Aerostatic & Aerodynamic Effects on

Long-Span Cable Stayed Bridge

Practical Case: Balang Island Bridge
[UI-Midas Civil Engineering Forum, Jakarta 3rd June 2021]

FX Supartono
Team Leader, Balang Bridge Design
CEO, PT Midasindo Teknik Utama
Aeroelasticity is a science that concerned with the relation between an elastic
structure and the dynamic forces resulting from a wind excitation (fluid movement)
around the structure, as explained by the Collar’s Aeroelastic Interaction Triangle.

Fluid Mechanics

Aerodynamic Aerostatic
Stability Stability

Structural Dynamics Structural Mechanics

[Source: A.R. Collar, 1946]

The structural dynamics of a long-span bridge show usually very long
fundamental periods (TB can be as long as several seconds) due to the
flexibility of their long-span deck, which is an aspect that differentiates them
from other structures.
T = 3.394 sec
In many cases, due to its long fundamental period, the structural response of a
long-span bridge under an earthquake excitation is not a critical point to its
failure risk, compared with the aerodynamic instability risk.

[Source: Web images]

In fact the flexibility and dynamic characteristics of the cable stayed bridges
depend on various parameters such as the main span length, deck width,
stay cable system and their arrangement, girder-pylons and girder-piers
support conditions, and many other aspects.

This structural typology is complex, consisting on several structural

components with different individual stiffness and damping properties, such
as pylon’s properties and deck’s properties. Therefore, it needs a more
comprehensive dynamic analysis for its wind resistance, and needs to
accurately evaluate their periods, mode shapes and damping
characteristics, particularly on the negative damping development during an
aerodynamic excitation.
Aerodynamic Effect on
the Long-Span Bridges
 Aerodynamic Wind Effect

Fv
M
Fh

Under aerodynamic wind attack, the long-span cable stayed bridges are
usually susceptible on the following risks:
 Cable vibration
 Limited amplitude response such as vortex-induced oscillations or buffeting
 Divergent amplitude response (increasing indefinitely) such as flutter
 Cable Vibration PYLON
WAX/GREASE

As the cables interact directly Internal damper

ANCHOR CAP

with air fluid movement, it located at the

SENSOR
E CABLE

can generate a cable extremity of the WEDGE

vibration that may occur in guide pipe LOAD CELL

(IF REQUIRED)
different velocity ranges and ADJUSTABLE
ANCHORAGE
may be influenced also by BEARING PLATE
HDPE PIPE
the associated rain. The CONNECTOR GUIDE PIPE

cable vibrations depend on HDPE PIPE HDPE SHEATHED STRAND

COLLAR
its geometry, stiffness, and COLLAR
INTERNAL DAMPER

mass. This phenomenon WATERPROOF CAP

INTERNAL DAMPER
does in general not cause GUIDE PIPE
EDGE BEAM

cable rupture, however some

events show large amplitude
and high frequency so that WAX / GREASE
BEARING PLATE
dampers have to be installed. FIXED ANCHORAGE
ANCHOR CAP WEDGE
External Damper PYLON
WAX/GREASE
ANCHOR CAP

SENSOR
E CABLE

WEDGE

LOAD CELL
(IF REQUIRED)

ADJUSTABLE
ANCHORAGE
BEARING PLATE
HDPE PIPE
CONNECTOR GUIDE PIPE

HDPE PIPE HDPE SHEATHED STRAND

COLLAR INTERNAL DAMPER

WATERPROOF CAP

INTERNAL DAMPER
EDGE BEAM
GUIDE PIPE

External damper,
usually installed for
BEARING PLATE
WAX / GREASE cable length > 250 m
FIXED ANCHORAGE
ANCHOR CAP WEDGE
 Limited Amplitude Response

Regarding the bridge deck, there are several types of vibration that may
occur due to aerodynamic excitation:
 Vortex-induced oscillations: These are oscillations of limited amplitude
excited by the periodic cross-wind forces arising from the shedding of
vortices alternatively from the upper and lower surfaces of the bridge
deck. They can occur over one or more limited ranges of wind speeds.
The frequency of excitation may be close enough to a natural
frequency of the structure to cause the resonance and, consequently,
cross-wind oscillations at that frequency.
These oscillations occur usually in isolated (minimal interacted)
vertical bending and torsional modes.
 Limited Amplitude Response

 Turbulence response: Because of its turbulent nature, the forces

and moments developed by wind on bridge decks fluctuate over a
wide range of frequencies. If sufficient energy is present in
frequency bands surrounding one or more natural frequencies of the
structure, vibration may occur such as buffeting, like irregular
shock/oscillation that can occur on an aircraft caused by turbulence.
How About the Divergent Amplitude Response?

