ARTICLE IN PRESS

JOURNAL OF
SOUND AND
VIBRATION
Journal of Sound and Vibration 290 (2006) 388–409
[Link]/locate/jsvi

Vibration analysis of the Second Saikai Bridge—a concrete

filled tubular (CFT) arch bridge
Mistuhiro Yoshimuraa, Qingxiong Wub, Kazuo Takahashib,,
Shozo Nakamurab, Kazuyoshi Furukawac
a
Nagasaki Shipyard & Machinery Works, Mitsubishi Heavy Industries, Ltd., 180, Akunoura-cho, Nagasaki, Japan
b
Department of Civil Engineering, Faculty of Engineering, Nagasaki University, 1-14, Bunkyo-machi, Nagasaki, Japan
c
Civil Engineering Information Office, Public Works Department, Nagasaki Prefectural Government, 2-13,
Edo-machi, Nagasaki, Japan
Received 30 April 2004; received in revised form 15 March 2005; accepted 2 April 2005
Available online 15 June 2005

Abstract

Concrete-filled steel tubular (CFT) arch bridges have been rapidly developing in China since 1990.
Research has focused on the static behavior, thermal stress and erection technique, however, and there has
been very little research of natural vibrations and dynamic responses of these bridges. Japan’s first CFT
arch bridge in highway bridges, the Second Saikai Bridge, is now under construction in Nagasaki
Prefecture. Furthermore, this bridge has a pedestrian bridge that is suspended under the girder, which is
rare. Therefore, the natural vibration properties are examined, and the influence of pedestrian bridge
structure on the natural vibration of main bridge is discussed first. Response analysis under a moving
vehicle and pedestrian is carried out, and the response characteristics and response level are clarified. The
natural vibrations and responses are compared to those of CFT arch bridges that have been constructed in
China. Results show the fine performance of both main bridge and pedestrian bridge of the Second Saikai
Bridge.
r 2005 Elsevier Ltd. All rights reserved.

Corresponding author. Tel.: +81 95 819 2610; fax: +81 95 819 2627.
E-mail address: takahasi@[Link] (K. Takahashi).

0022-460X/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/[Link].2005.04.004
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 389

1. Introduction

In arch bridges, it may be practical to use concrete filled steel tubes (CFT) for the arch ribs in
which the compression axial force is predominant. The infilled concrete delays local buckling of
the steel tube, and the steel tube reinforces the concrete against tension, bending moment, and
shear force. The tube also serves as a formwork for the concrete during construction of the bridge,
thus saving a major construction cost [1,2]. Moreover, the composite structural action between the
infilled concrete and the steel tube improves the member’s load-carrying capacity. As a result of
these advantages, CFT construction for bridges has become widespread in recent decades [3,4].
More than one hundred CFT arch bridges have been constructed since 1990 [5–7] in China, and a
CFT arch bridge with the span length of over 400 m is now being constructed [8]. The Second
Saikai Bridge, the first CFT arch highway bridge in Japan, is under construction in Nagasaki
Prefecture.
Research into CFT arch bridges has focused on the static behavior, thermal stress and erection
technique [7]. However, there has been very little research into vibrations and seismic responses of
CFT arch bridges. The dynamic properties of CFT arch bridges have not been examined in detail.
Therefore, it is necessary to evaluate the characteristics of the natural vibration, the dynamic
response under traffic loading and the nonlinear seismic response of the actual Second Saikai
Bridge. This paper focuses on the in-plane vibrations and discusses the properties of the natural
vibration and the traffic-induced vibration. The nonlinear seismic responses of this bridge were
discussed in a separate paper [9].
This bridge has a pedestrian bridge under the girder. The pedestrian bridge allows pedestrians
to cross the Second Saikai Bridge and visit a nearby public park. There is no other example in a
highway bridge with such a structure suspended under the girder so it is necessary to understand
the dynamic response characteristics of the pedestrian bridge. In particular, the response levels of
the pedestrian bridge when a vehicle passes over the girder or when a pedestrian walks on the
pedestrian bridge must be understood.
In this paper, the in-plane natural vibration properties of the bridge and the pedestrian bridge
are examined, and the influence of the pedestrian bridge structure on the natural vibration of the
bridge is discussed first. A response analysis induced by a moving vehicle is carried out, and the
response characteristics of the bridge and the pedestrian bridge are clarified. The natural
vibrations and responses are compared to those of CFT arch bridges that have been constructed
in China. Finally, the response induced by a pedestrian walking on the pedestrian bridge is
evaluated.

