applied

sciences
Article
Preliminary Flutter Stability Assessment of the Double-Deck
George Washington Bridge
Sebastiano Russo 1 , Gianfranco Piana 1, * , Luca Patruno 2 and Alberto Carpinteri 1

1 Department of Structural, Geotechnical and Building Engineering, Politecnico di Torino, Corso Duca Degli
Abruzzi 24, 10129 Torino, Italy; sebastiano.russo@polito.it (S.R.); alberto.carpinteri@polito.it (A.C.)
2 Department of Civil, Chemical, Environmental and Material Engineering, “Alma Mater Studiorum”
University of Bologna, Viale del Risorgimento 2, 40136 Bologna, Italy; luca.patruno@unibo.it
* Correspondence: gianfranco.piana@polito.it

Featured Application: This study provides a simplified approach to the flutter analysis of suspen-
sion bridges having two superposed decks. The George Washington Bridge engineering case is
analyzed, in consideration of its historical relevance and age. The obtained results are compared
with the predictions of simplified formulations available from other sources.

Abstract: We deal with the flutter analysis of the George Washington bridge, in both the single- and
double-deck configurations of 1931 and 1962, respectively. The influence of the additional lower
deck on the aerodynamic behavior is investigated. To overcome the lack of aerodynamic data, a
simplified approach is followed based on Fung’s formulation, in which the flutter derivatives are
expressed in terms of the real and imaginary parts of the Theodorsen function and of the steady-state
aerodynamic coefficients of the deck cross-section. The latter are obtained by Computational Fluid
Dynamics simulations conducted in ANSYS FLUENT, whereas the ANSYS Mechanical APDL finite
element package is used to perform the flutter analyses. Two different methods for the application of
the aeroelastic forces are employed for the double-deck configuration: (i) self-excited forces, based on
flutter derivatives related to the whole cross-section, acting on the upper deck; and (ii) self-excited
forces, based on flutter derivatives related to the single deck, simultaneously applied to the upper
and lower decks. The obtained results are critically compared with theoretical predictions of simple
formulas available from the literature; it is suggested that laboratory tests are needed since no
Citation: Russo, S.; Piana, G.;
Patruno, L.; Carpinteri, A.
experimental results seem to be available.
Preliminary Flutter Stability
Assessment of the Double-Deck Keywords: aeroelastic flutter; suspension bridge; double-deck; flutter derivatives; aerodynamic
George Washington Bridge. Appl. Sci. coefficients; computational fluid dynamics; finite element analysis; Fung formulation; stability;
2023, 13, 6389. https://doi.org/ safety assessment
10.3390/app13116389

Academic Editor: Cem Selcuk

Received: 4 April 2023 1. Introduction

Revised: 8 May 2023
Old bridges (conventionally, which are more than 50 years old), require special main-
Accepted: 17 May 2023
tenance, accurate inspections, and a careful assessment of their safety conditions.
Published: 23 May 2023
The conflict between economy and structural performance between the 19th and the
20th centuries led the design of long-span bridges to the development of very flexible
and slender structures. The use of the elastic deflection theory allowed for very slender
Copyright: © 2023 by the authors.
decks against static loads and shifted the design trend at that time from rigid trusses to
Licensee MDPI, Basel, Switzerland. slender edge girders. This evolution ended brutally with the Tacoma Narrows Bridge
This article is an open access article disaster due to the wind-induced flutter instability on 7 November 1940. From then on, any
distributed under the terms and design of a flexible structure must assure the structure itself to be stable under the dynamic
conditions of the Creative Commons effects of wind loads. In fact, in addition to the already known static divergence due to the
Attribution (CC BY) license (https:// steady-state wind loads, flutter stability has become a governing criterion in the design of
creativecommons.org/licenses/by/ long-span suspension bridges. The objective of a (linear) flutter analysis is to predict the
4.0/). lowest critical flutter wind velocity and the corresponding flutter frequency.

Appl. Sci. 2023, 13, 6389. https://doi.org/10.3390/app13116389 https://www.mdpi.com/journal/applsci

Appl. Sci. 2023, 13, 6389 2 of 21

A crucial point in flutter analysis is the definition of motion-dependent aerodynamic

loads. The first analytical solutions were developed by Wagner [1] and Theodorsen [2] for
thin airfoil. Theodorsen defined the self-excited forces as the superposition of circulatory
and non-circulatory contributions, the former depending on the oscillation frequency and
accounting for flow unsteady effects, and the latter independent of oscillation frequency
and including inertial effects due to the moving fluid mass. Subsequently, Scanlan and
Tomko exported some features of the Theodorsen’s results extending the formulation to
bluff bodies, as bridge cross-sections. In this formulation, the wind loads induced by sec-
tional harmonic motions are described by means of a linear format based on experimentally
evaluated parameters, called “flutter derivatives”, that supply the lack of closed-form
analytical formulations [3,4]. The relation between Wagner’s approach based on indi-
cial functions, Theodorsen’s theory based on circulatory function and Scanlan’s flutter
derivatives is well detailed in [5].
In the last decades, Scanlan’s formulation has been the most widely adopted, and
the calculation of flutter derivatives has become an important step for any flutter analysis.
Currently, the only method considered reliable for their calculation at the design stage is
that of carrying out experimental tests on scale models in wind tunnels. Some notable
examples are [6,7] for the Great Belt East Bridge and [8] for the Akashi Kaikyo Bridge; more
recently [9] for the Jianghai Channel Bridge and [10] for the Hardanger Bridge.
To overcome costs and difficulties arising from wind tunnel testing, several efforts
were made with the aim of developing some simplified methods for the calculation of
flutter derivatives. Although these methods are not appropriate for the final design stage,
they allow useful and versatile studies in early design phases. Some authors utilized the
derivatives of the thin airfoil, as [11] for the Humber bridge. This simplification leads to
relatively small errors when the bridges are characterized by streamlined cross-sections.
Al-Assaf [12] adopted an alternative approach for open-truss stiffened suspension bridges:
he estimated the aerodynamic derivatives based on the correlation between the thin airfoil
derivatives and those of other bridges having a similar deck configuration. The method
was applied to evaluate the flutter stability of the second Tacoma Narrows Bridge, with
special focus on the effects of the side grates. Another simplified approach is given by
the quasi-steady theory, in which the aerodynamic loads depend on the instantaneous
relative velocity between the flow and the cross-section. In this framework the flow-induced
forces are described by means of non-linear static relationships involving the wind angle
of attack and the displacements of the structure. A linearization of this model allows a
comparison with Scanlan’s semi-empirical approach and furnishes an expression of the
flutter derivatives as functions of the steady-state aerodynamic coefficients of the cross-
section [13]. These simplified expressions of the flutter derivatives are widespread for the
streamlined deck cross-sections. Some works on the validity of this approach are [14–16]
and some applications for the determination of the critical state can be found in [17–19]. In a
recent work [20], a linear superposition of flat plate aerodynamics was adopted to estimate
the flutter derivatives of streamlined multi-box deck sections. In that case, correction
factors were introduced into the proposed analytical formulas to better fit the available
experimental data. In the present paper, a simplified approach is used based on Fung’s
formulation [21], in which the flutter derivatives are expressed in terms of the steady-state
aerodynamic coefficients and of the real and imaginary parts of the Theodorsen function.
Two-dimensional Computational Fluid Dynamics (CFD) simulations were conducted
in ANSYS FLUENT for the determination of the lift and moment coefficients, varying the
wind attack angle. In recent years, the CFD approach gained importance with respect to
the traditional experimental investigation. Some applications related to the bridge cross-
sections are [22,23]. More recently, Brusiani et al. [24] pointed out that Reynolds-averaged
Navier-Stokes (RANS) approach coupled with the k−ω SST turbulence model, identify the
best compromise between accuracy and computational cost. The CFD framework also pro-
vides different approximate techniques for the direct calculation of flutter derivatives, some
applications are: [25–28]. Nevertheless, the calculation of the steady-state aerodynamic
 Appl. Sci. 2023, 13, 6389 3 of

Appl. Sci. 2023, 13, 6389 3 of 21

model, identify the best compromise between accuracy and computational cost. The C
framework also provides different approximate techniques for the direct calculation
flutter derivatives, some applications are: [25–28]. Nevertheless, the calculation of
steady-state
coefficients requires aerodynamiccompared
simpler simulations coefficients torequires simplerto
those required simulations
compute the compared
flutter to those
quired in
derivatives, especially to terms
compute the flutter derivatives,
of computational effort. especially
In addition, in terms of computational effort.
the approximations
introduced by a 2D addition, the approximations
geometrical representation, introduced
neglecting by several
a 2D geometrical
appendages representation,
that could neglect
severalbehavior,
affect the aerodynamic appendages do that could affect
not justify the usetheof
aerodynamic
more refined behavior,
methods. do not justify the use
Concerningmore refined
flutter methods.
analyses, several methods can be found in literature [29,30], a
method developed byConcerning
Hua and Chen flutter analyses,
[31], severalthe
which allows methods
analysiscantobebe found
carriedinoutliterature
with [29,30]
method developed by Hua and Chen [31], which allows
the FE package ANSYS Mechanical APDL, is adopted in this paper. The software, in fact, the analysis to be carried out w
the FE package ANSYS Mechanical APDL, is adopted
provides specific user-defined elements, through which it is possible to implement the in this paper. The software, in fa
provides specific user-defined elements, through
motion-dependent aeroelastic forces as expressed in Scanlan’s formulation. which it is possible to implement
motion-dependent aeroelastic forces as expressed in Scanlan’s formulation.
The George Washington suspension bridge was chosen as a case study, partly because
The George Washington suspension bridge was chosen as a case study, partly
of its historical significance. In fact, it was the first bridge whose main span exceeded one
cause of its historical significance. In fact, it was the first bridge whose main span exceed
kilometer. The bridge was opened to traffic in 1931. It is characterized by a total length
one kilometer. The bridge was opened to traffic in 1931. It is characterized by a total leng
of 1450 m and a mid-span of 1067 m. Initially it was composed only of the upper level,
of 1450 m and a mid-span of 1067 m. Initially it was composed only of the upper lev
whereas the lower deck was constructed from 1958 to 1962 because of the increasing traffic
whereas the lower deck was constructed from 1958 to 1962 because of the increasing tra
flow. Both configurations
flow. Bothwill be analyzedwill
configurations in the following:inthe
be analyzed theoriginal
following: onethe with the single
original one with the s
deck and the stiffened one with the two decks (Figure 1). The choice is also
gle deck and the stiffened one with the two decks (Figure 1). The choice is also motivated by motiva
the peculiarity ofby thethe
current bridge having two superposed decks. A similar configuration
peculiarity of the current bridge having two superposed decks. A similar config
can be detected ration
in thecanVerrazano-Narrows Bridge (USA), alsoBridge
be detected in the Verrazano-Narrows having a great
(USA), alsohistorical
having a great hist
relevance, and inical therelevance,
Yangsigang Yangtze River bridge (China), which
and in the Yangsigang Yangtze River bridge (China), which currently has the currently h
third longest span in the world after the Akashi Kaikyo bridge (Japan) and
the third longest span in the world after the Akashi Kaikyo bridge (Japan) the recent 1915
and the rec
Çanakkale bridge1915 (Turkey), the current
Çanakkale World record.
bridge (Turkey), the current World record.