Tacoma Narrows Bridge, USA (1940)

Opened: 1 July 1940
Collapsed: 7 November 1940
 Divergent Amplitude Response

 Classical flutter: this involves coupling (i.e. interaction) between the

vertical bending and torsional oscillations.
 Galloping and stall flutter: galloping instabilities arise on certain
shapes of deck cross-section because of the characteristic of
variation of the wind drag, lift and pitching moments (torsional up
and down) with angle of incidence.

Fv
M
Fh
Illustration of Flutter Phenomenon on the Bridge Deck
z Wind

Lift
t
θ
 Divergent Amplitude Response

Unfortunately, in certain case, for even a short-span bridge, a bad-

shaped cross section combined with a lightweight superstructure, can
bring the galloping type aerodynamic instability on the bridge deck, like
what happened on Yadagawa River Bridge in Nagoya, Japan, a three-
span steel box girder bridge with 84.2 m main span and two side spans
of 67.1 m, and 7.5 m of deck width. The cross section was consisting of
double box girder with h = 2.20 to 3.20 m, and a parapet of 1.50 m high.
(Source: Y.J. Ge et al / J. Wind Eng. Ind. Aerodyn. 90-2002)
Therefore we have to be very careful on the aerodynamic instability that
may happen not only in long-span bridges, usually with flutter type, but
also in short-span bridges with galloping type.
 Divergent Amplitude Response
Divergence (increasing
indefinitely) can occur if

Damping Ratio
the deck receives more
wind attack energy that
can no more be
dissipated by structural
damping, usually called
Positive Damping
as negative damping
effect, which can
produce a rapidly Negative Damping
increasing amplitude and
no limit until collapse of Wind Speed
the bridge.
 Positive Damping Effect
z

t
 Aerodynamic Negative Damping Effect
z

t
 Balang Bridge
Structural Modeling
Balang Island Bridge is a cable stayed concrete bridge having main
span length of 402 meters (total length: 804 meters) connecting
Tempadung (near Balikpapan) and Balang Island.
 CONSTRUCTION COMPLETED 2021
Penajam
Paser Utara

Balang Island

DED by FX Supartono & Team

Construction Completed 2021

DED by FX Supartono & Team

Construction Completed 2015
Balikpapan
 World Longest Cable Stayed Concrete Bridge

Longest Bridge Length Longest Span Length

SKARNSUNDET BR. 1010 m 240 530 240 1988-1991 SKARNSUNDET BR. 1010 m 240 530 240
ATLANTIC BRIDGE 990 m 230 530 230 2013-2019 ATLANTIC BRIDGE 990 m 230 530 230
VASCO DA GAMA BR. 826 m 203 420 203 1995-1998 SHANTOU BAY BRIDGE 760 m 154 452 154
BALANG ISLAND BR. 804 m 201 402 201 2015-2020 CARLOS F CASADO BR. 643.4 m 101.7 440 101.7
IROISE BRIDGE 800 m 200 400 200 1991-1994 HELGELAND BRIDGE 780 m 177. 5 425 177. 5
HELGELAND BRIDGE 780 m 177. 5 425 177. 5 1989-1991 VASCO DA GAMA BR. 826 m 203 420 203

WADI LEBAN BRIDGE 763 m 179 405 179 1993-1997 WADI LEBAN BRIDGE 763 m 179 405 179

SHANTOU BAY BRIDGE 760 m 154 452 154 1992-1995 BALANG ISLAND BR. 804 m 201 402 201
2nd WUHAN YANGTZE 760 m 180 400 180 1991-1995 IROISE BRIDGE 800 m 200 400 200
CARLOS F CASADO BR. 643.4 m 101.7 440 101.7 1981-1983 2nd WUHAN YANGTZE 760 m 180 400 180

SIDE-
SPAN
MAINSPAN SIDESPAN
 Structural Modeling
[Midas Civil v.2012]

For a long-span bridge such as cable stayed bridge or suspension bridge,

the structural modeling should be in 3D so that the structural detailing can
be well modeled.
108.75m

1.60m 1m
2.60m

2.96m 1.24m 3.50m 3.50m 1m 3.50m 3.50m 1.24m 2.96m

1.86m 18.48m 1.86m
0.60m 0.60m

Bridge Deck Section

Pylon
Longitudinal Free
Movable Bearing
Balanced cantilever
by using form traveler

Cast-in-place by shoring
 One side single
cantilever by using
form traveler

Increase the structural stability during the longest

cantilever construction stage, particularly against
the aerodynamic (wind) and seismic risk
Viscous dampers are installed in
the deck (edge beams) and to be
connected to the side span’s piers
 Aerostatic &
Aerodynamic
Susceptibility Analysis
Aerostatic Effect
 Static Wind Load [Aerostatic]

Wind
FD = ½ ρ Vg2 CD H
Vg = Gv · Vd
where:
FD = wind load at the wind attack direction (N/m)
ρ = mass density of air, usually taken as 1.25 (kg/m3)
Vg = geotropic wind speed at the bridge deck height z (m/sec)
Vd = design wind speed at the bridge deck height z (m/sec)
CD = drag coefficient, defined by Wind Tunnel Test or according to the Code
H = depth of the deck including railing fence for service state (m)
Gv = geotropic wind coefficient
 Combined Effects of Static & Dynamic Wind Loads
In general, the wind load can be calculated in 3 different effects,
namely the effect of static wind effect, dynamic wind effect, and
dynamic effect that causes vibrations in bridge components, as a
result of amplification of resonances that occur in some
components of the bridge.
By empirical approach, the combined effects of static and
dynamic wind loads can be calculated as the static wind load
multiplied by a coefficient of static wind Gv, which depends on the
site category and the span length or longest cantilever length.