2. A brief description of the Second Saikai Bridge

The Second Saikai Bridge (tentative name) is a highway bridge on the Nishisonogi Region
Expressway. The bridge crosses the Harioseto channel and connects the Sasebo urban area and
the Nagasaki urban area. The bridge is scheduled to be completed in 2006. As shown in Fig. 1(a),
the main span is 240 m in length and the side spans are 30 m in length. The bridge is 20.2 m in
width. The landscape of the site, the need to harmonize the design with that of the present Saikai
Bridge (a deck type steel arch bridge with a span of 216 m, completed in 1955) and the need to
 ARTICLE IN PRESS

390 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

Fig. 1. General view of the Second Saikai Bridge (unit: mm). (a) Main bridge; (b) Pedestrian bridge.

reduce the bridge’s construction costs together lead the selection of the half-through braced-rib
CFT arch bridge design for the main span. The bridge has two parallel arch ribs, each of which
has a triangular cross-section consisting of three steel tubes. High fluidity concrete fills the three
steel tubes. The three steel tubes have an outer diameter of 812.8 mm and a thickness that differs
depending upon the position of the arch rib. The deck system consists of I-beams arranged
longitudinally, upon which a composite slab is placed.
There is a pedestrian bridge under the bridge, which allows pedestrians to cross the Second Saikai
Bridge and visit the public park since the sidewalk cannot be built in the highway bridge. This
pedestrian bridge is 293.225 m long and 3 m wide. The pedestrian bridge itself has a girder made of
two H-type beams. This girder hangs by strand rope and steel tube from the lateral beams of the
bridge. A general view of the pedestrian bridge is shown in Fig. 1(b). The center of the pedestrian
bridge has been widened to allow visitors to enjoy the view from the Second Saikai Bridge.
The pedestrian bridge is designed by using static sidewalk live load specified in Design
Specifications for Highway Bridges in Japan [10]. The dynamic design considering moving
vehicles and walking pedestrians is not adopted in Design Specifications for Highway Bridges in
Japan. Therefore, the dynamic analysis is performed to check the dynamic safety of the bridge.

3. Analytical model

3.1. Basic assumptions

The basic assumptions of the present paper are as follows:

(1) The geometric nonlinearity is not considered to evaluate the response characteristics of this
bridge under a moving vehicle or pedestrian traffic.
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 391

(2) It is assumed that the stress–strain relationships of steel tube and filled concrete are linear. The
deformation of steel tube is taken equal to that of filled concrete along the boundary of two
materials. Consequently, the slip between steel tube and filled concrete does not occur.
(3) Equations of bridge-single-vehicle dynamic system subjected to road roughness are used to
evaluate the responses of main bridge and the pedestrian bridge. A two-degrees-of-freedom
vehicle model is employed. The model of road surface is modeled by a Gaussian process.

3.2. Integrated bridge model

Fig. 2 shows a three-dimensional FE model of the integrated bridge. This model takes into
account the rigidity and mass of the pedestrian bridge. The arch rib, stiffening girder, lateral
bracing, and pier are all modeled using three-dimensional beam elements based on actual cross-
sectional properties. The column and hanger are modeled using three-dimensional truss elements.
The lumped mass matrix is adopted for this analysis.
On the pedestrian bridge, the girder and lateral beams of the pedestrian bridge are modeled
using three-dimensional beam elements. The strand ropes and steel tubes are modeled using three-
dimensional truss elements. The girder of the pedestrian bridge takes into account the stiffness of
the floor system of pedestrian bridge.
The material properties of this CFT arch bridge (cross-sectional area A and elastic modulus E),
including the pedestrian bridge, are described in Table 1. Because the cross sections of the arch
rib, stiffening girder or etc. are different depending upon the position, Table 1 shows the basic
material properties at the center of those structures. The mass properties of the CFT arch rib
bridge and the pedestrian bridge are summarized in Table 2.
Regarding boundary conditions, the springing position of the arch rib and the pier bases are
fixed in all degrees of freedom. The earthquake shear force distribution rubber bearings are
installed in the joints between the stiffening girder and the piers (P4, P5 and P6), the joints between
the stiffening girder and abutment A2, and the joints between the stiffening girder and the lateral
beams (S3 and S4). The model uses linear springs to simulate these bearings. The points linking the
suspension members of the pedestrian bridge and the stiffening girder of the bridge have the same

Fig. 2. Integrated bridge model.

 ARTICLE IN PRESS

392 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

Table 1
Basic cross-sectional area A and elastic modulus E of the Second Saikai Bridge

Content A (m2) E (kN/m2)

Main bridge Arch rib 8.351
102 (Inside, outside rib) 2.00
108
1.537
102 (Lower rib)
Stiffening girder 1.164
101 2.00
108
Hanger 5.810
103 1.96
108
Pedestrian bridge Girder 1.577
102 (General part) 2.00
108
1.803
102 (Widened part)
Hanger 9.898
104 (Cable) 1.40
108
4.078
103 (Steel tube) 2.00
108

Table 2
Mass properties of the Second Saikai Bridge

Content Weight (kN)

Main bridge Arch rib 24,604

Stiffening girder 7,747
Floor slab 56,050
Lateral bracing etc. 11,139
Pedestrian bridge girder 5,858

longitudinal, vertical and out-of-plane displacements as the stiffening girder of the main
bridge. The vertical displacements in the original points and the end points of the pedestrian
bridge are restrained, as to the out-of-plane displacements in the lateral bracing of the pedestrian
bridge.

3.3. Original bridge model and separate pedestrian bridge model

In order to examine the influence of the pedestrian bridge on the natural vibrations of the
bridge, the original bridge model and a separate pedestrian bridge model are used in addition to
the integrated bridge model.
In the original bridge model in Fig. 3(a), the rigidities of the pedestrian bridge are not taken into
account, and the masses of the pedestrian bridge are included with those of the girder. This
original bridge model is adopted for the nonlinear seismic analysis of the main bridge.
The separate pedestrian bridge model is shown in Fig. 3(b). In this model, only the pedestrian
bridge’s structure is considered. Regarding the boundary condition of the pedestrian bridge
model, the points that link the hangers of the pedestrian bridge and the stiffening girder of the
bridge are fixed in the longitudinal, vertical and out-of-plane degrees of freedom.
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 393

(a) (b)

Fig. 3. (a) Original bridge model and (b) separate pedestrian bridge model.

