Wind and Structures, Vol. 14, No.

4 (2011) 359-381

359

Simplified formulations for flutter instability analysis of

bridge deck
Tan-Van Vu1, Young-Min Kim2, Tong-Seok Han3 and Hak-Eun Lee*4
1.4

School of Civil, Environmental and Architectural Engineering, Korea University, Seoul, South Korea
2
Institute of Construction Technology, DAEWOO E&C Co., Ltd., Suwon, South Korea
3
School of Civil and Environmental Engineering, Yonsei University, Seoul, South Korea
(Received May 11, 2010, Accepted January 8, 2011)

Abstract. This paper deals with the flutter instability problem of flexible bridge decks in the
framework of bimodal-coupled aeroelastic bridge system analysis. Based on the analysis of coefficients of
the polynomials deduced from the singularity conditions of an integral wind-structure impedance matrix, a
set of simplified formulations for calculating the critical wind velocity and coupled frequency are
presented. Several case studies are discussed and comparisons with available approximated approaches are
made and presented, along with a conventional complex eigenvalue analysis and numerical results. From
the results, it is found that the formulas that are presented in this study are applicable to a variety of
bridge cross sections that are not only prone to coupled-mode but also to single-mode-dominated flutter.
Keywords: bridges; eigenvalue; flutter; flutter derivatives; instability; simplified formulations.

1. Introduction
Long-span bridges are slender, low damping and flexible large-scale line-like structures that are
very susceptible to wind loads. The excitation mechanisms for aeroelastic interaction between wind
flow and structures may be distinguished as follows: extraneous-flow-induced excitations, flowinstability-induced excitations and movement-induced excitations. The latter excitation mechanism
is caused by fluctuating wind forces due to movements of the vibrating structural part. Small
deviations from the equilibrium position of the structure induce a re-distribution of impacting wind
forces, which further increase the initial disturbances. If these self-excited forces lead to a negative
damping threshold, the onset of flutter will occur and this directly induces structural failure, as was
the case in the collapse of the Tacoma Narrows Bridge in 1940. In streamlined bodies such as
airfoils, flutter occurs from a coupling of simultaneous vertical and torsional motion of which the
frequencies are in phase with one another. This type of flutter is a two degree-of-freedom or
coupled flutter, which was the cause of collapse of the Samuel Browns Brighton Chain Pier. For a
bluff body such as a plate girder, flutter occurs from a single degree-of-freedom, which is heavingmode aeroelastic instability or torsional flutter. Heaving-mode aeroelastic instability was the cause
of wild oscillation of the Deer Isle Bridge in 1942, while torsional flutter was the cause of the
damage that occurred in 1826 to Telfords Menail Bridge (Scott 2001).
* Corresponding Author, Professor, E-mail: helee@korea.ac.kr

360

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

Since these events, the importance of flutter analysis of long-span bridges has been strongly
recognized and has led to many research works and investigations on bridge aerodynamics. Early
pioneering work on bridge flutter analysis was presented by Bleich (1949), where stiffness-driven
flutter was addressed using airfoil aerodynamics theory. However, a wind speed with a value higher
than that, which occurred for the Tacoma disasters, was given in this approach. While Selberg
(1961) and Rocard (1963) proposed simplified empirical formulas for estimating flutter onset
velocity, these formulas nevertheless are only rigorously applied for a flat plate section.
From the point of view of aerodynamics, Scanlan and Tomko (1971) mathematically described the
self-excited forces in terms of flutter derivatives in the frequency-domain and offered an analysis on
bridge flutter. Recently, the aerodynamic instability problem of long-span bridges can be analyzed
by linear and nonlinear flutter approaches. Normally, the geometric nonlinearity of bridge structures
and effects of nonlinear wind-structure interaction are neglected in the linear flutter approach.
Nonlinear flutter analysis can be performed on the deformed configuration of bridge structures
under the static wind action as suggested by previous authors (e.g., Zhang 2006).
In frequency domain, the linear flutter analysis methods can be grouped into two broad categories
as follows: the mode-by-mode and multimode-coupled approach. The multimode-coupled approach
is often performed by using aerodynamic coupling of fundamental vertical bending and torsional
modes with secondary contributions from other modes. Hence, bimodal-coupled flutter analysis
consisting of two fundamental modes remains a useful tool for an expeditious evaluation of bridge
flutter performance at the preliminary design stage.
In the framework of binary flutter analysis, Nakamura (1978) suggested a set of approximated
formulas based on theoretical assumptions about the unsteady aeroelastic behavior of bridge decks,
in order to provide the frequency and rate of growth of oscillation, the position of the equivalent
center of rotation and the phase difference between bending and torsion near the critical flutter point
in the case of a variety of deck sections. Subsequently, Matsumoto (1999) proposed a step-by-step
iterative method attempting to highlight the characteristics of the torsional and heaving branch at
flutter onset. Based on closed-form solutions while assuming low-level damping, Chen and Kareem
(2007) introduced an approximated formula of which only H3*, A1*, A2* and A3* flutter derivatives
are needed to estimate the critical wind speed. From a manipulated and simplified process of the
equation of the eigenvalue problem, Bartoli and Mannini (2008) presented analytical expressions for
calculating critical frequency and reduced wind speed of which only H1*, A2*and A3* flutter
derivatives are used, and these were also applied in the case of a section prone to torsional flutter.
Accordingly, the approximated formulas can be successfully employed for capturing the physics of
flutter mechanisms such as stiffness-driven flutter or damping-driven mechanisms for the given
bridge deck section. Nevertheless, the limitation of these approaches is their inability to give
reasonable results when the frequency ratio tends to unity and even in zero-damping condition, as
well as in the case of cross sections that are prone to heaving-mode aeroelastic instability.
In the present paper, using an analysis process for coefficients of the polynomials deduced from
singularity conditions of an integral wind-structure impedance matrix, a simplified formulation is
presented in terms of flutter derivatives for calculating the critical wind velocity and coupled
frequency. Because the uncoupled and coupled flutter derivatives remain in the simplified formulas,
this considered framework seems to give reasonable results even if the frequency ratio is very close
to unity in the case of low damping. Furthermore, if the coefficients of non-dimensional mass and
moment of inertia associated with flutter derivatives were considered small or negligible, the
simplified formulas, which only consist of fully uncoupled flutter derivatives (H1*, H4* and A2*, A3*)

Simplified formulations for flutter instability analysis of bridge deck

361

extracted from the experimental system of 1-DOF sectional models, would no longer be tedious.
However, the proposed simplified formulation is still able to be effectively applied for a bridge
section that is not only prone to coupled-mode flutter but also single-mode instability (either
dominated by torsion or heaving mode).
In the second section of this paper, a mechanical model is introduced to deal with the bimodal
flutter problem and its solution is reviewed. Subsequently, simplified strategies are introduced, as
the results of the approximated formulations consisting of the flutter derivatives are constructed in
the third section. Finally, some results of numerical examples showing good agreement with those
obtained via other approximated formulas as well as with the conventional complex eigenvalue
analysis (CEA).

