GLOBAL STABILITY OF CABLE-STAYED DECKS

Roberto Ribeiro de Mendonça Feijóo

Department of Civil Engineering, Architecture and Mining, Instituto Superior Técnico,

Technical University of Lisbon – PORTUGAL
October 2011

Keywords: ABSTRACT
The elastic global stability analysis of cable-stayed decks is considered in this re-
Elastic stability search. The bending instability in this type of elements is studied, as well as the
influence of some specific aspects in the design of the deck, towers and stays ar-
Cable-stayed rangement. The linear and non-linear analysis models considered are presented,
bridges and the nonlinearities associated to cable-stay bridges included in this research are
defined. The linear elastic stability of the deck is evaluated based on a model, using
Linear analysis the analogy of a beam on an elastic foundation (BEF). A simplified approach still
using the BEF is developed. To compare the results, geometrical non-linear elastic
Non-linear
analysis made by the finite elements software SAP2000 and ANSYS are used. The
analysis
bending stability analysis of a 420 m main span length deck is performed. The in-
fluence of some design aspects in the global stability of the bridge are evaluated by
a parametric study that considers: The deck live load pattern; the stays arrange-
ment; the towers height and geometry; the stiffness of the deck; the connection be-
tween deck and towers; and the intermediate piers on the lateral spans.

1 INTRODUCTION decks. In the cases of slender decks, if the com-

pressive force is big enough, it may cause global
Cable-stayed structures pose as an efficient instability.
and visually appealing solution for bridges of var- In this research, the evaluation of cable-stayed
ied span length. In the last 50 years these types of decks is performed. As is standard in this type of
bridges have enjoyed a great development thanks investigation, the physical nonlinearity of materi-
to the progress of construction techniques and als is not accounted for, therefore it is assumed
analysis models. that the behaviour of steel and concrete remains
The global stability of cable-stayed bridges elastic for large load factors. This way, the time
decks has not received much attention from either dependent effects of concrete, just like the physi-
designers or researchers, because few studies cal nonlinearity of the connection between steel
have shown large load factors against it [8,9]. and concrete for composite decks, are not consid-
However, the susceptibility to buckling of the ered in this study. The geometrically nonlinear
deck on a cable-stayed bridge has increased, be- effects related to this type of structure are the on-
cause of the use of bigger span lengths and re- ly relevant ones for this type of analysis.
duced deck depth. Accounting for what was previously men-
The slanted stays drive vertical forces to the tioned, the following objectives are set:
bridge towers that lead to compression on the

1
 R.Feijóo / Global Stability of Cable-Stayed Decks (2011)

• Conception of a model to evaluate the

elastic stability of a cable-stayed deck
based on the analogy of a beam on an
elastic foundation.

• Development of parametric studies to as-

sess the influence of several conception
traits on the global stability of cable-
stayed decks. Traits such as: the cable-
stay arrangement; the height and geometry
of the towers; the distance between stays;
the influence of the load pattern; the con- Fig. 1 Vasco da Gama cable-stayed bridge, in Lisbon
nection between deck and towers; and the
intermediate piers on the lateral spans. two intermediate piers. The model used in this
research is a plain model and simulates half of the
2 CASE OF STUDY deck (Fig. 2). The main span is 420 m long and
the two lateral spans are 194.7 m long with two
The superstructure of a cable-stayed bridge is intermediate piers which divide them in three
composed of towers/piers, stays and a deck. smaller spans (one measuring 60.125 m and the
There can be many different types of arrange- other two 72.1875 m). There are two towers, 150
ments for each of these structural elements. This m long each (100 m above the deck), and four
makes it so this type of bridge can be used in planes of stays arranged in a semi-fan layout of
small, medium and large spans. In this research 4x16 cables. The deck is supported at the inter-
the most common configuration was taken into mediate and end piers but not at the towers. The
account - a configuration similar the one of the distance between stays, at deck level, is 13.125
Vasco da Gama bridge (Fig. 1) with three spans, m. The deck is simulated by a bar element pos-
towers in the shape of an H, and a twin girder sessing the same stiffness as half the deck shown
deck with side suspension. The stays are arranged in Fig. 3.
in a semi-fan layout and the lateral spans have

