Vo l . 40 N o .

1 Februar y 2007

Reynolds Number Effects on the Steady and Unsteady

Aerodynamic Forces Acting on the Bridge Deck Sections
of Long-Span Suspension Bridge

MATSUDA Kazutoshi : Dr. Eng., P. Eng., Manager, Heat and Fluid Dynamics Department, Research
Laboratory, Research & Development
TOKUSHIGE Masafumi : Design Dept., Bridge & Road Construction Division, Logistics Systems &
Structures
IWASAKI Tooru : Engineering, Galaxy Express Corporation

This paper gives the results of research on Reynolds number effects on steady and unsteady aerodynamic
forces on twin-box bridge section models of different scales in three different wind tunnels. The forces were
measured over a wide Reynolds number range from 1.1 × 104 to 1.5 × 106 based on the dimension of deck
height. In order to investigate Reynolds number effects on the steady and the unsteady aerodynamic forces,
wind-induced static displacement analysis and flutter analysis were carried out for a suspension bridge with a
main span of 2 500 m using a three-dimensional analytical model. Also, the paper considers amplitude effects
on both the unsteady aerodynamic forces and the predicted flutter speeds.

a bridge in which the aerodynamic characteristics or

1. Introduction wind response is influenced by Reynolds number. If the
The 1940 collapse accident of “the Tacoma Narrows aerodynamic characteristics or wind response differs for
Bridge” in the United States triggered the recognition different Reynolds numbers, there is a risk of erring in the
of the importance of dynamic wind effect in the wind estimation or evaluation of the response of a real bridge
resistant design of a bridge for the first time. Since then, to wind. Therefore, it is important for the wind resistant
a check by a wind tunnel test has come to be done to design of a long-span bridge to understand Reynolds
ensure the aerodynamic stability of a long-span bridge. If number effects on wind tunnel test results.
a wind tunnel test is to be done using a wind tunnel of low Using three different wind tunnels and bridge deck
wind speed, the aeroelastic similarity conditions must be models, this study measured the steady and unsteady
made equal between the real phenomenon and the wind aerodynamic forces acting on the bridge deck model
tunnel test. One such condition is Reynolds number, in Reynolds number range from 1.1 × 10 4 to 1.5 ×
which is the ratio of inertial force to viscous force of 106 (representative length: deck height) and looked at
a fluid; however, it is difficult to achieve equality in Reynolds number effects on each type of aerodynamic
ordinary cases. In a full bridge model wind tunnel test forces. As a study case in the past using a bridge deck
using a full aerostatic model, take a model scale of, for model in a high Reynolds number region, there are
example, 1/50 to 1/150 and assume that the aeroelastic steady aerodynamic force measurements(3) and spring-
similarity based on Froude number is satisfied. Then supported model tests.(4), (5) In this study, not only steady
the ratio of Reynolds number between the real bridge aerodynamic force, but also unsteady aerodynamic force
and wind tunnel test ranges from 354 (=50 × √50) to obtained by the forced excitation method, was measured.
1 837 (=150 × √ 150 ), that is, there is a difference of the Unsteady aerodynamic force measurement was made in a
order of 102 to 103. Therefore, in wind tunnel tests of a high Reynolds number region that was smaller by a factor
long-span bridge or other large steel structure, the test of 101 than that of the real bridge, and this was the world’s
conditions are set ignoring the similarity with respect to first attempt. Further, the obtained steady and unsteady
Reynolds number. When wind acts on a bridge or other aerodynamic force coefficients were applied to a three-
bluff structure, the separation point of the flow is thought dimensional analysis model of a long-span suspension
to be fixed. In practice, therefore, wind tunnels tests have bridge having a center span of 2 500 meters to investigate
been performed assuming that Reynolds number has little Reynolds number effects on the wind-induced static
influence on the wind resistance of the structure. displacements and flutter wind speed of the real bridge.
Recently, however, there are an increasing number of Results are also reported here. Flutter in this study means
cases reported in Refs(1) - (5) on the wind tunnel test of coupled flutter, which is a self-excited vibration consisting

12
 Vo l . 40 N o . 1 Februar y 2007

of a vertical bending motion and torsional motion coupled bridge deck

with a phase difference.

2. Wind Tunnel Test
From the definition equation of Reynolds number shown
by Equation (1), one can see that a higher Reynolds
number can be ensured for a higher mean wind speed and
a larger model.
Re = VD .............................................................. (1)
v
where
Re : Reynolds number
V : Mean wind speed (m/s) Fig. 3 1/80 scale bridge deck model mounted in the IHI 1.5 m
D : Representative length, or deck height in this × 2.5 m wind tunnel
case (m)
v : Kinematic viscosity coefficient (m2/s)
Therefore, this study used bridge deck models of 1/10, 1/30 and 1/80 were employed considering the test
different sizes and wind tunnels corresponding to them in section size of the wind tunnels used. These three models
size in order to measure steady and unsteady aerodynamic have the same cross-sectional shape. The three wind
forces in a wide range of Reynolds numbers while tunnels and models are shown in Figs. 1 to 3. A cross-
ensuring high Reynolds numbers. For the model scale, sectional view of the bridge deck under study is shown in
Fig. 4. Table 1 shows the major dimensions of the bridge
bridge deck
deck models. The steady and unsteady aerodynamic
forces acting on the model were measured using load cells
set at both ends of and inside the model. In addition, a
combined total of 60 pressure holes were provided on the
upstream and downstream halves of the deck to measure
the surface pressure distribution around the model cross-
section. The definition equations of mean and fluctuating
pressure coefficients are shown in Equations (2) and (3).
PS − P0
CPS = ...................... (2)
1 ρV 2
2
PD
(Note) *1 : National Research Council Canada CPD = ....................... (3)
Fig. 1 1/10 scale bridge deck model mounted in the NRC*1 9.1 m 1 ρV 2
× 9.1 m wind tunnel 2
( a ) Deck section A
Handrails Crash barrier Grating (solidity ratio 40%)
2% 2%
80°
0.4

B  4.0
bridge deck
( b ) Deck section B

80°

B  4.0

( c ) Deck section C

40°

B  4.6

(Note) Dimensions are values of a 1/10 scale bridge deck model.

Fig. 2 1/30 scale bridge deck model mounted in the IHI 6 m
× 3 m wind tunnel Fig. 4 Bridge deck cross section (unit: m)

13
Vo l . 40 N o . 1 Februar y 2007

Lift Ls
where
CPS : Mean pressure coefficient Pitching moment Ms
Drag Ds
CPD : Fluctuating pressure coefficient
PS : Mean pressure (= Total pressure) (N/m2) 80°
Angle of attack B
P0 : Reference static pressure (N/m2)
PD : RMS value of fluctuating pressure (N/m2) ection
dir
r : Air density (kg/m3) Wind