Vg = Gv · Vd
 Static Wind Load [Aerostatic]

For the other bridge components such as pylon, stay cables, piers, etc,
it can be calculated:

FH = ½ ρ Vg2 CH An
where:
FH = wind load at the wind attack direction (N)
CH = drag coefficient for each bridge component
An = effective area of each bridge component, and for the stay cable
is taken as the external diameter multiplied by the effective
length (m2)
 FL FV
FD

α
FH
M

Wind
 Wind Effect Site Categories & Coefficients

Vd = Vs10 (z/10)α m/sec

Category Exponential Value α

Class A 0.12
Class B 0.16
Class C 0.22
Class D 0.3

Class A : sea, sea shores, islands, lake and deserts

Class B : open fields, villages, forests, hills, sparsely-built town, and suburbs
of the cities
Class C : urban area of densely-populated cities
Class D : center of large city with closely spaced tall buildings
[Source: Chinese Standard JTG/T D60-01-2004]
 Design Wind Speed

Wind speed Bridge Deck Pylon

Basic wind speed Vs10 (m/sec)

30.00 30.00
[RSNI T-02-2005: Service State]

Site surface coefficient α 0.16 0.16

Design height z (meter) 32.96 70.04

Design wind speed at the height z

36.31 40.96
Vd = Vs10 (z/10)α (m/sec)
Design wind speed at construction stage
= η Vd (m/sec), η is a reduction factor 30.50 34.41
(return period of 10 years, η = 0.84)
 Static Design Wind Load
Design wind load on the bridge deck Completed Bridge Construction Stage
Design wind speed Vd (m/det) 36.31 30.50
Average height of deck section H (meter) 4.10 2.60
Drag coefficient CD (Indonesian Code / RSNI) 1.50 1.50
Drag coefficient CD (China Code) 1.68 1.68
Drag coefficient CD (Tsingsa River Bridge) 1.63 1.36
Drag coefficient CD (Ming River Bridge) 1.53 －
Geotropic wind speed Vg (m/sec), GV = 1.22 44.30 37.21
Static wind load FD (kN/m), CD=1.50 7.54 3.37

Design wind load on the pylon Completed Bridge Construction Stage

Design wind speed Vd (m/sec) 40.96 34.41
Average width of pylon B (meter) 5.9 5.9
Drag coefficient CD (China Code) 2.10 2.10
Geotropic wind speed Vg (m/sec), GV = 1.22 49.97 41.98
Static wind load FD (kN/m) 19.34 13.65
Wind Load Distribution in the Bridge Deck
Load Case 1
Wind Load Distribution in the Bridge Deck
Load Case 2
Wind Load Distribution in the Pylon
Load Case 1 & 2
Wind Load Distribution in the Pylon
Stress in the Deck due to Combined Forces
of Service Loads & Static Wind Loads (MPa)
Stress in the Pylon due to Combined Forces
of Service Loads & Static Wind Loads (MPa)
Stress in the Pylon due to Combined Forces
of Service Loads & Static Wind Loads (MPa)
Structural Dynamic Analysis
 Free Vibration Dynamic Analysis

