The Seventh International Colloquium on Bluff Body Aerodynamics and Applications (BBAA7)

Shanghai, China; September 2-6, 2012

Direct simulation method for flutter stability of bridge deck

Xiaobing Liua,b Zhengqing Chenb Zhiwen Liub
a
Shijiazhuang Tiedao University, Shijiazhuang, China
b
Hunan University, Changsha, China

ABSTRACT: A direct simulation method for flutter stability of bridge deck was developed
based on CFD software Fluent. Flutter stability of two representative sections, rectangular
section and streamlined section, was simulated by using this method. The validity of the
simulation method was verified by comparison of simulation result and wind tunnel test result.

KEYWORDS: Bridge Deck Flutter Stability Direct Simulation

1 INTRODUCTION

Bridges become more and more flexible with the increase of span. Wind–resistant design of
bridge becomes more and more important. Flutter is a kind of divergent vibration style which
may occur for long-span flexible bridges. The flutter of bridge decks can be classified into two
categories. One is single-degree torsional flutter which may occur for bridge decks with sharp
angles. The other is bending-torsional coupled flutter which may occur for bridges with
streamlined decks. Flutter may destroy bridge. Therefore, it is very necessary to ensure that the
flutter critical wind speed is higher than flutter checking wind speed in design stage of long-span
bridges. At present, there are two main simulation methods for flutter stability of bridge deck.
One is based on flutter derivatives which are gotten by simulated forced motion of bridge deck.
The flutter critical wind speed of bridge deck is calculated by using the flutter derivatives. The
other one is direct simulation method. The vibration displacements of bridge deck in different
wind speed are simulated. The flutter critical wind speed of bridge deck is gotten by checking
vibration displacements. In the past, researches paid more attention to the first simulation method
[1-7]
, while less attention to on direct simulation method [8-9]. The study on direct simulation
method is almost based on self-written program which is short of applicability.
In recent years, with rapid development of computer technique, some universal CFD software,
such as Fluent, CFX, and so on, were adopted by bridge design organization because of good
interface, convenient pre-processor and post-processor, open secondary developing function and
so on. It is meaningful to perform direct simulation of flutter stability based on these universal
CFD software to evaluate the flutter critical wind speed of the bridge deck in preliminary design
stage. In present paper, a direct simulation method for flutter stability of bridge deck was
developed based on CFD software Fluent. Flutter stability of two representative sections,
rectangular section and streamlined section, was simulated by using this method. The validity of
the simulation method was verified by comparison of simulation result and wind tunnel test
result.

1528
2. DIRECT SIMULATION METHOD

First of all, the bridge deck is simplified to bending-torsional two degree spring-mass-damping
system. The dynamic equation is as follows:
xx x xx x
m h(t )  ch h(t )  k h h(t ) L(t ) I D (t )  cD D (t )  kD D (t ) M (t ) (1)
xx x
where h (t ) , h (t ) and h(t ) are instantaneous bending acceleration, bending velocity and bending
xx x
displacement; D (t ) , D (t ) and D (t ) are instantaneous torsional acceleration, torsional velocity

and torsional displacement; L(t ) and M (t ) are instantaneous lift force and torque on bridge deck.

N-S ᮍ⿟∖㾷
Initial condition
Flow field ˖Simulation of flow over stationary deck
x xx
Bridge deck˖ h (0) 0, h (0) 0, h (0) 0
x xx
D (0) 0 D (0) 0, D (0) 0

Solution of N-S equation

Getting lift force L(t) and torque M(t)

Solving dynamical equation using Newmark-£method

xx x xx x Renew grid
m h  ch h k h h L (t ) I D  cD D  k D D M (t )

x xx x xx
h(t ), h(t ), h(t )ˈD (t ) D (t ), D (t )

NO
Is displacement divergent or decaying obviously?

YES
Ending simulation
Figure 1 Procedure of direct simulation in every wind speed

The bridge deck vibration displacements under several wind speeds need to be calculated in
direct simulation method. In order to improve efficiency, the flutter critical wind speed can be
estimated through Wind–resistant Design Specification for Bridge at first. Then the direct

1529
The Seventh International Colloquium on Bluff Body Aerodynamics and Applications (BBAA7)
Shanghai, China; September 2-6, 2012

simulation can be carried out around the estimated wind speed. The procedure of direct
simulation in every wind speed is shown in Figure 1. Firstly, flow around static bridge deck is
performed. When the flow is stable, the bending degree and torsional degree are unrestricted.
Before first calculating time step, the preliminary value of bending acceleration, bending velocity,
bending displacement, torsional acceleration, torsional velocity and torsional displacement are
set to zero. For every time step, the fluid control equation, Navier-Stocks equation is solved and
the lift force and torque on the bridge deck is gotten, then the bending and torsional
displacements can be obtained by solving equation (1) using Newmark-ȕ method. Simulation in
given wind speed ends if the displacement divergent or decaying obviously.