V : Mean wind speed (m/s) Fig. 5 Steady aerodynamic force

Table 2 shows the major dimensions and Reynolds
number range of the wind tunnels used. To ensure high
Reynolds numbers, the large wind tunnel of the National components of the aerodynamic forces acting on the
Research Council Canada (NRC) was used. The test bridge deck and can be classified into the three force
section size is 9.1 meters wide by 9.1 meters high by 23.9 components of drag, lift and pitching moment as shown in
meters long, and the maximum wind speed is 55 meters Fig. 5. When rendered dimensionless, they are called the
per second. For the airflow in the wind tunnels, a smooth drag coefficient, lift coefficient and moment coefficient,
flow with small turbulence intensity was employed in respectively, and expressed as the following definition
every case. equations:
In addition, Strouhal number was calculated by DS
measuring the frequency of the shed vortex street CD = .............................................. (5)
1 ρV 2A
generated in the wake area of the bridge deck model. The n
2
frequency of a shed vortex street is in proportion to the
LS
wind speed and in inverse proportion to the representative CL = .............................................. (6)
1 ρV 2B
length of the structure. As shown by Equation (4), the
2
Strouhal number is defined as the proportionality constant
MS
and is a dimensionless number specific to the cross- CM = .............................................. (7)
1 ρV 2B 2
sectional shape of the structure.
2
fD where
St = ............................................................ (4)
V CD
: Drag coefficient
where : Lift coefficient
CL
St : Strouhal number (—) CM
: Pitching moment coefficient
f : Vortex shedding frequency (1/s) DS
: Mean drag per unit length (N/m)
D : Deck height (m) : Mean lift per unit length (N/m)
LS
V : Mean wind speed (m/s) MS
: Mean pitching moment per unit length
(N·m/m)
3. Reynolds number effects on aerodynamic
r : Air density (kg/m3)
forces acting on bridge deck
V : Mean wind speed (m/s)
3.1 Steady aerodynamic forces An : Projected area per unit length of bridge
3.1.1 Relationship between Reynolds number and deck, or value of deck height only of real
steady aerodynamic coefficients bridge in this case (= 4 m2/m)
The steady aerodynamic forces are the time mean B : Value of deck width of real bridge (= 40 m)
The steady aerodynamic force coefficients at Deck
Table 1 Bridge deck cross section model properties
sections A, B and C shown in Fig. 4 are shown in Figs.
Scale
Total width B Height H Length L Aspect ratio 6, 7 and 8, respectively. Reynolds number effects on the
(m) (m) (m) L/B
steady aerodynamic force coefficients were conspicuous
1/10 4.0 0.4 8.0 2.0
at Deck section A in Fig. 6. At angles of attack below
1/30 1.3 0.13 3.0 2.3
−2 degrees and above +3 degrees, the lift and pitching
1/80 0.5 0.05 1.35 2.7
moment coefficients change as the Reynolds number

Table 2 Wind tunnel properties

Test section
Wind speed range Reynolds number range
Wind tunnel Width Height Length ( m/s ) Re = VD/v
(m) (m) (m)
the NRC 9.1 m s 9.1 m wind tunnel 9.1 9.1 23.9 5 - 55 1.4 s 105 - 1.5 s 106
the IHI 6 m s 3 m wind tunnel 6.0 3.0 24.0 0.5 - 15 4.5 s 103 - 1.4 s 105
the IHI 1.5 m s 2.5 m wind tunnel 1.5 2.5 8.0 0.5 - 20 1.7 s 103 - 6.8 s 104

14
 Vo l . 40 N o . 1 Februar y 2007

( a ) Drag coefficient ( b ) Lift coefficient (c) Pitching moment coefficient

2.0 0.4 0.08

Pitching moment coefficient CM (—)

0.2
Drag coefficient CD (—)

Lift coefficient CL (—)

1.5 0.04

1.0 0.0 0.00

0.5 0.2 0.04

0.0 0.4 0.08

6 4 2 0 2 4 6 6 4 2 0 2 4 6 6 4 2 0 2 4 6
Angle of attack B
(degree) Angle of attack B
(degree) Angle of attack B
(degree)

(Notes) : Scale 1/80, Re  5.1 s 104 : Scale 1/10, Re  1.1 s 106

: Scale 1/30, Re  1.0 s 105 : Scale 1/10, Re  1.3 s 106
: Scale 1/10, Re  2.7 s 105

Fig. 6 Relationship between angle of attack and steady aerodynamic force coefficients (Deck section A)

( a ) Drag coefficient ( b ) Lift coefficient ( c ) Pitching moment coefficient

2.0 0.4 0.08

Pitching moment coefficient CM (—)

Drag coefficient CD (—)

1.5 0.2
Lift coefficient CL (—)

0.04

1.0 0.0 0.00

0.5 0.2 0.04

0.0 0.4 0.08

6 4 2 0 2 4 6 6 4 2 0 2 4 6 6 4 2 0 2 4 6
Angle of attack B
(degree) Angle of attack B
(degree) Angle of attack B
(degree)

(Notes) : Scale 1/80, Re = 5.1 s 104 : Scale 1/10, Re = 1.1 s 106

: Scale 1/30, Re = 1.0 s 105 : Scale 1/10, Re = 1.3 s 106
: Scale 1/10, Re = 2.7 s 105

Fig. 7 Relationship between angle of attack and steady aerodynamic force coefficients (Deck section B)

(a) Drag coefficient (b) Lift coefficient ( c ) Pitching moment coefficient

2.0 0.4 0.08
Pitching moment coefficient CM (—)
Drag coefficient CD (—)

Lift coefficient CL (—)

1.5 0.2 0.04

1.0 0.0 0.00

0.5 −0.2 −0.04

0.0 −0.4 −0.08

−6 −4 −2 0 2 4 6 −6 −4 −2 0 2 4 6 −6 −4 −2 0 2 4 6
Angle of attack a (degree) Angle of attack a (degree) Angle of attack a (degree)

(Notes) : Scale 1/80, Re = 5.1 × 104 : Scale 1/10, Re = 1.1 × 106

: Scale 1/30, Re = 1.0 × 105 : Scale 1/10, Re = 1.3 × 106
: Scale 1/10, Re = 2.7 × 105

Fig. 8 Relationship between angle of attack and steady aerodynamic force coefficients (Deck section C)

