Zhou et al.

/ J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418 405

Journal of Zhejiang University-SCIENCE A (Applied Physics & Engineering)

ISSN 1673-565X (Print); ISSN 1862-1775 (Online)
www.zju.edu.cn/jzus; www.springerlink.com
E-mail: [email protected]

Seismic response study on a multi-span cable-stayed bridge scale

model under multi-support excitations. Part II: numerical analysis*

Rui ZHOU†1, Zhou-hong ZONG†‡2, Xue-yang HUANG3, Zhang-hua XIA3

(1State Key Lab for Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, China)
(2School of Civil Engineering, Southeast University, Nanjing 210096, China)
(3School of Civil Engineering, Fuzhou University, Fuzhou 350108, China)
†
E-mail: [email protected]; [email protected]
Received Oct. 16, 2013; Revision accepted Apr. 22, 2014; Crosschecked May 21, 2014

Abstract: In recent years, local component destruction, collapse failure and the control of long-span cable-stayed bridges under
strong multi-support excitations have received increasing attention. In this paper, two kinds of nonlinear finite element (FE)
models are established to simulate the seismic responses and failure modes of a multi-span cable-stayed bridge scale model under
multi-support excitations. One is the single girder model which is used to simulate the seismic response during four wave excita-
tions. It can be concluded that the FE analysis results of the scale model are a good fit with those from the shaking table tests. The
other one is the explicit dynamic FE model which is used to simulate the collapse and failure mechanisms of the scale model
during strong earthquakes. The aggressive failure processes of the scale model under two different types of wave excitations were
compared to reproduce the mechanisms in which the bearing at the middle tower failed under the El Centro (EC) wave (4.0 m/s2),
and we also observed the connection failure of the cables and towers, and the elements failure of the upper beam in the middle
tower, which occurred successively under a Jiangxin (JX) wave (4.0 m/s2). This simulation may be referenced as the basis for the
collapse failure of a cable-stayed bridge with a large span during a strong earthquake.

Key words: Multi-span cable-stayed bridge, Multi-support excitations, Nonlinear finite element model, Nonlinear dynamic
analysis, Collapse failure, Failure mode
doi:10.1631/jzus.A1300340 Document code: A CLC number: U448.27

1 Introduction site conditions should be considered. For example,

Wilson and Gravelle (1991), Ren et al. (2005), and
Over the past 20 years, rapid development has Pantaleon et al. (2010) studied the dynamic behaviors
been made on modern cable-stayed bridges (Gimsing of cable-stayed bridges through 3D finite-element
and Georgakis, 2011). Many researchers have studied (FE) models. In these studies, the modal behaviors
the dynamic behavior during ambient vibration and predicted by the FE models were compared to the
seismic response under multi-support excitations on measured characteristics from ambient vibrations of
long-span cable-stayed bridges, where the influences full-scale cable-stayed bridges. On the other hand, 3D
of the traveling wave effect, coherence loss, and local nonlinear seismic response analyses of long-span
cable-stayed bridges have been conducted (Abdel-
‡
Ghaffar and Nazmy, 1991; Nazmy and Abdel-
Corresponding author
*
Project supported by the National Natural Science Foundation of Ghaffar, 1992; Zavoni and Vanmarcke, 1994; Allam
China (Nos. 51178101 and 51378112), the Doctoral Program Project and Datta, 1999; Ren and Makoto, 1999). In China,
of the Ministry of Education (No. 20110092110011), and the National Fan et al. (2001), Li et al. (2005), Bao et al. (2008), as
Key Laboratory of Civil Engineering Disaster Prevention of Tongji
University (No. SLDRCE08-HZ-02), China well as Lin et al. (2009) studied the linear and
© Zhejiang University and Springer-Verlag Berlin Heidelberg 2014 nonlinear response characteristics of long-span
406 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418

