof. Chen Yingjun for his
constructive comments on this work.

REFERENCES

Deodatis, G., 1996, “Generation of Ground Motion Time Histories as Non-stationary Vector Process: Response
Spectrum Compatible Seismograms”, Probabilistic Mechanics and Structural and Geotechnical Reliability,
Proceedings of the 7th Specialty Conference on Probabilistic Mechanics and Structural Reliability, Aug. 7-9,
Worcester, MA, USA, pp616-619.
Ghosh, A. K., 1993, “On the Generation of Ground Motion Accelerogram Compatible with a Specified Response
Spectrum and a Fourier Amplitude Spectrum”, Transactions of the 11th International Conference on Structural
Mechanics in Reactor Technology, 08/18-08/23/91, Tokyo.
Hu Y., He X., 1986, “Phase Angle Consideration in Generating Response Spectrum-Compatible Ground Motions”,
Earthquake Engineering and Engineering Vibration (in Chinese), 6(2), 37-51.
Karabalis, D. L., Cokkinides, G. J., Rizos, D. C., Mulliken, J. S., 2000, “Simulation of Earthquake Ground Motions by
a Deterministic Approach”, Advances in Engineering Software, 31(5), 328-338.
Kaul, M. K., 1978, “Spectrum-consistent Time History generation”, J. of Engineering Mechanics, ASCE, 104(4),
781-788.
Kaul, M. K., 1978, “Stochastic Characterization of Earthquakes Through Their Response Spectrum”, Earthquake
Engineering & Structural Dynamics, 6(5), 497-509.
Mignolet, M. P., Harish, M. V., 1996, Comparison of Some Simulation Algorithms on Basis of Distribution, Journal
of Engineering Mechanics, 122(2), 172-176.
Nigam, N. C., 1982, “Phase Properties of a Class of Random Processes”, Earthquake Engineering & Structural
Dynamics, 10(6), 711-717.
Preumont, A., 1984, “The Generation of Spectrum Compatible Accelerograms for the Design of Nuclear Power
Plants”, Earthquake Engineering & Structural Dynamics, 12(4), 481-497.
Sabetta, F., Pugliese, A., 1996, “Estimation of Response Spectra and Simulation of Nonstationary Earthquake Ground
Motions”, Bulletin of the Seismological Society of America, 86(2), 337.
Shinozuka, M., Sato, Y., 1967, “Simulation of Nonstationary Random Process”, J. of Engineering Mechanics
Division, ASCE, 93(1), 11-40.
Shrikhande, M., Gupta, V. K., 1996, “On Generating Ensemble of Design Spectrum Compatible Accelerograms”,
European Earthquake Engineering, Bolona, Italy, X(3), 49-56(1996).
Shaw, D. E., Rizzo, P. C., Shukla, D. C., 1975, “Proposed Guideline for Synthetic Accelerogram Generation Method”,
Proceedings of the 2nd International Conference on Structural Mechanics in Reactor Technology, American
Nuclear Society, Berlin Germany.
Yang Qingshan, Jiang Haipeng, 2001, “Generation of Fully Nonstaionary Ground Motions Based on Phase Difference
Spectra”, Earthquake Engineering and Engineering Vibration (in Chinese), 21(3), 10-16.
Zhao Fengxin, 1992, “The Phase Difference Spectrum of Time Histories and the Generation of the Design Ground
Motions”, Ph.D. Dissertation (in Chinese), Institute of Geophysics, China Seismology Bureau.

118
 The Seismic Retrofit of the
Golden Gate Bridge
Charles Seim P.E.

ABSTRACT

The construction of record-breaking suspension bridges in the 1920’s and 1930’s led up
to the construction of the Golden Gate Bridge that opened to traffic on May 28, 1937. The
bridge is owned, operated, and maintained by the Golden Gate Bridge Highway and
Transportation District (GGBHTD). The District has kept the bridge in first-class condition and
has upgraded the bridge on several occasions to improve its performance, to reduce maintenance
costs, and to provide longer bridge-life.

The bridge had served the City of San Francisco for over 50 years when the Loma Prieta
Earthquake struck in 1989. This seismic event jolted the public’s attention into awareness of the
seismic vulnerability of the San Francisco Bay Area bridges. The damage to other San Francisco
Bay Area bridges and structures, however, prompted the State of California to call for rigorous
seismic evaluation and retrofit of all transportation structures, including the five long-span
bridges crossing San Francisco Bay, and the Golden Gate Bridge.

The GGBHTD acted within a week after the Loma Prieta earthquake by engaging T. Y.
Lin International, San Francisco, to evaluate the seismic vulnerability of the Golden Gate
Bridge. An interim report produced in 1990 stated that the main suspension span was vulnerable
to damage, some of which could be irreparable, and that the approach structures were vulnerable
to severe damage with the possibility of collapse.

This paper presents the seismic retrofit design of the Golden Gate Bridge, providing a
case study in engineering methodology that can be applied to the seismic retrofit of other major
long span suspension bridges. The methodology begins with an owner-developed performance
criteria followed by an engineer-developed design criteria to match the performance criteria.
The next steps are performing a seismic hazard analysis and developing a site-specific response
spectrum and at least three independent ground motions, for short return-period events as well as
for long-return period events. The last steps, which take the most time and are the most difficult,
are performing elastic and inelastic analyses; developing, designing, preparing plans; and the
writing of the construction specifications for the seismic retrofit measures.

__________
Vice President, T. Y. Lin International, 825 Battery St., San Francisco, CA., and 94111, USA
Phone 415 291 3771, Fax 415 433 0807, [email protected]

119
INTRODUCTION

One of the most famous, historic, and enduring structural achievements in the world, the
Golden Gate Bridge stands as a symbol of the City of San Francisco, California, and as a
monument to the achievement of the bridge engineering profession. The start of its construction
in January, 1933, culminated nearly a decade of bridge construction in the United States that
extended the world record for suspension bridge span lengths four times.

In 1924 the Bear Mountain Suspension Bridge across the Hudson River near West Point,
with a main span of 497.6 m (1632 ft), was completed and took the world record away from the
Williamsburg Bridge (487.8 m, 1903) in New York City by only 9.8 m (32ft). The Ben Franklin
Suspension Bridge, spanning the Delaware River at 533.5 m (1750 ft) near Philadelphia,
Pennsylvania took the record two years later in 1926. In 1929, the Ambassador Suspension
Bridge at Detroit, Michigan took the record for suspension bridge span-length and at 564 m
(1850 ft) it surpassed the 548.8 m (1800 ft) Quebec steel cantilever truss to take the record as the
longest span in the world.

However, the George Washington Suspension Bridge in New York City smashed all of
these records by almost doubling the Ambassador span-length with a 1,067 m (3500 ft) leap of
the Hudson River in 1931.

The Golden Gate Bridge opened to traffic on May 28, 1937 with a modest 213 m (700 ft)
increase in span-length (to 1280 m) over the George Washington Bridge. The Golden Gate
Bridge then held the title of the world’s longest bridge for 27 years, when it lost the record by
only 18 m to the Verrazzano-Narrows Bridge in New York. Today the Golden Gate Bridge has
slipped into seventh place behind the now-longest span in the World, the Akashi-Kaikyo Bridge
in Japan and the third place Jiangyin Bridge in China. China is now embarking on its own
record-span setting explosion.

Fig. 1 The Golden Gate Bridge in 1937

120
 The bridge is owned, operated, and maintained by the Golden Gate Bridge Highway and
Transportation District (GGBHTD). The District has kept the bridge in first-class condition and
has upgraded the bridge on several occasions to improve its performance, to reduce maintenance
costs, and to provide longer bridge life. In 1954, a bottom lateral-bracing system was added to
the bridge to improve its performance during high winds. In the mid 1970s, all of the corroding
vertical hanger cables were replaced. In 1985, a lightweight orthotropic steel deck, paved with
epoxy asphalt, replaced the original concrete deck, which was beginning to suffer chloride ion
corrosion. The need for maintenance and upgrading of the bridge continues as new challenges to
its structural integrity are discovered.

The public’s attention was jolted into awareness of the seismic vulnerability of the San
Francisco Bay Area bridges by the October 18, 1989 Loma Prieta earthquake. This moderate
Magnitude 7.1 and 100-km distant earthquake on the San Andreas Fault did not damage the
Golden Gate Bridge. The nearest seismograph indicated a peak ground-acceleration of 0.24 g,
about three times the value of 0.075 g, used for the original 1930 design.

The damage to other San Francisco Bay Area bridges and structures, however, prompted
the State of California to call for rigorous seismic evaluation and retrofit of all transportation
structures including five long-span bridges crossing San Francisco Bay. These included: the San
Mateo-Hayward Bridge steel box girder (228.7 m, 1967); the San Francisco-Oakland Bay Bridge
suspension spans (2 at 704 m, 1936); the Richmond-San Rafael Bridge steel cantilever truss (2 at
326 m, 1956); the Carquinez Strait Bridge steel cantilever trusses (335.4 m, 1927 and 1958); and
the Benica-Matrtinez Bridge continuous steel truss (190.5 m, 1963).

Seismic evaluation and retrofit of two bridges outside of the San Francisco Bay Area
were also completed. These were the San Pedro-Terminal Island Suspension Bridge near Los
Angeles (457 m, 1961) and the San Diego-Coronado Bay Bridge steel box girder (201 m, 1969).

The GGBHTD acted within a week after the Loma Prieta earthquake by engaging T. Y.
Lin International, San Francisco, to evaluate the seismic vulnerability of the Golden Gate Bridge.
An interim report produced in 1990 stated that the main suspension span was vulnerable to
damage, some of which could be irreparable, and that the approach structures were vulnerable to
severe damage with the possibility of collapse. In 1993 The GGBHTD engaged T. Y. Lin
International in Joint Venture with Imbsen & Associates, Inc., to design the seismic retrofit of
the main suspension span and the North approach structures. Another consultant was engaged to
design the seismic retrofit of the South approaches.

This paper will focus on the development of the design for the seismic retrofit of the
suspension span, which was the first seismic retrofit design of a major long span bridge in the
world. This unprecedented achievement is an excellent Case Study in applying the existing
technology in earthquake engineering and the technology developed since the earthquake. The
technologies used included performing seismic hazard analysis and developing ground motions;
adapting computer nonlinear-structural programming for modeling the suspension bridge; and
developing innovative seismic retrofitting concepts and devices to strengthen or modify the
response of the structure. Perhaps this interesting story may help others in some small way in

121
their quest of the seismic retrofitting of suspension bridges. These retrofit principles can and
should also be applied to the design of new long-span bridges.

BRIDGE DESCRIPTION

Although most people see the Golden Gate Bridge as a single structure, the 2790 m
overall length of the bridge actually consists of five different structure types. The bridge’s major
components are the North and South steel-truss approach viaducts, the Fort Point steel arch, the
arch's flanking concrete pylons that are purely architectural motifs, the two concrete cable-
anchorages housings, and the main-span steel suspension bridge. All of the foundations for these
structures are supported directly on rock, except for the northern viaduct, which is founded on
spread-footing bearings on competent soil.

The three-span 1,966-m suspension bridge has a 1280-m center span and two 343-m side
spans. The 227-m high towers, each with two shafts of multicellular steel plate configuration,
are tapered in steps from the base to the top, which supports cast-steel saddles. Above the
roadway the shafts are connected together with four struts forming portals. Below the roadway
they are braced together with two sets of cross bracing. At their bases, the tower shafts do not
have anchor bolts but depend on their own weight and the restraint provided by the cables for
anchorage.

The 0.92-m diameter cables are spaced 27.44- m apart with a cable sag-to-span ration of
1-to-9. At the shoreward end of each side span, the cables pass through concrete pylons, where
the cables are restrained vertically at the roadway level by steel cable tie-downs. The stiffening
trusses are 7.7-m deep and are spaced 27.44- m apart in the same plane as the cables. The
trusses are connected at the top with lateral cross bracing, which was part of the original design.
A lightweight bottom lateral cross bracing was added as a wind retrofit in 1954 after the bridge
suffered damage from strong winds. This cross bracing was designed for extreme lightness to
keep the added stress in the cable to the lowest amount possible.

Fig. 2 The Golden Gate Bridge under construction in 1936

122
 The originally 6-lane, 18.29-m wide concrete upper deck was replaced in 1985 with an
18.90-m lightweight orthotropic steel deck because of corroding deck-reinforcing bars. The
lighter weight orthotropic steel deck lowered the stress in the cables back to about the original
stress level before the addition of the lower lateral cross bracing.

The performance of a long-span bridge during, and the condition of that bridge, after a
predicted earthquake, is primarily the owner’s responsibility to determine in conjunction with the
design engineer. The GGBHTD determined that the GGB is a non-redundant bridge, as there is
no alternative route from San Francisco to the North. The Bridge is on a designated lifeline
corridor and the bridge should be opened immediately after an earthquake to emergency
vehicles.

In California, the performance of bridges is generally specified for two levels of

earthquakes that are based on the return periods of the seismic events. A return period from 100
to 475 years is generally specified for an FEE, a functional-evaluation earthquake. For these
smaller FEE events, the seismic performance should be nearly full-use of the facility with little
damage requiring repairs after the event.

A return period of 1000 to 2000 years is generally used for a SEE, a safety-evaluation
earthquake. For these large SEE events, the specified performance can range from no-collapse
with non-repairable damage (complete replacement of the bridge required), to repairable
damage, or as for the GGB, to nearly full-function of the bridge.

For the GGB, the lifeline determination required that the Bridge performance in a SEE be
essentially elastic with either no- or quickly-repairable damage. The bridge should be open to
emergency vehicles within hours and to traffic within days. At this high performance level for
the SEE event, the GGB will perform with no damage in an FEE event.

According to the United States Historic Preservation Act of 1966 and the Secretary of the
Interior’s Standards for Rehabilitation, special consideration must be given to any changes to an
older bridge that may affect the defining characteristics of the structure. For the GGB, the
seismic retrofit measures developed to upgrade the seismic performance of the Bridge followed
hierarchical guidelines. The first priority is to meet the seismic retrofit design criteria for the
safety of the structure and the traveling public. The next lower priority is to maintain the
architectural appearance of the bridge and to follow, as much as possible, the guidelines of the
U.S. Historic Preservation Act. Finally, the seismic retrofit measures should retain as much of
the original material that formed the structure as possible.

SEISMIC HAZARD AND GROUND MOTIONS

The Golden Gate Bridge spans the mouth of the San Francisco Bay, which was formed
by down-cutting rivers during a previous glacial period. The foundation material consists of
sandstone, shale, greenstone, and chert under the North Tower, and serpentine and melange
under the South Tower. Shear-wave velocities have been measured in the shale, greenstone, and
chert, ranging from 1500 m/sec to 2400 m/sec, and in the serpentine, about 1200 m/sec.

123
 The San Andreas Fault lies 10 km West of and parallel to the Bridge and the fault
became famous when it caused the San Francisco Earthquake of 1906. The 1000-2000 year
return period on the San Andreas Fault's is a Magnitude 8.3 event, comparable to the 1906 San
Francisco earthquake. The Hayward Fault is about 16 km East and is also in a direction parallel
to the Bridge. The 1000 to 2000-year event on the Hayward Fault is a Magnitude 7.0 event.
However, ground motions from this fault did not control any of the seismic retrofit decisions.

A site-specific target response spectrum was developed based on an 84 percentile, equal

probability, 5% damped spectrum and is representative of a 2,000-year return period earthquake
on the San Andreas Fault. Three independent design-ground motions were developed, based on
recorded events, and were made compatible with the target spectrum by adjustment of their
Fourier amplitudes. The compatible ground motions have peak ground-accelerations of about
0.65 g, peak velocities of about 115 cm/sec, peak displacements of about 60 cm, and durations of
60 to 90 seconds. The ground motions were developed by including possible fault-rupture
scenarios, wave-passage effects, extended-source effects, and the effect of ray-path incoherency.

A study was made of the response of the bridge to multiple-support excitation versus the
response to rigid base excitation. The only systematic trend observed in the study was that
vertical displacements of the stiffening trusses were larger for multiple-support excitation,
probably because differential movement between the bridge supports straightens the cables and
lifts the spans. In other respects, the differences in the response to multiple-support and rigid
base excitation were small and somewhat random over the three design earthquakes.

DESIGN CRITERIA

At the start of design, no design documents existed for the seismic retrofit design of long
span bridges so the design team had to develop a Design Criteria. It was developed as a guide
for the designers and was printed in loose-leaf notebooks so that revisions could easily be
inserted. The technical issues that the Design Criteria addressed are based on meeting the
Performance Criteria noted above. The seismic retrofit design is based on inelastic analysis of
the bridge. Therefore the criteria limit the displacement ductility demands on bridge members.
Limited repairable-damage that does not threaten structural safety and that can be repaired
without interrupting traffic is acceptable. These Criteria will allow portions of the bridge to
respond to limited inelastic action, but the primary response is to be essentially elastic wherever
possible.

Compression members carrying gravity loads are required to remain elastic. Width-to-
thickness ratios of steel plates in new members are limited to insure compactness and local
ductility. Bracing for buckling is required for most existing members to obtain global ductility.
The nominal strength of bolts or rivets in joints must be at least 25 percent greater than the
nominal strength of the member. For members that respond elastically to ground motions, the
member must be designed to ensure good ductility. Where inelastic response is permitted, the
member must be part of a redundant system and the member must be designed for large ductile
response values. Because of the non-redundancy of the cable system, inelastic deformations are

124
prohibited. Fortunately, the cable system responded elastically without any retrofitting
measures.

STRUCTURAL MODELING

The computer structural modeling was performed in incremental steps, each step adding
to the engineers' understanding of how the bridge would respond in a large earthquake. The first
step was to determine the dead-load state starting with the geometry and dead load at the
completion of the original construction, which determined the initial stresses. To this step was
superimposed the addition of the bottom lateral-bracing system and the removal and replacement
of the deck. The deformations and stresses of this dead-load state were computed by
geometrically nonlinear-analysis, with large displacement theory, and served as the initial
condition for subsequent dynamic analysis.

Performing a modal analysis was the next step for determining vibration properties. The
natural frequencies and modal shapes were compared with values obtained from ambient
vibration tests on the bridge. The good agreement with the first eight modes verified the
computer model. It is always a good idea, in the dynamic analysis of existing bridges, to have
measured vibration frequencies with which to compare calculated values. The next step was to
perform a linear-response spectrum analysis using the site-specific target response spectrum.
This analysis gave the engineers an initial insight into the magnitudes of seismic forces and
displacements and pointed out problem areas and components for detailed modeling and study.

The last step was performed after these learning stages gave the engineers a good
understanding of how the bridge was responding linearly. That last step used a dynamic
nonlinear finite element computer program that solved the integrated coupled equations of
motion in the time domain with multi-support excitation inputs. The ground motions at the six
primary supports were applied as a time-varying displacement boundary condition. Large
displacement effects were considered by establishing static or dynamic equilibrium of the
structure in its deformed configuration. The effects of limited displacement capacity at
expansion joints and tower uplift were modeled by gap elements.

Detailed three-dimensional finite element local models were used to study the stress
distribution, relative displacements, uplift, and moment-rotational relationships. These local
models were use for detailed study of the tower base uplift behavior, tower struts, and the
stiffening truss-tower interaction.

SEISMIC RETROFIT MEASURES

The design teams developed seismic retrofit measures for all of the vulnerable
components in the bridge that either would suffer damage or would not develop ductile action.
The seismic retrofits of these areas are presented below.

125
Tower Piers

The towers of the bridge are supported on reinforced concrete piers extending down to
bedrock. Although the piers are massive and very stable, they will be severely loaded when the
unanchored steel towers rock during an earthquake.

The retrofit of the piers is to drill in, grout, and post-tension high-strength threaded bars
from the surface of the pier and under the footprint of the tower base. These bars will prevent a
shear failure along the critical failure plane in the pier, which is beneath the loaded edge of the
tower.

Tower Bases

The towers are of multi-cellular construction and consist of plates riveted together with
corner angles. At the base, the cross-section consists of 103 cells, each 1.07 m×1.07m square that
are just large enough to work inside. The plates are 22 mm thick, giving a width-to-thickness
ratio of 48. Plates of this shape will buckle shortly after yielding. A finite-element analysis of a
cell showed the corner angles to be only minimally effective in restraining buckling of the plates.

The towers will rock during an earthquake; the magnitude of the uplift is about 60 mm at
the extreme fibers of the base. Fixing the bases of the towers caused higher stresses than
allowing uplift of the towers, and anchoring the tower bases would be very difficult. The uplift
causes concentrations of stress both at the base and above the first setback in the tower. The
peak-strains were about four times the yield-strain, assuming elastic-plastic behavior. Strains of
this magnitude can be accommodated by compact sections, but not by the non-compact sections
used in the original construction of the tower bases. The ductile behavior of the tower base will
be improved by adding retrofit stiffeners along the vertical centerlines of the plates at the base
between the horizontal diaphragms. The added stiffeners will prevent buckling of the plates until
after a displacement ductility of four is reached.

Hydraulic Dampers

Hydraulic viscous dampers will be will be installed between the stiffening trusses and the
towers to absorb energy, to reduce seismic forces of the stiffening truss impacting the tower, and
to control displacements at this critical location. These dampers are used because they will not
restrain the thermal and traffic-induced movements of the bridge, and because they can be built
with the large capacity needed. T he damper design is simply a piston moving in a cylinder that
displaces a viscous fluid passing through an orifice from one side of the piston to the other.

Based on an optimization study, dampers were chosen for the retrofit with a force-
relationship of:

F = ( 350 to 800 kip ⋅ sec1/2 / in 1/2 ) ⋅ V 1/ 2 (English Units)

126
 At a peak velocity of 180 cm/sec, the dampers will produce a control-peak force of
26,000 kN between the stiffening trusses and the towers.

The dampers dramatically reduce the displacement demands on the bridge wind-locks
and expansion joints, and eliminate actual impact between the stiffening trusses and the towers.
They also reduce the peak stresses in the stiffening truss chords and the towers, and reduce the
tower base shear-forces and uplift. The retrofit design also includes modifications to the bridge
wind-locks to increase their displacement capacity, in order to eliminate impact within the wind-
locks.

Lateral Bracing

The top and bottom lateral-bracing systems are over-stressed by about 50% in both
tension and compression. Because of the contribution of higher modes of vibration to the
response of the bridge, the overstress occurs over a large proportion of the length of the bridge,
and for a large percentage of members. The existing braces are of non-ductile construction;
consisting of four angles laced together into a box configuration. A finite-element analysis of a
typical lateral brace showed that the corner angles of the brace buckled locally at an overall
ductility-demand of only 1.15, which is the limit of usefulness of the member. Rapid degradation
of the strength and stiffness occur after the local buckling. An inelastic time-history analysis
showed that the deformation demands on the lateral braces were concentrated into those
members, which yielded or buckled first. The peak ductility demands from the inelastic analysis
were more than five times in excess of the design criteria limit of two.

The retrofit consists of replacing one-half of the top lateral braces with new members.
These will be ductile, compact members of tubular cross-section. The lateral bracing systems
are the primary means of resistance of the bridge, both to aftershocks and to wind, and these
loads must be provided for as part of the retrofit design.

CONCLUSIONS

The seismic retrofit design of the Golden Gate Bridge has provided a Case Study in
engineering methodology that can be applied to the seismic retrofit of other major long span
suspension bridges. The methodology begins with an owner-developed performance criteria
followed by an engineer-developed design criteria to match the performance criteria. The next
steps are performing a seismic hazard analysis and developing a site-specific response spectrum
and at least three independent ground motions, for short return-period events as well as for long-
return period events. The last steps, which take the most time and are the most difficult, are the
performing elastic and inelastic analyses, developing, designing, preparing plans, and the writing
of the construction specifications for the seismic retrofit measures.

Acknowledgments

The authors wish to acknowledge the Bridge Owner, the Golden Gate Bridge, Highway
and Transportation District; the General Manager, Celia G. Kuppersmith; the current District
Engineer, Denis Mulligan; and District Staff Engineer, Jerry Kao who acted as liaison for the

127
Project. Also acknowledged is the help of former District Engineers Dan Mohn, who started the
evaluation and seismic retrofit process, and Merv Giacomini, who continued the work. Also Tim
Ingham PhD., PE, Associate with T. Y. Lin International, who directed the analyses and seismic
retrofit design reported in this paper.

References
T. Y. Lin International, 1990, “Golden Gate Bridge Seismic Evaluation,” Golden Gate Bridge, Highway and
Transportation District, San Francisco, CA, November.
T. Y. Lin International, 1991, “Golden Gate Bridge Seismic Retrofit Studies,” Golden Gate Bridge, Highway and
Transportation District, San Francisco, CA, July.
Geospectra Incorporated, 1992, “Geological, Geotechnical and Ground Motion Studies for Seismic Retrofit of the
Golden Gate Bridge,” Golden Gate Bridge, Highway and Transportation District, San Francisco, CA, July.
T. Y. Lin International, 1992, “Golden Gate Bridge Design Criteria for Seismic Retrofit Measures,” Golden Gate
Bridge, Highway and Transportation District, San Francisco, CA, November.
Rodriguez, S., Seim, C., and Ingham, T. J., 1994, “Earthquake Protective Systems for the Seismic Upgrade of the
Golden Gate Bridge,” Proceedings of the Third U.S.-Japan Workshop on Earthquake Protective Systems for
Bridges, NCEER-94-0009, March.
Ingham, T., Rodriguez, S., Nader, M., Taucer, F., and Seim, C., 1994, “Golden Gate Bridge Seismic Retrofit
Design, Suspension Bridge Strategy Report,” Golden Gate Bridge, Highway and Transportation District, San
Francisco, CA, April.
Nader, M.N., and Ingham, T.J., 1995, “Seismic Retrofit of the Towers of the Golden Gate Bridge,” Proceedings of
the National Seismic Conference on Bridges and Highways, San Diego, CA, December.
Rodriguez, S., and Ingham, T.J., 1995, “Seismic Protective Systems for the Stiffening Truss of the Golden Gate
Bridge,” Proceedings of the National Seismic Conference on Bridges and Highways, San Diego, CA,
December.

128
 Response of Seismic Isolated Bridges
Using M-DOF Model and 2D Excitation

George C. Lee and Zach Liang

ABSTRACT

This paper discusses two important issues in modeling the seismic responses of
structures: the number of degree of freedom and the directions of excitation. By using a simple
M-DOF bridge model with two directional excitations, it is showing that the currently used
design spectrum method cannot always provide sufficiently accurate results for seismic design of
bridges. Rather, further studies should be carried out to develop or to refine design guidelines
based on dynamic analysis with multiple directional excitations, particularly for those bridges
with added response modification devices/systems.
The paper is prepared for discussion by the workshop participants. It is essentially a
progress report of an on-going research task at MCEER sponsored by the U.S. Federal Highway
Administration on development of retrofit strategies for special highway bridges.

INTRODUCTION

Maximum peak responses of structures subjected to earthquake ground motion are

important information in aseismic design, especially when seismic response modification
technologies are used to improve the structural performance. For M-DOF bridges with cross
effects introduced by geometrical and modal coupling as well as multidirectional ground
excitations, the response can be quite different from the results based on the currently used
design spectrum method, which strictly speaking applies only to bridges that can be
approximately by S-DOF models. In the latter cases, the response level can be obtained
deterministically by a mathematical model with given ground motion record using time history
analysis.
When the bridge is equipped with seismic isolation bearing or other types of response
reduction devices, the peak value of responses under random excitation cannot be adequately
estimated for the bridge-device systems that should be represented by M-DOF systems.
The basic approach to modify earthquake responses of a structure is to increase the
dynamic stiffness of the system (combined structure-device system). Increasing the dynamic
stiffness, however, is much more complex than increasing the static stiffness. The generic
concept of increasing dynamic stiffness is the reduction of a given structural response under a
given level of excitation. The responses can be acceleration, displacement, base shear, or other
parameters. When the dynamic stiffness is increased for acceleration reduction, the dynamic
__________________
George C. Lee, Multidisciplinary Center for Earthquake Engineering, University at Buffalo, Red
Jacket Quad, Buffalo, NY 14261, USA
Zach Liang, Multidisciplinary Center for Earthquake Engineering, University at Buffalo, Ketter
Hall, Buffalo, NY 14261, USA

129
stiffness for displacement may be decreased. Further, the dynamic stiffness of acceleration
reduction achieved in a certain frequency range may not necessarily mean the reduction of
acceleration will take place in all the other frequency ranges (Clough and Penzien, 1993).
Traditionally, the main goal of bridge seismic isolation design is to reduce the
acceleration of the deck or the superstructure. This concept can be shown in figure 1. In this
figure, the required acceleration level is represented by the dotted line. Originally, a given
bridge has a period T1 with small damping (with damping ratio ξs ) and an acceleration level of
A1, which is greater than the required level. By adding seismic isolation, the period is shifted to
T2, but then the acceleration level is still greater than the required value. Thus, additional
damping is used with damping ratio ξl , so that the acceleration level is reduced to A3 and the
design parameters of T2 and ξl are determined (AASHTO, 2001; ICBO, 2000; UBC, 1997).

Acc. Level

A1
Small Damping, ξs
A2
Required Level
A3 Large Damping, ξl

Period
T1 T2

Figure 1. Seismic Isolation Design Concept

In this approach, the corresponding dynamic stiffness of the structure is increased.

However, this is achieved by reducing the static stiffness of the bearing and resulting in a drastic
reduction of the dynamic stiffness of the displacement. The dilemma is that in order to increase
the dynamic stiffness for acceleration reduction, the bearing stiffness has to be very small; but
the smaller the bearing stiffness is, the weaker the dynamic stiffness for the displacement will be.
This logic further leads to adding damping to reduce the bearing displacement. The concept
described in Figure 1 implies that increasing damping can be beneficial for both acceleration and
displacement reduction.
This logic can be validated by using time history analysis. Consider the following
numerical example:

Suppose the supper structure of a bridge can be modeled by a single mass of 1,000,000
kg and the stiffness of the bearing is 10,000 kN/m, shown in Figure 2(a). The damping ratio is
16.4 % in direction X and 19.1% in direction Y. The isolation system then has the natural period
1.99 seconds in both directions. An earthquake recorded in Llollelo, Chile on March 3, 1985 will
be used for the analysis. In a latter section, the reason for choosing this special record will be
given.
The Llollelo earthquake was recoded at 100 degree, whose maximum level is 0.71g, as
the X directional input and recorded at 10 degree, whose maximum level is 0.45g, as the Y
directional input. The results of the time history analysis show that the maximum bearing
displacements are about 10.09 cm and 5.50 cm in the X and Y directions, respectively. The

130
accelerations of the deck are 0.13g and 0.06g in the X and Y directions, respectively. When the
damping ratio is doubled, the results are also obtained. They are given in Table 1.

Y
y’

x’

(a) S-DOF Model for

Seismic Isolation

Y
y’
(b) M-DOF X
Model for Seismic Isolation x’

Figure 2. Conceptual Models for Seismic Isolation of a Bridge

TABLE I. COMPARISON OF RESPONSES OF SDOF MODELS

WITH DIFFERENT DAMPING
X disp. (cm) X acc. (g) Y disp. (cm) Y acc. (g)
Original damping ratio 10.09 0.135 5.50 0.06
Damping ratio doubled 7.89 0.144 4.3 0.08
Change -27.6% 6.9% -21.7% 18.7%

These results show that, when the damping ratio is doubled, the bearing displacements
are reduced. However, the deck accelerations are increased, instead of reduced. The above only
illustrates a problem in response estimation using the S-DOF approach. The fundamental aspects
in this process are being investigated currently at MCEER under the research contract of the US
Federal Highway Administration. Two aspects of this current research will be briefly described
in the sections that follow.

EARTHQUAKE GROUND MOTIONS

A 3D ground motion time-history can be completely represented by any 3D Cartesian

coordinates. In practice, ground motions are often measured in three fixed directions: north-
south, east-west and up-down. For a given structure, a ground motion represented by different
coordinates will result in different peak value of the responses. The original record picked up by
instruments oriented in the fixed directions therefore may not yield the maximum responses. In
order words, the earthquake incident angle is an important parameter that affects the seismic
ground records. This issue therefore is how to select the coordinate that can provide the peak
value of the response.

131
 Penzien and Watabe (1975) defined the principal axes of ground motions as a set of
mutually perpendicular axes about which cross-correlation along such axes are zero-valued.
Their original idea is to identify the direction of the direction from the recording location to the
epicenter. However, this concept was subsequently used by many researchers, as a reference to
select the earthquake incident in the horizontal plan. If a pair of principal axes is in the horizontal
plan, then the vertical direction automatically defines the third axis. It is noted that, if only the
horizontal plane is considered, such pair of axes always exists. Therefore, the definition of
principal axes in the horizontal is legitimate.
In many of the existing earthquake records, the vertical component is relatively small. In
this case, the cross correlation between the horizontal components and the vertical one can be
quite small. Therefore, the above-mentioned horizontal principal axes can automatically become
a 3D set of principal axes under the definition of Penzien and Watabe. However, there are other
records that have large vertical components and the correlation function of the horizontal and
vertical signals may not be taken as zero.
Let x, y and z denote the two arbitrary horizontal and the vertical directions (see figure 2)
and Rxy denote the correlation function of the signal along x and y axes. We can establish a 2D
and a 3D covariance matrix as follows:

 R xx R yx 
h= and (1a)
 R xy R yy 

 R xx R yx R zx 
 
H =  R xy R yy R zy  (1b)
 R xz R yz R zz 

Where Rxy is the correlation function of signals measured from x and y directions, at time zero,
and so on. Both h and H are symmetric. Orthogonal eigenvectors are available to decouple the
symmetric matrices. For the 2D matrix, its eigenvector matrix can always be written in the
following form:

 cos( θ ) − sin( θ ) 
 sin( θ ) cos( θ ) 
(2a)


that indicates that turning an angle θ from the original Cartesian coordinates, the newly formed
correlation matrix h’ will have its off-diagonal entry equal to zero. This is the definition of the
principal direction.
Returning now to the numerical example of the M-DOF bridge model and examine the
concept of the principal axes of ground motion. First of all, the cross-correlation function of the
accelerations is equal to –7.66. Recall the maximum values of the Llollelo earthquake ground
acceleration are 0.71g and 0.45g. The correlation function seems to have a large value and not
close to zero. From equation (2a), if the angle θ is chosen to be 4.020, zero-valued correlation
functions can be obtained. Denoting the new direction as x’ and y’ as shown in Figure 1, the new
excitations yield the responses as shown in Table II

132
 TABLE II. COMPARISON OF RESPONSES OF A S-DOF MODEL
WITH DIFFERENT SEISMIC INCIDENT
X disp. (cm) X acc. (g) Y disp. (cm) Y acc. (g)
Original incident 10.09 0.135 5.50 0.062
Principal axes 10.73 0.131 5.28 0.064
Change -1.6% -4.0% -1.5% -4.1%

Table II shows that the input along the principal axes does not produce larger responses.
Instead, the responses become smaller. This fact motivated us to examine further the details
about the covariance matrix.
In order to decouple the 3D correlation matrix and to force the two horizontal correlation
functions zero-valued, the corresponding eigenvector matrix must have the following form:

cos(θ ) − sin(θ ) 0
 sin(θ ) cos(θ ) 0
 
 0 0 1

This requirement may not be satisfied by many earthquake records. Therefore, to use 3D
ground motions, the first problem is how to define the principal axes in the horizontal plan.
Consider the Llollelo earthquake, without the vertical acceleration, the 2x2 matrix is
given by

 1.9007 − 0.0767  0.9973 0.0734

h = 108   with its eigenvector matrix equal to  
 − 0.0767 0.8643   − 0.0734 0.9973

With the vertical data, however, the 3x3 matrix is

 1.9007 − 0.0767 0.4551 

H = 10  − 0.0767 0.8643 − 0.1117 , with its eigenvector matrix equal to
8
 
 0.4551 − 0.1117 1.2460 

 0.8836 0.4621 − 0.0752

 − 0.0929 0.3305 0.9392 
 
 0.4589 − 0.8229 0.3350 

From the above matrix, we can no longer extract the angle 4.020 corresponding to zero-
valued correlation functions. The corresponding angle for this case actually will be much larger
than 40.
López and his coworkers (1997, 2002) have shown that, even in a horizontal plan,
seismic incident along the above-defined principal axes will not yield the maximum peak value.
In their study, existence of critical incident angle is pointed out, which is measured from the

133
above-mentioned principal axes. Because their formulation is based on the design spectrum, no
information is provided about the variation of peak responses under different excitations with
varying seismic incident. The current MCEER study to-date using time history analysis has
shown that such a exact variation can be quite large (Yang, 2002; Lee and Liang, 2002 in print).
The maximum peak values can be several times larger than the minimum values. This implies
that, if a bridge is designed based on the original records or the record projections along the
principal axes defined by Penzein and Watabe (1975), the responses of the bridge may be several
times smaller than the actual maximum values.
To substantiate this point, results of the time history analysis using the S-DOF model of
the same example are given in Table III.

TABLE III. COMPARISONS OF RESPONSES UNDER EXCITATIONS

WITH DIFFERENT SEISMIC INCIDENCE
X disp. (cm) X acc. (g) Y disp. (cm) Y acc. (g)
Critical incident 12.18 0.142 11.28 0.141
Principal axes 10.73 0.133 5.28 0.062
Change 5.16% 5.8% 130.4% 128.8%

It is seen that more than 100% error exists.

The important conclusion of the above analysis is that the dynamic stiffness is differently
related to various response parameters. These response parameters can have their maximum
values at different incident angles. In other words, a given structure under a given earthquake
may have many critical seismic incidents.
In the next section, the effect of seismic incident angle will be given by using the same
numerical example.

DIRECTIONAL CROSS EFFECT AND MODAL COUPLING OF DYNAMICALLY

LOADED STRUCTURES

If a structure can be completely decoupled in individual modes, then using model

superposition (the S-DOF approach) will not introduce errors. If a structure does not has
responses due to perpendicular excitations, then we can compute the response in the two
individual directions. To classify the responses in three or two mutually perpendicular directions
is a useful approach, because seismic ground motion along a single direction of a building causes
responses at different floors of the building along the same direction. However, such idealized
consideration does not exist in real structures. It is, therefore, important to understand the cross
effects on the responses and to determine their magnitude.
To understand cross effects, the first question to consider is whether a given structure has
principal axes under dynamic load. If not, then a load acting along one axis will cause responses
in the perpendicular directions. In this case, directly estimate of the peak response by
superposition method is difficult because the relationship between the amplitude of excitation
and the peak response is non-linear.
The nonexistence of principal axes for dynamic loaded structures has been shown by the
authors (Lee and Liang, 1998). Assume that a structure consists two sub-structures and each of
them has its own principal axes. If these two pairs of principal axes are not in the identical
directions, then the total structure will not have principal axes. Experimentally, the authors have

134
quantitatively validated the theory with the shaking table using a 3 story steel frame structure.
The ground motion is single directional. The structure was tested in many positions, by turning it
from 00 to 1000 around a vertical axis. The experimentally observed cross effects are the
responses in the perpendicular direction of the single direction excitation (Liang and Lee, 2002;
Yang et al, 2002).
This cross effect is not only caused by structural rotation (torsional effect). Both
theoretical and experimental studies show that, without rotation, cross effect still exists and our
experimental results show that more than 50% enlargement of the responses can be easily
realized due to cross effect without detectable rotating of the test structure.
To substantiate the above, the previous numerical example will be again utilized. Figure
2(b) shows the possibility of adding more DOFs to the model if the column is flexible.
Suppose, two more DOFs in each directions are added with the corresponding modal
parameters listed in Table IV.

TABLE IV. MODAL PARAMETERS OF MDOF MODEL BRIDGE

Mode 6 5 4 3 2 1
Natural Frequency 46.6156 37.8707 22.5061 27.4108 0.5585 0.4486
(HZ)
Damping Ratio 0.0160 0.0556 0.0678 0.0356 0.3438 0.3656

Recall in our numerical example when the isolation system is modeled by S-DOF in each
direction, with the same damping coefficient, the damping ratios are 32.8 % in X direction and
38.2% in Y direction. If we have a more flexible pier and add more DOFs, the corresponding
damping ratios change slightly and have the values of 34.4% and 36.6%, respectively.
First, by ignoring the cross effect and by inputting the excitation in the X and Y direction
independently, the difference of results between M-DOF and S-DOF models are listed in Table
V.
Next, we consider the cross effect. In this case, the input contains the two perpendicular
excitations. The results are also given in Table V for the purpose of comparison.

TABLE V. COMPARISON BETWEEN SDOF AND MDOF MODEL

Y base
X disp. Y disp. X bases shear
(cm) X acc. (g) (cm) Y acc. (g) hear (kN) (kN)
SDOF (principal axes) 4.54 0.144 7.72 0.081 454 772
SDOF (critical incident) 8.47 0.145 9.09 0.145 847 909
MDO (single excitation) 8.06 0.146 9.12 0.141 1,016 884
MDOF (with cross effect) 7.83 0.147 9.43 0.151 1,007 900

From Table V, it is seen that different models can yield quite different results. Bearing
displacement is one of the most critical parameters. By means of the MOF model with the
consideration of slight cross effect as well as the critical seismic incident, we can compute the
largest bearing displacement as large as 9.4 cm.
More detailed comparisons can be made by examining Figure 3, where the comparison of
the base shear by using the S-DOF and M-DOF models is shown. The seismic incident angles

135
are chosen from 00 and 1800. It is seen that, different models and different directions will have
their own critical incident angles. The maximum and minimum values changes drastically. For
example, the maximum base shear of the M-DOF model in the X and Y direction are 68% and
91%, respectively, larger than that of the minimum value, whereas the factors of S-DOF model
are different.