15
Vo l . 40 N o . 1 Februar y 2007

increases. At angles of attack near 0 degree, these with increasing Reynolds number, the lift coefficient at
coefficients change a little with the Reynolds number. an angle of attack of +5 degrees shown in Fig. 6 can be
This tendency at an angle of attack of 0 degree is similar thought to increase in magnitude in the positive direction,
to the steady aerodynamic force coefficients of a 1/10 and the pitching moment coefficient can be thought to
scale model(3) of “the Normandy Bridge” in France. The change from negative to positive.
measured Reynolds number range at this time is, taking At any angle of attack, the peak of the negative pressure
the deck height as the representative length, from 0.2 × 106 at the bottom surface end of the upstream deck half shifts
to 1.7 × 106. to the upstream side as the Reynolds number increases.
3.1.2 Mechanism of Reynolds number effects on This tendency is also seen in past studies on an inverted
steady aerodynamic forces trapezoidal-shaped deck section(6) and p-shaped deck
The lift and pitching moment coeff icients at Deck section of a cable-stayed bridge.(7)
section A are influenced by Reynolds numbers at angles 3.1.3 Relationship between drag coefficient and
of attack below −2 degrees and above +3 degrees. The Strouhal number
mechanism for this is looked at here from the measured A representative example in which Strouhal number is
results of pressure distribution at the model surface. The influenced by Reynolds number is a cylindrical section.
mean pressure distribution at the model surface in Deck That is, the Strouhal number is about 0.2 in the subcritical
section A at rest is shown in Fig. 9 for angles of attack region and a critical region with the Reynolds number of
of −5, 0 and +5 degrees. At all the angles of attack, about 3 × 105 or less, while it is as large as about 0.5 in a
the pressure distribution locally changes at the top and higher critical region and the supercritical region.(8)-(10) In
bottom surfaces of the upstream deck half as the Reynolds this way the Strouhal number increases with the Reynolds
number increases. At an angle of attack of −5 degrees, for number, and this means from Equation (4) that the rate
example, as the Reynolds number increases, the magnitude of increase of vortex shedding frequency is higher than
of the negative pressure at the bottom surface of the the rate of increase of wind speed, namely, the width of
upstream deck half increases, so the downward lift and the shed vortices generated in the wake area of the bridge
negative steady aerodynamic pitching moment increase. deck model becomes narrow.
Therefore, the lift and pitching moment coefficients at an In past studies, Shu et al.(11) measured the Strouhal
angle of attack of −5 degrees in Fig. 6 can be thought to number of rectangular sections with B/D = 2, 4 and
increase in magnitude, but in the negative direction, as the 6 (B: Deck width, D: Deck height) and pointed out
Reynolds number increases. as a result that with a large aspect ratio with B/D, the
At an angle of attack of +5 degrees, as the Reynolds Reynolds number has a large influence on the Strouhal
number increases, the magnitude of the negative pressure number when the Reynolds number is 5 000 or less.
at the bottom surface of the upstream deck half decreases They concluded that a 1-box cross section showed
and the magnitude of the negative pressure at the top a similar tendency to a rectangular section. Schewe
surface of the upstream deck half increases, so the upward et al.(2) ascertained Reynolds number effects on the steady
lift and positive pitching moment increase. As a result, aerodynamic force coefficients for an inverted trapezoidal
cross-section and say that, for such a cross-section, the
( a ) Angle of attack B
 5° relationship between the Strouhal number St and drag
1.0
coefficient CD can be expressed as Equation (8).
5°dWind 0
irection
80° St × CD = Constant ............................................ (8)
0
The relationship between drag coeff icient and
1.0
Strouhal number at Deck section A, where Reynolds
( b ) Angle of attack B
 0° number effects were observed, is shown in Fig. 10. At
1.0
angles of attack of −5 and +5 degrees, the Strouhal
Wind
direction
0 number increases, while the drag coefficient decreases
80°
0° as the Reynolds number increases. At an angle of
0
attack of 0 degree, Reynolds number effects were hardly
1.0 observed, even for the same section, so the Strouhal
( c ) Angle of attack B
 5° number and drag coefficient were not influenced by the
1.0
Reynolds number and almost constant. Figure 11 shows
0 the relationship between the Reynolds number and St
Wind 0
80°
× CD . As with the past research results, it was verified
5° direction that Equation (8) holds true also for Deck section A.
1.0 Therefore, it was found that the Reynolds number effects
(note) : Re  2.7 s 105
: Re  6.8 s 105 on the drag coefficient can also be explained from the
: Re  1.3 s 106 viewpoint of the Strouhal number.
Fig. 9 Surface pressure distribution ( Deck section A )

16
 Vo l . 40 N o . 1 Februar y 2007

( a ) Angle of attack a = −5° ( b ) Angle of attack a = 0°

0.30 2.0 0.30 2.0

Drag coefficient CD (—)

Drag coefficient CD (—)

Strouhal number St (—)

Strouhal number St (—)

0.25 1.5 0.25 1.5

0.20 1.0 0.20 1.0

0.15 0.5 0.15 0.5

0.10 0.0 0.10 0.0

105 106 105 106
Reynolds number*1 Reynolds number*1

( c ) Angle of attack a = +3° ( d ) Angle of attack a = +5°

0.30 2.0 0.30 2.0

Drag coefficient CD (—)

Drag coefficient CD (—)

Strouhal number St (—)

Strouhal number St (—)

0.25 1.5 0.25 1.5

0.20 1.0 0.20 1.0

0.15 0.5 0.15 0.5

0.10 0.0 0.10 0.0

105 106 105 106
Reynolds number*1 Reynolds number*1

(note) : CD
: St
*1 : Re = VD/v

Fig. 10 Relationship between drag coefficients and Strouhal numbers ( Deck section A )

a full aerostatic model. However, the Reynolds number of

0.5 : Angle of attack a = +5° a wind tunnel test is smaller by a factor of 102 to 103 than
: Angle of attack a = −5°
that of a real bridge, so the aeroelastic similarity condition
: Angle of attack a = +3°
: Angle of attack a = 0°
with respect to the Reynolds number cannot be satisfied
0.4 in ordinary cases. The test expenses are high, including
the model fabrication cost, and the test has trouble with
St × CD*1 (—)

a parametric study of bridge deck section shape. For

0.3 these and other reasons, employment of this technique is
sometimes shelved.
In a technique replacing this full bridge model test, the
0.2 unsteady aerodynamic forces acting on the bridge deck
model are found by a wind tunnel test, then the obtained
aerodynamic forces are input as external forces of the
0.1 equations of motion of the whole bridge system, and the
105 106 flutter characteristics are estimated by a flutter analysis
Reynolds Number*2
program in an analytical way. This section pays attention
(note) *1 : St = Strouhal number
CD = Drag coefficient number
to the unsteady aerodynamic forces used in this method.
*2 : Re = VD/v The unsteady aerodynamic forces can be transformed into
Fig. 11 Relationship between Reynolds number and St × CD
dimensionless unsteady aerodynamic force coefficients,
( Deck section A ) which depend on the cross-sectional shape, Reynolds
number, reduced frequency, angle of attack and amplitude
in ordinary cases. Of these, there are few investigations
3.2 Unsteady aerodynamic forces into the effects of the Reynolds number.
3.2.1 Outline of unsteady aerodynamic forces In 1950, when Bleich(12) investigated the causes for
When examining the flutter characteristics of a long-span “the Tacoma Narrows Bridge” collapse accident, he
bridge, the most reliable technique will be, judging from applied the unsteady aerodynamic forces acting on the
the current technical level, a full bridge model test using plate wing as thought out by Theodorsen(13) and pointed