The model used for the dynamic analysis of Balang Bridge was 3D
model in order to obtain a clearer overview on the dynamic behavior of
the structure, both vertical and lateral as well as longitudinally.
An accurate analysis of natural frequencies and modal shapes on a
cable stayed bridge is very important, not only for the study of the
aerodynamic stability effects, but also for the other transient loads’
excitation such as seismic loads and traffic loads.
Completed Bridge Model
Mode Shape 1 (f = 0.295 Hz)
Mode Shape 2 (f = 0.320 Hz)
Mode Shape 3 (f = 0.392 Hz)
Mode Shape 6 (f = 0.663 Hz)
 Mode Shape 1 - 10 at Service State
Completed Bridge State
Mode Frequency [Hz] Mode shape type
1 0.295 First order vertical bending of the deck
2 0.320 Longitudinal drift
3 0.392 First order lateral bending of the deck
4 0.398 Longitudinal drift
5 0.558 Second order vertical bending of the deck
6 0.663 First order torsion of the deck
7 0.697 Longitudinal drift
8 0.742 Third order vertical bending of the deck
9 0.799 First order antisym vertical bending of the deck
10 0.928 Fourth order vertical bending of the deck
As the bridge will be built in a segmental way, therefore an aeroelastic
stability analysis should be also performed during the construction stage,
particularly for the longest double cantilever state (just before side closure)
and the longest single cantilever state (just before middle closure).
For that purpose, a dynamic
analysis was also performed on
those longest cantilever states.
Longest Double Cantilever State
Longest Single Cantilever State
Mode Shape 1 - 10 at the Longest Single Cantilever State
Construction Stage
Mode Frequency [Hz] Mode shape type
1 0.284 First order vertical bending of the deck
2 0.302 First order lateral bending of the deck
3 0.490 Second order lateral bending of the deck
4 0.655 First order torsion of the deck
5 0.710 Second order vertical bending of the deck
6 0.829 Third order vertical bending of the deck
7 0.986 Fourth order vertical bending of the deck
8 1.186 Second order torsion of the deck
9 1.265 Longitudinal drift
10 1.279 Third order torsion of the deck
Aerodynamic Susceptibility Analysis
 Aerodynamic Wind Effect
Fv
M
Fh

Under aerodynamic wind attack, the bridge deck can be susceptible

on the following risks:
 Limited amplitude response such as vortex-induced oscillations or buffeting
 Divergent amplitude response (increasing indefinitely) such as flutter
First Initial Checking of Aerodynamic Susceptibility

Undeformed State

Mode 1
Vertical Bending
fv = 0.295 Hz

Mode 6 Torsion
ft = 0.663 Hz
 First Initial Checking of Aerodynamic Susceptibility

Based on the results of the free vibration frequencies, the ratio of torsional
frequency to vertical bending frequency is:

.
R= = = 2.247 > 2.0
.

This is the first initial check point for the aerodynamic susceptibility
against flutter risk at completed bridge state (service state). According to
Jacques Mathivat [1979, Eyrolles French Edition], the bridge structure
shows a low susceptibility to the aerodynamic excitation. However, a
Section Model Wind Tunnel Test should be recommended in order to
confirm a more reliable level of aerodynamic stability.
 Second Initial Checking of Aerodynamic Susceptibility
Further checking can be performed according to BSI 2001 BD 49/01 “Design
Rules for Aerodynamic Effects on Bridges” for the purpose of preliminary
categorization. The following parameters’ values may be used to give an
indicative range of adequacy on the aerodynamic stability:
 Vr : between 20 & 40 m/sec (Balang Bridge: 36.31 m/sec)
 m/b : between 600 & 1200 kg/m2 (Balang Bridge: 1704 kg/m2)
 fB : between 50/L0.87 & 100/L0.87 that means between 0.271 & 0.542
(Balang Bridge: 0.295 Hz)
where: Vr = hourly mean wind speed, in our case was taken as Vd (m/sec)
m = mass per unit length of the bridge deck (kg/m)
b = overall width of the bridge deck (m)
fB = natural frequency in vertical bending (Hz)
 Aerodynamic Susceptibility

BSI 2001 BD 49/01 has also introduced some more comprehensive

criteria and rules, with the objective to verify analytically the aerodynamic
susceptibility risks, and to endeavor:
 The phenomenon of divergent amplitudes increasing rapidly to large
values must be avoided.
 The phenomenon of non-oscillatory divergence due to a form of
aerodynamic torsional instability must also be avoided.
 Aerodynamic Susceptibility
In that objective, BD49/01 introduced the aerodynamic susceptibility parameter
Pb that can be derived to categorize the structure using the equation:

𝜌𝑏 16𝑉
𝑃 =
𝑚 𝑏𝐿𝑓

where:
ρ = mass density of air (kg/m3)
b = overall width of the bridge deck (m)
m = mass per unit length of the bridge deck (kg/m3)
Vr = hourly mean wind speed, in our case was taken as Vd (m/sec)
L = length of the relevant maximum span of the bridge (meter)
fB = natural frequency in vertical bending (Hz)
 Aerodynamic Susceptibility

According to BSI 2001 BD 49/01, the bridge shall then be categorized as

follows:
1. Bridges design to carry the standard loadings, built of normal
construction, are considered to be subject to insignificant effects in
respect of all forms of aerodynamic excitation when Pb < 0.04.
However the rules can still be applied if required, provided the
constraints of positions of parapets, railings, central median barrier,
etc, as specified in §2.3 are satisfied.
 Aerodynamic Susceptibility