cable-stayed bridges during uniform excitations and during strong earthquakes have seldom been done, let
non-uniform excitations. These investigations con- alone analyzing the impact on multi-span cable-
centrated on two-tower cable-stayed bridges. stayed bridges.
A few researchers were involved in theoretical This paper focuses on the numerical simulation
and numerical simulation to understand the seismic of a seismic response for a multi-span cable-stayed
response mechanisms of multi-span cable-stayed bridge scale model under uniform excitations and
bridges subjected to multi-support excitations. Fang non-uniform excitations.
et al. (2011) explored the influence of traveling-wave
effects on the seismic responses of a multi-span 2 Scaled multi-span cable-stayed bridge
cable-stayed bridge with different constraint mecha-
nisms between the main towers and main girder. The Wuhan Erqi Yangtze River Bridge, a
Okamoto and Nakamura (2011) studied seismic re- three-tower and double-cable-plane cable-stayed
sponses of a multi-span cable-stayed bridge with bridge, was selected as the prototype bridge for a
steel/concrete hybrid towers subjected to medium and multi-span cable-stayed bridge. A 1:100 scale model
ultra-strong seismic waves specified in the “Japanese with a 17.32 m length for the main girder and a 2.05 m
Seismic Code for Highway Bridges”. This study height for the main towers was tested by using
revealed that a new steel/concrete hybrid tower is earthquake shaking tables (Fig. 1). The analytical
feasible for multi-span cable-stayed bridges and is modeling discussed in this study simulates the seis-
most effective for seismic forces when the main mic response of the scale model during multi-support
girder is connected with bi-linear springs. excitations up to failure. For a detailed description of
Meanwhile, the collapse analysis of cable-stayed the prototype bridge and scale model, please refer to
bridges under strong earthquakes has gradually the relevant study of ‘Part I: shaking table tests’
gained attention after the recent experience of many (Zong et al., 2014). Fig. 2 shows the scale model test
strong earthquakes. For example, Self and Dilger setup.
(1990) described the behavior of planar prestressed
concrete cable-stayed bridges until failure occurred,
taking into account the geometric and material 3 Finite element modeling of the scale model
nonlinearities under static loads. Song and Kim
3.1 Single girder model
(2007) discussed the analysis problem of the in-plane
overall collapse mechanisms during static loads, and A 3D single girder FE model of the scale model
the evaluation of the ultimate load-carrying capacities was established based on ANSYS software (Nazmy
of cable-stayed bridges with various cable layouts and Abdel-Ghaffar, 1990a; 1990b; Wilson and
using the bifurcation point instability approach. Gravelle, 1991). Spatial beam-column elements
However, they did not study the structural behavior (BEAM4) with six degrees of freedom at each node
under dynamic loads. Luo et al. (2010) compared the were adopted to model the main girder, cross beams,
collapse process of the concrete girder bridge (Baihua towers, and side piers. The truss elements (LINK10),
Bridge, China) during the Wenchuan earthquake having only tension stiffness, were used to model the
between a simulation analysis and a practical situa- cables. The main girder was modeled by using a sin-
tion, which shows that the local components of the gle central spine with offset links to accommodate
connecting devices were destroyed first. Yoo et al. lumped masses and cable anchor points, and stiffness
(2012) proposed a new and simple method for esti- was assigned to the spine to simulate the actual
mating the collapse load of a steel cable-stayed bridge stiffness of the main girder. The nonlinearity of the
under static loads. The results demonstrated that the cables was accounted for by an equivalent tangent
proposed method is a good substitute for a complex equation (Ernst, 1965; Nazmy and Abdel-Ghaffar,
nonlinear inelastic analysis to approximately evaluate 1990a; 1990b). The supplemental masses were mod-
the collapse loads as well as the failure modes of steel eled by using the mass elements (MASS21) defined
cable-stayed bridges. Note that the collapse analysis by a single node. The vertical bearings among the
and failure control of long-span cable-stayed bridges towers, side piers, and superstructure were modeled
 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418 407

(a)

(b) (c)

Fig. 1 Three-tower cable-stayed bridge scale model (unit: cm): (a) longitudinal arrangement of the scale model;
(b) dimension of the middle tower; (c) dimension of the side tower

by using spring-damper elements (COMBIN14) de-

fined by two nodes of a spring constant and damping
coefficients, and the other connections among the
towers, side piers, and superstructure were simulated
by coupling the corresponding rotational and lateral
degrees of freedom. The bottom of the towers and the
side piers were completely fixed, ignoring the influ-
ence of the soil-pile-structure interactions. The single
girder model of the scale model has 706 nodes and
Fig. 2 Test setup of the scale model 1600 elements (Fig. 3).
408 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418

The first 15 natural modal frequencies and vi- 3.2 Explicit dynamic model
bration modes of the scale model were calculated, as
In recent years, dynamic disasters, collapse
well as the partial measured frequencies from the
failures and their effect on long-span cable-stayed
shaking table tests were compared (Table 1). It can be
bridges under complex service environments, such as
seen that the measured frequencies of the scale model
earthquakes, typhoons explosions, and collisions,
are consistent with those of the numerical calculation
have been receiving increasing attention with the
by the single girder model. These results indicate that
rapid development of cable-stayed bridges around the
the single girder model is accurate enough for stud-
world. Because the collapsing process of a cable-
ying the dynamic characteristics and the seismic re-
stayed bridge is noticeably discontinuous, large de-
sponses of a long-span cable-stayed bridge, and for
formation, and highly nonlinear, traditional implicit
the requirements of related dynamic researches on
FE methods are hard to simulate the collapse and
full structures. However, the implicit FE model can-
failure mechanisms of the cable-stayed bridge during
not directly simulate the whole failure process of
a strong earthquake. On the other hand, explicit dy-
long-span cable-stayed bridges to demonstrate the
namic algorithms are using a small time step to meet
mechanisms of collapse and failure of the complex
the convergence conditions and save the calculation
structures during a strong earthquake.
costs by avoiding the stiffness matrix through inverse
in the implicit integration algorithm. In this study,
LS-DYNA, as one of the prominent explicit dynamic
programs, was adopted to simulate the collapse and
failure process of the multi-span cable-stayed bridge
scale model during multi-support excitations (LSTC,
2006). The main girder, main towers, and central
buckles were modeled by using spatial beam elements
BEAM161 which included finite transverse shear
strains. The cables were modeled by using elements
LINK167 which allowed elastic cables to be realis-
tically modeled. The vertical bearing among the
towers, side piers, and the main girder were modeled
Fig. 3 Spatial finite element model of the scale model by using elements COMBI165 which allowed us