Figure 3. Comparisons of the base shear M-DOF vs. S-DOF

The results of bearing displacement by using single and 2D excitations are shown in
Figure 4. The differences are measures of the cross effect. It is seen that the critical incident
angles between the single and 2D excitations are different. The maximum of the bearing
displacement with concern with the cross effect is 130% larger than that of the minimum value in
the X direction. Without consider the cross effect, it is 100% larger that value in Y directions.
About 30% different is visualized. It is noted that, in the example and with the Llollelo
earthquake, the different of the amplitude of the bearing displacement is not significant.
However, results using other earthquake records have shown that, with the same model bridge, a
much larger difference in the bearing displacement can result.
In Figure 5, the deck accelerations by using the M-DOF and S-DOF models are
compared. Again, it is seen that, the critical incident angles between these two models are
different. In addition, the maximum deck acceleration is more than 90% larger than that of the
minimum value in the X direction, and in the Y direction it is about 90% larger.
Figure 6 shows the comparison of the deck acceleration by using the single and 2D
excitations. Once again, the critical incident angles are different in these two cases. In addition,
Table VI lists the results of the critical angles for the M-DOF model under 2D excitations. It can

136
be seen that, first of all, different parameters do have different incident angles. In addition, the
incident angles for identical parameter along perpendicular axes are rarely perpendicular.

Figure 4. Comparisons of the bearing displacement, single and double excitations

Figure 5. Comparison of deck accelerations, M-DOF. vs S-DOF

137
 Figure 6. Comparison of deck accelerations, single and double excitations

TABLE VI. CRITICAL INCIDENT ANGLE (DEGREE)

X direction Y direction
Acc. of deck 15 64
Bearing disp 164 64
Base Shear 11 104

SUMMARY

One of the major difficulties in seismic response estimation of structures is to properly

handle the nonlinear relationship between the maximum response and the amplitude of the
ground excitation. The design response spectrum is a simple approach. However, it has its limit
in applications. This paper presents a numerical example on the responses of a seismic isolated
bridge by using M-DOF model and two directional excitations to illustrate the applicability of
the design response spectrum. Two major issues influencing the seismic responses are
discussed: (1) the choice of input ground motions, and (2) the geometrical and modal coupling
effect.
It is expected that time history analyses will be more and more intensively used as
computer technology improves to gradually replace the spectrum analysis in bridge design.
Accurate modeling and correctly selecting the forcing functions are critical in time history
analysis. This paper is prepared as a progress report of the current MCEER research on the
subject, for the purpose of inviting discussions and comments from the workshop participants.
Hopefully, future cooperative research projects can be developed with PRC researchers
interested in formulating design guidelines for seismic response modification of bridges.

138
REFERENCES

AASHTO Guide Specifications for Seismic Isolation Design. 2001.

Clough, R.W. and Penzien, J. 1993. Dynamics of Structures. Second Edition, New York:
McGraw-Hill.

Hernández, J. J. and López, O. A. 2002. Response to Three-Component Seismic Motion of

Arbitrary Direction. Earthquake Engineering and Structural Dynamics, 31:55-77.

International Conference of Building Officials. 2000.

Lee, G. C. and Liang, Z. 1998. “On Cross Effects of Seismic Response of Structures,”
Engineering Structures, 20(4-6): 503-509.

Lee, G. C. and Liang, Z. to appear 2002. “Towards next generation seismic isolation
technology.” Proceedings International Conference on Advances in Building Technology, Hong
Kong, Elsevier Science Ltd.

Liang, Z. and Lee, G. C. to appear 2002. Principal Axes of M-DOF Structures under Dynamic
Loading, Part I: Theoretical Considerations, MCEER Technical Report.

López, O. A. and Torres, R. 1997. “The Critical Angle of Seismic Incidence and the Maximum
Structural Response.” Earthquake Engineering and Structural Dynamics, 26:881-894.

Penzien, J. and Watabe, M. 1975. “Characteristics of 3-dimensional earthquake ground

motions.” Earthquake Engineering and Structural Dynamics, 3: 365-374.

Uniform Building Code (UBC), (1997)

Yang, T. 2002. Investigation of Principal Axes of M-DOF Structures under Dynamic Loadings,
M.S. Thesis, University at Buffalo, State University of New York, Buffalo, NY, June 2002.

Yang, T., Liang, Z. and Lee, G. C. to appear 2002 . Principal Axes of M-DOF Structures under
Dynamic Loading, Part II: Experimental Observations, MCEER Technical Report.

139
 Seismic Conceptual Design

for Bridge Tower of a Long-span Cable-stayed Bridge

Ai-jun Ye1,Shi-de Hu1 and Li-chu Fan1

ABSTRACT

The performance of a long-span cable-stayed bridge is highly dependent on the choice of

tower type. In this paper, three types of tower (inverted Y shape, diamond shape and A shape) was
compared from a seismic design viewpoint, the results indicate that tower of inverted Y shape or A
shape has better seismic behavior than that of diamond shape. Further, for the inverted Y shape
tower, the influence of height variation of the crossbeam section on seismic behavior was
analyzed, and the results indicate that the influence is light within a certain variation range.

__________________
1
State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, PRC

141
INTRODUCTION

The tower of a cable-stayed bridge is one of the key components of the whole structure.
The type and dimension of the tower determine the dynamic characteristics and seismic behavior
of the bridge to a great extent. Therefore, seismic conceptual design for bridge towers is discussed
herein.
The cable-stayed bridge under investigation has a steel box girder deck of 40.6m wide
overall and 2088m’s long. The layout of spans is 100+100+300+1088+300+100+100m. Figure1,
Figure 2 shows the elevation of the bridge and the cross-section of deck respectively. Three types
of concrete tower were designed, i.e. inverted Y shape, diamond shape and A shape, see Figure 3.
There is a little difference in the cross-section dimension between inverted Y shape, diamond
shape and A shape tower. Each tower of the bridge is founded on a caisson of about 80m’s deep.
The deck is supported by sliding bearings sitting on top of each auxiliary pier and anchored pier,
and there is no bearing between deck and crossbeam of each tower. The sectional area of the first
cable away from tower is 0.01134 m2, and that near the tower is 0.00518m2.
The emphasis of dynamic analysis was put on two aspects:
(1) The comparison of dynamic characteristics and seismic behavior between inverted Y shape,
diamond shape and A shape tower;
(2) The influence of height variation of the crossbeam section on seismic behavior of the tower.

-16.91
3-1

f f 14.4 0 -2 4.05
3.8 0 -25.0 8
f
-28.01 12.9 0 -2 8.58 7.20 -28.29
f
13. 10 - 31.38

Q
4

4
4
4
2 6.70 -47.79 38.6 0 -4 8.25
-50.91 34.2 0 -4 9.88 18.20 -50.0 9
33. 60 - 51.88 31. 50 -52.7 8
f f
f f 5-1 f
f 37. 80 -59.0 8 3 8.20 -59.29
-62.41 5- 1 5- 1
x x
z 51.0 0 -6 6.68 x x x x
50. 60 - 68.88 4 6.90 -67.99
x 38.85 -70.7 4 48. 80 -70.0 8 60.8 0 -7 0.45
x c c
z 6-1 c
z c 6- 1 f 6- 1 x
c 65.2 0 -8 0.88 70.3 0 -7 9.95
-82.51 67.5 0 x -8 3.18 60. 50 -81.7 8 6 1.90 -82.99
66. 50 - 84.78 51.80 x -83.6 9 63. 70 x -84.9 8
f -85.71 7

Q
x 54.65 -86.5 4 6 5.80 x -86.89

3
c 69. 10 - 87.38 7 c 7 x
c 8- 1 68. 30 -89.5 8 6 8.40 z -89.49
z c
8- 1 61.70 -93.5 9 8- 2 73. 20 f -94.4 8 7 2.50 x -93.59 83.8 0 -9 3.45
c x
-97.31 82.8 0 -9 8.48 79. 45 - 97.73 65.63 -97.5 2
82. 20 - 100.48 8- 3 z
86.9 0 f -102.58 c
f c c
-109.2 1 8-3 8-1 c
8- 1 x
95.7 0 -111.38 x 8- 1 x
8-3 c
z 100.0 0 -115.68 84.40 c -116. 29
-117.8 1 x 98. 50 - 116.78
105.0 0 -120.68 102. 40 x - 120.68 8- 2 88.80 x -120. 69 c 98. 20 -119.29
x -122.3 1 100. 60 x -121.69 111.8 0 -1 21.45
c c c 8-2 f
c c 8- 1 95.44 -127. 33 106.30 -1 27.58 105. 80 -126.89 117.2 0 -1 26.85
-128.8 1 114.2 0 -129.88 110. 60 - 128.88
f 9
9 103.65 -135. 54 9 114. 80 -135.89 126.0 0 -1 35.65
124.1 5 -139.83 f 10 129.3 0 f -1 38.95
123. 25 - 141.53 119.20 -1 40.48
f
112.10 x -143. 99 10 123.00 f -1 44.28
x 128. 10 f - 146.38 10 124. 50 -145.59 11
132.4 5 -148.13 11
-150.9 6 11 118.00 -149. 89 141.6 0 -1 51.25
135. 60 - 153.88 131. 50 -152.59
140.0 0 -155.68 f 134.50 -1 55.78
12 f

Q 2
f x f 12 f f
x
x 12 f
x
-171.9 1 x 150.20 x -1 71.48 162.8 0 -1 72.45
142.52 -174. 41 152. 20 x -173.29
159.9 0 z -175.58 157. 60 c - 175.88
13 145.80 -177. 69 13
163.9 0 -179.58 160. 30 - 178.58 170.6 5 -1 80.30
-182.5 1 14 149.50 -181. 39
162. 00 -183.09
167. 00 - 185.28
171.0 0 -186.68 f 14
-190.1 7 168. 00 -189.09 180.4 5 -1 90.10
x x 15
181.5 0 -197.18 x x
15
180. 70 - 198.98 168.00 -199. 89 177. 80 -198.89

Q
181. 70 c -202.79 16
16
1 c

Figure 1. Elevation of the cable-stayed bridge

Figure 2. Cross-section of the deck

142
 (a) inverted Y shape tower (b) diamond shape tower (c) A shape tower

Figure 3. Three types of tower

CONSTRUCTING OF DYNAMIC ANALYSIS MODEL

Based on the design drawings of the cable-stayed bridge, three 3-dimension finite element
models were constructed to make the dynamic characteristics and seismic response analysis.
Model 1 is for inverted shape tower (see Figure 4.), Model 2 is for diamond shape tower, and
Model 3 is for A shape tower.
In each FE model, the deck, towers and side piers were represented by beam elements with
six degrees of freedom (DOFs) for each node. The bridge deck was simplified using a single spine
passing through the shear center of the deck cross section. The stiffness and mass properties of the
deck were calculated and assigned to the spine. The cables and the deck spine were linked with
master-slave relation. Each tower was modeled using a 3-D frame. The cables were represented by
linear elastic truss elements with three DOFs for each node, considering the influence of cable’s
sag and dead load on the geometric stiffness. The nonlinear stiffness characteristic of the cables
due to sag was approximated by linearizing the cable stiffness using the concept of an equivalent
modulus of elasticity, such as the Ernst equation.
The deck and each tower were linked with master-slave relation only in transverse, while
the deck and each side pier (anchored pier or auxiliary pier) were linked with master-slave relation
in transverse and vertical. Each tower and side pier was fixed at the bases.

143
 Figure 4. Dynamic model of the cable-stayed bridge (inverted Y shape)

ANALYSIS OF DYNAMIC CHARACTERISTICS

Analyzing and understanding the dynamic characteristics of bridge is the base of making
seismic behavior analysis. Therefore, in this part, three FE models (mentioned above) were used to
achieve the dynamic characteristics of the bridge.
Assuming that the bridge vibrates around its dead-load static equilibrium position, the
natural modal properties were computed. Table I shows the results. One can see from the table, the
dynamic characteristics of the bridge with three types of tower are almost the same except the
transverse vibration of the tower its own. For the first tower-dominate lateral vibration mode,
frequency of diamond shape tower is 27% lower than that of inverted Y shape tower, and
frequency of A shape tower is 31% lower than that of inverted Y shape tower. Furthermore, the
torsional vibration frequency of A shape tower is also about 3% lower than that of the other two
types of tower.
In addition, the first vibration mode of the bridge is floating longitudinally, and the
corresponding period approaches 15s. Therefore, the response spectra and time history curves
used to carry out seismic response computation must include the information of a long period of
15s.
TABLE I. COMPARISION ON DYNAMIC CHARACTERISTICS OF THREE TYPES OF TOWER
Inverted Y shape tower Diamond shape tower A shape tower
Order
Frequency Frequency Frequency
Mode characteristics Mode characteristics Mode characteristics
(Hz) (Hz) (Hz)
1 0.0680 Floating longitudinally 0.0688 Floating longitudinally 0.0677 Floating longitudinally
Lateral bending Lateral bending Lateral bending
2 0.1044 0.1040 0.1043
symmetrically symmetrically symmetrically
Vertical bending Vertical bending Vertical bending
3 0.1801 0.1803 0.1787
symmetrically symmetrically symmetrically
Vertical bending Vertical bending Vertical bending
4 0.2227 0.2230 0.2213
anti-symmetrically anti-symmetrically anti-symmetrically
Lateral bending Lateral bending Lateral bending
5 0.2779 0.2718 0.2741
anti-symmetrically anti-symmetrically anti-symmetrically
Vertical bending Tower lateral bending Tower lateral bending
6 0.3150 0.3109 0.2920
symmetrically in same direction in same direction
Vertical bending Vertical bending Tower lateral bending in
7 0.3819 0.3156 0.2958
anti-symmetrically symmetrically reverse direction

144
 Tower lateral bending
8 0.4258 Vertical bending 0.3185 0.3130 Vertical bending
in reverse direction
Tower lateral bending
9 0.4281 0.3830 Vertical bending 0.3805 Vertical bending
in same direction
Tower lateral bending
10 0.4291 0.4277 Vertical bending 0.4220 Vertical bending
in reverse direction
Torsional bending Torsional bending Torsional bending
16 0.5406 0.5398 0.5223(14)
symmetrically symmetrically symmetrically
Torsional bending Torsional bending Torsional bending
27 0.7804 0.7806 0.7591
anti-symmetrically anti-symmetrically anti-symmetrically

SELECTING OF INPUT EARTHQUAKE MOTION AND ANALYSIS METHOD

Response spectrum method is selected in the seismic response computation, so two ground
response spectra (longitudinal and vertical respectively) from the site earthquake risk analysis
were selected as the input earthquake motion, see Figure 5. The corresponding damping ratio for
the two response spectra is 5%. The peak value of acceleration on ground surface is 0.194g in
horizontal, and 0.135g in vertical.
The first 300 modes of the bridge were included and combined with CQC method in the
analysis. A damping ration of 5% is used for all modes.
The structural responses under seismic inputs in longitudinal-vertical and in
transverse-vertical were computed respectively.
3

2. 5
— hor i zont al
－－ ver t i cal
Amplifying factor

1. 5

0. 5

0
0 2 4 6 8 10 12 14
Period( s)

Figure 5. Response spectrum curves

SEISMIC BEHAVIOR ANALYSIS OF THE CABLE-STAYED BRIDGE WITH

DIFFERENT TYPES OF TOWER

Based on the three models and earthquake input mentioned above, the seismic response
was predicted. Seismic response of the bridge with the three types of tower was analyzed with
comparison, as shown in Table II, Table III and Table IV. From which we can see, the difference

145
on longitudinal seismic response among three types of tower is small, while the difference is big
on transverse direction. Among which, as for inner forces, there is little difference between tower
of inverted Y shape and A shape; but for displacements on the top of the tower, A shape tower is
55% larger than the inverted Y shape tower. In particular, both inner forces and displacements of
diamond shape tower are very large, the bending moment on the bottom of the tower and the
displacement on the top of the tower are twice as large as that of inverted Y shape tower.
Table V shows the transverse inner forces on side piers. For the bending moment at the
bottom of 1# auxiliary pier, which controls the design, result of inverted Y shape tower is 6% less
than that of diamond shape tower, 7% more than that of A shape tower. So, the difference is small.
As a whole, inverted Y shape tower and A shape tower have a better aseismic behavior
than diamond shape tower.
TABLE II. MAXIMUM VALUES OF INNER-FORCE ON CONTROL SECTION OF MAIN TOWER
(LONGITUDINAL-VERTICAL INPUT)
Axial force Shear Force Bending Moment
Tower shape Section location of tower
P (kN) Q3 (kN) M2 (kN.m)
Bottom of tower 7.503E4 2.699E4 2.250E6
Inverted Y
Above the crossbeam 6.669E4 1.809E4 1.241E6
shape
Below the intersection 3.255E4 9.564E3 5.193E5
Bottom of tower 7.426E4 2.805E4 2.310E6
Diamond
Above the crossbeam 6.546E4 1.876E4 1.249E6
shape
Below the intersection 3.211E4 9.651E3 5.278E5
Bottom of tower 7.646E4 2.941E4 2.362E6
A shape Above the lower crossbeam 6.815E4 1.922E4 1.302E6
Below the upper crossbeam 3.587E4 1.020E4 5.874E5
where subscript 3 indicates the longitudinal direction , subscript 2 indicates the transverse direction

TABLE.III MAXIMUM VALUES OF INNER-FORCE ON CONTROL SECTION OF MAIN TOWER

(TRANSVERSE-VERTICAL INPUT)
Section Axial force Shear Force Bending Moment
Tower shape
location of tower P (kN) Q2 (kN) M3 (kN.m)
Bottom of tower 2.612E5 3.996E4 1.462E6
Above the crossbeam 2.047E5 3.777E4 1.235E6
Inverted Y
Below the intersection 1.939E5 9.114E3 5.118E5
shape
End of the crossbeam 1.018E4 7.507E4 1.720E6
Bottom of the tower 2.322E5 6.711E4 2.956E6
Diamond Above the crossbeam 1.822E5 3.095E4 7.410E5
shape Below the intersection 1.710E5 7.321E3 3.438E5
End of the crossbeam 1.622E4 7.114E4 1.628E6
Bottom of tower 1.974E5 4.179E4 1.488E6
Above the lower crossbeam 1.431E5 3.888E4 1.118E6
A shape
Below the upper crossbeam 1.298E5 1.019E4 7.784E5
End of the lower crossbeam 1.933E4 6.829E4 1.560E6
where for tower column: subscript 3 shows the longitudinal direction, subscript 2 shows the transverse
direction; for crossbeam: subscript 2 shows the vertical direction, subscript 3 shows the longitudinal direction

146
 TABLE IV. MAX. DISPLACEMENT ON KEYPOINTS (m)
Longitudinal-vertical input Transverse-vertical input
Tower shape Location
Longitudinal Vertical Transverse Vertical
Top of tower 1.292 / 0.222 /
Inverted Y
End of deck 1.198 / 0.118 /
shape
Mid-span 1.201 0.246 0.955 0.246
Top of tower 1.260 / 0.488 /
Diamond
End of deck 1.168 / 0.132 /
shape
Mid-span 1.171 0.247 1.002 0.246
Top of tower 1.309 / 0.345 /
A shape End of deck 1.213 / 0.113 /
Mid-span 1.216 0.241 0.983 0.241

TABLE V MAX. INNER-FORCE OF CONTROL SECTION OF SIDE PIERS

(TRANSVERSE-VERTICAL INPUT)
Axial Force Shear Force Bending Moment
Tower shape Section location
P (kN) Q2 (kN) M3 (kN.m)
Bottom of 1# auxiliary pier 6.415E3 1.024E4 5.339E5
Inverted Y shape Bottom of 2# auxiliary pier 4.869E3 8.543E3 4.261E5
Bottom of the anchored pier 2.113E3 7.723E3 3.639E5
Bottom of 1# auxiliary pier 7.083E3 9.708E3 5.040E5
Diamond shape Bottom of 2# auxiliary pier 4.978E3 8.400E3 4.152E5
Bottom of the anchored pier 2.336E3 8.813E3 4.086E5
Bottom of 1# auxiliary pier 5.248E3 1.094E4 5.713E5
A shape Bottom of 2# auxiliary pier 4.913E3 8.808E3 4.418E5
Bottom of the anchored pier 2.116E3 7.417E3 3.477E5
where axial 3 shows the longitudinal direction, axial 2 shows the transverse direction.
Numeral order is from tower to the anchored pier.

INFLUENCE ANAYLSIS OF THE HEIGHT VARIATION OF CROSSBEAM SECTION

ON THE SEISMIC RESPONSE

Since the inner forces due to earthquake are greatly more than that of the static load, so it
will control the design. Section checking shows that: when the earthquake motion is input in
transverse-vertical direction, bending moment of the crossbeam is considerably large, which
makes the design of section with 8m high very difficult. So herein we try to analyze the sensitivity
of seismic response of inverted Y shape tower by modifying the height of the section from7.0m up
to 10.0m.
Table VI and Table VII show the inner forces of control section of inverted shape tower
under longitudinal-vertical input and transverse-vertical input respectively.
Table VI shows that as section height of the crossbeam increase from 7.0m up to 10.0m, so
do the inner forces of tower column, however, the variation ratio is small, not more than 1%.
Table VII shows that seismic response of sections along the bridge tower varies
inconsistently under transverse-vertical input: results of column section above the crossbeam, and
section at the end of crossbeam have the trend to increase, while results of column section at the

147
bottom of the tower and below the intersection decreases. When the height of crossbeam varies
from 8.0m to 9.0m, the seismic responses of the four control sections mentioned above vary
slightly, the increase ratio of crossbeam bending moment is within 6%, while the section strength can
be greatly improved. Therefore, it is beneficial to the seismic behavior of the tower.

TABLE VI. THE INFLUENCE OF HEIGHT VARIATION OF CROSSBEAM SECTION

ON THE INNER FORCES OF TOWER (LONGITUDINAL-VERTICAL INPUT)
Height of Section location Axial Force Shear Force Bending Moment Variation
crossbeam (m) P (kN) Q3 (kN) M2 (kN.m) ratio
Bottom of tower 7.490E4 2.676E4 2.243E6 -0.3%
7.0 Above the crossbeam 6.670E4 1.808E4 1.240E6 -0.1%
Below the intersection 3.253E4 9.545E3 5.176E5 -0.3%
Bottom of tower 7.503E4 2.699E4 2.250E6 0
8.0 Above the crossbeam 6.669E4 1.809E4 1.241E6 0
Below the intersection 3.255E4 9.564E3 5.193E5 0
Bottom of tower 7.519E4 2.726E4 2.258E6 0.3%
9.0 Above the crossbeam 6.669E4 1.812E4 1.243E6 0.2%
Below the intersection 3.257E4 9.583E3 5.214E5 0.4%
Bottom of tower 7.537E4 2.753E4 2.266E6 0.7%
10.0 Above the crossbeam 6.672E4 1.815E4 1.244E6 0.2%
Below the intersection 3.260E4 9.602E3 5.236E5 0.8%
where 3 shows the longitudinal direction,2 shows the transverse direction; the variation ratio is based on
8.0m,”+”means increasing,”-“ means decreasing.

TABLE VII. THE INFLUENCE OF HEIGHT VARIATION OF CROSSBEAM SECTION

ON THE INNER FORCES OF TOWER (TRANSVERSE-VERTICAL INPUT)
Height of Axial Force Shear Force Bending Moment Variation
Section location
crossbeam (m) P (kN) Q2 (kN) M3(kN.m) ratio
Bottom of tower 2.435E5 3.876E4 1.496E6 2.3%
Above the crossbeam 1.897E5 3.629E4 1.185E6 4.0%
7.0
Below the intersection 1.784E5 9.089E3 5.181E5 1.2%
End of the crossbeam 9.556E3 6.872E4 1.574E6 -8.5%
Bottom of tower 2.612E5 3.996E4 1.462E6 0
Above the crossbeam 2.047E5 3.777E4 1.235E6 0
8.0
Below the intersection 1.939E5 9.114E3 5.118E5 0
End of the crossbeam 1.018E4 7.507E4 1.720E6 0
Bottom of tower 2.736E5 4.076E4 1.433E6 -2.0%
Above the crossbeam 2.162E5 3.891E4 1.265E6 2.4%
9.0
Below the intersection 2.057E5 9.121E3 5.050E5 -1.3%
End of the crossbeam 1.084E4 7.945E4 1.821E6 5.9%

148
 Bottom of tower 2.819E5 4.131E4 1.409E6 -3.6%
Above the crossbeam 2.245E5 3.976E4 1.283E6 3.9%
10.0
Below the intersection 2.143E5 9.116E3 4.987E5 -2.6%
End of the crossbeam 1.146E4 8.252E4 1.891E6 9.9%
where 3 shows the longitudinal direction , 2 shows the transverse direction; the variation ratio is based on
8.0m,”+”means increasing,”-“ means decreasing.

CONCLUSION

Based on the above discussion, conclusions can be made as follows:

The difference on longitudinal seismic response among the three types of tower is small,
while the difference is big in transverse direction. Among which, diamond shape tower has
the disadvantage of particular large bending moment on the bottom of tower and
displacement on the top of tower, while for inverted Y shape and A shape tower, the
bending moment and displacement are comparatively small. Therefore, tower of inverted
Y shape or A shape has better seismic behavior than that of diamond shape
Since the influence of height variation of crossbeam section on seismic behavior is small,
increasing the section height (1m) could be an alternative to achieve sufficient section
strength for the crossbeam, where reinforcing bars is difficult to be set.

REFERENCES

Li-chu Fan, Shi-de Hu and Ai-jun Ye. 2001. “Seismic Design For Long Span Bridge,” People’s Communication
Publishing House (in Chinese)
Li-chu Fan.1997. “Seismic Design Of Highway Bridge”, Huajie International Publishing Co. limited

149
 Analytical Investigation of the Response of Lu-Pu Bridge
with Added Viscous Dampers

Shi-de Hu1 , Zhi-qiang Wang1 and Li-chu Fan1

ABSTRACT

Energy dissipation devices have been widely used in the world since 1970s for the
seismic protection of bridges. Viscous passive damping devices have lately emerged as one of
the alternative technology devices that are available for using in the seismic design of bridge
structures. Different bridge structure types and site conditions may result in significantly
different strategy in applying the seismic protection concept with viscous dampers. The present
study examines the seismic response of LuPu bridge, with the arch span length of 550m long,
using supplemental viscous damping devices and some testing results of viscous damper
property are presented. For LuPu bridge, the use of large damping elements only at a few
locations within the supuerstructure to reduction local response of the bridge.
The series of analysis provided an understanding of the sensitivity of the bridge response
to different parameters. These included the effect of the damper non-linearity, such as velocity
exponent and damping constant on the response reduction.

__________________
1
State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, PRC

151
INTRODUCTION

LuPu bridge provides the direct highway link between PuDong and PuXi. With a 550m
main span and 2×100m side span the arch bridge is the longest span in the world at the moment.
It is a new symbol of shanghai city. The major components of the arch bridge are the steel arch
rib, steel bridge deck, as shown in figure 1.
Nonlinear time history analysis of LuPu bridge reveals that relative displacements at
expansion joints are large. Impact might take place under earthquake. Therefore, longitudinal
viscous dampers will be installed to limit relative displacements at expansion joints.

MODELING PARAMETERS

The present investigation use non-linear transient 3D dynamic analysis to determine the
response of LuPu arch bridge with varying damper parameters subjected to earthquake ground
motions.

Figure 1. Main span of LuPu bridge

east

west
Location of dampers

Figure 2. Damping element locations

152
Bridge analysis model

Longitudinal dampers will be installed at locations as shown in Figure 2. Two viscous

dampers will be installed at each of the expansion joints between the main span deck and
transverse girder of arch rib.
The main objective is to reduce the relative displacements at the expansion joints, thus
eliminate the possibility of impact forces and a reduction of arch rib stress due to longitudinal
displacements. At the same time, viscous damper can accommodate slow temperature
displacements without forces at the dampers.

Ground motion

Some spectrum compatible acceleration time histories were created, one of such time
history is used in this investigation. It was used to represent the site motions of LuPu bridge.
Figure 3 shows the pseudo acceleration response spectrum of the time history.

Damping devices

In this investigation, viscous dampers were utilized. The devices modeled were based on
those available commercially from SRIM. These are the damping devices currently being used in
some other bridges. They are fluid viscous devices that dissipate energy by forcing a silicon fluid
through orifices at high velocity. The behavior of the devices can be described by the following
equation:
ξ
F = C V sgn(V ) (1)

5
psuedo acceleration

Response spectrum
4

0
0 1 2 3 4 5
Period (sec)

Figure 3. Response spectra of ground motion

153
 where F is the damping force, V is the velocity across the device; C is a damping
coefficient and ξ is an velocity exponent in the range of 0.1 to 2.0. These damping devices can

be manufactured with a wide range of C and ξ value. The cost of the devices is generally

proportional to the maximum damping force required. Figure 4 gives the diagram of the damping
force-displacement relationship for the case of harmonic motion, keeps the coefficient
C constant and varies the exponent ξ from 0.15 to 1.0, the normal range of practical devices.

As ξ reduce from 1.0 to 0.15 the damping force reduces and the damping force-displacement

relationship trends from an elliptical toward a more rectangle form.

The result from a transient dynamic analysis can be quite voluminous. It is important to
identify relatively few parameters that will still represent the total behavior of the structure. For
these analyses, the maximum and minimum axial forces in a representative number of elements
were monitored.

Baseline analysis

The LuPu bridge in its existing configuration without dampers was subjected to the
ground motions to obtain a benchmark for evaluating the effectiveness of the damping elements.
table 1 show the response of LuPu bridge to the artificial wave. Both are plots the maximum and
minimumo0 axial force in the arch rib during the earthquake.

5
ξ =0.15 ξ =0.21 ξ =0.3
4
ξ =0.50 ξ =0.70 ξ =1.0
3
damping froce (10 kN)

2
3

-1

-2

-3

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6

displacement (m)
Figure 4. Harmonic damping force-displacement relationship

154
Analyses with dampers

The properties of the dampers were optimized to achieve an appropriate reduction of the
relative displacements and stresses, keeping the damper force within reasonable limits. Different
values for the exponent ξ were evaluated. Exponents ξ from 0.15 to 1.0 were considered. For

each of these exponents, several values of C were examined in a parametric study aimed
identifying the optimum dampers. This study was carried out by calculating the seismic response
to one of the ground motions with three-dimensional elastic models of the bridge. The same
damper properties were assumed for all dampers. Thirty different configurations of the damping
devices were utilized. The dampers were located at the same positions in the bridge for all cases,
but the parameters of the dampers themselves were varied. See table 2.

TABLE I ANALYSIS RESULT WITHOUT DAMPER

P4 (longitudinal+vertical)
Shear force （kN×103
Position of sections Axial force（kN×104） Moment（kN.m×104）
）
west east west east west east
Section of arch springing of
3.949 4.266 8.967 11.06 25.32 20.61
Side span
Section of arch springing of
2.963 2.960 9.578 10.75 17.61 27.04
Main span
Displacement of expansion
160.7 169.1
joints（mm）

TABLE II CASES OF PARAMETERS ANALYSIS

Cases of Ground motion
C 230 330 460 560 660
analysis level
0.15 0.15 0.15 0.15 0.15 P4
0.21 0.21 0.21 0.21 0.21
0.3 0.3 0.3 0.3 0.3
ξ
0.5 0.5 0.5 0.5 0.5
0.7 0.7 0.7 0.7 0.7
1.0 1.0 1.0 1.0 1.0

Damper parameters variations

Variation of velocity exponent ξ

The value of ξ in equation 1 was varied to determine its effect on response and on

maximum damper force. Figure 5, 6 and 7 show the change in the response parameters of
structure with changing ξ value.

155
 Figure 5, 6 and 7 display a reduction of the maximum relative displacement at the
expansion joint and damper force by increasing ξ , but axial force of arch bridge increases.

130
west

displacement (mm)
east
120

110

100

0.2 0.4 0.6 0.8 1.0

ζ ( C=460 )

Figure 5. Relative displacement of expansion joint

3.8

3.6
axial force ( 10 ) kN

3.4
side arch east
side arch west
4

3.2 main arch east

main arch west
3.0

2.8

2.6
0.0 0.2 0.4 0.6 0.8 1.0
ζ ( C=460 )

Figure 6. Arch rib axial force

2.0

west
damping froce ( 10 ) kN

1.8
east
3

1.6

1.4

1.2

1.0
0.0 0.2 0.4 0.6 0.8 1.0
ζ ( C=460 )

Figure 7. Damping force of damper

156
Variation of damping coefficient C

2.0

1.8 west

damping force (10 ) kN

east

3
1.6

1.4

1.2

1.0

0.8

0.6
200 300 400 500 600 700
C ( ζ=0.7 )

Figure 8. Damping force of damper

3.8

3.6
axial force ( 10 kN )

3.4
4

side arch east

side arch west
3.2 main arch east
main arch west
3.0

2.8

200 300 400 500 600 700

C ( ζ=0.7 )

Figure 9. Arch rib axial force

120

west
east
displacement (mm)

110

100

90

200 300 400 500 600 700

C ( ζ=0.7 )

Figure 10. Relative displacement of expansion joint

157
 Figure 11. Fluid damper
The value of C in equation 1 was varied to determine its effect on response and on
maximum damper force. Figure 8, 9 and 10 show the change in the response parameters of
structure with changing C .
Figure 8 9 and 10 show a reduction of the maximum relative displacement at the
expansion joint and axial force of arch bridge with increasing C . But damping force increases.

DAMPER TESTS

Fluid viscous dampers were identified as the type of energy dissipator that was most able
to meet the demanding load and displacement requirements for the earthquake. In the design
process, it was recognized that there is little precendent for the use of viscous damping devices
in the seismic design of LuPu bridge, at least in china, and that a comprehensive testing program
would be required to verify that fluid viscous dampers could provided the desired performance.
The damper test program of dynamic performance described in this paper was
undertaken at State Key Laboratory for Disaster Reduction in Civil Engineering of Tongji
University specifically to validate the assumed mechanical characteristics of reduced-scale
viscous dampers.
The damper in the testing program is unaxial fluid viscous damper. It is essentially
cylindrically-shaped units, see Figure 11 ,with a piston and piston rod moving relative to the
main body of the damper. The test damper is reduced-scale with respect to the full-size units that
will be used in the LuPu bridge. The damper needed to be scaled down to achieve a reasonable
size of device for the testing program.
The damper tested is designed for a force-velocity relationship of

F = 92 ⋅ V 0.21 kN (2)
The test damper is designed to produce approximately 200kN at a velocity of 200mm/s,
and to have a displacement range of ± 70 mm. These values represent scale factors of 1:10 for
force, 1:2.857 for displacement, and 1:3.2 for velocity.
In total, about 29 tests were performed on the damper. Some results for the damper are
presented and discussed here. A typical damper force-displacement plot for a constant-velocity
test is shown in Figure 12. The plot is for a five-cycle test with a target constant velocity of

158
35.2mm/s. It can be seen that the damper force output was very stable and repeatable in this test.
Note that because the damper force is a function of velocity and that the applied velocity was
approximately constant, the damper force is therefore also approximately constant across the test
displacement range. For each of the constant velocity tests, an “average” value of damper force
as a function of velocity was determined, and then compared against the target force-velocity

200
Damper Force (KN)

100

-100

-200
dam per t 0.167 Hz

-40 -20 0 20 40
displacem ent (m m )

Figure 12. Typical damper force verus displacement plot for a cyclic
constant-velocity test

300

250
Damper Force (KN)

200

150

Test
100 Design

50

0
0 5 10 15 20 25 30 35 40 45 50
velocity (mm/s)
Figure 13. Damper force verus velocity for all cyclic constant-velocity tests

159
relationship. Figure 13 summarizes all of the constant-velocity tests. The solid line shown in the
figure is the target force-velocity law (Eq.2). It can be seen that in general, there is good
agreement between the actual and target damper behavior.
Other damper characteristics that were evaluated in the test program include: the
predictability of the damper behavior under seismic loading, sinusoidal Tests.

CONCLUSIONS

Although this investigation has been far from thorough, the following conclusions can be
drawn:
Placing dampers in LuPu arch bridge can reduce the response of the bridge to earthquake
excitations.
The response sensitivity analysis that was performed to obtain the optimum design and
understand the behavior of the bridge with damper can be generally applied to other structures.
It shows that the importance that the velocity exponent and damping coefficient can have
on the optimum seismic design, the force , relative displacements, and on the damping force.
Considering these results, exponent ξ range from 0.2～0.5 was selected, damping coefficient is

about 500kN.mm/s.
Besides seismic excitation, the damper will be subjected to thermal movements, wind
excitation and small amplitude, high frequency vibrations caused by traffic on the bridge. The
final damper design must be able to perform well under all these environments.

REFERENCES

Lichu, Fan, 1997, Seismic Design for Highway Bridge, Huajie International Publishing Co. Limited, Hong Kong.
Priestley, M.J.N., Seible, F., and Calvi, G.M. 1996. Seismic Design and Retrofit of Bridges. John Wiley & Sons, Inc.,
New York.
Aiken, I.D. and Kelly, J.M., 1996, “ Cyclic Dynamic Testing of Fluid Viscous Dampers,” Proceedings, Caltrans;
Fourth Seismic Research Workshop, California Department of Transportation, Sacramento, California, July.
Parvin, A. and Ma, Z., 2001, “ the Use of Helical Spring and fluid Damper Isolation Systems for Bridge Structures
Subjected to Vertical Ground Acceleration,” Electronic Journal of structural Engineering, (2),98-110.

160
 Seismic Performance and Retrofit of a 24-
Span Freeway Bridge
M. Saiidi1, A. Itani1, Q. Yang2, and T. Isakovic3

ABSTRACT

Seismic retrofit screening of highway bridges in Nevada, USA has identified a major
viaduct in downtown Las Vegas as a structure with highest priority for retrofit. The bridge has
24 spans and 110 columns of various shapes. The seismic performance, vulnerability, and
retrofit details of the bridge and its adjacent ramps are being evaluated. Five major aspects of
study are in progress. The focus of this article is on the effects of incoherent ground motion on
the viaduct, detailed nonlinear dynamic analysis of the bridge, and shake table studies of single
column piers. It is shown in the paper that combination of geometric incoherency, wave passage,
and site characteristics effects can increase the base shears considerably. The columns in the
main bridge are found particularly vulnerable to the longitudinal motion of the bridge. The
shake table tests of the as-built octagonal columns have shown that the pedestals are unable to
work with the columns and separate from the column under large drifts.

INTRODUCTION

New concepts and issues have emerged in the past few years in seismic evaluation and
retrofit of structures. The underlying principle in utilizing new concepts is to ensure a reliable,
economical, and intelligent design that makes optimal use of innovative materials and
information technology tools. Some of the new tools, while may be well developed in the arena
of their origin require further research and development before they can be adopted in civil
engineering projects. It may be hard to justify the development work for specific projects of
relatively small magnitude. However, the potential benefit of such research can be significant
for large projects and hence undertaking the so-called “high risk” studies may be warranted.
In planning the seismic retrofit of major bridges, it is inevitable that the solution will
involve both conventional and non-conventional strategies. This article presents some of the
results of an on-going study to develop retrofit plans for a major freeway viaduct in Las Vegas,
Nevada. This bridge has been identified as the most critical bridge for retrofit in Nevada due to
its high average daily traffic. Several aspects of the structural response and retrofit methods are
being studied. The article will present a summary of the study of three of the critical issues.

1
Civil Engineering Department, University of Nevada, Reno, NV 89557, USA
2
Civil Engineering Department, Northern Jiaotong University, Beijing, CHINA
3
Civil Engineering Department, University of Ljubljana, Ljubljana, SLOVENIA

161
 Figure 1. Bridge aerial view.

DESCRIPTION OF THE VIADUCT

The aerial view of the freeway and the bridge is shown in Fig. 1. The viaduct consists of
two main 22-span continuous structures for the westbound and eastbound traffic on Interstate 95
in North Las Vegas (Fig. 2). In addition a 7-span on ramp and two off ramp bridges, one with
two spans and the other with three spans are connected to the viaduct. The superstructure is cast-
in-place multicell box girder in all the structures and has seven in-span hinges in the main
structure. The main bridges are supported on multi-column piers with two to four diamond
shape columns and the ramp bridges are supported on irregular octagonal single-column piers
some with a pedestal. Figures 3 and 4 show the cross section of the columns. Single columns

Figure 2. Dimensions of the viaduct.

162
 6'-8"

3'-4" 3'-4"

# 5 ties

2' 1 " typ.

24
# 5 tie
4'
CL Pier

2' # 11

2" clear cover

Colum n CL

Figure 3. Typical column cross section in multi-column piers (1 in.= 25.4 mm)

are detailed with a one-way hinge at connection to the footing and columns of the multi-column
bents are detailed with two-way hinges at the base.
The viaduct was constructed in 1969 and was widened in 1984 by adding one row of
columns and a new box cell on the north side with little seismic load and detailing consideration.
The columns are supported on spread footings. Based on the soil blow counts at the site of the
bridge, approximately the west third of the bridge is on a relatively soft soil while the rest of the
structure is on medium firm soil.
Seismic codes place Las Vegas in areas of moderate seismicity category. To determine
how critical a bridge is with respect to the need to upgrade its seismic performance, however, the
average daily traffic (ADT) is also factored in. As a result the Nevada Department of
Transportation has identified the viaduct as the most critical bridge in Nevada with respect to
priority for seismic retrofit because of its high ADT in addition to having seismic deficiencies.
Seismological studies of the Las Vegas area have pointed out the possibility of strong ground
motion due to soil amplification despite the fact that the area is categorized as having only a
moderate probability of experiencing high seismic forces (Su et al. 1998).

Figure 4. Typical column cross section in single-column piers (1 in. =25.4 mm)

163
CRITICAL ISSUES

The original structure and the subsequent expansion were not designed to resist
significant seismic forces. As a result there are many deficiencies in the bridge with respect to
seismic detailing. These include inadequate lateral steel in the columns (Fig. 3 and 4), short
anchorage length for longitudinal bars, insufficient shear steel in the beams, very low amount of
bottom steel in the beams at connections to columns, a lack of top mat of steel in footings, short
seat widths at hinges, and a lack of shear steel at beam column connections. Considering the
length of the viaduct and variation in soil properties, there is also uncertainty about the
performance of the bridge under incoherent ground motions at different supports.
The study to address these deficiencies include (1) a reduced linear modeling of the
bridge to study the effect of incoherent ground motions, (2) a three-dimensional detailed
nonlinear modeling of the viaduct subjected to earthquakes to determine the performance level
under different loading scenarios, (3) shake table studies of the as-built and retrofitted models of
single-column bents, (4) shake table studies of the as-built and retrofitted models of multi-
column bents, and (5) an exploratory study of the application of innovative materials for retrofit.
Due to space limitation only a summary of the research in Tasks 1, 2 and 3 is presented in this
article.

