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

net/publication/279910622

Vehicle-bridge interactions and applications to high speed rail bridges

Article · March 1998

CITATIONS READS
358 1,235

2 authors:

Y.B. Yang Jong-Dar Yau

Chongqing University Tamkang University
311 PUBLICATIONS 10,794 CITATIONS 98 PUBLICATIONS 3,794 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

vehicle scanning method View project

Vehicle-Bridge Interaction View project

All content following this page was uploaded by Y.B. Yang on 29 September 2020.

The user has requested enhancement of the downloaded file.

 ARTICLE IN PRESS

JOURNAL OF
SOUND AND
VIBRATION
Journal of Sound and Vibration 284 (2005) 205–226
www.elsevier.com/locate/jsvi

Vehicle–bridge interaction dynamics and potential applications

Y.B. Yang, C.W. Lin
Department of Civil Engineering, National Taiwan University, No. 1, Section 4, Roosevelt Road, Taipei 10617, Taiwan
Received 8 September 2003; received in revised form 1 June 2004; accepted 14 June 2004

Abstract

The dynamic interaction between a moving vehicle and the sustaining bridge is studied. By the method of
modal superposition, closed-form solutions are obtained for the vertical responses of both the bridge and
moving vehicle, assuming the vehicle/bridge mass ratio to be small. For both the bridge and vehicle
responses, it is confirmed that rather accurate solutions can be obtained by considering only the first mode.
The displacement, velocity, and acceleration of the bridge are governed at different extents by two sets of
frequencies, i.e., the driving frequency of the vehicle and natural frequencies of the bridge. From the
spectrum for the bridge displacement, the vehicle speeds can be shown to be associated with some low-
frequency pikes. On the other hand, the vehicle responses are governed by five distinct frequencies that
appear as driving frequencies, vehicle frequency, and bridge frequencies with shift. From the vehicle’s
acceleration spectrum, the first bridge frequency (with shift) is shown to have rather high visibility and can
be easily identified. The effects of damping of the vehicle and bridge are evaluated in the numerical studies.
Potential applications of the present results, as well as further researches required, are also indicated in the
paper.
r 2004 Elsevier Ltd. All rights reserved.

1. Introduction

The problem of a vehicle traveling over a bridge is commonly encountered in the transportation
facilities such as highway bridges, railroad bridges, aircraft/taxiway bridges in airports, and so on.
When a vehicle passes over a bridge, certain impact or dynamic amplification effect will be

Corresponding author. Tel.: +886 2 2363 2104; fax: +886 2 2363 7585.
E-mail address: [email protected] (Y.B. Yang).

0022-460X/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsv.2004.06.032
 ARTICLE IN PRESS

206 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

induced on the bridge, which needs to be taken into account in the design of bridges. Attention
paid to this subject dates back to the works of Willis [1] and Stokes [2] in the mid 19th century. It
has long been observed that when a bridge is subjected to moving loads, the induced dynamic
deflections and stresses can be significantly higher than those observed for the static case [3]. In
this aspect, the majority of the literature has been devoted to investigation of the bridge vibrations
using the so-called moving load [4–8], moving mass [9–15] and moving sprung mass models [6,16–19]
for the vehicles.
The moving load is the simplest model that can be conceived for a vehicle in studying the bridge
vibrations. With this model, the essential dynamic characteristics of the bridge caused by the
vehicle’s moving action can be captured with a sufficient degree of accuracy. However, it suffers
from the drawback that the interaction between the bridge and moving vehicle is ignored. For this
reason, the moving load model is strictly valid for the case where the mass of the vehicle is small
relative to that of the bridge, and only when the vehicle response is not desired. The moving mass
model represents an improvement over the moving load model in that the effect of inertia of the
vehicle is taken into account. Nevertheless, it does not allow consideration of the bouncing action
of the moving vehicle relative to the bridge. Such an effect is expected to be significant in the
presence of pavement roughness or for vehicles moving at rather high speeds.
To account for the bouncing or suspension action of the moving vehicle, a spring can be
attached to the moving mass to result in the sprung mass model shown in Fig. 1. This is the
simplest model that can be used to study the dynamic interaction between the moving vehicle and
the supporting bridge. It is true in the past two decades that researchers continue to develop
vehicle models of various complexities to account for the dynamic properties of the vehicle, see,
for instance, those listed in Refs. [18,20–25]. To the knowledge of the writers, however, a great
majority of the previous related works have been focused on the dynamic response of the bridge in
time domain, with little attention paid to the dynamic behavior of the moving vehicle or to the
frequency contents of the interaction between the two subsystems. Perhaps, one exception in this
regard is the series of works carried out by Yang and co-workers on the vehicle response, in which
equal emphases have been placed on the bridge and vehicle responses, using the latter as an
indicator for the riding comfort of passengers [17,18,24,26]. Recently, a preliminary study was

v
qv
mv
qb
kv
m , EI

L
Fig. 1. Vehicle moving over a simply supported bridge.
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 207