Table 1 Summaries of first 15 modes of the bridge model

Order Testing Simulation
Characteristics of modes
number frequency (Hz) frequency (Hz)
1 – 1.832 Longitudinal floating vibration of the deck
2 – 3.039 Antisymmetric lateral bending vibration of the side tower
3 – 3.545 Symmetric lateral bending vibration of the side tower
4 – 5.808 Lateral bending vibration of the middle tower
5 6.087 6.048 Antisymmetric lateral bending vibration of the deck
6 8.272 8.255 Symmetric lateral bending vibration of the deck
7 9.023 9.066 Antisymmetric vertical bending vibration of the deck
8 11.507 11.361 Symmetric vertical bending vibration of the deck
9 – 11.391 Antisymmetric torsion vibration of the side tower
10 – 12.528 Symmetric torsion vibration of the side tower
11 – 13.852 Torsion vibration of the middle tower
12 – 17.088 The second antisymmetric lateral bending vibration of the deck
13 – 17.647 The second symmetric lateral bending vibration of the deck
14 – 18.161 Antisymmetric lateral bending vibration accompanied by torsion
15 – 18.261 Symmetric lateral bending vibration accompanied by torsion
 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418 409

to model simple spring or damper systems, and the 4 Single girder finite element model
attached masses were modeled by using MASS166 verification
elements which are defined by a single node with
concentrated mass components. Based on the above single girder FE model, the
To preferably reflect the material property, an time-history analysis of the scale model during uni-
elastic-plastic model with kinematic hardening form excitations and non-uniform excitations was
(MAT_PLASTIC_KINEMATIC3) was used to sim- conducted by the Newmark-β method.
ulate the polymethyl methacrylate (PMMA). The Three synthetic seismic waves with the length
yielding stress and the failure strain of the PMMA of time of 20.5 s including the Wuchang (WC) wave,
were 5.5 MPa and 0.0325ε in the failure criteria def- Jiangxin (JX) wave, Hankou (HK) wave, and El
inition of the material, respectively. The modulus of Centro (EC) wave were selected as the input excita-
tions described in the relevant study of ‘Part I:
elasticity and the mass density of the PMMA were
shaking table tests’ (Zong et al., 2014) (Fig. 5,
2690 MPa and 1180 kg/m3, respectively. The discrete
p.411). Since only the acceleration and strain pa-
beam element model (MAT_CABLE_DISCRETE_
rameters were measured in the shaking table tests, the
BEAM) was used to model the high-tensile steel ca-
acceleration responses of the towers in the test and
bles, and the offset was set in the real constants of the the FE analysis (FEA) were compared (Table 2).
cable elements. The modulus of elasticity and the Table 2 indicates that the test results are in good
mass density of the cables were 195 000 MPa and agreement with the results from FEA, and most of the
7850 kg/m3, respectively. The rigid arm of the main errors are less than 10% on the whole. The numerical
girder was simply modeled by the rigid body model simulation results are a little higher than those from
through constraining the freedom degrees of the rigid the tests due to the simplified disposal of the deck in
body through the DEMP command, thus the compu- establishing the FE model and the errors of the
ting time was greatly reduced. The spring element equipment in the practical experimental measure-
model (MAT_SPRING_ELASTOPLASTIC) was ment. The seismic responses of three towers are ap-
adopted to model the vertical bearings among the proximately symmetric under horizontal earthquake
towers, side piers, and main girder. Through the excitations.
TBDATA command, the elastic stiffness and shear The time-history curves at the middle span of
stiffness were set as 1.5 MPa and 5.5 MPa, respec- the main girder are described in Fig. 6 (p.412). It can
tively. The explicit dynamic relaxation method was be seen that the response tendencies of the FE
adopted to consider the action of gravity, and the analysis are consistent with those from the shaking
single automatic contact was used to simulate the table tests (Zong et al., 2014). The seismic responses
dynamic contact problem in the process of the col- in the vertical direction are larger than those in the
lapse. The explicit dynamic model of the three-tower transversal direction. From these comparisons, it can
cable-stayed bridge scale model has 706 nodes and be concluded that the single girder FE model can be
1600 elements (Fig. 4). used to predict the seismic responses of the multi-
span cable-stayed bridge during uniform excitations
and non-uniform excitations.

5 Seismic response analysis of the single

girder model
5.1 Displacement responses
The longitudinal displacement at the top of the
tower, the displacement at the end of the main girder,
the relative displacement between the tower and the
main girder, and the vertical displacement of the
mid-span of the main span in the scale model were
Fig. 4 Explicit dynamic model of the scale model calculated under uniform excitations. The peak value
410 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418