17
Vo l . 40 N o . 1 Februar y 2007

out the possibility of flutter in a suspension bridge quasi-steady aerodynamic force in phase, and this phase
with a truss stiffening girder. When a two-dimensional lag has a large influence on the flutter characteristics
plate rigid wing is in harmonic oscillation with two of the plate. From the results of comparison between
degrees of freedom of vertical bending and torsion of wind tunnel experiment values and calculated values,
small constant amplitude, the unsteady lift and unsteady it has been reported (14) that there are unreasonable
pitching moment acting on this wing can be expressed points in applying the above Bleich flutter theory to any
by the following equations: suspension bridge. For example, the theoretical value of
flutter wind speed is considerably higher than the flutter
L = −πρb 2 {ḣ˙ + Vq˙ − baq˙˙} wind speed obtained from wind tunnel tests. As a major
reason for that, because the air flow separates from the
⎧ ⎛1 ⎞ ⎫ stiffened deck cross-section of the suspension bridge,
− 2p ρVbC ( k ) ⎨η˙ + Vθ + b ⎜ − a⎟ θ˙⎬ ................... (9)
⎩ ⎝2 ⎠ ⎭ the potential flow assumption fundamental to the flutter
theory does not hold true, and therefore the unsteady
⎧ ⎛1 ⎞ ⎛1 ⎞ ⎫ aerodynamic forces acting on the cross section differ from
˙˙ − V ⎜ − a⎟ q˙ − b ⎜ + a 2⎟ θ˙˙⎬
M = prb3 ⎨ah
⎩ ⎝2 ⎠ ⎝8 ⎠ ⎭ the theoretical equations of Equations (9) and (10).(15) To
find the unsteady aerodynamic forces at the bridge deck
⎛ 1⎞ ⎧ ⎛1 ⎞ ⎫ cross-section under consideration with high accuracy,
+ 2prVb2 ⎜ a + ⎟ C ( k ) ⎨h˙ + Vθ + b ⎜ − a⎟ θ˙⎬ ... (10)
⎝ 2⎠ ⎪⎩ ⎝ 2 ⎠ ⎪⎭ direct measurement using a wind tunnel test is mainly
used. Recently, computational fluid dynamics (CFD)
where is strenuously used to calculate unsteady aerodynamic
L : Unsteady lift (downward is positive) (N/m) forces in an analytical way.(16)
M : Unsteady pitching moment (pitch-up is This section examines the effects of the Reynolds
positive) (N·m/m) number on the unsteady aerodynamic forces acting on a
r : Air density (kg/m3) bridge deck cross-section of a long-span bridge.
b : Half chord length (m) 3.2.2 Relationship between Reynolds number and
V : Mean wind speed (m/s) unsteady aerodynamic force coefficients
a : See Fig. 12 When a sectional model is in coupled vibration of vertical
h : Vertical bending amplitude (m) bending and torsion with a circular frequency w in a wind
q : Torsional amplitude (rad) of speed V, the unsteady aerodynamic forces acting on the
C(k) : Theodorsen function {=F ( k ) + iG ( k )} model can be expressed as:
H1( 2 ) ( k ) ⎧ ⎫
= h
H1( 2 ) ( k ) + iH0( 2 ) ( k ) D = prB 2An w 2 ⎨(CDhR + iCDh i) + (CDqR + iCDq i )q ⎬ ... (11)
⎩ B ⎭
F(k) : Real part of Theodorsen function ⎧ ⎫
h
G(k) : Imaginary part of Theodorsen function L = prB 3w 2 ⎨(C LhR + iCLh i ) + (CL qR + iCLq i )q ⎬ ....... (12)
i : Imaginary unit ⎩ B ⎭
Hn(2)(k) : Hankel function of the first class ⎧ h ⎫
M = prB 4w 2 ⎨(CMhR + iCMh i) + (CMqR + iCMq i )q ⎬ ..... (13)
k : Reduced frequency (= wb/V) ⎩ B ⎭
w : Circular frequency (1/s)
These unsteady aerodynamic forces contain a where
Theodorsen function C(k), which represents the ratio of D : Unsteady drag (N/m)
quasi-steady aerodynamic force to unsteady aerodynamic L : Unsteady lift (N/m)
force. Therefore, the unsteady aerodynamic force lags the M : Unsteady pitching moment (N·m/m)
r : Air density (kg/m3)
B : Deck width (m)
h
An : Projected bridge deck area (m2/m)
w : Angular frequency (1/s)
–b 0 b
h : Vertical bending amplitude (amplitude at center
X
V ba of deck width) (m)
q : Torsional amplitude (rad)
M CXYZ : Unsteady aerodynamic force coeff icients
q (reduced frequency k = function of fB/V, f :
frequency, V: mean wind speed)
i : Imaginary unit
For measurement of the unsteady aerodynamic forces
L
acting on a bridge deck, the following two methods are
Fig. 12 Aerodynamic forces on oscillating flat plate available:

18
 Vo l . 40 N o . 1 Februar y 2007

(1) Forced oscillation method with box girder. (23)

This method was developed by Kawashima Of the measured unsteady aerodynamic coefficients
et al.(17) and applied to a suspension bridge model by at Deck section A and at an angle of attack of 0 degree
Ukeguchi and Sakata.(18), (19) Unsteady aerodynamic shown in Fig. 13, the coefficients CMhi and CLqR(24), (25) having
forces are measured while the model is in forced a large influence on flutter, hardly showed a discernible
oscillation. change with the Reynolds number. The coefficient that
(2) Free oscillation method showed a large change with Reynolds number was CLqi.
While a two-dimensional rigid model is set As the dependence on Reynolds number, the value of
in free oscillation in an air f low, the response this coefficient becomes negative for a Reynolds number
f requency, response loga r it h m ic decrement , of 9.5 × 104 or less, while it becomes positive for 3.4 ×
amplitude ratio of vertical bending to torsional 105 or more. However, the effect of this coefficient on
motion and phase difference are measured. The flutter characteristics is said to be small. Comparing the
unsteady aerodynamic forces are calculated back.(20) measured unsteady aerodynamic force coefficients with
In this study, the unsteady aerodynamic forces were the Theodorsen function, the two show about the same
measured by the forced oscillation method. This tendency, though they differ in magnitude.
measuring method is based on the assumption that the Of the measured unsteady aerodynamic coefficients at
bridge deck should be in harmonic oscillation when Deck section A and at an angle of attack of +3 degrees
it flutters. Therefore, the unsteady aerodynamic force shown in Fig. 14, the coefficients influenced greatly by
coefficients obtained by the forced oscillation method Reynolds number are CMhR, CMhi, CMqR and CMqi related to
may depend on amplitude. An investigation into the the pitching moment. That is, CMqi having a large influence
amplitude dependence is described in Section 3.2.3. on flutter is included in those coefficients. On the other
Measured results of unsteady aerodynamic forces(21), (22) hand, CLqR has a large influence on flutter; however, no
are shown in Figs. 13 to 16. Vertical axis represents the dependence on Reynolds number is seen. Considering the
unsteady aerodynamic force coefficients according to the measured results in this case and the above results at an
definition equations shown as Equations (12) and (13). angle of attack of 0 degree, it was found that at different
Horizontal axis represents reduced frequency k = fB/V angles of attack, different unsteady aerodynamic force
(f : frequency, B: deck width, V: mean wind speed). In coefficients have a dependence on Reynolds number even
these figures, the Theodorsen function is also shown for for the same bridge deck cross-section shape.
comparison. The unsteady drag coefficient is not shown Figures 15 and 16 show the measured unsteady
because it has little influence on the flutter of the bridge aerodynamic force coefficients at sections B and C, and they

Vibration Unsteady aerodynamic lift coefficient Unsteady aerodynamic moment coefficient

mode Real part Imaginary part Real part Imaginary part

0.5 0 0.2 0.00

−1 0.0 −0.25
0.0
CMh R (—)
CLh R (—)

CLh i (—)

CMh i (—)

Vertical −2 −0.2 −0.50

bending
−0.5
−3 −0.4 −0.75

−1.0 −4 −0.6 −1.00

0.000 0.075 0.150 0.000 0.075 0.150 0.000 0.075 0.150 0.000 0.075 0.150
Reduced frequency k = fB/V Reduced frequency k = fB/V Reduced frequency k = fB/V Reduced frequency k = f B/V

20 1.0 0.00
3.0

0.5
CLqR (—)

CMq i (—)
CMq R (—)
CLq i (—)