4. Natural vibration characteristics

The natural vibration analysis is carried out using the integrated bridge model, and the natural
vibration characteristics are examined. In this paper, the natural vibrations of the bridge except
for the local oscillation of the pedestrian bridge are called the ‘natural vibrations of the main
bridge’, and the local natural vibrations of the pedestrian bridge are called the ‘natural vibrations
of the pedestrian bridge’.

4.1. Natural vibration characteristics of main bridge

The natural frequencies and modal shapes of the in-plane natural vibrations of the main bridge
are demonstrated in Table 3.
The first in-plane natural vibration, which has a frequency of 0.472 Hz, corresponds to the
floating mode, as it does with cable-stayed bridges. This mode is a longitudinal pendulum-like
movement of the stiffening girder as illustrated in Table 3. This is due to the rubber bearings
installed in the longitudinal direction of the bridge that distribute earthquake shear forces. The
second in-plane natural vibration has a frequency of 0.639 Hz, which corresponds to the unique
antisymmetric mode of arch bridges.
Furthermore, there are in-plane natural vibration modes in which the vibrations of the side
spans are predominant. They have natural frequencies of 2.553 Hz (the 13th mode) and 2.644 Hz
(the 15th mode).

4.2. Natural vibration characteristics of pedestrian bridge

Table 4(a) shows in-plane natural vibrations of the pedestrian bridge using the integrated bridge
model. The lowest in-plane natural frequency of the pedestrian bridge natural vibration is
2.276 Hz, which is much higher than the lowest in-plane natural frequency of the natural vibration
of the main bridge.
 ARTICLE IN PRESS

394 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

Table 3
In plane natural vibrations of the main bridge

4.3. Influence of natural vibrations of pedestrian bridge on natural vibrations of main bridge

The weight of the pedestrian bridge on the Second Saikai Bridge is 8% of the steel weight of the
integrated bridge. Therefore, in the model of the integrated bridge, the original bridge and the
pedestrian bridge are treated as a monolithic structure. However, the original bridge is used for
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 395

Table 4
In-plane natural vibrations of the pedestrian bridge

the seismic analysis of the main bridge [9] and the separate pedestrian bridge model is used for the
natural vibration analysis of the pedestrian bridge itself. In other words, it is assumed that
the effect of the rigidity of the pedestrian bridge on the vibration of the main bridge is negligible.
 ARTICLE IN PRESS

396 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

4 4
Integrated bridge model Integrated bridge model
Original bridge model Separate Pedestrian bridge model

3 3

Frequency (Hz)
Frequency (Hz)

2 2

1 1

0 0
0 5 10 15 20 25 0 5 10 15 20 25
(a) Number of mode (b) Number of mode

Fig. 4. In-plane natural frequencies: (a) natural vibrations of the main bridge; (b) natural vibrations of the pedestrian
bridge.

The influence of the natural vibrations of the pedestrian bridge on the natural vibrations of the
integrated bridge is considered using the integrated bridge, original bridge, and independent
pedestrian bridge models in order to analyze the validity of the assumptions.
The in-plane natural frequencies of the main bridge vibrations using the integrated bridge
model and the original bridge model are plotted in Fig. 4(a). There are some differences in the
natural modes near the natural frequency of 2.5–2.8 Hz, though the results using the original
bridge model coincide with those using the whole bridge model. The vertical displacements in the
origin points and the end points of the pedestrian bridge in the integrated bridge model are
restrained, but these boundary conditions are not applied to the original bridge model. In other
words, the differences in the natural frequencies are due to the different boundary conditions. The
natural frequency of the original bridge model is about 8% lower than that of the integrated
bridge model. Therefore, the influence of the natural vibrations of the pedestrian bridge on the
natural vibration of the main bridge is seen to be small.
Fig. 4(b) and Table 4 show the natural vibrations of the pedestrian bridge for the integrated
bridge model and for the separate pedestrian bridge model. The natural frequencies of the
separate pedestrian bridge model agree well with those for the integrated bridge model. According
to the dense distribution of natural frequencies of the separate pedestrian bridge, the vibration
behavior of the pedestrian bridge installed in the main bridge is similar to that of an attached
beam connected with a main beam by uniformly elastic spring. Thus, the effect of the rigidity of
the pedestrian bridge on the vibration of the main bridge is negligible. Moreover, the natural
vibrations of the pedestrian bridge can be evaluated using the separate pedestrian bridge model.
However, when a moving vehicle passes by the girder of the main bridge, the dynamic analysis of
the pedestrian bridge must use the integrated bridge model.