(a) (b)
Figure 1. GeorgeFigure
Washington bridge
1. George configurations
Washington (dimensions (dimensions
bridge configurations in m). (a) Single-deck, 1931;
in m). (a) Single-deck, 1931;
Double-deck, 1962.
(b) Double-deck, 1962.

As a matter of fact, several investigations on the aerodynamic performances of trussed

girders have been made, but only a few for double-deck bridges [32]. Due to the lack of
aerodynamic data and studies on the subject, two simplified ways for the application of the
aeroelastic forces are followed in the present study with the aim of obtaining approximate
predictions. The first way considers the self-excited forces, based on the flutter derivatives
related to the whole cross-section, acting on the upper deck. In the second case, the self-
 As a matter of fact, several investigations on the aerodynamic performances of
trussed girders have been made, but only a few for double-deck bridges [32]. Due to the
Appl. Sci. 2023, 13, 6389 lack of aerodynamic data and studies on the subject, two simplified ways for the applica- 4 of 21
tion of the aeroelastic forces are followed in the present study with the aim of obtaining
approximate predictions. The first way considers the self-excited forces, based on the flut-
ter derivatives
excited related
forces, based ontothe
theflutter
wholederivatives
cross-section, acting
related onsingle
to the the upper
deck,deck. In the second
are simultaneously
case, the self-excited forces, based on the flutter derivatives related to the
applied to the upper and lower decks. Both approaches produce plausible results, single deck, are
which
simultaneously applied to the upper and lower decks. Both approaches produce
can serve as a reference for future comparisons with alternative investigation methods. In plausible
results, which
principle, can serveapproach
the proposed as a reference
can befor
usedfuture comparisons
for analyzing with
other alternative
double-deck investiga-
bridges, and
tion methods. In principle, the proposed approach
deserves further attention for proper validation. can be used for analyzing other double-
deck bridges, and deserves further attention for proper validation.
2. Motion Related Wind Load
2. Motion Relateddescribing
The equation Wind Load the motion of the bridge in the smooth flow can be expressed as:
The equation describing the motion
..
of the
.
bridge in the smooth flow can be expressed as:
M X + C X + KX = Fae . (1)
𝑀𝑋 + 𝐶𝑋 + 𝐾𝑋 = 𝐹 . (1)
where, in
where, in aa Finite
Finite Element
Elementframework,
framework, 𝑀, M, 𝐾,K, and
and 𝐶 C represent the global
represent the global mass,
mass, stiffness,
stiffness,
.. .
and damping matrices, respectively; 𝑋, 𝑋, and 𝑋 are the nodal acceleration, velocity,
and damping matrices, respectively; X, X, and X are the nodal acceleration, velocity,
and
and displacement
displacement vectors,
vectors, respectively;
respectively; andand 𝐹 Fdenotes
ae denotes thevector
the vectorofof nodal
nodal self-excited
forces. The
forces. The three
three degrees
degrees of
of freedom
freedom of the bridge deck cross-section, namely the vertical vertical
displacement ℎh,, the torsional
displacement rotation 𝛼
torsional rotation displacement 𝑝p,, to which
α,, and the horizontal displacement which
correspond the
correspond the aeroelastic forces of lift L𝐿SE , moment M
aeroelastic forces 𝑀SE,, and drag D
and drag 𝐷 SE,, are
are reported
reported in
Figure 2,
Figure where U
2, where 𝑈 indicates
indicatesthe
theundisturbed
undisturbedmean meanwindwindvelocity.
velocity.

Figure 2. Three-degrees-of-freedom section model.

Figure 2. Three-degrees-of-freedom section model.
According to Scanlan’s formulation [33], the self-excited aeroelastic forces per unit deck
According
length to Scanlan’s
can be expressed formulation
as a linear function[33], thedisplacement
of the self-excited aeroelastic
and velocityforces per unit
parameters of
deck length can be
the deck as follows: expressed as a linear function of the displacement and velocity param-
eters of the deck as "follows:
. . . #
1 2 h Bα h p p
1 L = ρU ℎ B KH𝐵 𝛼 + KH + K2 Hℎ α + K2 H 𝑝 + KH𝑝 + K2 H6 , (2a)
𝐿 = 𝜌𝑈 SE 𝐵 𝐾𝐻2 + 𝐾𝐻 1 U + 𝐾 𝐻2 𝛼U+ 𝐾 𝐻 3+ 𝐾𝐻 4+B𝐾 𝐻 5 U, B (2a)
2 𝑈 𝑈 𝐵 𝑈 𝐵
" . . . #
1 𝑝1 2 𝐵𝛼 p B α 𝑝2 ℎ 2 p ℎ h 2 h
𝐷 = 𝜌𝑈 𝐵DSE 𝐾𝑃= 2+ 𝐾𝑃B KP1+U𝐾+𝑃KP
ρU 𝛼 2+ U 𝐾 𝑃+ K +P3𝐾𝑃
α + K +P4𝐾 𝑃+ KP5, + K P6
B 𝐵 U B
, (2b)
(2b)
2 𝑈 𝑈 𝐵 𝑈
" . . . #
1 1 ℎ2 2 𝐵 𝛼h Bα ℎ 𝑝 h 𝑝 p p
𝑀 = 𝜌𝑈M𝐵 SE =𝐾𝐴 ρU +B𝐾𝐴KA1 ++𝐾KA 𝐴 2𝛼 + 𝐾+ 𝐴 K2 A3+α 𝐾𝐴
+ K2 A+ 4 𝐾+𝐴KA5 , + K A6
2
, (2c)
(2c)
2 2 𝑈 𝑈U U 𝐵 𝑈 B 𝐵 U B

where ρ is the air density, K = ωB

U is the reduced circular frequency, and Hi (K ), Pi (K ), Ai (K )
.
are the flutter derivatives. In most practical applications, p, p, and Pi (DSE ) can be ignored.
As mentioned in the Introduction, the most reliable way to calculate flutter derivatives
is based on experimental tests made in the wind tunnel. For the bridge chosen as a
case study, no data regarding flutter derivatives were found in the literature, therefore
a simplified approach was followed for their calculation. It was decided to utilize a
Appl. Sci. 2023, 13, 6389 5 of 21

formulation developed by Fung [21] for the thin airfoil because it allows one to input the
slopes of the steady-state aerodynamic coefficients, and, unlike the quasi-steady approach,
it provides a unique formulation for the determination of the derivatives H2 , A2 . The
expressions of the flutter derivatives are the following:

KH1# (K ) = −Cl0 Fe(K ), (3a)

" #
C0

1 G
e( K )
KH2# (K )= Cl0− a F (K ) +
e + l , (3b)
4 K 4
  
e( K ) 1 − a a
K2 H3# (K ) = Cl0 Fe(K ) − K G + Cl0 K2 , (3c)
4 4

Cl0 K2
K2 H4# (K ) = Cl0 K G
e( K ) + , (3d)
4
0 e
KA#1 (K ) = Cm F ( K ), (3e)
" #
Cl0
  
0 1 G
e( K ) 1
KA#2 (K ) = −Cm − a F (K ) +
e + a− , (3f)
4 K 4 4

a C0 K2
  
0 1
KA#3 (K ) = −Cm F (K ) − K G (K )
e e − a + Cl0 K2 + l , (3g)
4 4 128

0
K2 A#4 (K ) = Cm e( K ) − C 0 K 2 a ,
KG (3h)
l
4
where A#i (K ) and Hi# (K ) are the flutter derivatives for the thin airfoil as expressed by Fung
(K = ωB/U ), Fe(K ) and G e(K ) are respectively the real and the imaginary parts of the
0
Theodorsen function, Cl and Cm 0 are the derivatives of the lift and moment coefficients with

respect to the wind angle of attack and a is the distance between the shear center and the
centroid of the airfoil, normalized with respect to the chord B. For low values of reduced
frequency, Fe(K ) tends to one, G e(K ) tends to zero and the Fung formulation tends to the
quasi-steady one.
These formulas were tested for two bridges of which both steady-state aerodynamic
coefficients and flutter derivatives are available: the Great Belt East in Denmark, with
a streamlined deck [6,34], and the Akashi Kaikyo in Japan, with a truss girder [35,36].
The derivatives of the aerodynamic coefficients, expressed in rad−1 , are: Cl0 = −4.37
 

and Cm 0 = 1.17 for the Great Belt Bridge; and C 0 = −1.19 and C 0 = 0.3 for the Akashi
l m
Kaikyo Bridge. The curves in Figures 3 and 4 show a comparison between the flutter
derivatives: obtained experimentally (solid crossed line), calculated by the Fung formulas
(solid line), calculated by the quasi-steady formulation (dashed line) [26], and predicted
by the Theodorsen theory for the thin airfoil (dash dotted line). From the comparisons
in Figures 3 and 4, it is possible to notice that the Fung formulation leads generally to a
good alignment with the experimental curves: the average trend is always respected except
for H4 , that in some cases can be neglected [12]. A remarkable superposition is obtained
for A1 and A3 of both bridges, and also for H3 in the case of the Great Belt Bridge. Good
predictions are also obtained for A2 derivative.
As expected, for a truss-stiffened girder, Theodorsen formulation for thin airfoil
becomes unable to reliably predict the aerodynamic behavior. Even though the quasi-steady
formulation also shows a good agreement, the Fung formulation was chosen over the quasi-
steady approach because this latter is better suited for low values of reduced frequency and
high values of reduced wind velocities [16]. In fact, the bridge chosen as a case study is
relatively stiff and characterized by high values of torsional and vertical eigenfrequencies,
therefore flutter instability is expected to occur at a relatively low reduced velocity.
 6 of 21
Appl. Sci. 2023, 13, 6389 6 of 21

Figure 3. Flutter derivatives of the Great Belt Bridge deck.

