Engineering Structures 32 (2010) 3288–3300

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Parametric study on the dynamic response of cable stayed bridges to the sudden
failure of a stay, Part I: Bending moment acting on the deck
C.M. Mozos a,∗ , A.C. Aparicio b
a
Department of Civil Engineering and Building, UCLM University of Castilla-La Mancha, Av. Camilo Jose Cela s/n 13071, Ciudad Real, Spain
b
Department of Construction Engineering, UPC Technical University of Catalonia, C. Jordi Girona 1-3, 08034, Barcelona, Spain

article info abstract

Article history: The existing recommendations on cable stayed bridges propose the use of a static analysis with a dynamic
Received 15 December 2009 amplification factor of 2.0 for evaluating the response of cable stayed bridges to the sudden loss of a cable.
Received in revised form The application of this simplified procedure has been found to be critical and to control the design of some
25 May 2010
recently erected bridges. The present paper deals with the dynamic response of cable stayed bridges to
Accepted 1 July 2010
Available online 8 August 2010
the sudden loss of a stay. Its objectives are to quantify the relative importance of the accidental ultimate
limit state of failure of a stay in the design of the bridge, and to determine the safety level provided by the
Keywords:
simplified procedure of using a static analysis with a D.A.F. of 2.0. For this purpose, a parametric study has
Cable stayed bridges been carried out. In this study, ten cable stayed bridges have been analyzed and the effect of characteristics
Dynamic amplification factor such as the layout of the stays, either fan or harp pattern, the number of planes of stays and the stiffness
Loss of a cable of the deck have been studied. First, this paper describes the geometry and materials of the cable stayed
bridges studied, the numerical models and the basis of the static and dynamic analysis developed. Next,
the results related to the deck cross sections are presented and discussed. Finally, some conclusions
related to the design and analysis of the cable stayed bridges are provided. A second companion paper
(i.e. Part II) is focused on the response of the pylons and the stays.
© 2010 Elsevier Ltd. All rights reserved.