Overall length = 829 m

stay-cable

Cross-girder

Fig. 2 420 m main span model used and deck cross-section considered [3]

2
 R.Feijóo / Global Stability of Cable-Stayed Decks (2011)

Four more models were used in this research foundation, and the energy of the compressive
with the following main span lengths: 577.5 m; forces, allows, by solving an eigenvalues prob-
735 m; 892.5 m; and 1050 m. They were all lem, to determine the instability modes and the
based on the initial model (with 420 m main span respective critical loads [10]. However, Klein [2]
length), differing from it in that they include proposed a simplified method to estimate the first
higher towers and a greater number of stays, alt- buckling load factor based on the existence of a
hough the distance between stays remains the section along the main span, where the ratio be-
same at deck and tower level. Each new stay has tween the vertical elastic restraint due to the ca-
1.5 cm2 more than the last one. bles and the compressive load introduced by them
In the elastic stability analysis the only nec- is a minimum (Fig. 4). This critical section de-
essary parameter to define is Young’s modulus, fines the buckling mode of the “actual” BEF, in a
since it is assumed an elastic behavior to both way that the same behavior is attained by an
concrete and steel. The value of Young’s modu- “equivalent” BEF with constant elastic stiffness
lus for steel is Ea = 210 GPa and for the stay ca- and axial force of the critical section.
ble’s steel Ee = 195 GPa. The concrete’s Young’s The buckling axial load Ni,cr of the equivalent
modulus is Eco = 36 GPa, which refers to a BEF is given with sufficient accuracy, for a high
C45/55 concrete. number of half-waves of the buckling modes, by
Both the live and the permanent loads are de- the Engesser formula (eq. 4). The buckling verti-
fined as uniform throughout the deck. The first’s cal load (qcr) of the actual BEF is Ni,cr/Ni (eq. 5)
value is 171 kN/m and the latter’s is 54kN/m [3]. where Ni is the compressive load of the critical
Both loads are relative to half of the deck. section due to the vertical load q. Using the same
model, the number n of half waves of the buck-
3 LINEAR STABILITY OF CABLE- ling mode, and the equivalent buckling length Lcr
STAYED DECKS: BEAM-COLUMN ON are given by eq. 6 and eq. 7, correspondingly.
Table 1 presents the results for both method-
AN ELASTIC FOUNDATION ANALO- ologies when determining the critical load for all
GY decks’ lengths considered in this work, when the
decks are entirely loaded.
The buckling analysis of a cable-stayed deck
can be done based on the buckling model of a Table 1 BEF model and Klein’s simplified method results
beam on an elastic foundation (BEF). The main when the deck is entirely loaded
span of a cable-stayed bridge can be considered
as a simple supported beam with bending stiff- Main span
420 577.5 735 892.5 1050
ness EI and elastically supported along the span length (m)
by the cables, with variably vertical stiffness Kv, Klein simplified method
given by eq. 1, increasing towards the towers. For No,cr [kN] 403442 384159 336412 312865 286908
vertical loads q this beam is subjected to increas- BEF model
ing compressive forces introduced by the cables No,cr [kN] 407688 382575 343702 313647 289258
(eq. 2, Fig. 3). Since the cables are closely spaced Deviation (%)
(a = 13.125 m) a continuous vertical restraint βi 0.99 0.41 2.17 0.25 0.82
can be envisaged for the beam on elastic founda-
tion (eq. 3). The cables stiffness and axial forces
distribution are functions of the staying configu-
ration (Fig. 3).
An energy method for buckling analysis,
based on the deformation energy of the beam and