2. Bridges having 0.04 ≤ Pb ≤ 1.0 shall be considered to be within the

scope of these rules, provided the geometric constraints of §2.3 are
satisfied, and shall be considered adequate with regard to each
potential type of excitation if they satisfy the relevant criteria given in
§2.1.1 (vortex excitation), §2.1.2 (turbulence) and §2.1.3 (galloping
and flutter).
3. Bridges with Pb > 1.0 shall be considered to be potentially very
susceptible to aerodynamic excitation.
 Balang Bridge Aerodynamic Parameters
Parameter Description Value
Design wind speed (Vd) Completed bridge condition [m/s] 36.31
CS Design wind speed (Vd) Construction stage [m/s] 30.50
Bridge deck width (b) [meter] 23.40
Deck half-width (b/2) [meter] 11.70
Air density (ρ) [kg/m3] 1.25
Mass density of the bridge deck (m) [kg/m] 39883
Unit inertia (Im) [kg.m2/m] 2859750.65
Gyration radius (r) [m] 8.12
μf Correction factor 1.24
K Safety factor 1.2
η Reduction factor 0.5
Bending natural frequency (fb) Completed bridge [Hz] 0.295
Torsion mode frequency (ft) Completed bridge [Hz] 0.663
Frequency ratio (ε) Completed bridge 2.247
CS Bending natural frequency (fb) Construction stage [Hz] 0.334
CS Torsion mode frequency (ft) Construction stage [Hz] 0.637
CS Frequency ratio (ε) Construction stage 1.907
 Aerodynamic Susceptibility
For Balang Island Bridge:

1.25𝑥23.4 16𝑥36.31
𝑃 = = 0.44
39883 23.4𝑥402𝑥0.295

According to BSI 2001 BD 49/01, as having value of 0.04 ≤ Pb ≤ 1.0, it can be

considered that Balang Island Bridge satisfied the aerodynamic stability criteria
within the scope of the rules, and shall be considered adequate with regard to
aerodynamic excitation.
However, as some of the criteria such as for vortex excitation were not satisfied
(please see further slides), then the Wind Tunnel Test (Section Model) was
recommended for confirming the aerodynamic stability in a more accurate level.
 Limited Amplitude Response – Vortex Excitation

Another important rules for aerodynamic susceptibility is also given in BD

49/01 for vortex induced vibration by estimating the critical wind speed.
The critical wind speed for vortex excitation for both bending and torsion
(Vcr) can be derived according to BD 49/01 §2.1.1.2.
Vcr can be defined as the velocity of steady air flow or the mean velocity
of turbulent flow at which the aerodynamic excitation due to vortex
shedding occurs and can be calculated for both vertical bending and
torsional modes of vibration of box and plate girder bridges by the
formula in the next slide.
Alternatively Vcr may be determined by appropriate wind tunnel test on
suitable scale models.
 Limited Amplitude Response – Vortex Excitation

Vcr for bridge types 1, 1A, 3,

b*/d4 Vcr for bridge types 2, 5, 6
3A, 4, 4A
≤5 6.5·f·d4 6.5·f·d4
>5 and <10 f·d4 (1.1b*/d4+1.0) f·d4 (0.7b*/d4+3.0)
≥10 12·f·d4 10·f·d4
where:
b* = the effective width of the bridge
d4 = the depth of the bridge; if the depth is variable over the span, d4 shall be
taken as the average value over the middle third of the longest span
f = either fB or fT as appropriate, i.e. the natural frequencies in bending and
torsion respectively (Hz) calculated under dead and superimposed dead load
Limited Amplitude Response – Vortex Excitation
 Limited Amplitude Response – Vortex Excitation

For the case of Balang Bridge, there is no appropriate cross section in

the bridge type selection in BD 49/01. For a first approximation, the type
1A was selected for this case:

b* = 23.40 m
d4 = 2.60 m
Bridge Cross Section
b*/d4 = 9

Vcr = f·d4 (1.1b*/d4+1.0) = 0.295 · 2.60 (1.1 · 23.40 / 2.60 + 1.0)

= 8.36 m/sec < Vvs (please see the Acceptance Criteria)
 Vortex Excitation – Acceptance Criteria

The following conditions shall be used to determine the susceptibility of

a bridge to vortex excited vibrations.
(a) Any bridge whose fundamental frequency is greater than 5 Hz shall
be considered stable with respect to vortex excitation.
(b) Any bridge, including truss bridges (see also (c)), shall be
considered stable with respect to vortex excited vibrations if the
lowest critical wind speeds, Vcr, for vortex excitation in both
bending and torsion, exceed the value of reference wind speed Vvs

Vvs = 1.25 Vr
Vr = the hourly mean wind speed for relieving areas of the bridge
deck derived in accordance with BD37.
 Vortex Excitation – Acceptance Criteria