4.4. Comparison of CFT arch bridge and steel arch bridge

The Jianghan Wuqiao Bridge (240 m, CFT, China) [11] and the Saikai Bridge (216 m, steel,
Japan) [12], which have almost the same span length as the Second Saikai Bridge, are used to
compare the natural vibration properties of CFT arch bridges with those of steel arch bridges. The
in-plane natural frequencies of those three bridges are summarized in Table 5. The weigh of CFT
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 397

Table 5
Natural frequencies of CFT arch bridges and steel arch bridge

Bridge name Span (m) Type Natural frequencies of in-plane modes (Hz)

First First Second Second

antisymmetric symmetric antisymmetric symmetric

Second Saikai Bridge 230 CFT 0.639 0.929 1.509 1.474

Jianghan Wuqiao Bridge 240 CFT 0.724 1.229 1.878 2.041
Saikai Bridge 216 Steel 1.153 1.507 2.805 2.306

arch bridges is greater than that of the steel arch bridge. Therefore, the natural frequencies of the
CFT arch bridges are smaller than those of steel bridges as can be seen in Table 5. The arch action
is not effective in the out-of-plane direction. The seismic safety of CFT arch bridges should be
checked when the bridges are subjected to earthquake in the out-of-plane direction [9–13].

5. Response characteristics under a moving vehicle

In this section, the vibrations of the Second Saikai Bridge under traffic loading are examined
using the bridge–vehicle–road surface model based on Refs. [14–16]. For a long-span bridge, there
are often multiple vehicles traveling on it. The response level of bridge subjected to multiple
vehicles becomes large. However, the case where a single vehicle is traveling the bridge has the
large frequency in which a pedestrian receives the influence of traveling vehicle. Accordingly, in
the present paper study, a single moving vehicle is adopted to evaluate the response level of this
bridge.
The equations of motion of the bridge subjected to the loading can be expressed as
n o n o n o
½M y€ g þ ½C y_ g þ ½K yg ¼ fF g, (1)

where [M], [C] and [K] are the mass, damping and stiffness matrices of the bridge, respectively, fF g
is the load vector, fyg g is the displacement vector representing the displaced shape of the bridge,
fy_ g g is the velocity vector and fy€ g g is the acceleration vector [17].
A two-degrees-of-freedom vehicle is used as shown in Fig. 5 and the equation of motion of
vehicle is
½M v Z€ tþDt þ ½C v Z_ tþDt þ ½K v Z tþDt ¼ fF v g (2)
and the force from vehicle to bridge fFg is expressed as follows
fFg ¼ jfF v g, (3)
where
" # " #
MS C S1 þ C S2 r1 C S1  r2 C S2
½M v  ¼ ; ½C v  ¼
My r1 C S1  r2 C S2 r21 C S1 þ r22 C S2
 ARTICLE IN PRESS

398 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

r2 r1

M
Ms

Ms2
ZS
Ms1

KS2 CS2 KS1 C S1

RR RF

Fig. 5. Two-degrees-of-freedom vehicle model.

Table 6
Parameters of vehicle

MS (t) 25.000
My (t m2) 64.000
MS1 (t) 5.000
MS2 (t) 20.000
KS1 (kN/m) 1776.529
KS2 (kN/m) 7106.115
CS1 (kN s/m ) 28.274
CS2 (kN s/m) 113.097
r1 (m) 3.200
r2 (m) 0.800
Damping constant hc 0.030
First frequency (Hz) 3.000
Second frequency (Hz) 3.000

and
" #
K S1 þ K S2 r1 K S1  r2 K S2
½K v  ¼
r1 K S1  r2 K S2 r21 K S1 þ r22 K S2



are the mass, damping and stiffness matrices of the vehicle, Z€ , Z_ and fZ g are the

s
acceleration, velocity and displacement vectors of the vehicle and Z ¼ Z y , j is the coefficient
matrix which convert the load to the node load of the bridge, fF v g is the load vector subjected to
vehicle and is expressed as
" #( ) " #( )
C S1 C S2 R_ F K S1 K S2 RF
fF v g ¼ þ
r1 C S1 r2 C S2 R_ R r1 K S1 r2 K S2 RR
" #( ) " #( )
C S1 C S2 y_ F K S1 K S2 yF
þ þ . ð4Þ
r1 C S1 r2 C S2 y_ R r1 K S1 r2 K S2 yR

yF and yR are the displacements of the girder at the positions of the front and rear wheel. RF and
RR are the road surface roughness at the position of the front and rear wheel. The parameters of
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 399

the vehicle used in this paper are listed in Table 6. The power spectral density of the road surface
roughness is determined using Refs. [14–16] and can be expressed as
A
S r ðOÞ ¼ 2
, (5)
O þ a2
where O is the frequency of the road surface (cycle/m), A is a parameter expressing the level of
road surface roughness and a is a parameter based on the observed result. In the present paper,
the road surface is assumed to be in the best condition when A ¼ 0:001 and a ¼ 0:05. Just for
reference, a simulated wave of road surface roughness by AR model [18] is presented in Fig. 6.
According to ISO standard [19], the wave corresponds to the ‘best’ road surface (A ¼ 0:001).
Fig. 7 shows the power spectral density of this wave and the target spectral density. The power
spectral density of the simulated wave coincides with the target power spectral density.
The response of the bridge under a moving vehicle is obtained by solving Eqs. (1)–(3). The
Newmark b method (b ¼ 0:25) of direct integration is used.
The Rayleigh damping [20] is used to form the damping matrix of bridge [C] as a linear
combination of mass matrix [M] and stiffness matrix [K] (see Eq. 1)
½C ¼ a3 ½M þ a4 ½K, (6)

2
Road surface roughness (cm)

-1

-2
0 50 100 150 200 250 300 350 400
Space (m)

Fig. 6. Road surface roughness.