10 0.0 1.5 −0.25

Torsion

−0.5

0 −1.0 0.0 −0.50

0.000 0.075 0.150 0.000 0.075 0.150 0.000 0.075 0.150 0.000 0.075 0.150
Reduced frequency k = fB/V Reduced frequency k = fB/V Reduced frequency k = fB/V Reduced frequency k = f B/V

(note) : Re = 2.8 × 104 : Re = 3.4 × 105 : Re = 1.5 × 106

: Re = 4.1 × 104 : Re = 6.6 × 105 : Theodorsen function
: Re = 9.5 × 104 : Re = 1.0 × 106

Fig. 13 Unsteady aerodynamic force coefficients ( Deck section A, Angle of attack : a = 0 degree )

19
 Vo l . 40 N o . 1 Februar y 2007

Vibration Unsteady aerodynamic lift coefficient Unsteady aerodynamic moment coefficient

mode Real part Imaginary part Real part Imaginary part

0.5 0.0 0.2 0.0

0.1 −0.1
0.0 −0.5

CMh i (—)
CLh i (—)

CMh R (—)
CLh R (—)

Vertical 0.0 −0.2

bending
−0.5 −1.0
−0.1 −0.3

−1.0 −1.5 −0.2 −0.4

0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
Reduced frequency k = f B/V Reduced frequency k = f B/V Reduced frequency k = f B/V Reduced frequency k = f B/V

4 3 0.5 0.2

0.4 0.1
3 2
CLq i (—)

CMq R (—)

CMq i (—)
CLq R (—)

0.3 0.0
2 1
Torsion 0.2 −0.1
1 0 −0.2
0.1

0 −1 0.0 −0.3
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
Reduced frequency k = f B/V Reduced frequency k = f B/V Reduced frequency k = f B/V Reduced frequency k = f B/V

(note) : Re = 2.8 × 104 : Re = 3.4 × 105

: Re = 4.1 × 104 : Re = 6.6 × 105
: Re = 9.5 × 104 : Theodorsen function

Fig. 14 Unsteady aerodynamic force coefficients ( Deck section A, Angle of attack : a = +3 degrees )

Vibration Unsteady aerodynamic lift coefficient Unsteady aerodynamic moment coefficient

mode Real part Imaginary part Real part Imaginary part

0.5 0.0 0.2 0.0

0.1 −0.1
0.0 −0.5
CLhR (—)

CLh i (—)

CMh i (—)
CMhR (—)

Vertical 0.0 −0.2

bending
−0.5 −1.0
−0.1 −0.3

−1.0 −1.5 −0.2 −0.4

0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
Reduced frequency k = fB/V Reduced frequency k = f B/V Reduced frequency k = f B/V Reduced frequency k = fB/V

4 3 0.5 0.2

0.4 0.1
3 2
CMqR (—)

CMq i (—)
CLq i (—)
CLqR (—)

0.3 0.0
2 1
Torsion −0.1
0.2
1 0
0.1 −0.2

0 −1 0.0 −0.3
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
Reduced frequency k = fB/V Reduced frequency k = f B/V Reduced frequency k = f B/V Reduced frequency k = fB/V

(note) : Re = 2.8 × 104 : Re = 3.4 × 105

: Re = 4.1 × 104 : Re = 6.6 × 105
: Re = 9.5 × 104 : Theodorsen function

Fig. 15 Unsteady aerodynamic force coefficients ( Deck section B, Angle of attack : a = +3 degrees )

20
 Vo l . 40 N o . 1 Februar y 2007

Vibration Unsteady aerodynamic lift coefficient Unsteady aerodynamic moment coefficient

mode Real part Imaginary part Real part Imaginary part

0.5 0.0 0.2 0.0

0.1 −0.1
0.0 −0.5

CMhR (—)

CMh i (—)
CLhR (—)

CLh i (—)
Vertical 0.0 −0.2
bending
−0.5 −1.0
−0.1 −0.3

−1.0 −1.5 −0.2 −0.4

0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
Reduced frequency k = fB/V Reduced frequency k = f B/V Reduced frequency k = f B/V Reduced frequency k = f B/V

4 3 0.5 0.2

0.4 0.1
3 2

CMqR (—)

CMq i (—)
CLqR (—)

CLq i (—)

0.3 0.0
2 1
Torsion 0.2 −0.1
1 0
0.1 −0.2

0 −1 0.0 −0.3
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
Reduced frequency k = fB/V Reduced frequency k = f B/V Reduced frequency k = f B/V Reduced frequency k = f B/V

(note) : Re = 2.8 × 104 : Re = 3.4 × 105

: Re = 4.1 × 104 : Re = 6.6 × 105
: Re = 9.5 × 104 : Theodorsen function

Fig. 16 Unsteady aerodynamic force coefficients ( Deck section C, Angle of attack : a = +3 degrees )

are qualitatively similar to the Theodorsen function, and no −0.13

dependence on the Reynolds number is seen. As a reason for 0
this, it is thought that this is because these sections have a 80°
Wind 0
shape closer to a thin plate than Deck section A. +3° direction
−0.13
Of the unsteady aerodynamic force coefficients at (note) : Re = 3.9 × 104
Deck section A and angle of attack of +3 degrees in : Re = 3.4 × 105
: Re = 1.2 × 106
which Reynolds number effects were seen in Fig. 14,
CMqR and CMqi are the aerodynamic moment coefficients Fig. 17 Unsteady pressure coefficients : CPD
accompanied by torsional excitation. Paying attention to ( Deck section A, a = +3 degrees, q = 1 degree )
these coefficients, the mechanism of Reynolds number
effects is investigated from the measured results of nonlinearity may have influence on the flutter response
fluctuating pressure distribution at the model surface. characteristics, depending on the circumstances. However,
The fluctuating pressure distribution at the model surface flutter analysis ordinarily performed is premised on the
was measured for three different Reynolds numbers, and condition that the unsteady aerodynamic force coefficients
the results are shown in Fig. 17. As the Reynolds number are linear. Therefore, it is desirable to change the
increases, the fluctuating pressure peak on the bottom amplitude to see whether the unsteady aerodynamic force
surface of the upstream deck half shifts upstream. It can coefficients have amplitude dependence or not.(26), (27)
be thought that the magnitudes of CMqR and CMqi in Fig. As a past study on amplitude dependence, unsteady
14 increase with the Reynolds number for this reason. aerodynamic forces of “the Great Belt East Bridge,”
The tendency of the pressure distribution at the model Denmark, were measured using a 1/300-scale taut strip
surface changing with the Reynolds number was seen in model having a flattened hexagonal cross-section.(28) It is
Fig. 9, or in the mean pressure distribution measured when reported that no amplitude dependence was observed at
the model was at rest. As shown in Fig. 17, it was found a reduced wind speed Vr = V/fB (V: mean wind speed, f :
that the same tendency is seen in the fluctuating pressure frequency, B: deck width) of 10 or less, and some measure
distribution when the model is excited. of amplitude dependence was observed at more than 10. It
3.2.3 Amplitude dependence of unsteady must be noted, however, that the Reynolds number in this
aerodynamic force coefficients study is, taking the deck height D as the representative
In general, the unsteady aerodynamic force coefficients length, as small as Re = VD/v = 6.5 × 103 (V: mean wind
are not necessarily linear with respect to amplitude, and speed = 6.5 m/s, D : deck height = 4.4 m/300 = 0.014 7 m,