Figure 3. Flutter derivatives of the Great Belt Bridge deck.

 7 of 21

Appl. Sci. 2023, 13, 6389 7 of 21

Figure 4. Flutter derivatives of the Akashi Kaikyo Bridge deck.

Figure 4. Flutter derivatives of the Akashi Kaikyo Bridge deck.
3. Full-Order Flutter Analysis Using ANSYS Mechanical APDL
The main lack of almost all the FE packages commonly used in civil engineering is
As expected, for theaimpossibility
truss-stiffened girder,
of including Theodorsenaeroelastic
motion-dependent formulation forthe
loads into thin
modelairfoil be-
and then
comes unable to reliably predictathe
of performing aerodynamic
complex behavior.
eigenvalue analysis. HuaEven though
and Chen [31] the quasi-steady
proposed a method
that allows one to perform a flutter analysis using the commercial finite element package
formulation also shows a good agreement, the Fung formulation was chosen over the
ANSYS Mechanical APDL. The method is based on the definition of the aeroelastic loads
quasi-steady approach because
by means this latter
of a particular is better
user-defined suited
element for low
as shortly valuesbelow.
illustrated of reduced
Accordingfre-
to
quency and high values of reduced
the method developedwind velocities
by Hua [16].
and Chen Inthe
[31], fact, the bridge chosen
motion-dependent as atocase
effect due the
wind load is considered for each deck element by the element aeroelastic
study is relatively stiff and characterized by high values of torsional and vertical eigenfre- stiffness and
damping matrices modeled by the user defined MATRIX27 elements. Following Scanlan’s
quencies, therefore flutter instability is expected to occur at a relatively low reduced ve-
locity.

3. Full-Order Flutter Analysis Using ANSYS Mechanical APDL

 tion, the self-excited forces are expressed as functions of the flutter derivatives, as recall
in Section 2.
It is possible to manipulate Equation (2a)–(2c) in order to obtain the equivalent nod
forces acting on the ends of the generic deck element, hence a lumped formulation can
Appl. Sci. 2023, 13, 6389 8 of 21
used to derive the element aeroelastic stiffness and damping matrices [30]. Assembli
the element matrices into global aeroelastic stiffness (𝐾 ) and damping (𝐶 ) matrices, t
mathematical model of the system integrated with the effect of aeroelasticity is obtain
formulation, the self-excited forces are expressed as functions of the flutter derivatives, as
[31]:
recalled in Section 2.
It is possible to manipulate Equations (2a)–(2c) in order to obtain the equivalent nodal
𝑀𝑋 𝐶 𝐶 𝑋 𝐾 𝐾 𝑋 0
forces acting on the ends of the generic deck element, hence a lumped formulation can be
used to derive
The globaltheaeroelastic
element aeroelastic
stiffnessstiffness and damping
and damping matrices
matrices [30]. Assembling
contain flutter derivativ
the element matrices into global aeroelastic stiffness (Kae ) and damping (Cae ) matrices, the
which depend on the reduced wind velocity 𝑈 (𝑓 , oscillation frequenc
mathematical model of the system integrated with the effect of aeroelasticity is obtained [31]:
Therefore, the system expressed .. by Equation
. (4) is parameterized by wind velocity an
vibration frequency. By Equation M X + (C − Caea) X
(4), + (K − Keigenvalue
complex ae ) X = 0 analysis can be carried (4) out
determine the critical
The global windstiffness
aeroelastic velocityand and vibration
damping frequency.
matrices containBeingflutterthe conjugate pairs
derivatives,
complex eigenvalues
which depend 𝜆
on the reduced 𝜎 wind 𝑖𝜔 ,velocity
and theUconjugate U
R = f B ( f pairs
= 2π , of
ω
eigenvectors 𝜙
complexfrequency).
oscillation
𝑝 𝑖𝑞 , the system will be dynamically unstable if the real part of any eigenvalue
Therefore, the system expressed by Equation (4) is parameterized by wind velocity and is po
itive. Hence, the condition for the onset of flutter instability is stated as follows:toat a certa
vibration frequency. By Equation (4), a complex eigenvalue analysis can be carried out
determine the critical wind velocity and vibration frequency. Being the conjugate pairs
critical wind velocity 𝑈 , the system has only one eigenvalue 𝜆 with zero real part, an
of complex eigenvalues λ j = σj ± iω j , and the conjugate pairs of complex eigenvectors
theφj imaginary part
= p j ± iq j , the 𝜔 of
system willthe
be complex
dynamically eigenvalue
unstable if the𝜆 becomes
real part ofthe anyflutter frequency.
eigenvalue is It
necessary to provide
positive. Hence, the variation
the condition of both
for the onset wind instability
of flutter velocity and vibration
is stated frequency
as follows: at a in t
certain critical
complex eigenvaluewind velocity
analysis,U f , so
thethat
system has only one eigenvalue
a mode-by-mode tracking λ f with
methodzero real
canpart,
be employ
toand the imaginary
iteratively search part j of the complex
theωflutter frequency eigenvalue f becomes
and theλflutter the flutter
velocity [31].frequency. It
is necessary to provide the variation of both wind velocity and vibration frequency in the
The motion-dependent effect due to the wind is taken into account for each deck
complex eigenvalue analysis, so that a mode-by-mode tracking method can be employed
ement by the search
to iteratively elementthe aeroelastic
flutter frequency stiffness
and the and damping
flutter velocitymatrices.
[31]. These matrices shou
be properly compiled to implement
The motion-dependent effect due to thetheaeroelastic
wind is taken motion-dependent
into account for each loaddeck
by means
the flutterbyderivatives.
element User defined
the element aeroelastic MATRIX27
stiffness and damping element in ANSYS
matrices. Mechanical
These matrices shouldAPDL c
be properly
only model either compiled to implement
an aeroelastic the aeroelastic
stiffness matrixmotion-dependent
or an aeroelasticload by means
damping of inste
matrix
the flutter derivatives. User defined MATRIX27 element in ANSYS
of both of them simultaneously, so a pair of MATRIX27 elements must be attached to ea Mechanical APDL can
only model either an aeroelastic stiffness matrix or an aeroelastic damping matrix instead
node of a generic bridge deck element as illustrated in Figure 5. The MATRIX27 elemen
of both of them simultaneously, so a pair of MATRIX27 elements must be attached to each
𝑒1node
andof𝑒3 represent
a generic respectively
bridge deck elementthe aeroelastic
as illustrated stiffness
in Figure andMATRIX27
5. The of the node 𝑖,
dampingelements
the 𝑒2 and 𝑒4
e1 and e3 represent respectively the aeroelastic stiffness and damping of the node i, as stiffness
MATRIX27 elements represent respectively the aeroelastic the an
damping
MATRIX27 the nodee2 𝑗.
of elements and e4 represent respectively the aeroelastic stiffness and damping
of the node j.

Figure 5. Hybrid finite element model: “e”, structural element; “e1”, “e2”, “e3” and “e4”, MAYRIX27
aeroelastic
Figure element.
5. Hybrid finite element model: “e”, structural element; “e1”, “e2”, “e3” and “e4”, MAYRIX
aeroelastic element.
4. The George Washington Bridge
As mentioned in the Introduction, the two configurations of the George Washington
4. bridge
The George Washington
were chosen Bridge
as case studies (Figure 1). The deck cross-sections of both configura-
Aswere
tions mentioned
modeled in ANSYS
the Introduction,
FLUENT, where the two configurations
the steady-state of the George
aerodynamic Washingt
coefficients
were calculated via CFD analyses. Finally, full bridge models were developed in
bridge were chosen as case studies (Figure 1). The deck cross-sections of both configur ANSYS
Mechanical APDL to perform structural and flutter analyses. Flutter derivatives character-
tions were modeled in ANSYS FLUENT, where the steady-state aerodynamic coefficien
izing the aeroelastic load of full bridge models were calculated by means of Equation (3),
were
withcalculated
Cl0 and Cm via CFDfrom
0 resulting analyses.
the CFDFinally,
analyses.full bridge models were developed in ANSY
 Mechanical APDL to perform structural and flutter analyses. Flutter derivatives charac-
terizing the aeroelastic load of full bridge models were calculated by means of Equation
(3), with 𝐶𝑙′ and 𝐶𝑚′
resulting from the CFD analyses.

Appl. Sci. 2023, 13, 6389 4.1. Steady-State Aerodynamic Coefficients and Flutter Derivatives 9 of 21

Once the Fung formulation was validated, the steady-state aerodynamic coefficients
of both the George Washington deck cross-sections were calculated in ANSYS FLUENT.
4.1. Steady-State Aerodynamic Coefficients and Flutter Derivatives
For each section model and different wind attack angles, RANS simulations were per-
Once the Fung formulation was validated, the steady-state aerodynamic coefficients of
formed modeling the turbulence by the 𝑘 − 𝜔 𝑆𝑆𝑇 method, following the indications in
both the George Washington deck cross-sections were calculated in ANSYS FLUENT. For
[24,37]. In order not to interrupt the logical flow of the speech, more information is given
each section model and different wind attack angles, RANS simulations were performed
in modeling
Appendixthe A.turbulence
The resulting
by thesteady-state aerodynamic
k − ω SST method, following coefficients in 𝐶
of lift,
the indications 𝑙 , and mo-
[24,37].
ment, 𝐶𝑚 , are
In order not shown in Figure
to interrupt 6, where
the logical flowdashed lines represent
of the speech, the linearisregressions.
more information given in In
theAppendix
case of the
A. double deck (Figure
The resulting 6b), aerodynamic
steady-state static coefficients deviate
coefficients from
of lift, Cllinearity
, and moment,for higher
Cm , are
angles ofshown
attack,in so
Figure
the 6,linear
whereregressions
dashed lines are
represent the linear
restricted intervalsIn[−2°;
regressions.
to the the case2°] and
of the
[−1°; 1°]double
for liftdeck
and(Figure
moment 6b),coefficients,
static coefficients deviate from
respectively. The linearity for higherofangles
first derivatives static coef-
of attack,
ficients so origin,
at the the linear regressions
calculated asare
therestricted
slopes oftolinear
the intervals [−2◦ ; 2are:
regressions,
◦ ] and′ [−1◦ ; 1◦ ] for ′
𝐶𝑙 = −6.291, 𝐶𝑚 =
lift and moment coefficients, respectively. The first ′+ derivatives of static coefficients at the
1.902 for the single-deck and 𝐶𝑙′ = −11.024, 𝐶𝑚 = 1.954 for the double-deck configura-
origin, calculated as the slopes of linear regressions, are: Cl0 = −6.291, Cm 0 = 1.902 for the
tions.
single-deck and C 0 = −11.024, C 0+ = 1.954 for the double-deck configurations.
l m

(a) (b)
Figure
Figure6.6.Steady-state aerodynamic
Steady-state aerodynamic coefficients
coefficients of liftof
andlift and moment
moment for the
for the George George Bridge.
Washington Washington
Bridge. (a) Upper
(a) Upper deck;
deck; (b) (b) Double-deck.
Double-deck.