INCOHERENT GROUND MOTION EFFECTS

Although commonly assumed to be uniform, ground motion at different supports of a

structure may vary from one support to the next. Variation in support excitation may be due to
three causes: (1) geometric incoherency, (2) wave-passage, and (3) local site geotechnical
characteristics. The first effect is due to randomness of the earthquake motion even at a given
site with practically the same distance from the earthquake source. The second effect is caused
by delay in the motion from one support to the next as the earthquake wave passes through the
site. Finally the third source of variation is that, for the same bedrock motion, soils of different
properties and depth transfer the motion to the surface differently. The first two factors can be
particularly important for relatively long structures, whereas the third factor can be important
even for short structures with supports located on soils layers that are different from one another.
The total length of the bridge is 552.7 m. Furthermore soil investigation of the site prior
to construction have shown that the soil type over the western third of the bridge may be
categorized as soft, while the soil for the rest of the bridge is in the category of medium firm.
Considering the length and the variation in the soil type it was decided to include all three
sources of deficiencies in the study. No site-specific seismic studies and detailed soil
investigations have been conducted for the viaduct. As a result it was felt that the study of the
incoherency effects would have to be approximate. A reduced linear model of the viaduct was
used to determine the trends in the response relative to the response for uniform ground motions.
The structure was treated as a planar system subjected to in-plane ground motion loading in the
longitudinal direction of the bridge. Details of the study are presented by Yang et al. 2002.
Figure 5 shows the elevation of the reduced model of the viaduct. To develop the model,
each segment of the structure between adjacent hinges was represented by a single mass
supported on a column that had the same lateral stiffness as that of the bridge segment. Bridge
hinges were modeled as pinned connections. The coherency function was adopted from a

164
 Y 63.55m 67.06m 77.57m 73.15m 73.15m 71.17m 74.01m
M1 M2 M3 M4 M5 M6 M7 M8
X
Z 9 10 11 12 13 14 16 17 18 19 20 21 22 23
15
H 1 P1 H2 P1 H 2 P2 H 3 P3 H3 P1 H3 P1 H3 P1 H 4 P4
1 2 3 4 5 6 7 8

Soft soil (Type Ⅲ ) Medium soil (Type Ⅱ )

D H
A B E F G
C Bed rock

Figure 5. Reduced model of viaduct.

previous study that was based on random processes principles developed by Yang and Chen
(2000). Eight acceleration records were generated at bedrock with slightly different peaks in a
range of 0.38 to 0.41g. These motions were used to determine the displacement and acceleration
histories at the ground level of the soil column under each support of the reduced model. To
study the effect of wave passage, a wave velocity of 600 m/sec. was used for the entire viaduct
even though the soil varied from the west to the east. This was done because blow counts of the

western and eastern parts, although placed the soils in different categories, were not drastically
different.
The effect of different combinations of incoherency parameters were studied by focusing
on the “column” base shears in the reduced model. An artificial acceleration record generated
for a magnitude 7.4 at 40 km distance from the epicenter was used. To establish a benchmark,
the model was analyzed for two uniform ground motions obtained by combining the bedrock
motion near the middle of the viaduct one with amplification due to soft soil and the other with
5*
Ft/Ft
Case3 Case4 Case5
2.0 Case6 Case7 Case8

1.6

1.2

0.8

0.4
0 1 2 3 4 5 6 7 8
Node number

Figure 6. Effect of incoherent motion.

165
amplification with medium firm soil. The average of the two responses was used as the
representative response for uniform ground motion.
The ratios of base shear at different column base nodes for the non-uniform and uniform
motions are shown in Fig. 6. In case 3 only the effect of site characteristics was included. Cases
4 and 5 included the wave passage effect in addition to the local site effect, one for waves
traveling from east to west and the other traveling in the opposite direction. In case 6 the
combination of local site effect and geometric incoherency was included. Finally cases 7 and 8
accounted for all three parameters, for waves traveling from east to west and west to east,
respectively. A ratio exceeding one on the vertical axis indicates an increase in force due to
incoherent ground motion. It can be seen in Fig. 6 that cases 3, 4 and 5 were not generally
critical. However, for cases 6 to 8 the base shear ratios exceeded 1 in many piers. The
maximum ratio occurred in Pier 7 with ratios being 1.59, 1.72, and 1.84 for case 6, 7, and 8,
respectively. It was concluded that the inclusion of all three incoherency effects can be the most
critical combination and even eliminating the wave passage effect (case 6) led to force ratios that
exceeded one in one-half of the piers.

NONLINEAR DYNAMIC ANALYSIS OF COMPLETE SYSTEM

The study of possible effects of incoherent ground motion described in the previous
section was conducted on a reduced, 2-dimensional, linear version of the structure to keep the
model manageable and to allow for the study of combination of several parameters. In the part
described in this section a comprehensive model of the bridge was developed for analysis on
program Drain-3DX (Prakash, et al., 1993). This part of the study is still in progress, but
preliminary results have been obtained and are presented. The essential features used in the
model were nonlinear fiber elements for the columns, hinges, shear keys, restrainers, and
abutments. A three-dimensional model of the bridge was developed for analysis subjected to
uniform support excitation in two horizontal orthogonal directions.

0.34’ 0.66’
0.31’ 0.52’
y 0.34’ 0.66’ 0.52’ 0.31’

1 0.66’
1
0.52’
2
1.33’ 2
0.86’
0.84’ y 3
3 0.86’
4 0.5’ 4
5 0.52’
z
z

(a) Concrete (b)

Figure 7. Fiber elements for diamond columns.

166
Nonlinear Elements

The columns in the bridge are of two general types, diamond shape sections used in
multi-column bents and irregular octagonal sections used in single column bents. There is
considerable variation in the longitudinal steel ratio in diamond shape columns ranging from 2.2
to 5.6 percent. As can be seen in Figs. 3 and 4, the longitudinal steel is distributed around the
section. In developing the fiber elements for plastic hinges, the bars were lumped as shown in
Figs. 7 for the diamond shape columns. Similar simplifications were made in octagonal
columns. A trilinear curve with properties based on the average measured properties was used
for concrete and steel fibers.
Several nonlinear components had to be included at the hinges. Figure 8 shows the
nonlinear hinge model that included unidirectional springs for shear keys, restrainers, and impact
effect. The impact and restrainer springs are activated in the longitudinal direction after closure
of a gap (in the case of the impact spring) or tightening of the restrainer rods (in the case of the
restrainers springs). The shear key elements act in the transverse direction of the bridge after the
initial gap between the edge of the shear key and the superstructure is closed. At connections
between the ramps and the main structure, the hinge element is more complicated but has the
same spring types as those shown in Fig. 8.
The eastbound and westbound structures in the main bridge are connected at a
longitudinal hinge. This hinge was modeled only with compression springs. Therefore, the two
structures were allowed to separate during the earthquake without resistance.

Dynamic Analyses and Results

Two acceleration records were used in the analysis, the 1941 El Centro record and an
artificial earthquake based on Eurocode 8. The peak ground acceleration for the former was
0.35g and for the latter was 0.2g. The spectra for the two records are shown in Fig. 9. The peak
accelerations are varied for different runs to simulate the effect of earthquakes with different
intensities.
A modal analysis of the structure was first conducted. Because of the hinges and the
nonlinearity they introduce two sets of vibration modes were calculated with two different
boundary conditions, one with hinges disconnected and the other with hinges fully connected.
restrainers
shear keys
impacts

1 4 1 4 Legend:

comp. elements used

to model the impact

2 2
restrainers
5 5
shear key - compression

3 3
6 6
shear key - tension

Figure 8. Hinge components.

167
 ElCentro Artificial accelerogram

1 0.6
0.5

acceleration [a/g]
acceleration [a/g]

0.8
0.4
0.6
0.3
0.4
0.2
0.2 0.1
0 0
0 0. 5 1 1.5 2 2.5 3 3. 5 4 0 0. 5 1 1.5 2 2.5 3 3. 5 4
period T [s] period T [s]

Figure 9. Acceleration spectra for 5% damping.

The former led to 61 important modes with period ranging from 0.12 to 1.77 sec., and the latter
lead to 17 important modes with a period range of 0.14 to 0.62 sec. The actual behavior of the
bridge is in between these two sets of boundary conditions and depends on the ground motion
characteristics and the response amplitude. Because of differences in column heights and pier
stiffnesses, the model with disconnected joints had a particularly large contribution from in-plane
rotational modes. The results of the modal analysis were used to estimate the mass and stiffness
proportional damping coefficients used in the response history analyses. A damping ratio of 5
percent was used.
Figure 10 shows the longitudinal displacement envelopes for analysis in the longitudinal
direction of the bridge subjected to the El Centro record amplified by a factor of two (PGA=
0.7g). The maximum displacement was 101 mm (3.96 in.), corresponding to a column drift ratio
of 1 percent. This level of drift led to yielding of the columns and the development of plastic
hinges with relatively low ductility demand. Given the low level of confinement in the columns,
however, they are expected to have a ductility capacity of approximately 1.5 and are likely to be
damaged under this motion. Considering the uncertainty in the abutment capacity, the analysis
was repeated with abutments removed. Because of the low flexibility of the diamond shape

Displacement envelopes - EAST part

0.4
0.3
displacem ent [ft]

0.2
0.1
0
-0.10.00 500.00 1000.00 1500.00
-0.2
-0.3
-0.4
sta tion [ft]

Figure 10. Maximum longitudinal displacements for 2x El Centro (1 ft.=305mm).

168
 Figure 11. Shake table setup.
columns in the longitudinal direction of the bridge the maximum displacement increased to 244
mm (9.6 in.) with a column drift ratio of 2.6 percent. It was concluded that if the abutments fail
during the earthquake, the high deformation demand on the columns would lead to column
failure and collapse of the structure.

SHAKE TABLE PERFORMANCE OF OCTOGONAL SINGLE COLUMNS

As-Built Specimen

Three, quarter-scale models of the octagonal single column piers were constructed for
shake testing in the strong direction. One specimen has been tested in the as-built condition and
the other two will be retrofitted and tested to determine the effectiveness of different retrofit
techniques. The models represented the most critical single columns in the ramp structure with
respect to the shear demand.

Shake Table Testing

Figure 11 shows the shake table set up for the as-built specimen. Preliminary nonlinear
dynamic analysis of the column for a variety of earthquakes indicated that the most demanding
motion would be the Sylmar record of the 1994 Northridge earthquake. This motion was
simulated on the table with increasing amplitudes in successive runs until the column failed. The
envelope of the measured load-displacement hysteresis relationships is shown in Fig. 12. The
primary damage occurred in the pedestal. Vertical cracking of the pedestal began at relatively
small motions. As the intensity of the motion increased, the pedestal cracks widened and led to
separation of the pedestal from the column (Fig. 13). Several important observations were made.
Figure 12 shows that the overall load-displacement response is ductile. The measured
displacement ductility capacity exceeded 6. The longitudinal bar strain data indicated substantial
strains without bar pullout despite the fact that the bar anchorage lengths do not meet the current
code requirements. The early cracking of the pedestal at a displacement ductility of
approximately 0.5 is of concern because the pedestals are buried and their damage after moderate

169
 Deflection (cm)
0 1.27 2.54 3.81 5.08 6.35
50 222.4
1.5xSylmar (1.96,44.5)
(0.2568,43.32)
1.25xSylmar 1.75xSylmar 2.0xSylmar
37.5 1.0xSylmar 166.8

Load (Kips)

Load (kN)
0.75xSylmar (Pedestal Cracking)
25 111.2
0.5xSylmar

12.5 55.6
0.25xSylmar
test values
bilinear curve
0 0
0 0.5 1 1.5 2 2.5
Deflection (in)

Figure 12. Load-deflection envelope.

earthquakes may go undetected. The damage in the column above the pedestal was minimal as
can be observed in Fig. 13.

Retrofit Details

The bottom of the pedestal is detailed with one-way hinges. The present and past studies
of one-way hinges have indicated a relatively low level of energy dissipation even when the
dowels are properly anchored. In the course of designing the retrofit for the pedestal it was
decided to enlarge and strengthen the pedestal to the point that plastic hinging would shift from
the bottom of the pedestal to the column immediately above the pedestal. The retrofit work is
currently in progress on the second as-built column. Additional steel bars are being placed to
enlarge the pedestal while maintaining one-way hinge action according to the design shown in
Fig. 14. The ends are rounded so that fiber-reinforced plastic composite jackets can be placed to

Figure 13. Damage to octagonal column after the test.

170
improved the connectivity between the column and pedestal in the horizontal direction. The
column above the pedestal will be kept as-is. Shake table testing of the retrofitted model will
reveal the adequacy of the column and the necessary level of retrofit for the column. The retrofit
details for the third as-built column will be designed after testing of the second column.

CONCLUSIONS

Several aspects of the behavior and retrofit of a major viaduct are being studied. The
results of the study of the incoherent ground motion effects suggest that internal forces can be
significantly higher than those based on uniform ground motion when all three parameters,
namely, geometric incoherency, soil characteristics under different supports, and the wave
passage effects are included. The increase in the force even for earthquakes with peak ground
acceleration (PGA) of approximately 0.4g exceeded 80 percent. The nonlinear dynamic analysis
of the entire bridge has revealed that severe column damage for response in the longitudinal
direction of the bridge could occur in an earthquake with PGA of 0.7g. Should the abutments
suffer significant damage, the severe column damage could lead to collapse. The study of single
column bents showed that, contrary to the assumed behavior, the pedestal damage can be the
main source of nonlinearity and that the damage could initiate under relatively small motions
(with PGA of 0.3g). The nonlinear dynamic analysis of the bridge, retrofit of single columns,
and study of multi-column bents are in progress.

6.75" 12" 30"

[17.15cm] [30.5cm] [76.2cm]
GENERAL NOTES
- Hatched area is existing concrete
- fy = 60 ksi [414 MPa]
13.5" - f'c = 6000 psi [41.4 MPa]
[34.3cm] - Epoxy is Power-Fast adhesive

See Plan Detail

PLAN

See Elevation Detail

6"
[15.2cm]

12"
[30.5cm]

ELEVATION SIDE

Figure 14. Retrofit details for the pedestal.

171
ACKNOWLEDGEMENT

The study reported in this article is funded by the US Federal Highway Administration
and the Nevada Department of Transportation (NDOT). The advice of Messrs Bill Crawford and
Troy Martin of NDOT is much appreciated. Several graduate students have been working on
different aspects of the project and they are thanked for their dedication. The students involved
in the parts of the study presented in this article are: C. Ayoub, D. Gang, N. Johnson, S.
Kandasamy, H. Wang, and H. Mohamad.

REFERENCES

Prakash, V., G.H. Powell, and S. Cambell, (1993) “Drain-3DX, Base Program Users Guide Version 1.0,”
Earthquake Engineering Research Center, University of California, Berkeley, November.
Su, F., J. Anderson, S. Ni, and Y. Zeng, (1998) “Effect of Site Amplification and Basin Response on Strong
Motion in Las Vegas, Nevada,” Earthquake Spectra, V. 14, N. 2, pp. 357-376.
Yang, Q., M. Saiidi, H. Wang, and A. Itani, (2002) ”Influence of Ground Motion Incoherency on Earthquake
Response of Multi-Support Structures,” Civil Engineering Department, University of Nevada, Reno, Report
No. CCEER 02-2.
Yang Q., and Y. Chen, (2000) “A Practical Coherency Model for Spatially Varying Ground Motions”, Structural
Engineering and Mechanics, 9(2), pp. 141-152.

172
 Nonlinear Seismic Response Analysis of Effects of Sliding
and Pounding of Urban Interchange Bridges with Rubber
Bearing

Li-ying Nie1, Jian-zhong Li1, Li-chu Fan1

ABSTRACT
In design and construction of urban interchange bridges, the rubber bearing is usually placed
on the top of pier directly and the shear force transmitting from superstructure depends on friction
between the interfaces. The slide might occur on the interfaces between rubber bearing and girder
bottom or pier top under an earthquake. The excessive bearing displacement caused by sliding
might induce pounding between adjacent components. In this paper, the effects of sliding for
rubber bearing and pounding between adjacent components are investigated.

_________________
1
State Key Laboratory for Disaster Reduction in Civil Engineering , Tongji University, Shanghai 200092, PRC

173
INTRODUCTION

Rubber bearing is commonly adopted in bridge construction. In design and construction of

urban interchange bridges the rubber bearings are placed on the top of pier directly and the shear
force transmitting from superstructures depends on friction between the interfaces. The slide
might occur on the interfaces between rubber bearing and girder bottom or pier top under an
earthquake. The slide occur on the interfaces between rubber bearing and girder bottom or pier top
is called rubber bearing sliding in the latter in brief.
The effects of rubber bearing sliding in the former seismic analysis were usually not
considered [1,2]. Bearing’s sliding might induce not only excessive displacement but also pounding
between adjacent components. In this paper, the effects of sliding of rubber bearing and pounding
between adjacent components are investigated.

FINITE ELEMENT MODEL

Finite element model of sliding bearing

To simulate the sliding behavior of a bearing, a three-dimensional sliding bearing element

which is assumed to be lateral isotropy based on the bilinear restoring force characteristics has
been used. The three-dimensional sliding bearing element with three translation degrees is
depicted in Figure 1(a). The element vertical deformation is assumed to exhibit linear elastic
behavior and horizontal restoring force characteristics is depicted in Figure 1(b).
In Figure1(b), the bearing force F is the horizontal resultant force ,and the displacement d
is the horizontal resultant displacement corresponding to F . The k1 is stiffness of bearing before

sliding , k 2 is stiffness of bearing after sliding , Fcr is the critical frictional force which is

expressed as
Fcr = Nµ （1）

In which N is the vertical resistant force of bearing include dynamic and dead vertical
resistant force ; µ is the frictional coefficent of sliding interface.
In three-dimensional sliding bearing element, the frictional coefficent of sliding interface µ
is assumed to be constant ,so the critical frictional force is dominated by the vertical resistant force
of bearing N .

174
 z Fy
f zj F
J y
Fcr k2
f xj f yj x
Fcr = N µ

k1

f yi f xi d Fx
I
non-sliding
f zi
phase

a. Sliding bearing element model b. Restoring force characteristics c. Criteria for sliding and non-sliding phases

Figure 1 Analytical element model of three-dimensional sliding bearing

The criteria for sliding and non-sliding phases[3] which is depicted in Figure.1.(c) is used to
judge the sliding phase of element .The resultant of the frictional forces F is expressed as:

F = Fx2 + Fy2 (2)

where Fx 、 Fy is frictional force of bearing in x- and y-direction.

In non-sliding phase the resultant of the frictional forces which mobilized at the sliding
interface is less than the critical frictional force, i.e. F < Fcr , The system will start sliding as soon as

this resultant exceeds the critical frictional force. Thus, the sliding phase of the system will take place
if F = Fcr .

Contact element model

element
fy force
f f

K K1

J d
I x
C d0 element
deformation

a. Contact element model b. Restoring characteristics

Figure 2 Analytical model of contact element

The contact element is used to simulate the pounding which is the boundary nonlinear
phenomenon by changing from one stage to the other[4,5] as depicted in Figure 2a. The two
segments moving independently at first and the element become active if the relative displacement
between adjacent components is smaller than the initial gap. So incremental equation of contact
element can be expressed as

175
 {∆f s } = k ⋅ ∆(d i − d j ) (3)

0 di − d j < d0
where {∆f s } is the incremental element force, k =  , k is stiffness of element
K1 d0 < di − d j

, d 0 is initial gap and K 1 is impact stiffness.

NONLINEAR SEISMIC ANALYSIS FOR BRIDGE WITH RUBBER BEARING

Outline of analysis

Rubber bearings for bridges have been widely used in china for more than 30 years. The
sliding might occur on interface if there isn’t any connection between rubber bearing and top
surface of pier or bottom surface of slab.
To analyze the effects of rubber bearing sliding in urban interchange bridges, dynamic
analysis of multi-span simple supported girder bridge has been carried out under minor, design
and severe earthquake. The minor, design and severe earthquake are defined as ground motion
with 63%,10% and 2% probabilities of exceedance in 50 years, respectively.
To analyze the dynamic behavior, which considered effects of rubber bearing sliding and
the pounding of bridge structure, two analytical models are established, i.e.:
Model I: the analytical model without considering the effects of pounding.
Model II: the analytical model with considering the effects of pounding.
The multi-span simple supported girder bridge consisted of 36 spans, in which 0#, 36# are
abutment and expansion joints are set on the top of 0#, 6＃, 12＃, 18＃, 24＃, 30＃ piers. In
addition, the bearings are adopted rubber bearings or PTFE. The important aspects of analytical
model is given as following:
(1) All the bottoms of pier were assumed fixed ends.
(2) The inelastic characteristics of piers were represented by the fiber beam-column element.
(3) Rubber bearings and PTFES were simulated by
three-dimensional sliding bearing element introduced in this gird gird
paper respectively, in which the frictional coefficient between
Cap bent
rubber bearing and concrete is 0.15 and frictional coefficient
between PTEF and concrete is 0.02.
(4) Girders of superstructure were simulated by the elastic
beam element.
(5) The masses of structure system were discrete as
lumped-mass. Figure 3 Convex cap bent
(6) In multi-span simple supported girder bridge, all cap bents
which are depicted in Figure 3 are convex and have retaining block on each transversal side. There
are gaps between ends of girder and bent caps or retaining blocks. Pounding might occur under an
earthquake in this situation. In analytical model II with considering effects of pounding, the

176
pounding was represented by contact element .The width of initial gap between cap bents and
girders end are 0.08m at extension joint and 0.05m where extension joint don’t exist. The width of
gaps between retaining blocks and girders end are 0.05m in the same way.
The analytical model of multi-span simple supported girder described above is depicted in
Figure 4.

detail
drawing B

detail
drawing A

a. Dynamic analytical model of multi-span simple supported girder

contact element contact element

elastic contact element contact element
beam

elastic beam
element
lumped mass sling bearing sliding bearing
fiber beam of bent cap element fiber beam element
element element
lumped mass
of bend cap

model in longitudinal model in transverse model in longitudinal model in transverse

visual angle visual angle visual angle visual angle

b. Detail drawing A of single column c. Detail drawing B of double column

Figure 4. Analytical model of multi-span simple supported girder

There are two cases for seismic wave exciting during analysis:
Case I: Combine the response resulting from 100% of the transverse input acceleration
with the corresponding response from 30% of the longitudinal input acceleration.
Case II: Combine the response resulting from 100% of the longitudinal input acceleration
with the corresponding response from 30% of the transverse input acceleration.
The typical site-specific time histories of input acceleration are provided.

177
Analysis results

Seismic response for analytical model I—without consideration of pounding

(1) Bearing displacement under minor, design and severe earthquake

Acceleration history time input along transverse axis
The rubber bearings of entire bridge don’t slide under minor earthquake. The maximum
displacement for rubber bearing which occurred at top of 33# pier is 0.7263cm.
The rubber bearings of entire bridge have slid under design earthquake. The maximum
displacement for rubber bearing which occurred at top of 9# pier is 2.995cm. Because the
maximum displacement for rubber bearing is smaller than initial gap between girder end and
retaining block, pounding will not occur.
The rubber bearings of entire bridge have slid under severe earthquake. The maximum
displacement for rubber bearing which occurred at top of 9# pier is 13.76cm. Because the
displacements for rubber bearings are larger than initial gap between girder end and retaining
block, pounding will occur.
Acceleration history time input along longitudinal axis
The rubber bearings of entire bridge don’t slide under minor earthquake. The maximum
displacement for rubber bearing which occurred at top of 1# pier is 0.6696cm.
The parts of rubber bearings on the each end of bridge have slid under design earthquake.
The maximum displacement for rubber bearing which occurred on top of 1# pier is 2.194cm.
Because the maximum displacement for rubber bearing is smaller than initial gap between girder
end and convex bent cap, pounding will not occur.
The parts of rubber bearings on the each end of bridges have slid under severe earthquake.
The maximum displacement for rubber bearing which occurred on top of 1# pier is 10.95cm.
Because the displacements for rubber bearings are larger than initial gap between girder end and
convex bent cap, pounding will occur.
(2) Peak seismic force response for piers
If the pier is under the elastic range, the ratio of the maximum moment to yield moment at
pier bottom is used to reflect on the force response behavior. The calculating results is shown in
Table I, Table II.

TABLE I. PEAK SEISMIC RESPONSE UNDER DESIGN EARTHQUAKE

Seismic wave excite in transverse Seismic wave excite in longitude

NO. Transverse Transverse
Ratio of
NO. Longitudinal Longitudinal Ratio of
Pier shear moment Pier shear moment
moment moment
(KN) (KN-m) (KN) (KN-m)
16# 5.080E+02 6.595E+03 0.2558 7# 4.424E+02 3.105E+03 0.6677
17# 5.017E+02 6.589E+03 0.2595 8# 3.947E+02 3.010E+03 0.6259
21# 5.817E+02 7.453E+03 0.2901 29# 3.924E+02 2.874E+03 0.5471

178
 TABLE II. PEAK SEISMIC RESPONSE UNDER SEVERE EARTHQUAKE

Seismic wave excite in transverse Seismic wave excite in longitude

NO. Transverse Transverse
Ratio of
NO. Longitudinal Longitudinal Ratio of
Pier shear moment Pier shear moment
moment moment
(KN) (KN-m) (KN) (KN-m)
16# 6.230E+02 7.951E+03 0.3728 7# 6.007E+02 4.181E+03 0.9573
17# 6.175E+02 7.966E+03 0.3785 27# 5.138E+02 4.668E+03 0.9243
21# 6.615E+02 8.380E+03 0.4012 28# 6.016E+02 4.917E+03 0.9993

As can be seen from the above analytical results, due to sliding of rubber bearing, the piers
are in elastic range under design and severe earthquake. The force response for piers become
smaller and the rubber bearing displacements become larger which maybe induce the pounding
between the adjacent components.

Seismic response for analytical model II—with consideration of pounding

From the above results of analysis model I, rubber bearing sliding results in the larger
bearing displacement which maybe induce the pounding between the adjacent components under
severe earthquake. So the analysis of model II is carried out.
If the pier is under the inelastic range, the ratio of the maximum curvature to yield curvature
bottom at pier bottom is used to reflect on the deformation response behavior. The calculating
results are shown in Table III, Table IV:

TABLE III. PEAK SEISMIC RESPONSE FOR MODEL II UNDER SEVERE EARTHQUAKE

Force of Moment of Shear of Curvature of

Seismic wave NO. Ratio of
pounding pier bottom pier bottom bottom of pier
input direction pier curvature
(KN) (KN-m) (KN) (1/m)
14# 3.146E3 1.330E4 1.052E3 6.203E-04 0.7794
In transverse 16# 3.491E3 1.426E4 1.061E3 6.630E-04 0.8269
22# 1.400E3 1.235E4 1.147E3 5.874E-04 0.7321
1# 7.880E3 6.396E3 2.087E3 5.881E-03 1.8965
2# 7.935E3 6.524E3 1.560E3 4.456E-03 1.4314
In longitude
34# 5.070E3 4.980E3 1.486E3 3.073E-03 1.0233
35# 8.161E3 6.880E3 2.185E3 5.637E-03 1.7334

To compare the pounding effects of bridge structure, the piers response of model I
corresponding to the Table III is given in Table IV.

179
 TABLE IV. SEISMIC RESPONSE OF MODEL I UNDER SEVERE EARTHQUAKE

Force of Moment of Curvature of Ratio of

Seismic wave No.
pounding pier bottom bottom of pier curvature
input direction pier
(KN) (KN-m) (1/m)
14# 6.706E3 5.473E2 2.459E-04 0.3165
In transverse 16# 7.951E3 6.230E2 2.926E-04 0.3728
22# 6.804E3 6.119E2 2.187E-04 0.2780
1# 1.486E3 5.641E2 4.534E-04 0.1622
2# 1.909E3 5.754E2 6.935E-04 0.2473
In longitude
34# 1.651E3 4.954E2 6.248E-04 0.2269
35# 1.307E3 4.846E2 4.070E-04 0.1479

TABLE V. MAXIMUM POUNDING FORCE FOR MODEL II UNDER SEVERE EARTHQUAKE

Seismic wave Pounding force

NO. pier
input direction (kN)
7.650E3 0#
3.339E3 7#
In transverse
3.828E3 11#
1.927E4 36#
7.880E3 1#
7.935E3 2#
In longitude
5.070E3 34#
8.161E3 35#

From the analytical results for model II, pounding of adjacent components results in
redistribution of response force for piers and the peak response of piers have large increment.
Especially while acceleration input along longitudinal axis, the short piers yield seriously. The
effects of shear, moment, curvature for piers bottom caused by pounding is depicted in Figure 5
and Figure 6.

CONCLUSION

In design and construction of urban interchange bridges, the rubber bearings are usually
placed on the top of pier directly and the shear force transmitting from superstructures depends on
friction between the interfaces. Bearing’s sliding might induce not only excessive displacement
but also pounding between adjacent components. In this paper, the analysis of this dynamic
behavior in urban interchange bridges is carried out through a multi-span simple supported girder
bridges. It can be known from the analytical results :
(1) From the seismic response for analytical model without considering pounding effects,
due to sliding of rubber bearing, the piers are in elastic range under design and severe earthquake.
The force response for piers become smaller and the rubber bearing displacements become larger
which maybe induce the pounding between the adjacent components.

180
 3000

2500
2500

pier bottom shear (KN)

2000

pounding force (KN)

2000

1500

1500
1000

1000 500

0
500
-500

0 -1000
0 10 20 30 40 50 0 10 20 30 40 50

time history (s) time history (s)

a. Response history for pounding force at the top of pier b. Response history for shear at the bottom of pier
10000
0.00035
pier bottom moment (KN.m)

8000
0.00030

pier bottom curvature (1/m)

6000 0.00025

0.00020
4000
0.00015

2000 0.00010

0.00005
0
0.00000

-2000
-0.00005

0 10 20 30 40 50 0 10 20 30 40 50

time history (s) time history (s)

c. Response history for moment at the bottom of pier d. response history for curvature at the bottom of pier

Figure 5 Analytical response history for No. 2 pier under severe earthquake with acceleration
input along transverse axis

(2) From the seismic response for analytical model with considering pounding effects,
pounding of adjacent components results in redistribution of response force for piers and the peak
response of pier have large increment . Especially while acceleration input along longitudinal axis,
the short piers yield seriously.

REFERENCES

Fan Li-chu. 1997. “Seismic design of highway bridge ”,Hong Kong :Huajie International Publishing Co.Limited，
Fan Li-chu,Hu Shi-de, Ye Ai-jun. 2001, “Seismic design of Long-span bridge” ,.People’s Communication Publishing
House (in Chinese)
R S Jangid. 1996. “Seismic response of sliding structures to bidirectional earthquake” excitation.Earthquake
Engineering and Structural Dynamics.,25,1301-1305
Parveen K. Malhotra , Moh J.Huang and Anthony F.Shakal .1995. “Seismic Interaction at Separation Joints of An
Instrumented Concrete Beidge”, Earthquake Engrg. Struct. Dyn.,24,1055-1067
Robert Jankowski, Krzysztof wilde and Yozo Fujino.1998. “Pounding of Superstructure Segments in Isolated
Elevated Bridge During Earthquakes”, Earthquake Engrg. Struct. Dyn., 27, 487-502,

181
 2000
8000

pier bottom moment (KN-m)

0
pounding force (KN)

6000

-2000

4000
-4000

2000
-6000

-8000
0
0 10 20 30 40 50
0 10 20 30 40 50

time history (s) time history (s)

a. Response history for pounding force at the top of pier b. Response history for moment at the bottom of pier

1500 0.001
pier bottom shear force (KN)

pier bottom curvature (1/m) 0.000

1000

-0.001
500

-0.002

0
-0.003

-500
-0.004

-1000 -0.005
0 10 20 30 40 50 0 10 20 30 40 50
time history (s) time history (s)

c. Response history for shear at the bottom of pier d. Response history for curvature at the bottom of pier

Figure 6 Analytical response history of No.1 pier under severe earthquake with acceleration input
along longitudinal axis

182
 Seismic Design and Analysis for Urban Viaducts With a
Double Deck

Jian-zhong Li1, Shi-de Hu1 and Li-chu Fan1

ABSTRACT

The urban viaducts with a double deck in Shanghai are now under construction.
According to the China seismic design code for of urban bridges (Draft 2001), the seismic
design and analysis for urban viaducts has been finished by State Key Laboratory for Disaster
Reduction of Civil Engineering in Tongji University. The three-level seismic design approach
corresponding the performance levels is adopted. The analysis results for different column cross
section sizes and different reinforcement ratios were investigated.

__________________
1
State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, PRC

183
INTRODUCTION

Seismic design and analysis method for bridges have been improving and advancing
based on research findings and lessons learned from past earthquakes. After recent earthquakes,
especially the 1989 Loma Prieta, 1994 Notrhridge and 1995 Hyogoken-Nanbu earthquakes,
seismic design for highway bridges has been undergoing a critical reappraisal, with the emphasis
change from “strength” to “performance”. In several countries, seismic design of highway
bridges is in the process of fundamental change [1]. In China, a new seismic design code for
urban bridges (draft 2001) [2] has been complied by State Key Laboratory for Disaster
Reduction of Civil Engineering in Tongji University. The three-level seismic design approach
corresponding the performance levels is adopted in this new code.
Based on the new code, the seismic design of the urban viaduct with a double deck in
Shanghai has been carried out. The viaduct with a double deck in Shanghai is shown in Figure 1.
The pier is Y type with an upper deck for six lanes of urban highway and a lower deck with two
lanes of urban light railway system. The superstructure for the upper deck is simple concrete
hollow plate girders and for the lower deck is simple box girders, supported on the elastomeric
bearings with a span of 30m. The upper and lower columns of the pier are rectangular columns
with initial cross section 1.5m ×2.0m and 2.0m ×2.0m, respectively. The pier has a specified
concrete compressive strength of 40 Mpa with initial longitudinal steel reinforcement ratio 1.6%.
The volumetric ratio of lateral steel reinforcement provided inside the plastic hinge range of the
columns is about 0.5%.

SEISMIC ANALYSIS METHODS AND PROCEDURES

Design Criteria

Based on the China seismic design code for urban bridges, the three-level seismic design
approach corresponding the performance levels as shown in Figure 2 was adopted for the
seismic design of the urban viaduct with a double deck. Three levels of earthquake loads, minor,
design and severe earthquakes are defined as ground motion with 63%, 10% and 2% probability
of exceedance in 50 years, respectively. Corresponding the three levels of earthquake loads, The
earthquake performance levels of I, II, and III are defined as following:
Performance I: Minor earthquakes should be resisted in the elastic range of structural
components without damage.
Performance II: The structural components may be limited damage, but full access to normal
traffic is available almost immediately following a design earthquake.
Performance III: Severe earthquakes should not cause collapse of all or parts of a viaduct
bridge; damage that does occur should be readily detectable and accessible for inspection
and repair.

184
 Figure 1. The urban viaduct with a double deck in Shanghai (unit: cm)
Earthquake load

Severe earthquake
(2%/50yrs)
Design earthquake Level II
(10%/50yrs) Level III
Minor earthquake
(63%/50yrs)
Level I Lateral Deformation

Figure 2. Relationship between earthquake load levels and performance levels

Design Methods and Procedures

For a double-deck viaduct, the upper and lower columns are chosen as the ductile
member .The potential plastic hinges should be formed in upper and lower columns of bridges
(Figure 3). Plastic hinges may occur at the top and bottom of each upper and lower column when
the seismic forces act in the transverse direction, but only at the bottom of the lower columns
when seismic forces act in the longitudinal direction.
The three-level seismic design approach for the viaduct bridge with a double deck has
evolved as outlined below:
The multi-modal spectral analysis was used to estimate longitudinal and transverse seismic
effects for structures under minor earthquakes. The total stress in concrete and
reinforcements for bridge columns under a minor earthquake, weight of structure and 50%
loads of train should be limited permissible values to insure the structure components in
elastic ranges.

185
 Nonlinear time history analysis was used determined longitudinal and transverse seismic
effects for structures under design and severe earthquakes, respectively. The elastomeric
bearings should be designed to resist the maximum shear deformation corresponding to the
design earthquake actions. The plastic rotation capacity of the plastic hinge for the upper
and
lower columns of a pier should be design to resist the maximum plastic rotation
corresponding to the severe earthquake actions.
The demands for foundations, cap beams, joints and shear forces for columns were
determined in accordance with capacity design principles. These components should be
designed to remain essentially elastic when the columns reach its over strength capacity.
The flow charts outlining the steps in the three-level seismic design approach was
presented in Figure 4.

Figure 3. Potential plastic hinge locations for double-deck viaduct

Determine Seismic Design Force under Minor Earthquakes

and Design Structure Components

Strength Verification of Structure Components No

under Minor Earthquakes

Determine Seismic Demand under Design Earthquakes and

Check the Capacities of Bearings, Connection Devices

Determine Seismic Demands under Severe Earthquake

Check Deformation Capacities of Structure No

Components under Severe Earthquake

Detailing

Figure 4. Flow charts outlining the steps in the seismic design

186
ANALYTICAL MODEL

The analytical model for the viaduct bridge with a double deck is shown in Figure 5. The
upper and lower decks were modeled by three-dimensional linear elastic beam-column elements
placed at the geometric centroid of the cross section. The inelastic three-dimensional
beam-column elements were used to model each upper and lower column of the piers. The
general interaction yield surface for a concrete section under the action of axial force combined
with biaxial bending suggested by Bresler [3] was adopted. Each elastomeric bearing in viaduct
bridges was modeled by a linear spring. The effects of rail and rail fastener were considered in
seismic response analysis for lower decks of the viaduct bridge as following:
The rails were idealized as 3D elastic beam elements;
The transverse and vertical stiffness of the rail fastener were simulated by linear spring
elements;
The longitudinal interactions between the rails and a lower deck were simulated by
nonlinear spring elements with bilinear model.

Figure 5. Analytical model for the viaduct bridge with a double deck

0.45
Severe earthquake
0.40
Acceleration coefficient

0.35
0.30 Design earthquake
0.25
0.20
0.15
Minor earthquake
0.10
0.05

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Period(s)

Figure 6. The site-specific acceleration coefficient

EARTHQUAKE LOADING

The site-specific acceleration coefficient and typical site-specific time histories of input
acceleration for a minor, design and severe earthquakes were provided. Figure 6 shows the
acceleration coefficient for a minor, design and severe earthquakes, respectively.

187
RESULTS

The computer program IPSABS developed by State Key Laboratory for Disaster
Reduction of Civil Engineering in Tongji University was used to perform linear spectral analysis
and the nonlinear dynamic time history analysis.
In order to investigate the seismic performance of the double-deck viaduct for different
column cross section sizes and reinforcement ratios, the double-deck viaduct with three types of
the column cross section size and longitudinal steel reinforcement ratios as shown in TABLE I
are analyzed and compared.
TABLE I. THE COLUMN CROSS SECTION SIZES AND STEEL REINFORCEMENT RATIOS FOR
DOUBLE-DECK VIADUCT

Cross section size Longitudinal reinforcement ratio Volumetric ratio of lateral

(cm) (%) reinforcement (0.5%)
Upper column 200×150 1.6 0.5
Type 1
Lower column 200×200 1.6 0.5
Upper column 180×150 1.2 0.5
Type 2
Lower column 180×150 1.2 0.5
Upper column 180×130 1.2 0.5
Type 3
Lower column 180×130 1.2 0.5

Based on the each column cross section and reinforcement ratio of three types that shown
in TABLE I, the shear, flexural strengths and rotation capacities at plastic hinge ranges of the
upper and lower columns were determined according to China seismic design code for urban
bridges. To assess the response, column bending moments, the maximum relative displacements
between the girder and the pier, and plastic rotation in plastic hinge ranges were calculated with
the site-specific input acceleration of a minor, design and severe earthquakes, respectively. Parts
of import results are discussed as following.

Upper and Lower Column Capacity

Mander’s model for confined concrete and unconfined concrete [4] was used to calculate
the yield bending moments, yield curvature, ultimate curvature for three types of columns. The
rotation capacities of plastic hinge of columns are estimated by [5]:
θ p = (φ u − φ y ) L P
(1)
Where φ y = yield curvature corresponding to the first yield of reinforcements; φ u = curvature
capacity at the failure limit state and L p =equivalent analytical plastic hinge length.
The calculating results for yield bending moments, yield curvature, ultimate curvature
and plastic rotation capacities for plastic hinge of columns in the direction of perpendicular the
bridge axis are presented in TABLE II.
The demands for cap beams, and shear force demands of columns for three type viaducts
determined in accordance with capacity design principles, corresponding to gravity loads are
shown in TABLE III.