conducted by the authors to extract the fundamental frequency of a bridge from the dynamic
response of a passing vehicle [27].
Unlike most of the previous works, this paper is focused exclusively on the dynamic interaction,
especially the frequency aspects of such an interaction, between the moving vehicle and the bridge,
with equal emphases placed on the bridge and vehicle responses. To this end, two sets of second-
order differential equations of motion will be written, one for the vehicle and the other for the
bridge. It is the interaction or contact force existing between the two subsystems that makes the
two sets of equations nonlinear and coupled, as the contact force moves (and thereby varies) from
time to time, even though the two subsystems by themselves may be linear.
In order to highlight the effect of interaction, a simply supported beam subjected to a moving
sprung mass as shown in Fig. 1 will be adopted. The mass of the vehicle is assumed to be small
relative to that of the bridge. By the method of modal superposition, together with the use of
convolution integrals, closed-form solutions will be obtained for both the bridge and the moving
vehicle, which are approximate in the sense that iterations are not performed for updating the
contact force existing between the two subsystems to consider the mutual interaction. The
accuracies of the solutions obtained are compared with those obtained by independent finite
element analyses. By transforming all the time-history responses into the frequency domain, some
interesting phenomena can be observed. For example, from the displacement response of the
bridge, one can identify the vehicle speeds as those appearing as low-frequency pikes. On the other
hand, from the vehicle’s acceleration spectrum, it is seen that the first bridge frequency can be
identified with high visibility. Such implications are important and have their potential areas of
applications. It is suggested that future research be carried out along the lines of interaction
dynamics to investigate the practical aspects of the ideas presented herein.

2. Physical modelling and formulation

Consider a simply supported beam subjected to a vehicle moving at speed v, as shown in

Fig. 1. The vehicle is represented as a concentrated mass mv supported by a spring of stiffness kv
with the effect of damping of the suspension system neglected. The beam is assumed to be
of the Bernoulli–Euler type with constant cross sections. No consideration will be made
of the damping property or pavement irregularity of the bridge. The vehicle–bridge system as
shown in Fig. 1 allows us to investigate various interaction phenomena involved, as will be
presented below.
The equations of motion governing the transverse or vertical vibration of the bridge and
moving vehicle are
m̄u€ þ EIu0000 ¼ pðx; tÞ; (1)

mv q€ v þ kv qv ¼ kv ujx¼vt ; (2)
where m̄ denotes the mass per unit length, E the elastic modulus, and I the moment of inertia of
the beam, mv and kv denote the mass of vehicle and stiffness of the suspension system,
respectively, qv is the vertical displacement of the vehicle, uðx; tÞ is the vertical displacement of the
bridge, and pðx; tÞ is the applied force acting on the bridge through the contact point at position vt,
 ARTICLE IN PRESS

208 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

which moves with the vehicle. The applied force pðx; tÞ can be expressed as follows:
pðx; tÞ ¼ f c ðtÞ dðx  vtÞ; (3)
where dðx  vtÞ is the Dirac delta function evaluated at the contact point, x ¼ vt; and the contact
force f c ðtÞ is equal to the sum of the vehicle weight and the elastic force of the suspension system,
i.e.


f c ðtÞ ¼ mv g þ kv qv  ujx¼vt (4)
with g denoting the gravity of acceleration.
It should be noticed that the vertical displacement of the vehicle is measured from its static
equilibrium position. Therefore, there is no additional force acting on the vehicle except the elastic
force resulting from the shortening or elongation of the supporting spring. The solution to Eq. (1)
will be expressed in terms of the modal shapes, fn ðxÞ; and associated modal coordinates, qbn ðtÞ; as
X
uðx; tÞ ¼ fn ðxÞqbn ðtÞ: (5)
n

For the simply supported beam considered in this study, the mode shapes of the bridge are
known to be of the sinusoidal type. Therefore, Eq. (5) becomes
Xh npx i
uðx; tÞ ¼ sin qbn ðtÞ : (6)
n
L
Substituting Eq. (6) for the displacement uðx; tÞ into Eq. (1), multiplying both sides
of the equation by fm ðxÞ; and integrating with respect to x over the length L of the beam, one
obtains
Z L X Z L X Z L
0000
m̄fm ðfn q€ bn Þ dx þ EI fm ðfn qbn Þ dx ¼ f c ðtÞdðx  vtÞfm dx: (7)
0 n 0 n 0

By use of the orthogonality conditions for the modal shapes and changing the subscript m into
n, Eq. (7) reduces to
RL
2 f c ðtÞ 0 dðx  vtÞfn ðxÞ dx
q€ bn þ obn qbn ¼ RL ; (8)
m̄ 0 f2n ðxÞ dx
where obn is the frequency of vibration of the nth mode of the bridge,
rffiffiffiffiffiffi
n2 p2 EI
obn ¼ 2 : (9)
L m̄
Further, through manipulation of the right-hand side of Eq. (8), the nth modal equation of the
bridge can be obtained as follows:
2o2 mv npvt X jpvt
q€ bn þ o2bn qbn þ v sin sin q
m̄L L j
L bj
2o2v mv npvt 2mv g npvt
 sin qv ¼ sin : ð10Þ
m̄L L m̄L L
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 209

In a similar way, Eq. (2) can be rewritten as

X npvt
q€ v þ o2v qv ¼ o2v sin q ; (11)
n
L bn
where ov is the frequency of vibration of the vehicle
sffiffiffiffiffiffi
kv
ov ¼ : (12)
mv
By the assumption that the vehicle mass mv is much less than the bridge mass m̄L; i.e.,
mv =m̄L51; the governing equation in Eq. (10) can be approximated as follows:
2mv g npvt
q€ bn þ o2bn qbn ¼sin ; (13)
m̄L L
which is identical to the one for a bridge subjected to a single moving load with constant
speed v [27].