1
10
Power spectral density (cm /cycle/m)

Target value
Simulation value
0
10
2

-1
10

-2
10

-3
10

-4
10
0.01 0.1 1 10
Wave Number Ω (cycle/m)

Fig. 7. Power spectrum of road surface roughness.

 ARTICLE IN PRESS

400 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

in which a3 and a4 are arbitrary proportional factors. These factors are given in terms of two
different frequencies (oi ; oj ) and the corresponding critical damping ratios (hi ; hj ) (ioj) and
expressed as
2f i f j ðhi =2pf j  hj =2pf i Þ 2ðhj f j =2p  hi f i =2pÞ
a3 ; a4 . (7)
f 2j  f 2i f 2j  f 2i

In this paper, the damping constants hi and hj are assumed to be equal to 0:01 and the frequencies
f i and f j are set to 0.639 Hz and 0.929 Hz, respectively.

5.1. Response characteristics of girder

Fig. 8 shows the maximum velocities of the girder and arch rib when a vehicle passes by the
girder of the main bridge at a speed of 80 km/h. The vertical axis corresponds to the maximum
velocities and the horizontal axis corresponds to the coordinate in the longitudinal direction of the
bridge.
On the main span, the maximum velocity of the girder of the main span is about 0.8 cm/s.
There is an investigation for the response level of a CFT arch bridge built in China [21]. The
Shitanxi Bridge is a half-through CFT arch with a span length of 136 m. The bridge has two
parallel arch ribs, each of which has a rectangular cross-section consisting of four steel tubes. The

P4 P5 P6 A2
0.02
G1004
Velocity (m/sec)

G1058

0.01 G1031

0.00
-40 0 40 80 120 160 200 240
(a) Coordinate of longitudinal direction (m)

0.02
Outside upper chord member
Inside upper chord member
Lower chord member
Velocity (m/sec)

0.01

0.00
-40 0 40 80 120 160 200 240
(b) Coordinate of longitudinal direction (m)

Fig. 8. Maximum velocities under moving vehicle: (a) girder; (b) arch rib.
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 401

response level of this bridge under the vehicle loading is investigated in 2000. The maximum
velocity of the girder in the Shitanxi Bridge is about 5.7 cm/s.
Comparing the maximum velocity of the Second Saikai Bridge with that of the Shitanxi Bridge,
the vibration level in the girder of the main span of the Second Saikai Bridge under traffic loading
is smaller than that of the Shitanxi Bridge.
The maximum velocity in the arch rib of the Second Saikai Bridge is less than 0.5 cm/s as shown
in Fig. 8(b), while that of the Shitanxi Bridge is about 3.9 cm/s. Therefore, the vibration level in
the arch rib of the Second Saikai Bridge is smaller than that of the Shitanxi Bridge.
However, a vibration with a maximum velocity of 1.5 cm/s is generated in the side spans, which
have a span length of 30 m. Time histories of displacement, velocity and acceleration at the center
points (G1004 and G1058) of the side spans are shown in Fig. 9. When a vehicle passes by the
center point of the side span (t ¼ 0:79 s), the displacement of the center point G1004 of the side
span is the largest. The displacement of the girder at t ¼ 0:79 s is shown in Fig. 10. The girder in
the main span generates small displacements even if the displacement of the side span is
predominant. In other words, the vibration of the side span is local, and the in-plane natural
vibration, in which the vibration of the side span with natural frequency of 2.553 Hz (the 13th
mode in Table 3) is predominant, is induced. This means that the response of the side span is large
due to the resonance, since the in-plane natural frequency of 2.553 Hz in the side spans
corresponds with the natural frequency of 3.0 Hz of the vehicle.

5.2. Response characteristics of pedestrian bridge

Fig. 11 shows the maximum velocities of the pedestrian bridge when a vehicle passes by the
girder of the main bridge at a speed of 80 km/h. Fig. 12 shows the time histories for center points
P21009 and P21109 of the side spans. Fig. 13 shows the time histories for center points P21057 of
the main span. The maximum velocity of the main span is about 0.8 cm/s and that of side spans is
about 1.8 cm/s, which are the same as those of the girder of the main bridge.
The fundamental vibration frequency of the added pedestrian bridge is much higher than that
of the main bridge so that the added pedestrian bridge does not behave like a tuned mass damper
in the present bridge.