2. Bimodal approach to flutter problem

The natural circular frequencies of the first vertical bending and torsional modes in the aeroelastic
bridge system are denoted as z and , respectively. Only aerodynamic forces generally dominate
the aerodynamic performance of the bridge. The self-excited lift and pitching moment acting on the
bridge deck section per unit length are given by Simiu and Scanlan(1996)
2

V B
* z
* B
ae
2 *
2 *z
L ( t, K ) = ------------- KH 1 --- + KH 2 ------- + K H 3 + K H 4 --2
V
V
B
2

V B
* z
* B
ae
2 *
2 *z
M ( t, K ) = ------------- KA 1 --- + KA 2 ------- + K A 3 + K A 4 --2
V
B
V

(1)
(2)

where z and are the vertical and torsional displacements, V = mean wind speed, B = bridge deck
*
*
(Fig. 1), = air density, the reduced frequency K = 2k = B/V, and H i , A i ( i = 1, 4 ) = flutter
derivatives, which are functions of reduced frequency and can be extracted from a wind-tunnel test.
The governing equations of the combined bridge system in terms of the generalized modal
coordinates under approaching crosswind flow are expressed as
ae

M + C + K = F

Fig. 1 Sign convention for the displacements and self-excited forces

(3)

362

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

where M, C, K are generalized modal mass, damping and stiffness matrices, respectively; Faeis the
generalized self-excited force vectors and the over-dot denotes partial differentiation with respect to
time. Taking the Fourier transform on both sides of Eq. (3) leads to Eq. (4), which is considered as
the modal equilibrium equation of motion in the frequency domain (Strmmen 2006),
2

ae

ae

( M + iC + K )X ( ) = ( iC + K )X ( )

(4)

where i2 = -1, matrices of Cae and Kae contain the coefficients that are required for the vertical and
torsional motions. These coefficients are normalized by 2/2 and B4/2, wherein is the inwind frequency dependent on mean wind velocity (V)
C

ae

B H 1 BH 2
B H 4 BH 3
ae
= ------------, K = ---------------2
2
2 *
2 *
*
*
BA 1 B A 2
BA 4 B A 3

(5)

The impedance matrix (Strmmen, 2006) can be obtained by,

[ S ( r, V ) ] =

ae

1 zz

ae

ae

ae

z 1

r
---- z
0

0
r
----

+ i2

-----r 0
z

0 ------r

ae

ae

z zz z
ae

(6)

ae

With
2
B 2 *
ae
zz = --------- -----r H 4 G z z, z,
z z
2m
4
B r 2 *
ae
= --------- ------ A 3 G , ,

2m
3
B
ae
*
z = --------- ------r A 1 G z, ,

4m

3
3
B 2 *
B r 2 *
ae
ae
z = --------- -----r H 3 G , , z = --------- ------ A 4 G z, z
z z

2m
2m
2
B r *
ae
zz = --------- ----- H 1 G z z, z ,
z z
4m
4
B
ae
*
= --------- ------r A 2 G , ,

4m

3
B r *
ae
z = --------- ----- H 2 G z , z
z z
4m

G m n, p =

L exp

m n dx / p dx
L

z and m
are the modal mass associated
where G m n, p [ m, n, p = ( z, ) ] are modal integrals; m
with the vertical and torsional modes; L and Lexp are length and wind exposed lengths. The effects
ae
ae
of xy and xy ( x, y = z, ) change the damping and stiffness properties of the combined bridge
system, respectively.
Introducing the simplifications of Lexp = L and z = , any stability limit can be found by setting
the determinant of the coefficient matrix of Eq. (6) equal to zero, resulting in cubic or quartic
polynomials in the real or imaginary parts with respect to the in-wind frequency ratio as follows,
z * *
1
*
*
* *
* *
* *
*
*
2
4
3
1 + --- ( z H 4 + A 3 ) + ---------( A 1 H 2 A 2 H 1 + A 3 H 4 A 4 H 3 ) + ( z H 1 + z A 2 )
2
4
1
*
2
2 *
2
1 + + 4 z + --- ( z H 4 + A 3 ) + 1 = 0
2

(7)

Simplified formulations for flutter instability analysis of bridge deck

z * *
*
*
* *
* *
* *
2 1
3
( H1 A3 H2 A4 H 3 A1 + H 4 A2 )
--- ( z H 1 + A 2 ) + ---------2
4
*
*
* *
2
2 1
2 *
2 z ----- A 3 + + -----z H 4 + 1 --- ( z H 1 + A2 ) + 2 ( z + ) = 0
2

363

(8)

2
z and = B 4 m
represent
where = z is the structural frequency ratio; z = B m
the non-dimensional mass and the polar moment of inertia, respectively; = r is the inwind frequency ratio for bridge sections prone to coupled flutter and is expressed as follows,

11

(9)

It can be seen that the solution of these equations requires searching for the lowest identical roots with
respect to , which do not easily allow practical calculation by hand, from both the fourth and third degree
polynomial. Moreover, if this analysis is performed, it is rather difficult to identify in which branch the
coupled flutter instability occurs. This is because the formula to determine the in-wind frequency ratio, for
a torsional and heaving branch is symmetric in terms of torsional and heaving motion.

3. Approximated solutions to bimodal flutter problem

3.1. Approximated 2-DOF flutter derivatives formulations
In order to simplify the bimodal flutter equation, let us set the coefficients of the fourth order
polynomial obtained by imposing the vanishing of the real part from the determinant of the
impedance matrix in Eq. (7) as follows
z
1
*
*
2
R 1 = 1 + --- ( z H 4 + A 3 ) + ---------C
2
4 R
*

(10a)

R 2 = ( z H 1 + z A 2 )

(10b)

1
*
2
2 *
R 3 = 1 + + 4 z + --- ( z H 4 + A 3 )
2

(10c)

where C R = H 1 A 2 + H 2 A 1 H 3 A 4 + H 4 A 3
For an analogous reason, Eq. (8) obtained by imposing the vanishing of the image part from the
determinant of the impedance matrix, the setting coefficients are as follows
z
*
*
2 1
C
I 1 = --- ( z H 1 + A 2 ) + ---------2
4 I

(11a)

*
*
2
I 2 = 2 z ----- A 3 + + -----z H 4 + 1
2

(11b)

1
*
2 *
I 3 = --- ( z H 1 + A 2 )
2

(11c)

364

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

I4 = 2 ( z + )
*

(11d)

where C I = H 1 A 3 H 2 A 4 H 3 A 1 H 4 A 2
Once the terms of I i = ( i = 1, , 4 ) and R j ( j = 1, , 3 ) are introduced, Eqs. (7) and (8) can
be rewritten as follows
3

I1 + I2 + I3 + I4 = 0
4

(12)

R1 + R2 + R3 + 1 = 0

(13)