3. Dynamic responses of the bridge

Assuming zero initial conditions, the solution to the second-order linear differential equation of
the bridge in Eq. (13) can be obtained as
Dstn h npvt i
qbn ðtÞ ¼ sin  S n sin ðo bn tÞ ; (14)
1  S2n L
where Dstn is the static deflection caused by the vehicle with respect to the nth mode,
2mv gL3
Dstn ¼ (15)
n4 p4 EI
and Sn is a non-dimensional speed parameter,
npv
Sn ¼ : (16)
Lobn
Hence, the total displacement response of the bridge to a vehicle moving at speed v is
X Dstn n npx h npvt io
uðx; tÞ ¼ 2
sin sin  S n sin o bn t : (17)
n 1  Sn
L L
Differentiating Eq. (17) with respect to time t, we obtain the velocity response of the bridge as
follows:
X Dstn n npx h npv npvt io
_ tÞ ¼
uðx; 2
sin cos  ðo S
bn n Þ cos o bn t (18)
n 1  Sn
L L L
which can further be differentiated to yield the acceleration response of the bridge as
X Dstn  npx
2  npv2 npvt

€ tÞ ¼
uðx; 2
sin o S
bn n sin o bn t  sin : (19)
n 1  Sn
L L L
 ARTICLE IN PRESS

210 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

As can be seen, each of the responses of the bridge given in Eqs. (17)–(19) can be divided into
two components pertaining to the driving frequencies of the vehicle, npv/L, and natural frequencies
of the bridge, obn : By substitution of x ¼ L=2 into Eqs. (17)–(19), we obtain the mid-point
responses of the bridge as
  X
L Dstn np h npvt i
u ;t ¼
 sin sin  S n sin obn t ; (20)
2 n 1  S 2n 2 L
  X
L Dstn np hnpv npvt pv i
u_ ;t ¼ 2
sin cos  cos o bn t ; (21)
2 n ð1  S n Þ
2 L L L
  X
L Dstn np npv2 npvt pvobn
u€ ;t ¼ 2
sin  sin þ sin obn t : (22)
2 n ð1  S n Þ
2 L L L

By relating Dstn to Dst1 using the definition in Eq. (15), the coefficients of the two components
mentioned above for the bridge responses pertaining to the driving frequencies and natural
frequencies have been listed in Table 1. Because the midpoint of the beam happens to be a
stationary point for the modal shapes of even order, the amplitudes corresponding to all the even-
order modes are merely equal to zero. Thus, only the odd-order modes need to be considered in
Table 1. The contributions of the two frequency components to the midpoint displacement,
velocity and acceleration of the bridge were also plotted in Fig. 2(a)–(c) for the case of S1=0.1,
which means that the time for the vehicle to pass over the bridge is 10 times the fundamental
period of the bridge. As can be seen, the amplitudes of each individual mode pertaining to the two
frequency components decrease drastically as the order of the mode increases. For practical
applications, it is concluded that solutions of sufficient accuracy can be obtained for the bridge if
only the first mode of vibration is considered, as was pointed out by Biggs [16]. Besides, the
displacement amplitude is dominated mainly by the driving frequency component. The velocity

Table 1
Coefficients of frequency terms in bridge responses

Driving frequency npv=L Natural frequency obn

Displacement
Dstn1 Dst1 S 1



n4 1  S21 =n2 n6 1  S 21 =n2

Velocity
Dst1 S1 Dst1 S 1



n3 1  S 21 =n2 n4 1  S 21 =n2

Acceleration
Dst1 S21 Dst1 S 1



n2 1  S21 =n2 n2 1  S 21 =n2
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 211

1.2 1.2
Bridge Displacement Bridge Velocity
Driving Frequency Set Driving Frequency Set

Dimensionless Amplitude
1 1
Dimensionless Amplitude

Natural Frequency Set Natural Frequency Set

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
1 3 5 7 9 11 13 15 17 19 1 3 5 7 9 11 13 15 17 19
(a) Mode n (b) Mode n

1.2
Bridge Acceleration
Driving Frequency Set
1
Dimensionless Amplitude

Natural Frequency Set

0.8

0.6

0.4

0.2

0
1 3 5 7 9 11 13 15 17 19
(c) Mode n

Fig. 2. Bridge response amplitude for two frequency components (S1=0.1): (a) displacement, (b) velocity, and (c)
acceleration.

amplitude, however, is nearly equally governed by both frequency components. On the contrary,
the acceleration amplitude is dominated by the natural or bridge frequency component.
Another parameter that affects the amplitude of the individual frequency component is the
speed parameter S. As defined in Eq. (16), the speed parameter represents the ratio of the driving
frequency to the natural frequency of bridge, which is normally less than 0.5 in practice [28]. As
can be seen from Figs. 3 and 4, the displacement amplitude of each mode, particularly of
the first mode, increases as the speed parameter of the vehicle increases, due to the fact
that the energy input to the bridge is higher as the vehicle moves over the bridge at a higher
speed. On the other hand, a greater speed parameter also implies that a vehicle passes over a
bridge with a lower natural frequency or a longer span with the same speed. A bridge
with a longer span allows the vehicle to pass at a larger acting time, which in turn allows more
energy to be accumulated on the bridge. This explains why the amplitude of the displacement of
the bridge increases as the speed parameter of the vehicle increases. The same is also true for the
velocity and acceleration responses of the bridge, which are not shown here for the sake of saving
the paper length.
 ARTICLE IN PRESS

212 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

Fig. 3. Bridge displacement amplitude for driving frequency component.