Table 2 Comparison of acceleration responses of the towers between test and FEA (unit: m/s2)
EC wave of 1.0 m/s2 (Case 1) WC wave of 1.0 m/s2 (Case 2)
Location Longitudinal Transverse Longitudinal Transverse
Test FEA Error Test FEA Error Test FEA Error Test FEA Error
L1 4.050 4.285 5.49% 2.957 3.044 2.87% 5.515 5.631 2.06% 4.345 4.416 1.60%
L2 2.565 2.684 4.44% 1.905 1.911 0.31% 3.667 3.788 3.21% 2.381 2.523 5.60%
M1 5.125 5.266 2.68% 3.629 3.786 4.16% 6.333 6.589 3.89% 4.856 4.946 1.83%
M2 3.202 3.158 −1.39% 2.129 2.145 0.75% 4.195 4.192 −0.07% 2.587 2.702 4.27%
R1 4.315 4.406 2.06% 3.184 3.291 3.24% 5.321 5.593 4.88% 4.274 4.398 2.83%
R2 2.560 2.634 2.81% 1.917 1.911 −0.29% 3.613 3.726 3.05% 2.421 2.485 2.60%
JX wave of 1.0 m/s2 (Case 3) HK wave of 1.0 m/s2 (Case 4)
Location Longitudinal Transverse Longitudinal Transverse
Test FEA Error Test FEA Error Test FEA Error Test FEA Error
L1 6.875 6.892 0.24% 5.072 5.193 2.32% 6.048 6.158 1.78% 4.648 4.678 0.65%
L2 4.615 4.885 5.52% 3.562 3.175 −12.18% 4.143 4.260 2.75% 2.789 2.900 3.83%
M1 7.926 8.052 1.57% 6.126 6.267 2.25% 7.082 7.283 2.76% 5.639 5.652 0.23%
M2 5.035 5.152 2.26% 3.108 3.160 1.65% 4.619 4.691 1.53% 2.731 2.933 6.91%
R1 6.766 6.842 1.12% 5.255 5.323 1.29% 6.071 6.183 1.81% 4.653 4.737 1.77%
R2 4.768 4.847 1.64% 3.084 3.172 2.78% 4.216 4.241 0.57% 2.777 2.893 4.03%
EC wave of 1232 m/s (Case 5) EC wave of 616 m/s (Case 6)
Location Longitudinal Transverse Longitudinal Transverse
Test FEA Error Test FEA Error Test FEA Error Test FEA Error
L1 15.424 15.840 2.62% 10.960 11.329 3.25% 9.426 9.504 0.82% 6.698 6.865 2.43%
L2 9.250 9.363 1.21% 6.688 6.773 1.25% 5.905 6.081 2.90% 4.204 4.289 1.98%
M1 18.578 18.749 0.92% 13.231 13.401 1.27% 11.585 12.006 3.50% 8.330 8.506 2.07%
M2 11.137 11.294 1.39% 7.721 7.860 1.76% 6.947 7.068 1.71% 4.719 4.805 1.81%
R1 15.786 15.885 0.62% 11.847 12.541 5.54% 9.369 9.531 1.70% 7.240 7.437 2.65%
R2 9.482 9.773 2.98% 6.880 7.104 3.15% 5.794 5.864 1.18% 4.204 4.317 2.61%
EC wave of 308 m/s (Case 7) EC wave of 2.0 m/s2 (Case 8)
Location Longitudinal Transverse Longitudinal Transverse
Test FEA Error Test FEA Error Test FEA Error Test FEA Error
L1 7.284 7.503 2.92% 5.176 5.532 6.44% 7.532 7.658 1.64% 5.461 5.579 2.12%
L2 4.563 5.100 10.54% 3.249 3.542 8.27% 5.037 5.156 2.32% 3.582 3.718 3.65%
M1 8.952 9.841 9.03% 6.437 6.669 3.48% 10.053 10.187 1.32% 6.757 7.061 4.31%
M2 5.368 6.227 13.79% 3.646 4.029 9.51% 6.316 6.405 1.39% 4.029 4.116 2.12%
R1 7.490 8.410 10.93% 5.594 6.150 9.04% 8.412 8.563 1.77% 6.258 6.497 3.67%
R2 4.477 5.082 11.89% 3.249 3.563 8.81% 5.068 5.208 2.70% 3.522 3.648 3.46%
EC wave of 3.0 m/s2 (Case 9) EC wave of 4.0 m/s2 (Case 10)
Location Longitudinal Transverse Longitudinal Transverse
Test FEA Error Test FEA Error Test FEA Error Test FEA Error
L1 11.911 12.055 1.20% 8.591 8.690 1.13% 16.296 16.385 0.54% 11.218 11.694 4.07%
L2 7.381 7.553 2.28% 5.073 5.261 3.56% 9.736 9.856 1.22% 5.876 7.018 16.27%
M1 15.080 15.203 0.81% 10.359 10.484 1.19% 19.064 19.997 4.67% 13.815 14.046 1.64%
M2 8.736 8.886 1.69% 6.345 6.485 2.16% 12.362 12.540 1.42% 8.379 8.531 1.78%
R1 12.222 12.325 0.84% 9.287 9.538 2.63% 16.624 16.865 1.43% 12.163 12.695 4.19%
R2 7.714 7.883 2.15% 5.373 5.500 2.31% 10.254 10.410 1.50% 7.364 7.418 0.72%
L1, M1, R1: acceleration responses at the top of the left tower, the middle tower, and the right tower; L2, M2, R2: acceleration responses at
the lower cross beam of the left tower, the middle tower, and the right tower
 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418 411