3
 R.Feijóo / Global Stability of Cable-Stayed Decks (2011)

Ni [kN] 𝐸𝐸𝑒𝑒 𝐴𝐴𝑖𝑖 𝑁𝑁𝑖𝑖,𝑐𝑐𝑐𝑐 2�𝐸𝐸𝐸𝐸𝛽𝛽𝑖𝑖

Ni 𝐾𝐾𝑣𝑣,𝑖𝑖 = 𝑠𝑠𝑠𝑠𝑠𝑠2 𝛼𝛼𝑖𝑖 (1) 𝑞𝑞𝑐𝑐𝑐𝑐 = = 𝑎𝑎 (5)
500 𝑙𝑙𝑖𝑖 𝑁𝑁𝑖𝑖 ∑𝑛𝑛º 𝑡𝑡𝑡𝑡𝑡𝑡
𝑗𝑗 =𝑖𝑖 𝑡𝑡𝑡𝑡𝑡𝑡𝛼𝛼
stay i, EeAi 𝑗𝑗
450 Harp configuration
450 Li , α i
400 𝑛𝑛º 𝑡𝑡𝑡𝑡𝑡𝑡
400 Semi-fan configuration 𝑞𝑞 𝑎𝑎 𝛽𝛽𝑖𝑖 𝐿𝐿4
350
350 𝑁𝑁𝑖𝑖 = � (2) 𝑛𝑛4 ≈ (6)
300 Fan configuration tan 𝛼𝛼𝑗𝑗 𝜋𝜋 4 𝐸𝐸𝐸𝐸
300 𝑗𝑗 =𝑖𝑖
250
250
200 𝐾𝐾𝑣𝑣,𝑖𝑖 𝜋𝜋 2 𝐸𝐸𝐸𝐸 𝐿𝐿
200
150 𝛽𝛽(𝑥𝑥) = (3) 𝐿𝐿𝑐𝑐𝑐𝑐 = � ≈ (7)
150 𝑎𝑎 𝑁𝑁𝑖𝑖,𝑐𝑐𝑐𝑐 𝑛𝑛√2
100
100 q = vertical unitary load
50
50
0 𝑁𝑁𝑖𝑖,𝑐𝑐𝑐𝑐 = 2�𝐸𝐸𝐸𝐸𝛽𝛽𝑖𝑖 (4)
0
cable i =11 1 2 33 3 4 5
55 6 7
77 8 9
99 10 11 12
11
11
13
13
13 14
15
15
15 16
0
N0,cr EI, L
Ni,cr N(x) N0
5000

harp configuration
0,1β0 β(x)
10000
β0 βi
Semi-fan configuration critical section
N(x)/β(x)
15000
Fan configuration
EI, L
20000 Ni,cr
Kv,1=45556
kN/m
25000

K v,i [kN/m] βi

Fig. 3 Axial force and cables vertical stiffness for a 420 m

Fig. 4 Equivalent beam on an elastic foundation model
long span deck with 100 m high towers

𝐿𝐿𝑖𝑖 𝑇𝑇 𝐿𝐿𝑖𝑖 𝑇𝑇
4 BEF MODEL MODIFICATION 𝛥𝛥1,𝑣𝑣 = (8) 𝛥𝛥2,ℎ = (10)
𝐸𝐸𝑖𝑖 𝐴𝐴𝑖𝑖 𝑠𝑠𝑠𝑠𝑠𝑠 𝛼𝛼𝑖𝑖 𝐸𝐸𝑖𝑖 𝐴𝐴𝑖𝑖 𝑐𝑐𝑐𝑐𝑐𝑐 𝛼𝛼𝑖𝑖

When an increasing uniform load is applied 𝛥𝛥2,ℎ 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝛼𝛼𝑖𝑖 𝐸𝐸𝑒𝑒 𝐴𝐴𝑖𝑖