188
 TABLE II. COLUMN CAPACITIES

Axial force（ Yield moment Yield curvature Ultimate curvature Rotation

kN） （kN-m） （kN-m） (1/m) capacity
Upper 0.0 9398.0 0.00167 0.082 0.0250
Column 20000.0 18910.0 0.00226 0.040 0.0110
Type 1
Lower 0.0 14320.0 0.00120 0.059 0.0153
Column 20000.0 27540.0 0.00153 0.036 0.0091
Upper and lower 0.0 5246.00 0.00152 0.069 0.0217
Type 2
Column 20000.0 14104.0 0.00210 0.029 0.0089
Upper and lower 0.0 4299.0 0.00178 0.080 0.0245
Type3
Column 20000.0 10085.0 0.00270 0.032 0.0092

TABLE III. THE DEMANDS FOR CAP BEAMS, JOINTS AND SHEAR FORCES FOR COLUMN

Type 1 Type 2 Type3

Shear demand for upper column (kN) 3230.0 2893.4 2283.0
Shear demand for lower column (kN) 7794.0 4785.6 3576.4
Bending moment demand for upper Max 24890.0 Max 15270.2 Max 13699.4
cap beam (kN-m) Min -11628.0 Min -3755.84 Min -2510.0
Bending moment demand for lower Max 25830.0 Max 15674.34 Max 14268.0
cap beam (kN-m) Min -25838.0 Min -3104.0 Min -10543.0

Response Analysis Results

The upper and lower columns in viaduct bridge for all three types are worked generally
in the elastic range under minor earthquakes and the shear deformations of elastomeric bearings
are adequate to resist the design earthquake. The bending moments for columns under minor
earthquakes in the direction of perpendicular the bridge axes for three type columns are showing
in TABLE IV. The upper and lower columns enter to plastic work range and the maximum
plastic rotation at plastic hinge range for the upper and lower columns under severe earthquakes
are shown in Figure 7 and TABLE V.

Figure 7. The plastic hinge at upper and lowe column under severe earthque

189
 TABLE IV. THE BENDING MOMENTS FOR COLUMNS UNDER MINOR EARTHQUAKES

Upper column Lower column

Max. Moment Max. Axial Min. Axial Max. Moment Max. Axial Min. Axial
(kN-m) force (kN) force (kN) (kN-m) force (kN) force (kN)
Type 1 5515.6 11113.0 9297.7 3477.67 15901.3 10988.6
Type 2 5328.0 8627.0 6460.0 2463.0 13295.0 8364.0
Type 3 5295.0 8123.0 6361.0 2243.0 12372.0 8988.0

TABLE V. THE ROTATION IN PLASTIC RANGE UNDER A SEVERE EARTHQUAKE

θ 1 andθ 2 θ 3 andθ 4 θ 5 andθ 6 θ 7 andθ 8

Type 1 2.15e-3 1.30e-3 8.01e-4
Type 2 4.10e-3 1.98e-3 8.01e-4 5.65e-4
Type 3 5.60e-3 2.70e-3 1.25e-3 7.99e-4

DISCUSSION

As mentioned as above, the demands for foundations, cap beams, joints and shear forces
for columns are determined when the columns reach its over strength capacity. A proper column
cross section size and longitudinal steel reinforcement ratio are very important to design
foundation, cap beams and joints. From TABLE III, according to Type 1 design (initial design)
of the viaduct bridge, the bending moment and shear force demands for upper and lower cap
beams are very high. As the column cross section sizes and longitudinal reinforcement ratios are
reduced, the demands for cap beams, and shear forces for columns are reduced largely.
However, the analysis results of three-level seismic design approach show that the Type
2 and Type 3 are adequate to resist minor, design and severe earthquakes, respectively,
according to China seismic code for seismic design of urban bridges.

CONCLUSIONS

Based on the new code for seismic design of urban bridges, the three-level seismic design
approach corresponding the three-level performances is adopted for the seismic design of the
urban viaduct with a double deck. The analysis methods and analysis results were investigated.
The analysis results show that a proper column cross section size and longitudinal steel
reinforcement ratio are very important to design foundation, cap beams and joints if capacity
design principles are adopted to calculate demand for these components.

190
REFERENCES

Kazuhiko KAWASHIMA, ` Seismic Design and Retrofit of Bridge’, Paper No, 2828, 12WCEE, Auckland, New
Zealand, 2000.
Stare Key Laboratory for Disaster Reduction of Civil Engineering. `China Seismic Code for Urban Bridges (Draft
2001)’, Shanghai, P.R.C, 2001.
Boris Bresler, `Design Criteria for Reinforced Column under Axial Load and Biaxial Bending’, ACI Journal.
66(4)(1960) 481-490.
J. B. Mander, M. J. N. Priestley and R. Park.. `Theoretical Stress-Stain Model for Confined Concrete’, ASCE
Journal of Structure Engineering. 114 (8)(1988) 1804-1826.
M. J. N. Priestley, F. Seible, G. M. Calvi, ` Seismic Design And Retrofit Of Bridges’, John Wiley &Sons, New
York, 1996.

191
 A Pseudodynamic Test of an Urban Viaduct with a
Double-Deck

Tian-bo Peng1, Shi-de Hu1, Jian-zhong Li1 and Li-chu Fan1

ABSTRACT

A pseudodynamic test of an urban viaduct with a double-deck was carried out in the
structural laboratory of State Key Laboratory for Disaster Reduction in Civil Engineering in Tongji
University. The test model and the pseudodynamic method are introduced. The results are shown
and investigated.

_________________
1
State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, PRC

193
INTRODUCTION

An urban viaduct with a double-deck is now under construction in Shanghai. Its upper deck
is urban arterial highway with six lanes and the lower deck is urban light railway system with two
lanes. The pier is Y type, and the superstructure for the upper deck is simple concrete hollow plate
girders and for the lower deck is simple box girders, supported on the elastomeric bearings with a
span of 30 m.
A pseudodynamic test was performed in the structural laboratory of State Key Laboratory
for Disaster Reduction in Civil Engineering in Tongji University to investigate the seismic
performance and damage condition of the viaduct. This method is a relatively new structural
seismic test method, which combines computers with actuators on line and has some advantages:
1) It’s cheaper than shaking table test;
2) Large scale or even full scale model test of large size structures is available;
3) The situation of the specimen can be observed in detail at any moment under test;
4) The effects of damage on the behavior of the structure are physically modeled;
5) Almost all kinds of structures can be tested integrated with the substructuring technique.
Pseudodynamic test method has many advantages compared with other seismic test
methods, and has become a popular and credible method abroad.
This method has some disadvantages too:
1) High performance test hardware is required;
2) Strain rate effect cannot be considered in an ordinary pseudodynamic test;
3) Excessive testing time is needed;
4) Test response is often specific to a particular input motion.

SPECIMEN AND TEST SETUP

According to the test capacity of the laboratory, a one fifth scale model shown in figure 1
was adopted. The total height of the specimen is 4.18 m including a foundation 0.44 m thick. The
distances from the bottom of lower columns to the center lines of the upper and lower cap beam are
1.76 m and 3.56 m, respectively. The distance between the two lower columns is 1.64 m. Two jacks
acted on the top of the upper cap beam and the prestressing wire strand imposed on the lower cap
beam are used to simulate the influence of the dead load on the axial force of the four columns. The
whole specimen is fixed on the floor with ten screws through the foundation, and cement mortar is
also paved between the underside of the foundation and the floor to increase the frictional force.
Based on the test data, the foundation’s movement is no more than 0.05 mm and is negligible in the
test.
According to the simulation law, a one fifth scale specimen was designed. The columns of
the specimen are designed on the basis of the height, dimensions and reinforcement content of the
prototype. The designs of cap beams and joints are in accordance with the philosophy of capacity
design, that is the columns are designed to be ductile members and the two cap beams and the joints

194
are designed to be capacity protected members, so it’s a type of strong-beam weak-column
mechanism. The sections of all the components of the specimen designed are shown in figure 2.

j acks

act uat or s

pr est r essi ng
st r and

Figure 1. Elevation of the specimen

The design strength of concrete of the prototype is 23.0 Mpa, and the design yield strength
of longitudinal reinforcements and hoops are 340 Mpa. In the specimen, the strength of concrete is
44.22 Mpa, and the yield and ultimate strength of longitudinal reinforcements are 355 and 520 Mpa,
and which of stirrups are 360 and 455 Mpa, respectively.

upper column lower column upper cap beam

foundation lower cap beam

Figure 2. Details of components of the specimen

195
 A site specific time history of input acceleration, called Gong2 wave, is shown in figure 3a).
The seismic behavior of the prototype structure is investigated systematically by increasing the
maximum of the acceleration history gradually. Because the wave is used to investigate the
prototype, the time scale is condensed in the pseudodynamic test. In order to reduce the period of a
test, the first two seconds of the condensed earthquake history is removed. The condensed Gong2
wave is shown in figure 3b). El Centro (NS) and Hanshin earthquake record not condensed are also
used. The cases of the test are shown in table I.
0 . 2 0 . 2

0 . 1 0 . 1

ac c el er at i on( g)
ac c el er at i on( g)

0 . 0 0 . 0

- 0 . 1 - 0 . 1

- 0 . 2 - 0 . 2
0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8 3 0 0 1 2 3 4 5 6 7 8 9 1 0

t i me ( s ) t i me ( s )

a) Initial wave b) The condensed Gong2 wave

Figure 3. Gong2 wave
TABLE I. CASES OF THE TEST
Case Wave Peak of the acceleration history (g)
1 Gong2 wave 0.13
2 Gong2 wave 0.20
3 El Centro wave 0.40
4 Gong2 wave 1.00
5 El Centro wave 0.60
6 Hanshin wave 0.60
7 Gong2 wave 1.20

TEST RESULTS

In the test, data of the displacement and restoring force history of the two cap beams were
recorded. Strains of some longitudinal reinforcement bars and stirrups at critical locations were also
collected. Some typical results are as follows:
Before the test, the initial tangent stiffness of the specimen was measured, and the shear
stiffnesses of the upper and lower structure were 50.0E3 kN/m and 73.5E3 kN/m, respectively.
According to the simulation law, the lumped masses on the upper and lower cap beam were 56 ton
and 34 ton, respectively. The tolerance limit for every freedom was 0.05 mm, and this was
maintained throughout the test.
In case 1, a minor and short crack appeared on the outside of the right lower column, 20 cm
above from the bottom. But the crack closed immediately. The displacement histories and hysteretic
loops of the interstory shears and drifts of the two parts are shown in figure 4. If there is no specific
note, real line and dotted line are used to indicate the data of lower and upper part of the structure

196
respectively in all the figures of this paper. All the components remained intact, until the maximum
displacement of the upper cap beam reached 7.83 mm. With the increase of the displacement, the
stiffness of the specimen declined. The equivalent shear stiffnesses of the upper and lower part were
37.2E3 kN/m and 36.9E3 kN/m, from the slope of the line between the positive and negative
maximum displacement points of each loop.

8
de[ 1]
7
de[ 2]

h i s t o r y ( mm)
6
5
4
3
2
1
0
d i s p l a c e me n t
-1
-2
-3
-4
-5
-6
-7
0 1 2 3 4 5

t i me ( s )

a) Displacement history
200 200

150 150
i n t e r s t o r y s h e a r ( k N)

i n t e r s t o r y s h e a r ( k N)
100 100

50 50

0 0

- 50 - 50

- 100 - 100

- 150 - 150

- 200 - 200
-4 -2 0 2 4 6 -4 -2 0 2 4 6

i n t e r s t o r y d r i f t ( mm) i n t e r s t o r y d r i f t ( mm)

b) Lower hysteretic loop c) Upper hysteretic loop

Figure 4. Data of case 1
In the next case, residual cracks developed first on the outside of lower columns near the
bottom, then on the inside near the top of lower columns, and last on the outside of upper columns
near the bottom. The displacement histories and hysteretic loops of the case 2 are shown in figure 5,
the equivalent shear stiffnesses of the upper and lower part are 18.3E3 kN/m and 19.9E3 kN/m.
Because more cracks were observed on the lower columns, the measured stiffness of the lower part
reduced more.

25 de[ 1]
20 de[ 2]
h i s t o r y ( mm)

15

10

0
d i s p l a c e me n t

-5

- 10

- 15

- 20

- 25

- 30
0 1 2 3 4 5

t i me ( s )

a) Displacement history

197
 300 300

200 200

i n t e r s t o r y s h e a r ( k N)

i n t e r s t o r y s h e a r ( k N)
100 100

0 0

- 100 - 100

- 200 - 200

- 300 - 300

- 15 - 10 -5 0 5 10 15 - 15 - 10 -5 0 5 10 15

i n t e r s t o r y d r i f t ( mm) i n t e r s t o r y d r i f t ( mm)

b) Lower hysteretic loop c) Upper hysteretic loop

Figure 5. Data of case 2
Greater nonlinearity was exhibited in the hysteretic loops in case 3, as shown in figure 6. In
this case, the observed surface cracks on the outside of the two lower columns near the bottom
developed in the form of fish scales. Some of the surface shelled off, but stirrups were not exposed.
Cracks distributed relatively uniformly, and 6 cracks with the space of 7 to 8 cm appeared on the
outside of the left lower column near the bottom. On the other locations of potential plastic hinge
range, which are at the top of lower columns and the bottom of upper columns, cracks appeared
relatively densely too.

100
de[ 1]
80
h i s t o r y ( mm)

de[ 2]
60

40

20
d i s p l a c e me n t

- 20

- 40

- 60

- 80
0 1 2 3 4 5 6

t i me ( s )

a) Displacement history
400 400

300 300
i n t e r s t o r y s h e a r ( k N)

i n t e r s t o r y s h e a r ( k N)

200 200

100 100

0 0

- 100 - 100

- 200 - 200

- 300 - 300

- 400 - 400
- 40 - 20 0 20 40 60 - 40 - 20 0 20 40 60

i n t e r s t o r y d r i f t ( mm) i n t e r s t o r y d r i f t ( mm)

b) Lower hysteretic loop c) Upper hysteretic loop

Figure 6. Data of case 3

198
 Along with the increase of the acceleration peak, the specimen damaged more severely. The
concrete surface on the lower columns spalled and several stirrups and longitudinal reinforcements
were exposed first. But damage of upper columns developed more quickly, and in case 5, the
interstory drift of the upper part exceeded the lower, as shown in figure 8.
300 300

200 200
i n t e r s t o r y s h e a r ( k N)

i n t e r s t o r y s h e a r ( k N)
100 100

0 0

- 100 - 100

- 200 - 200

- 300 - 300

- 50 - 40 - 30 - 20 - 10 0 10 20 30 40 - 50 - 40 - 30 - 20 - 10 0 10 20 30 40

i n t e r s t o r y d r i f t ( mm) i n t e r s t o r y d r i f t ( mm)

a) Lower hysteretic loop b) Upper hysteretic loop

Figure 7. Data of case 4
300 300

200 200
i n t e r s t o r y s h e a r ( k N)

i n t e r s t o r y s h e a r ( k N)

100 100

0 0

- 100 - 100

- 200 - 200

- 300 - 300

- 400 - 400
- 80 - 60 - 40 - 20 0 20 40 60 80 - 80 - 60 - 40 - 20 0 20 40 60 80

i nt er s t or y d r i f t ( mm) i nt er s t or y d r i f t ( mm)

a) Lower hysteretic loop b) Upper hysteretic loop

Figure 8. Data of case 5
After the concrete surface on the outside of the left upper column near the bottom crushed,
bars were bared. In case 6, damage was concentrated on the outside of the left upper column, and
the maximum interstory drift of the upper reached 136.78 mm. The bared longitudinal
reinforcements on the outside of the left upper column became buckled outward lightly in this case,
and part of core concrete crushed, which could indicate that the specimen had been destroyed. The
displacement histories and hysteretic loops of this case are shown in figure 9.

200
180 de[ 1]
160 de[ 2]
h i s t o r y ( mm)

140
120
100
80
60
40
20
d i s p l a c e me n t

0
- 20
- 40
- 60
- 80
- 100
- 120
- 140
- 160
0 1 2 3 4 5

t i me ( s )

a) Displacement history

199
 400 400

300 300

i n t e r s t o r y s h e a r ( k N)

i n t e r s t o r y s h e a r ( k N)
200 200

100 100

0 0

- 100 - 100

- 200 - 200

- 300 - 300

- 400 - 400
- 120 - 100 - 80 - 60 - 40 - 20 0 20 40 60 80 100 120 140 160 - 120 - 100 - 80 - 60 - 40 - 20 0 20 40 60 80 100 120 140 160

i n t e r s t o r y d r i f t ( mm) i n t e r s t o r y d r i f t ( mm)

b) Lower hysteretic loop c) Upper hysteretic loop

Figure 9. Data of case 6
At the last case, and soft story effect was found at the upper structure, but the lower part was
far from the ultimate state, as shown in figure 10. One of the buckled longitudinal reinforcements of
the each upper column fractured, and the structure could not return the initial position. The fractured
bar of the upper left column and its location are shown in photograph 1. In all the cases, all the cap
beams and joints remained almost intact, and almost no cracks could be found on the surface of all
these components.
300 300

200 200
i n t e r s t o r y s h e a r ( k N)

i n t e r s t o r y s h e a r ( k N)

100 100

0 0

- 100 - 100

- 200 - 200

- 300 - 300
- 60 - 40 - 20 0 20 40 60 80 100 120 140 - 60 - 40 - 20 0 20 40 60 80 100 120 140

i n t e r s t o r y d r i f t ( mm) i n t e r s t o r y d r i f t ( mm)

a) Lower hysteretic loop b) Upper hysteretic loop

Figure 10. Data of case 7

Photograph1. The fractured bar and its location

200
CONCLUSIONS

The conclusions below can be drawn from the test results of the pier model of the viaduct:
1) The potential plastic hinge locations are at the top and bottom of each column of a pier of a
double-deck viaduct, and more plastic deformation and damage are concentrated at the
bottoms.
2) Cracks on the lower columns were observed first, but damage of the upper columns was more
concentrated, and finally one of the upper columns was destroyed first.
3) The hysteretic loops of the upper and lower structure look like a bow with quite good capacity
of energy dissipation, and a certain extent of pinching effect is also shown.

REFERENCES

Zhu, B.L. 1989, Structural Seismic Test, Seismic Publishing House (in Chinese)
Fan, L.C., Li, J.Z., and Wang, J.J.2001, Seismic Design of Viaduct, People’s Communication Publishing House (in
Chinese)
Fan, L.C., Li, J.Z., and Hu, S.D.2001, Seismic Design Guidebook of Shanghai Gonghexin Road Viaduct (in Chinese)

201
 Seismic Design and Retrofit Strategies of Cable-Supported
Bridges: An Overview of Current U.S. Practice
George C. Lee, John Sun and Chuck Seim

ABSTRACT

This paper is a brief review of the current practice of seismic design and retrofit of cable-
supported long-span bridges in the United States. It may be regarded as a progress report of the review of
typical design practices and retrofit methods of long-span bridges. Full information on this subject is
currently being developed by MCEER under a research contract from the U.S. Federal Highway
Administration (FHWA).
A major objective of this paper is to provide a basis for discussion by the workshop participants
and to call to the attention of PRC and U.S. engineers and researchers about the U.S. FHWA project by
requesting their cooperation to supply available information on this subject area (case studies).

____________
George C. Lee, Multidisciplinary Center for Earthquake Engineering Research, University at Buffalo,
Red Jacket Quad, Buffalo, NY 14261, USA
John Sun, T. Y. Lin International, 825 Battery St., San Francisco, CA 94111, USA
Chuck Seim, T. Y. Lin International, 825 Battery St., San Francisco, CA 94111, USA

203
INTRODUCTION

Long-span bridges are a very significant part of the physical infrastructures of our society. In
bridge design, long-span bridges are essential structures in the bridge importance category and always
need special design attention that is different from regular bridges. In AASHTO specifications, bridges
with span length more than 500 feet are considered long-span bridges.
Typical long-span bridges are: cable-stayed bridges, suspension bridges, arch bridges and multi-
span continues deck or truss bridges. For the past two decades, more and more long-span bridges were
designed and constructed around the world, and some of them are located in high seismic areas. Multi-
support excitation, geometrical non-linearity, material non-linearity, and soil-foundation nonlinear
interaction have been considered in the analysis, and design.
The investigation of bridge damage in recent large earthquakes have occurred in high seismic
prone areas (1994 North-Ridge earthquake and 1995 Kobe earthquake) (Yashinsky 1995, Ritchie 1999)
indicate most of the long-span bridges have less seismic vulnerability than short-span regular bridges, and
no major damage or collapse had happened to long-span bridges. Because more design attention had been
given to long-span bridges, that does not guarantee the long-span bridge function or damage free if a
higher intensity earthquake (maximum credible earthquake) occurs near the bridges.
Seismic retrofit strategies of long-span bridges are very different from those used for regular
short-medium span highway bridges. Different parts of a long-span bridge may require different strategies
(hybrid approach).
In the United States, the AASHTO specifications apply to all highway bridges. Three design
methods are included in the provision, which are: coefficient and strength modification factor R (1-3.0/
maximum 5), response spectra design method with one fundamental mode or with multi-mode procedure,
and time history method. The design method selection will depend on regularity (bridge with more than 7
spans are considered irregular), importance, and seismic zone.
The hazard risk level is defined by the design spectra spectrum or design ground motion. The
design earthquake motions (10% 50 years or 500 return period) and forces specified herein are based
upon a low probability of their being exceeded during the normal life expectancy of a bridge. Bridges that
are designed and detailed in accordance with the provisions of the specification may suffer damage, but
should have low probability of collapse due to seismically-induced ground motions. A return period of
2500 years for a maximum credible earthquake might be used for critical structures such as long-span
bridges.
Each state is entitled to implement its own details, design procedure or provisions. In California,
CALTRANS has its own guideline for bridge seismic design. And in some states, 2500 return period
earthquake (maximum credit earthquake) is used for strength design for essential design. In most states,
the AASHTO provision (Division I-A) applies to bridges of conventional slab, beam girder, box girder,
and truss superstructure construction with span not exceeding 500 feet. For other types of construction
and bridges with span exceeding 500 feet, the owner (state DOT) shall specify and/or approve
appropriate provisions for design.
The AASHTO specifications are based the following principles:

1. Small to moderate earthquake should be resisted within the elastic range of the structural components
without significant damage.
2. Realistic seismic ground motion intensities and force should be used in the design procedures.
3. Exposure to shaking from large earthquake should not cause collapse of all or part of the bridge.
Where possible, damage that does occur should be readily detectable and accessible for inspection
and repair.
4. The probability of the coefficient is based on 10% exceeding probability in 50-years or 500-year
return period.
5. Acceptable damage is restricted to inelastic hinges in the columns. The foundation should therefore
remain in its elastic range.

204
 In California, the majority of the California highway bridges need dynamic analysis design. The
seismic performance criteria established by CALTRANS for design and evaluation of bridge is shown in
Table I.

TABLE I. SEISMIC PERFORMANCE CRITERIA OF CALTRANS

Ground Motion at Site Minimum Performance Important Bridge
Level Performance level
Function Evaluation Immediate Service Level Immediate Service
Repairable Damage Minimum Damage
Safety Evaluation Limited Service Level Immediate Service Level
Significant Damage Repairable Damage

Seismic performance is assessed at the following two levels of earthquake ground motion:
1. Function evaluation earthquake having 40% probability of occurring during the useful life of the
bridge (50, 100, 200).
2. Safety evaluation earthquake which is either the maximum credible earthquake based on the
conventional deterministic assessment or an earthquake with an average return period of 1000 to
2000 years.

The seismic requirement for important bridges (long-span bridges are important bridges) are
significantly raised to the following level:

1. The structure should remain essentially elastic under the functional evaluation earthquake.
2. The structure should provide service to normal traffic almost immediately following the safety
evaluation earthquake. Any damage incurred should be repairable with limited loss of service,
i.e., short closure time.

In this paper, seismic design and certain retrofit strategies of cable-supported bridges will be
reviewed and discussed.

SOME FUNDAMENTAL ISSUES OF SEISMIC DESIGN

During the past two decades, several cable-supported long-span bridges, approximately 200
meters or longer in span length, have been either seismically retrofitted or constructed in the United
States. There are many such bridges constructed throughout the world, particularly in Europe and Asia.
In essence, the design of a long-span cable-supported bridge under seismic loading, like the
design of any important bridge, must address a sequence of key issues:

• Determine the site's seismic exposure by developing the seismic hazard evaluation, seismic response
spectra, and ground motion time-histories either by deterministic methods of evaluating site-specific
historical, seismologic, and geological data, or by probabilistic methods for the region.
• Develop multi-level seismic design events based on the data developed above, in the case of the often
adopted two-level (or bi-level) approach, a lower level event that has a 50 percent probability of
occurring during the life of the bridge, and a higher level event that has a 5 percent probability of
occurring during the life of the bridge. For the usual 75-year life of an ordinary highway bridge in the
U. S., this would require developing both 150-year and 1500-year return period seismic design
events. Most long-span bridges designed today will survive with good maintenance for 150 years or
longer, so developing higher levels of the two-level seismic design event may be warranted.

205
• Establish a Bridge Performance Policy (BPP) with the owners or operators of the facility. The BPP
should consider the economic, social impacts to the local community and the surrounding areas, and
the cost to the owner of the bridge under multi-level design seismic events. The BPP is a statement of
how the owner wants the bridge to perform in small earthquakes and in large ones with the imposed
financial constraints and limited other resources.
• Transfer the bridge performance policies into a site-specific seismic design criteria document to
augment the prevailing bridge design specifications.
• Develop several concepts for seismic resisting systems (SRS) based on the gravity resisting system
developed for the bridge during the bridge preliminary design phase.
• Design a bridge structural system for these seismic demands and for the seismic design criteria.
• Analyze the structure for global, regional, and local action with linear and nonlinear analysis.
• Refine the design and detailing.
• Verify the design with critical large-scale structural laboratory testing.

More detailed discussion on the seismic design issues listed above are given in the following sections.

Design Seismic Motions

Except for a situation in which a bridge is founded on a single block of rock or for cases that the
nearest credible fault line is located at least 50 kilometers away, a rigorous structural time history
response evaluation of a long-span cable-supported bridge is fully warranted by the bridge designer. Site-
specific multiple-support rock motions and site-specific geotechnical data are essential to performing a
realistic linear and nonlinear time history analysis for a long-span cable-supported bridge.
In evaluating the site-specific multiple-support rock motions, the following factors that should be
considered are:

• Near-field fault rupture directivity effects: From the recent observations in bridge performance during
earthquakes such as the 1989 Loma Prieta Earthquake (USA), the 1994 Northridge Earthquake
(USA), the 1995 Kobe Earthquake (Japan), and the 1999 Chi-Chi Earthquake (Taiwan), bridge
structures are vulnerable to velocity pulses (called in the U.S. the "Fling Effect" because it results in a
large rapid ground displacement). A realistic set of site-specific rock time history motions should be
characterized by including a velocity pulse and by its frequency diversity.
• Vertical acceleration effects: A bridge located near an active fault (< 10 kilometers) can experience
significant vertical accelerations, which must be accounted for in the design.
• Wave-passage effects to account for traveling seismic waves.
• Compatibility with a coherency function to account for scattering and complex wave propagation
phenomena.
• Cross-correlation between the fault-normal and fault-parallel components.

When multiple-support rock motions are used in a free-field analysis to generate inputs for the
soil-structural interaction analyses and for inputs to the global bridge response model, site-specific
geophysical and geotechnical field and laboratory test data are essential to obtain realistic analyses.

Bridge Performance Policy and Seismic Design Criteria

Bridge Performance Policy is directly related to the level of importance of a bridge to a local
community and to the regions it serves. A long-span, cable-supported bridge is usually a landmark
Bridge and is often classified as “important” or “critical” infrastructure to the community it serves.
For example, in the case of the San Francisco–Oakland Bay Bridge, the State of California
classifies it as a “lifeline” structure. This level of importance requires that the bridge must provide full

206
service immediately after both a 90-year return period event, called a Functional Evaluation Earthquake,
and a 1500-year return period, called a Safety Evaluation Earthquake. The Bridge Performance Policy
further states that the bridge may experience minimal or no damage during a Functional Evaluation
Earthquake and repairable damage (that can be repaired quickly with minimum interruption of traffic
flow) during a Safety Evaluation Earthquake.
Depending on social, political, and economic factors, other landmark bridges may have different
requirements stated in the Bridge Performance Policy. Nevertheless, a Bridge Performance Policy is the
key item that needs to be fully defined at the beginning so that the seismic design criteria for the bridge
can next be established. The performance requirements of key structural elements or sub-structural
groups are defined in terms of allowable strains, deformations, and the demand-to-capacity ratios.
Again, taking the case of the San Francisco–Oakland Bay Bridge East Bay Replacement Project
as an example, “repairable damage” of the Bridge Performance Policy under Safety Evaluation
Earthquakes is ensured by the “Limited Ductility Structure” requirements of the Seismic Design Criteria.
The “Limited Ductility” rules set the limits of the maximum concrete compression strain to less than 2/3
of the ultimate concrete strain, and the maximum tensile strain in the reinforcement to less than 2/3 of the
defined ultimate steel tensile strain. In addition, the Limited Ductility Structure clauses require that the
design provides a clearly defined ductile mechanism for the response to seismic load and with limited
nonlinear deformations concentrated to a few specifically chosen elements, such as the tower shear links,
the main pier hinging regions, and the top of foundation piles. Maximum residual displacement of any of
the elements of the bridge is limited to 300 mm.
Under Functional Evaluation Earthquakes, the “minimal damage” of the Bridge Performance
Policy is ensured by the “essentially elastic” requirements of the Seismic Design Criteria. In the Seismic
Design Criteria, the term “essentially elastic” is mainly characterized by the elastic response of bridge
superstructure and the main tower with strain limits of 0.004 for the concrete in compression and 0.001
for the reinforcing steel in tension.

Seismic Bridge Concept Development

Much like the seismic design of conventional bridges, the seismic design concept for long-span
cable-supported bridges should have a clearly identifiable earthquake resisting system (ERS). The
bridge's ERS should provide a reliable and uninterrupted load path transmitting seismically induced
inertia forces into the ground. The ERS should also provide stable elastic and inelastic displacement
capacities to withstand the displacement demands generated by seismic motions. In developing the
bridge’s optimal ERS system, main span towers and approach span piers should be considered as an
integrated seismic system. It may be preferable that devices such as bearings or isolators not be used for
considerations of structural simplicity, functional reliability, and continuous maintenance. However,
seismic response modification devices can be used effectively when properly employed and when the
stand-alone ERS system can not be made adequate to meet the force and displacement demands induced
by seismic motions.
Since the main towers of a long-span, cable-supported bridge are the primary gravity-load-
carrying members, it is preferable that the seismic lateral loads from the cable-supported system are
resisted by the flanking piers and not by the main towers. This concept may require a partial or a fully-
integrated design between the approach piers and the main span piers and towers. This consideration is
further justified by the fact that post-earthquake repairs of the flanking piers are much easier to repair than
the main span towers. However, the towers must be designed for the seismic loading produced at the top
of the towers from the seismic forces in the cables.
While towers or pylons are designed traditionally to meet strength demands, their displacement
capacities become crucial in ensuring the safety of the bridges under seismic loadings. A recent
development using a multi-shaft system for the main tower design presents a new structural system
alternative to meet high demands of both axial load carrying capacity and lateral displacement capacity.
This concept is discussed in more detail below.

207
Structural Analysis to Evaluate Demands and Capacities of Bridge Elements

Overall approach of the structural analysis for a cable-supported bridge should include an
estimate of force and deformation demands from practical computer models of the bridge- i.e. models that
are as simple as possible but that incorporate all significant aspects of ground motion input and structural
response.

Global Analysis

In general, for a long-span, cable-supported bridge, the global analysis should be a three-
dimensional linear or nonlinear computer program. Both response spectrum and time history analysis
methods should be used for design of the long-span, cable-supported bridge.
Time history analysis should be used for models that are largely linear, but that include all
significant nonlinearities. These nonlinearities might include:
• Globally nonlinear geometry
• Stay-cable or suspension-cable geometric nonlinearities
• Plastic hinging of piles, piers, and towers
• Rocking of pile caps or piers
• Pounding between structural units
• The action of nonlinear devices such as dampers and restrainers
• The nonlinear behavior of critical areas of the structures such as the bases of towers or the
rocking of towers. Simplified representations (e.g. simplified finite element models) of these
areas may be included in the global model.

A global model should include appropriate representations of the bridge foundations and soil-
structure interaction. From the simple to the complex, the different approaches that might be used to
model soil-structure interaction are:
• Linear impedance-matrices and corresponding scattered input motions
• Secant impedance-matrices considering nonlinear soil and/or foundation behavior
• Scattered input motions from time history analysis of the foundation
• If nonlinear soil or foundation behavior or rocking of the foundations is particularly severe, these
behaviors may be included in the global model itself

Local Analysis

Local finite element analysis should generally be used to investigate unusual or critical portions
of the structure and to determine the behavior of the region studied, in order to incorporate that behavior
in the global model in a simplified form.
Parts of the bridge that might be the subject of local analysis are:
• Pile/pile cap connections
• Tower bases, and tower heads
• Joints between the deck and towers

Local models may also be utilized to develop the actual design of unusual portions of the
structure, for which design rules are unavailable, or to which code provisions may not apply. We have
taken this approach for the seismic retrofit of the Golden Gate Bridge to design the strengthening of the
tower bases and the supporting piers, and also for the design of the tower shear links of the new San
Francisco–Oakland Bay Bridge, which is discussed below.

208
Critical Structural Detailing

Proper detailing of structural elements is the link between the structural concept and the
completion of a successful project. Proper detailing can never be over-emphasized for design of essential
infrastructures, especially for the design of long-span, cable-supported bridges and for seismic designs.
Among the many important structural details to ensure the bridge safety against seismic actions, lateral
confinement in concrete members and compactness for post-yielding behavior in steel members are
probably the two most decisive factors that can improve bridge performance under earthquake motions.
Lateral confinement. Moderate increase of lateral reinforcement in concrete compression
members significantly enhances the post-yield displacement capacity and prevents brittle type shear
failures that have been experienced in many concrete bridges in earthquakes during recent decades. This
was shown experimentally in the early 1980s at the University of Canterbury, New Zealand, in the late
1980s and in the early 1990s at the University of California, San Diego, and in many other institutions in
the U. S.
In addition to the minimum specified transverse reinforcement ratio, currently Caltrans requires
that all hoop bars in columns, piers, towers and pylons be spliced with ultimate strength splices or
certified butt-welded hoops or be connected by certified mechanical couplers. Traditional “lap-splice”
detail that has been used for many years on numerous bridge projects is prohibited. The cost of the lateral
confinement reinforcement often is less than 1% of the overall project cost, yet this small cost will have
the single most positive effect on good performance of a bridge structure under seismic loading.
Steel plate compactness for post-yielding behavior. Buckling of steel plates is probably the
dominant failure mode in steel design under seismic loading. The key design index to control or limit this
type of failure is the plate width-to-depth ratios or “b/t” ratios. There appears to be some confusion in
current design codes for this critical design index. An example of this confusion would be the definition
categories of steel section compactness. Currently, there are four section compactness categories defined
in various codes:
Slender section. In this category, local plate buckling will occur before the yielding of the full
section occurs.
Compact Section Type A (AASHTO, AISC). In this category, no local plate buckling will occur
before the yielding of the full section occurs.
Compact Section Type B (AASHTO, AISC). In this category, yielding can be ensured for nearly
the full section and plastic capacity can be developed before the occurrence of local plate buckling. This
category can reach a strain level of 2 to 3 times the steel yield strain before any local plate buckling
occurs.
Compact section for large strain (ATC–32, AISC Addendum). In this category, the section can
endure deformations corresponding to a strain level of 5 times steel yield strain or more before local plate
buckling occurs.
In seismic design for steel members of the long-span, cable-supported bridge, the b/t ratios
corresponding to the “compact section for large strain” should be used for member sections that are
expected to endure significant non-linear or ductile deformations for displacement ductility demand equal
to or larger than 2. Extensive damage to steel frames in buildings during the Northridge Earthquake
shows how non-ductile steel members can crack if not properly detailed.
The discussion of these decisive factors is to highlight the importance of structural detailing in
bridge design, especially for long-span bridges, which are held to high standards in terms of seismic
performance and long-term structural durability.

Large Scale Design Verification Laboratory Tests

Often in long-span, cable-supported bridge design, the common state of practice in bridge
engineering is pushed, stretched, even broken to ensure the demands arising from social, aesthetic,
economic, and public considerations as well as demands from natural and man-induced forces. This

209
pushing of the edge of the state of practice cannot merely be characterized by uncommonly large size
bridge elements, but also by the selected structural system and the materials chosen.
In the case of the new East Bay replacement span of the San Francisco–Oakland Bay Bridge, a
panel of engineers and architects that represented the public interest selected a record-setting long-span
self-anchored suspension bridge. The bridge engineers, presented with the formidable challenge of
designing this unusual structure, developed a new and innovative structural system for the seismic
resisting systems. New concepts for key elements in the bridge structural system, such as a four-shaft
steel tower connected by deformable steel shear-links, a looped cable anchorage system, multi-shaft
seismic force resisting concrete piers, and ductile steel and concrete detailing are all part of essential
elements in making the bridge concept a reality.
To ensure high standards in structural safety and the durability of the 150-year design life for the
structure, key elements in the bridge structural system were tested in the structural testing laboratory in a
scale that is representative of the bridge elements.
Considering the immense investment typically associated with long-span, cable-supported
bridges, a verification laboratory test program should be an inherent part of the overall project program,
essential not only to ensure structural safety but also to control and reduce the overall cost of the project
investment and to advance the state of practice in bridge engineering.

SEISMIC RETROFIT USING RESPONSE MODIFICATION TECHNOLOGIES

Seismic retrofit for cable-supported bridges typically uses “seismic response modification
technologies. The two most important factors that affect the structural response of a bridge are stiffness
and mass distribution, which determine the frequency content and structural damping of the bridge. As
the stiffness and mass of the bridge are not uniformly distributed, irregular responses occur under random
earthquake ground motions. The key issue is to first understand the dynamic responses and develop a
retrofit strategy to modify or reduce the bridge responses. To a large extent, much of the existing efforts
in structural response modifications have been concentrated on:

• “Regularizing the irregularity”: Equalizing the stiffness of supporting elements or adding sub-
structural systems to equalize the effective seismic demands and the structural capacity.
• Isolation: Increasing the fundamental periods of vibration of the bridge to a point on the design
spectrum at which the power of the input motions is lowered.
• Dampers: Increasing the energy-dissipation capacity of the structure, thereby increasing the
effective equivalent structural damping to reduce or to limit the base shear and the displacement
demands on the substructure.

Engineers and researchers have been accomplishing these objectives with two main approaches:

• Modifying or optimizing the bridge structural response by designing innovative and engineered sub-
structural systems or elements, such as multi-shaft piers, sleeved piles, rocking piers, and ductile
shear links. This approach is often favored by bridge designers in new construction projects.
• Modifying or optimizing the bridge structural response by using structural response modification
“devices” such as isolators, which increase the structure's fundamental period and dampers, which
increase the effective damping. Additionally, there are many new developments such as active or
semi-active control systems which use devices and/or smart materials. These “smart structure”
approaches have not yet been seriously used in bridge engineering practice in the U.S. today.

Obviously, combinations of these two basic approaches can be in most cases cost-effective. This
is the basic objective of the current MCEER research task to develop retrofit strategies for long-span

210
bridges. The following are some typical structural response modification approaches. Some of them are
taken from information supplied to MCEER by U.S. bridge designers on cable-supported bridges that are
either retrofitted or in the planning and/or construction state (see Figure 1). Specific examples on
isolation bearings and various dampers are not provided because their application to cable-supported
bridges in the U.S. is only in the very preliminary stage. Many references on their implementation are
available, particularly in Japan and China. They are not reviewed in this paper.

Figure 1. Some earthquake-resisting cable-supported bridges in U.S. (new or retrofitted by response

modification technologies)

Multi-Shaft (Column) Concrete Piers

The idea of Multi-Shaft Piers is to split a classic single column pier into a multiple-column pier to
increase its flexibility. This leads to significantly increased fundamental periods of the bridge and the
displacement capacity of the pier without compromising its shear capacity. This concept was actively
explored and used in the final design of the San Francisco–Oakland Bay Bridge Suspension Span. The
resulted four-column west-anchorage piers increase the first fundamental period of vibration from 1.5
seconds to 4.0 seconds. This successfully limits the high-density power input of seismic motions within
the range of 1.0 to 2.5 seconds. Compared to a classic single-column pier design, the displacement
capacity is also enhanced from about 0.3 meters to about 2.0 meters for the 45-meter tall pier.

Multi-Shaft Steel Towers

The Multi-Shaft Concept was also used to improve seismic performance of the single tower
supporting the suspension cables of the San Francisco–Oakland Bay Bridge. The single tower was split
into four steel shafts connected together with strategically placed ductile shear links.

Sleeved-Pile Foundations
The basic idea of the sleeved pile is to introduce a “soft layer” between the bridge foundation and
the ground motion so ground-to-foundation isolation is achieved. This is especially useful in situations in
which the geotechnical profile varies under the same bridge foundation. In other situations, this concept

211
has been used to lower a pier foundation in order to achieve optimal structural response under seismic
loads.

Rocking Foundation
It has been recognized that rocking response of a tall bridge pier limits the moment applied to the
foundation pier and, consequently, the uplift to one side of the pier. The rocking is also effective in
dissipating energy, thus increasing the effective damping in the structure. This concept was studied in the
1960s, and was further studied and applied to a railroad bridge design in New Zealand in the 1970s.
Recently, the rocking response mechanism was used for the retrofit design of the steel towers of the
Golden Gate Bridge. Special detailing was required to adequately reinforce the lower section of the
towers to withstand the rocking pressure. The tops of the two concrete piers were also reinforced by pre-
stressing cables to withstand tower base impact loading.

Metallic Dampers
There are numbers of damping devices available today. The underlying energy dissipation
mechanism is derived from the inelastic deformation of a metal, usually steel, or an alloy such as lead, to
achieve a stable elastic-plastic behavior.

Friction Dampers and Friction Bearings

Friction dampers utilize the mechanism of Coulomb friction between sliding plates of special
material to provide energy dissipation. A special class of friction damper, using high-strength bolts,
depends on galling action on soft metal inserts. One friction-bearing device combines the concept of a
sliding bearing with that of pendulum action, which both dissipates energy and lengthens the fundamental
structural period. Even though friction dampers have not been used to any extent in large bridges because
they have small energy-absorbing capacities, the friction/pendulum bearing has been used with great
economy and structural efficiency.