of the displacement responses are presented selected to calculate the displacement response ratios
(Table 3). It can be seen that the displacement re- among the non-uniform excitations and the corre-
sponses at the top of the three towers are different and sponding uniform excitations. The longitudinal and
the D2 value is the largest (exceed 10 mm), because vertical displacement ratios of the main girder at
there is no auxiliary pier and anchor cable in the two mid-span are described (Fig. 7).
middle spans, leading to the lack of effective assis- Fig. 7 indicates that all the vertical displacement
tance from the stiffness of the main girder and ca- ratios of the main spans are larger than 1.0 while all
bles. Therefore, some measures should be adopted to the longitudinal ratios of the scale model are smaller
improve the stiffness of the middle tower to control than 1.0 no matter how much the apparent wave ve-
the larger deformation of the middle tower in the locity is. In other words, the vertical displacement of
three-tower cable-stayed bridge as much as possible. the main girder increases while the longitudinal dis-
Due to the restriction of the bearing, the displace- placement at the top of the tower and at the end of the
ment response of D6 is smaller than other dis- deck decreases after considering the traveling wave
placement responses for the same case. Note that the effects. The displacement responses become larger as
vertical displacement at the mid-span of the main the apparent wave velocity increases, especially for
span cannot be ignored during horizontal earthquake the prominent amplification of 1232 m/s. Among the
excitations. displacement response ratios at the top of the three
Meanwhile, the influence of the traveling wave towers, the variation of the middle tower is the
on the displacement responses of the deck and tower smallest, and the variation of the left tower is the
should not be ignored in the seismic design of the largest. The similar phenomenon is also occurring in
floating system cable-stayed bridge. Three different the relative displacement between the tower and deck.
apparent wave velocities of the EC wave, which were Therefore, the influence of the apparent wave velocity
308 m/s, 616 m/s, and 1232 m/s, respectively, were on the displacement of the side tower where the

Table 3 Displacement responses of the scale model under uniform excitations (unit: mm)
Displacement response
Case
D1 D2 D3 D4 D5 D6 D7 D8 D9 D10
1 6.6 7.6 6.5 2.4 2.2 2.2 2.2 2.3 5.6 5.9
2 7.0 10.9 7.1 3.1 2.8 2.9 2.8 3.1 7.4 7.6
3 8.0 12.4 7.8 4.1 3.8 3.9 3.8 4.2 10.0 10.2
4 7.6 11.5 7.4 3.6 3.3 3.4 3.3 3.5 8.3 8.5
D1, D2, D3: top displacements of the left tower, the middle tower, and the right tower; D4, D8: left and right end displacements of the
main girder; D5, D6, D7: relative displacements between the main girder and the left tower, between the main girder and the middle tower,
and between the main girder and the right tower; D9, D10: vertical displacements of the left mid-span and the right mid-span

Fig. 5 Three types of multi-support excitations (unit: cm)

412 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418

4 3
FEA FEA
3 Test Test

Transverse accelaration (m/s2)

Vertical accelaration (m/s2)

2
2

1 1

0 0

-1
-1
-2
-2
-3
(a) (b)
-4 -3
00 55 10
10 15
15 20
20 25
25 3030 35
35 4040 4545 5050 5555 00 55 10
10 15
15 20
20 2525 3030 35
35 40
40 45
45 50
50 5555
Time（s） Time（s）
Time (s) Time (s)

15 10
FEA FEA
Test 8 Test
10 Transverse accelaration 2(m/s2) 6
Vertical accelaration (m/s2)

Transverse accelaration（m/s
）
Vertical accelaration（m/s
）

4
2

5 2
0
0
-2
-5 -4
-6
-10 -8
(c) -10 (d)
-15 -12
00 22 44 66 88 1010 1212 1414 1616 18
18 2020 2222 00 22 44 66 88 1010 1212 1414 1616 18
18 20
20 22
22
Time（s） Time（s）
Time (s) Time (s)
8 4
FEA FEA
6 Test 3 Test
Transverse accelaration (m/s2)
Vertical accelaration (m/s2)

4 2
Vertical accelaration（m/s
2
）

2 1

0 0

-2 -1

-4 -2

-6 (e) -3
(f)
-8 -4
0
0 10
10 20
20 30
30 40
40 50
50 6060 00 10
10 20
20 3030 40
40 50
50 6060
Time（s） Time（s）
Time (s) Time (s)

15 12
FEA FEA
Test 10 Test
Transverse accelaration (m/s2)

10 8
Vertical accelaration (m/s2)

6
5
4
2
0
0
-2
-5
-4
-10 -6
(g) -8 (h)
-15 0 -10
0 55 10
10 15
15 20
20 2525 3030 35
35 40
40 45
45 50
50 55
55
00 55 10
10 15
15 20
20 25
25 30
30 35
35 40
40 45
45 50
50 5555
Time（s） Time（s）
Time (s) Time (s)

Fig. 6 Acceleration responses comparison of deck

(a), (c), (e), and (g): vertical comparison under Case 1, 3, 7, and 10; (b), (d), (f), and (h): transverse comparison under Case 1, 3, 7, and 10
 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418 413

earthquake wave arrives first is the largest, and the left tower under non-uniform excitations are de-
impact on the middle tower is the smallest for the scribed (Fig. 8).
three-tower cable-stayed bridge. Table 4 shows that the bending moment re-
sponses of the side piers are smaller than those of the
5.2 Bending moment responses
towers and the values of the middle tower are the
Since internal forces at the bottom of the tower largest among those of the three towers under four
vary widely for different structural systems of cable- uniform excitations. However, the bending moment
stayed bridges, the bending moments of the tower and increment of the left tower is the largest among those
pier were analyzed for the scale model. The detailed of the three towers and the variation of the middle
bending moments under four uniform excitations are tower is the smallest during non-uniform excitations
recorded (Table 4), and the ratios of the three towers (Fig. 8). In addition, the bending moment ratios of all
as well as the time history curves at the bottom of the the towers become larger as the apparent wave