1. Introduction
Stays in a cable stayed bridge, see [1–5], are exposed to corro-
sion, abrasion and fatigue processes which may cause a reduction
in their section and a decrease in their resistance capacity [6,7].
This decrease in their capacity to withstand the forces transferred
by the deck and the pylon can cause the fracture of the cable, as it
happened in the Zárate-Brazo Largo Bridge (Argentina) in 1996 [8].
Therefore, the accidental event of the sudden failure of a stay has
been incorporated into some design codes and recommendations.
Up until now, the codes and recommendations on the topic
tackle this accidental situation by means of an equivalent static
analysis in which a dynamic amplification factor (D.A.F.) is used.
The D.A.F. is defined as the dimensionless ratio of the maximum
dynamic response to the static response of a structure when it is
subjected to dynamic loading. The dynamic response of a cable
stayed bridge due to a sudden failure of a stay was studied in [9].
The study was focused on the Saame bridge, on the border of
Finland and Norway, and it concluded that the maximum D.A.F. is
equal to 1.8 for critical sections of the bridge. In the same direction,
∗
Corresponding author. Tel.: +34 926295300; fax: +34 926295391.
E-mail addresses: carlosmanuel.mozos@uclm.es (C.M. Mozos),
angel.carlos.aparicio@upc.edu (A.C. Aparicio).
0141-0296/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2010.07.003
the recommendations edited by the P.T.I., [10], indicate that, in the
equivalent static analysis, the dynamic effect of the loss of a stay
will be equal to the tension force that the stay bears multiplied
by an amplification factor equal to 2.0. The French SETRA has
also published a set of recommendations, [11], for the design of
cable stayed bridges in which the same method was suggested, but
placing the amplification factor between 1.5 and 2.0.
On the other hand, the dynamic response of a single degree of
freedom (SDOF) system to a rectangular pulse load has been stud-
ied in many classical references such as [12,13]. For these systems,
the maximum D.A.F. that can be reached under the action of a pulse
load is equal to 2.0, which makes it necessary for the ratio of the
pulse duration to the natural period of the system to be equal to
or larger than 0.5. The case of multiple degree of freedom (MDOF)
systems under a pulse load of infinite duration was studied in [14]
and it was concluded that, under certain conditions, the D.A.F. of
MDOF systems can reach values larger than 2.0. The relation be-
tween the maximum D.A.F. and the strain energy of the system
when the load is statically applied was also derived in [14]. Cable
stayed bridges are very redundant structures with many degrees
of freedom and their dynamic behavior is very different from the
behavior of a SDOF system. Hence, recent studies such as, [15–19],
focused on conventional cable stayed bridges, and [20] for the case
of the unconventional ones, have demonstrated that the simplified
approach suggested in [10,11] becomes unsafe in some cases.
 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300
Fig. 1. Longitudinal layout of the fan and harp pattern bridges and location of the cross sections studied. Those dimensions corresponding to bridges with the deck d0.40
are underlined. Dimensions in meters.
The responses of four different cable stayed bridges with a
length of 403 m, which were obtained combining two concrete
decks with a fan or a harp arrangement of the stays, to the sudden
failure of a stay were studied in [15]. Simplified 2D numerical mod-
els of the bridges were used for obtaining their dynamic responses
and self-weight and fifteen live load distributions were considered.
Results detailed in [15] show that the dynamic amplification fac-
tors for bending moments on the deck and bending moments on
the pylons can reach values much larger than 2.0 such as 4.18 and
10.61 for deck and pylons, respectively. In the case of the forces act-
ing on the stays, the dynamic amplification factors obtained were
around 2.0. Another general result summarized in [15] is that only
D.A.F. smaller than 2.0 were observed at cross sections which were
located close to the broken stay.
A five-span cable stayed bridge with a total length of 1140 m,
steel deck and concrete pylons was studied in [17–19]. The
influence of cable modeling and damping on the dynamic response
of the bridge to the loss of a stay was obtained from a 3D numerical
model of the bridge under permanent loads and two possible
distributions of live loads. In addition, the envelopes of the D.A.F.
were obtained from the responses of the undamped system to the
loss of each one of the stays under self-weight and without live
loads. The authors of these works concluded that a D.A.F. smaller
than 2.0 seems possible for bending moments in the stiffening
girder, a D.A.F. of 2.0 is necessary for the design of the cables and
significantly larger D.A.F. were observed for the bending moments
acting on the pylons. High values of D.A.F. were observed at
locations of the deck placed farther away from the ruptured cable,
however, the authors concluded that the dynamic responses at
locations farther away from the broken cable are irrelevant when
considering all cable loss load cases, because only responses in the
vicinity of the ruptured cable are design governing.
Thus, results attained in [15] and in [17–19] were obtained for
specific bridges and assuming some simplifications. Consequently,
and as is suggested in [19], it is necessary to investigate to
3289
what degree they are representative for cable stayed bridges in
general.
This paper is the first of two papers dealing with the response of
cable stayed bridges to the sudden rupture of one stay. The objec-
tives are to quantify the relative importance of the accidental ulti-
mate limit state of failure of a stay in the design of the bridge, and
to determine the safety level provided by the simplified approach
proposed in [10,11]. To this purpose, the dynamic and static behav-
ior of ten cable stayed bridges, which are obtained combining three
decks with four layouts of the stays, are numerically analyzed.
Detailed 3D numerical models of the bridges, which include the
corresponding live and variable loads, are developed for obtaining
the responses of the bridges. The live load distributions are deter-
mined from the influence lines corresponding to each cross section
and bridge studied. Results for the failure of the stay which is crit-
ical for the internal force and section studied are obtained in order
to quantify and compare the ultimate limit states of permanent and
transitory loads with the accidental limit states of failure of a stay.
The influence of damping, stiffness of the deck, number of planes
of stays and arrangement of the stays, fan or harp, on the response
of the bridge to the loss of a stay arises from the parametric study
allowing to determine which ones of these characteristics reduce
the effects of the loss of a cable and, consequently, avoid that the
accidental event governs the design of the bridge. In addition, the
error incurred when a D.A.F. equal to 2.0 is used for evaluating the
loss of a cable is expressed in terms of the envelope of ultimate
limit states of permanent and transitory loads in order to quantify
the relative importance of this error.
In Section 2, this paper describes the geometry of the bridges,
their finite element models and the basis of the numerical analysis
developed. The main results related to the response of the deck are
discussed in Section 3. Finally, Section 4 provides some of the con-
clusions related to the response of the deck. The responses related
to the pylons and stays are detailed in the second companion paper
(i.e. Part II), [21].
3290 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300

Fig. 2. Dimensions (meters) of the decks.

a b
Table 1
Decks. Cross section, A, second moment of area, Iy , and torsion constant, Jx .
Fig. 3. Pylon for bridges with two lateral planes of stays, (a), and for bridges with
Deck Depth (m) A (m2 ) Iy (m4 ) Jx (m4 ) a central plane of stays, (b). Dimensions in meters.