In the design of cable-supported bridges, three basic section configurations bluff girder,
streamlined box and slotted girder sections, are widely used in the design (Yang and Ge 2009).
Hence, it is important to analyze dynamic and aerodynamic parameters from existing bridges with
three sections to derive the approximated formulas. Table 1 summarizes the data of four Edge girder
bridge sections with dimensions and cross sections taken into account: Busan-Geoge Bridge (Lee et
al. 2004), Seohae Grand Bridge (KHC 1998), Krkinen (Kiviluoma 2001), and Golden Gate (Simiu
and Scanlan 1996). Data of the streamline bridge sections is shown in the second part of the Table
1: Great Belt Bridge (Larsen 1993, Nissen et al. 2004), 2nd Geo-Germ Bridge (Larsen 2002), Hoga
Kusten Bridge (Livesey 1995), Tsurumi Fairway Bridge (Sarkar et al. 1992). At the end of the
table, three slotted girder bridge sections are considered: Messina Straits Bridge (DAsdia and Sepe
1998), Xihoumen Bridge (Yang et al. 2007) and Shanghai Bridge (Zhou and Ge 2009).
From the data gathered in the last three columns of Table 1, it can be seen that the value of nondimensional mass is smaller than that of the moment of inertia. Furthermore, the value of the free
term I4 in Eq. (12) is much smaller than unity. In addition to investigating the impact of terms in
Table 1 Geometric and dynamics properties with coefficient of I4 of the different bridges
z ( kg m ) m
( kgm 2 m ) z
B (m) fz (Hz)

z = m

I4

Edge girder bridge section

Busan-Geoge Grand 22.0

0.334

3.00

0.005

26,854

1,430,000

0.023 0.205 0.0400

Seohae Grand

34.0

0.250

1.84

0.005

28,789

2,756,000

0.050 0.606 0.0284

K a rkinen

13.28 0.4646 1.51

0.0064

17,193

306,218

0.013 0.127 0.0321

Golden Gate

27.5 0.1638 1.17

0.005

5,208

3,680,000

0.182 0.194 0.0217

Streamline bridge section

Great Belt

31.0

0.099

2.75

0.005

22,700

2,470,000

0.053 0.467 0.0375

2nd Geo-Germ

16.9

0.185

2.99

0.003

11,699

295,250

0.031 0.345 0.0240

Hoga Kusten

22.0 0.1198 2.10

0.005

10,588

603,209

0.057 0.485 0.0310

Tsurumi

38.0

2.38

0.005

32,220

2,880,100

0.056 0.905 0.0338

Messina Straits

60.4 0.0605 1.32

0.01

55,000

28,000,000

0.083 0.594 0.0463

Xihoumen

36.0 0.1005 2.31

0.01

27,511

4,002,800

0.059 0.525 0.0662

Shanghai

51.5 0.2520 2.64

0.005

35,151

10,076,000

0.094 0.873 0.0364

0.204

Slotted box girder bridge section

Simplified formulations for flutter instability analysis of bridge deck

365

Eqs. (12) and (13) on their own, let us take into account the aerodynamic behavior of two groups of the
bridge deck section, one of these groups consists of two bridge deck sections that are quite different from
each other: the thin flat plate (Theodorsen 1934) and the rectangular sections with width to depth ratio of
B/D = 15, R15 (Matsumoto 1996); another of these groups consisting of two bridge deck sections are
often used in super long bridges: section of 2TF and 2TFGP (Matsumoto et al. 2004).
For several levels of reduced wind velocity of V/fB, these terms that are gathered in groups are
calculated in the case of low (z = z = 0.0025) and high level damping ratios (z = = 0.01).
Furthermore, this investigation is also performed on sets of structural frequency ratios with the case
of the fundamental vertical bending frequency close to torsional frequency (case of Golden Gate
bridge, = 1.17 ) and vice versa (case of Krkinen bride, = 1.51 ), both of which are in the inwind frequency ratios of = 0.95 ; other sets of in-wind frequency ratios around the flutter onset
( = 0.80 and = 0.85) that satisfied the inequality (9) for the section of 2TF, 2TFGP with theirs
structural frequency ratio assumed are 1.72 and 1.30, respectively.
Finally, the results obtained from the mentioned cases are presented in the form of graphs in order to

Fig. 2 Variation in Ri, Ii of R15 with V/fB

(a) z = = 0.0025, = 1.17 , (b)z = = 0.0025, = 1.51,
(c) z = = 0.01, = 1.17 and (d) z = = 0.01, = 1.51

366

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

facilitate comparisons. For example, Fig. 2 for the rectangular section of R15 (z = 0.035, = 0.174);
Fig. 3 for the theoretical thin Flat Plate section (z = 0.027, = 0.163) ; Fig. 4 for the 2TF section
z = 0.023, = 0.143; Fig. 5 for the 2TFGP section (z = 0.027, = 0.163). It is evident that the
term R23 is definitely negligible with respect to the other terms in Eq. (13). The results from
figures show that the value of Eq. (12) is dependent on the sum of the terms I13 and I3 in case
of the low damping ratios. Nevertheless, several linear differences were observed in the case of a
high damping ratio when the structural frequency ratio is close to unity.
In practice, the hypothesis of low-level damping is generally implied in the modeling of selfexcited forces (Chen and Kareem 2007). In addition, much of the existing bridge is characterized by
frequency ratios far from the unity (Bartoli and Mannini 2008). Hence, Eqs. (12) and (13) were
rewritten as follows
3

0 = I1 + I2 + I3 + I4 I1 + I3
and

Fig. 3 Variation in Ri, Ii of Flat Plate with V/fB:

(a) z = = 0.0025, = 1.17 , (b) z = = 0.0025, = 1.51,
(c) z = = 0.01, = 1.17 and (d) z = = 0.01, = 1.51

(14)

Simplified formulations for flutter instability analysis of bridge deck

367

Fig. 4 Variation in Ri, Ii of Flat Plate with V/fB:

(a) z = = 0.0025, = 0.80, (b) z = = 0.0025, = 0.85,
(c) z = = 0.01, = 0.80 and (d) z = = 0.01, = 0.85
4

0 = R1 + R2 + R3 + 1 R1 + R3 + 1

(15)

From these equations, Eqs. (21) and (22), an expression for the in-wind frequency ratio 2 can be
found
2

( z H1 + A2 )
= --------------------------------------------------------------------------*
*
2
[ ( z H 1 + A 2 ) + 0.5 z C 1 ]
2

(16)

and
2

R 3 R 3 4R 1
2
1 = -----------------------------------2R 1

R 3 R 3 4R 1
2
2 = -----------------------------------2R 1

(17a, b)

368

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

Fig. 5 Variation in Ri, Ii of Flat Plate with V/fB:

(a) z = = 0.0025, = 0.80 , (b) z = = 0.0025, = 0.85,
(c) z = = 0.01, = 0.80 and (d) z = = 0.01, = 0.85