1.6 1.05
308 m/s (b)
(a)
616 m/s
1.4
1232 m/s Longitudinal displacement ratio 1.00
Vertical displacement ratio

1.2
0.95
1.0
0.8 0.90

0.6
0.85
0.4 308 m/s
0.80 616 m/s
0.2
1232 m/s
0.0 0.75
Left main span Right main span D1 D2 D3 D5 D6 D7
Displacements of different locations

Fig. 7 Displacement response ratios of the scale model under non-uniform excitations: (a) vertical displacement
ratios; (b) longitudinal displacement ratios

Table 4 Bending moment responses under uniform excitations (unit: kN·m)

Bending moment response
Case
Left pier Left tower Middle tower Right tower Right pier
1 33.72 248.15 350.12 254.49 23.20
2 46.24 330.78 454.86 324.77 32.77
3 50.16 439.17 616.81 444.27 70.89
4 48.09 364.77 521.48 377.03 44.08

1.6 400 Uniform

(a) 308 m/s (b)
Bending moment of left tower (kN·m)
Bending moment ratio of the tower

616 m/s 1232 m/s

1.4 300 616 m/s
1232 m/s
1.2 308 m/s
200
1.0 100
0.8 0
0.6 -100
0.4 -200
0.2 -300
0.0 -400
Left tower Middle tower Right tower 0 10 20 30 40 50 60
Time (s)
Fig. 8 Bending moment responses at the bottom of towers under non-uniform excitations: (a) bending moment
ratios of three towers; (b) bending moment time history of left tower
414 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418

velocity increases. Therefore, the traveling wave kind of phenomenon was consistent with the shaking
effect has an adverse impact on the internal force of table test results (Zong et al., 2014).
the towers. The influence on the side tower where the
4
earthquake wave arrives first is also the largest, and
3
that on the middle tower is also the smallest for the 2

Acceleration (m/s )
three-tower cable-stayed bridge.

2
1
0
-1
6 Collapse and failure analysis
-2
-3
The explicit dynamic model established in -4

LS-DYNA was also an FE model, which was similar -5

0 5 10 15 20 25 30 35 40 45 50 55
to the above single girder model in ANSYS without Time (s)
the modal validation. Due to space limitation, only (a)
part of the numerical simulation analysis about the 4

collapse and failure mode of the scale model during 3

only two different seismic excitations were reported 2

Acceleration (m/s )
2

and the corresponding discussions were given in this 1

study. To explore the failure mode of the scale model 0
during strong earthquakes and to demonstrate the -1
collapse process of the components, two earthquake -2
waves of the JX wave and EC wave with the accel- -3
eration peak value of 4.0 m/s2 were selected as the
-4
bi-direction input excitations in the numerical simu- 0 3 6 9 12 15 18 21
Time (s)
lation. The time history curves of the two waves are
(b)
described (Fig. 9).
Fig. 9 Time history curves of two input excitations:
6.1 Failure mode of the scale model under the EC (a) EC wave; (b) JX wave
wave
6.2 Failure mode of the scale model under the JX
The whole process of the components failure of
wave
the scale model can be demonstrated in an animation
form using the software of LS-PREPOST and the Considering that the seismic response of the
failure conditions at several key times during the EC scale model under the JX wave among the four wave
wave excitation are displayed (Fig. 10). The failure excitations was the largest, the whole process for the
process indicated that the slight vibrations of the cable components failure of the scale model under the JX
can be visible at 2.2 s which was the moment for the wave was also simulated. The failure conditions of
EC wave with the peak value, at the same time, which several key times are presented (Fig. 11), and two
can also arouse the small horizontal displacement of typical failure modes for the scale model under the JX
the towers. However, the vibration effect can be re- wave with the LS-DYNA are shown (Fig. 12).
duced when the earthquake was passing away. The These results showed that the small deformation
vertical vibration of the main girder aggravated and of the bearing spring elements between the middle
the torsion phenomenon of the cables emerged at tower and the main girder had occurred and the slight
53.7 s. During the whole action process, no obvious visible vibrations of the cables appeared at the time of
vibration of the towers occurred. The bearing springs 0.8 s, and then the deformation of the bearing devel-
of the middle tower were stretched sharply at 20.5 s, oped and the vibration of the cables gradually in-
according to the profile chart, and the failure state of creased. The cable at the highest point of the upper
the bearing was assumed to occur at this time. This tower column of the middle tower was separated from
 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418 415

Fig. 10 Failure mode of the scale model under the EC wave

(a) 0 s; (b) 2.2 s (peak value point): slight vibration of the cables; (c) 20.5 s: failure of the bearing at the middle tower;
(d) 53.7 s: torsional vibration of the cables

Connection failure Connection failure

Element failure
Connection failure Connection failure

Element failure
Connection failure Connection failure

Element failure Element failure Element failure

Connection failure Connection failure

Fig. 11 Failure mode of the scale model under the JX wave

(a) 0 s; (b) 0.8 s: vibration of the cables; (c) 2.2 s: connection failure of the cables and middle tower; (d) 8.4 s (peak value
point): element failure of the inner upper column of the middle tower; (e) 19.7 s: connection failure of all the cables and
middle tower, and all element failure of the upper column of the middle tower; (f) 20.5 s: connection failure of all the
cables and three upper towers, and all element failure of the upper column and cross beam of the middle tower
416 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418