𝛥𝛥2,𝑣𝑣 = (9) 𝐾𝐾𝑣𝑣 = = 𝑠𝑠𝑠𝑠𝑠𝑠2 𝛼𝛼𝑖𝑖 (11)
𝑡𝑡𝑔𝑔 𝛼𝛼𝑖𝑖 𝛥𝛥1,𝑣𝑣 + 𝛥𝛥2,𝑣𝑣 2𝐿𝐿𝑖𝑖
all over the deck, the analysis of the main span is
accurate enough using just the BEF model. How-
𝛽𝛽 2�𝐸𝐸𝐸𝐸𝐸𝐸 𝑁𝑁𝑐𝑐𝑐𝑐 𝑞𝑞𝑐𝑐𝑐𝑐
ever, if only the main span is fully or partially 𝑁𝑁𝑐𝑐𝑐𝑐 ′ = 2�𝐸𝐸𝐸𝐸𝛽𝛽 ′ = 2�𝐸𝐸𝐸𝐸 = = → 𝑞𝑞 ′ 𝑐𝑐𝑐𝑐 = (12)
2 √2 √2 √2
loaded, a large influence of tower and lateral span
displacements occur, and the BEF’s results are
not accurate. To be able to use this model when lateral stay i
Δ2,h
central stay i
T T
only the main span is loaded, the towers dis-
, Li Ei A
placements must be considered in the definition E iA i i, Li
of the BEF model. These displacements decrease
the foundation’s stiffness value to half of it. αi
T
T sen(αi)
The following hypotheses are considered in Δ1,v
this modification:
Fig. 5 Effect of the tower displacements on the foundation’s
• the configuration of the stays is stiffness of the BEF model
symmetric;
• the tower’s bending stiffness is not • supports on all lateral stays are con-
considered, since it is very inferior in sidered at deck level.
comparison with the stay’s stiffness; Equations (8) and (10) show the vertical dis-
• the deformability control of the tow- placement on the deck (Δ1,v), and the horizontal
er is done by the lateral stays; displacement on the tower (Δ1,h) induced by a pair
of stays, when only the main span is loaded, as

4
 R.Feijóo / Global Stability of Cable-Stayed Decks (2011)

illustrated in Fig.5. The hor-

izontal displacement on the 1200
qcr[kN/m]
1200
1200

tower is related to a vertical qcr

1000
1000 1000
displacement on the deck
(Δ2,v) as shown in eq. 9 when main span Load scenario b)
800
800 800
taking into account only
small displacements. Know-
600
600 600
ing the total displacement of
the deck, the stiffness as- 400
400 400
signed by a pair of stays
(lateral and central stay) is Results [3]:Load scenario b)
200
200 200
given by eq. 11, which Klein’s simp. method: Load scenario b)

shows that the foundation’s BEF model: Load scenario b)

00 0

stiffness is now half of what 400

400
400
400
500
500
500
500
600
600
600
600
600 700
700
700
700
700
800
800
800
800
900
900
900
900
1000
1000
1000
1000
1100
1100
1100
1100
it used to be when the deck Main span [m]
was entirely loaded (eq. 1).
Fig. 6 Buckling loads for load scenario b) when considering the modifications on
The critical load ob- both BEF model and Klein’s simplified method
tained with Klein’s simpli-
fied method can also be modified to consider the sible to conclude that the inclusion of the con-
effect previously explained. The relation between structive process in the stability analysis does not
the critical load of an entire loaded deck (qcr), and alter the results in any significant way.
the one of a deck that is only loaded at the main 5.2 Load pattern influence
span (q’cr) is √2 (eq. 12).
The buckling of the deck is dependent on the
Fig. 6 shows that the results considering this load vector [8; 3]. Three load patterns were con-
modification, for both BEF model and Klein’s sidered using the nonlinear SAP2000 model, and
simplified method, agree with those obtained by both BEF and Klein’s simplified method (Fig 7).
Pedro [3] with a nonlinear analysis. The results show that when the bridge’s deck is
loaded at the main span, the buckling load is re-
5 PARAMETRIC STUDY
duced. When only half of the main span is loaded,
5.1 Constructive process influence the critical load decreases to half the value of the
load scenario a).
The inclusion of the constructive process on
a nonlinear stability analysis induces an initial Table 2 Constructive process influence on the stability
compressive force in the deck which may affect analysis of the deck
its critical load. The SAP2000 software simulates
a cantilever constructive method in simple fash- Nmáx,inicial Ncr qcr
Analysis Type
ion. The insertion of an initial compressive force [kN] [kN] [kN/m]
in the deck is due to the pulling force on the stays With Constr.
47905 346186 1445
during the constructive process, which allows for Process
control of the span deformation. Table 2 shows Without Constr.
42904 329229 1442
the results for a nonlinear analysis of a fully load- Process
ed deck of the 420 m main span model. It’s pos-