The graphically approximated solution of bimodal flutter problems can easily be found by the
intersection point of the 2 root defined by Eq. (16) and Eq. 17(a) or Eq. 17(b) within the
respective range of interesting of reduced wind speed, V/fB . It is clear that the curve of 12 defined
the heaving branch coupled flutter, otherwise the curve of 22 defined the torsional branch coupled
2
2
flutter. In addition, the intersection point between the curve of [ i ] ( i = 1, 2 ) and the reduced
frequency axis V/fB with the lowest value of Vc / fcB defines the critical flutter wind speed. Fig. 6
illustrates the heaving and torsional branch coupled flutter characteristics of the slotted girder bridge
sections of 2TF and 2TFGP with an angle of attack of 30 (Matsumoto et al. 2004). It also showed
the frequency ratio effects on flutter instability that were obtained by changing the natural frequency
of the torsion mode.
On the other hand, it is easy to extract the value of 2 from Eq. (15), then substitute its value in
* *
* *
Eq. (16) and by the aid of Eqs. 10(a) and (c) and ij = ( H i A j H j A i ) ( i, j = 1, 2, 3, 4 ) , after
rearranging some terms, leads to the following expression that could be used to determine the

Simplified formulations for flutter instability analysis of bridge deck

369

Fig. 6 Coupled flutter analysis for slotted girder section: (a) 2TF and (b) 2TFGP

critical reduced wind speed

*

8 z ( A 2 + H 1 z ) [ 2 ( A 2 + H 1 z ) + z C 1 ]
2

2 2

+ z [ ( H 2 A 4 + H 3 A 1 )C I 2H 1 A 2 C R ] +
2

+ 2 z ( 1 )
3

2H 1 A 2 ( 1 ) + ( H 2 A 4 + H 3 A 1 ) ( H 1 z A 2 )
*

+ 2 H1 A2 ( A3 z H4 )
*

+ z A 2 [ A 2 12 + A 3 13 + A 4 32 ] + z H 1 [ H 1 12 + H 3 14 + H 4 42 ] = 0

(18)

Eq. (18) consists of coupled and uncoupled flutter derivatives that are extracted from the
experimental system of 2-DOF sectional models and aerodynamic parameters of the given bridge
section. Within the range of interest of reduced velocity, approximated solutions for the critical
reduced wind speed and frequency can be found from Eqs. (16) and (18). The results for the four
section cases previously investigated are gathered in Table 2, which shows the accuracy of the
approximated formulas in estimating the flutter onset frequency and reduced wind speed in the case
of low, moderate and high structural damping ratios. The structural frequency ratio assumed for case
of the rectangular R15(z = 0.035, = 0.174) and Flat Plate (z = 0.027, = 0.163) are 1.51 and
1.17, respectively.
It is evident that the results from the approximated formulas are in good agreement in terms of
critical frequency solutions with those of the CEA approach. The degree of difference between these
proposed formulas and the CEA approach depends on the levels of damping ratio. When the
frequency ratio is small and the damping levels are high, as expected, the accuracy of the
approximation is inferior. Finally, these approximated formulas give under-predicted results in
critical wind speeds.
3.2. Approximated 1-DOF flutter derivatives formulations
A simplified strategy was introduced based on investigating the influence of individual terms on

370

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

Table 2 Comparison between the solution of the approximated 2-DOF flutter derivatives formulations and that
of CEA
Onset flutter
Error

Damping ratio

appox
V rc
approx
fc
CEA
V rc

z = = 0.0025

CEX

fc

V c (%)
f c (%)

Flutter derivative set

Flat Plate

R15

2TF

2TFGP

15.38

40.81

47.38

33.31

0.0628

0.1166

0.0762

0.0560

16.06

41.78

47.69

34.35

0.0625

0.1159

0.0760

0.0558

-4.45

-2.36

-0.66

-3.10

0.45

0.62

0.24

0.38

appox
V rc
approx
fc
CEA
V rc

15.41

40.82

47.38

33.33

0.0627

0.1166

0.0762

0.0560

16.71

42.60

47.98

35.29

CEX
fc

0.0622

0.1153

0.0759

0.0555

V c (%)

-8.44

-4.36

-1.26

-5.88

f c (%)

0.82

1.12

0.46

0.81

15.50

40.87

47.41

33.40

0.0627

0.1166

0.0762

0.0560

17.96

44.02

48.55

36.92

0.0618

0.1144

0.0756

0.0550

-15.81

-7.72

-2.41

-10.55

1.45

1.89

0.87

1.69

z = = 0.005

appox

V rc

approx
fc
CEA
V rc
CEA
fc

z = = 0.010

V c (%)
f c (%)

both the imaginary and real parts at certain in-wind frequency ratios when using different damping
levels and within a large range of reduced velocities. This strategy leads to a presented
approximated method as a result of the previous part.
In this section, we investigate terms related to parameters I1 from the imaginary part and R1 from
the real part, where coupled and uncoupled flutter derivatives are combined with the ingredients of
0.25 2 CI and 0.25 2 CR.
For the several reduced wind speeds V/fB, the terms appearing in the I1and, R1 are compared in
Table 3 for the Flat Plate section, and in Table 4 for the rectangular section R15. It is clear that the
coefficients of 0.25 2CR combined with coupled and uncoupled flutter derivatives, are neglected
*

with respect to [ 1 + 0.5 ( h H 4 + A 3 ) ] . Hence, the value of term R1 in Eq. 10(a) becomes
R1 =

z
1
1
*
*
*
*
2
C 1 + --- ( z H 4 + A 3 )
1 + --- ( z H 4 + A 3 ) + ---------2
2
4 R

(19)

Simplified formulations for flutter instability analysis of bridge deck

371

It is also shown that the coefficients of 0.252CI combined with coupled and uncoupled flutter derivatives
*
*
2
are neglected with respect to 0.5 ( h H 1 + A 2 ) . Hence, the value of term I1 in Eq. 11(a) becomes
2
z

*
*
*
*
2 1
C I ----- ( z H 1 + A 2 )
I 1 = --- ( z H 1 + A 2 ) + ---------2
4
2

(20)

Therefore, Eq. (16) can be rewritten as

2

I
( z H1 + A2 )
= ----3 = -----------------------------------------*
*
2
I1
( z H1 + A2 )
2

(21)

Substituting Eq. (21) into Eq. (15) while taking into account Eqs. 10(c) and (19), and after
rearranging some terms, leads to the simple equation as follows
Table 3 Variation of terms related to parameters I1, R1 with V / fB for Flat Plate ( = 1.17, z = 0.027, = 0.163 )
V / fB 0.25 2CI

0.5 ( z H 1 + A 2 )

0.252CR [ 1 + 0.5 ( z H 4 + A 3 ) ]