d0.40 0.40 5.20 0.069 0.277

d1.15 1.15 5.20 0.817 2.334
d2.25 2.25 5.20 3.710 9.755 Table 2
Characteristics of the ten bridges studied.
Bridge Stay Stay Distance between Deck’s depth
2. Description of the parametric study pattern planes stays

F 2P d0.40 Fan 2 6.20 0.40

2.1. Geometry of the cable stayed bridges F 1P d1.15 Fan 1 12.40 1.15
F 2P d1.15 Fan 2 12.40 1.15
F 1P d2.25 Fan 1 12.40 2.25
Ten cable stayed bridges are studied in the parametric study. F 2P d2.25 Fan 2 12.40 2.25
The geometry of these bridges is defined on the basis of the
H 2P d0.40 Harp 2 6.20 0.40
parametric study included in [1]. Each one of the bridges has a total H 1P d1.15 Harp 1 12.40 1.15
length of 403 m spaced out in one principal span and two lateral H 2P d1.15 Harp 2 12.40 1.15
spans, and it is defined in relation to the rest through the section of H 1P d2.25 Harp 1 12.40 2.25
its deck, through the layout of its stays (either in a fan or in a harp H 2P d2.25 Harp 2 12.40 2.25

pattern) and through the number of cable planes (one central plane
or two lateral planes). The length of the spans, the layout of stays
and the distance between stays are detailed for both fan and harp
patterns in Fig. 1. The dimensions of the three decks considered are
shown in Fig. 2 and their main mechanical properties are detailed
in Table 1. In addition to the webs and slabs, a diaphragm with
0.35 m of thickness is placed at every cross section where a stay
is connected to the deck.
An ‘H’ shaped pylon is considered for those bridges with two
planes of stays, Fig. 3(a), and an equivalent, in terms of stiffness,
an ‘I’ shaped pylon, Fig. 3(b), is considered for the bridges with
a central plane of stays. The deck is connected to both pylons by
an articulating joint with no possibility of relative displacement
between them. Table 2 summarizes the characteristics of the ten
bridges studied.
2.2. Materials
A 0.600 steel strand has been considered for the stays with an
ultimate tensile strength fu = 1860 N/mm2 , a modulus of elasticity
Es = 190 000 N/mm2 and a self-weight γ = 76.93 kN/m3 .
The mechanical properties of the concrete used for pylons and
decks are: fck = 40 N/mm2 , E0.28 = 36 342.4 N/mm2 and
γ = 24.5 kN/m3 .
2.3. Design of the stay cables
The cross section of the stays is established considering strength
and fatigue criteria. Therefore, 45% of fu is considered as the
maximum axial stress and 200 MPa as the stress range under every
load combination with characteristic values.
2.4. Loads
The loads are defined for highway bridges in [22]. Thus, the self-
weight and a uniformly distributed load of 3.750 kN/m2 are con-
sidered as permanent loads, G1 and G2 , respectively. The prestress-
ing force on stays, P2 , is incorporated into the model in order to
compensate 100% of the permanent loads. The traffic load, Q , com-
prises a uniformly distributed load of 4.0 kN/m2 and a three-axes
special vehicle of 600 kN. The temperature effect, Temp, is taken
into account with a uniform temperature increment of ±22 °C,
with a differential temperature change from the top to the soffit
of the deck of 8.5 °C and with a differential temperature between
the stays and the rest of the bridge of +18 and −10 °C.
Loads are combined following the criteria established in [22]
and considering the traffic load as the dominant load. Thus, per-
manent and transitory ultimate limit states, ULSp,t , are defined by
 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300
Table 3
Partial and combination factors.
Limit state Load effect Factor
γG1 γG2
Favorable 1.00 1.00
ULSp,t
Unfavorable 1.35 1.35
Favorable 1.00 1.00
AULS
Unfavorable 1.00 1.00
γG1 G1,k + γG2 G2,k + γP2 P2,k + γQ Qk
+ γTemp ψTemp Tempk ,
and the accidental ultimate limit states, AULS, are defined by
γG1 G1,k + γG2 G2,k + γP2 P2,k + γQ ψQ Qk
+ γTemp ψTemp Tempk + γA Ak ,
where the partial and combination factors are detailed in Table 3.
The selected design criteria and loads, according to [22], allow
us to determine the size and prestress load of the stay cables and
to define the accidental scenarios where the failure of a stay takes
place. These criteria and loads are close to those suggested in other
national or international codes. Thus, the qualitative analysis of the
obtained results is not affected by them.
2.5. Numerical models and analysis
Numerical models of the bridges are developed and studied by
means of linear static and dynamic analysis using the finite ele-
ment code SAP2000NL v. 8.0.8. The deck of the bridge is modeled
using shell elements while beam elements are considered for the
pylons and the stays. The nonlinear behavior of the stays is taken
into account with Ernst’s modulus of elasticity. Newmark’s step-
by-step method is used for obtaining the undamped and damped
dynamic responses to the sudden rupture of the critical stay.
The rupture time of a seven wire steel strand of 0.6000 was
determined in [16] after an experimental program concluding that
0.005 s is the rupture time of the strand. After a preliminary