21
Vo l . 40 N o . 1 Februar y 2007

v : kinematic viscosity = 1.46 × 10-5 m2/s. a full bridge model of a long-span bridge, and static
In this study, Reynolds number dependence was observed displacement analysis and flutter analysis in strong winds
at Deck section A and an angle of attack of +3 degrees, are performed. By this, the effects of the Reynolds number
so the amplitude dependence of unsteady aerodynamic on the static displacement under wind load and flutter
force coefficients was investigated in this condition by characteristics are investigated.
changing the oscillation amplitudes of vertical bending and 4.1 Wind-induced static displacement analysis
torsion. For the model, a 1/10-scale bridge deck sectional 4.1.1 Analysis model and method
model was used. For the oscillation amplitude, a vertical The long-span bridge analyzed is a 3-span, 2-hinge
bending amplitude hm = Bm/100 (Bm : model deck width) stiffened box girder suspension bridge with a main span of
and torsional amplitude q = 1.0 degree in the Wind Tunnel 2 500 meters. Its structural dimensions are shown below.
Test Specification(29) were taken as reference, and a total Suspension structure type
of three sets of values, including one third and two thirds 3-span, 2-hinge stiffened box
of these values, were used. It is important to grasp the girder suspension bridge
unsteady aerodynamic force coefficient characteristics at Span divisions
small amplitudes because they have influence from the Cable 1 250 m + 2 500 m + 1 250 m
initial state of flutter. Thus, these smaller amplitudes than Deck 1 226 m + 2 480 m + 1 226 m
the reference were chosen. Measured results are shown Sag ratio 1/10
in Fig. 18. In any measured unsteady aerodynamic force Cable
coefficient, no significant difference due to different Tension 586 289 kN
oscillation amplitudes was observed. In a high Reynolds Cable-to-cable distance
number region which is smaller by a factor of 101 than the 35.5 m
Reynolds number of a real bridge, no amplitude dependence Diameter 1.324 m
of unsteady aerodynamic force coefficients was observed. Cross-sectional area
From this, the vertical bending amplitude hm = Bm/100 and 0.871 m2 per cable
the torsional amplitude q = 1 degree stipulated in the Wind Mass per unit length
Tunnel Test Specification are thought to be valid. Stiffening girder 23.09 t/m
(uniform over all spans)
4. Wind-induced response analysis of real Cable 7.66 t/m per cable
bridge (uniform over all spans)
Here, the steady and unsteady aerodynamic force Suspension structure
coeff icients obtained in Section 3 are applied to 38.41 t/m per bridge

Vibration Unsteady aerodynamic lift coefficient Unsteady aerodynamic moment coefficient

mode Real part Imaginary part Real part Imaginary part

0.5 0.0 0.2 0.0

0.0 0.1
CLh R (—)

−0.5
CMh R (—)

−0.1
CLh i (—)

CMh i (—)

Vertical
bending 0.0
−0.5 −1.0 −0.2
−0.1

−1.0 −1.5 −0.2 −0.3

0.10 0.15 0.20 0.10 0.15 0.20 0.10 0.15 0.20 0.10 0.15 0.20
Reduced frequency k = fB/V Reduced frequency k = fB/V Reduced frequency k = fB/V Reduced frequency k = f B/V

2 0.4 0.4 0.0

CMqR (—)
CLqi (—)

CMq i (—)
CLq R (—)

Torsion 1 0.2 0.2 −0.1

0 0.0 0.0 −0.2

0.10 0.15 0.20 0.10 0.15 0.20 0.10 0.15 0.20 0.10 0.15 0.20
Reduced frequency k = fB/V Reduced frequency k = fB/V Reduced frequency k = fB/V Reduced frequency k = f B/V

(note) : h = B/100 or 1 degree

: 2/3h or 2/3 degree
: 1/3h or 1/3 degree
: Theodorsen function

Fig. 18 Amplitude effects on unsteady aerodynamic force coefficients ( Deck section A, a = +3 degrees )

22
 Vo l . 40 N o . 1 Februar y 2007

(uniform over all spans) capable of finite displacement analysis including wind
Polar moment of inertia per unit length load imposition, removal or addition of structural members
Stiffening girder 2 712 t·m2/m and other shape change history of the structure. Using a
(uniform over all spans) method that increases load on a step-by-step incremental
Cable 4 827 t·m2/m per cable basis,(33) a wind load equivalent to a wind speed of 70 m/s
(uniform over all spans) was divided into 10 equal parts and successively imposed
Suspension structure on the structural members of the bridge deck, main towers,
7 539 t·m2/m per bridge cables and hangers of the analysis model. The bridge
(uniform over all spans) deck’s steady aerodynamic force coefficients used for
Section properties the analysis are the wind tunnel test values measured at
Cross-sectional area different Reynolds numbers as shown in Fig. 6. For the
1.314 m2 (uniform over all spans) main towers, cables and hangers, the steady aerodynamic
Vertical bending rigidity force coefficients in the Wind Resistant Design Code for
3.837 m4 (uniform over all spans) Honshu-Shikoku Bridges(34) were used.
Horizontal bending rigidity 4.1.2 Analysis results
146.4 m4 (uniform over all spans) Figure 20 shows the results of the wind-induced static
Torsional rigidity
9.700 m4 (uniform over all spans)
Main tower Stiffening girder Main tower Cable
The analysis model is an FEM analysis model of a
solid framework. The bridge deck and cable members
were divided into 20 parts in the main span and 10 parts
in the side spans, giving consideration to ensuring that
significant natural oscillation modes could be reproduced
in flutter analysis. A schematic of the long-span bridge
analysis model is shown in Fig. 19.
The wind-induced static displacements were analyzed
using the f inite element structural analysis program
AERODYNA(30) - (32) developed by IHI. This program is Fig. 19 Analytical model

Air flow
inclination
angle Horizontal bending displacement Vertical bending displacement Torsional displacement*1
(degree)
Horizontal bending displacement ( m )

Vertical bending displacement ( m )

25 2 0.0
Torsional displacement (degree)

Position of Position of
the main the main 1 −0.5
20 tower tower

15 0 −1.0
0
10 −1 −1.5

5 −2 −2.0

0 −3 −2.5
−3 000 −1 500 0 1 500 3 000 −3 000 −1 500 0 1 500 3 000 −3 000 −1 500 0 1 500 3 000
Distance from the center of the main span ( m ) Distance from the center of the main span ( m ) Distance from the center of the main span ( m )
Horizontal bending displacement ( m )

Vertical bending displacement ( m )

25 1
Torsional displacement (degree)

Position of Position of 0.0

the main the main
20 tower tower 0 −0.5
15
−1 −1.0
−3 10
−2 −1.5
5

0 −3 −2.0
−3 000 −1 500 0 1 500 3 000 −3 000 −1 500 0 1 500 3 000 −3 000 −1 500 0 1 500 3 000
Distance from the center of the main span ( m ) Distance from the center of the main span ( m ) Distance from the center of the main span ( m )

(note) : Re = 5.1 × 104 : Re = 4.7 × 105 : Re = 1.3 × 106

: Re = 1.0 × 105 : Re = 6.8 × 105 *1 : Pitch-down is negative
: Re = 2.7 × 105 : Re = 1.1 × 106

Fig. 20 Wind-induced static deformations at a wind speed of 70 m/s at deck height ( Inclination angle : 0 degree and -3 degrees )