Fig. 4. Bridge displacement amplitude for natural frequency component.

 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 213

3.1. Bridge response to a single moving vehicle

Fig. 5 shows the midpoint acceleration response of a bridge to a vehicle of mv=1200 kg and
kv=500,000 N/m moving at three different speeds, 5, 10 and 20 m/s. The following
properties are adopted for the bridge: L ¼ 25 m; m̄ ¼ 4800 kg=m and EI ¼ 3:3  109 N m2 : As
was mentioned above, the displacement response of a bridge is predominated by the
driving frequency component, and the velocity response is equally governed by the driving
frequency and bridge frequency components, but the acceleration response is predominated
by the bridge frequency component. Such a fact is confirmed by the spectrum plotted in
Figs. 6–8 for the displacement, velocity and acceleration of the midpoint displacement
of the bridge. Clearly, the influence of the driving frequency component relative to the
bridge frequency component decreases as we go from displacement, to velocity, and then
to acceleration.
It is interesting to note that the three driving frequencies (pv=L) identified from Figs. 6–8 are
0.1, 0.2 and 0.4 Hz, which correspond exactly to the three speeds 5, 10, 20 m of the moving
vehicle. The three driving frequencies all fall below 1 Hz, much less than the bridge
frequency. Besides, higher response amplitudes are generated by vehicles moving at higher
speeds, as indicated by the velocity and acceleration spectrum in Figs. 7 and 8, respectively.
One implication herein is that if the dynamic response of a bridge under moving loads
can be accurately monitored, the recorded data, when processed in real time, can be
used for detecting the moving speeds of vehicles over the bridge, at least from the theoretical
point of view. The feasibility of such an idea may be of interest to the traffic patrolpersons for
chasing speeding vehicles. It is suggested that further research be conducted in this regard
concerning the technical aspects.

0.1
Vehicle Speed
0.08 5 m/s
10 m/s
0.06 20 m/s
Acceleration (m/s2)

0.04

0.02

-0.02

-0.04

-0.06
0 5 10 15 20 25
Vehicle Position (m)
Fig. 5. Bridge midspan acceleration response.
 ARTICLE IN PRESS

214 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

0.0015

Amplitude of Displacement (m) Driving Frequency Vehicle Speed

0.0012 5 m/s
10 m/s
20 m/s

0.0009

0.0006

Bridge Frequency
0.0003

0
0.01 0.1 1 10 100
Frequency (Hz)
Fig. 6. Spectrum of bridge displacement response.

0.005
Vehicle Speed
5 m/s
Amplitude of Velocity (m/s)

0.004 10 m/s
20 m/s
Driving Frequency
0.003 Bridge Frequency

0.002

0.001

0
0.01 0.1 1 10 100
Frequency (Hz)
Fig. 7. Spectrum of bridge velocity response.

3.2. Bridge response to five moving vehicles

The same properties as those used in Section 3.1 are adopted for the bridge and vehicle. In order
to examine the effect of multi-vehicle loadings, however, we shall consider an arbitrary case of five
identical vehicles moving over the bridge at different speeds, i.e., 15, 8, 5, 12, 10 m/s, and at
different entrance times. The solution to such a multi-vehicle case can be obtained by superposing
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 215

0.04

Amplitude of Acceleration (m/s2) Vehicle Speed

5 m/s
0.03 10 m/s Bridge Frequency
20 m/s

0.02
Driving Frequency

0.01

0
0.01 0.1 1 10 100
Frequency (Hz)
Fig. 8. Spectrum of bridge acceleration response.

0.002

0.001

-0.001

-0.002
Vehicle Acting Duration on Bridge
15 m/s
-0.003 8 m/s
Vehicle Acting Duration
5 m/son Bridge
12 m/s
-0.004 10 m/s

-0.005

0 1 2 3 4 5 6

Fig. 9. Bridge midspan displacement response (five moving vehicles).

the results obtained for each individual vehicle, with due account taken of the effect of time lag
and different initial conditions [8]. The displacement response obtained for the midpoint of the
bridge has been plotted in Fig. 9, along with the acting duration of each vehicle on the bridge
indicated.
The spectrum of the displacement, velocity and acceleration of the midpoint of the bridge has
been shown in Figs. 10–12. From these figures, it is observed that the five driving frequencies
 ARTICLE IN PRESS

216 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

0.0015

Driving Frequency
Five Moving Vehicles
0.0012

0.0009

Bridge Frequency
0.0006

0.0003

0.01 0.1 1 10 100

Fig. 10. Spectrum of bridge displacement response (five moving vehicles).

0.008
Bridge Frequency
Five Moving Vehicles
0.006

0.004

Driving Frequency
0.002

0.01 0.1 1 10 100

Fig. 11. Spectrum of bridge velocity response (five moving vehicles).

implied by the five vehicle speeds can be clearly identified from the displacement spectrum, not so
clearly from the velocity spectrum, but almost vanish in the acceleration spectrum. Such an
observation is consistent with the statement that the driving frequency component loses its
influence as we go from displacement, to velocity and then to acceleration of the bridge.
Therefore, if the speed of a moving vehicle is to be detected from the dynamic response of the
supporting bridge, it is suggested that we work on the displacement or velocity response. In this
regard, further works should be conducted to develop techniques that are suitable for monitoring
the displacement or velocity, rather than acceleration response, of a bridge in the low-frequency
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 217

0.1

Five Moving Vehicles Bridge Frequency

0.08

0.06

0.04

0.02
Driving Frequency

0.01 0.1 1 10 100

Fig. 12. Spectrum of bridge acceleration response (five moving vehicles).

range, as most vibration sensors available nowadays are of the acceleration or velocity type and
are good for high-frequency vibrations.