Flutterderivatives,
Flutter derivatives, calculated
calculated accordingly by means
accordingly of Equations
by means (3a)–(3h), are
of Equations shown are
(3a)–(3h),
in Figures 7 and 8. These were used to calculate the aerodynamic lift and
shown in Figures 7 and 8. These were used to calculate the aerodynamic lift and moment moment actions
to be applied to the bridge section.
actions to be applied to the bridge section.
For the double-deck section, the lift and moment actions, based on static coefficients
of the whole section, are applied to the top deck only, as sketched in Figure 9a. In addition,
a second alternative method was adopted, which allows avoiding the use of data in
Figure 6b and provides an additional model for comparison and validation of results.
This latter method permits the same aeroelastic forces calculated for the single deck to
be simultaneously applied to both upper and lower decks, as illustrated in Figure 9b.
This procedure assumes that there is no aerodynamic interference between the two decks,
whereas the mechanical interference is correctly modeled by the elements composing the
truss. Furthermore, the pressure coefficients of the lower deck are assumed to be equal to
those of the upper one and the aerodynamic influence of the secondary truss elements is
also neglected. Despite the previous simplifications, the advantage of this procedure lies in
the greater reliability of the results of CFD analysis for the single deck. Moreover, it can be
conjectured that the simplified formulation previously introduced for the calculation of
flutter derivatives (see Equations (3a)–(3h)), being based on the aerodynamics of the thin
Figure 7. Flutter
airfoil, is morederivatives
suitable forofthe
thedescription
George Washington Bridge’s
of the behavior of aupper
singledeck
deckaccording
rather thanto of
Equations
a
(3a)–(3h).
double deck.
 Figure 6. Steady-state aerodynamic coefficients of lift and moment for the George Washington
Bridge. (a) Upper deck; (b) Double-deck.

Flutter derivatives, calculated accordingly by means of Equations (3a)–(3h), are

Appl.2023,
Appl. Sci. Sci. 2023, 13, 6389
13, 6389 10 ofmoment
shown in Figures 7 and 8. These were used to calculate the aerodynamic lift and 21
10 of 21
actions to be applied to the bridge section.

H i (-)

Ai (-)
Appl. Sci. 2023, 13, 6389 Figure 8. Flutter derivatives of the George Washington bridge’s double deck according to Equations
10 of 21
(3a)–(3h).
Figure
Figure 7. Flutter
7. derivatives
Flutter of the
derivatives of George Washington
the George Bridge’s
Washington upper deck
Bridge’s upperaccording to Equations
deck according to
(3a)–(3h). (3a)–(3h).
Equations
For the double-deck section, the lift and moment actions, based on static coefficients of
the whole section, are applied to the top deck only, as sketched in Figure 9a. In addition, a
second alternative method was adopted, which allows avoiding the use of data in Figure 6b
and provides an additional model for comparison and validation of results. This latter
method permits the same aeroelastic forces calculated for the single deck to be simultane-
ously applied to both upper and lower decks, as illustrated in Figure 9b. This procedure
assumes that there is no aerodynamic interference between the two decks, whereas the me-
chanical interference is correctly modeled by the elements composing the truss. Further-
more, the pressure coefficients of the lower deck are assumed to be equal to those of the
upper one and the aerodynamic influence of the secondary truss elements is also neglected.
Despite the previous simplifications, the advantage of this procedure lies in the greater reli-
ability of the results of CFD analysis for the single deck. Moreover, it can be conjectured that
the simplified formulation previously introduced for the calculation of flutter derivatives
Figure
Figure 8. Flutter
(see Equations derivatives
Flutter
(3a)–(3h)), of the
derivatives
being George
ofbased onWashington
the George bridge’s
Washington
the aerodynamics ofdouble
bridge’s deck
the thindoubleaccording
airfoil, deck to Equations
according
is more suita- to
(3a)–(3h).
Equations (3a)–(3h).
ble for the description of the behavior of a single deck rather than of a double deck.

For the double-deck section, the lift and moment actions, based on static coefficients of
the whole section, are applied to the top deck only, as sketched in Figure 9a. In addition, a
second alternative method was adopted, which allows avoiding the use of data in Figure 6b
and provides an additional model for comparison and validation of results. This latter
method permits the same aeroelastic forces calculated for the single deck to be simultane-
ously applied to both upper and lower decks, as illustrated in Figure 9b. This procedure
assumes that there is no aerodynamic interference between the two decks, whereas the me-
chanical interference is correctly modeled by the elements composing the truss. Further-
more, the pressure coefficients of the lower deck are assumed to be equal to those of the
upper one and the aerodynamic influence of the secondary truss elements is also neglected.
Despite the previous simplifications,
(a) the advantage of this procedure lies in the greater reli-
ability of the results of CFD analysis for the single deck. Moreover, it can be conjectured that
the simplified formulation previously introduced for the calculation of flutter derivatives
(see Equations (3a)–(3h)), being based on the aerodynamics of the thin airfoil, is more suita-
ble for the description of the behavior of a single deck rather than of a double deck.

(b)

Figure 9. Sketch of the alternative methods adopted for the definition and application of aeroelastic
loads to the double-deck cross-section. (a) Lift and moment actions, based on static coefficients of the
whole section, applied to the top deck only; (b) Lift and moment actions, based on static coefficients
(a)to both upper and lower decks.
of the single-deck section, applied
 loads to the double-deck cross-section. (a) Lift and moment actions, based on static coefficients of
the whole section, applied to the top deck only; (b) Lift and moment actions, based on static coeffi-
cients of the single-deck section, applied to both upper and lower decks.

Appl. Sci. 2023, 13, 6389 4.2. Finite Element Models 11 of 21

The ANSYS Mechanical APDL numerical models were realized with the geometrical
properties found in literature [38–40]. The suspension cables were modeled as beam ele-
ments (BEAM188),
4.2. Finite Element Modelswith a circular cross-section of radius 0.573 m. An equivalent elastic
modulus of 1.07 × 10
The ANSYS Mechanical 11 Pa was adopted in order to take in to account the influence of the side
APDL numerical models were realized with the geometrical
spans. The hangers
properties found in were modeled
literature using The
[38–40]. LINK180 elements,
suspension cablesthat transmit
were modeledonlyastensile
beamforces
elements
between the(BEAM188),
deck and the withmain
a circular cross-section
cables. The floor of radiuswas
system 0.573modeled
m. An equivalent
by meanselastic
of equiva-
modulus of 1.07 × 10 11 Pa was adopted in order to take in to account the influence of
lent beams having the properties listed in Table 1, wherein the V-shaped up-winds and the
the side
lower spans.refer
elements The hangers were modeled
to the double-deck using LINK180
configuration. elements,
The thatmaterial
following transmit properties
only
tensile forces between the deck and the main cables. The floor system was modeled by
were assigned to each element except the suspension cables: Young’s modulus = 2.1 × 1011
means of equivalent beams having the properties listed in Table 1, wherein the V-shaped
Pa,up-winds
Poisson’sand ratiothe= lower
0.3 and mass density
elements refer to=the
7860 kg/m3. Point
double-deck mass elements
configuration. The(MASS21)
following were
attached
materialtoproperties
the upperwere chords and to
assigned tothe
eachequivalent longitudinal
element except beams,cables:
the suspension both lower
Young’sand up-
per,
modulus = 2.1 × 10 Pa, Poisson’s ratio = 0.3 and mass density = 7860 kg/m . Point mass and
in order to model 11 the inertial contribution of the sidewalk slabs, roadway
3 slabs
other secondary
elements (MASS21) elements. According
were attached to upper
to the Dana et al. [39],
chords andatototal dead loadlongitudinal
the equivalent of 417 kN/m and
569beams,
kN/mboth were lower and upper,
obtained in order
for the to model
single-deck thethe
and inertial contribution
double-deck of the sidewalk
configurations, respec-
slabs, roadway slabs and other secondary elements. According to Dana
tively. Lastly, the MATRIX27 elements were attached to each node of the equivalent longi- et al. [39], a total
dead load
tudinal beams of 417 kN/mthe
to model andaeroelastic
569 kN/m were
forcesobtained
in termsforofthe single-deck
aeroelastic and theand
stiffness double-
damping.
deck configurations, respectively. Lastly, the MATRIX27 elements were attached to each
node of the equivalent longitudinal beams to model the aeroelastic forces in terms of
Table 1. Geometrical properties of the truss girder of the double-deck George Washington Bridge.
aeroelastic stiffness and damping.
Area Bending Moments of Inertia [m4]
Elements
Table 1. Geometrical properties of the truss girder of the
2 double-deck George Washington Bridge.
[m ] Horizontal Axis Vertical Axis
Equivalent longitudinal upper beam 0.4465 Iy = 0.1867
Bending Iz =[m
Moments of Inertia 38.6601
4]
Elements Area [m2 ]
Upper chord 0.0549 Iy = 0.0049
Horizontal Axis Vertical IzAxis
= 0.0060
Equivalent transversalupper
Equivalent longitudinal upper beam
beam 0.237
0.4465 Iy I=x =0.1867
0.9498 Iz = 3.8102
Iz = 38.6601
Upper chord
V-shaped up-winds 0.0549
0.0374 IyI= 1 =0.0049
0.0016 Iz = 0.0060
I2 = 0.0006
Equivalent transversal upper beam 0.237 Ix = 0.9498 Iz = 3.8102
Equivalent longitudinal lower beam
V-shaped up-winds
0.3439
0.0374
Iy = 0.1108
I1 = 0.0016
Iz = 29.3241
I2 = 0.0006
Equivalent Lower chord
longitudinal lower beam 0.0484
0.3439 Iy I=y =0.1108
0.0029 Iz = 0.0051
Iz = 29.3241
Lower chord
Equivalent transversal lower beam 0.0484
0.0816 Iy I=x =0.0029
0.1559 I z = 0.0051
Iz = 1.3645
Equivalent transversal lower beam 0.0816 Ix = 0.1559 Iz = 1.3645

According to the two different methods introduced in the previous section for the
According
double-deck to the two different
configuration, methods
two different introduced
models wereinrealized.
the previous section
In the firstfor the the
model,
double-deck configuration, two different models were realized. In the first model, the pairs
pairs of MATRIX27 elements, compiled with the flutter derivatives in Figure 8, were at-
of MATRIX27 elements, compiled with the flutter derivatives in Figure 8, were attached
tached only to the upper deck. Whereas, according to the second method, MATRIX27 el-
only to the upper deck. Whereas, according to the second method, MATRIX27 elements,
ements, compiled
compiled with thewith
flutterthe flutter derivatives
derivatives plotted
plotted in Figure in Figure
7, were 7, to
attached were
bothattached
upper and to both
upper and lower decks. The finite element models are shown
lower decks. The finite element models are shown in Figure 10. in Figure 10.