3. SIMULATION MODEL

In present paper, flutter stability of two representative sections, rectangular section and
streamlined section, was simulated. The dimensions of these two representative sections are
shown in Figure 2. The width and height of rectangular section are 30cm and 6cm. The width
and height of streamlined section are 40.392cm and 6cm. The numerical simulation region and
boundary condition are shown in Figure 3. The simulation region is two dimensional rectangular
region. The distance between inlet and middle point of simulation region is 10times the width of
section. The distance between outlet and middle point of simulation region is 20 times the width
of section. The distance between top wall and bottom wall is 20 times the width of section The
boundary conditions are set as follows: velocity inlet boundary condition for inlet; outflow outlet
boundary condition for outlet; free-slip wall boundary condition for top and bottom wall; no-slip
wall boundary condition for surface of deck section.
e
%

E

E
% %

(a) rectangular section (b) streamlined section

Figure 2 Dimension of two representative sections

t o p w a ll

deck
in le t o u t le t
%

% %

b o t t o m w a ll

Figure 3 Numerical simulation region and boundary condition

1530
Figure 4 and Figure 5 shows numerical grid of the two sections. Rigid mesh grid is used in inner
region. Stationary mesh grid is used in outer region. Between the rigid mesh grid and stationary
mesh grid is dynamic mesh grid. The width and height of rigid mesh are about two times and one
time the width of section, respectively. Both the width and height of dynamic mesh grid are
about six times the width of section. When deck section vibrates, the rigid mesh grid follows the
deck section synchronously; the static grid keeps stationary; the shape and dimension of dynamic
mesh changes constantly. The number of mesh grid for rectangular section and streamlined
section are about 76800 and 62300, respectively. In region which is far away the section, the size
of mesh grid is big comparatively. In region which is near the section, the size of mesh grid is
small comparatively.

Deck section
Stationary grid

Dynamic grid

Rigid grid

Figure 4 illustration of numerical grid division

(a) rectangular section

(b) streamlined section

Figure 5 Numerical grid of two representative sections
In order to validate the direct simulation result, wind tunnel tests were performed to measure the
flutter stability of these two sections. The photos of wind tunnel tests are shown in Figure 6. The
length of sectional model is 1.5 meter. The test parameters, including model mass, mass moment
of inertia, bending vibration frequency, torsional vibration frequency and damping ratio, are

1531
The Seventh International Colloquium on Bluff Body Aerodynamics and Applications (BBAA7)
Shanghai, China; September 2-6, 2012

listed in Table 1. The direct numerical simulation parameters are in accordance with wind tunnel
test parameters. The only difference is that the mass and mass moment of inertia per unit length
are used in direct numerical simulation.

(a) rectangular section (b) streamlined section

Figure 6 Photo of wind tunnel test

Table 1 Parameters of wind tunnel test

section type m(kg) I(kg m2) f h (Hz) f D (Hz) [h [D

rectangular section 11.89 0.1280 3.13 6.64 0.3% 0.3%

streamlined section 13.38 0.1505 2.54 5.27 0.3% 0.3%

4 SIMULATION RESULT

Figure 7 and Figure 8 show curves of torsional angular displacement versus time history of
streamlined section and rectangular section, respectively. Figure 9 shows amplitude spectrum of
curves of torsional angular displacement versus time history. In these figures, the reduced wind
speed is defined as follows:
U
Vr (2)
ft B
Where U is incoming wind speed, ft is torsional vibration frequency of model, B is width of
model.

1532
  


WRUVLRQDODQJXODUGLVSODFHPHQW˄e˅

WRUVLRQDODQJXODUGLVSODFHPHQW˄e˅






 








 
             

WLPH˄ V˅ WLPH˄ V˅

(a) Vr=10.80 (b) Vr=11.27



WRUVLRQDODQJXODUGLVSODFHPHQW˄e˅



WRUVLRQDODQJXODUGLVSODFHPHQW˄e˅






 









     
      
WLPH˄ V˅
WLPH˄ V˅

(c) Vr=11.51 (d) Vr=11.74

Figure 7 Curves of torsional angular displacement versus time history of streamlined section

 
WRUVLRQDODQJXODUGLVSODFHPHQW˄e˅

WRUVLRQDODQJXODUGLVSODFHPHQW˄e˅

 

 

 

 

 

 
                 
WLPH˄ V˅ WLPH˄ V˅

(a) Vr=3.29 (b) Vr=3.77

 

 
WRUVLRQDODQJXODUGLVSODFHPHQW˄e˅

WRUVLRQDODQJXODUGLVSODFHPHQW˄e˅

 

 

 

 

 

 

 
           
WLPH˄ V˅ WLPH˄ V˅

(c) Vr=4.00 (d) Vr=4.24

Figure 8 Curves of torsional angular displacement versus time history of rectangular section

1533
The Seventh International Colloquium on Bluff Body Aerodynamics and Applications (BBAA7)
Shanghai, China; September 2-6, 2012

 




 ; < 
; < 


DPSOLWXGH



DPSOLWXGH







      
     