4. Dynamic response of the vehicle

Substituting the expression in Eq. (14) for the bridge displacement qbn ðtÞ into the right-hand
side of Eq. (11), the equation of motion for the vehicle can be written as
X1
Dstn o2v n npvt h npvt io
q€ v þ o2v qv ¼ 2
sin sin  S n sin ðo bn tÞ : (23)
n¼1 1  S n
L L
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
gðtÞ

The term denoted as gðtÞ on the right-hand side above represents the interaction effect between
the bridge and moving vehicle, which is a function of time. The solution to Eq. (23) can be
obtained by the Duhamel integral as
Z
1 t
qv ðtÞ ¼ gðtÞ sin ov ðt  tÞdt: (24)
ov 0
Using the expression for gðtÞ in Eq. (23), the displacement response of the vehicle can be
integrated as
X1    
ðn  1Þpv ðn þ 1Þpv
qv ðtÞ ¼ A1n cos t þ A2n cos t þ A3n cosðov tÞ
n¼1
L L
npv npv
þA4n cos obn  t þ A5n obn þ t ; ð25Þ
L L
 ARTICLE IN PRESS

218 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

where the coefficients in the brackets are as follows

Dstn o2v
A1n ¼

; (26)
2ð1  S2n Þ ov þ ðn  1Þpv=L ov  ðn  1Þpv=L

Dstn o2v
A2n ¼

; (27)
2ð1  S2n Þov þ ðn þ 1Þpv=L ov  ðn þ 1Þpv=L

pv2
2Dstn o2v
n L
A3n ¼





1 S2nov þ ðn  1Þpv=L ov  ðn  1Þpv=L ov þ ðn þ 1Þpv=L ov  ðn þ 1Þpv=L


2Dstn Sn o2v npv=L obn




 ; ð28Þ
ov  obn þ npv=L ov þ obn  npv=L ov þ obn þ npv=L ov  obn  npv=L

S n Dstn o2v
A4n ¼

; (29)
2ð1  S 2n Þ ov  obn þ npv=L ov þ obn  npv=L

Sn Dstn o2v
A5n ¼

: (30)
2ð1  S2n Þ ov þ obn þ npv=L ov  obn  npv=L
The solution as derived here for the response of the moving vehicle is approximate in the sense
that the bridge displacement solved from Eq. (13) has been directly used in computing the acting
force for the vehicle equation in Eq. (23). In fact, the vehicle response computed from Eq. (23)
may affect again the acting force on the bridge, and therefore the response of the bridge.
Obviously, the interaction between the bridge and moving vehicle is an issue of iterative nature,
but in this study, only the first cycle of iteration was considered. In the numerical study, it will be
demonstrated that even with such an approximation, the closed-form solutions derived appear to
be quite accurate compared with the finite element analysis that takes into account the full effect
of interaction. Compared with the numerical approaches, the present approximate approach has
the advantage that the key parameters involved in each phenomenon can be clearly identified,
while clear physical meanings can be easily appreciated. It should be added that the solution
presented in Eq. (25) reduces to that previously derived [27], if only the first mode of vibration is
considered.
As can be seen from Eq. (25), the number of terms involved in the vehicle response increases as
the number of vibration modes considered for the bridge increases. There are five types of
frequencies involved in the vehicle response, which can further be categorized into three groups as:
(1) driving frequencies, including ðn  1Þpv=L and ðn þ 1Þpv=L; (2) vehicle frequency ov ; and (3)
bridge-related frequencies, obn  npv=L and obn þ npv=L; where the index n relates to the
vibration mode number of the bridge. In particular, the third group represents the fact that the
bridge frequencies obn are shifted by an amount equal to the vehicle speed npv=L:
In this study, we are particularly interested in the terms containing the bridge frequencies with
shift, i.e., obn  npv=L and obn þ npv=L; in Eq. (25), as they provide the clues for extracting the
bridge frequencies from the dynamic response of a passing vehicle, at least from the theoretical
point of view [27]. Before this can be realized, however, we like to have some idea about the
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 219

relative influence of the terms containing obn  npv=L and obn þ npv=L in the vehicle response of
Eq. (25) with respect to the remaining terms, and particularly to see if they are practically visible
from the response spectrum. To this end, we shall use mn to denote the ratio of the vehicle
frequency ov to the bridge frequency obn of the nth mode, mn ¼ ov =obn ; and rewrite the
coefficients in Eqs. (29) and (30) accordingly,
Dstn S n
A4n ¼  ; (31)
2ð1  S 2n Þ 1  m2n ð1  Sn Þ2

Dstn S n
A5n ¼  : (32)
2ð1  S 2n Þ 1  m2n ð1 þ Sn Þ2
In Table 2, the magnitudes of all the coefficients A1n ; A2n ; . . . ; A5n have been listed. It can be
observed that the coefficient A5n, i.e., the term associated with the frequency obn þ npv=L; remains
generally the largest for the practical ranges of m and S considered, i.e., for 0pmp5 and for
0pSp0:5: Such a fact offers us a theoretical means to extract the bridge frequency from the
vehicle response, if due account is taken of the shifting effect. Furthermore, the magnitude for the
second mode is about one quarter of that of the first mode, and even smaller for higher modes,
indicating that modes higher than the first one can generally be neglected in computing the vehicle
response, as will be demonstrated later on. By setting the denominator for the coefficient A5n
equal to 0 for n=1, one can obtain the condition for the resonance to occur, as indicated by the
vehicle–bridge frequency ratio m; which turns out to be equal to 1 for small S, but slightly less than
1 for large S. Under the condition of resonance, the coefficient A5n reaches a local maximum. The