Viscoelastic Bearing/Dampers
The basic material used in viscoelastic bearing/dampers are layers of copolymer materials
interlayered with steel plates to form a stacked bearing that can carry vertical loads, and, when deformed
by shear action, they deform and dissipate energy and increase the fundamental period of the structure.
They have been used in a number of short-span bridges and in buildings, but rarely in long-span bridges
probably because the heavy weight of the bridge requires very large diameter and very thick bearings that
are difficult to manufacture.

Viscofluid Dampers
Unlike the viscoelastic dampers that use inelastic deformation as the mechanism for energy
dissipation, a viscofluid damper converts the mechanical energy into heat as the piston of a cylindrical
damper forces a highly viscous fluid from one chamber to another through an orifice. The viscous fluid
also can be placed in a rectangular container and be heat generated by vanes moving through the viscous
fluid by mechanical action.
The viscofluid dampers have one other characteristic: they allow movements at slow rates and
will provide resistance and energy dissipation when imposed by displacement at higher rates. These
characteristics allow them to be used in situations where movements induced by creep, shrinkage,
temperature, or traffic should not be restrained under normal service conditions.
Viscofluid dampers will be installed on the 1280-meter main suspension span of the Golden Gate
Bridge seismic retrofit between the towers and the stiffening trusses for the main span and for the side
spans. These types of dampers also will be installed at the same locations for the seismic retrofit of the
twin 704-meter main suspension spans of the San Francisco-Oakland Bay Bridge. However, the new
728-meter main suspension span currently under construction across the Carquinez Straits near the San

212
Francisco Bay does not utilize viscofluid dampers at the towers because the steel box girder is suspended
continuously from end to end and passes through the tower legs.

Tuned Mass Dampers

The principle of a Tuned Mass Damper (TMD) is that the out-of-phase motion of a small spring-
mass system can reduce or “damp” out the fundamental period of vibration of a much larger spring-mass
system. TMD applications for major bridges in seismic active zones have not been wide spread. This is
probably due to the fact that the frequency bandwidth of a typical seismic motion is rather wide. This
high-frequency content tends to generate higher modes of vibrations in the structure and the TMD, tuned
to the fundamental frequency of the structure, can suppress little of the dynamic response of higher modes
in the structure.
TMD was used very effectively for the erection of the towers of the Akaski Kaikyo Bridge and
are still in service inside the towers since the bridge was completed. A torsion TMD was installed in
1987 on the Bronx-Whitestone Bridge to effectively suppress the first asymmetric torsion mode by wind
excitation.

Tuned Liquid Damper

The basic principle in applying a Tuned Liquid Damper (TLD) to mitigate the structural response
is similar to that of a Tuned Mass Damper (TMD). Unlike TMDs, the response of a TLD is highly
nonlinear, partially because of liquid sloshing or the presence of orifices. In practice, a TLD does have
several advantages over a TMD such as low installation and maintenance cost and long-term reliability of
the damping mechanism.
The above are only typical examples. There are many other approaches used in the U.S. that are
given in the References without further explanation in this preliminary review.

SUMMARY

This paper is a preliminary review of the recent seismic design and retrofit practice in the U.S. It
is presented with two specific purposes: (1) to provide a platform for discussion by the workshop
participants to identify future research needs and subject areas suitable for joint research projects to be
carried out by PRC and U.S. researchers, and (2) to inform the PRC and U.S. earthquake engineering
community of the MCEER project on the development of seismic response modification technologies for
long-span bridges and to request their cooperation to provide case studies to MCEER so that a more
comprehensive technical volume can result from this study.
The authors apologize for the incompleteness in the review of published information of U.S.
efforts on this subject matter.

ACKNOWLEDGEMENTS

The authors acknowledge the financial support of the U.S. Federal Highway Administration
(Contract #DTFH61-98-C-00094). Additional support has also been provided by NSF (Contract #EEC-
9701471) and State University of New York (Contract #C-000591). They also wish to express their
appreciation to those organizations which provide the design and retrofit information on the bridges
indicated in Figure 1 (Buckland and Taylor, Ltd.; Caltrans; Modjeski and Masters; Parsons Brinckerhoff;
Texas Department of Transportation; T.Y. Lin International; and Washington State Department of
Transportation).

213
REFERENCES

AASHTO. 1998. LRFD Bridge Design Specifications, 2nd Edition, Washington, DC: American Association of State
Highway and Transportation Officials.
AASHTO. 2001. Guide Specifications for Seismic Isolation Design, Washington, DC: American Association of
State Highway and Transportation Officials.
Abbas, H., S.P. Singh, and J. Uzarski. 1988. “Seismic Evaluation and Retrofit of Sacramento River Bridge at Rio
Vista,” 6th US National Conference on Earthquake Engineering, Seattle, Washington.
Aiken, I.D. and J. M. Kelley. 1990. “Earthquake Simulator Testing and Analytical Studies of the Two Energy-
Absorbing Systems for Multistory Structures,” Report No. UCB/EERC-90/03, University of California,
Berkeley.
Giacomini, M.C. and J.E. Woelfel. 1997, “Retrofitting the Golden Gate Bridge,” J. Civil Engineering, 67(10).
Ingham, T.J., S. Rodriguez, M.N. Nader, F. Taucer and C. Seim. 1995. “Seismic Retrofit of the Golden Gate
Bridge,” Proc. National Seismic Conference on Bridges and Highways, San Diego, CA, December 10-13.
Jung, H.J., B.F. Spencer, Jr. and I.W. Lee. 2002. “Seismic Protection of a Benchmark Cable-Stayed Bridge Using
Magnetorheological Dampers,” Seventh U.S. National Conference on Earthquake Engineering (7NCEE) –
Urban Earthquake Risk, Boston, MA, July 21-25.
Matson, D. 1998. “Experience with Seismic Retrofit of Long-span Bridges,” Structural Engineering World Wide,
Elsevier Science Ltd.
Moon, S.J., L.A. Bergman and P.G. Voulgarios. 2002. “Sliding Mode Control of a Cable-Stayed Bridge Subjected
to Seismic Excitation,” Seventh U.S. National Conference on Earthquake Engineering (7NCEE) – Urban
Earthquake Risk, Boston, MA, July 21-25.
Murphy, T.P. and J.M. Kulicki. 2002. “Seismic Issues Affecting the Proposed New Mississippi River Bridge at St.
Louis, Missouri,” 3rd International Workshop on Performance-Based Seismic Design and Retrofit of
Transportation Facilities, Tokyo Institute of Technology, Tokyo, Japan, July 9-11.
Nader, M. and T.J. Ungham. 1995, “Seismic Retrofit of the Towers of the Golden Gate Bridge,” Proc. National
Seismic Conference on Bridges and Highways, San Diego, CA.
Park, K.S., S.W. Cho, I.W. Lee and H.J. Jung. 2002. “Hybrid Control Strategies for Seismic Protection of a
Benchmark Cable-stayed Bridge,” Seventh U.S. National Conference on Earthquake Engineering (7NCEE) –
Urban Earthquake Risk, Boston, MA, July 21-25.
Priestley, M.J.N., F. Seible and G.M. Calvi. 1996. Seismic Design and Retrofit of Bridges, John Wiley & Sons, Inc.
Reno, M.L. and M. Pohll. 1997. “Seismic Retrofit of the San Francisco-Oakland Bay Bridge, West Crossing,” Proc.
National Seismic Conference on Bridges and Highways, Sacramento, CA, July.
Somerville, P. 2002. “Characterizing Near Fault Ground Motions for Design and Evaluation of Bridges, Proc.
Third National Seismic Conference and Workshops for Highways and Bridges, April, pp. 137-148.
Stroh, S. and T. Lovett. 1990, “The Houston Ship Channel Cable-Stayed Bridge,” Proc. 6th US-Japan Bridge
Engineering Workshop, Lake Tahoe, NV, May.
Tang, M.C. 1995. “Seismic Design and Isolation of Long-Span Cable-Stayed and Suspension Bridges.” Proc.
Annual Seminar on Construction for Earthquake Hazard Mitigation – Seismic Isolation Retrofit of Structures,
Buffalo, NY, pp. 167-176.
Tang, M.C. 1998. “Cable-stayed Bridges,” Structural Engineering World Wide, Elsevier Science Ltd.
Trachtenberg, E. 1990, “Brooklyn Bridge Rehabilitation Design and Construction,” Proc. 6th US-Japan Bridge
Engineering Workshop, Lake Tahoe, NV, May.
T.Y. Lin International/Moffat & Nichol Joint Venture. 2000. Description of Design and Analysis Approach to the
Self Anchored Suspension Span of the New San Francisco-Oakland Bay Bridge, prepared for California
Department of Transportation, September.

214
 Seismic Response Analysis of Wuhu Yangtze River Bridge
Xi Zhu1

ABSTRACT

The Wuhu Bridge is a cable-stayed bridge with steel truss girder over the main navigable
pass of Yangtze River. The seismic ground motion transformed problems are studied at Wuhu
bridge site in this paper. The analysis methods of soil-pile-pylon interactions are investigated.
The number of modes to provide sufficiently accurate results in this complex bridge structure is
discussed. The response spectrum method and the time history analysis method using the
identical seismic exciting inputs are compared. The time history analysis of multi-support
exciting and response spectrum method for incoherent support motions are implemented. A new
improved multi-support response spectrum (IMSRS) method is proposed and comparing the
results with that of MSRS method, it is reliable, feasible and timesaving.

_________________________
1
College of Civil Engineering and Architecture, Northern Jiaotong University, Beijing, 100044, China

215
INTRODUCTION

The long-span (180m+312m+180m) cable-stayed bridge (Fig. 1) is built at Wuhu city to

cross the main navigable pass of Yangtze River. In fact, it is a continuous truss bridge with the
extra doses. It is a highway and railway two-usage bridge. The height of pylon over the bridge
deck is only 33.2m.The ratio of pylon height over the length of main span is only 0.106.It is less
than the normal value of cable-stayed bridge. The engineering geologic condition of the bridge
is more complicated and the conditions of every pier (of Pylon) basis are variant remarkably.
The covering soil of the long pylon is 27m depth which is from –43m to –16m in altitude. The
bridge foundations are all using large diameter bored piles through the covering soils to reach
the rock bed. It is obvious that the soil-pile-pylon interaction of each pylon is different. In order
to calculate the seismic responses and dynamic behaviors of each pylon considering the
interaction of soil-pile-pylon respectively, the simpler model called Single Pylon Model (SPM)
is used. In order to attest the rationality of SPM, more elaborate model called Whole Bridge
Model (WBM) is used also.

Figure 1 Model of whole bridge

SEISMIC GROUND MOTION AND SOIL-STRUCTURE INTERACTION

The analysis of soil-pile-pylon interaction in SPM mainly consists of two steps. The first
step is the free field analysis of earthquake response and the second step is the soil-pile-pylon
interaction analysis in which the pile-foundation elasticity-confined to the field is considered.
The boundary conditions of far field are provided by the results of the free field analysis.

Seismic response analysis of the free field

The assumptions about the soils are：the surface of the site is horizontal; the soil in one
layer is homogeneous; the soil is boundless. The one-dimension soil column model is adapted to
simulate the free field. The kinetics equation of the free field is
M GU&& G + C GU& G + K GU G = − MI G u&&g (1)

216
in which, UG displacement vector of 3-dimensional seismic response; u&&g acceleration of the base

rock; MG mass matrix with the diagonal elements m G i = 1 ( ρ i hi + ρ i +1 hi +1 ) ;KG stiffness matrix with
2
2Gi 1 − γ i Gi G
vertical and horizontal elements respectively K G wi = ( ) and K G ui = ;C Rayleigh
hi 1 − 2γ i hi

damping matrix. In preceding formula: ρ i , hi , Gi , γ i are mass per meter, height, shear modulus,

Poisson ratio of the ith layer respectively.

The non-linearity of soil is dealt with equivalent linearity method in which the
equivalent stiffness and equivalent damping ratio were approached by iteration. The special
program is compiled for this purpose.

Design Seismic and input motions

According to Chinese code for Aseismic Design of Railway Engineering at referring to

other codes at home and abroad. The seismic input ground motions at the bed rock are provided
by the Seismic Bureau of Anhui Province, where the bridge is located. The earthquake ground
motion inputs with the exceeding probability in 100 years is 10% was taken for the Functional
Evaluation Earthquake. The earthquake ground motion inputs with the exceeding probability in
100 years is 2% was taken for the Safety Evaluation Earthquake.
As the results of seismic risk analysis at the bridge site, there are 12 input motions
provided, which include 6 in horizontal and 6 in vertical respectively. One of six input motions
in each direction is shown in Figure 2. Its maximum acceleration is 0.952 m in horizontal
s.s
direction and 0.436m/ss in vertical direction respectively.

Single Pier(Pylon)Model

A fictitious pile simulates the group piles. The equivalent spring’s stiffness at the bottom
Q
of cap slab caused by pile’s support is computed according K Φu = ∑ x 2 i k pi ,where, Q the total
i =1

number of piles, kpi the axial stiffness of the ith pile according to Sato assumption, xi the
coordinate of the ith pile. The SPM is shown in Figure3.The characters of the pylon cross
sections are shown in TABLE I. The stiffness matrix is assembled by beam elements. In order to
reduce the total freedom of the system for accounting soil-pile-pylon is located reasonably with
lumped mass in SPM. In order to ensure the comparability, the calculating data in SPM is
conformed with those in WBM.

217
 The masses and stiffness influences of the stiffened steel truss and the stayed cables
were simply transformed to the piers and pylons. It was built the Single Pier (Pylon) Model.
The allocated mass of main truss and other auxiliary in SPM are shown in TABLE I.

TABLE I.THE ALLOCATED MASS OF MAIN TRUSS AND OTHER AUXILIARY IN SPM(T)

Long pylon Short pylon

Long. Ver. Tran. Long. Ver. Tran.

Joint of pylon and truss 1771 6833 3574 5597 7565 10179
Top of pylon 3330.7 168.3 382.6 4916.0 168.3 548.57

Seismic time history responses for three orthogonal directions located at the pile cap
considering soil-pile-pylon interaction were calculated, to get the seismic response spectra at the
cap slab in three orthogonal directions respectively. That will be the earthquake inputs when the
seismic response of whole bridge is investigated. It will be also provided the spring’s stiffness
coefficients of the fore cap slabs to account the effects of soil-pile-pylon. It was iterated
accurately to approach the constrain condition of the pile foundation, when it was calculated the
natural vibration characters and the seismic response with the whole bridge model.

Equation of Soil-Pile-Pylon interaction

The equations of motion considering the soil-pile-pylon interaction are

n
(mi + mi )u&&i + ∑ cij u& j + ci u& i + k i u i = − mi u&&g + mi u&&i + ci u& i + k i u i
p s p s s p s G s G s G
(2)
j =1

where, ui the relative displacement of the pile and pylon; uiG the relative displacement of the soil.

Equivalent Parameter of Soil-Pile interaction

The equivalent horizontal stiffness between soil and fictitious pile are calculated by
Mindlin formula and Elasticity Winkler Assumption. The equivalent vertical stiffness between
soil and fictitious pile are calculated according Sato assumption. The equivalent masses of the
soil-pile interaction were calculated by the energy equivalent theory.

Seismic Response Spectra

The response spectrum analysis method is used as one method to calculate the seismic
response in WBM. Using those 12 seismic input motions of bed rock, 18 acceleration time
histories-six respectively in each of longitude, vertical and transverse directions-at top of cap
slabs of the pile foundation are analyzed by SPM to get the response spectra at the same

218
location. Six response spectra are obtained in each direction by Duhamel Integral. The envelope
curve of the six response spectra is used as the input motion spectra.

Analysis Results

The Stiffness Coefficient of the Foundation Cap Slab

The foundation spring stiffness of translation, torsion and the coupling each other term
that are calculated by SPM are shown in TABLE II.These spring’s coefficients are the constraint
conditions at the cap slab to alternate the pile foundation in the WBM.

TABLE II. THE CONSTRAINT COEFFICIENTS AT CAP SLAB

Long pylon Short pylon

Length of piles M 30 20
Number of piles 19 17
Diameter of piles M 3.0 3.0
VE translation KN/m 1.683E8 4.085E8
TR translation KN/m 0.116E9 0.439E9
TR coupling KN/rad -1.730E9 -7.053E9
TR rotation KN*m/rad 34.72E9 151.47E9
AL translation KN/m 0.115E9 0.439E9
AL coupling KN/rad 1.718E9 7.503E9
AL rotation KN*m/rad 34.86E9 151.47E9

Envelope Curve of the Response spectra

The envelope curves of the response spectrum at the top of cap slabs which are shown in
Figure 4 and TABLE III were calculated by SPM. Those spectra were the input spectra in the
WBM.

219
 TABLE III.THE ENVCELOPE CURVES OF RESPONSE SPECTRUM AT THE TOP OF CAP SLABS

α max β max × α max β min

Displacement
β min × α max
m ( s * s) β max α max at period start Period
(mm)
m (s * s)
m ( s * s) s s
Long AL 3.717 1.4476 5.01996 7.26689 0.20 0.46 0.4343
pylon VE 2.785 1.0735 4.29193 4.60738 0.20 0.45 0.3221
TR 3.240 1.2041 4.55865 5.4888 0.12 0.52 0.3612
Short AL 1.073 0.4034 4.22372 1.7038 0.28 0.56 0.1210
pylon VE 0.389 0.2784 4.02578 1.1208 0.08 0.44 0.0835
TR 1.180 0.4103 5.02317 2.0610 0.28 0.60 0.1231

Note: AL-along the axis of the bridge; TR-transverse the axis of the bridge; VE-verticality

Whole Bridge Model

The input locations of seismic ground motions in the Whole Bridge Model are put on the
pile cap slabs of four pylons (or piers). The input seismic response spectra are taken from the
SPM which is considered the soil-pile-pylon interaction and is enveloped six spectra with same
probability condition. The geometric nonlinear effects for cable-stayed are included in the
WBM.

220
Free Vibration Characteristics

The dynamic behaviors of the bridge are calculated by the two proposed models
respectively. The natural periods are shown in the TABLE IV and the SPM results are identical
with those of WBM.

TABLE IV. THE PERIODS OF BRIDGE IN WBM AND SPM(S)

WBM Character of WBM Long Short pylon Character of SPM

pylon

1 2.6843 Truss, TR, symmetrical bending

2 2.4420 Two pylons, AL, floating 2.4369 2.55240 Two pylon, AL
3 2.2460 Truss and pylon, VE, symmetrical bending
4 1.7088 Long pylon and relevant beam, TR 1.73718 Long pylon, TR
5 1.5718 short pylon and relevant beam, TR 1.68529 Short pylon, TR
6 1.3606 short pylon and relevant beam, TR, torsion
7 1.3446 long pylon and relevant beam, TR, torsion
8 1.2982 Short pylon, TR 1.32631 Short pylon’s limb TR
9 1.1666 Long pylon, TR 1.23953 long pylon’s limb TR
10 1.1624 Russ and pylon, VE, anti-symmetrical
bending
11 1.1484 Truss torsion

The tangent stiffness matrix of the bridge in its dead load deformed configuration,
abstained through an iterative nonlinear static analysis, is utilized in solving the eigenvalue
problem. For comparison, four programs were used to calculate the free vibration
characteristics. They are the special code 1 by ourselves to account geometric nonlinear effect
under dead load, the special code 2 by ourselves and SAP90 and SAP93.The last three codes
can’t include the nonlinear effects. By examining the computed modes of vibration, the
following comments can be made:
(1) Most modes can be categorized as longitudinal, vertical, translateral and purely cable
modes.
(2) Strong coupling in the three orthogonal directions occurs in a number of modes. The
dominate lateral motion of the bridge truss is strongly coupled with torsion deck and tower
vibrations, more over the primarily torsion vibrations are associated with lateral truss motion, and
cable vibrations are also present within these modes.
(3) The bridge towers’ longitudinal dominant motion is associated with vertical truss
vibrations.

221
 (4) The nonlinear effects are not obvious. Because this bridge has a strong stiffen truss, so the
natural period of first three modes dominated by the stiffen truss are prolonged less than 3% on
considering that the geometric nonlinearity is compared with another. But the natural period from
the fourth to ninth which are dominated by the tower are prolonged about 7% to 17% respectively,
and the effects of geometric nonlinearity have increased.

SEISMIC RESPONSE OF CABLE-STAYED BRIDGE

Modal contribution ratio

The mode superposition method is widely recognized as a powerful method for

calculating the dynamic response of linear structural system with classical damping. The method
is attractive because the response of a multi-degree-of-freedom system is expressed as the
superposition of modal response, and each modal response is determined from the dynamic
analysis of a single-degree-of-freedom system, and these dynamic analysis needs to implement
only a few of modes which are significant to the response of total system. Then lots of the
calculating efforts will be safe. The number of modes to be included partly depends on modal
contribution factor. That will refer to these three ideas, namely, modal participation coefficient
vector, modal mass and modal contribution ratio.
The modal participation coefficient vectors are defined as
Fx = Φ T ME x , Fy = Φ T ME y , Fz = Φ T ME z (3)

in which, M the mass matrix of system, Φ modal vector, and E x , E y , E z as

E x = [1,0,0,1,0,0...]T , E y = [0,1,0,0,1,0...]T , E z = [0,0,1,0,0,1...]T

The modal masses are the important parameters implying that it is a measure of the
degree to which the ith mode participant in the dynamic response.
The modal masses of each mode and each direction are defined as
2 2 2
M xi = Fxi , M yi = Fyi , M zi = Fzi (i = 1,2,..., N ) (4)

The modal contribution ratio is defined as a ratio, each modal mass is normalized by
total mass.
The modal contribution ratio of the ith mode in x direction as: M xi
γ xi = Nx
, i = 1,2,..., N
∑m
j =1
xj

Combining the response contributions of the three orthogonal direction, the total modal
contribution ratio of the ith mode:

222
 M xi + M yi + M zi (5)
γi = Nx Ny Nz
, i = 1,2,..., N
∑m
j =1
xj + ∑ m yj + ∑ m zj
j =1 j =1

In practice, on using mode superposition the modal analysis method can be truncated,
and depends on a measure of contribution of the mode to a response quantity. If we take the first
n modes to calculate the dynamic response, the total contribution ratio for the first n modes are,
respectively
n n n n
R x = ∑ γ xi , R y = ∑ γ yi , R z = ∑ γ zi , R = ∑ γ i (6)
i =1 i =1 i =1 i =1

The value of R approaching 1, means the contribution of first n modes are enough.
The relationship of the number of accumulated free vibration modes and the value of
accumulated total contribution ratio for the Wuhu Yangtze River main navigable bridge is
shown in Fig. 5. The total accumulated contribution ratio is reached 0.62 when 30 modes are
used, and the R is reached 0.88 when 100 modes are used. R is reached more than 0.90 when
130 modes are used. For this complex long-span bridge, it may be adequate to require R=0.9 or
to cut off at first 100 modes.

Response spectrum analysis under identical excitations

Though we use same seismic bed rock inputs with same exceeding probability of 0.1 in
100 years, but the obtained response spectrum respectively at 4 cap slabs of pile foundation are
different. The response spectrum of three orthogonal directions at the cap slabs of 10# pylon was
selected to conservatively act as the identical seismic inputs for whole bridge. The response
spectrum analysis is a procedure for dynamic analysis of a structure subjected to earthquake
excitation, but it reduces a series of static analysis. The peak value r0 of the total response r(t) is
estimated by combining the peak modal response rio(I=1,2,…n) according to the complete
quadratic combination(CQC) rule. The major results calculated by response spectrum analysis
with CQC rule are shown in the TABLE Ⅴ, in order to attest the modes cut off number at 100
is appreciable.

223
TABLE V.THE RESULTS OF BRIDGE’S IMPORTANT LOCATION WITH VARIOUS MODES’ NUMBER

q30 q100
q130 q130 q130
Items

10#trMtrans.(t-m) 0.878 1.000 52266

10#trQtrans.(t) 0.358 1.000 1947.9
10#trMlog.(t-m) 0.986 1.000 77406
10#trQlog.(t) 0.827 1.000 1267.4
10#trN (t) 0.650 0.967 429.48
11#trMtrans.(t-m) 0.756 1.000 57393
11#trQtrans.(t) 0.300 1.000 2229.6
11#trMlog.(t-m) 0.907 1.000 95432
11#trQlog.(t) 0.453 1.000 2888.5
11#trN (t) 0.530 0.995 529.02
10#tp ∆ trans.(cm) 0.979 1.000 4.939
10#tp ∆ log.(cm) 0997 1.000 7.848
11#tp ∆ trans.(cm) 0.994 1.000 5.285
11#tp ∆ log.(cm) 0.974 1.000 8.360
BM ∆ trans.(cm) 0.936 1.000 11.65
BM ∆ log.(cm) 0.999 1.000 7.879
BM ∆ ver.(cm) 0.809 1.000 3.228
(note：tr-tower root; tp-tower top; trans-tranverse; log-longitudial; ver-vertical; M-moment; Q-shear force;
∆ -displacement; BM-middle of bridge; q30 means the contribution of the first 30 modals, and so forth
q100, q130; the note of TABLE Ⅵ and TABLE Ⅶ is same to TABLE Ⅴ.)

Time history analysis of multi-supported linear structures

Using 12 acceleration response spectra of degrees of freedom at 4 top cap slabs of pile
foundation, to produce the correspondent artificial seismic records, estimating the time history
for every modal coordinate in the first m modal subspace, it has calculated the response values
of the modal coordinate time history Z(t) as following expression
m m n
Z (t ) = ∑ a k u k (t ) + ∑∑ bki s ki (t ) (7)
k =1 k =1 i =1

in which, a k = q T γ k , k = 1,2,..., m , bki = q T Φ i β ki , k = 1,2,..., m, i = 1,2,..., n

224
 The first part of this response equation represents the components of pseudo static
response, and the second part represents the components of dynamic response. It is concerned
about the effects of ‘local site’ through the various exciting inputs.

TABLE VI. THE MAXIMUM VALUE UNDER IDENTICAL EXCITATION COMPARED WITH THE VALUE
OF RESPONSE SPECTRUM ANALYSIS

Items THistory Qh RSpectrum Q s Qh Q s

10#trMtrans.(t-m) 46383 45903 1.010
10#trQtrans.(t) 669.49 696.62 0.961
10#trMlog.(t-m) 72043 76323 0.944
10#trQlog.(t) 1076.9 1047.9 1.028
10#trN (t) 205.82 279.12 0.737
11#trMtrans.(t-m) 40946 43401 0.943
11#trQtrans.(t) 646.47 699.27 0.924
11#trMlog.(t-m) 70849 86513 0.819
11#trQlog.(t) 1063.90 1309.40 0.812
11#trN (t) 284.95 280.36 1.016
10#tp ∆ trans.(cm) 4.363 4.837 0.902
10#tp ∆ log.(cm) 6.424 7.826 0.821
11#tp ∆ trans.(cm) 4.408 5.253 0.839
11#tp ∆ log.(cm) 6.179 8.143 0.759
BM ∆ trans.(cm) 12.816 10.9 1.176
BM ∆ log.(cm) 6.202 7.875 0.788
BM ∆ ver.(cm) 3.588 2.610 1.375

225
 TABLE VII. THE MAXIMUM RESPONSE VALUE COMPARED WITH IDENTICAL INPUTS AND
MULTI-EXCITED INPUTS

Items Q1 Q2 Q2 Q1
10#trMtrans.(t-m) 46383 37004 0.798
10#trQtrans.(t) 669.49 755.13 1.128
10#trMlog.(t-m) 72043 62005 0.861
10#trQlog.(t) 1076.9 1574.1 1.437
10#trN (t) 205.82 226.01 1.098
11#trMtrans.(t-m) 40946 25197.82 0.615
11#trQtrans.(t) 646.47 579.38 0.896
11#trMlog.(t-m) 70849 58976 0.832
11#trQlog.(t) 1063.90 1474.41 1.386
11#trN (t) 284.95 174.60 0.613
10#tp ∆ trans.(cm) 4.363 4.742 1.087
10#tp ∆ log.(cm) 6.424 3.822 0.595
11#tp ∆ trans.(cm) 4.408 1.925 0.437
11#tp ∆ log.(cm) 6.179 4.433 0.717
BM ∆ trans.(cm) 12.816 8.998 0.702
BM ∆ log.(cm) 6.202 3.982 0.642
BM ∆ ver.(cm) 3.588 3.719 1.037

RESPONSE SPECTRUM METHOD TO SPATIALLY VARYING GROUND MOTION

Multi-support response spectrum (MSRS) method

Recent awareness of the spatial variation of earthquake ground motion, and observations
during recent earthquake, notably the Loma Prieta earthquake of October 17, 1989, have clearly
demonstrated that seismic ground motions can vary significantly over distance which are of the
same order of magnitude as the dimensions of some extended structures, such as long-span
bridges. Three phenomena are responsible for these variations：(1) the difference in the arrival
time of seismic waves at different stations, denoted as the “wave passage ”effect; (2) the loss of
coherence of the motion due to reflections and refraction of the waves in the heterogeneous
medium of the ground, as well as due to the difference in the manner of superposition of waves
arriving from an extended source at various stations, denoted as “incoherence” effect; (3) the
difference in the local soil conditions at each support and the manner in which they influence
the amplitude and frequency content of the bedrock motion, denoted as the “local” effect.
Nakamura et al. (1993) used the approximate multi-support response spectrum method (MSRS)
developed by Der Kiureghian and Neuenhofer (1992), which accounts for spatially varying
earthquake ground motion to analyze a complex three-dimensional (3D) model of the Golden
Gate Bridge. The method is based on fundamental principles of random vibration theory and

226
properly accounts for the effects of correlation between the support motions as well as between
the modes of the structure. The method works best when the significant segment of the
excitation is quasi-stationary ant it is several times longer than the fundamental period of the
structure. The combination rule for the mean of the absolute maximum response is given as
follows:
1
m m m m n
 2

 ∑∑
E[max z (t )] = k =1 l =1
a k a l ρ uk u l u k , max u l , max + 2 ∑∑∑ a k b1 j ρ uk s1 j u k ,max Dl (ω j , ζ j ) + 
(8)
 k =1 l =1 j =1

 ∑∑∑∑ bki b1 j ρ ski s1 j Dk (ω i ,ζ i )Dl (ω j , ζ j ) 
in which, the double sum represents the square of the pseudo-static response, the quadruple sum
represents the square of the dynamic response, and the triple sum represents a coupling term
between the pseudo-static and dynamic response that arises from the covariance between the two
components.

Improved multi-support response spectrum (IMSRS) method

An improved multi-support response spectrum (IMSRS) method[7] is deduced by (X.Zhu

and H.Liu, 2001) for the structural systems subjected to spatially varying seismic excitations. In
the formula, only pseudo-static terms and dynamic terms are included, neglecting the cross
contributions of the pseudo-static response and dynamic response and based on the spectral
parameters defined by E.H.Zavoni, the analytical solution of those spectral parameters are
worked out with the assumptions of the ideal white-noise model and the approximate square
frequency transfer function model. The analytical solutions are adopted to analyze the seismic
response of this long-span bridge, and the influences of the spatial vibration on the Wuhu bridge
are discussed. The improved Multi-supported Response Spectrum (IMSRS) formula can show
as following:
1
m m m m n
 2

∑∑ a k al Θ kl u k ,max u l ,max + ∑∑∑ (α kli Γ0,k1i + β k1i Γ2,k1i ) Dk (ω j , ξ j ) Dl (ω j , ξ j ) −

E[max z (t )] =  k =1 l =1 m m n
k =1 l =1 i =1

 
 ∑∑∑
k =1 l =1 i =1
( λ Λ
kli l , k 1i + χ Λ
k 1i 3, k 1i ) D k (ω i , ξ i ) Dl (ω j , ξ j )


The results of the Wuhu bridge show that the IMSRS method with the analytical solutions of
spectra parameters is more convenient and time saving than that of numerical computing, and it
is reliable and feasible.
(1) The total seismic response of Wuhu Bridge by MSRS method and IMSRS method are in
good agreement with each other. There are approximate analytical solutions in the IMSRS method,
so it is simple, obvious and convenient (TABLE VIII).

227
 TABLE VIII. TOTAL RESPONSE OF WUHU BRIDGE BY TWO METHODS

Vs = Vapp = 200 m Vs = Vapp = 2000 m

s s
T T
Items dl2
dr2
T
M l2
B
M r2
B dl2 dr2
T
M l2
B
M r2
B

(cm) (cm) (kn.m) (kn.m) (cm) (cm) (kn.m) (kn.m)

IMSRS 6.58 5.88 658807 490518 7.10 6.51 676021 538790
MSRS 6.34 5.71 700884 466811 6.59 6.06 717461 510590
Error(%) 3.65 2.89 -6.39 483 7.18 6.91 -6.13 5.23

(2) There are relative important influences on the moment response of Wuhu Bridge when the
values of Vs and Vapp are within 100m/s to 600m/s. If the values of Vs and Vapp are very large, the
influences of the ‘wave passage’ effect and the ‘incoherence’ effect can be neglected (Figure
6) .The first part of this response equation represents the components of pseudo static response, and
the second part represents the components of dynamic response. It is concerned about the effects of
‘local site’ through the various exciting inputs.
(3) The effect of the ‘local site’ can enlarge or reduce the seismic response of Wuhu Bridge
(TABLE IX). It is reliable to response the influence of the variety of local site for long-span bridge
by the multi-supported excited response spectrum method. Fig. 6 shows the curves of the moment
response at the bottom of the towers of Wuhu cable-stayed bridge with the varying Vs or Vapp. Case
1: All the support motions are correlative, i.e. γkl(iω)=1; Case 2: Only wave passage effects are
considered, i.e.α=0; Case 3: Only incoherence effects are considered; Case 4: Both wave passage
and incoherence effects are considered. The loval site effects are included in all the four cases.

Figure 6. The curves of the moment response with the varying Vs or Vapp

228
 TABLE Ⅸ. LOCAL SITE EFFECTS OF DISPLACEMENT RESPONSE OF WUHU BRIDGE

T
d lh
T
d rh
T
dh
M
d lz
T
d rz dz
M
dv
M

Items
(cm) (cm) (cm) (cm) (cm) (cm) (cm)
Multi-excites 8.91 3.26 9.18 7.29 6.68 6.70 6.47
Min-excites 3.10 3.09 4.27 3.30 3.35 3.32 2.14
Max-excites 9.10 9.07 12.53 11.84 12.04 11.90 8.18
Min/Multi 0.35 0.95 0.47 0.45 0.50 0.50 0.33
Max/Multi 1.02 2.78 1.36 1.65 1.80 1.78 1.26

CONCLUSION

Based on results of this study, the following conclusion were reached:

1. It is necessary that the effects of soil-pile-pier (pylon) interaction should be considered in
seismic response analysis for long-span complex bridge with pile foundation. The thickness
of the soft soil deposit is variable. Through the difference of deposit the variety response of
ground motions is reached. The components of higher frequency will be attenuation and
lower frequency will be amplification in the soft deposit layers. The response spectra of
each cap slab of pile foundation are different.
2. The soil-pile-pylon interaction is possible and expedient to make these considerations into
reality in SPM. The elastic constraints at the top of cap slabs are calculated by SPM for
replacing pile foundation in WBM. It is convenient to calculate the response spectrum at
the top of cap slabs by SPM. The envelope curves of those spectra are used as input spectra
in WBM to evaluate the seismic response.
3. Comparing the results from the accumulated contribution ratio side and the seismic
response side, the modes cut off at first 100 modes is better in mode superposition method
for seismic response of Wuhu Bridge, to require the accumulated contribution ratio reached
0.9 is appreciate.
4. There are important influences on the long-span bridge subjected to multi-support
excitations. An improved multi-support response spectrum (IMSRS) method is proposed
for the structure systems subjected to spatially varying seismic excitations. The results are
compared with that of the MSRS method. It is shown that the results produced by the two
methods are well in agreement with each other, but the proposed method is reliable,
feasible and timesaving.

229
ACKNOWLEDGMENT

This study was supported by the Railing Ministry of P. R. China, GRANT

No.96G35©.This support is greatfully acknowledged. The authors wish to thank the Ph.D
students Jianping Sun, Kehai Wang, Hongbing Liu and the colleges of the Major Bridge
Engineering Bureau for their support in complete process of this project.

REFERENCES

Wiegel, r. l.(et al), 1970. Earthquake Engineering (chapter 14). Prentice-Hall.

R. A. Imbsen, W. D. Liu, “Seismic performance evaluation of long-span bridges”, Proceedings of seminar on new
developments in earthquake ground motion estimation and implication for engineering design practice, ATC
35-1, March 1994.
Y. Nabamura, A. Derkiureghian, D. Liu, May 1993, “Multiple-support response spectrum analysis of the Golden
Gate Bridge”, Report No. UCB/EERC-93105.
A. K. Chopra, 1996. “Modal analysis of linear dynamic systems：Physical interpretation”, Journal of Structural
Engineering, ASCE, 122(5).
Transportation Ministry of P. R. China, 1990, Aseismic Design Code of Highway Engineering,(in Chinese).
Railway Ministry of P. R. China, 1988, Aseismic Design Code of Railway Engineering,(in Chinese).
X.Zhu, & H. Liu: “Seismic Analysis of Cable-stayed Bridge for Multi-supported Excitations”, IABSE Conference
(84), Seoul, Korean 2001.
H.Liu, X.Zhu, 2000. “Response srectrum analysis of long-span bridge for multi-supported seismic excitations”,
Proceeding of International Conference on Engineering and Technological Science, 2000, Sience Press,
Beijing.

230
 A Study on Shear Strengthening of Reinforced Concrete
Beams Using ±45°Glass Fiber Reinforced Polymer

Jie Li1,Yuan-de Xue2 and Wen-xiao Li3

ABSTRACT

This paper focuses on the effects of fiber orientation and thickness on cracking and
ultimate load, stiffness and deflection of the reinforced concrete beams which are strengthened
using GFRP in ±45°direction with respect to the beam axis. Two series of the specimens
consist of three simply supported beams and three simply supported beams with overhang. The
research has presented that effective strain of the FRP is concerned with the FRP thickness and
the failure mode varies with the quantities of the FRP. The validity of an anchorage method,
using steel plate bonded with epoxy and fixed with bolt, has been verified. A reduced equation of
shear strengthening has been put forward. From the analysis of the experimental results the
following conclusions can be obtained: The ultimate load and ductility are greatly enhanced if
the reinforced concrete beams are strengthened using GFRP in ±45°direction, the stiffness
will increase and accordingly the deflection will decrease. It is obvious that the GFRP can
effectively confine the diagonal cracking width. Due to the improvement of shear resistance of
the reinforced concrete beam, the failure mode will be changed and the shear failure will
translate into flexural mode. The more the GFRP quantities are, the less is the effective strain of
GFRP. It is worth noting that good anchorage method must be considered in order to prevent
FRP fabric from the peel failure, and the reduction of the ultimate capacity may be observed if
the thickness of FRP increases.

__________________
1
College of Civil Engineering, Tongji University, 200092,Shanghai, PRC
2
State Key Laboratory of Concrete Materials Research, Educational Department Key Laboratory of Solid
Mechanics, Tongji University, Shanghai 200092,PRC
3
Department of Engineering Mechanics and Technology, Tongji University, Shanghai 200092,PRC

231
INTRODUCTION

In practical projects it is often addressed that reinforced concrete members are

strengthened for enhancement of shearing resistance. Existing structures could exhibit
insufficient shear-resist capacity due to the change of usage, the mistakes of the design or
construction and the earthquake and so on. Besides the flexural failure mode may be changed
into diagonal failure if the flexural strength increases because of bonding the FRP to the beam
tensile region. Especially in seismic zone the strong shear resistance is expected to prevent from
brittle failure. Therefore the research of shear retrofit is significant. Many researchers have
extensively investigated the mechanism and effects of reinforced concrete beam shear
strengthening with FRP composite. [1][2]。A serials of parameters, for example, load condition,
support pattern, the ratio of span to depth, the ratio of shear span to depth, concrete strength,
flexural reinforcement ratio, shear reinforcement ratio, nominal FRP reinforcement ratio, bond
pattern, FRP thickness and orientation, modulus of elasticity, the anchor length, shear strength
and thickness of adhesion agent, ultimate elongation and so forth, have effects on efficiency of
shear strengthening. Amir M. et. al [3] have proposed a method of shear calculation for RC
beam strengthened with FRP in different orientations when the beam have no cracking and have
only vertical cracking. The results have a good agreement with those of finite element method
and agree with the experimental results. Michael J. Chajes et. al[4] carried out shearing
experiment of RC beam strengthened with graphite FRP in ±45°orientation. The shear
capacity increased to 151.2%。However it is generally considered that the effects on cracking
and ultimate capacity of RC beam strengthened with FRP in ±45° direction cannot be known
clearly now. This paper presents a study on the above-mentioned issues and tends to provide
experimental database to obtain a reduced equation of shear strength.

THE EXPERIMENTAL PROGRAM

The design of specimens

Two series of specimens are used in the experiment. Three simply supported beams
and three simply supported beams with overhang are designated SBF SBF1 SBF2 and EBF
EBF1 EBF2 respectively. The size and reinforcement are shown in Figure 1 and Figure 2. It
should be noted that the T section simulates the action of floor and the cantilever region is
considered the effects of positive and negative moment on the shear strengthening. FRP
Composite can provided tension in ±45°direction because there may be three diagonal
cracking on the beam with overhang. A steel plate of 6 mm thickness is placed at the upper side
of the beam, which is bonded with adhesion agent and anchored by the bolts simultaneously. The
thickness of the glass fiber fabric is 0.24 mm. The orientation of the fiber is braided in ±45°

232
direction. The adhesion agent TYFO produced in America is adopted in the test. All specimens are
listed in table I and the concrete mix is shown in table II.

The sketch of test

The load is applied according to four-point flexural test. The location of the P1 and P2,
shown in Figure 2, is adjusted so as to produce the same shear force at the two side of the
support B. At the same time the shear capacity is designed equally at left and right sections of
the support B. One or two hydraulic jacks reacting on a structural frame applied the loads. A
steel I-beam distributed the load from the jack to two point load on the specimen.