(c) In addition, truss girder bridges shall be considered stable with

regard to vortex excited vibrations provided φ < 0.5, where φ is the
solidity ratio of the front face of the windward truss, defined as the
ratio of the net total projected area of the truss components to the
projected area encompassed by the outer boundaries of the truss
(i.e. excluding the depth of the deck). For truss with φ ≥ 0.5, refer to
the above criteria (b).
If any one of (a), (b) or (c) is satisfied, then the bridge shall be deemed
stable with respect to the effects of vortex excitation. If none of these
conditions is satisfied, then the effects of vortex excitation shall be
checked with the additional requirements of BD 49/01, or to be tested in
a wind tunnel test.
 Wind Tunnel Test for
Wind-Induced Vibration
 Wind Tunnel Test
The experimental method for
simulating aeroelastic or
aerodynamic excitation is to
perform a reduced scale model test
that we called “wind tunnel test”.
Wind tunnel test simulates the wind
attacks on the structure, on which
its response can be determined
using different instruments such as
accelerometers and strain gages. In bridge aerodynamics, there are
three types of tests can be performed, i.e. aerostatic section model test,
aerodynamic section model test, and full-scale model test.
 Aerostatic Section Model Test

a) Aerostatic Section Model Test

The Section Model Test is usually
used to study the effect of the wind on
the bridge deck section in a limited
length (not full bridge model),
particularly to define the aerodynamic
coefficients for drag force, lift force,
and pitching moment in various angles
of attack (usually in the range of -15o
to +15o).
Static Equivalent of Aerodynamic Forces & Coefficients

Fh = Ch q H [kN/m]
Fv = Cv q H [kN/m]
M = Cm q B H [kN.m/m]

q = wind pressure (kN/m2)

H = height of the deck incl railing fence (at service state) (meter)
B = deck width (meter)
Ch , Cv , Cm = aerodynamic coefficients depending on the deck shape
and the wind attack angle, obtained by Wind Tunnel Test
 Aerodynamic Section Model Test

b) Aerodynamic Section Model Test

A rigid model for deck cross section
with a scaled geometry in limited length
is exposed to a uniform air flow in the
wind tunnel. The aim of this test is to
study the magnitude of critical wind
speed that may induce a divergent
amplitude response such as flutter
on the bridge deck under wind attack of various angles.
 Aerodynamic Section Model Test

The 2D spring-suspended
rigid section model was
usually employed in the test.
The section model is
suspended with 8 springs
from the two frames installed
outside the wind tunnel.
 Basic Similarity in the Concept of Section Model Test
The geometric scale (λL) was determined to be a certain ratio according
to the dimensions of the prototype deck cross-section and the size of
wind tunnel.
Besides of the geometric similarity, the following three groups of
dimensionless parameters should be kept in consistence between the
prototype and the model in the wind tunnel of rigid spring-suspended
section model test:

Frequency parameters: , or (frequency ratio)

Inertia parameters: , or (ratio of gyration radius)

Damping parameters: ξv , ξt (damping ratio）

where:
V = Mean wind speed (m/sec)
𝑓 = Natural frequency of vertical vibration (1/sec)
𝑓 = Natural frequency of torsional vibration (1/sec)
B = Deck width (meter)
b = Deck half-width (meter)
𝑚 = Equivalent mass of the bridge deck per unit length (kg/m)
𝐽 = Equivalent mass moment of inertia per unit length (kg/m)
𝑟 = Equivalent gyration radius (meter)
v = Damping ratio of vertical vibration
ξt = Damping ratio of torsional vibration
The model damping ratios are set to be 2% for an initial analysis due to
prestressed concrete structure used in Balang Bridge, as suggested by the
“China Code of Wind Resistant Design Specification for Highway Bridges”.
Basic Similarity in the Concept of Section Model Test

The designed and measured parameters of the sectional model and the
corresponding parameters of the prototype, obtained in a principle of the
similarity requirements mentioned above, are presented in the next table for
the service state, on which the fundamental natural frequencies of vertical
and torsional vibrations were selected for the simulation of elastic
parameters.
The mass and mass moment of inertia of the sectional model were designed
according to the equivalent mass and mass moment of inertia of the
prototype bridge deck to consider the spatial behavior of the vibration of the
prototype bridge and the effects of the vibrations of the pylons and cables.
 Basic Similarity in the Concept of Section Model Test

The equivalent mass and equivalent mass moment of inertia of the prototype
bridge can be determined as follows:

~
L ϕ d2 ( x)dx
d
m =M
eq
g
~
 ϕθ2x ( x)dx
x
J meq =M
Lg
where:
M = the generalized mass of the corresponding mode
d = x, y, z
ϕx(x), ϕy(x), ϕz(x) and ϕθx(x) = the mode function values of the deck in the
longitudinal, vertical, lateral and torsional
directions at the coordinate of x, respectively
Lg = the total length of the bridge deck
Similarity to be performed for the Sectional Model
Scale ratio Model
Name Sym Unit Prototype
Flutter Vortex Flutter Vortex
Deck length L m 69.6 1/40 1/40 1.74 1.74