5
 R.Feijóo / Global Stability of Cable-Stayed Decks (2011)

To be able to use the BEF

1600
1600
model in load scenario c), an- qcr [kN/m]
qcr
other hypothesis has to be ad- 1400
1400

main span Load scenario a)

mitted: The foundation’s stiff- 1200
1200

ness and compressive force dis- qcr

1000
1000
tribution are regarded as the main span Load scenario b)

same as in load scenario a) and 800

800

qcr
b) in this model. The maximum 600
600
main span Load scenario c)
vertical stiffness is provided by
400
400
the stay closer to the load pat-
tern border (Fig.8). The BEF 200
200 BEF model:
Coluna
Coluna s/ Fund.Load
s/ Fund. scenario
Elástica
Elástica a)
-- Carreg.1
Carreg.1 Hip.
Hip. Kleinsimp.
Klein’s
Klein -- Carreg.1
Carreg.1
meth.: Load scenario a)
BEF model:
Coluna
Coluna s/ Fund.Load
s/ Fund. scenario
Elástica
Elástica b)
-- Carreg.2
Carreg.2 Klein’s
Hip.
Hip. Kleinsimp.
Klein meth.: Load scenario b)
-- Carreg.2
Carreg.2
model considering this hypothe- 00
BEF model:
Coluna
Coluna s/ Fund.Load
s/ Fund. scenario
Elástica
Elástica c)
-- Carreg.3
Carreg.3 SAP2000:
SAP2000
SAP2000 Load scenario a)
-- Carreg.1
Carreg.1

400 500 600 700 800 900 1000 1100

sis and the foundation’s stiff- 400 500 600 700 800 900 1000
Main span [m]
1100

ness modification explained

earlier lead to results that agree Fig. 7 Buckling loads for different load patterns
with those obtained by Pedro
[3] using a nonlinear analysis Load scenario c)
for load scenario c). N0

5.3 Stays arrangement in-

β0 Compressive force and foundation’s stiffness
fluence main span configuration considered:
Load pattern N0
border
Different stay’s arrange-
ments change the distribution of
the compressive force and βo
a main span
foundation’s stiffness, as shown K
Kv,i →βo = av,i
in Fig.3. A nonlinear analysis
Fig. 8 Hypothesis considered in the BEF model for load scenario c)
using SAP2000 for three mod-
els, each one with a different
2000
stay configuration (fan, semi- qcr [kN/m] Load scenario b) Load
qcr
h = tower
height

1800 scenario a)
fan and harp configurations),
shows that when a bridge has a 1600
L = main span length

fan arrangement, the critical 1400 BEF

torre:model:5%h=
h = 25% x L25% x L;1Load scenario a)
; carreg.
BEF
torre:model:0%h=
h = 20% x L20% x L;1Load scenario a)
load of the deck is higher than 1200 BEF
torre:model:5%h=
h = 15%
; carreg.
x L15% x L;1Load scenario a)
; carreg.

in any other case (50% higher 1000

BEF
torre:model:
h = 25%h=
x L 25% x L;2 Load scenario b)
; carreg.
BEF
torre:model:
h = 20%h=
x L 20% x L;2 Load scenario b)
; carreg.
that the semi-fan case). On the 800
torre: h = 15% x L ; carreg. 2
BEF model: h= 15% x L; Load scenario b)

other side the harp arrangement

600
gives the lowest critical load
400
(almost 40% lower that the
200
semi-fan case). In fact, the fan
arrangement is the one that pro- 0
400 500 600 700 800 900 1000 1100
vides the highest “vertical sup- Main span [m]

port” throughout most of the

deck’s length, while the harp Fig. 9 Buckling loads for different tower heights
arrangement provides the low-
est.

6
 R.Feijóo / Global Stability of Cable-Stayed Decks (2011)

5.4 Towers height and stays

1600
qcr [kN/m] Load scenario b) Load
qcr scenario a)
spacing influence 1400

The height of each of the a Main span

1200
bridge’s towers has direct im- BEF model - a = 13.125 m: Load scenario a)
1000 BEF model - a = 13.125 m: Load scenario b)
pact on the critical load of the
BEF model - a = 26.25 m: Load scenario a)
deck. As the towers become 800 BEF model - a = 26.25 m: Load scenario b)
higher, so does the critical load
600
of the deck, since the stays be-
come more slanted and induce 400

greater “vertical support” to the 200

deck (similarly to the fan ar-
0
rangement). As tower height 400 500 600 700 800 900 1000 1100

becomes lower, with less slant- Main span [m]

ed stays, the “vertical support” Fig. 10 Buckling loads for different stays spacing
becomes smaller and the com- qcr [kN/m]
2800
pressive forces on the deck in- qcr stay-cable
@ 13.125 m
crease in value (similarly to the 2400
main span Load scenario a)
0.25 m
harp arrangement).