2 5 6
1
3 4

number of left fiber

strain gagues
bolt M 12 tie 2 16
GFRP
3
2 2 6@150
1 5 6
1 4 3 4

strain 3 25
gague LDVT
5
6

7
number of right rebar
strain gagues

Figure 1. Specimens SBF SBF1 SBF2

strain
gague

7 4
6
5 6 5 4 3 1
8 2

1 2
5 6
3 4 1 GFRP
LVDT
2
number of left fiber 3
strain gagues
number of right rebar
strain gagues

Figure 2. Specimens EBF EBF1 EBF2

233
 The increment of load was 10 kN before cracking and 20 kN after cracking. If the width
of cracking or peel of the FRP is obvious, the load was applied monotonically until failure. Test
data were collected by a computer acquisition system.

TABLE I LIST OF THE SPECIMENS

The shear FRP

Retrofit Transverse
specimens span to depth reinforcement fcu (MPa)
condition steel ratio
ratio ratio
SBF Control - 40.98
SBF1 One layer 2.32 0.24% 40.98
SBF2 Two layers 0.48% 40.98
EBF Control 1.82 - 0.2% 46.85
EBF1 One layer cantilever 0.24% 46.85
1.73 46.85
EBF2 Two layers mid-span 0.48%

TABLE II CONCRETE MIX

Serial Water Cement Sand Aggregate Fly-ash Water reducer
No. 1 0.45 1 1.55 3.11 / /
No. 2 0.500 1 1.526 2.380 0.167 0.012

Figure 3 Diagonal white lines Figure 4 Flexural failure of SBF1

THE OBSERVED BEHAVIOR

Specimen SBF: When the load was applied to 140 kN the first flexural crack appeared at
the soffit near the mid-span. The diagonal crack was observed firstly at the load of 150 KN,
substantially several flexural cracks occurred. The diagonal cracks gradually increased in

234
intensity and depth. Critical diagonal crack formed and extended to the point of the load. Finally
the crush of the concrete occurred in the shear-compression zone at the ultimate load 313 kN.
The width of the major diagonal crack is approximately 2~3 mm., and the deflection is not
apparent.
Specimen SBF1: Firstly flexural crack occurred at the mid-span when the load reached
160 kN. Several white lines under the FRP fabric, shown in figure 3, were found near the support
at the load of 200 kN, which indicated that concrete had cracked. As the load increased the white
lines developed and formed a white strip to location of the bolts. At the load of 429 kN the sound
emitted from FRP could be heard and it may be considered that fiber would be peeled from the
surface of concrete. Increasing the load to 573 kN the concrete in compression zone crushed,
shown in figure 4, and the deflection at the mid-span was greatly obvious. Shear failure changed
into bending failure. After the test, the FRP fabric was removed and fine shallow cracks were
found on the surface of concrete, as shown in figure 5. Just only small parts of the white lines
represented the cracks. There were no major diagonal cracks. And the width of crack is about
0.25 mm.
Specimen SBF2: Flexural and diagonal crack occurring are similar to the specimen
SBF2.The deflection in the mid-span is apparent and concrete in compression crushed somewhat
at the load 520 kN. Afterwards, the steel plate anchored in shear-compression zone seemed to
peel from the concrete. When the load increased to 557 kN cracking was heard and concrete
pieces were pull out seriously. The depth of failure was reached under the tie position. The bolts
were also dragged out, shown in Figure 6.
Specimen EBF: Two diagonal cracks were shown in the middle of beam depth at the
simple span when the load is 200 kN. At the load of 260 kN diagonal crack occurred at the
cantilever region. The No. 6 rebar yielded when the load is 310 kN, and the diagonal crack width
reached 5 mm. It is indicated that test beam failed in shear mode.
Specimen EBF1: As the load increased to 160 kN the diagonal crack appeared in the
simple span where the concrete was not strengthened. White lines could be seen and cracking
from the FRP be heard at the load 240 kN and 380 kN respectively. At the ultimate load of 580
kN the fabric adjacent to beam flange was peeled off. But it is regretful that data file which was
acquired by computer failed when it is transformed into excel file. Fortunately the ultimate load
and maximum fiber strain of No. 4 and No. 6 were recorded.
Specimen EBF2: Much phenomenon was similar to specimen EBF2. At the ultimate
load of 529 kN FRP fabric was peeled from concrete near the flange.

235
 Figure5 Fine diagonal cracks

Figure 5 fine diagonal cracks Figure 6 Concrete pieces of SBF2

EXPERIMENTAL RESULTS

Cracking and ultimate load

Accurate cracking load cannot be obtained due to cover of FRP fabric. The cracking
load has not being enhanced significantly in terms of the deflection- load curve shown in Figure
7 and Figure 8.
However, the ultimate load increased greatly because the diagonal cracks were confined
by FRP in ±45°direction. It is worth noting that the specimen SBF1 test indicated the failure
mode was changed from shear to flexural pattern. The specimen SBF2 initially showed the
tendency of flexural failure but finally failed in peel mode. It is thought that the enhancement of
FRP thickness increases the stiffness of FRP, reduces the length of effective bond and adds the
peel stress at the end of FRP fabric. If the peel failure is avoided the higher ultimate load would
be expected with increasing the quantities of the FRP reinforcement.

Behavior of ductility

Seen from Figure 7 and Figure 8, the ductility of the beam, which was strengthened with
FRP in shear, improves significantly. The ratio of SBF2 maximum deflection to that of SBF is
2.5. SBF1 ultimate deflection cannot be obtained due to loose installation of LVDTs, but it is
estimated that the ultimate deflection is greater than that of SBF2 on the basis of test observation
and the mode of flexural failure.

Analysis of the rebar stress

From figure 9, 10 and 11 it is seen that the rebar stress between the different specimens is
almost the same before cracking. After cracking, the rebar strain of beam strengthened is smaller

236
than that of the beam not strengthened at the same load, especially the beam that was
strengthened with two layers FRP fabrics. Hence GFRP can also reduce the rebar stress and
share the external force in spite of low modulus of elasticity. Besides if peel failure can be
avoided and anchor measurements are adopted suitably the shear strength can be improved
greatly.

Fiber strain

Average fiber strains in principle stress direction at the ultimate load are shown as
follow: SBF1 is 5865 µε ，SBF2 is 4232 µε and EBF2 is 3553 µε . The strains measured are
discrete because of random distribution and non-smoothness of fiber surface. The relationship
between the effective strain and the thickness of the FRP will be further researched by
experiments.

Measurement of anchor

The method using steel plate and bolt has the function for preventing FRP fabric from
debonding. Experiment shows that anchorage function is more effective in rectangular section
than T section. It is mentioned that although the steel plate provides another surface of bond to
FRP fabric, there are some weak zones that is resulted from concentration of stress in concrete.

THEORETICAL ANALYSES

The summation method is approximately adopted in theoretical analysis for calculating

the shear strength of beam. The shear capacity consists of three parts (concrete, rebar and FRP)
and the equation is shown as follows:
EBF-跨中 EBF2-跨中 EBF-支座 EBF2-支座
SBF SBF1 SBF2

60
60

50
50

40
40
load(10kN)
load(10kN)

30
30

20 20

10 10

0 0
0 2 4 6 8 10 12 14 -2 0 2 4 6 8 10 12

d fl ti ( )
deflection(mm)

Figure 7 Deflection-load curve Figure 8 Deflection-load curve

237
 SBF SBF1 SBF2

60

50

40

load(10kN)
30

20

10

0
- 500 0 500 1000 1500 2000 2500

No . 1 s t r a i n o f bar ( ? ? ?

Figure 9 Strain-load curve

60

50

40
load(10kN0

SBF
30 SBF1
SBF2

20

10

0
0 200 400 600 800 1000 1200 1400 1600

Figure 10 strain-load curve

SBF SBF1 SBF2

60

50

40
load(10kN)

30

20

10

0
-500 0 500 1000 1500 2000 2500

No.7 strain of bar

Figure 11 strain-load curve

238
 Vu = Vc + Vs + V f (1)

where Vc , Vs and V f are shear strength of concrete ,rebar and GFRP respectively. Vc , Vs are
obtained according to current RC code, V f can be calculated by the following equation:
A f E f ε fv (sin α + cos α )h f
Vf = (2)
sf

where A f , E f , ε fv , h f , S f are area, the modulus of elasticity of elasticity, effective strain, depth
and span of FRP respectively. The effective strains used in calculation obtain from test. The
results are listed in table 3. It is worth mentioning that a great deal of aggregation may be exist
because the diagonal cracks is constrained by fiber in ±45° direction. Considering that the
final failure mode is not governed by shear mode, of course, the theoretical results may be
approximate. Therefore it is necessary that a suitable equation of shear peel strength should be
established, in the view of point of possibility, especially when large amounts of FRP material
are used.

TABLE III
Specimen Experiment Theory Experiment/ Increment Failure mode
KN KN Theory
SBF 313 281 1.11 - Shear
SBF1 573 456 1.26 1.83 flexural
SBF2 557 528 1.05 1.78 peel
EBF 310 329 0.94 - Shear
EBF1 580 lost lost 1.87 peel
EBF2 529 608 0.87 1.71 peel

CONCLUSIONS

（1）If the RC beam is strengthened with GFRP in ±45°direction for purpose of shear
resistance, the shear strength and stiffness will be increased significantly.
（2）The ultimate deformation and the ductility of the RC beam can be enhanced if using
GFRP in ±45°direction to retrofit.
（3）The diagonal cracks are constrained effectively by fiber in ±45°direction, and the
depth and width of cracks can be reduced.
（4）After cracking glass fiber can share the stress of rebar obviously, the more the
quantities are, the better the effect is.
（5）The effective strain of GFRP will decrease with increasing GFRP quantities.

239
ACKNOWLEDGEMENT

Financial support for this study was provided by 863 project No. 2001AA33610

REFERENCES

QuanWang Li , ShuHong Zhao, LiePing Yie, The Experimental Study of Reinforced Concrete Beam Strengthened
with CFRP. Proceeding of the first China FRP-concrete structure symposium 2000:242-246
YueXuan Zhang, YuanDe Xue Shear Behavior Research of Concrete-Filled FRP Tubes. Proceeding of the first
China FRP-concrete structure symposium 2000:138-145
Amir M.Malek, Hamid Saadatmanesh. Analytical Study of Reinforced Concrete Beams strengthened With
Web-Bonded Fiber Reinforced Plastic Plates or Fabrics. ACI Structural Journal. May-June 1998:343-352
Michael J. Chajes, Ted F. Januszka, Dennis R.Mertz .Shear Strengthening of Reinforced Concrete Beams Using
Externally Applied Composite Fabrics. ACI Structural Journal. May-June 1995.

240
 Research on Flexural Properties of Hybrid GFRP/CFRP
Tube Confined Concrete Beams

Hua Yuan1,Yuan-de Xue2 and Wen-xiao Li3

ABSTRACT

This paper reports a study on flexural properties of hybrid beams which consist of
concrete-filled ±45ºglass filament-wound FRP square tube with carbon FRP laminates bonded
to its tensile side. In order to optimize the utility of characteristic mechanical properties of each
element materials, the material and structure designs of that beams are base on such a rule that
carbon composite is mainly arranged for bearing tension, confined concrete for compression and
±45º plies glass-fiber composite for shear. The constitutive relations of FRP and confined
concrete under compression are obtained experimentally and Vecchio and Collins’ tensile model
of confined concrete is used. The flexural behaviors of beams, which have different hybrid
structures, were studied through four-point bending tests. The result shows the synergetic effect
of the combination of composite and concrete.

INTRODUCTION

Recently, the usage of fiber-reinforced composites combined with traditional materials in

primary structural applications has emerged as one of the most promising developments in the
construction industry due to corrosion problems. Diverse concrete-filled FRP tubes (CFFT) are
studied for beams. A number of studies about his paper presents a study on a hybrid beam
consisted of concrete filled ±45ºglass filament-wound FRP tube with a carbon FRP laminate
bonded to its tension side.
The most common characteristics of combined beams are follows: 1). The designable
system allows optimization based on material properties of each component and the state of load
carrying. 2) The CFFT replaces conventional reinforcing steel. The enhanced behavior of CFFT
__________________
1
State Key Laboratory of Concrete Materials Research, Tongji University, 200092, Shanghai, PRC
2
State Key Laboratory of Concrete Materials Research, Educational Department Key Laboratory of Solid
Mechanics, Tongji University, shanghai 200092, PRC
3
Department of Engineering Mechanics and Technology, Tongji University, Shanghai 200092, PRC

241
beams could allow the use of smaller sections than those which would be required for
conventionally reinforced concrete beams . It also increases durability while providing enhanced
confinement to the concrete core. 3) FRP shape acts as permanent form for concrete. It greatly
enhances ease of handling and erection speeds. Furthermore, FRP jacket reinforced concrete would
be more durable than conventional reinforced concrete beam and therefore would require less
maintenance and have longer service life. Overall cost is lower.
In 2000, Mirmiran and Shahawy[1] tested round concrete filled 356-mm-diameter and
6.6-mm-wall thickness GFRP winding tube by a series of four point bending tests. Karbhari et al
pursued the experimental characterization of the concrete filled carbon shell for girder on short
and medium span bridge. The carbon shell geometry used in all the test units of the experimental
program is that of a cylindrical tube with an inside diameter of 343 mm and a wall thickness of
10 mm. The carbon/epoxy laminated shell has a lay-up architecture with approximately 80%
longitudinal (±10º helical) and 20% transverse (90º) fiber reinforcement with an average fiber
volume ratio of 55%. In this paper, a novel concrete filled square hybrid GFRP/CFRP tube is
studied. The constitutive relations are set up. Applicability of the constitutive relations to hybrid
beam is examined.

Experimental analyses on concrete filled hybrid GFRP/CFRP tube beams

Hybrid GFRP/CFRP tube confined concrete beam, shown in Fig. 1, consists of

concrete filled ±45ºglass fiber winding FRP tube with a thin carbon fiber laminate bonded to
its tension side. The thick of GFRP is 2.7 mm and the nominal thick of CFRP 0.11 mm.
t b t

Figure 1 The cross-sectional shape of hybrid GFRP/CFRP tube confined concrete beam

where h=b=50mm; Concrete Ec=28GPa; Carbon fiber E=210 GPa;

90mm 120mm 90mm

Figure 2. The schematic of beam test setup

242
 The specimens were subjected a four-point flexure tests, the schematic of beam test setup
shown in Figure 2. The result of test are shown in Figure 3 and Figure 4.

50
with carbon fiber reinforced
40

load(KN)
30 no carbon fiber reinforced
20

10
0
0 5 10 15 20
deflection(mm)
Figure 3. load – deflection diagram

0
height of neutral

0 0.005 0.01 0.015 0.02

-5
Without carbon fiber
axis(mm)

-10
-15
With carbon fiber
-20
-25
-30
strain of tension side

Figure 4 Strain of tension side –height of neutral axis diagram

As shown in Figure 3 and Figure 4, it was noted that when the thick of the GFRP tubes is same,
the limit load of the member with carbon fiber reinforced is as 1.2 times as that with no carbon fiber
reinforced. The failure bending moment increased from 1.485KN• m to 1.827 KN• m. At the
same time, the neutral axis transfers to the tension side due to contribution of carbon fiber. Figure 5
and Figure 6 show the failure of the members.

Figure 5. Picture of concrete filled GFRP tube after failure

243
 Figure 6 Picture of concrete filled hybrid GFRP/CFRP tube after failure

THE CALCULATION MODELS

Tensile model of GFRP

Three slices, dimensions shown in TABLEI, intercepted from a hollow GFRP tube are
tested. The specimens after failure are in figure 7

TABLE I DIMENSION OF SPECIMENS

Number 1 2 3
Width (mm) 1.144 1.1964 1.1672
Thickness (mm) 0.1136 0.1184 0.098

Figure 7 specimens after failure

244
 In order to simplify calculation, the results of test are fitted to get the curve (Figure 8) and
Eqn.1.

80

stress （Mpa)
60

40

20

0
0 0.005 0.01 0.015
strain

Figure 8 curve of FRP tensile stress-strain

10000 ε FRP (MPa) ε FRP ≤ 0.0072 (1)

σ =
FRP
72 (MPa) ε FRP > 0 . 0072

Compressive model of GFRP

Two tubes about 80mm high (dimension shown in TABLE II) intercepted from hollow
GFRP are tested. The specimens after failure are shown in Figure 9.

Figure 9 specimens after failure

TABLE II DIMENSION OF SPECIMENS

Number 1 2
Width (mm) 5.325 5.340
Thickness (mm) 0.134 0.136
Height (mm) 80.1 80.5

245
The results of test are fitted to obtain Figure 10 and Eqn. 2.

150

stress(MP)
100

50

0
0 0.005 0.01 0.015 0.02
strain

Figure 10 Curve of FRP compressive stress-strain

10000 ε FRP
(MPa) ε FRP ≤ 0.0097 (2)
σ =
FRP
97 (MPa) ε FRP > 0 . 0097

Tensile model of CFRP

CFRP was bonded to the tension side of a beam in order to bear tensile stress, so carbon
fiber was laid parallel to the axis of the beam. Because the thickness of CFRP is much smaller
than that of GFRP and dimension of the beam, the strain of the CFRP is thought equal to tension
side of GFRP. In calculation stress equals to modulus of elasticity multiplied by strain, namely
σ=2.1×105εFRP MPa, while the elastic modulus of carbon.is2.1×105 MPa

Compressive model of confined concrete

According to design guidelines in our country, the constitutive relation of concrete in

compressive section is shown as Figure.

fcm

0.002 0.0033 ε

Figure 11 σ − ε diagram of compressive concrete

246
If 0 < ε < 0.002 ，
σ = f cm 1 − (1 − ε ε ) 2  （3）
 0 

If 0.002 < ε < 0.0033 ，

σ = f cm （4）

where f cm :flexural strength of concrete； ε0：yield strain

According to Amir Mirmiran[12] et. al’s research , strength of concrete will improve when
it is confined. Magnifying coefficient is given by test.
Two blocks of same height were cut from a beam. One block was shucked off FRP. (The
dimensions are shown in TABLE III). Two blocks were tested under compression. The results
are shown in Figure 12.and the specimens after failure are shown in figure 3

TABLE III DIMENSION OF SPECIMENS

Plain concrete block FRP confined concrete block

Height (mm) Side (mm) Height (mm) Side (mm) Thick of FRP tube (mm)
76.0 52.5 76.0 57.82 2.66

60
50
stress(Mpa)

40
30 FRP confined concrete
20 plain concrete
10
0
0 5000 10000 15000 20000 25000
strain(*e-6)
Fig. 12 Curve of stress-strain of plain concrete and FRP confined concrete in compression

247
 According to the test, compressive strength of FRP confined concrete is much higher
than that of plain concrete. The ratio of two failure stresses is 2.5.

Figure 13 Specimens after failure

According to Amir Mirmiran[13]and Xue Yuande et al[6] ˊ s researches ， and the

manuscript of Technology Code of Fiber Reinforced Composite Strengthening Concrete
Structure, a calculation equation was put forward：

 E f ε c f nt f  
N u = α 1  f c + 8r1  Ac + As ' f y '

 h   （5）
Nu
where： ―carrying load of strengthened member；
α 1 ―Strength coefficient ，according to regulations (GBJ10),taken 0.9；
fc
―Compressive strength of concrete core；
r1 ―Strength of concrete effecting coefficient 60− c ；
r1 = 0.8 +
150
c ―Strength grade of concrete；
Ef
―Elastic modulus of fabric, taken elastic modulus of FRP tube；
ε cf
―Permitting strain of fabric under compression. According to the test, take longitudinal
strain 0.004488 as permitting strain. Permitting strain of fabric under compression is 0.004488×
0.48=0.0021 due to Poisson’s ratio of ±45ºwinding FRP tube is 0.48
n ―Layers of fabric；
tf
―Thickness of single layer of fabric；
h ―Side of member section；
Ac
―Area of concrete section；
Ac ' Area of steel section. It is 0 in this paper
― ；
fy'
―Compressive strength of steel. It is 0 in this paper；

248
 When tf=1.65mm, reinforcement ratio is 1.36 on calculation. The value is smaller than
that obtained from test. 1.36 is adopted in this paper on account of safety. Compressive model of
confined concrete is shown in Fig. 14.

αfcm

0.002α ε
Figure 14 curve of concrete σ − ε under compression

σ = α f cm 1 − (1 − ε ε ) 2 
When 0 < ε < 0.002α ，  0  （6）

When ε > 0.002α ， σ = αf cm （7）

Where： α : compressive reinforcing coefficient of confined concrete
f cm ：flexural compressive strength of concrete

Tensile model of confined concrete

According to guidelines, concrete in tension section does not work after crack, so
carrying load capacity of concrete in tension section usually is neglected.
But according to Amir Mirmiran、Vecchio and Collins [14], FRP Confined concrete in
tension section works all the time. The result is also proved in our test.
The model Vecchio and Collins[14]obtained is used in this paper. It is shown in Figure14.
σ

0.00015 ε

Figure 15 Tensile σ − ε of confined concrete diagram

When 0 < ε < 0 .00015 时 ， σ = E c × ε （8）

β f cr
σ =
When ε > 0 . 00015 时， 1+ 200 ε （9）

249
Where：Ec elastic modulus of concrete；
fcr : Crack stress of concrete；
β : Reinforcing coefficient of concrete under tension（ β =1 in order for safety）

CALCULATING CARRYING LOAD OF HYBRID GFRP/CFRP TUBE CONFINED

CONCRETE BEAM

Specimen is same as above

Assumption of calculation

1. Plane sections remain plane and normal to the neutral axis after bending.
2. Perfect bond exists between concrete and the FRP tube.
3. The constitutive relations above are suitable to the calculation.

Calculating program frame

Figure 16 shows the Calculating program frame.

Result analysis

Moments in different strains are shown in table IV by test and calculation respectively.
From the table IV, it is found that the results tested are close to but not equal to that calculated.
The main reasons are:(1)there are perhaps some errors using nominal thick of the carbon fiber
instead of real thick of CFRP in calculation .(2) Carbon fiber is controlled in elastic range in
practical use. When elastic model is used, there is bigger error in state of larger strain than
smaller strain.

TABLE IV COMPARISON THE RESULT WITH THAT OF CALCULATION

Item Result of test Result of calculation

ε=0.0075 0.810 0.790
moment/KN.m ε=0.014 1.305 1.049
ε=0.0176 1.485 1.16

250
 Input dimensions of the member and material parameter h, b, t , E c , E frp

10000ε frp Mpa ε frp ≤ 0.0072

Input GFRP tensile constitutive relations σ frp =
72 Mpa ε frp > 0.0072

10000 ε FRP (MPa) ε FRP ≤ 0.0097

GFRP compressive
constitutive relations σ =
FRP
97 (MPa) ε FRP > 0 . 0097

Input CFRP tensile constitutive relations σ=2.1e11εMPa

σ = α f cm 1 − (1 − ε ε ) 2 
Input concrete  0  MPa 0 < ε < 0 . 002 α
compressive σ = α f cm ε > 0 . 002 α
MPa

Input concrete tensile constitutive relations σ = Ec × ε MPa 0 < ε < 0 . 00015

β f cr
σ =
1+ 200 ε MPa ε > 0 . 00015

Input height (x) of compressive section

Output resultant force: N

N=0

No
Yes
Output moment: M and the height of compressive section: X

Figure 16 FRP confined concrete beam calculating program frame

251
CONCLUSION

1. Hybrid beams were designed, manufactured and tested. This innovative design produces highly
optimized behavior with a pronounced synergetic effect. In hybrid beam, carbon composite
works in tension, concrete filling works mainly in compression and ±45ºglass fiber composite
works in shear
2. The numerical model used for this analysis was implemented. This model predicts with
reasonable accuracy the moments. This model is still in progress.

ACKNOWLEDGEMENT

Financial support for this study was provided by 863 projects No.2001AA33610

REFERENCES

Amir Mirmiran, et al. Large Beam-column Tests on Concrete-Filled Composite Tubes [A]. ACI Structural Journal,
2000,97(2):268-276
Karbhari V M, Seible F, et al. Structural characterization of fiber-reinforced composite short-and medium-span
bridge system [J]. Applied Composite Materials, 2000, (7): 151-182
Zhao J study on FRP confined concrete [D]. Shanghai: Department of Engineering Mechanics and Technology,
Tongji University, 1997.
Zhao J and Xue Y D. Primary Research on FRP Tube Confined Concrete [A]. Proceedings of the 12th Glass fiber
Reinforced Plastics/Composite[C] , Qindao: Glassfiber reinforced plastics/composite agent,1997:232-237.
Xue Y D, Zhao, J. Primary Research on FRP Confined Concrete[A]. Proceedings of International Conference of
the Mechanical Test on Material and Structure[C], Hongkong, 1998: 274-278
Zhang Y X. Study on property of FRP confined concrete [D]. Shanghai: Department of Engineering Mechanics and
Technology, Tongji University 2000.
Zhang Y X, Xue Y D. The basic property of FRP tube confined concrete [A] ，Glassfiber reinforced
plastics/composite, 1999,6:21-23
Xue Y D, Zhang Y X. Study on shear property of FRP tube confined concrete [A], Proceedings of the 1st Chinese
Academic Conference of FRP/Concrete Structure [C], Beijing. 2000,6:138-145
Yuan H, Xue Y D. Study on basic property of FRP tube confined concrete [J]. Concrete and cement
products,2000(6):34-36.
Fan L.C, Zhuo W. D, Xue Y. D., Research on Seismic Property of FRP Jacket RC pier [A], Proceedings of the 1st
Chinese Academic Conference of FRP / Concrete Structure [C], Beijing. 2000,6:113-117.
Xue Y D, Yuan H ,et al. Fiber reinforced polymer confined concrete combined component [J].journal of building
materials,2002,1:1-8.
Amir Mirmiran，Mohsen Shahawy and Michael Samaan . Strength and Ductility of Hybrid FRP-Concrete
Beam-Columns [J] ，Journal of Structural Engineering, 1999; 10

252
Amir Mirmiran, Mohsen Shahawy. Effect of Column Parameters on FRP-Confined Concrete [J]. Journal of
Composites for Construction, 1998, 175-185.
Vecchi F J and Collins M P. The modified compression-field theory for reinforced concrete elements subjected to
shear [J]. ACI Structural Journal, 1998, (3-4): 219-185

253
 PRC-US Workshop Agenda

October 8, AM: Four invited papers (30 minutes for each)

9:00-9:10: Opening speech

9:10-10:10: Two papers
1. Brief introduction of the study trends of SLDRCE, by Li-chu Fan
2. Performance-based seismic design of highway systems, by Ian G. Buckle
10:10-10:30: Coffee break
10:30-11:30: Two papers
3. Mitigate earthquake hazard & risk for highway bridges through planning, design and
retrofitting (extended abstract), by W. Phillip Yen
4. Seismic safety evaluation of large scale interchange system in Shanghai,
introduced by Li-chu Fan, and presented by Li-ying Nie

October 8, PM: Seven papers (20 minutes for each)

2:30-3:50: Four papers

5. A new approach to analysis of soil-pile-structure interactions for long-span bridges,
by Jamshid Ghaboussi
6. Preliminary study of hydrodynamic effects on seismic response of bridges, by Jun-jie wang
7. Observed pile and pipeline performance in the full-scale lateral spread experiment,
by Scott A. Ashford
8. Seismic response of railway bridges considering track restriction, by Gui-ping Yan
3:50-4:10: Coffee break
4:10-5:10: Three papers
9. Seismic design and retrofit guidelines for bridges in New Jersey, a low-to-moderate seismic
hazard area, by H.A. Capers
10. Generation of response spectrum compatible non-stationary ground motions based on phase
difference spectra, by Qing-shan Yang
11. The seismic retrofit of the Golden Gate Bridge, by Charles Seim, P.E.
5:10-6:00: Visit to SLDRCE (wind tunnel, bridge laboratory and shaking table)

October 8, Evening, 20:00: Reception (Radission hotel)

October 9, AM: Seven papers(20 minutes for each)

9:00-10:20: Four papers

12. Response of seismic isolated bridges using M-DOF model and 2D excitation, by G.C. Lee
13. Aseismic conceptual design of bridge tower of long-span cable-stayed bridges, by Ai-jun Ye
14. Analytical investigation of the response of LuPu bridge with added viscous dampers,
by Shi-de Hu
15. Seismic performance and retrofit of a 24-span freeway bridge, by M. Saiidi
10:20-10:40: Coffee break
10:40-11:40: Three papers

255
 16. A study on shear strengthening of reinforced concrete beams using ±45º glass fiber
reinforced polymer, by Yuan-de Xue
17. Seismic design and analysis for urban viaducts with a double deck, by Jian-zhong Li
18. A pseudodynamic test of an urban viaduct with a double-deck, by Tian-bo Peng

October 9, PM: Two invited papers (30 minutes for each) plus summary and recommendations for future
research
2:30-3:30: Two invited papers
19. Seismic design and retrofit strategies of cable-supported bridge: An overview of current
practice, introduced by G.C. Lee, presented by J. Sun
20. Seismic response analysis of Wuhu Yangtze River Bridge, by Xi zhu
3:30-3:50: Coffee break
3:50-5:20: Summary and recommendations for future research (8th floor of new bridge building)

October 9, Evening, 19:00: Banquet (Radisson hotel)

October 10: Technical tour to Lupu bridge and sightseeing in Shanghai

256
 Participants

U.S. Participants

Scott A. Ashford M.Saiidi

Department of Structural Engineering Civil Engineering Department
University of California, San Diego University of Nevada, Reno
La Jolla, CA 92093-0085 Reno, NV 89557
[email protected] [email protected]

Ian G. Buckle Charles Seim, P.E.

Department of Civil Engineering T.Y.Lin International
MS 258 825 Battery St.
University of Nevada, Reno San Francisco, CA., 94111
Reno, NV 89557 [email protected]
[email protected]
John Sun
Harry Allen Capers T.Y.Lin International
Structural Engineering (State Bridge Engineer) 825 Battery St.
New Jersey Department of Transportation San Francisco, CA., 94111
PO Box 615 [email protected]
Trenton, New Jersey 08625-0615
[email protected] Jane Stoyle
MCEER
Jamshid Ghaboussi University of Buffalo
Department of Civil and Environmental Red Jacket Quad
Engineering Buffalo, NY 14261
University of Illinois at Urbana-Champaign [email protected]
Urbana, IL 61801
W. Phillip Yen
George C. Lee Office of Infrastructure, R&D Federal Highway
MCEER Administration
University of Buffalo 6300 Georgetown,Pike
Red Jacket Quad McLean, VA 20121
Buffalo, NY 14261 [email protected] [email protected]

257
PRC Participants

Suwen Chen Tianbo Peng

College of Civil Enginering Department of Bridge Engineering
Tongji University Tongji University
Shanghai,China Siping Road, 1239#,
Current address: 200092, Shanghai, China
MCEER [email protected]
University of Buffalo
Red Jacket Quad Junjie Wang
Buffalo, NY 14261 Department of Bridge Engineering
[email protected] Tongji University
Siping Road, 1239#,
Lichu Fan 200092, Shanghai, China
Department of Bridge Engineering [email protected]
Tongji University
Siping Road, 1239#, Zhiqiang Wang
200092, Shanghai, China Department of Bridge Engineering
[email protected] Tongji University
Siping Road, 1239#,
Shidi Hu 200092, Shanghai, China
Department of Bridge Engineering [email protected]
Tongji University
Siping Road, 1239#, Yan Xu
200092, Shanghai, China Department of Bridge Engineering
[email protected] Tongji University
Siping Road, 1239#,
Jianzhong Li 200092, Shanghai, China
Department of Bridge Engineering [email protected]
Tongji University
Siping Road, 1239#, Guiping Yan
200092, Shanghai, China College of Engineering and Architecture
[email protected] Northern Jiaotong University
Bejing,100044,PRC
Liying Nie [email protected]
Department of Bridge Engineering
Tongji University
Siping Road, 1239#,
200092, Shanghai, China
[email protected],[email protected]

258
Qingshan Yang
College of Engineering and Architecture
Northern Jiaotong University
Bejing,100044,PRC
[email protected]

Aijun Ye
Department of Bridge Engineering
Tongji University
Siping Road, 1239#,
200092, Shanghai, China
[email protected]

Xi Zhu
College of Engineering and Architecture
Northern Jiaotong University
Bejing,100044,PRC
[email protected]

259
 Multidisciplinary Center for Earthquake Engineering Research
List of Technical Reports
The Multidisciplinary Center for Earthquake Engineering Research (MCEER) publishes technical reports on a variety of subjects
related to earthquake engineering written by authors funded through MCEER. These reports are available from both MCEER
Publications and the National Technical Information Service (NTIS). Requests for reports should be directed to MCEER
Publications, Multidisciplinary Center for Earthquake Engineering Research, State University of New York at Buffalo, Red
Jacket Quadrangle, Buffalo, New York 14261. Reports can also be requested through NTIS, 5285 Port Royal Road, Springfield,
Virginia 22161. NTIS accession numbers are shown in parenthesis, if available.

NCEER-87-0001 "First-Year Program in Research, Education and Technology Transfer," 3/5/87, (PB88-134275, A04, MF-
A01).

NCEER-87-0002 "Experimental Evaluation of Instantaneous Optimal Algorithms for Structural Control," by R.C. Lin, T.T.
Soong and A.M. Reinhorn, 4/20/87, (PB88-134341, A04, MF-A01).

NCEER-87-0003 "Experimentation Using the Earthquake Simulation Facilities at University at Buffalo," by A.M. Reinhorn
and R.L. Ketter, to be published.

NCEER-87-0004 "The System Characteristics and Performance of a Shaking Table," by J.S. Hwang, K.C. Chang and G.C.
Lee, 6/1/87, (PB88-134259, A03, MF-A01). This report is available only through NTIS (see address given
above).

NCEER-87-0005 "A Finite Element Formulation for Nonlinear Viscoplastic Material Using a Q Model," by O. Gyebi and G.
Dasgupta, 11/2/87, (PB88-213764, A08, MF-A01).

NCEER-87-0006 "Symbolic Manipulation Program (SMP) - Algebraic Codes for Two and Three Dimensional Finite Element
Formulations," by X. Lee and G. Dasgupta, 11/9/87, (PB88-218522, A05, MF-A01).

NCEER-87-0007 "Instantaneous Optimal Control Laws for Tall Buildings Under Seismic Excitations," by J.N. Yang, A.
Akbarpour and P. Ghaemmaghami, 6/10/87, (PB88-134333, A06, MF-A01). This report is only available
through NTIS (see address given above).

NCEER-87-0008 "IDARC: Inelastic Damage Analysis of Reinforced Concrete Frame - Shear-Wall Structures," by Y.J. Park,
A.M. Reinhorn and S.K. Kunnath, 7/20/87, (PB88-134325, A09, MF-A01). This report is only available
through NTIS (see address given above).

NCEER-87-0009 "Liquefaction Potential for New York State: A Preliminary Report on Sites in Manhattan and Buffalo," by
M. Budhu, V. Vijayakumar, R.F. Giese and L. Baumgras, 8/31/87, (PB88-163704, A03, MF-A01). This
report is available only through NTIS (see address given above).

NCEER-87-0010 "Vertical and Torsional Vibration of Foundations in Inhomogeneous Media," by A.S. Veletsos and K.W.
Dotson, 6/1/87, (PB88-134291, A03, MF-A01). This report is only available through NTIS (see address
given above).

NCEER-87-0011 "Seismic Probabilistic Risk Assessment and Seismic Margins Studies for Nuclear Power Plants," by Howard
H.M. Hwang, 6/15/87, (PB88-134267, A03, MF-A01). This report is only available through NTIS (see
address given above).

NCEER-87-0012 "Parametric Studies of Frequency Response of Secondary Systems Under Ground-Acceleration Excitations,"
by Y. Yong and Y.K. Lin, 6/10/87, (PB88-134309, A03, MF-A01). This report is only available through
NTIS (see address given above).

NCEER-87-0013 "Frequency Response of Secondary Systems Under Seismic Excitation," by J.A. HoLung, J. Cai and Y.K.
Lin, 7/31/87, (PB88-134317, A05, MF-A01). This report is only available through NTIS (see address given
above).

NCEER-87-0014 "Modelling Earthquake Ground Motions in Seismically Active Regions Using Parametric Time Series
Methods," by G.W. Ellis and A.S. Cakmak, 8/25/87, (PB88-134283, A08, MF-A01). This report is only
available through NTIS (see address given above).

Formerly the National Center for Earthquake Engineering Research 261

NCEER-87-0015 "Detection and Assessment of Seismic Structural Damage," by E. DiPasquale and A.S. Cakmak, 8/25/87,
(PB88-163712, A05, MF-A01). This report is only available through NTIS (see address given above).

NCEER-87-0016 "Pipeline Experiment at Parkfield, California," by J. Isenberg and E. Richardson, 9/15/87, (PB88-163720,
A03, MF-A01). This report is available only through NTIS (see address given above).

NCEER-87-0017 "Digital Simulation of Seismic Ground Motion," by M. Shinozuka, G. Deodatis and T. Harada, 8/31/87,
(PB88-155197, A04, MF-A01). This report is available only through NTIS (see address given above).

NCEER-87-0018 "Practical Considerations for Structural Control: System Uncertainty, System Time Delay and Truncation of
Small Control Forces," J.N. Yang and A. Akbarpour, 8/10/87, (PB88-163738, A08, MF-A01). This report is
only available through NTIS (see address given above).

NCEER-87-0019 "Modal Analysis of Nonclassically Damped Structural Systems Using Canonical Transformation," by J.N.
Yang, S. Sarkani and F.X. Long, 9/27/87, (PB88-187851, A04, MF-A01).

NCEER-87-0020 "A Nonstationary Solution in Random Vibration Theory," by J.R. Red-Horse and P.D. Spanos, 11/3/87,
(PB88-163746, A03, MF-A01).

NCEER-87-0021 "Horizontal Impedances for Radially Inhomogeneous Viscoelastic Soil Layers," by A.S. Veletsos and K.W.
Dotson, 10/15/87, (PB88-150859, A04, MF-A01).

NCEER-87-0022 "Seismic Damage Assessment of Reinforced Concrete Members," by Y.S. Chung, C. Meyer and M.
Shinozuka, 10/9/87, (PB88-150867, A05, MF-A01). This report is available only through NTIS (see address
given above).

NCEER-87-0023 "Active Structural Control in Civil Engineering," by T.T. Soong, 11/11/87, (PB88-187778, A03, MF-A01).

NCEER-87-0024 "Vertical and Torsional Impedances for Radially Inhomogeneous Viscoelastic Soil Layers," by K.W. Dotson
and A.S. Veletsos, 12/87, (PB88-187786, A03, MF-A01).

NCEER-87-0025 "Proceedings from the Symposium on Seismic Hazards, Ground Motions, Soil-Liquefaction and Engineering
Practice in Eastern North America," October 20-22, 1987, edited by K.H. Jacob, 12/87, (PB88-188115, A23,
MF-A01). This report is available only through NTIS (see address given above).

NCEER-87-0026 "Report on the Whittier-Narrows, California, Earthquake of October 1, 1987," by J. Pantelic and A.
Reinhorn, 11/87, (PB88-187752, A03, MF-A01). This report is available only through NTIS (see address
given above).

NCEER-87-0027 "Design of a Modular Program for Transient Nonlinear Analysis of Large 3-D Building Structures," by S.
Srivastav and J.F. Abel, 12/30/87, (PB88-187950, A05, MF-A01). This report is only available through NTIS
(see address given above).

NCEER-87-0028 "Second-Year Program in Research, Education and Technology Transfer," 3/8/88, (PB88-219480, A04, MF-
A01).

NCEER-88-0001 "Workshop on Seismic Computer Analysis and Design of Buildings With Interactive Graphics," by W.
McGuire, J.F. Abel and C.H. Conley, 1/18/88, (PB88-187760, A03, MF-A01). This report is only available
through NTIS (see address given above).

NCEER-88-0002 "Optimal Control of Nonlinear Flexible Structures," by J.N. Yang, F.X. Long and D. Wong, 1/22/88, (PB88-
213772, A06, MF-A01).

NCEER-88-0003 "Substructuring Techniques in the Time Domain for Primary-Secondary Structural Systems," by G.D.
Manolis and G. Juhn, 2/10/88, (PB88-213780, A04, MF-A01).

NCEER-88-0004 "Iterative Seismic Analysis of Primary-Secondary Systems," by A. Singhal, L.D. Lutes and P.D. Spanos,
2/23/88, (PB88-213798, A04, MF-A01).

Formerly the National Center for Earthquake Engineering Research 262

NCEER-88-0005 "Stochastic Finite Element Expansion for Random Media," by P.D. Spanos and R. Ghanem, 3/14/88, (PB88-
213806, A03, MF-A01).

NCEER-88-0006 "Combining Structural Optimization and Structural Control," by F.Y. Cheng and C.P. Pantelides, 1/10/88,
(PB88-213814, A05, MF-A01).

NCEER-88-0007 "Seismic Performance Assessment of Code-Designed Structures," by H.H-M. Hwang, J-W. Jaw and H-J.
Shau, 3/20/88, (PB88-219423, A04, MF-A01). This report is only available through NTIS (see address given
above).

NCEER-88-0008 "Reliability Analysis of Code-Designed Structures Under Natural Hazards," by H.H-M. Hwang, H. Ushiba
and M. Shinozuka, 2/29/88, (PB88-229471, A07, MF-A01). This report is only available through NTIS (see
address given above).

NCEER-88-0009 "Seismic Fragility Analysis of Shear Wall Structures," by J-W Jaw and H.H-M. Hwang, 4/30/88, (PB89-
102867, A04, MF-A01).

NCEER-88-0010 "Base Isolation of a Multi-Story Building Under a Harmonic Ground Motion - A Comparison of
Performances of Various Systems," by F-G Fan, G. Ahmadi and I.G. Tadjbakhsh, 5/18/88, (PB89-122238,
A06, MF-A01). This report is only available through NTIS (see address given above).