23
Vo l . 40 N o . 1 Februar y 2007

displacement analysis of a stiffening girder. It can be seen a technique based on the modal analysis method (35)
that the horizontal bending displacement is not influenced was employed. For the unsteady aerodynamic force
by the Reynolds number. This is ascribable to the fact that coefficients, the values shown in Fig. 13 were used.
the drag coefficient is hardly influenced by the Reynolds 4.2.2 Analysis results
number in Fig. 6. On the other hand, the vertical bending Figure 21 shows the results of real bridge flutter
displacement and torsional displacement are influenced analysis.(21), (36) Up to a Reynolds number of 0.8 × 106, the
by the Reynolds number. That is, at an air flow inclination flutter wind speed was about 70 m/s independent of the
angle of 0 degree, the results of analysis using the steady Reynolds number. At higher Reynolds numbers, however,
aerodynamic force coefficients at a low Reynolds number the flutter wind speed somewhat increased. This tendency
tend to show a larger displacement than the results of of flutter wind speed’s increasing from a Reynolds
analysis using the coefficients in the high Reynolds number of 1.0 × 106 is the same as the results of a flutter
number region. This is due to the fact that, as shown in Fig. analysis(22) performed on a structural system in a spring-
6, the lift and moment coefficients in the low Reynolds supported test and using an analysis model. The latter is
number region are negative and their magnitudes are large the analysis result at an angle of attack of +3 degrees and
in the range of angle of attack from −2.5 to 0 degrees. At the analysis results in Fig. 21 are in the case of an air flow
an air flow inclination angle of −3 degrees, the results of inclination angle (angle of attack) of 0 degree. As above,
analysis using the steady aerodynamic force coefficients although the unsteady aerodynamic force coefficients
in the low Reynolds number region contrariwise tend to used for analysis were different, Reynolds number effects
show a smaller displacement than the results using the on flutter wind speed showed the same tendency not
coefficients in the high Reynolds number region. This dependent on angle of attack.
is ascribable to the fact that, as shown in Fig. 6, the lift That is, because the flutter wind speed shows a rising
and moment coefficients in the low Reynolds region are tendency from a Reynolds number of 1.0 × 106 and an
negative and their magnitudes are small in the range of ordinary flutter check is done by a wind tunnel test or
angle of attack from −5 to −3 degrees. flutter analysis in a low Reynolds number region of the
As stated above, it was found that the horizontal order of 104 to 105, such a check is thought to give an
bending displacement of a long-span suspension bridge evaluation on the safe side in the wind-resistant design.
is hardly influenced by the Reynolds number. Therefore, However, this is a conclusion restricted to the bridge deck
even if a static wind-resistant design is done using the cross-sections taken up in this study.
drag coefficient obtained by a wind tunnel test in a low
5. Conclusion
Reynolds number region, its results are thought to be valid
as a wind-resistant design. On the other hand, it was found As a result of the wind tunnel test and wind-induced
that the vertical bending and torsional displacements are response analysis performed in this study, the following
influenced by the Reynolds number. Thus, for the vertical findings were obtained. Note, however, that these findings
bending and torsional displacements, analysis should be are restricted to the bridge deck cross-sections investigated
performed using steady aerodynamic coefficients obtained in this study.
in a high Reynolds number region as close as possible to (1) A flutter analysis performed using the unsteady
the real bridge.
This also affects the analysis accuracy of aerodynamic 100 : Scale 1/80
stability of a long-span bridge in a strong wind. That : Scale 1/30
is, the vertical bending displacement of the bridge : Scale 1/10
deck contributes to the vertical bending rigidity of the 80
Flutter wind speed ( m/s )

whole suspension bridge system and thereby affects

the calculation of the natural frequencies and natural
60
oscillation modes of the whole suspension bridge system
in a strong wind. As to torsional displacement, the
unsteady aerodynamic force coefficient characteristics 40
may depend on the angle of attack when doing flutter
analysis, so it is very important to evaluate this torsional
displacement with high accuracy. 20
4.2 Flutter analysis
4.2.1 Analysis model and techniques
0
The analysis model is the same as the one used for the 104 105 106
wind-induced static displacement analysis in the previous Reynolds number*1
section. An eigenvalue analysis was performed in the *1 : Re = VD/v
state under static wind load, and the values obtained first Fig. 21 Flutter analysis results using a three-dimensional
to the 50th natural frequencies and natural oscillation analytical model with a main span of 2 500 m ( Inclination
modes were used for the analysis. For the flutter analysis, angle : 0 degree )

24
 Vo l . 40 N o . 1 Februar y 2007

aerodynamic coefficients obtained in an ordinary a rou nd more- or-less bluf f bod ies, Jou r nal of
low Reynolds number region gives an evaluation on Wind Engineering and Industrial Aerodynamics
the safe side in the wind-resistant design. 89 (2001) pp. 1 267-1 289
(2) Because no amplitude dependence of unsteady (7) U.Y. Jeong, H. M. Koh and H. S. Lee : Finite
aerodynamic coefficients was observed in a high element formulation for the analysis of turbulent
Reynolds number region which is smaller by a wind flow passing bluff structures using the RNG
factor of 101 than the Reynolds number of a real k- e model, Jou r nal of Wind Engineer ing and
bridge, the reference amplitude in Wind Tunnel Test Industrial Aerodynamics 90 (2002) pp. 151-169
Specification for Honshu-Shikoku Bridges (2001) is (8) P. W. Bear man : On vor tex shedding from a
valid. circular cylinder in the critical Reynolds number
The tasks for the future include measurement and regime, Journal of Fluid Mechanics 37 part 3 (1969)
investigation of the steady and unsteady aerodynamic pp. 577-585
force characteristics in a turbulent flow in a high Reynolds (9) G. Schewe : On the force fluctuation acting on a
number region. circular cylinder in crossflow from subcritical up
to transcritical Reynolds numbers, Journal of Fluid
– Acknowledgments – Mechanics 133 (1983) pp. 265-285
(10) K . M a t s u d a a n d S . Z a n : A e r o d y n a m i c
The authors sincerely thank Prof. Hiroshi Tanaka at characteristics of 2-dimensional circular cylinders
the University of Ottawa, Mr. Kevin Cooper, a former at high Reynolds numbers, Proceedings of the 33rd
NRC Senior Research Officer, Prof. Hitoshi Yamada Fluid Dynamics Conference Sep. 2001
at Yokohama National University Graduate School, (11) K. Shu, Y. Kubo, T. Hironaka and K. Tazaki :
Prof. Hiromichi Shirato at Kyoto University Graduate Reynolds number effects on Strouhal number of
School and Prof. Kichiro Kimura at Kyushu Institute of bluff structures, Proceedings of JSCE West Division
Technology Graduate School for their generous advice and Annual Meeting(I)-47 March 1992 pp. 114-115
cooperation in the implementation of this study. (12) F. Bleich : Dynamic instability of truss-stiffened
suspension br idges under wind action, Trans.
REFERENCES American Society of Civil Engineers 114 (1949)
(13) T. Theodorsen : General theory of aerodynamic
(1) Subcommittee on Similarity of Wind Tunnel instability and the mechanism of flutter, NACA TR
Experiments : Is a bridge oscillating due to winds? 496 (1935) pp. 413-433
Subcom m it tee on Si m ila r it y of Wi nd Tu n nel (14) F. Kaneko : Numerical Study of the FLUTTER
Experiments(the second period), Proceedings of the T h e o r y fo r Ae r o d y n a m i c S t a b i l i t y of L o n g
workshop of subcommittee on Similarity of Wind Span Suspension Bridges, Ishikawajima-Harima
Tunnel Experiments(the second period) July 1996 Engineering Review Vol.2 No.4 Jan. 1962 pp. 69-83
pp. 60-61 (15) H. Tanaka and M. Ito : The characteristics of the
(2) G. Schewe and A. Larsen : Reynolds number aerodynamic forces in self-excited oscillations of
effects in the flow around a bluff bridge deck cross bluff structures, Journal of Structural Mechanics
section, Journal of Wind Engineering and Industrial and Earthquake Engineering, JSCE, No.168 Aug.
Aerodynamics 74-76 (1998) pp. 829-838 1969 pp. 15-24
(3) C. Barre and G. Barnaud : High Reynolds number (16) Se e, fo r ex a m ple, S. Ku r o d a : Nu me r ic a l
simulation techniques and their application to shaped computation of unsteady aerodynamic forces for
structures model test, Proceedings of the 1st IAWE long-span bridge with two-equation turbulence
European and African Regional Conference on Wind model, Jou r nal of St r uct u r al Mecha n ics a nd
Engineering Guernsey UK (1993) pp. 83-93 Earthquake Engineering, JSCE No.654/(I)-52 July
(4) M. C. H. Hui and A. Larsen : Aerodynamic 2000 pp. 377-387
i n v e s t i g a t i o n fo r t h e d e c k of S t o n e c u t t e r s (17) S. Kawashima, H. K imu ra and T. Shibato :
Bridge emphasizing Reynolds number effects, An experiment for unsteady aerodynamic forces
Proceedings of the 2nd International Symposium act i ng on a wi ng, P roceed i ngs of 13th Japa n
on Advances in Wind & Structures (AWAS ‘02) National Congress for Applied Mechanics Sep.
Busan Korea (2002) pp. 649-656 1963 pp. 19-20
(5) G. L. Larose, S. V. Larsen, A. Larsen, M. Hui (18) N. Ukeguchi and H. Sakata : An investigation
and A. G. Jensen : Sectional model experience at of aeroelastic instability of suspension bridge,
high Reynolds number for the deck of a 1 018 m Journal of JSASS / Journal of the Japan Society for
span cable-stayed bridge, Proceedings of 11th Aeronautical and Space Science Vol.133 Feb.1965
International Conference on Wind Engineering pp. 27-36
Lubbock Texas USA (2003) pp. 373-380 (19) N. Ukeguchi, H. Sakata and H. Nishitani : An
(6) G. Schewe : Reynolds number effects in f low investigation of aeroelastic instability of suspension