Connection failure and all the element failures of the upper columns and
upper cross beams of the middle tower happened at
around 20.5 s. Meanwhile, the fracture deformations
of the upper tower column of the side towers ap-
peared, the significant deformation of the low tower
column of the middle tower emerged, owing to the
1 2
large deformation of the low cross beam. The whole
structural system collapsed. Note that the failure of
the main girder did not take place because of the
integral effects between the main girder and the
rigid arm.
3 4

(a)

7 Conclusions
Connection failure Element failure

In this paper, seismic responses of a three-tower

cable-stayed bridge scale model under multi-support
excitations and failure modes during two strong
earthquake actions were investigated by using two
11 22 kinds of nonlinear finite element models, respective-
ly. The main conclusions can be extracted from the
simulation and listed as follows:
1. The acceleration response results of the nu-
merical simulation are consistent with those from the
33 44 shaking table tests (Zong et al., 2014), indicating that
the established single girder model can predict the
(b)
seismic response of the scale model during uniform
Fig. 12 Two typical states of the scale model under the excitations and non-uniform excitations.
JX wave: (a) destruction state at the time of 8.4 s; (b)
destruction state at the time of 20.5 s 2. The acceleration responses of the scale model
1: enlarged view of side tower; 2: enlarged view of middle under non-uniform excitations become larger as the
tower; 3: elevation; 4: panoramagram
apparent wave velocity increases; travelling wave
effects have adverse impact on the vertical dis-
the tower and fell down, and then the connection placement of the main girder and the bending moment
failure of the stay cables in the upper tower column at the bottom of the tower, and the effects have ben-
were gradually coming down, beginning at 2.2 s. The eficial impact on the displacement at the top of the
inboard elements and upper cross beam elements in tower and that at the end of the girder. In addition, the
the upper tower column of the middle tower first influence of the travelling wave on the side tower
experienced failure at 8.4 s as the peak value time of where the earthquake wave arrives first is the largest
the JX wave. The stay cables in the upper tower and the influence on the middle tower is the smallest.
columns between the side towers and middle tower 3. The displacement responses of the middle
also started to crack at the same time. When it came to tower are the most significant in the three towers, due
19.7 s, all the stay cables of the upper tower columns to the lack of effective assistance on the stiffness of
of the middle tower were snapped, and all the out- the main girder and cables. Therefore, some measures
board elements of the upper column of the middle should be adopted to improve the stiffness of the
tower also failed. The connection failure of all the middle tower to control the larger deformation of the
cables and the upper columns of all the three towers, middle tower of the three-tower cable-stayed bridge.
 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418 417