5.3. Response level

The ergonomic evaluation method developed by Kobori and Kajikawa [22] is used to appraise
the vibration sensibility of the traffic-induced vibration on the Second Saikai Bridge. The
vibration stimulation S (cm/s) is calculated using the effective value of the maximum velocity Vmax
(cm/s). It is described by the following equation:
pffiffiffi
S ¼ V max = 2. (8)
The vibration greatness level VGL(dB) is given as described by the following equation:
VGL ¼ 20 log10 ðS=S0 Þ ðS0 ¼ 1:4
102 cm=sÞ. (9)
The vibration greatness VG can then be obtained from the following expressions:
log10 VG ¼ 0:05ðVGL  40Þ when VGLp40 dB,
 ARTICLE IN PRESS

402 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

0.002

Displacement (m)
0.000

-0.002

-0.004 G1004
G1058

-0.006
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(a) Time (sec)

0.02
Velocity (m/sec)

0.01

0.00

-0.01 G1004
G1058

-0.02
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(b) Time (sec)

0.4
Accelation (m/sec )
2

0.2

0.0

-0.2 G1004
G1058

-0.4
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time (sec)
(c)

0.003
Power spectral density (m2/sec3)

0.002
2.55 Hz

0.001 3.02 Hz

0.000
0 1 2 3 4 5 6 7 8 9 10
(d) Frequency (Hz)

Fig. 9. Time histories at the center point of side span (G1004 and G1058): (a) displacement; (b) velocity; (c)
acceleration; (d) power spectral density of G1004 acceleration.
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 403

0.002

Displacement (m)
0.000

-0.002

-0.004

-0.006
-40 0 40 80 120 160 200 240
Coordinate of longitudinal direction (m)

Fig. 10. Displacement of the girder at t ¼ 0.79 s.

P4 P5 P6 A2
0.02
P21009
P21109
Velocity (m/sec)

P21057
0.01

0.00
-40 0 40 80 120 160 200 240
Coordinate of longitudinal direction (m)

Fig. 11. Maximum velocities of pedestrian bridge under moving vehicle.

log10 VG ¼ 0:03ðVGL  40Þ when VGL440 dB: (10)

The relationship between the category and the lower limit of the VG for pedestrians is shown in
Table 7.
Table 8 summarizes the maximum velocities and corresponding VG of the Second Saikai
Bridge.
Regarding the main span, both the maximum velocity of the girder and that of the pedestrian
bridge are about 0.014 m/s, which is equal to VG ¼ 0.73. From the viewpoint of vibration
sensation, its VG is in the category of ‘definitely perceptible’, and the vibration of the main span
under traffic loading is small.
In the side spans, the maximum velocity of the girder of the main bridge and that of the
pedestrian bridge are about 0.016 m/s and 0.013 m/s, which are equal to VG ¼ 0.80 and
VG ¼ 0.66, respectively. These VG are in the category of ‘definitely perceptible’. Therefore, the
vibration of the side spans under traffic loading is small, too.

6. Response characteristics under pedestrian traffic

The lowest natural frequency of the pedestrian bridge is 2.276 Hz. This is close to 2.0 steps/s,
which corresponds to the pace of an adult walking down a street [23]. There is some apprehension
 ARTICLE IN PRESS

404 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

0.002

Displacement (m)
0.000

-0.002 P21009
P21109

-0.004

-0.006
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(a) Time (sec)

0.02

0.01
Velocity (m/sec)

0.00

P21009
-0.01 P21109

-0.02
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(b) Time (sec)
0.4
Accelation (m/sec 2)

0.2

0.0

P21009
-0.2 P21109

-0.4
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(c) Time (sec)

Fig. 12. Time histories of points (P21009 and P21109) in the pedestrian bridge: (a) displacement; (b) velocity;
(c) acceleration.

about the resonance induced by a pedestrian walking on the pedestrian bridge when the
predominant frequency of the pedestrian bridge is near the pace of the pedestrian. Accordingly,
the dynamic response characteristics of the pedestrian bridge induced by pedestrian traffic are
examined.

6.1. Load induced by pedestrian

The following walking load, which takes into account the dead load of the pedestrian [23,24], is
used to examine the dynamic response characteristics that are induced by a pedestrian walking on
the pedestrian bridge.
F ðtÞ ¼ W ð1 þ a cos otÞ, (11)
where W is the weight of the pedestrian and is set to 0.588 kN (60 kgf) in this paper, o ¼ 2pf is the
circular frequency of the pace, a is the ratio of the impulsive force according to the pace and the
relationship between a and pace is shown in Fig. 14 according to Ref. [23]. The relationship
between velocity and pace is also shown in the same figure.
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 405

0.002
0.001
Displacement (m)

0.000
-0.001
-0.002
-0.003
-0.004
-0.005
-0.006
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(a) Time (sec)

0.02

0.01
Velocity (m/sec)

0.00

-0.01

-0.02
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(b) Time (sec)

0.4
Accelation (m/sec 2 )

0.2

0.0

-0.2

-0.4
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(c) Time (sec)

Fig. 13. Time histories of point P21057 in the pedestrian bridge: (a) displacement; (b) velocity; (c) acceleration.

Table 7
Category and lower limit of VG for pedestrain

No. Content of category VG

1 Slightly perceptible 0.32

2 Definitely perceptible 0.61
3 Lightly hard to walk 1.12
4 Extremely hard to walk 1.48
 ARTICLE IN PRESS

406 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

Table 8
Response level of the Second Saikai Bridge

Position Vmax (m/s) VG Category

Girder N1004 0.01449 0.73 Definitely perceptible

N1031 0.00502 0.25 —
N1058 0.00942 0.48 Slightly perceptible
Pedestrian bridge N21009 0.01577 0.80 Definitely perceptible
N21057 0.00707 0.36 Slightly perceptible
N21109 0.01297 0.66 Definitely perceptible

2 6


1.6

Velocity (m/sec)
V
1 3


0.4

0 0
Pace (step/sec)

Fig. 14. Relationship between a and v [23].