Deck height
Doubling the spacing be-

Girder height
2000
tween stays does not affect the cross-girders
@ 4.375 m
stability of the deck (Fig. 10),
1600
therefore, replacing a stay dur- BEF model h = 2.25m : Load scenario a)
BEF model h = 2.75m : Load scenario a)
ing the lifetime of the bridge BEF model h = 3.25m : Load scenario a)
1200
will not risk its stability [10]. BEF model = 3.75m : Load scenario a)

5.5 Bending stiffness of the 800

deck influence
400
400 500 600 700 800 900 1000 1100
The bending stiffness of the Main span [m]
deck affects the critical load as Fig. 11 Buckling loads for different girder heights
shown in eq. 5 This equation
and the results shown in Fig 11 make it possible Table 3 Bending stiffness of the deck considered
to conclude that the greater the bending stiffness
of the deck is, the greater it’s critical load is as girder height (m) EI [kN/m2]
well. This increase of the bending stiffness of the
2.25 6.487E+07
deck is obtained taking into account higher gird-
2.75 9.901E+07
ers (Table 3).
3.25 1.417E+08
5.6 Towers geometry influence 3.75 1.936E+08
The geometry of the tower’s cross section the analysis’ results are similar for all types of
used in the initial model has the shape of an H. geometry [7].
The use of a nonlinear analysis, performed with However some authors [8] refer that towers
SAP2000, on tridimensional models with A and with A or inverted Y shapes provide an improved
inverted Y shapes leads to the conclusion that the global stability to the structure.
geometry of the tower’s cross section does not
seem to affect the critical load of the deck, since

7
 R.Feijóo / Global Stability of Cable-Stayed Decks (2011)

5.7 Towers bending stiffness

2500
influence qcr[kN/m]
qcr

Some types of cable-stay 2000 EI

main span
EI
Load scenario a)
3 x EI

bridges, like the multiple span 2 x EI

cable-stay bridges, require tow- 1500

ers with greater longitudinal EI

EI
bending stiffness. A nonlinear 1000 EI
2

4
analysis performed with EI deck instability
8
SAP2000 on the 420 m model 500
shows that a higher value on the EI
16
tower instability
tower’s stiffness leads to the 0
increase on the critical load of 0 2000 4000 6000 8000 10000 12000 14000 16000 18000
the deck (Fig. 12). However, in EItower [kN/m2]

case the towers are too flexible

(with about an eight times Fig. 12 Buckling loads for different bending stiffness of the tower
smaller EI), they lose their sta-
bility first. 1600
qcr [kN/m]
qcr Load scenario b)
Load
5.8 Connection between 1400 scenario a)

deck and towers influ- 1200 Main span

SAP2000- c/
SAP2000 – 2with piers:
pilares BEF- model
BEF – with piers:
c/ 2 pilares
1000 intermédios - carreg.1 intermédios
ence load scenario a) load scenario2- carreg.1
b)
SAP2000- sem
SAP2000 – nopilares
piers: Results [3][3]– -no
Resultados sempiers:
pilares
800 load scenario
intermédios a)
- carreg.1 load scenario
intermédios b)
- carreg.2
On the initial model, the
600
deck is only supported by the
stays and intermediate and final 400

piers. The inclusion of a simple

200
support between the deck and
the towers is analyzed through a 0
400 500 600 700 800 900 1000 1100
nonlinear analysis performed Main span [m]