Table 2
Amplitude of vehicle acceleration response

Frequency Amplitude First term (n=1) Second term (n=2)

2 1
ðn  1Þpv Dstn ðn  1Þpv=L 1
L   1  ðm1 S1 Þ2
2ð1  S2n Þ 1  ðn  1Þmn S1 2

ðn þ 1Þpv
2 1 1
Dstn ðn þ 1Þpv=L
L   1  ð2m1 S1 Þ2 1  ð3m1 S 1 Þ2
2ð1  S2n Þ 1  ðn þ 1Þmn S1 2

npv
2
2
obn  Dstn S 1 m2n 1  S1 =n2 S1 m21 ð1  S 21 Þ S1 m21 1  S 1 =4
L
  1  m21 ð1  S 1 Þ2
2 
2ð1  S 2n Þ n2 1  n2 mn 1  S 1 =n2 2 4 1  4m21 1  S1 =4

npv
2
2
obn þ Dstn S 1 m2n 1 þ S1 =n2 S1 m21 ð1 þ S 1 Þ2 S 1 m21 1 þ S1 =4
L
  1  m21 ð1 þ S 1 Þ2
2 
2ð1  S 2n Þ n2 1  n2 mn 1 þ S 1 =n2 2 4 1  4m21 1 þ S1 =4
 ARTICLE IN PRESS

220 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

same is also true for the coefficient A4n, except that the vehicle–bridge frequency ratio m is equal to
1 for small S, but slightly greater than 1 for large S.

5. Numerical verification

In order to evaluate the effect of approximation made in deriving the solution to Eqs. (10) and
(11), i.e., by assuming the mass of the vehicle to be small compared with that of the bridge, an
independent finite element analysis [27] that does not rely on any such assumption will be
conducted. The following properties are adopted for the bridge: cross-sectional area A=2.0 m2,
moment of inertia I=0.12 m4, elastic modulus E=2.75  1010 N/m2, length L=25 m, and mass
per-unit-length m̄ ¼ 4800 kg=m: The vehicle is assumed to have a mass mv=1200 kg and a spring
constant kv=500,000 N/m, which implies a vibration frequency of ov ¼ 20:41 rad=s ð¼ 3:25 HzÞ:
And the beam is divided into 20 beam elements. The first three eigenvalues ob computed for the
bridge are: 13:09 rad=s ð¼ 2:08 HzÞ; 52:36 rad=s ð¼ 8:33 HzÞ; 117:81 rad=s ð¼ 18:75 HzÞ:
The time-history responses computed for the displacement, velocity and acceleration of the
vehicle passing through the bridge at speed 10 m/s considering either a single mode or two modes
have been plotted in Figs. 13–15, together with the solutions obtained from the finite element
analysis. The first observation is that the solution obtained by considering only the first mode is
very close to that obtained by considering two modes, indicating that the single-mode approach is
reliable for simulating the vehicle response, and that the high modes can generally be neglected.
The discrepancy between the present analytical solutions and the finite element solutions is owing
to the fact that iteration was not performed in this study for updating the interaction forces
existing between the moving vehicle and bridge, in addition to the assumption of a small mass

0.0004
Multiple Bridge Modes
Mode 1 Only
0 Mode 1 & Mode 2
FEM
Displacement (m)

-0.0004

-0.0008

-0.0012

-0.0016

0 0.5 1 1.5 2 2.5

Time (sec)
Fig. 13. Displacement response of vehicle.
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 221

0.004

0.002
Velocity (m/s)

-0.002

Multiple Bridge Modes

-0.004 Mode 1 Only
Mode 1 & Mode 2
FEM
-0.006

0 0.5 1 1.5 2 2.5

Time (sec)
Fig. 14. Velocity response of vehicle.

0.06
Multiple Bridge Modes
Mode 1 Only
0.04 Mode 1 & Mode 2
FEM
Acceleration (m/s2)

0.02

-0.02

-0.04

0 0.5 1 1.5 2 2.5

Time (sec)
Fig. 15. Acceleration response of vehicle.

ratio for the vehicle. Such a discrepancy is generally negligible when observed from the frequency
domain, as will be discussed below.
The Fourier transforms of the acceleration response of the vehicle obtained by the present
analytical approach and the finite element method have been plotted in Figs. 16 and 17,
respectively, in which several peaks can be clearly seen. An observation is that the frequencies
 ARTICLE IN PRESS

222 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

0.025
3
Amplitude of Acceleration (m/s2) Spectrum of Analytical Solution
0.02
πv
1 Driving freq.
L
2 Bridge 1st freq. (-) ω b1 − π v
0.015 L
st v
3 Bridge 1 freq. (+) ω b1 +
2 L
4 Vehicle freq. ω v
0.01 2π v
4 5 Bridge 2nd shifted freq. ω b2 − L
2π v
ω b2 +
L
1
0.005
5

0
0.1 1 10 100
Frequency (Hz)
Fig. 16. Spectrum of vehicle acceleration response (analytical).