numerical study on the dynamic response of the bridges to pulse
loads of infinite duration, 0.05 s is established as the time step to be
used in Newmark’s method. This time step allows us to achieve an
important reduction in computing time and to maintain adequate
accuracy in the results. A Rayleigh damping of 2% of the critical
damping, obtained with the first two longitudinal bending mode
shapes, is considered in the damped responses.
The numerical analysis is focused on obtaining the static and
dynamic bending moments at five cross sections of the deck of
the bridges studied which are critical for the design of the deck.
Four of these cross sections, which are placed at a lateral or at
the central span, are located at 24.80, 74.40, 173.60 and 201.00 m
from one of the ends of the bridge, respectively. The fifth studied
cross section coincides with the pylon. As the decks are modeled
with 2D shell elements, the bending moments are obtained by
integration of the internal normal forces acting on the elements.
The load combinations for positive and negative bending moments
are studied at these five cross sections and the stay which fails
is the critical one for the corresponding cross section and load
combination, i.e. the most unfavorable. Positive bending moments
are defined as those moments that cause compression at the upper
slab of the deck.
The envelopes of the ultimate limit states for permanent and
transitory loads, MULSp,t , are obtained from static analysis. The
static bending moment with the stay before its failure, M0 and the
static bending moment without the stay, Mw/o s are also obtained
from static analysis but with the load combination of AULS. The
dynamic bending moments produced by the rupture of the stay
along the first 20 s of movement and their envelopes, MAULSdyn ,
3291
γP 2 γQ γTemp γA ψQ ψTemp
1.00 0.00 0.00 0.00
– 0.60
1.35 1.50 1.50 0.00
1.00 0.00 0.00 1.00
0.50 0.20
1.00 1.00 1.00 0.00
are obtained from the dynamic analysis with the corresponding
load combination of AULS. In addition, the static bending moment
(1)
obtained by applying a D.A.F. equal to 2.0, MAULSDAF =2 , as is
suggested in [10,11] for studying the accidental ultimate limit state
of loss of a cable, is calculated with the same load combination
of AULS than the dynamic bending moments for comparison with
(2) them.
3. Results and discussion
Fig. 4 includes, as an example, the static and dynamic bending
moments along the deck for two different bridges and load combi-
nations which are representative of the obtained results. The fig-
ure shows how a dynamic load, produced by the failure of the stay,
generates dynamic bending moments at cross sections which are
located far away from the point where the dynamic load is applied.
Otherwise, if the same load is statically applied, it only produces a
bending moment at the cross sections which are close to the point
where the load is applied. These results illustrate why the approach
suggested in [10,11] for studying the accidental ultimate limit state
of loss of a cable, in which the dynamic analysis is substituted by a
static one and a D.A.F., becomes unsafe especially when the critical
stay is not located close to the cross section studied. Table 4 sum-
marizes the bending moments acting on the cross sections studied
for the cases plotted in Fig. 4. The table also includes the D.A.F. ob-
tained from the dynamic analysis defined as
MAULSdyn − M0
D.A.F . = . (3)
Mw/o S − M0
In the case shown in Fig. 4(a), the D.A.F. reaches the value of 5.28
in the undamped movement and 4.14 when damping is considered.
These values widely surpass the D.A.F. proposed in [10,11]. This
case is especially relevant since the dynamic bending moments
reached in both damped and undamped movements are larger
than the envelopes of the ultimate limit states for permanent and
transitory loads, MULSp,t , and the accidental limit state of loss of a
stay becomes critical for the design of the cross section of the deck.
Next, the obtained results for all of the deck cross sections
studied and bridges are summarized and compared using figures
whose corresponding values are detailed in tables included in the
Appendix.
3.1. Relative importance of the accidental limit state of loss of a stay
In order to determine the relative importance of the accidental
limit state of loss of a stay, the maxima dynamic bending moments
reached at the cross sections studied of the decks are compared
with the envelopes of the ultimate limit states for permanent and
transitory load combinations. This allows us not only to determine
the relative importance of the accidental event, but also to detect
where it becomes critical for the design of the cross section of the
deck.
3.1.1. Load combinations for negative bending moments
Fig. 5 shows graphically the ratios of MAULSdyn to MULSp,t at the
cross sections studied of the ten studied bridges for the load combi-
nations for negative bending moments. The corresponding values
3292 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300