IUHTXHQF\˄ +]˅
IUHTXHQF\˄ +]˅

(a) Streamlined section Vr=11.74 (b) rectangular section Vr=4.0

Figure 9 Amplitude spectrum of curves of torsional angular displacement versus time history

From Figure7 and Figure 9(a), it can be seen that: (1) when reduced wind speed is less than11.51,
torsional angular displacement of streamlined section is decreasing with increase of time. This
illustrates that the total damping of model is positive. (2)When reduced wind speed is less
than11.51, with increase of reduced wind speed, torsional angular displacement of streamlined
section is decreasing slowly. This suggests that although the total damping of model is positive,
the absolute value of total damping becomes small with increase of reduced wind speed.
(3)When reduced wind speed reaches to 11.74, torsional angular displacement of streamlined
section is increasing with the increase of time. This illustrates that the total damping of model
changes from positive to negative. Flutter critical wind speed is around 11.74. (4) When flutter
occurs, the torsional vibration frequency is 3.10Hz. This frequency is between natural bending
vibration frequency and torsional vibration frequency. This suggests that flutter style for
streamlined section is bending-torsional coupled flutter.
From Figure8 and Figure 9(b), it can be seen that: (1) When reduced wind speed is less than3.77,
torsional angular displacement of rectangular section is very small and is decreasing with the
increase of time. This illustrates that the total damping of model is positive. (2) When reduced
wind speed reaches to 4.0, torsional angular displacement of rectangular section is increasing
dramatically with increase of time. What is different from streamlined section is that vibration
amplitude of rectangular section keeps a constant value at last. When reduced wind speed
continues to increase, the amplitude increases to a greater constant value. This illustrates that the
flutter critical wind speed of rectangular section is around 4.0. The rectangular section shows
obvious soft flutter characteristic. (3) When flutter occurs, the torsional vibration frequency is
6.52Hz. This frequency is very close to nature torsional vibration frequency. This suggests that
flutter style for rectangular section is single-degree torsional flutter.
In Table 2, direct numerical simulation result and wind tunnel test result were compared. It can
be seen that direct simulation results are in good agreement with wind tunnel test results. This
shows that the direct simulation method used in present paper is correct.

1534
 Table2. Compare of wind tunnel test result and direct simulation result
section type flutter stability wind tunnel test direct simulation
non dimensional flutter critical wind
4.00 3.77~4.00
rectangular section speed
flutter frequency (Hz) 6.63 6.52
non dimensional flutter critical wind
10.67 11.51~11.74
streamlined section speed
flutter frequency (Hz) 3.99 3.10

5 CONCLUSION

In present paper, a direct simulation method for flutter stability of bridge deck was developed
based on CFD software Fluent. Flutter stability of two representative sections, rectangular
section and streamlined section, was simulated by using this method. The validity of the
simulation method was verified by comparison of simulation result and wind tunnel test result. In
further study, the flutter stability of bridge deck with more complex geometry, such as Ȇ-shaped
section, separate twin box section and so on, will be calculated using this simulation method.

6 REFERENCES

[1] Walther J H, Larsen A. Two dimensional discrete vortex method for application to bluff body aerodynamics.
Journal of Wind Engineering and Industrial Aerodynamics, 1997, 67-68: 183-193
[2] Larsen A, Walther J H. Aeroelastic analysis of bridge girder sections based on discrete vortex simulations.
Journal of Wind Engineering and Industrial Aerodynamics, 1997, 67-68: 253-265
[3] Fujiwara A, Kataoka H, Ito M. Numerical simulation of flow field around an oscillating bridge using finite
difference method. Journal of Wind Engineering and Industrial Aerodynamics, 1993, 46-47: 567-575
[4] Vairo G. A numerical model for wind loads simulation on long-span bridges. Simulation Modelling Practice
and Theory, 2003, 11(5-6): 315-351
[5] Cao Fengchan, Xiang Haifan, Chen Airong. Numerical Assessment of Aerodynamic Derivatives of Thin Flat
Plates. Journal of Tongji University. 1999, 27(2): 131-135
[6] Zhu Zhiwen, Chen Zhengqing, Chen Weifang. Assessment of Aerodynamic Derivatives of the Ideal Flat
Plate Based on the Moving Grid Method., Journal of National University Of Defense Technology. 2002,
24(3): 13-17
[7] Zhou Zhiyong, Chen Airong Xiang Haifan. Assessment of Aerodynamic Derivatives and Critical Wind
Speed of Flutter of Bridge Decks by Discrete Vortex Method. Journal of Vibration Engineering. 2002, 15(3):
327-331
[8] Selvam R P, Govindaswamy S, Bosch H. Aeroelastic analysis of bridges using FEM and moving grids. Wind
and Structures, 2002, 5(2-4): 257-266
[9] Frandsen J B. Numerical bridge deck studies using finite elements. Partĉ˖flutter. Journal of Fluids and
Structures, 2004, 19: 171-191

1535