4. The constructed explicit dynamic model can neering Vibration, 29(6):154-161 (in Chinese).
predict the failure modes and collapse process of the LSTC (Livermore Software Technology Corporation), 2006.
LS-DYNA (V970) Keyword User’s Manual. LSTC,
scale model by using the LS-DYNA. Only the bearing
Livermore, USA.
of the middle tower failure occurred under the EC Luo, R., Xu, H.R., Li, H.J., et al., 2010. Collapse simulation
wave excitation of 4.0 m/s2. Under the JX wave ex- analysis of the Baihua bridge during the Wenchuan
citation of 4.0 m/s2, the connection failure started earthquake. The 8th Academic Conference Symposium
between the cables and upper columns of all three on the National Earthquake Engineering (I), Chongqing,
towers, and then the elements failure occurred in the China (in Chinese).
Nazmy, A.S., Abdel-Ghaffar, A.M., 1990a. Non-linear earth
upper tower column and the upper cross beam in the
quake-response analysis of long-span cable-stayed
middle tower. The whole structural system collapsed bridges: applications. Earthquake Engineering & Struc-
in the end. This kind of analysis may be referenced as tural Dynamic, 19(1):63-76. [doi:10.1002/eqe.4290190
the basis for the collapsed failure control of the cable- 107]
stayed bridge with large span during a strong Nazmy, A.S., Abdel-Ghaffar, A.M., 1990b. Three-dimensional
nonlinear static analysis of cable-stayed bridges. Com-
earthquake.
puter and Structures, 34(2):257-271. [doi:10.1016/0045-
7949(90)90369-D]
References Nazmy, A.S., Abdel-Ghaffar, A.M., 1992. Effects of ground
Abdel-Ghaffar, A.M., Nazmy, A.S., 1991. 3D nonlinear seis- motion spatial variability on the response of cable-stayed
mic behavior of cable-stayed bridge. Journal of Struc- bridges. Earthquake Engineering & Structural Dynamics,
tural Engineering, 117(11):3456-3476. [doi:10.1061/ 21(1):1-20. [doi:10.1002/eqe.4290210101]
(ASCE)0733-9445(1991)117:11(3456)] Okamoto, Y., Nakamura, S., 2011. Static and seismic studies
Allam, S.M., Datta, T.K., 1999. Seismic behavior of cable- on steel/concrete hybrid towers for multi-span cable-
stayed bridges under multi-component random ground stayed bridges. Journal of Constructional Steel Research,
motion. Engineering Structures, 21(1):62-74. [doi:10. 67(2):203-210. [doi:10.1016/j.jcsr.2010.08.008]
1016/S0141-0296(97)00141-7] Pantaleon, M.J., Ramos, O.R., Ortega, G., et al., 2010. Dy-
Bao, L.X., Li, X.Z., Wei, X., et al., 2008. Evaluation of seismic namic analysis of a composite cable-stayed bridge:
resistance capacity for Yibin Yangtze River Cable-stayed Escaleritas viaduct. Journal of Bridge Engineering, 15(6):
Bridge. Engineering Mechanics, 25(2):174-182. 653-660. [doi:10.1061/(ASCE)BE.1943-5592.0000116]
Ernst, H.J., 1965. Der e-modul von seilen unter Ren, W.X., Makoto, O., 1999. Elastic-plastic seismic behavior
berücksichtigung des durchhanges. Der Bauingenieur, of long span cable-stayed bridges. Journal of Bridge En-
40(2):52-55 (in German). gineering, 4(3):194-203. [doi:10.1061/(ASCE)1084-
Fan, L.C., Wang, J.J., Chen, W., 2001. Response characteris- 0702(1999)4:3(194)]
tics of long-span cable-stayed bridges under non-uniform Ren, W.X., Peng, X.L., Lin, Y.Q., 2005. Experimental and
seismic action. China Journal of Computational Me- analytical studies on dynamic characteristics of a large
chanics, 18(3):359-364 (in Chinese). [doi:1007-4708 span cable-stayed bridge. Engineering Structures, 27(4):
(2001)03-0358-06] 535-548. [doi:10.1016/j.engstruct.2004.11.013]
Fang, Y., Li, J.Z., Peng T.B., et al., 2011. Influence of traveling- Self, S.P., Dilger, W.H., 1990. Nonlinear analysis and collapse
wave effect on seismic response of a long-span multi- load of PC cable-stayed bridge. Journal of Structural
span cable stayed bridge. Journal of Vibration and Shock, Engineering, 116(3):829-849. [doi:10.1061/(ASCE)
29(10):148-152 (in Chinese). [doi:10.3969/j.issn.1000- 0733-9445(1990)116:3(829)]
3835.2010.10.030] Song, W.K., Kim, S.E., 2007. Analysis of the overall collapse
Gimsing, N.J., Georgakis, C.T., 2011. Cable Supported mechanism of cable-stayed bridges with different cable
Bridges: Concept and Design, 3rd Edition. Wiley, layouts. Engineering Structures, 29(9):2133-2142.
Chichester, UK, p.150-160. [doi:10.1016/j.engstruct.2006.11.005]
Li, Z.X., Huang, J., Ding, Y., et al., 2005. Seismic responses of Wilson, J.C., Gravelle, W., 1991. Modeling of a cable-stayed
long-span cable-stayed bridges under different patterns of bridge for dynamic. Earthquake Engineering & Struc-
earthquake excitations. China Journal of Highway and tural Dynamic, 20(8):707-721. [doi:10.1002/eqe.4290
Transport, 18(3):48-53 (in Chinese). [doi:1001-7372 200802]
(2005)03-0048- 06] Yoo, H., Na, H.S., Choi, D.H., 2012. Approximate method for
Lin, J.Q., Bai, C.X., Chen, Y.S., et al., 2009. Study on the estimation of collapse loads of steel cable-stayed bridges.
seismic response of cable-stayed bridges with multi- Journal of Constructional Steel Research, 72:143-154.
support excitations. Earthquake Engineering and Engi- [doi:10.1016/j.jcsr.2011.12.003]
418 Zhou et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2014 15(6):405-418

Zavoni, E.H., Vanmarcke, E.H., 1994. Seismic random vibra- response study on a multi-span cable-stayed bridge scale
tion analysis of multi-support structural system. Journal model under multi-support excitations. Part I: shaking
of Engineering Mechanics, 120(10):1107-1128. [doi:10. table tests. Journal of Zhejiang University-SCIENCE A
1061/(ASCE)0733-9399(1994)120:5(1107)] (Applied Physics & Engineering), 15(5):351-363. [doi:10.
Zong, Z.H., Zhou, R., Huang, X.Y., et al., 2014. Seismic 1631/jzus.A1300339]

中文概要：
本文题目：多点激励下多跨斜拉桥缩尺模型的地震响应研究. 部分 2：数值分析
Seismic response study on a multi-span cable-stayed bridge scale model under multi-support
excitations. Part II: numerical analysis
研究目的：通过非线性动力有限元分析，模拟多塔斜拉桥在多点激励下的地震响应以及揭示其在强震下
的倒塌破坏模式。
创新要点：解决了多塔斜拉桥的非线性动力数值模拟技术，并首次模拟了多塔斜拉桥在强震下的破坏模
式。
研究方法：首先采用隐式的 Newmark-β 法求解多点地震激励下的非线性动力方程（图 3），随后采用显
式的中心差分格式对强震作用下桥梁的倒塌破坏过程进行数值仿真（图 4）。
重要结论：1.单主梁模型的数值结果可以很好地反映地震模拟振动台试验的地震响应结果；2.在加速度
峰值为 4.0 m/s2 的 El-Centro 波时，多塔斜拉桥模型只出现了中塔支座的破坏；3.在加速度峰
值为 4.0 m/s2 的江心波时，多塔斜拉桥模型先后出现了支座破坏、拉索与塔的连接破坏以及
中塔上塔柱段的单元失效。
关键词组：多塔斜拉桥；非线性有限元分析；多点激励；破坏模式