P4 P5 P6 A2
0.006

P21057 P21061
P21009 P21109
Velocity (m/sec)

0.004

0.002

0.000
-40 0 40 80 120 160 200 240
Coordinate of longitudinal direction

Fig. 15. Maximum velocity of the pedestrian bridge in the pace of 2.55 steps/s.

6.2. Response characteristics of pedestrian bridge

The maximum velocities of the pedestrian bridge at the pace of 2.55 steps/s are shown in Fig. 15.
Among all the points shown, the largest velocities are generated at the center point P21009 of the
side span between pier P4 and pier P5, at center point P21109 of the side span between pier P6 and
abutment A2, and at points P21057 and P21061 in the wide part of the main span.
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 407

2.55Hz 2.62Hz 2.70Hz 2.71Hz 2.83Hz

0.010 0.010
P21009 P21057
P21109 P21061
0.008 0.008
Velocity (m/sec)

Velocity (m/sec)
0.006 0.006

0.004 0.004

0.002 0.002

0.000 0.000
2.0 2.2 2.4 2.6 2.8 3.0 2.0 2.2 2.4 2.6 2.8 3.0
(a) Pace (steps/sec) (b) Pace (steps/sec)

Fig. 16. Relationship between the velocity and the pace of pedestrian: (a) center point of side spans; (b) wide part.

0.003 0.30

Power spectral density (m2/sec)

0.002 0.25
Velocity (m/sec)

0.001 0.20

0.000 0.15

-0.001 0.10

-0.002 0.05

-0.003 0.00
0 20 40 60 80 100 120 140 160 2 3 4
Time (sec) Frequency (Hz)

Fig. 17. Time history of velocity and power spectrum of P21009 in the pace of 2.55 steps/s.

The relationship between the pace of a walking pedestrian and the velocity at those four points
is shown in Fig. 16.
Regarding point P21009, shown in Fig. 16(a), the peak of maximum velocity appears near
2.55 steps/s. Fig. 17 shows the time history of the velocity and the power spectrum at the pace. The
power spectrum indicates that the frequencies of 2.51 and 2.55 Hz are predominant. 2.51 Hz
corresponds to the natural frequency of the 12th vibration mode of the main bridge (see Table 3),
and 2.55 Hz corresponds to the natural frequency of the 13th vibration mode of the main bridge
(see Table 3) and the frequency of the pedestrian’s pace. In other words, a pace of 2.55 steps/s
induces vibrations at the frequencies of the 12th and 13th modes of the main bridge.
The peaks of the maximum velocities of points P21057 and P21061 in the main span appear
near 2.70, 2.71 and 2.83 steps/s, as shown in Fig. 13(b). In the response spectrum shown in Fig. 18,
the frequencies of 2.72 and 2.67 Hz are predominant. The frequency of 2.72 Hz corresponds to the
natural frequency of the 16th mode of the pedestrian bridge (see Table 4), and the frequency of
2.67 Hz corresponds to that of the pedestrian’s pace. Consequently, vibrations at the frequencies
of the 16th mode of the pedestrian bridge are induced at a pace of 2.70 steps/s.
Regarding the response level induced by a pedestrian walking on the pedestrian bridge, the
maximum velocity of the pedestrian bridge induced by a walking pedestrian is less than 0.008 m/s
(VG ¼ 0.39), which is in the category of ‘slightly perceptible’. Therefore, the responses induced by
a pedestrian walking on the pedestrian bridge are small.
 ARTICLE IN PRESS

408 M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409

0.010 0.30

Power spectral density (m2/sec)

0.25
0.005
Velocity (m/sec)

0.20

0.000 0.15

0.10 2.72
-0.005 2.67
0.05

-0.010 0.00
0 20 40 60 80 100 120 140 160 2 3 4
Time (sec) Frequency (Hz)

Fig. 18. Time history of velocity and power spectrum of P21061 in the pace of 2.70 steps/s.

7. Conclusions

This paper examined the natural vibrations, vehicle- and pedestrian-induced vibrations of the
Second Saikai Bridge, which is the first CFT arch highway bridge constructed in Japan. The main
findings are as follows:

(1) The floating mode and the unique antisymmetric mode of the arch bridge are the first two in-
plane vibrations of the main bridge, and the first natural frequency of the pedestrian bridge is
higher than the lower frequencies of the main bridge.
(2) The independent pedestrian bridge model can be used to evaluate the natural vibrations of the
pedestrian bridge, since the natural frequencies of the pedestrian bridge are separate from
those of the main bridge. The influence of the rigidity of the pedestrian bridge on the main
bridge is small, since the first in-plane natural frequency is higher than the fundamental
frequency of the main bridge.
(3) The vehicle-induced responses of the main span are small, while the responses in the side
spans are larger than those of the main span. The maximum velocity of the side span is
about 1.8 cm/s.
(4) The dynamic response induced by the pedestrian walking on the pedestrian bridge increases
when the predominant frequencies of the pedestrian bridge are close to the pace of the
pedestrian. However, the response level of the pedestrian bridge is small.

Acknowledgements

The authors wish to express their gratitude to Chodai Co. Ltd. for providing the materials used
in this study.