Deck width B m 23.4 1/40 1/40 0.585 0.585

Deck height H m 2.4 1/40 1/40 0.06 0.06

Equivalent
Meq kg/m 5.42x104 1/402 1/402 33.88 33.88
Service state

mass

Equivalent
mass kg×m2/
J 3.51x106 1/404 1/404 1.371 1.371
moment of meq m
inertia
Similarity to be performed for the Sectional Model

Scale ratio Model

Name Sym Unit Prototype
Flutter Vortex Flutter Vortex
Equivalent
gyration re m 8.12 1/40 1/40 0.203 0.203
radius

Frequency f
Service state

2.247 1/1 1/1 +2.25 +2.25

ratio f

Vertical v 1～2% 1/1 1/1 +1.5% +1.5%

damping

Torsional t 1～2% 1/1 1/1 +1.5% +1.5%

damping
 Full Model Test

c) Full Model Test

In this model, the whole bridge
and major surrounding topo-
graphy are scaled down. The
bridge model is scaled to a
certain ratio (usually in the range
of 1/200 to 1/1000 depending on
the bridge size and the wind
tunnel size). This full model is built to fulfill the aerodynamic stability
requirements, and to prevent flutter instability occurrence. Also, it
helps to evaluate the effect of site conditions on the wind load.
 Full Model Test

However, due to cost and time limitation, the full model wind tunnel
test is not necessary to apply to all types of long-span bridges. The
full model is usually needed for the bridges with high risk of aero-
dynamic instability such as a long-span suspension bridge using steel
deck (with lower damping ratio).
For a long-span bridge using concrete deck (with higher damping
ratio), and in particular for the cable stayed concrete bridge, a section
model wind tunnel test would be generally sufficient to well simulate
its response under aerodynamic excitation.
Wind Tunnel Test for
Balang Bridge
 Wind Tunnel Test
The section model wind tunnel
test for Balang Bridge was
carried out in the TJ-1 Boundary
Layer Wind Tunnel of the State
Key Laboratory for Disaster
Reduction in Civil Engineering,
Tongji University, Shanghai,
China.
TJ-1 Boundary Layer Wind Tunnel is an open-circuit low-speed tunnel
with a test-section of being 1.8 m wide, 1.8 m high and 14 m long. The
fan power is 90kW, and the wind speed ranges from 0.5 to 30 m/sec.
 Section Model test

The geometric size scale (λB) was determined to be 1/40 according to

the dimensions of the prototype deck cross-section of the bridge.
To diminish the influence of the 3D flow around the two ends of the
sectional model on the test results, the length of the sectional model
was designed to be 1.74m and the gaps between the model ends and
the side walls of wind tunnel were about 3cm. Therefore the ratio of
length over width of the section model was about 2.974.
The rigid sectional model was made of a metal framework and wood
plate covers, and the deck shape was elaborately simulated.
The barriers and the handrails were made of plastic plates, by means
of computer controlled sculpturing.
Three different states of the bridge were tested in the wind tunnel for flutter check:

Completed Bridge at Service State

Longest Double
Cantilever State

Longest Single Cantilever State

 Ultimate Design Wind Speed

Wind speed Bridge Deck

Basic wind speed Vs10 (m/sec)

35.00
[RSNI T-02-2005: Ultimate State]

Site surface coefficient α 0.16

Design height z (meter) 32.96

Design wind speed at the height z for

42.36
service state Vd = Vs10 (z/10)α (m/sec)
Design wind speed at construction stage
= η Vd (m/sec), η is a reduction factor 35.58
(return period of 10 years, η = 0.84)
 Ultimate Design Wind Speed for Flutter Check
Ultimate design wind speed at the deck level Vd = Vs10 (z/10)α = 42.36 m/sec
For flutter risk identification, the ultimate design wind speed for flutter check can
be defined as:
Vuf = K μf Vd
K is a comprehensive safety factor considering the uncertainties in the wind tunnel
test and the design & construction of the bridge, which was set to 1.2.
μf is a coefficient considering the influence of turbulence on wind speed and the
incomplete correlation of winds along the bridge span (depending also on the span
length and the bridge site category), which was set to 1.24.
Then: Vuf = K μf Vd = 1.2 x 1.24 x Vd = 1.488 Vd = 63.03 (m/sec)
For flutter stability requirement, the critical flutter wind speed obtained later by
the wind tunnel test should exceed the above ultimate design wind speed.
Aerostatic Section Model Test
Service State
Service
State
Results of Aerodynamic Coefficients with Various
Wind Attack Angles in Service State
2

Cd
1.5

Ch
Aerodynamic coefficient

1 Ch
Cv
Cm(*10)
Cd
0.5 Cl

-0.5
-15 -10 -5 0 5 10 15
Wind Direction(deg)
where:
Ch = aerodynamic coefficient of lateral force
Cv = aerodynamic coefficient of vertical force
Cm = aerodynamic coefficient of torsion
Cd = aerodynamic coefficient of drag
Cl = aerodynamic coefficient of lift-up
FL FV
FD