(a) (b)

Figure 10. ANSYS Mechanical APDL Finite element model. (a) Single-deck configuration (MATRIX27
elements are shown in red color); (b) Double-deck configuration (MATRIX27 elements are not shown).
Appl.
Appl.Sci. 2023,
Sci. 13,13,
2023, 6389
6389 1212of of2121

Figure
Figure10.10.ANSYS
ANSYSMechanical
MechanicalAPDL
APDLFinite
Finiteelement
elementmodel.
model.(a)(a)Single-deck
Single-deckconfiguration
configuration(MA-
(MA-
Appl. Sci. 2023, 13, 6389 TRIX27 elements are shown in red color); (b) Double-deck configuration (MATRIX27
TRIX27 elements are shown in red color); (b) Double-deck configuration (MATRIX27 elements12are
elements ofare
21
not shown).
not shown).

5.5.Flutter Analyses
Flutter Analyses
5. Flutter Analyses
Before
Before introducing
introducing the
theaeroelastic
aeroelastic nodal
nodal forces
forces bybyMATRIX27
MATRIX27 elements,
elements, a modal anal-
Before introducing the aeroelastic nodal forces by MATRIX27 elements, aa modal
modal anal-
anal-
ysis with
ysis with no
with no wind
no wind was
wind was performed:
was performed:
performed: the the results
the results are
results are summarized
are summarized
summarized in in Figures
in Figures
Figures 1111 and
11 and 12,
and 12, where
12, where
where
ysis
the
the
thelabels
labelsV,V,T,T,
V, LL
T, stand
Lstand
standfor vertical,
for
for vertical, torsional
vertical, torsional and
torsional and lateral,
andlateral,respectively;
respectively;
lateral, respectively; andand the
the
and labels
labels
the AA and
labels and
A
S and
stand
S stand for antisymmetric
for antisymmetric
S stand for antisymmetric and symmetric,
and symmetric,
and symmetric,respectively.
respectively. Since it is
Since Since
respectively. well
it is well known
it isknown that
well known flutter
that flutter
that
instability
instability
flutter usually
usually
instability involves
involves
usually mainly
mainly
involves thethefirst
mainly first
themodes
modes atatlower
first modes at frequencies
lower frequencies
lower [31],
frequencies[31],only
onlythe
[31], thefirst
only first
the
six modes
six modes
first were
six modes extracted
werewere
extracted for both
for both
extracted configurations.
configurations.
for both The
configurations. geometric
The geometric nonlinearity
The geometric nonlinearity of the
of the
nonlinearity struc-
struc-
of the
ture
turewas
wastaken
structure was into
intoaccount
takentaken account
into running
running
account the
running themodal
modal analysis
the modal analysis inina apreloaded
analysis preloaded
in a preloadedconfiguration
configuration
configurationac-ac-
counting
counting forforthe
thegravity
gravity loads.
loads. Once
Once eigenfrequencies
eigenfrequencies and
and
accounting for the gravity loads. Once eigenfrequencies and eigenmodes have been eigenmodes
eigenmodes have
have been
been evalu-
evalu-
ated,
ated,the
ated, theflutter
the flutteranalysis
flutter analysiswas
analysis wasperformed
was performedbybythe
performed themethod
the methoddescribed
method describedinin
described in[31].
[31].The
[31]. Theaeroelastic
The aeroelastic
aeroelastic
loads were
loads were
loads modeled
were modeled as nodal
modeled as nodal forces
nodal forces affecting
forces affecting the
affecting the stiffness
the stiffness and
stiffness and damping
and damping matrices
matricesofof
damping matrices ofthethe
the
element
element to
element to which
to which the
which the nodes
the nodes belong,
nodes belong, so
belong, so MATRIX27
so MATRIX27 elements
MATRIX27 elements were
elements were incorporated
were incorporated
incorporated into into
intothethe
the
structural
structuralmodel
structural modeltoto
model toperform
performthe
perform thedamped
the dampedeigenvalue
damped eigenvalueanalyses.
eigenvalue analyses.
analyses.

Mode
Mode1,1,LS,
LS,0.061
0.061Hz
Hz Mode
Mode2,2,VA,
VA,0.116
0.116Hz
Hz

Mode
Mode3,3,TA,
TA,0.143
0.143Hz
Hz Mode
Mode4,4,LA,
LA,0.152
0.152Hz
Hz

Mode
Mode5,5,VS,
VS,0.159
0.159Hz
Hz Mode
Mode6,6,TS,
TS,0.198
0.198Hz
Hz
Figure
Figure11.
Figure Eigenfrequencies
11.
11. and
Eigenfrequencies
Eigenfrequencies eigenmodes
and
and ofof
eigenmodes
eigenmodes the
of thesingle-deck
the configuration
single-deck
single-deck (1931).
configuration
configuration (1931).
(1931).

Mode
Mode1,1,LS,
LS,0.063
0.063Hz
Hz Mode
Mode2,2,VA,
VA,0.126
0.126Hz
Hz

Mode
Mode3,3,LA,
LA,0.161
0.161Hz
Hz Mode
Mode4,4,VS,
VS,0.164
0.164Hz
Hz

Mode
Mode5,5,TA,
TA,0.262
0.262Hz
Hz Mode
Mode6,6,TS,
TS,0.274
0.274Hz
Hz
Figure 12.12.
Figure Eigenfrequencies and
Eigenfrequencies eigenmodes
and ofof
eigenmodes the double-deck
the configuration
double-deck (1962).
configuration (1962).
Figure 12. Eigenfrequencies and eigenmodes of the double-deck configuration (1962).

Finally,
Finally,the
Finally, theiterative
the iterativeprocedure
iterative procedurefor
procedure forthe
for thedetermination
the determinationofof
determination offlutter
flutterspeed
flutter speedand
speed andfrequency
frequency
was
was
was carried out.
carried The
out. The previous
previous computational
computational steps
stepswere
were followed
followed for the
for
out. The previous computational steps were followed for the eigenmodes eigenmodes
the eigenmodes col-
col-
lected in
collectedFigures
lected in Figures 11
in Figures and
11 11 12.
andand The
12.12. damped
TheThe
damped complex
damped complex eigenvalue
complexeigenvalue analyses
eigenvalue analyses were
analyseswereconducted
conducted
were for
conductedfor
the
themodel
for the
model under
model
under different
under different wind
different velocities.
wind
wind The
velocities.
velocities. step
The
The increment
step
step increment
increment ofof
wind
of wind
windvelocity
velocitywas
velocity wasset
set
set
variable, from a maximum value of 10 m/s in the ranges far from the flutter instability to a
minimum of 1 m/s close to the instability value. Accordingly, the resulting flutter wind
velocity has the accuracy of one m/s. The variation of the complex eigenvalues vs. wind
 1. The first method consists of the application of the aeroelastic forces to the upper deck,
where the latter are expressed via the flutter derivatives evaluated using the steady-
state aerodynamic coefficients obtained for the whole two-deck cross-section (see
Figures 6b, 8 and 9a).
2. The second method consists of the application of the aeroelastic forces to both decks,
Appl. Sci. 2023, 13, 6389 13 of 21
where the latter are expressed via the flutter derivatives evaluated using the steady-
state aerodynamic coefficients obtained for the upper deck (see Figures 6a, 7 and 9b).
The analyses furnished a critical flutter speed of 40 m/s (144 km/h) and a critical flut-
velocity is plotted
ter frequency in Figures
of 0.130 13–15,
Hz for where the labels
the single-deck within parentheses
configuration. As regards refer
theto the mode
double-deck
shapes under wind
configuration, flutteraction:
speedsforofinstance, the label
107 m/s (385.2 VT indicates
km/h) and 79 m/sa mode
(284.4 characterized by
km/h), and flutter
both vertical and torsional components, in which the vertical component
frequencies of 0.203 Hz and 0.195 Hz, were provided by the first and the second method,is predominant.
The real and imaginary parts of the eigenvalues represent the logarithm decay rates and
respectively.
damped As vibration frequencies
already said, of these
the variations of modes, respectively.
the imaginary part ofAs
thestated above,
complex the flutterin
eigenvalues
condition occurs when the real part of any eigenvalue becomes positive.
Figures 13–15 describe the change in the oscillation frequencies with the increasing The results
wind
obtained for the 1962
speed, whereas configuration
the variations of therefer to the
real part two
tell us methods previously
about stability of eachintroduced and
single vibration
recalled hereafter:
mode, and thus of the bridge deck. Accordingly, the first torsional-vertical anti-symmetric
1.mode The first ismethod
(TVA) consists
responsible of the instability
for flutter applicationin of
allthe
the aeroelastic forces
analyzed cases. tohigher
For the upper
wind
deck,
speeds, where the
a second latter
flutter are is
mode expressed
attained via thetorsional-vertical
in the flutter derivatives evaluated
symmetric using
mode the
(TVS);
steady-state
see Figures 13–15.aerodynamic coefficients
For the double-deck obtained for the whole
configuration, two-deck
first and cross-section
the second flutter ve-
(seeare
locities Figures
rather6b, 8 and
close, 9a). close also the corresponding eigenfrequencies; see Figures
being
2.14 and
The15.
second
In relation to that, weofcan
method consists thesay
application of the aeroelastic
that interaction forces to
between modes both decks,
having similar
where the latter are expressed via the flutter derivatives evaluated
frequencies could be a triggering factor for instability, thus lowering the critical using the steady-
wind
state aerodynamic coefficients obtained for the upper deck (see Figures 6a, 7 and 9b).
speed.