0.025
Amplitude of Acceleration (m/s2)

3 Spectrum of FEM
0.02
πv
1 Driving freq. L

2 Bridge 1st freq. (-) ω b1 − π v

0.015 L
st πv
2 3 Bridge 1 freq. (+) ω b1 +
L
4 Vehicle freq. ω v
2π v
0.01 nd ω b2 −
5 Bridge 2 shifted freq. L
2π v
ω b2 +
L
4

0.005 1
5

0.1 1 10 100
Frequency (Hz)
Fig. 17. Spectrum of vehicle acceleration response (FEM).

corresponding to all the peaks in the two figures appear to be in good agreement, indicating that
the use of the single-mode approach can yield sufficiently accurate results, particularly when the
frequency distribution is of concern.
In each figure, the first peak (counted from the left hand side) relates to the driving frequency,
pv=L ¼ 0:2 Hz: The frequencies associated with the second and third peaks relate to the
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 223

0.0016
Driving Frequency Vehicle Speed 10 m/s
Amplitude of Velocity (m/s) Closed-form Solution (No Damping)
Numerical Solution (5% Damping)
0.0012
Bridge Frequency

0.0008

0.0004

0.01 0.1 1 10 100 1000

Frequency (Hz)
Fig. 18. Effect of bridge damping on bridge velocity spectrum.

fundamental frequency of the bridge with shift, i.e., ob1  pv=L ¼ 1:88 Hz and ob1 þ pv=L ¼
2:28 Hz: By removing the shifting effect, the fundamental frequency of the bridge can be recovered
as ob1 ¼ 2:08 Hz: The frequency identified from the figures for the fourth peak is the vehicle
frequency, i.e., ov ¼ 3:25 Hz: Of interest are the two small peaks on the right-hand side of the
figures, which are associated with the second frequency of the bridge with shift.
To evaluate the effect of damping, we assume the bridge damping to be of the Rayleigh type
and assign a 5% damping ratio to each of the first two modes of the bridge. The midpoint velocity
spectrum of the bridge solved by the finite element method has been compared with the closed-
form solution for the zero damping case in Fig. 18. As can be seen, the amplitude of the bridge
frequency decreases slightly due to the existence of damping, while basically no change can be
observed on the magnitude of the vehicle frequency. It is concluded that both the driving and
bridge frequencies can be clearly identified regardless of the presence of damping on the bridge.
As for the vehicle damping, a 10% damping ratio is assigned to the vehicle. Fig. 19 shows the
vehicle acceleration spectrum computed for this case by the finite element method, along with the
closed-form solution obtained for the case of zero vehicle damping. Although the existence of
vehicle damping has caused a significant drop in the vehicle frequency component, its influence on
the bridge frequencies is generally small. Clearly, the bridge frequency can be extracted with no
difficulty from the vehicle response, notwithstanding the existence of the vehicle damping.
It should be noted that all the above results have been based on the assumption of smooth
pavement surface for the bridge. Should this factor be taken into account, it is likely that some
new peaks will be generated by irregularities in the pavement. Under such a situation, it is
suspected that the third peak associated with the fundamental frequency, ob1 þ pv=L ¼ 2:28 Hz;
can still be observed. Before closing this section, we like to say that further research is required in
this regard, since it presents lots of advantages if the bridge frequency can be extracted indirectly
from the acceleration response of a passing vehicle, compared with the conventional approaches
of measuring the bridge frequencies directly from the bridge response.
 ARTICLE IN PRESS

224 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

0.025
3 Vehicle Speed 10 m/s
Amplitude of Acceleration (m/s2) Closed-form Solution (No Dampling)
0.02 FEM Solution (10% Damping)
πv
1 Driving freq. L
πv
2 Bridge 1st freq. (-) ω b1 − L
0.015 st πv
2 3 Bridge 1 freq. (+) ω b1 + L
4 Vehicle freq. ω v
2π v
0.01 nd ω b2 −
4 5 Bridge 2 shifted freq. L
2π v
ωb2 +
L
1
0.005
5

0
0.1 1 10 100
Frequency (Hz)
Fig. 19. Effect of vehicle damping on vehicle acceleration spectrum.

5. Concluding remarks

In this study, a modal superposition approach was adopted to derive the solutions for a
vehicle–bridge system, with the vehicle simulated as a sprung mass and the bridge as a simply
supported beam. Such an approach is approximate in that no iteration was conducting for
updating the interaction force existing between the two subsystems. The solutions obtained,
however, were shown to be quite accurate compared with those obtained by the finite element
method. For both the bridge and vehicle responses, it is concluded that sufficiently accurate
solutions can be obtained by considering only the first mode of vibration of the bridge. As for the
bridge, the displacement response is predominated by the driving frequency component; the
velocity is equally governed by the driving frequency and bridge frequency components; and the
acceleration predominated by the bridge frequency component. There exists a possibility, at least
theoretically, that the vehicle speed can be identified from the displacement or velocity spectrum
of the bridge as it appears as a low-frequency peak. Further investigation is required in this
regard. As for the vehicle response, there are three groups of frequencies involved, i.e., the driving
frequency, vehicle frequency, and bridge frequency with shift. For the case where zero damping
and smooth pavement surface are assumed for the bridge, all the three groups of frequencies can
be identified from the vehicle response. Of interest is the fact that the shifted fundamental
frequency of the bridge appears as the highest peak in the spectrum. If the effects of damping and
irregular pavement are taken into account, it is suspected that the fundamental frequency of the
bridge can still be identified from the vehicle response. Both further theoretical and experimental
studies are required in this regard.
 ARTICLE IN PRESS

Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226 225

Acknowledgment

The research reported herein was supported in part by the National Science Council of the
ROC through the grant No. NSC91-2211-E-002-018.