α
FH
M

Wind
Aerodynamic Section Model Test for Flutter Stability
Aerodynamic Section
Model Test for Service
State
Service
State
Service
State
Service
State
Service
State
Construction
Stage
[Longest
Cantilever
State]
Construction
Stage
[Longest
Cantilever
State]
Construction
Stage
[Longest
Cantilever
State]
Wind Tunnel Test Control Room
Vibration at
Positive
Damping
Condition
Wind Speed
Monitor
Zero
Damping
Vibration
Starting of
Negative
Damping
Ratio
Damping Ratio

Positive Damping

Damping Ratio
Negative Damping decreases as a
function of Wind
Speed
Wind Speed
Wind Speed
Monitor
Vibration at
Negative
Damping
Condition
[Prior to
Instability]
Vibration at
Negative
Damping
Condition
[Prior to
Instability]
 Flutter Critical Wind Speed by Section Model Test

Flutter wind
Longest single-
speed (m/s) Service state
cantilever state
Attack angle
+3° >104.3 95.9
0° >104.6 >105.1
-3° 93.8 >105.5
Ultimate Design Wind
63.03 52.94
Speed for Flutter Check
 Summary of Flutter Stability Test Result
The result of flutter test demonstrates that the critical flutter wind speed Vcrf in service
state is 93.8 m/sec at a wind attack angle (to the bridge deck) of -3°.

Vuf = 1.488 Vd = 63.03 m/sec Vuf < Vcrf (adequate)

As the critical flutter wind speed obtained by the wind tunnel test (93.8 m/sec) is
much higher than the ultimate wind speed at the bridge deck level (63.03 m/sec), that
means the bridge is safe and adequate for the flutter risk.
In the construction stage, the test result shows a critical flutter wind speed of 95.9
m/sec in the longest cantilever state with wind attack angles of +3°.
The above result of wind tunnel test is consistent with the result of theoretical
aerodynamic susceptibility analysis as mentioned previously.
Therefore, Balang Island Bridge is proven to be adequate for the aerodynamic
excitation, in both service state and construction stages.
Aerodynamic Test for Vortex-Induced Resonance
Service
State
 Vortex-Excited Resonance Test

Similar to those for the flutter sectional model test, the mass and mass
moment of inertia of the sectional model for the vortex-excited resonance test
should also be simulated according to the equivalent mass and equivalent
mass moment of inertia of the prototype deck, to consider the effect of the
spatial vibration behavior of the prototype bridge in the 2D sectional model
test.
In this wind tunnel, the vortex-excited resonance tests for the service state
was carried out in smooth flows with various attack angles of +3°, 0° and -3°.
The tests were conducted for the case of structural damping ratios between
1.0% and 2.0% for both vertical vibration and torsional vibration.
Vortex Induced Vibration v/s Wind Speed in Prototype Bridge

Vertical Bending（ cm.Single Peak）

0
α=+3
4 α=0
0

0
α=-3
3

11 12 13 14 15 16 17
Prototype Bridge Wind Speed(m/s)
 Summary of Vortex-Excited Resonance Test Result

Maximum Amplitude
Attack Lock-in region
Vortex-excited Damping ratio Wind
Angle of wind speed Amplitude
resonance （％） speed
（o） (m/sec) （m）
(m/sec)
+3 12.4 ～ 16.6 Vertical 1.5～2.0 14.1 0.0270
0 11.4 ～ 12.4 Vertical 1.5～2.0 11.6 0.0409
-3 11.9 ～ 14.5 Vertical 1.5～2.0 8.5 0.0461
Note: All parameters in the table are for the prototype bridge on Vertical Vortex-Excited
Resonance in the Service Sate (Uniform flow)

According to “China Code of Wind Resistant Design Specification for Highway Bridges”,
the allowable amplitudes of the vertical vortex-excited resonance ha = 0.04/fv = 0.04/0.295
= 0.135 m (adequate).
Concluding Remarks
1. In the design of long-span cable stayed bridge, the aerodynamic instability
risks must be carefully analyzed, using the comprehensive criteria and
rules, with objective to be always in the safe range of adequacy on the
aerodynamic susceptibility parameters.
2. Although we achieves a good result in the above Point 1, the wind tunnel test is
always suggested for long-span bridges, because it is a good way to simulate
more accurately the aerodynamic excitation on a long-span bridge, and to
determine “close to reality” of the bridge’s responses under wind attack.
3. However, due to cost and time limitation, the full model wind tunnel test will not
be applied to all types of long-span bridges. The full model is usually needed
for the bridges with high risk of aerodynamic susceptibility such as a long-span
suspension bridge using steel deck (with lower damping ratio).
4. For a long-span bridge using concrete deck (with higher damping ratio), and in
particular for the cable stayed concrete bridge, a section model wind tunnel test
would be generally sufficient to well simulate its responses under aerodynamic
excitation.
Thank you