Appl. Sci. 2023, 13, 6389 14 of(la-

Figure 13. Variation of complex eigenvalues vs. wind velocity for the single-deck configuration 21
Figure 13. Variation of complex eigenvalues vs. wind velocity for the single-deck configuration
bels within parentheses refer to mode shapes under wind action).
(labels within parentheses refer to mode shapes under wind action).

Figure
Figure 14.14.Variation
Variationofof complexeigenvalues
complex eigenvaluesvs.vs.wind
windvelocity
velocityfor
forthe
thedouble-deck
double-deckconfiguration:
configuration:
first method (labels within parentheses refer to mode shapes under wind
first method (labels within parentheses refer to mode shapes under wind action). action).
Appl. Sci. 2023, 13, 6389 14 of 21
Figure 14. Variation of complex eigenvalues vs. wind velocity for the double-deck configuration:
first method (labels within parentheses refer to mode shapes under wind action).

Figure15.
Figure 15.Variation
Variationofofcomplex
complexeigenvalues
eigenvaluesvs.
vs.wind
windvelocity
velocityfor
forthe
thedouble-deck
double-deckconfiguration:
configuration:
second method (labels within parentheses refer to mode shapes under wind action).
second method (labels within parentheses refer to mode shapes under wind action).

Theanalyses
The two different methods
furnished adopted
a critical for the
flutter speeddouble-deck
of 40 m/sconfiguration
(144 km/h) and furnish a dif-
a critical
ference of about 26% and 4% in the estimated critical wind speed and
flutter frequency of 0.130 Hz for the single-deck configuration. As regards the double- flutter frequency,
respectively.
deck Moreover,
configuration, flutterthey
speedsprovide
of 107similar frequency
m/s (385.2 km/h)trends
and 79asm/s
wind speed
(284.4 increases
km/h), and
and the
flutter same prediction
frequencies of 0.203 ofHz
theandflutter mode.
0.195 Hz, Both
wereintroduce
provided simplifications
by the first andfrom the aer-
the second
odynamic
method, standpoint: in addition to the approximate evaluation of the flutter derivatives,
respectively.
the former
As alreadymethod
said,imposes the torsion
the variations of therotation
imaginaryaxis part
of theofcross-section
the complexto coincide with
eigenvalues in
the upper
Figures deck
13–15 axis; the
describe latter
the changeassumes
in thethat there is frequencies
oscillation no aerodynamic with interference
the increasing between
wind
the two
speed, decks and
whereas neglects the
the variations of influence of secondary
the real part tell us about truss-elements. Lastly,
stability of each singleit vibration
should be
noted and
mode, thatthus
the use of flutter
of the bridgederivatives itself represents
deck. Accordingly, the firstatorsional-vertical
strong simplification of the com-
anti-symmetric
plex aerodynamic
mode behavior
(TVA) is responsible forofflutter
bluff bodies, as bridge
instability decks.
in all the analyzed cases. For higher wind
speeds, a second flutter mode is attained in the torsional-vertical symmetric mode (TVS); see
Simplified
Figures Formulations
13–15. and Comparisons
For the double-deck configuration, the first and the second flutter velocities
are rather close, being close also
Given the absence of reference the corresponding
data for a directeigenfrequencies; see Figures
validation, different 14 and for-
simplified 15.
In relation to that, we can say that interaction between modes having
mulas available in the literature for predicting the flutter wind velocity were employedsimilar frequencies
could be a triggering
for a comparison. The factor forare
results instability,
collectedthus lowering
in Table 2. Thethe critical wind
simplified speed.
formulas adopted are
The two different methods adopted for the double-deck configuration furnish a dif-
ference of about 26% and 4% in the estimated critical wind speed and flutter frequency,
respectively. Moreover, they provide similar frequency trends as wind speed increases and
the same prediction of the flutter mode. Both introduce simplifications from the aerody-
namic standpoint: in addition to the approximate evaluation of the flutter derivatives, the
former method imposes the torsion rotation axis of the cross-section to coincide with the
upper deck axis; the latter assumes that there is no aerodynamic interference between the
two decks and neglects the influence of secondary truss-elements. Lastly, it should be noted
that the use of flutter derivatives itself represents a strong simplification of the complex
aerodynamic behavior of bluff bodies, as bridge decks.

Simplified Formulations and Comparisons

Given the absence of reference data for a direct validation, different simplified formulas
available in the literature for predicting the flutter wind velocity were employed for a
comparison. The results are collected in Table 2. The simplified formulas adopted are
in Equations (5a)–(5f). The first one is based on the reduction of the divergence wind
velocity (Ud ) and is the only one involving the aerodynamics of the deck by means of the
derivative of the moment coefficient (Cm0 ). Equations (5b)–(5d) are empirical formulas fitted

for thin airfoils, while Equations (5e) and (5f) contain a coefficient fitted for suspension
bridge decks.
Appl. Sci. 2023, 13, 6389 15 of 21

Table 2. Flutter velocity according to different simplified formulations.

Flutter Velocity [m/s]

Bridge
Configuration Present
Frandsen [41] Selberg [42] Rocard [42] Matsumoto [43] Van der Put [44] Fu and Wang [45]
Work
Single-deck 39.56 36.08 36.21 36.09 38.70 46.45 40
Double-deck 121 113.86 114.44 113.87 89.23 97.76 107–79

Frandsen [41]:
s !s  2
2I ωv
Uf = Bωt 0
1− (5a)
ρB4 Cm ωt

Selberg [42]: s 
1
U f = 2.623 f t B 1− 2
r α µ1 . (5b)
γω
Rocard [42]: s
rα2 µ1

1
U f = 6.282 f t B 1− 2
. (5c)
γω 1 + 8rα2
Matsumoto [43]:
u√ "
v
 #
fv 2
u Im 
U f = 3.71B f t t 1− . (5d)
ρB3 ft

Van der Put [44]

"  r #
ωt 2r B
Uf = η 1 + − 0.5 0.72µ2 ωv . (5e)
ωv B 2

Fu and Wang [45]: !

r
2r
U f = η 2.5 µ2 B f t . (5f)
B
q q
ωt I 2m 4m I
where γω = ω v
, r α = mB2
, µ 1 = ρB2
, µ 2 = πρB2
, r = m . I is the polar mass moment
of inertia per unit span length, m is the mass per unit span length, ωv , ωt are the angular
frequencies of the first vertical mode and the first torsional mode, f v , f t are the respective
frequencies and η is an empirical coefficient representing the difference between the flat
plate and a generic profile (equal to 0.7 for suspension bridges [44]). Other simplified
formulations based on flutter derivatives can be found in [46].
Results provided by the different formulations for the single-deck configuration are similar
to the ones obtained in the present work within the range [−10%; 15%]. Larger variations are
found for the double-deck configuration. In particular, formulas fitted for the thin airfoil provide
a considerable overestimation of the critical speed with respect to the formulas fitted for the
suspension bridges. The flutter limit provided by the first method is close to the prediction
obtained by Equations (5b)–(5d), whereas the flutter velocity provided by the second method is
closer to the predictions of Equations (5e) and (5f), and is the most conservative prediction.
Hence, although for a single-deck configuration having the aerodynamic coefficients simi-
lar to those of the airfoil it is possible to obtain reasonable results using simplified formulations,
for a double-deck configuration it is necessary to perform more detailed analyses.

6. Discussion
The flutter stability of the George Washington bridge was investigated for the two
different configurations, i.e., the original single-deck and the current double-deck. Due to
the lack of literature data on the aerodynamic parameters for the bridge under consideration,
two simplifications were introduced for the description of the motion-related wind loads:
Appl. Sci. 2023, 13, 6389 16 of 21

one for the calculation of flutter derivatives and the other for the application of aerodynamic
loads on the two decks of the current configuration.
The method adopted for the calculation of flutter derivatives allows them to be ob-
tained by analytical formulations based on the steady-state aerodynamic coefficients. The
latter parameters were calculated by CFD simulations using the finite element software
ANSYS FLUENT. The application of the method provided good results on the bridges
chosen for validation, namely the Great Belt and the Akashi Kaykio. The main advantage
lies in the relative simplicity of the calculation of the steady-state aerodynamic coefficients,
on which the formulation is based. In fact, the computational cost required is much lower
than that which characterizes the CFD analyses necessary for a direct numerical calculation
of flutter derivatives. Of course, this approach furnishes approximate predictions. To
further validate the method, comparisons should be extended to several bridge decks for
which both steady-state aerodynamic coefficients and flutter derivatives are available. Once
the flutter derivatives were defined, the flutter analysis was performed by a finite element
ANSYS Mechanical APDL model endowed with MATRIX27 elements [31].
With regards to the double-deck configuration, two alternative methods were adopted
for the definition and application of aeroelastic loads: a first one where lift and moment
actions, based on static coefficients of the whole section, are applied to the top deck
only; and a second one where lift and moment actions, based on static coefficients of the
single-deck section, are applied to both upper and lower decks. Both methods introduce
simplifications from the aerodynamic viewpoint. The former provides predictions similar
to those of the simplified formulas calibrated for the airfoil. The latter is more aligned
with the results provided by simplified formulations accounting for the difference between
airfoil and deck cross-sections. With respect to the modal frequencies, the two methods
provide similar trends for increasing wind speed, a 4% difference in the critical frequency
estimation, and the same prediction of the flutter mode. According to the results of this
study, the presence of the lower deck has raised the flutter wind velocity by more than 200%.
Although a definitive validation of the numerical results was not possible, due to the
lack of experimental data, the authors believe that the results obtained here can be used for
future comparisons with others obtained by different methods. In addition, the proposed
approach provides an approximate way for estimating the flutter velocity of double-deck
long-span bridges in the absence of more detailed analyses.

7. Concluding Remarks
Two simplified methods were adopted to achieve a preliminary flutter stability assess-
ment of the double-deck George Washington Bridge. Both methods provided a realistic
outcome, the second one seeming the most reliable and being the most conservative. The
predicted critical flutter wind velocity is about 79 m/s (284 km/h). According to [47], for the
bridge site, the wind speed corresponding to approximately a 1.6% probability of exceedance
in 50 years is 129 mph (about 57.7 m/s). Given the lack of experimental data, it is suggested
that wind tunnel tests are needed to obtain the flutter derivatives of the bridge deck.