25
Vo l . 40 N o . 1 Februar y 2007

bridges, International Symposium on Suspension pp. 967-976

Bridges Laboratorio Nacional de Engenharia Civil (28) G. L. Larose, A. G. Davenport and J. P. C. King
Lisbon (1966) pp. 273-284 : On the unsteady aerodynamic forces on a bridge
(20) R. H. Scanlan : Airfoil and bridge deck f lutter deck in turbulent flow, Proceedings of the 7th US
derivatives, Jour nal of Engineering Mechanics National Conference on Wind Engineering UCLA
Division, Proceedings of the American Society of USA (1993)
Civil Engineering 97 EM6 Dec. 1971 pp. 1 717-1 737 (29) Honshu-Shikoku Bridge Authority: Wind Tunnel
(21) K. Matsuda, K. R. Cooper and H. Tanaka : The Test Specification for Honshu-Shikoku Bridges
analysis of wind-induced static displacements and (2001) Aug. 2001
f lutter for long-span suspension bridges using (30) T. Yu k i a n d K . A n d o : St u d ie s o n Fi n it e
steady and unsteady aerodynamic forces measured Element Method for Structural Analysis - 4th
at high Reynolds numbers, Proceedings of 11th Report Analysis of Vibration, Buckling and Large
International Conference on Wind Engineering, Deformation for Framed Structure -, Ishikawajima-
Lubbock Texas USA (2003) pp. 649-656 Harima Engineering Review Vol.10 No.4 July 1970
(22) K. Matsud a, K. R. Cooper, H. Ta na ka, M. pp. 327-336
Tokushige and T. Iwasaki : An investigation of (31) T. Yuki, T. Shimada and Y. Hikami : Studies on
Reynolds number effects on the steady and unsteady Finite Element Method for Structural Analysis -
aerodynamic forces on a 1:10 scale bridge deck Large Deformation Structure Analsysis Program
section model, Journal of Wind Engineering and for Suspension Bridge and Plane Frame - , IHI
Industrial Aerodynamics 89 (2001) pp. 619-632 Engineering Review Vol.6 No.2 Sep. 1973 pp. 24-29
(23) Y. H i k a m i a n d K . M a t s u d a : O n t h e (32) K. Matsuda, H. Uejima, M. Tokushige and T.
characteristics of unsteady aerodynamic forces Iwasaki : Numerical Computation for Aeroelastic
on f lutter of long-span bridges, Proceedings of Stability of Long-Span Bridges, Ishikawajima-
the 50th JSCE Annual Meeting (I)-687 Sep. 1995 Harima Engineering Review Vol.37 No.6 Nov. 1997
pp. 1 374-1 375 pp. 411-417
(24) M . M a t s u m o t o , Y. K o b a y a s h i a n d H . (33) Y. H i k a m i , K . M a t s u d a a n d T. S u z u k i :
Hamasaki : On mechanism of coupled flutter for Nonlinear geometric and aerodynamic analysis
fundamental bluff bodies, Proceedings of the 13th fo r a lo n g- s p a n c a ble - s t aye d b r id ge d u r i n g
National Symposium on Wind Engineering Nov. construction, Wind Engineering Proceedings of
1994 pp. 359-364 the 1st IAWE European and African Regional
(25) M. Mat su mot o, Y. Da it o, F. Yosi z u m i, Y. Conference Guernsey UK (1993) pp. 431-440
Ichikawa and T. Yabutani : Torsional flutter of bluff (34) Ho n s hu - Sh i kok u Br id ge Au t ho r it y: Wi nd
bodies, Journal of Wind Engineering and Industrial Resistant Design Standard for Honshu-Shikoku
Aerodynamics 69-71 (1997) pp. 871-882 Bridges (2001) Aug. 2001
(26) Y. Kubo, M. Ito and T. Miyata : Nonlinear (35) T. J. A. Agar : Aerodynamic f lutter analysis
analysis of aerodynamic response of suspension of su s pe n sion br idge by a mod a l t ech n ique,
bridges in wind, Journal of Structural Mechanics Engineering Structure Vol.11 April 1989 pp. 75-82
and Earthquake Engineering, JSCE No.252 1976 (36) K. Matsuda, K. R. Cooper and H. Tanaka :
pp. 35-46 Reynolds number effects on wind-induced static
(27) H. Yamada, T. Miyata, H. Katsuchi, T. Suzuki displacements and flutter for long-span suspension
and H. Sugiura : A study on non-linear extension bridges, Journal of Wind Engineering, JAWE, No.95
of unsteady aerodynamic force definition, Journal April pp. 109-110
of Structural Engineering Vol.47A March 2001

26