0.0000

0.000

0.000

0.0000

1.368

0.0001

-0.022

-0.022

-0.0004

1.377

0.0002

-0.041

-0.041

-0.0011

1.396

0.0005

-0.066

-0.065

-0.0020

1.428

0.0009

-0.095

-0.095

-0.0032

1.475

11

0.0012

-0.129

-0.128

-0.0046

1.537

13

0.0016

-0.167

-0.165

-0.0060

1.615

15

0.0019

-0.208

-0.206

-0.0073

1.711

17

0.0022

-0.252

-0.250

-0.0085

1.823

19

0.0024

-0.299

-0.296

-0.0095

1.953

21

0.0025

-0.348

-0.346

-0.0102

2.100

R1

1.368
1.377
1.395
1.426
1.471
1.532
1.609
1.703
1.814
1.943
2.090

Table 4 Variation of terms related to parameters I1, R1, with V / fB for ( = 1.51, z = 0.035, = 0.174 )
V / fB

0.25 2CI

0.0000

3
5

0.25 2CR

[ 1 + 0.5 ( z H 4 + A 3 ) ]

R1

0.000

0.000

0.0000

2.278

2.278

-0.0017

-0.094

-0.096

-0.0013

2.347

2.346

-0.0037

-0.172

-0.175

-0.0027

2.438

2.435

-0.0055

-0.261

-0.267

-0.0038

2.563

2.559

-0.0079

-0.361

-0.369

-0.0032

2.712

2.709

0.5 ( z H 1 + A 2 )

11

-0.0117

-0.470

-0.482

0.0002

2.888

2.888

13

-0.0013

-0.614

-0.615

0.0063

3.194

3.201

15

0.0229

-0.780

-0.757

0.0154

3.610

3.625

17

-0.0031

-0.952

-0.955

0.0411

4.168

4.209

19

-0.0147

-1.125

-1.139

0.0563

4.710

4.767

21

0.0078

-1.297

-1.289

0.0377

5.219

5.257

372

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

*

( 1 ) z [ ( H 1 z + A 2 ) ( A 3 H 1 H 4 A 2 ) + 2H 1 A 2 ( 1 ) ]
*

+ 8 z ( A 2 + H 1 z ) ( A 2 + H 1 z ) = 0

(22)

Eqs. (21) and (22) can be used to estimate the approximated solutions of the critical frequency and
reduced wind speed of the combined bridge system. Eq. (21) is the same as the simplified formula for
critical frequency proposed by Bartoli and Mannini (2008), which is based on the assumption that the
damping ratio is neglected for both bending and torsional modes. These approximated formulas depend
on the dynamic parameters of structural bridge: damping ratios i(i = z, ), structural frequency ratio ,
non-dimensional mass, moment of inertia z, , and the aerodynamic parameters: uncoupled flutter
derivative Ai (i = 2,3) and Hi (i=1,4) of the given deck bridge section.
The approximations of coefficients R1 and R2 are given by Eqs. (19) and (20), respectively. They
are considered as reasonable simplifications, if terms CR and CI are small amounts to neglect. It can
be obtained by utilizing four sets of inter-relationships between coupled flutter derivatives and
uncoupled flutter derivatives, which are validated throughout physical measurements for rectangular
cylinders as presented in Matsumoto (1996)
*

H 1 kH 3 ,
*
A1

*
kA 3 ,

H4 k H2

(23a, 23b)

*
A4

(23c, 23d)

*
k A2

Scanlan et al. (1997) shown that the expressions in Eqs. 23(a)-(d) are approximate results for the
*
airfoil case, and some streamlined bridge decks if contribution of A 1 is very small. Once reductions
of the coupled flutter derivative are made in terms R1 and R2, Eq. (22) can be derived by equivalent
substitutions in the previous simplified flutter equation of real part.
From the aerodynamic point of view, it is noted that Eq. (22) can obtained consistently with the
physical phenomenon. Thus, it is not directly derived by setting to zero all the coupled flutter
derivatives. Although, all damping and stiffness flutter derivatives still remain in this equation, so it
can be concluded that the proposed approximate solutions of critical reduced wind speed for the
cross sections are prone to not only the coupled-mode, but also the single-mode flutter (either
dominated by torsion or heaving mode).

4. Numerical validations
In this section, numerical computations of the critical reduced wind speed and frequency are
performed for a variety of deck sections to validate the degree of approximation as well as to point
out the limitations of the approximated approaches that are based on the 1-DOF and 2-DOF flutter
derivatives for the combined bridge system. Subsequently, solutions of the CEA for flutter stability
are compared between the solutions from Eqs. (16) and (18) [Approx. 2-DOF forms] and Eqs. (21)
and (22) [Approx. 1-DOF forms].
4.1. Coupled-mode flutter simulations
Thin flat plate section and the rectangular section of R15 are prone to coupled-mode flutter. Sensitivity

Simplified formulations for flutter instability analysis of bridge deck

373

analyses of onset flutter are carried out at both sections for different damping levels and various
frequency ratios. Besides that, the results of critical wind speed given by the Selberg (1961) and Rocard
(1963) empirical formulas are regarded as reference solutions for case of thin flat plate section.
Results of this analysis are illustrated in Fig. 7 for the rectangular section of R15 (z = 0.027, = 0.163) and
Fig. 8 for the classical Flat Plate (z = 0.035, = 0.174) section. The first conclusion drawn from
these graphs for both approximated formulas gives results, which are very close to those obtainable
from the solution of the traditional CEA for a wide range of frequency ratios both in the case of
structural zero-damping as well as moderate structural damping ( z = = 0.005).
It also shows that when the structural torsional frequency reaches close to the vertical bending
frequency, both CEA and Approx. 2-DOF forms shows the critical reduced wind speeds, that tend
to become high (Fig. 7) even if approaching infinity (Fig. 8), depending on the type of cross section
and structural damping levels. These results are in good agreement with those reported by Dyrbye
and Hansen (1997) and Bartoli and Mannini (2008) for the flat plate and rectangular section.
Meanwhile, the Approx. 1-DOF forms give opposite results from those of the Approx. 2-DOF
forms and the CEA.
Finally, for a thin flat plate the results given by the proposed formula are very similar to those
calculated by the empirical Selberg (1961) and Rocard (1963) formulas, as can be seen in Fig. 8,
these empirical formulas do not give the infinity critical wind speed when the frequency ratio, is
very close to unity. However, when the structural torsional frequency is far from vertical bending
frequency, all formulas give the same results.
In addition, to validate the accuracy and effectiveness of the proposed approximated approaches,
Nine case studies are taken into account, two of which that are reported in Table 5 originate from
the existing bridges: case study 3 for the 2nd Geo-Germ Bridge (Larsen 2002) and case study 8 for
the Busan-Geoge Grand Bridge (Lee et al. 2004). The other case studies show the dynamic
parameters of bridges listed in Table 1 and the flutter derivatives set measured from the cross
sections prone to the coupled-mode flutter. The geometric and aerodynamic properties of all case
studies are listed in Table 5. The solution of critical wind speeds and frequencies calculated from
approximated formulations are compared with those of CEA, while approximated results calculated
from the Chen and Kareem (2007) and Bartoli and Mannini (2008) formulas are regarded as