Author Contributions: Conceptualization, A.C., G.P. and L.P.; data curation, G.P. and S.R.; formal
analysis, S.R.; investigation, S.R.; methodology, G.P., L.P. and S.R.; resources: A.C., G.P. and L.P.;
supervision, A.C.; validation, S.R.; visualization, S.R.; writing—original draft preparation, G.P. and
S.R.; writing—review and editing, A.C. and L.P. All authors have read and agreed to the published
version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Some or all data, models or code that support the findings of this study
are available from the corresponding author upon reasonable request.
Conflicts of Interest: The authors declare no conflict of interest.
Appl. Sci. 2023, 13, 6389 17 of 21

Appendix A. RANS Simulations

The finite element package ANSYS FLUENT was chosen both for meshing and sim-
ulating the fluid flow around the rigid body, with the aim of integrating the pressure
distribution and finally obtain the steady-state aerodynamic coefficients. For each deck
cross-section model, a 2D RANS simulation was performed. The k − ω SST turbulence
model, developed by [48], was adopted. This method consists in a sort of combination of
the two simpler, widely used k − ε and k − ω models, whose weighting is controlled by the
wall distance. Both are two-equation models, whose differences can be summarized as fol-
lows: the k − ε model does not allow for a direct integration of the field equations through
the boundary layer because the ε parameter tends to zero close to the wall. Conversely, the
k − ω model allows for directly integrating through the boundary layer, thus permitting to
improve the goodness of the wall boundary layer unsteady-state solution, as demonstrated
in [49]. As a drawback, the k − ω model has proved to be highly sensitive to inlet turbu-
lence boundary conditions: this can sensibly affect the solution even in large computational
domains [48]. To overcome this drawback, the k − ω SST model can conveniently be used.
It preserves the main advantages of the classical k − ω model, but it has proved to be less
sensitive to the inlet conditions.
After several tests regarding the turbulence model, the mesh sizing, the extension
of the computational domain, and the boundary conditions, the following settings were
adopted. Reference was made to [24,37] as a guide. For what concerns the computational
domain, the outer rectangle in Figure A1 delimits the fluid region, the left and right sides
representing the velocity inlet and the pressure outlet, respectively. The dimensions of the
computational domain were set following the indication in [24]: around 12B (360 m) in the
along-wind direction and 5.5B (180 m) in the transverse one. The cross-section centroid is
located at a horizontal distance of 120 m from the velocity inlet. The domain is subdivided
into different regions as shown in Figure A1. The circle and the inner rectangle were created
as auxiliary geometric entities so to allow for a differentiated meshing of the region closer
to the obstacle and of the wind wake. The circle and the inner rectangle have radius and
height equal to 35 m for the single deck model and 40 m for the double-deck. Being the
dimensions of the two models rather similar, the same mesh size was set for both the single-
and the double-deck versions. As for the mesh, in the inner regions, the mesh dimension
ranges from an edge size of 1 cm (order B × 10−3 ) to a maximum size of 25 cm (order
B × 10−2 ), with a growth ratio equal to 1.1 [24]. In the fluid region between the outer
boundary and the inner parts, a maximum size of 3 m (order B × 10−1 ) and a growth ratio
of 1.2 were set. A quadrilateral mesh was chosen for each model to favor convergence and,
Appl. Sci. 2023, 13, 6389 thus, reduce the computational time. Figure A2 shows an example of meshed model18ofofthe 21
double-deck George Washington Bridge, with different zoom levels.

Figure A1.
Figure A1. Computational
Computational domain
domain for
for RANS
RANS simulation
simulation in
in ANSYS
ANSYS FLUENT
FLUENT (double-deck
(double-deck George
George
Washington Bridge).
Washington Bridge).
Appl. Sci. 2023, 13, 6389 18 of 21
Figure A1. Computational domain for RANS simulation in ANSYS FLUENT (double-deck George
Washington Bridge).

(a) (b)

(c) (d)
Figure
Figure A2. Mesh for RANSA2. Mesh forinRANS
simulation ANSYS simulation
FLUENTin(double-deck
ANSYS FLUENT
George(double-deck
WashingtonGeorge Washington
Bridge).
Bridge). (a) whole domain; (b) transition between sub-domains; (c) mesh around the deck; (d) detail
(a) whole domain; (b) transition between sub-domains; (c) mesh around the deck; (d) detail around
around appendages.
appendages.
Regarding the boundary conditions, the inlet was given a velocity of 30 m/s, a turbu-
Regarding the boundary conditions, the inlet was given a velocity of 30 m/s, a turbu-
lent intensity of 0.5%, according to standard wind tunnel conditions for laminar flow, and
lent intensity of 0.5%, according
a turbulent to standard
viscosity wind
ratio equal to 2.tunnel conditions
The outlet was givenforalaminar flow, and
null pressure andthe same
a turbulent viscosity ratio equal to 2. The outlet was given a null pressure and the same
turbulent intensity and viscosity as the inlet. The bridge deck edges were selected for the
turbulent intensity and viscosity
integration of the as the inlet.
pressure The bridge
distribution. Thedeck edges
model were selected
parameters for the
as the wind velocity, the
integration of thechord
pressure distribution.
length, The model
and the air density parameters
were set as the wind
so that dimensionless velocity,
pressure the
coefficients were
chord length, anddirectly
the air given
density
by were set so that
the program. dimensionless
Moment coefficientspressure coefficients
are referred weredeck cen-
to the upper
directly given by troid. To investigate
the program. Momentdifferent angles ofare
coefficients attacks, several
referred analyses
to the upperwere
deckperformed
centroid.by rotat-
To investigate different angles of attacks, several analyses were performed by rotatingthe
ing the bridge cross-section while keeping the boundary conditions and thecomputa-
tional domain unaltered. Results are summarized in Table A1.
bridge cross-section while keeping the boundary conditions and the computational domain As an example, Figures A3
and A4 show the distribution of velocity magnitude and pressure
unaltered. Results are summarized in Table A1. As an example, Figures A3 and A4 show in the case of horizontal
flow, with 𝑈 = 30 m/s, for the single-deck and the double-deck configurations of the
the distribution of velocity magnitude and pressure in the case of horizontal flow, with
George Washington Bridge, respectively. The simulations were run in a machine having
U = 30 m/s, for the single-deck and the double-deck configurations of the George Wash-
the following main features: Intel® Core™ i9-12900K (30 MB cache memory, 8 P-core + 8
ington Bridge, respectively. The simulations were run in a machine having the following
main features: Intel® Core™ i9-12900K (30 MB cache memory, 8 P-core + 8 E-core, from
3.2 GHz to 5.2 GHz, 125 W); RAM 32 GB DDR5 memory, up to 4400 MHz; video card
Nvidia RTX A4000, 16 GB, 4 DP; 64 bit operating system. The computational time required
for meshing of each model was approximatively 40 min, while the computational time for
RANS simulation was comprised between 5 and 10 min.

Table A1. Resulting steady aerodynamic coefficients for different wind angles of attack.

Single-Deck Double-Deck
Angle of Attack (◦ )
Cl Cm Cl Cm
−3 0.123 −0.09 0.107 −0.035
−2.5 0.030 −0.075 0.128 −0.038
−2 −0.028 −0.058 0.112 −0.045
−1.5 −0.042 −0.041 0.064 −0.050
−1 −0.159 −0.021 −0.043 −0.047
−0.5 −0.163 −0.007 −0.120 −0.038
0 −0.252 0.013 −0.287 −0.023
0.5 −0.274 0.026 −0.357 −0.004
1 −0.334 0.047 −0.516 0.018
1.5 −0.427 0.061 −0.529 0.035
2 −0.441 0.075 −0.730 0.074
2.5 −0.522 0.097 −0.712 0.074
3 −0.538 0.103 −0.762 0.090
 0.5
0.5 −0.274
−0.274 0.026
0.026 −0.357
−0.357 −0.004
−0.004
11 –0.334
–0.334 0.047
0.047 −0.516
−0.516 0.018
0.018
1.5
1.5 −0.427
−0.427 0.061
0.061 −0.529
−0.529 0.035
0.035
22 −0.441
−0.441 0.075
0.075 −0.730
−0.730 0.074
0.074
2.5
2.5 −0.522
−0.522 0.097
0.097 −0.712
−0.712 0.074
0.074
Appl. Sci. 2023, 13, 6389 19 of 21
33 −0.538
−0.538 0.103
0.103 −0.762
−0.762 0.090
0.090

(a)
(a) (b)
(b)
Figure A3.
Figure A3. RANS
A3. RANS simulation contours
RANS simulation contoursfor
contours forsingle-deck
for single-deckGeorge
single-deck GeorgeWashington
George WashingtonBridge
Washington Bridgesection,
Bridge section,
section, zero
zero attack
zero
attack
attack angle,
angle, angle, and 𝑈 = 30
= 30𝑈m/s.
and U and m/s.
= 30(a)
m/s. (a) Velocity;
(a) Velocity;
Velocity; (b) Pressure.
(b) Pressure.
(b) Pressure.

(a)
(a) (b)
(b)
Figure A4.
Figure A4. RANS
RANS simulation
simulation contours
contours for
for double-deck
double-deck George
George Washington
Washington Bridge
Bridge section,
section, zero
zero at-
at-
Figure
tack angle, RANS
A4. and 𝑈 =simulation
30 m/s. (a)contours
Velocity;for
(b) double-deck
Pressure. George Washington Bridge section, zero
tack angle,
attack angle, and𝑈U= =
and 3030m/s.
m/s.(a)(a)
Velocity; (b)(b)
Velocity; Pressure.
Pressure.