NCEER-88-0011 "Seismic Floor Response Spectra for a Combined System by Green's Functions," by F.M. Lavelle, L.A.
Bergman and P.D. Spanos, 5/1/88, (PB89-102875, A03, MF-A01).

NCEER-88-0012 "A New Solution Technique for Randomly Excited Hysteretic Structures," by G.Q. Cai and Y.K. Lin,
5/16/88, (PB89-102883, A03, MF-A01).

NCEER-88-0013 "A Study of Radiation Damping and Soil-Structure Interaction Effects in the Centrifuge," by K. Weissman,
supervised by J.H. Prevost, 5/24/88, (PB89-144703, A06, MF-A01).

NCEER-88-0014 "Parameter Identification and Implementation of a Kinematic Plasticity Model for Frictional Soils," by J.H.
Prevost and D.V. Griffiths, to be published.

NCEER-88-0015 "Two- and Three- Dimensional Dynamic Finite Element Analyses of the Long Valley Dam," by D.V.
Griffiths and J.H. Prevost, 6/17/88, (PB89-144711, A04, MF-A01).

NCEER-88-0016 "Damage Assessment of Reinforced Concrete Structures in Eastern United States," by A.M. Reinhorn, M.J.
Seidel, S.K. Kunnath and Y.J. Park, 6/15/88, (PB89-122220, A04, MF-A01). This report is only available
through NTIS (see address given above).

NCEER-88-0017 "Dynamic Compliance of Vertically Loaded Strip Foundations in Multilayered Viscoelastic Soils," by S.
Ahmad and A.S.M. Israil, 6/17/88, (PB89-102891, A04, MF-A01).

NCEER-88-0018 "An Experimental Study of Seismic Structural Response With Added Viscoelastic Dampers," by R.C. Lin, Z.
Liang, T.T. Soong and R.H. Zhang, 6/30/88, (PB89-122212, A05, MF-A01). This report is available only
through NTIS (see address given above).

NCEER-88-0019 "Experimental Investigation of Primary - Secondary System Interaction," by G.D. Manolis, G. Juhn and
A.M. Reinhorn, 5/27/88, (PB89-122204, A04, MF-A01).

NCEER-88-0020 "A Response Spectrum Approach For Analysis of Nonclassically Damped Structures," by J.N. Yang, S.
Sarkani and F.X. Long, 4/22/88, (PB89-102909, A04, MF-A01).

NCEER-88-0021 "Seismic Interaction of Structures and Soils: Stochastic Approach," by A.S. Veletsos and A.M. Prasad,
7/21/88, (PB89-122196, A04, MF-A01). This report is only available through NTIS (see address given
above).

NCEER-88-0022 "Identification of the Serviceability Limit State and Detection of Seismic Structural Damage," by E.
DiPasquale and A.S. Cakmak, 6/15/88, (PB89-122188, A05, MF-A01). This report is available only through
NTIS (see address given above).

Formerly the National Center for Earthquake Engineering Research 263

NCEER-88-0023 "Multi-Hazard Risk Analysis: Case of a Simple Offshore Structure," by B.K. Bhartia and E.H. Vanmarcke,
7/21/88, (PB89-145213, A05, MF-A01).

NCEER-88-0024 "Automated Seismic Design of Reinforced Concrete Buildings," by Y.S. Chung, C. Meyer and M.
Shinozuka, 7/5/88, (PB89-122170, A06, MF-A01). This report is available only through NTIS (see address
given above).

NCEER-88-0025 "Experimental Study of Active Control of MDOF Structures Under Seismic Excitations," by L.L. Chung,
R.C. Lin, T.T. Soong and A.M. Reinhorn, 7/10/88, (PB89-122600, A04, MF-A01).

NCEER-88-0026 "Earthquake Simulation Tests of a Low-Rise Metal Structure," by J.S. Hwang, K.C. Chang, G.C. Lee and
R.L. Ketter, 8/1/88, (PB89-102917, A04, MF-A01).

NCEER-88-0027 "Systems Study of Urban Response and Reconstruction Due to Catastrophic Earthquakes," by F. Kozin and
H.K. Zhou, 9/22/88, (PB90-162348, A04, MF-A01).

NCEER-88-0028 "Seismic Fragility Analysis of Plane Frame Structures," by H.H-M. Hwang and Y.K. Low, 7/31/88, (PB89-
131445, A06, MF-A01).

NCEER-88-0029 "Response Analysis of Stochastic Structures," by A. Kardara, C. Bucher and M. Shinozuka, 9/22/88, (PB89-
174429, A04, MF-A01).

NCEER-88-0030 "Nonnormal Accelerations Due to Yielding in a Primary Structure," by D.C.K. Chen and L.D. Lutes,
9/19/88, (PB89-131437, A04, MF-A01).

NCEER-88-0031 "Design Approaches for Soil-Structure Interaction," by A.S. Veletsos, A.M. Prasad and Y. Tang, 12/30/88,
(PB89-174437, A03, MF-A01). This report is available only through NTIS (see address given above).

NCEER-88-0032 "A Re-evaluation of Design Spectra for Seismic Damage Control," by C.J. Turkstra and A.G. Tallin, 11/7/88,
(PB89-145221, A05, MF-A01).

NCEER-88-0033 "The Behavior and Design of Noncontact Lap Splices Subjected to Repeated Inelastic Tensile Loading," by
V.E. Sagan, P. Gergely and R.N. White, 12/8/88, (PB89-163737, A08, MF-A01).

NCEER-88-0034 "Seismic Response of Pile Foundations," by S.M. Mamoon, P.K. Banerjee and S. Ahmad, 11/1/88, (PB89-
145239, A04, MF-A01).

NCEER-88-0035 "Modeling of R/C Building Structures With Flexible Floor Diaphragms (IDARC2)," by A.M. Reinhorn, S.K.
Kunnath and N. Panahshahi, 9/7/88, (PB89-207153, A07, MF-A01).

NCEER-88-0036 "Solution of the Dam-Reservoir Interaction Problem Using a Combination of FEM, BEM with Particular
Integrals, Modal Analysis, and Substructuring," by C-S. Tsai, G.C. Lee and R.L. Ketter, 12/31/88, (PB89-
207146, A04, MF-A01).

NCEER-88-0037 "Optimal Placement of Actuators for Structural Control," by F.Y. Cheng and C.P. Pantelides, 8/15/88,
(PB89-162846, A05, MF-A01).

NCEER-88-0038 "Teflon Bearings in Aseismic Base Isolation: Experimental Studies and Mathematical Modeling," by A.
Mokha, M.C. Constantinou and A.M. Reinhorn, 12/5/88, (PB89-218457, A10, MF-A01). This report is
available only through NTIS (see address given above).

NCEER-88-0039 "Seismic Behavior of Flat Slab High-Rise Buildings in the New York City Area," by P. Weidlinger and M.
Ettouney, 10/15/88, (PB90-145681, A04, MF-A01).

NCEER-88-0040 "Evaluation of the Earthquake Resistance of Existing Buildings in New York City," by P. Weidlinger and M.
Ettouney, 10/15/88, to be published.

NCEER-88-0041 "Small-Scale Modeling Techniques for Reinforced Concrete Structures Subjected to Seismic Loads," by W.
Kim, A. El-Attar and R.N. White, 11/22/88, (PB89-189625, A05, MF-A01).

Formerly the National Center for Earthquake Engineering Research 264

NCEER-88-0042 "Modeling Strong Ground Motion from Multiple Event Earthquakes," by G.W. Ellis and A.S. Cakmak,
10/15/88, (PB89-174445, A03, MF-A01).

NCEER-88-0043 "Nonstationary Models of Seismic Ground Acceleration," by M. Grigoriu, S.E. Ruiz and E. Rosenblueth,
7/15/88, (PB89-189617, A04, MF-A01).

NCEER-88-0044 "SARCF User's Guide: Seismic Analysis of Reinforced Concrete Frames," by Y.S. Chung, C. Meyer and M.
Shinozuka, 11/9/88, (PB89-174452, A08, MF-A01).

NCEER-88-0045 "First Expert Panel Meeting on Disaster Research and Planning," edited by J. Pantelic and J. Stoyle, 9/15/88,
(PB89-174460, A05, MF-A01).

NCEER-88-0046 "Preliminary Studies of the Effect of Degrading Infill Walls on the Nonlinear Seismic Response of Steel
Frames," by C.Z. Chrysostomou, P. Gergely and J.F. Abel, 12/19/88, (PB89-208383, A05, MF-A01).

NCEER-88-0047 "Reinforced Concrete Frame Component Testing Facility - Design, Construction, Instrumentation and
Operation," by S.P. Pessiki, C. Conley, T. Bond, P. Gergely and R.N. White, 12/16/88, (PB89-174478, A04,
MF-A01).

NCEER-89-0001 "Effects of Protective Cushion and Soil Compliancy on the Response of Equipment Within a Seismically
Excited Building," by J.A. HoLung, 2/16/89, (PB89-207179, A04, MF-A01).

NCEER-89-0002 "Statistical Evaluation of Response Modification Factors for Reinforced Concrete Structures," by H.H-M.
Hwang and J-W. Jaw, 2/17/89, (PB89-207187, A05, MF-A01).

NCEER-89-0003 "Hysteretic Columns Under Random Excitation," by G-Q. Cai and Y.K. Lin, 1/9/89, (PB89-196513, A03,
MF-A01).

NCEER-89-0004 "Experimental Study of `Elephant Foot Bulge' Instability of Thin-Walled Metal Tanks," by Z-H. Jia and R.L.
Ketter, 2/22/89, (PB89-207195, A03, MF-A01).

NCEER-89-0005 "Experiment on Performance of Buried Pipelines Across San Andreas Fault," by J. Isenberg, E. Richardson
and T.D. O'Rourke, 3/10/89, (PB89-218440, A04, MF-A01). This report is available only through NTIS (see
address given above).

NCEER-89-0006 "A Knowledge-Based Approach to Structural Design of Earthquake-Resistant Buildings," by M. Subramani,

P. Gergely, C.H. Conley, J.F. Abel and A.H. Zaghw, 1/15/89, (PB89-218465, A06, MF-A01).

NCEER-89-0007 "Liquefaction Hazards and Their Effects on Buried Pipelines," by T.D. O'Rourke and P.A. Lane, 2/1/89,
(PB89-218481, A09, MF-A01).

NCEER-89-0008 "Fundamentals of System Identification in Structural Dynamics," by H. Imai, C-B. Yun, O. Maruyama and
M. Shinozuka, 1/26/89, (PB89-207211, A04, MF-A01).

NCEER-89-0009 "Effects of the 1985 Michoacan Earthquake on Water Systems and Other Buried Lifelines in Mexico," by
A.G. Ayala and M.J. O'Rourke, 3/8/89, (PB89-207229, A06, MF-A01).

NCEER-89-R010 "NCEER Bibliography of Earthquake Education Materials," by K.E.K. Ross, Second Revision, 9/1/89,
(PB90-125352, A05, MF-A01). This report is replaced by NCEER-92-0018.

NCEER-89-0011 "Inelastic Three-Dimensional Response Analysis of Reinforced Concrete Building Structures (IDARC-3D),
Part I - Modeling," by S.K. Kunnath and A.M. Reinhorn, 4/17/89, (PB90-114612, A07, MF-A01). This
report is available only through NTIS (see address given above).

NCEER-89-0012 "Recommended Modifications to ATC-14," by C.D. Poland and J.O. Malley, 4/12/89, (PB90-108648, A15,
MF-A01).

NCEER-89-0013 "Repair and Strengthening of Beam-to-Column Connections Subjected to Earthquake Loading," by M.

Corazao and A.J. Durrani, 2/28/89, (PB90-109885, A06, MF-A01).

Formerly the National Center for Earthquake Engineering Research 265

NCEER-89-0014 "Program EXKAL2 for Identification of Structural Dynamic Systems," by O. Maruyama, C-B. Yun, M.
Hoshiya and M. Shinozuka, 5/19/89, (PB90-109877, A09, MF-A01).

NCEER-89-0015 "Response of Frames With Bolted Semi-Rigid Connections, Part I - Experimental Study and Analytical
Predictions," by P.J. DiCorso, A.M. Reinhorn, J.R. Dickerson, J.B. Radziminski and W.L. Harper, 6/1/89, to
be published.

NCEER-89-0016 "ARMA Monte Carlo Simulation in Probabilistic Structural Analysis," by P.D. Spanos and M.P. Mignolet,
7/10/89, (PB90-109893, A03, MF-A01).

NCEER-89-P017 "Preliminary Proceedings from the Conference on Disaster Preparedness - The Place of Earthquake
Education in Our Schools," Edited by K.E.K. Ross, 6/23/89, (PB90-108606, A03, MF-A01).

NCEER-89-0017 "Proceedings from the Conference on Disaster Preparedness - The Place of Earthquake Education in Our
Schools," Edited by K.E.K. Ross, 12/31/89, (PB90-207895, A012, MF-A02). This report is available only
through NTIS (see address given above).

NCEER-89-0018 "Multidimensional Models of Hysteretic Material Behavior for Vibration Analysis of Shape Memory Energy
Absorbing Devices, by E.J. Graesser and F.A. Cozzarelli, 6/7/89, (PB90-164146, A04, MF-A01).

NCEER-89-0019 "Nonlinear Dynamic Analysis of Three-Dimensional Base Isolated Structures (3D-BASIS)," by S.

Nagarajaiah, A.M. Reinhorn and M.C. Constantinou, 8/3/89, (PB90-161936, A06, MF-A01). This report has
been replaced by NCEER-93-0011.

NCEER-89-0020 "Structural Control Considering Time-Rate of Control Forces and Control Rate Constraints," by F.Y. Cheng
and C.P. Pantelides, 8/3/89, (PB90-120445, A04, MF-A01).

NCEER-89-0021 "Subsurface Conditions of Memphis and Shelby County," by K.W. Ng, T-S. Chang and H-H.M. Hwang,
7/26/89, (PB90-120437, A03, MF-A01).

NCEER-89-0022 "Seismic Wave Propagation Effects on Straight Jointed Buried Pipelines," by K. Elhmadi and M.J. O'Rourke,
8/24/89, (PB90-162322, A10, MF-A02).

NCEER-89-0023 "Workshop on Serviceability Analysis of Water Delivery Systems," edited by M. Grigoriu, 3/6/89, (PB90-
127424, A03, MF-A01).

NCEER-89-0024 "Shaking Table Study of a 1/5 Scale Steel Frame Composed of Tapered Members," by K.C. Chang, J.S.
Hwang and G.C. Lee, 9/18/89, (PB90-160169, A04, MF-A01).

NCEER-89-0025 "DYNA1D: A Computer Program for Nonlinear Seismic Site Response Analysis - Technical
Documentation," by Jean H. Prevost, 9/14/89, (PB90-161944, A07, MF-A01). This report is available only
through NTIS (see address given above).

NCEER-89-0026 "1:4 Scale Model Studies of Active Tendon Systems and Active Mass Dampers for Aseismic Protection," by
A.M. Reinhorn, T.T. Soong, R.C. Lin, Y.P. Yang, Y. Fukao, H. Abe and M. Nakai, 9/15/89, (PB90-173246,
A10, MF-A02). This report is available only through NTIS (see address given above).

NCEER-89-0027 "Scattering of Waves by Inclusions in a Nonhomogeneous Elastic Half Space Solved by Boundary Element
Methods," by P.K. Hadley, A. Askar and A.S. Cakmak, 6/15/89, (PB90-145699, A07, MF-A01).

NCEER-89-0028 "Statistical Evaluation of Deflection Amplification Factors for Reinforced Concrete Structures," by H.H.M.
Hwang, J-W. Jaw and A.L. Ch'ng, 8/31/89, (PB90-164633, A05, MF-A01).

NCEER-89-0029 "Bedrock Accelerations in Memphis Area Due to Large New Madrid Earthquakes," by H.H.M. Hwang,
C.H.S. Chen and G. Yu, 11/7/89, (PB90-162330, A04, MF-A01).

NCEER-89-0030 "Seismic Behavior and Response Sensitivity of Secondary Structural Systems," by Y.Q. Chen and T.T.
Soong, 10/23/89, (PB90-164658, A08, MF-A01).

NCEER-89-0031 "Random Vibration and Reliability Analysis of Primary-Secondary Structural Systems," by Y. Ibrahim, M.
Grigoriu and T.T. Soong, 11/10/89, (PB90-161951, A04, MF-A01).

Formerly the National Center for Earthquake Engineering Research 266

NCEER-89-0032 "Proceedings from the Second U.S. - Japan Workshop on Liquefaction, Large Ground Deformation and
Their Effects on Lifelines, September 26-29, 1989," Edited by T.D. O'Rourke and M. Hamada, 12/1/89,
(PB90-209388, A22, MF-A03).

NCEER-89-0033 "Deterministic Model for Seismic Damage Evaluation of Reinforced Concrete Structures," by J.M. Bracci,
A.M. Reinhorn, J.B. Mander and S.K. Kunnath, 9/27/89, (PB91-108803, A06, MF-A01).

NCEER-89-0034 "On the Relation Between Local and Global Damage Indices," by E. DiPasquale and A.S. Cakmak, 8/15/89,
(PB90-173865, A05, MF-A01).

NCEER-89-0035 "Cyclic Undrained Behavior of Nonplastic and Low Plasticity Silts," by A.J. Walker and H.E. Stewart,
7/26/89, (PB90-183518, A10, MF-A01).

NCEER-89-0036 "Liquefaction Potential of Surficial Deposits in the City of Buffalo, New York," by M. Budhu, R. Giese and
L. Baumgrass, 1/17/89, (PB90-208455, A04, MF-A01).

NCEER-89-0037 "A Deterministic Assessment of Effects of Ground Motion Incoherence," by A.S. Veletsos and Y. Tang,
7/15/89, (PB90-164294, A03, MF-A01).

NCEER-89-0038 "Workshop on Ground Motion Parameters for Seismic Hazard Mapping," July 17-18, 1989, edited by R.V.
Whitman, 12/1/89, (PB90-173923, A04, MF-A01).

NCEER-89-0039 "Seismic Effects on Elevated Transit Lines of the New York City Transit Authority," by C.J. Costantino,
C.A. Miller and E. Heymsfield, 12/26/89, (PB90-207887, A06, MF-A01).

NCEER-89-0040 "Centrifugal Modeling of Dynamic Soil-Structure Interaction," by K. Weissman, Supervised by J.H. Prevost,
5/10/89, (PB90-207879, A07, MF-A01).

NCEER-89-0041 "Linearized Identification of Buildings With Cores for Seismic Vulnerability Assessment," by I-K. Ho and
A.E. Aktan, 11/1/89, (PB90-251943, A07, MF-A01).

NCEER-90-0001 "Geotechnical and Lifeline Aspects of the October 17, 1989 Loma Prieta Earthquake in San Francisco," by
T.D. O'Rourke, H.E. Stewart, F.T. Blackburn and T.S. Dickerman, 1/90, (PB90-208596, A05, MF-A01).

NCEER-90-0002 "Nonnormal Secondary Response Due to Yielding in a Primary Structure," by D.C.K. Chen and L.D. Lutes,
2/28/90, (PB90-251976, A07, MF-A01).

NCEER-90-0003 "Earthquake Education Materials for Grades K-12," by K.E.K. Ross, 4/16/90, (PB91-251984, A05, MF-
A05). This report has been replaced by NCEER-92-0018.

NCEER-90-0004 "Catalog of Strong Motion Stations in Eastern North America," by R.W. Busby, 4/3/90, (PB90-251984, A05,
MF-A01).

NCEER-90-0005 "NCEER Strong-Motion Data Base: A User Manual for the GeoBase Release (Version 1.0 for the Sun3)," by
P. Friberg and K. Jacob, 3/31/90 (PB90-258062, A04, MF-A01).

NCEER-90-0006 "Seismic Hazard Along a Crude Oil Pipeline in the Event of an 1811-1812 Type New Madrid Earthquake,"
by H.H.M. Hwang and C-H.S. Chen, 4/16/90, (PB90-258054, A04, MF-A01).

NCEER-90-0007 "Site-Specific Response Spectra for Memphis Sheahan Pumping Station," by H.H.M. Hwang and C.S. Lee,
5/15/90, (PB91-108811, A05, MF-A01).

NCEER-90-0008 "Pilot Study on Seismic Vulnerability of Crude Oil Transmission Systems," by T. Ariman, R. Dobry, M.
Grigoriu, F. Kozin, M. O'Rourke, T. O'Rourke and M. Shinozuka, 5/25/90, (PB91-108837, A06, MF-A01).

NCEER-90-0009 "A Program to Generate Site Dependent Time Histories: EQGEN," by G.W. Ellis, M. Srinivasan and A.S.
Cakmak, 1/30/90, (PB91-108829, A04, MF-A01).

NCEER-90-0010 "Active Isolation for Seismic Protection of Operating Rooms," by M.E. Talbott, Supervised by M.
Shinozuka, 6/8/9, (PB91-110205, A05, MF-A01).

Formerly the National Center for Earthquake Engineering Research 267

NCEER-90-0011 "Program LINEARID for Identification of Linear Structural Dynamic Systems," by C-B. Yun and M.
Shinozuka, 6/25/90, (PB91-110312, A08, MF-A01).

NCEER-90-0012 "Two-Dimensional Two-Phase Elasto-Plastic Seismic Response of Earth Dams," by A.N. Yiagos, Supervised
by J.H. Prevost, 6/20/90, (PB91-110197, A13, MF-A02).

NCEER-90-0013 "Secondary Systems in Base-Isolated Structures: Experimental Investigation, Stochastic Response and
Stochastic Sensitivity," by G.D. Manolis, G. Juhn, M.C. Constantinou and A.M. Reinhorn, 7/1/90, (PB91-
110320, A08, MF-A01).

NCEER-90-0014 "Seismic Behavior of Lightly-Reinforced Concrete Column and Beam-Column Joint Details," by S.P.
Pessiki, C.H. Conley, P. Gergely and R.N. White, 8/22/90, (PB91-108795, A11, MF-A02).

NCEER-90-0015 "Two Hybrid Control Systems for Building Structures Under Strong Earthquakes," by J.N. Yang and A.
Danielians, 6/29/90, (PB91-125393, A04, MF-A01).

NCEER-90-0016 "Instantaneous Optimal Control with Acceleration and Velocity Feedback," by J.N. Yang and Z. Li, 6/29/90,
(PB91-125401, A03, MF-A01).

NCEER-90-0017 "Reconnaissance Report on the Northern Iran Earthquake of June 21, 1990," by M. Mehrain, 10/4/90, (PB91-
125377, A03, MF-A01).

NCEER-90-0018 "Evaluation of Liquefaction Potential in Memphis and Shelby County," by T.S. Chang, P.S. Tang, C.S. Lee
and H. Hwang, 8/10/90, (PB91-125427, A09, MF-A01).

NCEER-90-0019 "Experimental and Analytical Study of a Combined Sliding Disc Bearing and Helical Steel Spring Isolation
System," by M.C. Constantinou, A.S. Mokha and A.M. Reinhorn, 10/4/90, (PB91-125385, A06, MF-A01).
This report is available only through NTIS (see address given above).

NCEER-90-0020 "Experimental Study and Analytical Prediction of Earthquake Response of a Sliding Isolation System with a
Spherical Surface," by A.S. Mokha, M.C. Constantinou and A.M. Reinhorn, 10/11/90, (PB91-125419, A05,
MF-A01).

NCEER-90-0021 "Dynamic Interaction Factors for Floating Pile Groups," by G. Gazetas, K. Fan, A. Kaynia and E. Kausel,
9/10/90, (PB91-170381, A05, MF-A01).

NCEER-90-0022 "Evaluation of Seismic Damage Indices for Reinforced Concrete Structures," by S. Rodriguez-Gomez and
A.S. Cakmak, 9/30/90, PB91-171322, A06, MF-A01).

NCEER-90-0023 "Study of Site Response at a Selected Memphis Site," by H. Desai, S. Ahmad, E.S. Gazetas and M.R. Oh,
10/11/90, (PB91-196857, A03, MF-A01).

NCEER-90-0024 "A User's Guide to Strongmo: Version 1.0 of NCEER's Strong-Motion Data Access Tool for PCs and
Terminals," by P.A. Friberg and C.A.T. Susch, 11/15/90, (PB91-171272, A03, MF-A01).

NCEER-90-0025 "A Three-Dimensional Analytical Study of Spatial Variability of Seismic Ground Motions," by L-L. Hong
and A.H.-S. Ang, 10/30/90, (PB91-170399, A09, MF-A01).

NCEER-90-0026 "MUMOID User's Guide - A Program for the Identification of Modal Parameters," by S. Rodriguez-Gomez
and E. DiPasquale, 9/30/90, (PB91-171298, A04, MF-A01).

NCEER-90-0027 "SARCF-II User's Guide - Seismic Analysis of Reinforced Concrete Frames," by S. Rodriguez-Gomez, Y.S.
Chung and C. Meyer, 9/30/90, (PB91-171280, A05, MF-A01).

NCEER-90-0028 "Viscous Dampers: Testing, Modeling and Application in Vibration and Seismic Isolation," by N. Makris
and M.C. Constantinou, 12/20/90 (PB91-190561, A06, MF-A01).

NCEER-90-0029 "Soil Effects on Earthquake Ground Motions in the Memphis Area," by H. Hwang, C.S. Lee, K.W. Ng and
T.S. Chang, 8/2/90, (PB91-190751, A05, MF-A01).

Formerly the National Center for Earthquake Engineering Research 268

NCEER-91-0001 "Proceedings from the Third Japan-U.S. Workshop on Earthquake Resistant Design of Lifeline Facilities and
Countermeasures for Soil Liquefaction, December 17-19, 1990," edited by T.D. O'Rourke and M. Hamada,
2/1/91, (PB91-179259, A99, MF-A04).

NCEER-91-0002 "Physical Space Solutions of Non-Proportionally Damped Systems," by M. Tong, Z. Liang and G.C. Lee,
1/15/91, (PB91-179242, A04, MF-A01).

NCEER-91-0003 "Seismic Response of Single Piles and Pile Groups," by K. Fan and G. Gazetas, 1/10/91, (PB92-174994,
A04, MF-A01).

NCEER-91-0004 "Damping of Structures: Part 1 - Theory of Complex Damping," by Z. Liang and G. Lee, 10/10/91, (PB92-
197235, A12, MF-A03).

NCEER-91-0005 "3D-BASIS - Nonlinear Dynamic Analysis of Three Dimensional Base Isolated Structures: Part II," by S.
Nagarajaiah, A.M. Reinhorn and M.C. Constantinou, 2/28/91, (PB91-190553, A07, MF-A01). This report
has been replaced by NCEER-93-0011.

NCEER-91-0006 "A Multidimensional Hysteretic Model for Plasticity Deforming Metals in Energy Absorbing Devices," by
E.J. Graesser and F.A. Cozzarelli, 4/9/91, (PB92-108364, A04, MF-A01).

NCEER-91-0007 "A Framework for Customizable Knowledge-Based Expert Systems with an Application to a KBES for
Evaluating the Seismic Resistance of Existing Buildings," by E.G. Ibarra-Anaya and S.J. Fenves, 4/9/91,
(PB91-210930, A08, MF-A01).

NCEER-91-0008 "Nonlinear Analysis of Steel Frames with Semi-Rigid Connections Using the Capacity Spectrum Method,"
by G.G. Deierlein, S-H. Hsieh, Y-J. Shen and J.F. Abel, 7/2/91, (PB92-113828, A05, MF-A01).

NCEER-91-0009 "Earthquake Education Materials for Grades K-12," by K.E.K. Ross, 4/30/91, (PB91-212142, A06, MF-
A01). This report has been replaced by NCEER-92-0018.

NCEER-91-0010 "Phase Wave Velocities and Displacement Phase Differences in a Harmonically Oscillating Pile," by N.
Makris and G. Gazetas, 7/8/91, (PB92-108356, A04, MF-A01).

NCEER-91-0011 "Dynamic Characteristics of a Full-Size Five-Story Steel Structure and a 2/5 Scale Model," by K.C. Chang,
G.C. Yao, G.C. Lee, D.S. Hao and Y.C. Yeh," 7/2/91, (PB93-116648, A06, MF-A02).

NCEER-91-0012 "Seismic Response of a 2/5 Scale Steel Structure with Added Viscoelastic Dampers," by K.C. Chang, T.T.
Soong, S-T. Oh and M.L. Lai, 5/17/91, (PB92-110816, A05, MF-A01).

NCEER-91-0013 "Earthquake Response of Retaining Walls; Full-Scale Testing and Computational Modeling," by S.
Alampalli and A-W.M. Elgamal, 6/20/91, to be published.

NCEER-91-0014 "3D-BASIS-M: Nonlinear Dynamic Analysis of Multiple Building Base Isolated Structures," by P.C.
Tsopelas, S. Nagarajaiah, M.C. Constantinou and A.M. Reinhorn, 5/28/91, (PB92-113885, A09, MF-A02).

NCEER-91-0015 "Evaluation of SEAOC Design Requirements for Sliding Isolated Structures," by D. Theodossiou and M.C.
Constantinou, 6/10/91, (PB92-114602, A11, MF-A03).

NCEER-91-0016 "Closed-Loop Modal Testing of a 27-Story Reinforced Concrete Flat Plate-Core Building," by H.R.
Somaprasad, T. Toksoy, H. Yoshiyuki and A.E. Aktan, 7/15/91, (PB92-129980, A07, MF-A02).

NCEER-91-0017 "Shake Table Test of a 1/6 Scale Two-Story Lightly Reinforced Concrete Building," by A.G. El-Attar, R.N.
White and P. Gergely, 2/28/91, (PB92-222447, A06, MF-A02).

NCEER-91-0018 "Shake Table Test of a 1/8 Scale Three-Story Lightly Reinforced Concrete Building," by A.G. El-Attar, R.N.
White and P. Gergely, 2/28/91, (PB93-116630, A08, MF-A02).

NCEER-91-0019 "Transfer Functions for Rigid Rectangular Foundations," by A.S. Veletsos, A.M. Prasad and W.H. Wu,
7/31/91, to be published.

Formerly the National Center for Earthquake Engineering Research 269

NCEER-91-0020 "Hybrid Control of Seismic-Excited Nonlinear and Inelastic Structural Systems," by J.N. Yang, Z. Li and A.
Danielians, 8/1/91, (PB92-143171, A06, MF-A02).

NCEER-91-0021 "The NCEER-91 Earthquake Catalog: Improved Intensity-Based Magnitudes and Recurrence Relations for
U.S. Earthquakes East of New Madrid," by L. Seeber and J.G. Armbruster, 8/28/91, (PB92-176742, A06,
MF-A02).

NCEER-91-0022 "Proceedings from the Implementation of Earthquake Planning and Education in Schools: The Need for
Change - The Roles of the Changemakers," by K.E.K. Ross and F. Winslow, 7/23/91, (PB92-129998, A12,
MF-A03).

NCEER-91-0023 "A Study of Reliability-Based Criteria for Seismic Design of Reinforced Concrete Frame Buildings," by
H.H.M. Hwang and H-M. Hsu, 8/10/91, (PB92-140235, A09, MF-A02).

NCEER-91-0024 "Experimental Verification of a Number of Structural System Identification Algorithms," by R.G. Ghanem,
H. Gavin and M. Shinozuka, 9/18/91, (PB92-176577, A18, MF-A04).

NCEER-91-0025 "Probabilistic Evaluation of Liquefaction Potential," by H.H.M. Hwang and C.S. Lee," 11/25/91, (PB92-
143429, A05, MF-A01).

NCEER-91-0026 "Instantaneous Optimal Control for Linear, Nonlinear and Hysteretic Structures - Stable Controllers," by J.N.
Yang and Z. Li, 11/15/91, (PB92-163807, A04, MF-A01).

NCEER-91-0027 "Experimental and Theoretical Study of a Sliding Isolation System for Bridges," by M.C. Constantinou, A.
Kartoum, A.M. Reinhorn and P. Bradford, 11/15/91, (PB92-176973, A10, MF-A03).

NCEER-92-0001 "Case Studies of Liquefaction and Lifeline Performance During Past Earthquakes, Volume 1: Japanese Case
Studies," Edited by M. Hamada and T. O'Rourke, 2/17/92, (PB92-197243, A18, MF-A04).

NCEER-92-0002 "Case Studies of Liquefaction and Lifeline Performance During Past Earthquakes, Volume 2: United States
Case Studies," Edited by T. O'Rourke and M. Hamada, 2/17/92, (PB92-197250, A20, MF-A04).

NCEER-92-0003 "Issues in Earthquake Education," Edited by K. Ross, 2/3/92, (PB92-222389, A07, MF-A02).

NCEER-92-0004 "Proceedings from the First U.S. - Japan Workshop on Earthquake Protective Systems for Bridges," Edited
by I.G. Buckle, 2/4/92, (PB94-142239, A99, MF-A06).

NCEER-92-0005 "Seismic Ground Motion from a Haskell-Type Source in a Multiple-Layered Half-Space," A.P. Theoharis, G.
Deodatis and M. Shinozuka, 1/2/92, to be published.

NCEER-92-0006 "Proceedings from the Site Effects Workshop," Edited by R. Whitman, 2/29/92, (PB92-197201, A04, MF-
A01).

NCEER-92-0007 "Engineering Evaluation of Permanent Ground Deformations Due to Seismically-Induced Liquefaction," by

M.H. Baziar, R. Dobry and A-W.M. Elgamal, 3/24/92, (PB92-222421, A13, MF-A03).

NCEER-92-0008 "A Procedure for the Seismic Evaluation of Buildings in the Central and Eastern United States," by C.D.
Poland and J.O. Malley, 4/2/92, (PB92-222439, A20, MF-A04).

NCEER-92-0009 "Experimental and Analytical Study of a Hybrid Isolation System Using Friction Controllable Sliding
Bearings," by M.Q. Feng, S. Fujii and M. Shinozuka, 5/15/92, (PB93-150282, A06, MF-A02).

NCEER-92-0010 "Seismic Resistance of Slab-Column Connections in Existing Non-Ductile Flat-Plate Buildings," by A.J.
Durrani and Y. Du, 5/18/92, (PB93-116812, A06, MF-A02).

NCEER-92-0011 "The Hysteretic and Dynamic Behavior of Brick Masonry Walls Upgraded by Ferrocement Coatings Under
Cyclic Loading and Strong Simulated Ground Motion," by H. Lee and S.P. Prawel, 5/11/92, to be published.

NCEER-92-0012 "Study of Wire Rope Systems for Seismic Protection of Equipment in Buildings," by G.F. Demetriades,
M.C. Constantinou and A.M. Reinhorn, 5/20/92, (PB93-116655, A08, MF-A02).

Formerly the National Center for Earthquake Engineering Research 270

NCEER-92-0013 "Shape Memory Structural Dampers: Material Properties, Design and Seismic Testing," by P.R. Witting and
F.A. Cozzarelli, 5/26/92, (PB93-116663, A05, MF-A01).

NCEER-92-0014 "Longitudinal Permanent Ground Deformation Effects on Buried Continuous Pipelines," by M.J. O'Rourke,
and C. Nordberg, 6/15/92, (PB93-116671, A08, MF-A02).

NCEER-92-0015 "A Simulation Method for Stationary Gaussian Random Functions Based on the Sampling Theorem," by M.
Grigoriu and S. Balopoulou, 6/11/92, (PB93-127496, A05, MF-A01).

NCEER-92-0016 "Gravity-Load-Designed Reinforced Concrete Buildings: Seismic Evaluation of Existing Construction and
Detailing Strategies for Improved Seismic Resistance," by G.W. Hoffmann, S.K. Kunnath, A.M. Reinhorn
and J.B. Mander, 7/15/92, (PB94-142007, A08, MF-A02).

NCEER-92-0017 "Observations on Water System and Pipeline Performance in the Limón Area of Costa Rica Due to the April
22, 1991 Earthquake," by M. O'Rourke and D. Ballantyne, 6/30/92, (PB93-126811, A06, MF-A02).

NCEER-92-0018 "Fourth Edition of Earthquake Education Materials for Grades K-12," Edited by K.E.K. Ross, 8/10/92,
(PB93-114023, A07, MF-A02).

NCEER-92-0019 "Proceedings from the Fourth Japan-U.S. Workshop on Earthquake Resistant Design of Lifeline Facilities
and Countermeasures for Soil Liquefaction," Edited by M. Hamada and T.D. O'Rourke, 8/12/92, (PB93-
163939, A99, MF-E11).

NCEER-92-0020 "Active Bracing System: A Full Scale Implementation of Active Control," by A.M. Reinhorn, T.T. Soong,
R.C. Lin, M.A. Riley, Y.P. Wang, S. Aizawa and M. Higashino, 8/14/92, (PB93-127512, A06, MF-A02).

NCEER-92-0021 "Empirical Analysis of Horizontal Ground Displacement Generated by Liquefaction-Induced Lateral

Spreads," by S.F. Bartlett and T.L. Youd, 8/17/92, (PB93-188241, A06, MF-A02).

NCEER-92-0022 "IDARC Version 3.0: Inelastic Damage Analysis of Reinforced Concrete Structures," by S.K. Kunnath, A.M.
Reinhorn and R.F. Lobo, 8/31/92, (PB93-227502, A07, MF-A02).

NCEER-92-0023 "A Semi-Empirical Analysis of Strong-Motion Peaks in Terms of Seismic Source, Propagation Path and
Local Site Conditions, by M. Kamiyama, M.J. O'Rourke and R. Flores-Berrones, 9/9/92, (PB93-150266,
A08, MF-A02).

NCEER-92-0024 "Seismic Behavior of Reinforced Concrete Frame Structures with Nonductile Details, Part I: Summary of
Experimental Findings of Full Scale Beam-Column Joint Tests," by A. Beres, R.N. White and P. Gergely,
9/30/92, (PB93-227783, A05, MF-A01).

NCEER-92-0025 "Experimental Results of Repaired and Retrofitted Beam-Column Joint Tests in Lightly Reinforced Concrete
Frame Buildings," by A. Beres, S. El-Borgi, R.N. White and P. Gergely, 10/29/92, (PB93-227791, A05, MF-
A01).

NCEER-92-0026 "A Generalization of Optimal Control Theory: Linear and Nonlinear Structures," by J.N. Yang, Z. Li and S.
Vongchavalitkul, 11/2/92, (PB93-188621, A05, MF-A01).

NCEER-92-0027 "Seismic Resistance of Reinforced Concrete Frame Structures Designed Only for Gravity Loads: Part I -
Design and Properties of a One-Third Scale Model Structure," by J.M. Bracci, A.M. Reinhorn and J.B.
Mander, 12/1/92, (PB94-104502, A08, MF-A02).

NCEER-92-0028 "Seismic Resistance of Reinforced Concrete Frame Structures Designed Only for Gravity Loads: Part II -
Experimental Performance of Subassemblages," by L.E. Aycardi, J.B. Mander and A.M. Reinhorn, 12/1/92,
(PB94-104510, A08, MF-A02).

NCEER-92-0029 "Seismic Resistance of Reinforced Concrete Frame Structures Designed Only for Gravity Loads: Part III -
Experimental Performance and Analytical Study of a Structural Model," by J.M. Bracci, A.M. Reinhorn and
J.B. Mander, 12/1/92, (PB93-227528, A09, MF-A01).

Formerly the National Center for Earthquake Engineering Research 271

NCEER-92-0030 "Evaluation of Seismic Retrofit of Reinforced Concrete Frame Structures: Part I - Experimental Performance
of Retrofitted Subassemblages," by D. Choudhuri, J.B. Mander and A.M. Reinhorn, 12/8/92, (PB93-198307,
A07, MF-A02).

NCEER-92-0031 "Evaluation of Seismic Retrofit of Reinforced Concrete Frame Structures: Part II - Experimental
Performance and Analytical Study of a Retrofitted Structural Model," by J.M. Bracci, A.M. Reinhorn and
J.B. Mander, 12/8/92, (PB93-198315, A09, MF-A03).

NCEER-92-0032 "Experimental and Analytical Investigation of Seismic Response of Structures with Supplemental Fluid
Viscous Dampers," by M.C. Constantinou and M.D. Symans, 12/21/92, (PB93-191435, A10, MF-A03). This
report is available only through NTIS (see address given above).

NCEER-92-0033 "Reconnaissance Report on the Cairo, Egypt Earthquake of October 12, 1992," by M. Khater, 12/23/92,
(PB93-188621, A03, MF-A01).

NCEER-92-0034 "Low-Level Dynamic Characteristics of Four Tall Flat-Plate Buildings in New York City," by H. Gavin, S.
Yuan, J. Grossman, E. Pekelis and K. Jacob, 12/28/92, (PB93-188217, A07, MF-A02).

NCEER-93-0001 "An Experimental Study on the Seismic Performance of Brick-Infilled Steel Frames With and Without
Retrofit," by J.B. Mander, B. Nair, K. Wojtkowski and J. Ma, 1/29/93, (PB93-227510, A07, MF-A02).

NCEER-93-0002 "Social Accounting for Disaster Preparedness and Recovery Planning," by S. Cole, E. Pantoja and V. Razak,
2/22/93, (PB94-142114, A12, MF-A03).

NCEER-93-0003 "Assessment of 1991 NEHRP Provisions for Nonstructural Components and Recommended Revisions," by
T.T. Soong, G. Chen, Z. Wu, R-H. Zhang and M. Grigoriu, 3/1/93, (PB93-188639, A06, MF-A02).

NCEER-93-0004 "Evaluation of Static and Response Spectrum Analysis Procedures of SEAOC/UBC for Seismic Isolated
Structures," by C.W. Winters and M.C. Constantinou, 3/23/93, (PB93-198299, A10, MF-A03).

NCEER-93-0005 "Earthquakes in the Northeast - Are We Ignoring the Hazard? A Workshop on Earthquake Science and
Safety for Educators," edited by K.E.K. Ross, 4/2/93, (PB94-103066, A09, MF-A02).

NCEER-93-0006 "Inelastic Response of Reinforced Concrete Structures with Viscoelastic Braces," by R.F. Lobo, J.M. Bracci,
K.L. Shen, A.M. Reinhorn and T.T. Soong, 4/5/93, (PB93-227486, A05, MF-A02).