Fig. 7 Comparison of flutter onset of R15 for varying of the structural frequency ratio:
(a) critical reduced wind speed and (b) critical frequency

374

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

Fig. 8 Comparison of flutter onset of Flat Plate for varying of the structural frequency ratio:
(a) critical reduced wind speed and (b) critical wind speed

reference solutions. Finally, all results of these formulas are gathered in the Table 6.
From the results, it can be concluded that both proposed approximated formulations show
solutions that are close to those given by the complex eigenvalue analysis, particularly in the flutter
frequencies solution. However, when the frequency ratio is less than 1.3, not only the Approx. 1DOF forms, but also the Chen and Kareem (2007) and Bartoli (2008) formulas give inaccurate
results, as expected (Case 6, Case 7). Meanwhile, the Approx. 2-DOF forms show reasonable
results that are in good agreement with those of the CEA in all case studies.
4.2. Single modal flutter simulations
4.2.1. Torsional flutter
*
A non-streamlined cross section having values of A 2 tends to change its sign from negative to
positive, causing negative aerodynamic damping which will lead to instability in pure torsion
(Strmmen 2006). To deal with torsional flutter a mode-by-mode approach is performed, since the
resonance frequency associated with this mode is ( V ) then = ( V cr ) and the impedance
Table 5 Case study of coupled-mode flutter simulations
Case study

Flutter derivative set

B(m)

fz (Hz)

z (%)

(%)

2TF

51.5

0.252

2.64

0.50

0.50

0.0943

0.8727

2TF

36

0.232

2.31

1.00

1.00

0.0589

0.5245

2nd Geo-Germ

16.9

0.185

2.99

0.30

0.30

0.0305

0.3454

2nd Geo-Germ

31

0.099

2.75

0.50

0.50

0.0529

0.4674

Flat plate

38

0.204

2.38

0.50

0.50

0.0560

0.9050

Flat plate

60.4

0.080

1.32

1.00

1.00

0.0829

0.5942

Flat plate

27.5

0.164

1.17

0.20

0.20

0.1815

0.1943

Busan-Geoge Grand

22

0.334

3.00

0.50

0.50

0.0225

0.2048

R15

13.28

0.465

1.51

0.64

0.64

0.1815

0.1943

Simplified formulations for flutter instability analysis of bridge deck

375

Table 6 Results for coupled-mode flutter simulations

Analysis
methods
Approx.
1-DOF
forms

Approx.
2-DOF
forms
Chens
forms
Bartolis
forms
CEA

Onset
flutter
error

Vc ( m s )
f c ( Hz )
V c (%)
f c (%)
Vc ( m s )
f c ( Hz )
V c (%)
f c (%)
Vc ( m s )
f c ( Hz )
Vc ( m s )
f c ( Hz )
Vc ( m s )
f c ( Hz )

Case study
1

274.47

82.57

130.22

106.91

124.65

24.63

19.08

55.39

114.61

0.4509

0.1623

0.4231

0.2008

0.3170

0.0723

0.1895

0.9888

0.5902

-7.98

-6.71

-3.09

-0.18

-3.78

-22.35

-61.13

-8.41

-1.78

-0.71

0.91

2.81

1.57

0.32

1.70

0.35

0.45

0.98

293.82

86.34

131.23

106.59

128.29

27.13

27.98

55.40

112.42

0.4566

0.1630

0.4206

0.1985

0.3188

0.0722

0.1893

0.9708

0.5949

-0.87

-2.05

-2.30

-0.49

-0.84

-11.06

-9.85

-8.40

-3.77

0.54

1.35

2.22

0.43

0.90

1.63

0.2

-1.39

1.77

284.80

86.08

159.75

113.04

126.20

24.78

17.85

89.51

116.92

0.4504

0.1601

0.5540

0.2720

0.3139

0.0713

0.1890

1.0030

0.5848

286.48

86.97

134.64

109.96

126.22

24.74

17.69

60.90

119.51

0.4501

0.1608

0.4190

0.2029

0.3137

0.0713

0.1891

0.9858

0.5870

296.38

88.11

134.24

107.11

129.37

30.13

30.74

60.05

116.66

0.4541

0.1608

0.4112

0.1976

0.3160

0.0711

0.1889

0.9843

0.5844

matrix (Eq. (6)) is reduced as follows

ae

ae

S ( r, V cr ) = 1 ( r ) + i2 ( ) r

(24)

Setting the real and imaginary parts of Eq. (24) equal to zero, a dynamic stability limit may then
be identified at an in-wind resonance frequency
4

B *
r = 1 + --------- A 3

2m

0.5

(25)

when the damping properties are

4
*
B
ae
= = --------- ------r A 2 ( K c )

4m

(26)

Based on a cross section prone to torsional flutter, the sensitivity analyses are also performed to
validate the approximation degree of the proposed formulations and 1-DOF Torsion forms Eq. (25)
and Eq. (26) that was first presented by Scanlan (Simiu and Scanlan 1996) along with the CEA
approach. Let us consider that the previous rectangular section characterized by dynamic properties
includes the bridge deck width is B = 35.5 m; non-dimensional mass, moment of inertia are z =
0.035, = 0.174; fundamental torsional frequency f = 0.192(Hz) and the aerodynamics properties
of a rectangular cylinder R5 with width to depth ratios of 5 (Matsumoto 1996).
In Fig. 9, the critical reduced wind speed and frequency are plotted against the frequency ratio in
the case of zero damping and moderate and high level damping ratios. It is concluded that for the 1DOF Torsion forms, conservative values of the critical reduced wind speeds are shown when

376

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

structural frequency ratios are changed. Meanwhile, both the critical frequency and reduced velocity
curves defined by the Approximated 1-DOF and 2-DOF formulations show the same trend as that
of CEA. When the frequency ratios are close to unity, the Approx. 2-DOF forms show the peak
value of critical reduced velocity. Nevertheless, the errors between solution of the approximated
formulations and that of CEA are observed in case of high-level damping.
Further numerical validation of the case studies presented in Table 7 is provided at the end of this
section, where one case study is of the existing structure, the Seohae Grand bride (Case 5), and the
other case studies are of the Edge-girder bridge sections combined with the prototypes of the
geometric parameters of the existing bridges such as Busan-Geoge Grand (Case1, Case 3), Golden
Gate (Case 2, Case 6), Krkinen (Case 4) and the aerodynamics parameters of the rectangular
section of R5, R10 with width to depth ratios of 5, 10 (Matsumoto 1996).
From the results gathered in Table 8, it seems possible to conclude that both proposed
approximate formulas give results that agreed well with those of CEA as well as with those of the
Bartoli and Mannini (2008) and Chen and Kareem (2007) formulas. For the case of the Golden
Gate bridge, characterized by a very small frequency ratio ( = 1.17), not only the Approx. 1-DOF
forms but also reference formulas show poor accuracy results when compared with the those of
CEA, while the Approx. 2-DOF forms still give accurate results (Case 2, Case 6).
4.2.2. Heaving-mode aeroelastic instability
*
A cross section having bluffness with values of H 1 tends to change its sign from the negative to
positive causing negative aerodynamic damping which will lead to instability in pure bending
(Strmmen 2006). Since the resonance frequency associated with this mode is z(V) then r =
z(Vcr) and the impedance matrix (Eq. (6)) becomes
ae

ae

S ( r, V cr ) = 1 zz ( r z ) + i2 ( z zz ) r z

(27)