with SAP2000. The results

show very similar critical loads Fig. 13 Buckling loads when intermediate piers on the lateral spans are consid-
ered or not
in both cases, with or without
the supports’ inclusion. Conse-
quently, the support of the deck on the towers However, for load scenario b, and according to
does not affect its stability. Pedro [3,5], when no piers are considered, there
is a decrease in the critical load of nearly 50%.
5.9 Intermediate piers influence For longer spans, the impact of intermediate piers
Under service conditions, the intermediate on buckling loads is not so important.
piers at the lateral spans of cable-stayed bridges
6 Conclusions
improve the deck’s stiffness and stability [4]. In
order to examine the impact of those piers on the The beam on elastic foundation (BEF) mod-
deck’s overall stability, the two intermediate piers el, and the Klein’s simplified method provided
of the case study (Fig. 2) were fully eliminated. similar results to a nonlinear analysis. Therefore,
Fig. 13 shows the result of a nonlinear analysis these two types of methodologies are capable of
performed with SAP2000.
analyze the global stability of the deck. The mod-
For load scenario a) this change does not
seems to affect the critical load value of the deck. ification on both models taking in account the

8
 R.Feijóo / Global Stability of Cable-Stayed Decks (2011)

effect of the towers displacement, on the founda- 5 Pedro, J. J. O.; Reis, A. − « Stability of
tion’s stiffness of the deck, enabled the analysis Composite Cable-Stayed Bridges » − EU-
when only the main span was loaded. For a load ROSTEEL 2011. Hungary, August 2011.
scenario when only half of the deck’s main span
6 Reis, A.; Camotim, D. − « Estabilidade
was loaded, another modification, considering a
Estrutural » . Editora McGraw-Hill de
simplified distribution of the compressive force Portugal, 2001.
and foundation’s stiffness, provided results in
agreement with other researches. 7 Shu, H-S.; Wang, Y-C. − « Stability Analysis
With the parametric study presented, it was of Box-Girder Cable-Stayed Bridges » −
possible to identify the influence of: 1) the deck ASCE - Journal of Bridge Engineering, Jan-
load pattern, 2) the stays arrangement, 3) the tow- uary 2001.
ers height an stays spacing, 4) the stiffness of the
8 Taylor, P. R. − « What are the limiting crite-
deck, 5) the towers geometry and bending stiff- ria governing the maximum span of compo-
ness, 6) the connection between deck and tower site cable-stayed bridges? » − Composite
and 7) the intermediate piers. It is shown that the Bridges – State of the Art in Technology and
deck stability is sensitive to some of these param- Analysis. Proceeding of the 3rd International
eters, but in any case modern long-span cable- Meeting. Madrid, January 2001.
stayed bridges are more likely to be governed by
strength than by deck instability [3]. These con- 9 Taylor, P.; Kaneko, S.; Bergman, D. – «
Buckling stability and secondary stress ef-
clusions are in agreement with those by Taylor
fects in the deck girders of cable-stayed
[8] and Pedro [3]. bridges » – International Conference AIPC-
FIP, Proceedings - Vol. 2, pg. 223-230.
REFERENCES Deauville, October 1994.

1 Adeli, Y.; Zhang, J. − « Fully Nonlinear 10 Timoshenko, S.; Gere, J. − « Theory of elas-
Analysis of Composite Girder: Cable-Stayed tic stability » − pg. 94-113, 2nd edition,
Bridges » − Computers & Structures – Vol. McGraw-Hill Book Company, 1961.
54, Nº2, pg. 267-277. Elsevier Science, 1995.
11 Timoshenko, S. − « Resistência dos
2 Klein, J-F. − « Ponts Haubanés: Comporte- Materiais » − pg. 326-328, 1ª edição, Ao
ment et Stabilité des Tablier Minces » - Livro Técnico S.A.. Rio de Janeiro, 1969.
Ecole Polytechnique Fédéral de Lausanne,
These Nº 833. Lausanne, Avril 1990. 12 Wang, Y-C. − « Number of Cable Effects on
Buckling Analysis of Cable-Stayed Bridges »
3 Pedro, J. J. O. − « Pontes Atirantadas Mistas: − ASCE - Journal of Bridge Engineering,
Estudo do Comportamento Estrutural » − November 1999.
Tese de Doutoramento, Instituto Superior
Técnico. Lisboa, Julho 2007.

4 Pedro, J. J. O. − « Pontes de Tirantes –

Concepção, Dimensionamento e Construção
» − Instituto Superior Técnico. Lisboa, 2010.