References

[1] R. Willis, Appendix to the Report of the Commissioners Appointed to Inquire into the Application of Iron to Railway
Structures, H.M. Stationary Office, London, UK, 1849.
[2] G.G. Stokes, Discussion of a differential equation relating to the braking of railway bridges, Transactions of the
Cambridge Philosophical Society 8 (5) (1849) 707–735.
[3] C.E. Inglis, A Mathematical Treatise on Vibration in Railway Bridges, Cambridge University Press, Cambridge,
UK, 1934.
[4] S.P. Timoshenko, On the forced vibrations of bridges, Philosophical Magazine, Series 6 43 (1922) 1018–1019.
[5] C.P. Tan, S. Shore, Response of horizontally curved bridge to moving load, Journal of Structural Engineering 94
(9) (1968) 2135–2151.
[6] L. Frýba, Vibration of Solids and Structures under Moving Loads, Noordhoff International, Groningen,
Netherlands, 1972.
[7] M. Olsson, On the fundamental moving load problem, Journal of Sound and Vibration 145 (2) (1991) 299–307.
[8] Y.B. Yang, J.D. Yau, L.C. Hsu, Vibration of simple beams due to trains moving at high speeds, Engineering
Structures 19 (11) (1997) 936–944.
[9] H.H. Jeffcott, On the vibration of beams under the action of moving loads, Philosophical Magazine, Series 7 8
(1929) 66–97.
[10] E.C. Ting, J. Genin, J.H. Ginsberg, A general algorithm for the moving mass problem, Journal of Sound and
Vibration 33 (1) (1974) 49–58.
[11] M.M. Stanišić, On a new theory of the dynamic behavior of the structures carrying moving masses, Ingenieur-
Archiv 55 (1985) 176–185.
[12] S. Sadiku, H.H.E. Leipholz, On the dynamics of elastic systems with moving concentrated masses, Ingenieur-
Archiv 57 (1987) 223–242.
[13] J.E. Akin, M. Mofid, Numerical solution for response of beams with moving mass, Journal of Structural
Engineering 115 (1) (1989) 120–131.
[14] H.P. Lee, Dynamic response of a beam with a moving mass, Journal of Sound and Vibration 191 (2) (1996) 289–294.
[15] M.A. Foda, Z. Abduljabbar, A dynamic Green function formulation for the response of a beam structure to a
moving mass, Journal of Sound and Vibration 210 (3) (1998) 295–306.
[16] J.M. Biggs, Introduction to Structural Dynamics, McGraw-Hill, New York, 1964.
[17] Y.B. Yang, J.D. Yau, Vehicle–bridge interaction element for dynamic analysis, Journal of Structural Engineering
123 (11) (1997) 1512–1518 (Errata: 124(4), 479).
[18] Y.B. Yang, Y.S. Wu, A versatile element for analyzing vehicle–bridge interaction response, Engineering Structures
23 (2001) 452–469.
[19] A.V. Pesterev, B. Yang, L.A. Bergman, C.A. Tan, Response of elastic continuum carrying multiple moving
oscillators, Journal of Engineering Mechanics 127 (3) (2001) 260–265.
[20] K.H. Chu, V.K. Garg, T.L. Wang, Impact in railway prestressed concrete bridges, Journal of Structural
Engineering 112 (5) (1986) 1036–1051.
[21] P.K. Chatterjee, T.K. Datta, C.S. Surana, Vibration of suspension bridges under vehicular movement, Journal of
Structural Engineering 120 (1994) 681–703.
[22] G.H. Tan, G.H. Brameld, D.P. Thambiratnam, Development of an analytical model for treating bridge–vehicle
interaction, Engineering Structures 20 (1–2) (1998) 54–61.
[23] M. Fafard, M. Laflamme, M. Savard, M. Bennur, Dynamic analysis of existing continuous bridge, Journal of
Bridge Engineering 3 (1) (1998) 28–37.
 ARTICLE IN PRESS

226 Y.B. Yang, C.W. Lin / Journal of Sound and Vibration 284 (2005) 205–226

[24] Y.B. Yang, C.H. Chang, J.D. Yau, An element for analysing vehicle–bridge systems considering vehicle’s pitching
effect, International Journal for Numerical Methods in Engineering 46 (1999) 1031–1047.
[25] H. Xia, Y.L. Xu, T.H.T. Chan, Dynamic interaction of long suspension bridges with running train, Journal of
Sound and Vibration 237 (2) (2000) 263–280.
[26] J.D. Yau, Y.B. Yang, S.R. Kuo, Impact response of high speed rail bridges and riding comfort of rail cars,
Engineering Structures 21 (9) (1999) 836–844.
[27] Y.B. Yang, C.W. Lin, J.D. Yau, Extracting the bridge frequencies from the dynamic response of a passing vehicle,
Journal of Sound and Vibration 272 (2003) 471–493.
[28] Y.B. Yang, B.H. Lin, Vehicle–bridge interaction analysis by dynamic condensation method, Journal of Structural
Engineering 121 (11) (1995) 1636–1643.

View publication stats