Setting the real and imaginary parts of this equation equal to zero, when the damping properties are

Fig. 9 Comparison of flutter onset of R5 for varying of the structural frequency ratio
(a) critical reduced wind speed and (b) critical frequency

Simplified formulations for flutter instability analysis of bridge deck

377

Table 7 Case study of torsional flutter simulations

Case study

Flutter derivative set

B (m)

fz (Hz)

z (%)

(%)

R5

22

0.334

3.00

0.5

0.5

0.0225

0.2048

R5

27.5

0.164

1.17

0.5

0.5

0.1815

0.1943

R10

22

0.334

3.00

0.5

0.5

0.0225

0.2048

R10

13.28

0.465

1.51

0.64

0.64

0.0128

0.1270

Seohae Grand

34

0.250

1.84

0.5

0.5

0.0502

0.6061

Seohae Grand

27.5

0.164

1.17

0.5

0.5

0.1815

0.1943

Table 8 Results for torsional flutter simulations

Analysis
methods
Approx.
1-DOF
forms

Approx.
2-DOF
forms
1-DOF
Torsion forms
Chens forms
Bartols
forms
CEA

Onset flutter
error

Vc ( m s )
f c ( Hz )
V c (%)
f c (%)
Vc ( m s )
f c ( Hz )
V c (%)
f c (%)
Vc ( m s )
f c ( Hz )
Vc ( m s )
f c ( Hz )
Vc ( m s )
f c ( Hz )
Vc ( m s )
f c ( Hz )

Case study
1

109.23

15.29

158.15

69.29

45.72

15.81

0.8674

0.1833

0.8247

0.6182

0.4165

0.1861

-5.86

-28.85

-6.43

-10.48

-4.00

-21.09

-6.46

1.83

-8.58

-5.24

-1.08

-0.05

110.37

20.10

160.57

70.33

46.57

17.72

0.9290

0.1800

0.9083

0.6579

0.4215

0.1869

-4.87

-6.46

-4.99

-9.13

-5.26

-8.04

0.19

0.02

0.69

0.84

0.31

0.38

118.95

28.62

179.84

85.19

52.52

18.68

0.9121

0.1752

0.8535

0.6327

0.4110

0.1851

117.70

20.15

169.56

77.61

51.88

15.46

0.9270

0.1802

0.8983

0.6520

0.4600

0.1916

113.27

10.02

165.88

76.96

48.23

17.02

0.9306

0.1853

0.9069

0.6580

0.4275

0.1868

116.02

21.49

169.01

77.39

47.55

19.15

0.9273

0.1800

0.9021

0.6524

0.4210

0.1862

4
B *
ae
z = zz = --------- -------r H 1
4 m z
a dynamic stability limit may then be identified at an in-wind resonance frequency
2

B *
r = z 1 + --------- H 4

z
2m

(28)

0.5

(29)

As shown in Fig. 10, the critical reduced velocity and frequency of the previous rectangular
section (z = 0.035, = 0.174) with the fundamental vertical bending frequency f = 0.087(Hz) and
the aerodynamics properties of a rectangular cylinder with width to depth ratios of 1 (B/D = 1), R1
(Matsumoto 1996) obtained by the approximated approaches or CEA are compared for various
frequency ratios under different damping levels. It seems possible to conclude that the Approx. 2-

378

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

DOF forms and 1-DOF Bending forms Eqs. (28) and (29) show results that agree fairly well with
those obtained by the CEA method.
Table 9 shows that the case studies of the prototypes using the flutter derivative set of R1, R2 are
rectangular prisms with B/D = 1,2 and have heaving divergent type instability (Matsumoto 1996)
and the geometric parameters of the previous bridge are listed in Table 1. Table 10 summarizes the
results of the critical reduced velocity and frequency using the proposed formulas and the
conventional CEA and 1-DOF Bending forms.
From the results, the differences in the critical flutter speed and frequency given by the proposed
Approx. 2-DOF forms and CEA are between three and five percent. Meanwhile, it seems that when
the frequency ratio is greater than 1.5, the approximate 1-DOF forms are sufficient to obtain an
acceptable accuracy.

5. Conclusions
In this paper, a set of simplified formulations for estimating flutter critical and frequency has been
proposed. Firstly, a properly simplified strategy is used by neglecting some damping ratio terms in
coefficients of the real and imaginary equation part of flutter equation, as the results, approximated
formulas, consist of the uncoupled and coupled flutter derivatives are derived. These formulas are
able to show accurate results when compare with those obtain by CEA method, even if structural
close unity in case of low damping level. Furthermore, the characteristics of the torsional branch
(TB) and heaving branch (HB) are also highlighted in order to easily identify in which branch the
coupled flutter instability occurs because the frequency ratio, for TB and HB is asymmetric in
terms of torsional and heaving motion.
Next, when the non-dimensional mass and polar moment of inertia coefficients associated with
flutter derivatives are neglected, the simplified formulations are to be compacted of the uncoupled
*
*
*
*
flutter derivatives ( H 1, H 4 and A 2, A 3 ) . These uncoupled flutter derivatives could be extracted
from the one-degree-of freedom test set-up. This assumption is in agreement with the interrelationship among flutter derivatives (Matsumoto 1996), which seems to be suitable for the flutter

Fig. 10 Comparison of flutter onset of R1 for varying of the structural frequency ratio
(a) critical reduced wind speed and (b) critical frequency

Simplified formulations for flutter instability analysis of bridge deck

379

Table 9 Case study of heaving-mode aeroelastic instability simulations

Case study

Flutter
derivative set

B(m)

fz (Hz)

z (%)

(%)

R1

22

0.334

3.00

0.50

0.50

0.0225

0.2048

R1

13.28

0.465

1.51

0.64

0.64

0.0128

0.1270

R1

34

0.250

1.35

0.50

0.50

0.0502

0.6061

R2

22

0.334

3.00

0.50

0.50

0.0225

0.2048

R2

13.28

0.465

1.51

0.64

0.64

0.0128

0.1270

R2

27.5

0.164

1.17

0.50

0.50

0.1815

0.1943

Table 10 Results for heaving-mode aeroelastic instability simulations

Analysis
methods
Approx.
1-DOF
forms

Approx.
2-DOF
forms
1-DOF
Bending forms
CEA

Onset flutter
error

Vc ( m s )
f c ( Hz )
V c (%)
f c (%)
Vc ( m s )
f c ( Hz )
V c (%)
f c (%)
Vc ( m s )
f c ( Hz )
V c (%)
f c (%)