References
1. Wagner, H. Über Die Entstehung Des Dynamischen Auftriebes von Tragflügeln. ZAMM-Z. Angew. Math. Mech. 1925, 5, 17–35.
[CrossRef]
2. Theodorsen, T. General Theory of Aerodynamic Instability and the Mechanism of Flutter; NASA: Washington, DC, USA, 1935.
3. Scanlan, R.H. The Action of Flexible Bridges under Wind, I: Flutter Theory. J. Sound Vib. 1978, 60, 187–199. [CrossRef]
4. Scanlan, R.H.; Tomko, J.J. Airfoil and Bridge Deck Flutter Derivatives. J. Eng. Mech. Div. 1971, 97, 1717–1737. [CrossRef]
5. de Miranda, S.; Patruno, L.; Ubertini, F.; Vairo, G. Indicial Functions and Flutter Derivatives: A Generalized Approach to the
Motion-Related Wind Loads. J. Fluids Struct. 2013, 42, 466–487. [CrossRef]
6. Larsen, A. Aerodynamic Aspects of the Final Design of the 1624 m Suspension Bridge across the Great Belt. J. Wind. Eng. Ind.
Aerodyn. 1993, 48, 261–285. [CrossRef]
7. Larsen, A. Aerodynamics of Large Bridges; Routledge: Oxfordshire, UK, 2017; ISBN 9781315136950.
8. Miyata, T.; Yamaguchi, K. Aerodynamics of Wind Effects on the Akashi Kaikyo Bridge. J. Wind Eng. Ind. Aerodyn. 1993, 48,
287–315. [CrossRef]
9. Chen, Z.-S.; Zhang, C.; Wang, X.; Ma, C.-M. Wind Tunnel Measurements for Flutter of a Long-Afterbody Bridge Deck. Sensors
2017, 17, 335. [CrossRef]
10. Siedziako, B.; Øiseth, O.; Rønnquist, A. An Enhanced Forced Vibration Rig for Wind Tunnel Testing of Bridge Deck Section
Models in Arbitrary Motion. J. Wind Eng. Ind. Aerodyn. 2017, 164, 152–163. [CrossRef]
11. Brancaleoni, F.; Brotton, D.M. Analysis and Prevention of Suspension Bridge Flutter in Construction. Earthq. Eng. Struct. Dyn.
1981, 9, 489–500. [CrossRef]
12. Al-Assaf, A. Flutter Analysis of Open-Truss Stiffened Suspension Bridges Using Synthesized Aerodynamic Derivatives. Ph.D.
Thesis, Washington State University, Whitman, WA, USA, 2006.
13. Caracoglia, L.; Jones, N.P. Time Domain vs. Frequency Domain Characterization of Aeroelastic Forces for Bridge Deck Sections. J.
Wind Eng. Ind. Aerodyn. 2003, 91, 371–402. [CrossRef]
14. Diana, G.; Bruni, S.; Cigada, A.; Collina, A. Turbulence Effect on Flutter Velocity in Long Span Suspended Bridges. J. Wind Eng.
Ind. Aerodyn. 1993, 48, 329–342. [CrossRef]
Appl. Sci. 2023, 13, 6389 20 of 21

15. Larose, G.L.; Livesey, F.M. Performance of Streamlined Bridge Decks in Relation to the Aerodynamics of a Flat Plate. J. Wind Eng.
Ind. Aerodyn. 1997, 69–71, 851–860. [CrossRef]
16. Wu, T.; Kareem, A. Bridge Aerodynamics and Aeroelasticity: A Comparison of Modeling Schemes. J. Fluids Struct. 2013, 43,
347–370. [CrossRef]
17. Borri, C.; Costa, C. Quasi-Steady Analysis of a Two-Dimensional Bridge Deck Element. Comput. Struct. 2004, 82, 993–1006.
[CrossRef]
18. Salvatori, L.; Spinelli, P. A Discrete 3D Model for Bridge Aerodynamics and Aeroelasticity: Nonlinearities and Linearizations.
Meccanica 2007, 42, 31–46. [CrossRef]
19. Øiseth, O.; Rönnquist, A.; Sigbjörnsson, R. Simplified Prediction of Wind-Induced Response and Stability Limit of Slender
Long-Span Suspension Bridges, Based on Modified Quasi-Steady Theory: A Case Study. J. Wind Eng. Ind. Aerodyn. 2010, 98,
730–741. [CrossRef]
20. Andersen, M.S.; Eriksen, M.B.; Larsen, S.V.; Brandt, A. The Influence of Gap- and Chord-Widths for Multi-Box Girders: Superpo-
sition of Flat Plate Flutter Derivatives and Section Model Tests. J. Fluids Struct. 2022, 109, 103489. [CrossRef]
21. Fung, Y.-C. An Introduction to the Theory of Aeroelasticity; Dover Publications: Mineola, NY, USA, 2008.
22. Vairo, G. A Numerical Model for Wind Loads Simulation on Long-Span Bridges. Simul. Model. Pract. Theory 2003, 11, 315–351.
[CrossRef]
23. Huang, L.; Liao, H.; Wang, B.; Li, Y. Numerical Simulation for Aerodynamic Derivatives of Bridge Deck. Simul. Model. Pract.
Theory 2009, 17, 719–729. [CrossRef]
24. Brusiani, F.; de Miranda, S.; Patruno, L.; Ubertini, F.; Vaona, P. On the Evaluation of Bridge Deck Flutter Derivatives Using RANS
Turbulence Models. J. Wind Eng. Ind. Aerodyn. 2013, 119, 39–47. [CrossRef]
25. Keerthana, M.; Harikrishna, P. Numerical Investigations on the Effect of Mean Incident Wind on Flutter Onset of Bridge Deck
Sections. Struct. Eng. Mech. 2022, 82, 517–542.
26. Montoya, M.C.; Nieto, F.; Hernández, S.; Kusano, I.; Álvarez, A.J.; Jurado, J.Á. CFD-Based Aeroelastic Characterization of
Streamlined Bridge Deck Cross-Sections Subject to Shape Modifications Using Surrogate Models. J. Wind Eng. Ind. Aerodyn. 2018,
177, 405–428. [CrossRef]
27. Li, J.; Quan, Z.; Zhang, Y.; Cao, L.; Li, C. Computational Fluid Dynamics Based Kriging Prediction on Flutter Derivatives of Flat
Steel Box Girders. Symmetry 2022, 14, 1304. [CrossRef]
28. Zhuo, L.; Liao, H.; Zhou, Q.; Li, M. Identification of Aerodynamic Derivatives of a Box-Girder Bridge Deck with Twin Active
Flaps Using CFD Simulations. J. Bridge Eng. 2022, 27, 04022002. [CrossRef]
29. Ding, Q.; Chen, A.; Xiang, H. A State Space Method for Coupled Flutter Analysis of Long-Span Bridges. Struct. Eng. Mech. 2002,
14, 491–504. [CrossRef]
30. Namini, A.H. Analytical Modeling of Flutter Derivatives as Finite Elements. Comput. Struct. 1991, 41, 1055–1064. [CrossRef]
31. Hua, X.G.; Chen, Z.Q. Full-Order and Multimode Flutter Analysis Using ANSYS. Finite Elem. Anal. Des. 2008, 44, 537–551.
[CrossRef]
32. Li, Y.; Tang, H.; Wu, B.; Zhang, J. Flutter Performance Optimization of Steel Truss Girder with Double-Decks by Wind Tunnel
Tests. Adv. Struct. Eng. 2018, 21, 906–917. [CrossRef]
33. Scanlan, R.H. Motion-Related Body-Force Functions in Two-Dimensional Low-Speed Flow. J. Fluids Struct. 2000, 14, 49–63.
[CrossRef]
34. Larsen, A. Advances in Aeroelastic Analyses of Suspension and Cable-Stayed Bridges. J. Wind Eng. Ind. Aerodyn. 1998, 74–76,
73–90. [CrossRef]
35. Katsuchi, H.; Jones, N.P.; Scanlan, R.H. Multimode Coupled Flutter and Buffeting Analysis of the Akashi-Kaikyo Bridge. J. Struct.
Eng. 1999, 125, 60–70. [CrossRef]
36. Scanlan, R.H. Reexamination of Sectional Aerodynamic Force Functions for Bridges. J. Wind Eng. Ind. Aerodyn. 2001, 89,
1257–1266. [CrossRef]
37. Tang, H.; Li, Y.; Wang, Y.; Tao, Q. Aerodynamic Optimization for Flutter Performance of Steel Truss Stiffening Girder at Large
Angles of Attack. J. Wind Eng. Ind. Aerodyn. 2017, 168, 260–270. [CrossRef]
38. Ammann, O.H. George Washington Bridge: General Conception and Development of Design. Trans. Am. Soc. Civ. Eng. 1933, 97,
1–65. [CrossRef]
39. Dana, A.; Aksel, A.; Rapp, G.M. George Washington Bridge: Design of Superstructure. Trans. Am. Soc. Civ. Eng. 1933, 97, 97–163.
[CrossRef]
40. Dicker, D. Aeroelastic Stability of Unstiffened Suspension Bridges. J. Eng. Mech. Div. 1971, 97, 1677–1701. [CrossRef]
41. Frandsen, A. Wind Stability of Suspension Bridges. Application of the Theory of Thin Airfoils. In Proceedings of the International
Symposium on Suspension Bridges, Lisbon, Portugal, November 1966; pp. 609–627. Available online: https://trid.trb.org/view/
105424 (accessed on 1 May 2023).
42. Bartoli, G.; Mannini, C. A Simplified Approach to Bridge Deck Flutter. J. Wind Eng. Ind. Aerodyn. 2008, 96, 229–256. [CrossRef]
43. Matsumoto, M.; Matsumiya, H.; Fujiwara, S.; Ito, Y. New Consideration on Flutter Properties Based on Step-by-Step Analysis. J.
Wind Eng. Ind. Aerodyn. 2010, 98, 429–437. [CrossRef]
44. Van der Put, T.A.C.M. Rigidity of Structures against Aerodynamic Forces; IABSE: Zurich, Switzerland, 1976.
Appl. Sci. 2023, 13, 6389 21 of 21

45. Fu, C.; Wang, S. Computational Analysis and Design of Bridge Structures; Taylor & Francis Group: Abingdon, UK; CRC Press: Boca
Raton, FL, USA, 2017.
46. Vu, T.-V.; Kim, Y.-M.; Han, T.-S.; Lee, H.-E. Simplified Formulations for Flutter Instability Analysis of Bridge Deck. Wind Struct.
Int. J. 2011, 14, 359–381. [CrossRef]
47. Minimum Design Loads and Associated Criteria for Buildings and Other Structures; American Society of Civil Engineers: Reston, VA,
USA, 2021; ISBN 9780784415788.
48. Menter, F.R.; Kuntz, M.; Langtry, R. Ten Years of Industrial Experience with the SST Turbulence Model. In Proceedings of the 4th
International Symposium on Turbulence, Heat and Mass Transfer, Antalya, Turkey, 12–17 October 2003.
49. Wilcox, D.C. Turbulence Modeling for Cfd, 3rd ed.; D C W Industries: La Canada, CA, USA, 1998.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.