NCEER-93-0007 "Seismic Testing of Installation Methods for Computers and Data Processing Equipment," by K. Kosar, T.T.
Soong, K.L. Shen, J.A. HoLung and Y.K. Lin, 4/12/93, (PB93-198299, A07, MF-A02).

NCEER-93-0008 "Retrofit of Reinforced Concrete Frames Using Added Dampers," by A. Reinhorn, M. Constantinou and C.
Li, to be published.

NCEER-93-0009 "Seismic Behavior and Design Guidelines for Steel Frame Structures with Added Viscoelastic Dampers," by
K.C. Chang, M.L. Lai, T.T. Soong, D.S. Hao and Y.C. Yeh, 5/1/93, (PB94-141959, A07, MF-A02).

NCEER-93-0010 "Seismic Performance of Shear-Critical Reinforced Concrete Bridge Piers," by J.B. Mander, S.M. Waheed,
M.T.A. Chaudhary and S.S. Chen, 5/12/93, (PB93-227494, A08, MF-A02).

NCEER-93-0011 "3D-BASIS-TABS: Computer Program for Nonlinear Dynamic Analysis of Three Dimensional Base Isolated
Structures," by S. Nagarajaiah, C. Li, A.M. Reinhorn and M.C. Constantinou, 8/2/93, (PB94-141819, A09,
MF-A02).

NCEER-93-0012 "Effects of Hydrocarbon Spills from an Oil Pipeline Break on Ground Water," by O.J. Helweg and H.H.M.
Hwang, 8/3/93, (PB94-141942, A06, MF-A02).

NCEER-93-0013 "Simplified Procedures for Seismic Design of Nonstructural Components and Assessment of Current Code
Provisions," by M.P. Singh, L.E. Suarez, E.E. Matheu and G.O. Maldonado, 8/4/93, (PB94-141827, A09,
MF-A02).

NCEER-93-0014 "An Energy Approach to Seismic Analysis and Design of Secondary Systems," by G. Chen and T.T. Soong,
8/6/93, (PB94-142767, A11, MF-A03).

Formerly the National Center for Earthquake Engineering Research 272

NCEER-93-0015 "Proceedings from School Sites: Becoming Prepared for Earthquakes - Commemorating the Third
Anniversary of the Loma Prieta Earthquake," Edited by F.E. Winslow and K.E.K. Ross, 8/16/93, (PB94-
154275, A16, MF-A02).

NCEER-93-0016 "Reconnaissance Report of Damage to Historic Monuments in Cairo, Egypt Following the October 12, 1992
Dahshur Earthquake," by D. Sykora, D. Look, G. Croci, E. Karaesmen and E. Karaesmen, 8/19/93, (PB94-
142221, A08, MF-A02).

NCEER-93-0017 "The Island of Guam Earthquake of August 8, 1993," by S.W. Swan and S.K. Harris, 9/30/93, (PB94-
141843, A04, MF-A01).

NCEER-93-0018 "Engineering Aspects of the October 12, 1992 Egyptian Earthquake," by A.W. Elgamal, M. Amer, K.
Adalier and A. Abul-Fadl, 10/7/93, (PB94-141983, A05, MF-A01).

NCEER-93-0019 "Development of an Earthquake Motion Simulator and its Application in Dynamic Centrifuge Testing," by I.
Krstelj, Supervised by J.H. Prevost, 10/23/93, (PB94-181773, A-10, MF-A03).

NCEER-93-0020 "NCEER-Taisei Corporation Research Program on Sliding Seismic Isolation Systems for Bridges:
Experimental and Analytical Study of a Friction Pendulum System (FPS)," by M.C. Constantinou, P.
Tsopelas, Y-S. Kim and S. Okamoto, 11/1/93, (PB94-142775, A08, MF-A02).

NCEER-93-0021 "Finite Element Modeling of Elastomeric Seismic Isolation Bearings," by L.J. Billings, Supervised by R.
Shepherd, 11/8/93, to be published.

NCEER-93-0022 "Seismic Vulnerability of Equipment in Critical Facilities: Life-Safety and Operational Consequences," by
K. Porter, G.S. Johnson, M.M. Zadeh, C. Scawthorn and S. Eder, 11/24/93, (PB94-181765, A16, MF-A03).

NCEER-93-0023 "Hokkaido Nansei-oki, Japan Earthquake of July 12, 1993, by P.I. Yanev and C.R. Scawthorn, 12/23/93,
(PB94-181500, A07, MF-A01).

NCEER-94-0001 "An Evaluation of Seismic Serviceability of Water Supply Networks with Application to the San Francisco
Auxiliary Water Supply System," by I. Markov, Supervised by M. Grigoriu and T. O'Rourke, 1/21/94,
(PB94-204013, A07, MF-A02).

NCEER-94-0002 "NCEER-Taisei Corporation Research Program on Sliding Seismic Isolation Systems for Bridges:
Experimental and Analytical Study of Systems Consisting of Sliding Bearings, Rubber Restoring Force
Devices and Fluid Dampers," Volumes I and II, by P. Tsopelas, S. Okamoto, M.C. Constantinou, D. Ozaki
and S. Fujii, 2/4/94, (PB94-181740, A09, MF-A02 and PB94-181757, A12, MF-A03).

NCEER-94-0003 "A Markov Model for Local and Global Damage Indices in Seismic Analysis," by S. Rahman and M.
Grigoriu, 2/18/94, (PB94-206000, A12, MF-A03).

NCEER-94-0004 "Proceedings from the NCEER Workshop on Seismic Response of Masonry Infills," edited by D.P. Abrams,
3/1/94, (PB94-180783, A07, MF-A02).

NCEER-94-0005 "The Northridge, California Earthquake of January 17, 1994: General Reconnaissance Report," edited by
J.D. Goltz, 3/11/94, (PB193943, A10, MF-A03).

NCEER-94-0006 "Seismic Energy Based Fatigue Damage Analysis of Bridge Columns: Part I - Evaluation of Seismic
Capacity," by G.A. Chang and J.B. Mander, 3/14/94, (PB94-219185, A11, MF-A03).

NCEER-94-0007 "Seismic Isolation of Multi-Story Frame Structures Using Spherical Sliding Isolation Systems," by T.M. Al-
Hussaini, V.A. Zayas and M.C. Constantinou, 3/17/94, (PB193745, A09, MF-A02).

NCEER-94-0008 "The Northridge, California Earthquake of January 17, 1994: Performance of Highway Bridges," edited by
I.G. Buckle, 3/24/94, (PB94-193851, A06, MF-A02).

NCEER-94-0009 "Proceedings of the Third U.S.-Japan Workshop on Earthquake Protective Systems for Bridges," edited by
I.G. Buckle and I. Friedland, 3/31/94, (PB94-195815, A99, MF-A06).

Formerly the National Center for Earthquake Engineering Research 273

NCEER-94-0010 "3D-BASIS-ME: Computer Program for Nonlinear Dynamic Analysis of Seismically Isolated Single and
Multiple Structures and Liquid Storage Tanks," by P.C. Tsopelas, M.C. Constantinou and A.M. Reinhorn,
4/12/94, (PB94-204922, A09, MF-A02).

NCEER-94-0011 "The Northridge, California Earthquake of January 17, 1994: Performance of Gas Transmission Pipelines,"
by T.D. O'Rourke and M.C. Palmer, 5/16/94, (PB94-204989, A05, MF-A01).

NCEER-94-0012 "Feasibility Study of Replacement Procedures and Earthquake Performance Related to Gas Transmission
Pipelines," by T.D. O'Rourke and M.C. Palmer, 5/25/94, (PB94-206638, A09, MF-A02).

NCEER-94-0013 "Seismic Energy Based Fatigue Damage Analysis of Bridge Columns: Part II - Evaluation of Seismic
Demand," by G.A. Chang and J.B. Mander, 6/1/94, (PB95-18106, A08, MF-A02).

NCEER-94-0014 "NCEER-Taisei Corporation Research Program on Sliding Seismic Isolation Systems for Bridges:
Experimental and Analytical Study of a System Consisting of Sliding Bearings and Fluid Restoring
Force/Damping Devices," by P. Tsopelas and M.C. Constantinou, 6/13/94, (PB94-219144, A10, MF-A03).

NCEER-94-0015 "Generation of Hazard-Consistent Fragility Curves for Seismic Loss Estimation Studies," by H. Hwang and
J-R. Huo, 6/14/94, (PB95-181996, A09, MF-A02).

NCEER-94-0016 "Seismic Study of Building Frames with Added Energy-Absorbing Devices," by W.S. Pong, C.S. Tsai and
G.C. Lee, 6/20/94, (PB94-219136, A10, A03).

NCEER-94-0017 "Sliding Mode Control for Seismic-Excited Linear and Nonlinear Civil Engineering Structures," by J. Yang,
J. Wu, A. Agrawal and Z. Li, 6/21/94, (PB95-138483, A06, MF-A02).

NCEER-94-0018 "3D-BASIS-TABS Version 2.0: Computer Program for Nonlinear Dynamic Analysis of Three Dimensional
Base Isolated Structures," by A.M. Reinhorn, S. Nagarajaiah, M.C. Constantinou, P. Tsopelas and R. Li,
6/22/94, (PB95-182176, A08, MF-A02).

NCEER-94-0019 "Proceedings of the International Workshop on Civil Infrastructure Systems: Application of Intelligent
Systems and Advanced Materials on Bridge Systems," Edited by G.C. Lee and K.C. Chang, 7/18/94, (PB95-
252474, A20, MF-A04).

NCEER-94-0020 "Study of Seismic Isolation Systems for Computer Floors," by V. Lambrou and M.C. Constantinou, 7/19/94,
(PB95-138533, A10, MF-A03).

NCEER-94-0021 "Proceedings of the U.S.-Italian Workshop on Guidelines for Seismic Evaluation and Rehabilitation of
Unreinforced Masonry Buildings," Edited by D.P. Abrams and G.M. Calvi, 7/20/94, (PB95-138749, A13,
MF-A03).

NCEER-94-0022 "NCEER-Taisei Corporation Research Program on Sliding Seismic Isolation Systems for Bridges:
Experimental and Analytical Study of a System Consisting of Lubricated PTFE Sliding Bearings and Mild
Steel Dampers," by P. Tsopelas and M.C. Constantinou, 7/22/94, (PB95-182184, A08, MF-A02).

NCEER-94-0023 “Development of Reliability-Based Design Criteria for Buildings Under Seismic Load,” by Y.K. Wen, H.
Hwang and M. Shinozuka, 8/1/94, (PB95-211934, A08, MF-A02).

NCEER-94-0024 “Experimental Verification of Acceleration Feedback Control Strategies for an Active Tendon System,” by
S.J. Dyke, B.F. Spencer, Jr., P. Quast, M.K. Sain, D.C. Kaspari, Jr. and T.T. Soong, 8/29/94, (PB95-212320,
A05, MF-A01).

NCEER-94-0025 “Seismic Retrofitting Manual for Highway Bridges,” Edited by I.G. Buckle and I.F. Friedland, published by
the Federal Highway Administration (PB95-212676, A15, MF-A03).

NCEER-94-0026 “Proceedings from the Fifth U.S.-Japan Workshop on Earthquake Resistant Design of Lifeline Facilities and
Countermeasures Against Soil Liquefaction,” Edited by T.D. O’Rourke and M. Hamada, 11/7/94, (PB95-
220802, A99, MF-E08).

Formerly the National Center for Earthquake Engineering Research 274

NCEER-95-0001 “Experimental and Analytical Investigation of Seismic Retrofit of Structures with Supplemental Damping:
Part 1 - Fluid Viscous Damping Devices,” by A.M. Reinhorn, C. Li and M.C. Constantinou, 1/3/95, (PB95-
266599, A09, MF-A02).

NCEER-95-0002 “Experimental and Analytical Study of Low-Cycle Fatigue Behavior of Semi-Rigid Top-And-Seat Angle
Connections,” by G. Pekcan, J.B. Mander and S.S. Chen, 1/5/95, (PB95-220042, A07, MF-A02).

NCEER-95-0003 “NCEER-ATC Joint Study on Fragility of Buildings,” by T. Anagnos, C. Rojahn and A.S. Kiremidjian,
1/20/95, (PB95-220026, A06, MF-A02).

NCEER-95-0004 “Nonlinear Control Algorithms for Peak Response Reduction,” by Z. Wu, T.T. Soong, V. Gattulli and R.C.
Lin, 2/16/95, (PB95-220349, A05, MF-A01).

NCEER-95-0005 “Pipeline Replacement Feasibility Study: A Methodology for Minimizing Seismic and Corrosion Risks to
Underground Natural Gas Pipelines,” by R.T. Eguchi, H.A. Seligson and D.G. Honegger, 3/2/95, (PB95-
252326, A06, MF-A02).

NCEER-95-0006 “Evaluation of Seismic Performance of an 11-Story Frame Building During the 1994 Northridge
Earthquake,” by F. Naeim, R. DiSulio, K. Benuska, A. Reinhorn and C. Li, to be published.

NCEER-95-0007 “Prioritization of Bridges for Seismic Retrofitting,” by N. Basöz and A.S. Kiremidjian, 4/24/95, (PB95-
252300, A08, MF-A02).

NCEER-95-0008 “Method for Developing Motion Damage Relationships for Reinforced Concrete Frames,” by A. Singhal and
A.S. Kiremidjian, 5/11/95, (PB95-266607, A06, MF-A02).

NCEER-95-0009 “Experimental and Analytical Investigation of Seismic Retrofit of Structures with Supplemental Damping:
Part II - Friction Devices,” by C. Li and A.M. Reinhorn, 7/6/95, (PB96-128087, A11, MF-A03).

NCEER-95-0010 “Experimental Performance and Analytical Study of a Non-Ductile Reinforced Concrete Frame Structure
Retrofitted with Elastomeric Spring Dampers,” by G. Pekcan, J.B. Mander and S.S. Chen, 7/14/95, (PB96-
137161, A08, MF-A02).

NCEER-95-0011 “Development and Experimental Study of Semi-Active Fluid Damping Devices for Seismic Protection of
Structures,” by M.D. Symans and M.C. Constantinou, 8/3/95, (PB96-136940, A23, MF-A04).

NCEER-95-0012 “Real-Time Structural Parameter Modification (RSPM): Development of Innervated Structures,” by Z.

Liang, M. Tong and G.C. Lee, 4/11/95, (PB96-137153, A06, MF-A01).

NCEER-95-0013 “Experimental and Analytical Investigation of Seismic Retrofit of Structures with Supplemental Damping:
Part III - Viscous Damping Walls,” by A.M. Reinhorn and C. Li, 10/1/95, (PB96-176409, A11, MF-A03).

NCEER-95-0014 “Seismic Fragility Analysis of Equipment and Structures in a Memphis Electric Substation,” by J-R. Huo and
H.H.M. Hwang, (PB96-128087, A09, MF-A02), 8/10/95.

NCEER-95-0015 “The Hanshin-Awaji Earthquake of January 17, 1995: Performance of Lifelines,” Edited by M. Shinozuka,
11/3/95, (PB96-176383, A15, MF-A03).

NCEER-95-0016 “Highway Culvert Performance During Earthquakes,” by T.L. Youd and C.J. Beckman, available as
NCEER-96-0015.

NCEER-95-0017 “The Hanshin-Awaji Earthquake of January 17, 1995: Performance of Highway Bridges,” Edited by I.G.
Buckle, 12/1/95, to be published.

NCEER-95-0018 “Modeling of Masonry Infill Panels for Structural Analysis,” by A.M. Reinhorn, A. Madan, R.E. Valles, Y.
Reichmann and J.B. Mander, 12/8/95, (PB97-110886, MF-A01, A06).

NCEER-95-0019 “Optimal Polynomial Control for Linear and Nonlinear Structures,” by A.K. Agrawal and J.N. Yang,
12/11/95, (PB96-168737, A07, MF-A02).

Formerly the National Center for Earthquake Engineering Research 275

NCEER-95-0020 “Retrofit of Non-Ductile Reinforced Concrete Frames Using Friction Dampers,” by R.S. Rao, P. Gergely and
R.N. White, 12/22/95, (PB97-133508, A10, MF-A02).

NCEER-95-0021 “Parametric Results for Seismic Response of Pile-Supported Bridge Bents,” by G. Mylonakis, A. Nikolaou
and G. Gazetas, 12/22/95, (PB97-100242, A12, MF-A03).

NCEER-95-0022 “Kinematic Bending Moments in Seismically Stressed Piles,” by A. Nikolaou, G. Mylonakis and G. Gazetas,
12/23/95, (PB97-113914, MF-A03, A13).

NCEER-96-0001 “Dynamic Response of Unreinforced Masonry Buildings with Flexible Diaphragms,” by A.C. Costley and
D.P. Abrams,” 10/10/96, (PB97-133573, MF-A03, A15).

NCEER-96-0002 “State of the Art Review: Foundations and Retaining Structures,” by I. Po Lam, to be published.

NCEER-96-0003 “Ductility of Rectangular Reinforced Concrete Bridge Columns with Moderate Confinement,” by N. Wehbe,
M. Saiidi, D. Sanders and B. Douglas, 11/7/96, (PB97-133557, A06, MF-A02).

NCEER-96-0004 “Proceedings of the Long-Span Bridge Seismic Research Workshop,” edited by I.G. Buckle and I.M.
Friedland, to be published.

NCEER-96-0005 “Establish Representative Pier Types for Comprehensive Study: Eastern United States,” by J. Kulicki and Z.
Prucz, 5/28/96, (PB98-119217, A07, MF-A02).

NCEER-96-0006 “Establish Representative Pier Types for Comprehensive Study: Western United States,” by R. Imbsen, R.A.
Schamber and T.A. Osterkamp, 5/28/96, (PB98-118607, A07, MF-A02).

NCEER-96-0007 “Nonlinear Control Techniques for Dynamical Systems with Uncertain Parameters,” by R.G. Ghanem and
M.I. Bujakov, 5/27/96, (PB97-100259, A17, MF-A03).

NCEER-96-0008 “Seismic Evaluation of a 30-Year Old Non-Ductile Highway Bridge Pier and Its Retrofit,” by J.B. Mander,
B. Mahmoodzadegan, S. Bhadra and S.S. Chen, 5/31/96, (PB97-110902, MF-A03, A10).

NCEER-96-0009 “Seismic Performance of a Model Reinforced Concrete Bridge Pier Before and After Retrofit,” by J.B.
Mander, J.H. Kim and C.A. Ligozio, 5/31/96, (PB97-110910, MF-A02, A10).

NCEER-96-0010 “IDARC2D Version 4.0: A Computer Program for the Inelastic Damage Analysis of Buildings,” by R.E.
Valles, A.M. Reinhorn, S.K. Kunnath, C. Li and A. Madan, 6/3/96, (PB97-100234, A17, MF-A03).

NCEER-96-0011 “Estimation of the Economic Impact of Multiple Lifeline Disruption: Memphis Light, Gas and Water
Division Case Study,” by S.E. Chang, H.A. Seligson and R.T. Eguchi, 8/16/96, (PB97-133490, A11, MF-
A03).

NCEER-96-0012 “Proceedings from the Sixth Japan-U.S. Workshop on Earthquake Resistant Design of Lifeline Facilities and
Countermeasures Against Soil Liquefaction, Edited by M. Hamada and T. O’Rourke, 9/11/96, (PB97-
133581, A99, MF-A06).

NCEER-96-0013 “Chemical Hazards, Mitigation and Preparedness in Areas of High Seismic Risk: A Methodology for
Estimating the Risk of Post-Earthquake Hazardous Materials Release,” by H.A. Seligson, R.T. Eguchi, K.J.
Tierney and K. Richmond, 11/7/96, (PB97-133565, MF-A02, A08).

NCEER-96-0014 “Response of Steel Bridge Bearings to Reversed Cyclic Loading,” by J.B. Mander, D-K. Kim, S.S. Chen and
G.J. Premus, 11/13/96, (PB97-140735, A12, MF-A03).

NCEER-96-0015 “Highway Culvert Performance During Past Earthquakes,” by T.L. Youd and C.J. Beckman, 11/25/96,
(PB97-133532, A06, MF-A01).

NCEER-97-0001 “Evaluation, Prevention and Mitigation of Pounding Effects in Building Structures,” by R.E. Valles and
A.M. Reinhorn, 2/20/97, (PB97-159552, A14, MF-A03).

NCEER-97-0002 “Seismic Design Criteria for Bridges and Other Highway Structures,” by C. Rojahn, R. Mayes, D.G.
Anderson, J. Clark, J.H. Hom, R.V. Nutt and M.J. O’Rourke, 4/30/97, (PB97-194658, A06, MF-A03).

Formerly the National Center for Earthquake Engineering Research 276

NCEER-97-0003 “Proceedings of the U.S.-Italian Workshop on Seismic Evaluation and Retrofit,” Edited by D.P. Abrams and
G.M. Calvi, 3/19/97, (PB97-194666, A13, MF-A03).

NCEER-97-0004 "Investigation of Seismic Response of Buildings with Linear and Nonlinear Fluid Viscous Dampers," by
A.A. Seleemah and M.C. Constantinou, 5/21/97, (PB98-109002, A15, MF-A03).

NCEER-97-0005 "Proceedings of the Workshop on Earthquake Engineering Frontiers in Transportation Facilities," edited by
G.C. Lee and I.M. Friedland, 8/29/97, (PB98-128911, A25, MR-A04).

NCEER-97-0006 "Cumulative Seismic Damage of Reinforced Concrete Bridge Piers," by S.K. Kunnath, A. El-Bahy, A.
Taylor and W. Stone, 9/2/97, (PB98-108814, A11, MF-A03).

NCEER-97-0007 "Structural Details to Accommodate Seismic Movements of Highway Bridges and Retaining Walls," by R.A.
Imbsen, R.A. Schamber, E. Thorkildsen, A. Kartoum, B.T. Martin, T.N. Rosser and J.M. Kulicki, 9/3/97,
(PB98-108996, A09, MF-A02).

NCEER-97-0008 "A Method for Earthquake Motion-Damage Relationships with Application to Reinforced Concrete Frames,"
by A. Singhal and A.S. Kiremidjian, 9/10/97, (PB98-108988, A13, MF-A03).

NCEER-97-0009 "Seismic Analysis and Design of Bridge Abutments Considering Sliding and Rotation," by K. Fishman and
R. Richards, Jr., 9/15/97, (PB98-108897, A06, MF-A02).

NCEER-97-0010 "Proceedings of the FHWA/NCEER Workshop on the National Representation of Seismic Ground Motion
for New and Existing Highway Facilities," edited by I.M. Friedland, M.S. Power and R.L. Mayes, 9/22/97,
(PB98-128903, A21, MF-A04).

NCEER-97-0011 "Seismic Analysis for Design or Retrofit of Gravity Bridge Abutments," by K.L. Fishman, R. Richards, Jr.
and R.C. Divito, 10/2/97, (PB98-128937, A08, MF-A02).

NCEER-97-0012 "Evaluation of Simplified Methods of Analysis for Yielding Structures," by P. Tsopelas, M.C. Constantinou,
C.A. Kircher and A.S. Whittaker, 10/31/97, (PB98-128929, A10, MF-A03).

NCEER-97-0013 "Seismic Design of Bridge Columns Based on Control and Repairability of Damage," by C-T. Cheng and
J.B. Mander, 12/8/97, (PB98-144249, A11, MF-A03).

NCEER-97-0014 "Seismic Resistance of Bridge Piers Based on Damage Avoidance Design," by J.B. Mander and C-T. Cheng,
12/10/97, (PB98-144223, A09, MF-A02).

NCEER-97-0015 “Seismic Response of Nominally Symmetric Systems with Strength Uncertainty,” by S. Balopoulou and M.
Grigoriu, 12/23/97, (PB98-153422, A11, MF-A03).

NCEER-97-0016 “Evaluation of Seismic Retrofit Methods for Reinforced Concrete Bridge Columns,” by T.J. Wipf, F.W.
Klaiber and F.M. Russo, 12/28/97, (PB98-144215, A12, MF-A03).

NCEER-97-0017 “Seismic Fragility of Existing Conventional Reinforced Concrete Highway Bridges,” by C.L. Mullen and
A.S. Cakmak, 12/30/97, (PB98-153406, A08, MF-A02).

NCEER-97-0018 “Loss Asssessment of Memphis Buildings,” edited by D.P. Abrams and M. Shinozuka, 12/31/97, (PB98-
144231, A13, MF-A03).

NCEER-97-0019 “Seismic Evaluation of Frames with Infill Walls Using Quasi-static Experiments,” by K.M. Mosalam, R.N.
White and P. Gergely, 12/31/97, (PB98-153455, A07, MF-A02).

NCEER-97-0020 “Seismic Evaluation of Frames with Infill Walls Using Pseudo-dynamic Experiments,” by K.M. Mosalam,
R.N. White and P. Gergely, 12/31/97, (PB98-153430, A07, MF-A02).

NCEER-97-0021 “Computational Strategies for Frames with Infill Walls: Discrete and Smeared Crack Analyses and Seismic
Fragility,” by K.M. Mosalam, R.N. White and P. Gergely, 12/31/97, (PB98-153414, A10, MF-A02).

Formerly the National Center for Earthquake Engineering Research 277

NCEER-97-0022 “Proceedings of the NCEER Workshop on Evaluation of Liquefaction Resistance of Soils,” edited by T.L.
Youd and I.M. Idriss, 12/31/97, (PB98-155617, A15, MF-A03).

MCEER-98-0001 “Extraction of Nonlinear Hysteretic Properties of Seismically Isolated Bridges from Quick-Release Field
Tests,” by Q. Chen, B.M. Douglas, E.M. Maragakis and I.G. Buckle, 5/26/98, (PB99-118838, A06, MF-
A01).

MCEER-98-0002 “Methodologies for Evaluating the Importance of Highway Bridges,” by A. Thomas, S. Eshenaur and J.
Kulicki, 5/29/98, (PB99-118846, A10, MF-A02).

MCEER-98-0003 “Capacity Design of Bridge Piers and the Analysis of Overstrength,” by J.B. Mander, A. Dutta and P. Goel,
6/1/98, (PB99-118853, A09, MF-A02).

MCEER-98-0004 “Evaluation of Bridge Damage Data from the Loma Prieta and Northridge, California Earthquakes,” by N.
Basoz and A. Kiremidjian, 6/2/98, (PB99-118861, A15, MF-A03).

MCEER-98-0005 “Screening Guide for Rapid Assessment of Liquefaction Hazard at Highway Bridge Sites,” by T. L. Youd,
6/16/98, (PB99-118879, A06, not available on microfiche).

MCEER-98-0006 “Structural Steel and Steel/Concrete Interface Details for Bridges,” by P. Ritchie, N. Kauhl and J. Kulicki,
7/13/98, (PB99-118945, A06, MF-A01).

MCEER-98-0007 “Capacity Design and Fatigue Analysis of Confined Concrete Columns,” by A. Dutta and J.B. Mander,
7/14/98, (PB99-118960, A14, MF-A03).

MCEER-98-0008 “Proceedings of the Workshop on Performance Criteria for Telecommunication Services Under Earthquake
Conditions,” edited by A.J. Schiff, 7/15/98, (PB99-118952, A08, MF-A02).

MCEER-98-0009 “Fatigue Analysis of Unconfined Concrete Columns,” by J.B. Mander, A. Dutta and J.H. Kim, 9/12/98,
(PB99-123655, A10, MF-A02).

MCEER-98-0010 “Centrifuge Modeling of Cyclic Lateral Response of Pile-Cap Systems and Seat-Type Abutments in Dry
Sands,” by A.D. Gadre and R. Dobry, 10/2/98, (PB99-123606, A13, MF-A03).

MCEER-98-0011 “IDARC-BRIDGE: A Computational Platform for Seismic Damage Assessment of Bridge Structures,” by
A.M. Reinhorn, V. Simeonov, G. Mylonakis and Y. Reichman, 10/2/98, (PB99-162919, A15, MF-A03).

MCEER-98-0012 “Experimental Investigation of the Dynamic Response of Two Bridges Before and After Retrofitting with
Elastomeric Bearings,” by D.A. Wendichansky, S.S. Chen and J.B. Mander, 10/2/98, (PB99-162927, A15,
MF-A03).

MCEER-98-0013 “Design Procedures for Hinge Restrainers and Hinge Sear Width for Multiple-Frame Bridges,” by R. Des
Roches and G.L. Fenves, 11/3/98, (PB99-140477, A13, MF-A03).

MCEER-98-0014 “Response Modification Factors for Seismically Isolated Bridges,” by M.C. Constantinou and J.K. Quarshie,
11/3/98, (PB99-140485, A14, MF-A03).

MCEER-98-0015 “Proceedings of the U.S.-Italy Workshop on Seismic Protective Systems for Bridges,” edited by I.M. Friedland
and M.C. Constantinou, 11/3/98, (PB2000-101711, A22, MF-A04).

MCEER-98-0016 “Appropriate Seismic Reliability for Critical Equipment Systems: Recommendations Based on Regional
Analysis of Financial and Life Loss,” by K. Porter, C. Scawthorn, C. Taylor and N. Blais, 11/10/98, (PB99-
157265, A08, MF-A02).

MCEER-98-0017 “Proceedings of the U.S. Japan Joint Seminar on Civil Infrastructure Systems Research,” edited by M.
Shinozuka and A. Rose, 11/12/98, (PB99-156713, A16, MF-A03).

MCEER-98-0018 “Modeling of Pile Footings and Drilled Shafts for Seismic Design,” by I. PoLam, M. Kapuskar and D.
Chaudhuri, 12/21/98, (PB99-157257, A09, MF-A02).

Formerly the National Center for Earthquake Engineering Research 278

MCEER-99-0001 "Seismic Evaluation of a Masonry Infilled Reinforced Concrete Frame by Pseudodynamic Testing," by S.G.
Buonopane and R.N. White, 2/16/99, (PB99-162851, A09, MF-A02).

MCEER-99-0002 "Response History Analysis of Structures with Seismic Isolation and Energy Dissipation Systems:
Verification Examples for Program SAP2000," by J. Scheller and M.C. Constantinou, 2/22/99, (PB99-
162869, A08, MF-A02).

MCEER-99-0003 "Experimental Study on the Seismic Design and Retrofit of Bridge Columns Including Axial Load Effects,"
by A. Dutta, T. Kokorina and J.B. Mander, 2/22/99, (PB99-162877, A09, MF-A02).

MCEER-99-0004 "Experimental Study of Bridge Elastomeric and Other Isolation and Energy Dissipation Systems with
Emphasis on Uplift Prevention and High Velocity Near-source Seismic Excitation," by A. Kasalanati and M.
C. Constantinou, 2/26/99, (PB99-162885, A12, MF-A03).

MCEER-99-0005 "Truss Modeling of Reinforced Concrete Shear-flexure Behavior," by J.H. Kim and J.B. Mander, 3/8/99,
(PB99-163693, A12, MF-A03).

MCEER-99-0006 "Experimental Investigation and Computational Modeling of Seismic Response of a 1:4 Scale Model Steel
Structure with a Load Balancing Supplemental Damping System," by G. Pekcan, J.B. Mander and S.S. Chen,
4/2/99, (PB99-162893, A11, MF-A03).

MCEER-99-0007 "Effect of Vertical Ground Motions on the Structural Response of Highway Bridges," by M.R. Button, C.J.
Cronin and R.L. Mayes, 4/10/99, (PB2000-101411, A10, MF-A03).

MCEER-99-0008 "Seismic Reliability Assessment of Critical Facilities: A Handbook, Supporting Documentation, and Model
Code Provisions," by G.S. Johnson, R.E. Sheppard, M.D. Quilici, S.J. Eder and C.R. Scawthorn, 4/12/99,
(PB2000-101701, A18, MF-A04).

MCEER-99-0009 "Impact Assessment of Selected MCEER Highway Project Research on the Seismic Design of Highway
Structures," by C. Rojahn, R. Mayes, D.G. Anderson, J.H. Clark, D'Appolonia Engineering, S. Gloyd and
R.V. Nutt, 4/14/99, (PB99-162901, A10, MF-A02).

MCEER-99-0010 "Site Factors and Site Categories in Seismic Codes," by R. Dobry, R. Ramos and M.S. Power, 7/19/99,
(PB2000-101705, A08, MF-A02).

MCEER-99-0011 "Restrainer Design Procedures for Multi-Span Simply-Supported Bridges," by M.J. Randall, M. Saiidi, E.
Maragakis and T. Isakovic, 7/20/99, (PB2000-101702, A10, MF-A02).

MCEER-99-0012 "Property Modification Factors for Seismic Isolation Bearings," by M.C. Constantinou, P. Tsopelas, A.
Kasalanati and E. Wolff, 7/20/99, (PB2000-103387, A11, MF-A03).

MCEER-99-0013 "Critical Seismic Issues for Existing Steel Bridges," by P. Ritchie, N. Kauhl and J. Kulicki, 7/20/99,
(PB2000-101697, A09, MF-A02).

MCEER-99-0014 "Nonstructural Damage Database," by A. Kao, T.T. Soong and A. Vender, 7/24/99, (PB2000-101407, A06,
MF-A01).

MCEER-99-0015 "Guide to Remedial Measures for Liquefaction Mitigation at Existing Highway Bridge Sites," by H.G.
Cooke and J. K. Mitchell, 7/26/99, (PB2000-101703, A11, MF-A03).

MCEER-99-0016 "Proceedings of the MCEER Workshop on Ground Motion Methodologies for the Eastern United States,"
edited by N. Abrahamson and A. Becker, 8/11/99, (PB2000-103385, A07, MF-A02).

MCEER-99-0017 "Quindío, Colombia Earthquake of January 25, 1999: Reconnaissance Report," by A.P. Asfura and P.J.
Flores, 10/4/99, (PB2000-106893, A06, MF-A01).

MCEER-99-0018 "Hysteretic Models for Cyclic Behavior of Deteriorating Inelastic Structures," by M.V. Sivaselvan and A.M.
Reinhorn, 11/5/99, (PB2000-103386, A08, MF-A02).

Formerly the National Center for Earthquake Engineering Research 279

MCEER-99-0019 "Proceedings of the 7th U.S.- Japan Workshop on Earthquake Resistant Design of Lifeline Facilities and
Countermeasures Against Soil Liquefaction," edited by T.D. O'Rourke, J.P. Bardet and M. Hamada,
11/19/99, (PB2000-103354, A99, MF-A06).

MCEER-99-0020 "Development of Measurement Capability for Micro-Vibration Evaluations with Application to Chip
Fabrication Facilities," by G.C. Lee, Z. Liang, J.W. Song, J.D. Shen and W.C. Liu, 12/1/99, (PB2000-
105993, A08, MF-A02).

MCEER-99-0021 "Design and Retrofit Methodology for Building Structures with Supplemental Energy Dissipating Systems,"
by G. Pekcan, J.B. Mander and S.S. Chen, 12/31/99, (PB2000-105994, A11, MF-A03).

MCEER-00-0001 "The Marmara, Turkey Earthquake of August 17, 1999: Reconnaissance Report," edited by C. Scawthorn;
with major contributions by M. Bruneau, R. Eguchi, T. Holzer, G. Johnson, J. Mander, J. Mitchell, W.
Mitchell, A. Papageorgiou, C. Scaethorn, and G. Webb, 3/23/00, (PB2000-106200, A11, MF-A03).

MCEER-00-0002 "Proceedings of the MCEER Workshop for Seismic Hazard Mitigation of Health Care Facilities," edited by
G.C. Lee, M. Ettouney, M. Grigoriu, J. Hauer and J. Nigg, 3/29/00, (PB2000-106892, A08, MF-A02).

MCEER-00-0003 "The Chi-Chi, Taiwan Earthquake of September 21, 1999: Reconnaissance Report," edited by G.C. Lee and
C.H. Loh, with major contributions by G.C. Lee, M. Bruneau, I.G. Buckle, S.E. Chang, P.J. Flores, T.D.
O'Rourke, M. Shinozuka, T.T. Soong, C-H. Loh, K-C. Chang, Z-J. Chen, J-S. Hwang, M-L. Lin, G-Y. Liu,
K-C. Tsai, G.C. Yao and C-L. Yen, 4/30/00.

MCEER-00-0004 "Seismic Retrofit of End-Sway Frames of Steel Deck-Truss Bridges with a Supplemental Tendon System:
Experimental and Analytical Investigation," by G. Pekcan, J.B. Mander and S.S. Chen, 7/1/00.

MCEER-00-0005 "Sliding Fragility of Unrestrained Equipment in Critical Facilities," by W.H. Chong and T.T. Soong, 7/5/00.

MCEER-00-0006 "Seismic Response of Reinforced Concrete Bridge Pier Walls in the Weak Direction," by N. Abo-Shadi, M.
Saiidi and D. Sanders, 7/17/00.

MCEER-00-0007 "Low-Cycle Fatigue Behavior of Longitudinal Reinforcement in Reinforced Concrete Bridge Columns," by
J. Brown and S.K. Kunnath, 7/23/00.

MCEER-00-0008 "Soil Structure Interaction of Bridges for Seismic Analysis," I. PoLam and H. Law, 9/25/00.

MCEER-00-0009 "Proceedings of the First MCEER Workshop on Mitigation of Earthquake Disaster by Advanced
Technologies (MEDAT-1), edited by M. Shinozuka, D.J. Inman and T.D. O'Rourke, 11/10/00.

MCEER-00-0010 "Development and Evaluation of Simplified Procedures for Analysis and Design of Buildings with Passive
Energy Dissipation Systems," by O.M. Ramirez, M.C. Constantinou, C.A. Kircher, A.S. Whittaker, M.W.
Johnson, J.D. Gomez and C. Chrysostomou, 11/16/01.

MCEER-00-0011 "Dynamic Soil-Foundation-Structure Interaction Analyses of Large Caissons," by C-Y. Chang, C-M. Mok,
Z-L. Wang, R. Settgast, F. Waggoner, M.A. Ketchum, H.M. Gonnermann and C-C. Chin, 12/30/00.

MCEER-00-0012 "Experimental Evaluation of Seismic Performance of Bridge Restrainers," by A.G. Vlassis, E.M. Maragakis
and M. Saiid Saiidi, 12/30/00.

MCEER-00-0013 "Effect of Spatial Variation of Ground Motion on Highway Structures," by M. Shinozuka, V. Saxena and G.
Deodatis, 12/31/00.

MCEER-00-0014 "A Risk-Based Methodology for Assessing the Seismic Performance of Highway Systems," by S.D. Werner,
C.E. Taylor, J.E. Moore, II, J.S. Walton and S. Cho, 12/31/00.

MCEER-01-0001 “Experimental Investigation of P-Delta Effects to Collapse During Earthquakes,” by D. Vian and M.
Bruneau, 6/25/01.

MCEER-01-0002 “Proceedings of the Second MCEER Workshop on Mitigation of Earthquake Disaster by Advanced
Technologies (MEDAT-2),” edited by M. Bruneau and D.J. Inman, 7/23/01.

Formerly the National Center for Earthquake Engineering Research 280

MCEER-01-0003 “Sensitivity Analysis of Dynamic Systems Subjected to Seismic Loads,” by C. Roth and M. Grigoriu,
9/18/01.

MCEER-01-0004 “Overcoming Obstacles to Implementing Earthquake Hazard Mitigation Policies: Stage 1 Report,” by D.J.
Alesch and W.J. Petak, 12/17/01.

MCEER-01-0005 “Updating Real-Time Earthquake Loss Estimates: Methods, Problems and Insights,” by C.E. Taylor, S.E.
Chang and R.T. Eguchi, 12/17/01.

MCEER-01-0006 “Experimental Investigation and Retrofit of Steel Pile Foundations and Pile Bents Under Cyclic Lateral
Loadings,” by A. Shama, J. Mander, B. Blabac and S. Chen, 12/31/01.

MCEER-02-0001 “Assessment of Performance of Bolu Viaduct in the 1999 Duzce Earthquake in Turkey” by P.C. Roussis,
M.C. Constantinou, M. Erdik, E. Durukal and M. Dicleli, 5/8/02.

MCEER-02-0002 “Seismic Behavior of Rail Counterweight Systems of Elevators in Buildings,” by M.P. Singh, Rildova and
L.E. Suarez, 5/27/02.

MCEER-02-0003 “Development of Analysis and Design Procedures for Spread Footings,” by G. Mylonakis, G. Gazetas, S.
Nikolaou and A. Chauncey, 10/02/02.

MCEER-02-0004 “Bare-Earth Algorithms for Use with SAR and LIDAR Digital Elevation Models,” by C.K. Huyck, R.T.
Eguchi and B. Houshmand, 10/16/02.

MCEER-02-0005 “Review of Energy Dissipation of Compression Members in Concentrically Braced Frames,” by K.Lee and
M. Bruneau, 10/18/02.

MCEER-03-0001 “Experimental Investigation of Light-Gauge Steel Plate Shear Walls for the Seismic Retrofit of Buildings”
by J. Berman and M. Bruneau, 5/2/03.

MCEER-03-0002 “Statistical Analysis of Fragility Curves,” by M. Shinozuka, M.Q. Feng, H. Kim, T. Uzawa and T. Ueda,
6/16/03.

MCEER-03-0003 “Proceedings of the Eighth U.S.-Japan Workshop on Earthquake Resistant Design f Lifeline Facilities and
Countermeasures Against Liquefaction,” edited by M. Hamada, J.P. Bardet and T.D. O’Rourke, 6/30/03.

MCEER-03-0004 “Proceedings of the PRC-US Workshop on Seismic Analysis and Design of Special Bridges,” edited by L.C.
Fan and G.C. Lee, 7/15/03.

Formerly the National Center for Earthquake Engineering Research 281

University at Buffalo, State University of New York
Red Jacket Quadrangle ■ Buffalo, New York 14261
Phone: (716) 645-3391 ■ Fax: (716) 645-3399
E-mail: [email protected] ■ WWW Site http://mceer.buffalo.edu

University at Buffalo The State University of New York