Case study
1

64.00

54.44

52.78

56.81

58.34

8.26

0.3137

0.4479

0.1973

0.2391

0.4048

0.1703

-0.97

-2.77

-27.17

-7.74

-4.76

33.37

-0.03

-0.29

-13.67

-11.43

-2.21

-3.11

64.20

55.25

72.14

60.51

59.05

6.18

0.3137

0.4489

0.2283

0.2726

0.4128

0.1671

-0.66

-1.31

-0.46

-1.74

-3.59

-0.21

-0.03

-0.08

-0.11

0.95

-0.27

-4.93

63.96

55.84

66.27

55.49

56.35

7.53

0.3105

0.4463

0.2087

0.2333

0.3911

0.1699

64.62

55.99

72.47

61.58

61.25

6.19

0.3138

0.4492

0.2285

0.2700

0.4139

0.1758

derivatives of a wide range of cross-section geometries, hence the physical consistency is still
guaranteed in the approximate formulations. Besides that, its accuracy was verified for a variety of
bridge cross sections showing that the approximate solutions in onset flutter is acceptable when
structural frequency is larger than 1.5, wherein first torsion mode and vertical bending mode are
enough far from each other.
Finally, the pragmatic feature of these approximate formulations is that they are able to capture all
aerodynamic instabilities such as coupled-mode flutter as well as single-mode-dominated flutter
cases where the previous empirical as well as approximate formulas are no longer applicable.

Acknowledgments
The support of the research reported here by the grant from the Ministry of Land, Transport and Maritime
Affairs of Korean Government through Super Long Span Bridge R&D Center is gratefully acknowledged.

380

Tan-Van Vu, Young-Min Kim, Tong-Seok Han and Hak-Eun Lee

References
Bartoli, G. and Mannini, C. (2008), A simplified approach to bridge deck flutter, J. Wind Eng. Ind. Aerod.,
96(2), 229-256.
Bleich, F. (1949), Dynamic instability of truss-stiffened suspension bridge under wind action, Proc. ASCE,
74(8), 1269-1314.
Chen, X and Kareem, A. (2007), Improved understanding of bimodal coupled bridge flutter based on
closedform solutions, J. Struct. Eng., 133(1), 22-31.
DAsdia, P. and Sepe, V. (1998), Aeroelastic instability of long span suspended bridges: a multimode
Approach, J. Wind Eng. Ind. Aerod., 74-76, 849857.
Dyrbye, C. and Hansen, S.O. (1997), Wind Loads on Structures, Wiley, New York.
Korea Highway Corporation (KHC). (1998), Wind tunnel report for the Seohae Grand bridge, Kyunggi-do, Korea.
Kiviluoma, R. (2001), Frequency-domain approach for calculating wind induced vibration and aeroelastic
stability characteristics of long span bridges. PhD thesis, Acta Polytechnica Scandinavica.
Larsen, A. (1993), Aerodynamic aspects of the final design of the 1624m suspension bridge across the Great
Belt, J. Wind Eng. Ind. Aerod., 48(2-3), 261-285.
Larsen S.V. (2002), Section model tests for the design of the 2nd Geo-Germ Bridge, Force Technology-DMI,
Lyngby, Denmark.
Lee, H.E., et al. (2004), Section model tests for the design of the cable stayed bridge, Lot1, Busan-Geoge Fixed
Link, Korea University Research Centre for Disaster Prevention Science and Technology, Seoul, Korea
Livesey, F.M. (1995), Wind tunnel studies of the Hoga Kusten Bridge during construction, Part II: Aeroelastic
model tests, DMI Report 94124, Danish Maritime Institute, Denmark.
Matsumoto, M. (1996), Aerodynamic damping of prisms, J. Wind Eng. Ind. Aerod., 59(2-3), 159-175.
Matsumoto, M. (1999), Recent study on bluff body aerodynamics and its mechanism, Proceedings of the10th
Int. Conf. on Wind Engineering, Wind Engineering into the 21st Century, (Eds., A. Larsen, G.L. Larose, and
F.M. Livesey), Balkema, Rotterdam, The Netherlands.
Matsumoto, M., Shijo, S., Eguchi, A., Hikida, T., Tamaki, H. and Mizuno, K. (2004), On the flutter
characteristics of separated two box girders, Wind Struct., 7(4), 281-291.
Nakamura, Y. (1978), An analysis of binary flutter of bridge deck sections, J. Sound Vib., 57 (4), 471-482.
Nissen, H.D., Srensen, P.H. and Jannerup, J. (2004), Active aerodynamic stabilisation of long suspension
bridges, J. Wind Eng. Ind. Aerod., 92(10), 829-847.
Rocard, Y. (1963), Instabilite des ponts suspendus dans le vent-experiences sur modele reduit, Nat. Phys. Lab.
Paper, 10, England.
Sarkar, P.P., Jones, N.P. and Scanlan, R.H. (1992), System identification for estimation of flutter derivatives, J.
Wind Eng. Ind. Aerod., 42(1-3), 1243-1254.
Scott, R. (2001), In the wake of tacoma, suspension bridges and the quest for aerodynamics stability, American
society of civil engineers, reston VA.
Selberg, A. (1961), Oscillation and aerodynamic stability of suspension bridges, Acta. Polytech. Scand., 13,
308-377
Scanlan, R.H. and Tomko, A. (1971), Airfoil and bridge deck flutter derivatives, J. Eng. Mech,-ASCE, 97(6),
1717-1737.
Scanlan, R.H., Jones, N.P., and Singh, L. (1997), Inter-relation among flutter derivatives , J. Wind Eng. Ind.
Aerod., 69-71, 829-837.
Strmmen, E. (2006), Theory of Bridge Aerodynamics, Springer Verla Publications.
Simiu, E. and Scanlan, R.H. (1996), Wind Effects on Structures, Wiley, New York.
Theodorsen, T.(1934), General theory of aerodynamic instability and the mechanism of flutter, Technical
Report. 496, N.A.C.A., Washington, 413-433
Yang, Y.X., Ge, Y.J and Cao, F.C. (2007), Flutter performance of central-slotted box girder section for Long
span suspension Bridges, China J. Highway & Transport, 20(3), 35-40.
Yang, Y.X., Ge, Y.J. (2009), Aerodynamic flutter control for typical girder sections of long-span cable-supported
bridges, Wind Struct., 12(3), 205-217.
Zhang X.J. (2006), Study of design parameters on flutter stability of cable-stayed-suspension hybrid bridges,

Simplified formulations for flutter instability analysis of bridge deck

381

Wind and Structures, Wind Struct., 9(4), 331-344.

Zhou, L. and Ge, Y.J. (2009), Sectional model test study on vortex-excited resonance of vehicle-bridge system
of Shanghai Bridge over Yangtse River, Frontiers of Arch. & Civil Eng. in China, 3(1), 67-72.
CC