References

[1] C.W. Roeder, B. Cameron, C.B. Brown, Composite action in concrete filled tubes, Journal of Structural
Engineering 125 (5) (1999) 477–484.
 ARTICLE IN PRESS

M. Yoshimura et al. / Journal of Sound and Vibration 290 (2006) 388–409 409

[2] A.H. Varma, J.M. Ricles, R. Sause, L.W. Lu, Experimental behavior of high strength square concrete-filled steel
tube beam-columns, Journal of Structural Engineering 128 (3) (2002) 309–318.
[3] W.C. Clawson, Bridge applications of composite construction in the US, Structural Engineering in the 21st
Century, Proceedings of the Structures Congress, 1999, pp. 544–547.
[4] S. Nakamura, New structural forms for steel/concrete composite bridges, Structural Engineering International
1 (2000) 45–50.
[5] Z. Zhen, B. Chen, Q. Wu, Recent development of CFST arch bridge in China, Proceeding of 6th ASCCS
Conference, USA, 2000, pp.205–212.
[6] Y. Liu, B. Chen, H. Hikosaka, Recent developments in concrete-filled tubular arch bridge and horizontal swing
erection method in China, Bridge and Foundation Engineering 33 (2) (1999) 41–44 (in Japanese).
[7] Q. Wu, B. Chen, K. Takahashi, S. Nakamura, Construction and technical subjects of concrete filled steel tubular
arch bridges in China, Bridge and Foundation Engineering 35 (10) (2001) 40–45 (in Japanese).
[8] D. Peng, Q. Wu, K. Takahashi, S. Nakamura, Recent construction and development of long-span bridges in
China, Bridge and Foundation Engineering 37 (2) (2003) 43–49 (in Japanese).
[9] Q. Wu, M. Yoshimura, K. Takahashi, S. Nakamura, H. Fujita, K. Furukawa, Vibration and nonlinear seismic
analysis of the Second Saikai Bridge-concrete filled tubular (CFT) arch bridge, Proceedings of the 5th
Japan–German Joint Symposium, Osaka, Japan, 2003, pp.133–142.
[10] Japan Road Association, Design Specifications for Highway Bridges, Part I: Steel Bridge, Maruzen Co., Ltd.,
Japan, 2002 (in Japanese).
[11] Q. Wu, K. Takahashi, H. Matsuzaka, B. Chen, S. Nakamura, Study on natural vibrations and nonlinear seismic
response of CFT arch bridge constructed in China, Journal of Constructional Steel 11 (2003) 177–184 (in Japanese).
[12] Q. Wu, K. Takahashi, H. Kobayashi, S. Nakamura, Natural vibration properties and nonlinear seismic responses
of the Saikai Bridge, Journal of Constructional Steel 11 (2003) 185–192 (in Japanese).
[13] Q. Wu, K. Takahashi, H. Matsuzaka, B. Chen, S. Nakamura, Study on dynamic properties of partially concrete
filled steel tubular arch bridge, Journal of Constructional Steel 10 (2002) 141–148 (in Japanese).
[14] T. Okabayashi, Mean square response analysis of highway bridges under a single moving vehicle, Journal of
Structural Mechanics and Earthquake Engineering 286 (1979) 15–27 (in Japanese).
[15] J. Xu, T. Okabayashi, Optimal design of a dynamic damper for highway bridge vibration under a moving vehicle,
Proceedings of the Fifth PaciSteel Conference, Seoul, Korea, 1998, pp. 367–372.
[16] Q. Wu, K. Takahashi, T. Okabayashi, S. Nakamura, Response characteristics of local vibrations in stay cables on
an existing cable-stayed bridge, Journal of Sound and Vibration 261 (3) (2003) 403–420.
[17] R.W. Clough, J. Penzien, Dynamics of Structure, McGraw-Hill Kogakusha, 1975.
[18] Y. Iwatani, Simulation of multidimensional wind fluctuations having any arbitrary power spectra and cross
spectra, Journal of Wind Engineering 11 (1982) 5–18 (in Japanese).
[19] ISO (International Organization for Standardization), Proposals for generalized road inputs to vehicles, ISO/TC
108/WG9 Document No.5, 1972.
[20] T.K. Caughey, Classical normal modes in damped linear dynamic systems, Journal of Applied Mechanics 27 (1960)
269–271.
[21] M. Okatani, B. Chen, Q. Wu, T. Okabayashi, Dynamic behavior of a concrete filled steel tubular arch bridge under
traffic road, Proceedings of Annual Conference of the Japan Society of Civil Engineers, vol. 56(1) 2001, pp. 288–289.
[22] T. Kobori, Y. Kajikawa, Ergonomic evaluation methods for bridge vibrations, Journal of Structural Mechanics
and Earthquake Engineering 230 (1974) 23–31 (in Japanese).
[23] M. Yoneda, Dynamic response characteristics of the stress ribbon pedestrian bridges due to a walking human and
a simplified method for evaluating maximum amplitude of this type of bridges, Journal of Structural Engineering
47A (2001) 351–362 (in Japanese).
[24] M. Yoneda, Relationships between numerical dynamic responses using various human walking force models and
full scale measurements of single span pedestrian bridges, Journal of Structural Engineering 46A (2000) 439–448
(in Japanese).