Table 4
Static and dynamic bending moments (m kN) and the corresponding D.A.F. Cross section x = 24.80 m of the bridge F 1P d1.15 with load combination for negative bending
moments and cross section x = 173.60 m of the bridge H 1P d2.25 with load combination for positive bending moments.
Bridge
F 1P d1.15 H 1P d2.25

M0 −2478.89 10 063.05
Mw/o S −4355.12 37 393.59
Static analysis MAULSDAF =2 −6231.34 64 724.12
MULSp,t minima envelope −7327.49 −18 494.55
MULSp,t maxima envelope 20 641.04 37 617.95
MAULSdyn −12 388.53 10 063.05
Minima envelope
D.A.F. 5.28 0.00
ξ = 0%
MAULSdyn 4215.80 50 366.51
Maxima envelope
Dynamic analysis D.A.F. −3.57 1.47
MAULSdyn −10 254.88 10 063.05
Minima envelope
D.A.F. 4.14 0.00
ξ = 2%
MAULSdyn 1794.97 44 485.89
Maxima envelope
D.A.F. −2.28 1.26

Fig. 4. Bending moments (m kN) along the deck. Envelopes of the ultimate limit states for permanent and transitory loads, MULSp,t , envelopes of dynamic bending moment
produced by the rupture of the stay, MAULSdyn , without and with damping, static bending moment obtained applying a D.A.F. equal to 2.0, MAULSDAF =2 , and static bending
moment without the stay which fails, Mw/o s . Bridge F 1P d1.15. with the load combination corresponding to the AULS for negative bending moment at the cross section
x = 24.80 m, (a), and bridge H 1P d2.25 with the load combination corresponding to the AULS for positive bending moment at the cross section x = 173.60 m, (b).

are detailed in Table 9. For these load combinations, it has been
observed that the critical stay, whose failure produces the largest
bending moments on the cross section studied, is placed at a cer-
tain distance from the deck cross section.
The accidental event produces larger bending moments than
the envelope of the ultimate limit states for permanent and transi-
tory loads in 54% of the studied cases without damping and in 34%
when damping is considered. The averages of the ratios are 1.24
and 1.01 for the undamped and damped movements, respectively.
Larger values of the ratio are reached in the cases of bridges with
fan pattern and especially in those bridges with a stiffer deck. The
ratio is also larger in those bridges with only one plane of stays.
This result is expected due to the fact that the tension force act-
ing in the stays is larger on those bridges with a smaller number of
cables.
The influence of damping is studied with the ratio of the
damped to the undamped response, Fig. 6. It is shown that the
influence of the damping increases with the distance of the deck
cross section to the pylon. The reduction of the dynamic response
is also larger in those bridges with fan pattern and in those with
two planes of stays. Otherwise, the effect of the damping is not in-
fluenced by the stiffness of the deck.
 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300 3293

Fig. 5. Ratio of MAULSdyn to MULSp,t at the five deck cross sections of the ten bridges studied in the undamped (ξ = 0%) and damped (ξ = 2%) movements. Load combinations
for negative bending moments at the deck cross sections.

3.1.2. Load combinations for positive bending moments

Fig. 6. Ratio of the dynamic bending moments obtained in the damped movement
to the undamped at the five deck cross sections studied of the bridges. Load
combinations for negative bending moments at the deck cross sections.
The critical stay for the load combinations for positive bending
moments is the one close to the cross section studied in most of
the cases. The relative magnitude of MAULSdyn to MULSp,t obtained
in the case of load combinations for positive bending moments
is similar to the one obtained for the negative bending moments:
MAULSdyn is larger than MULSp,t in 54% of the cases without damping
and in 44% when damping is considered. The ratios of MAULSdyn to
MULSp,t are plotted in Fig. 7 and detailed in Table 10. The average
of these ratios is 1.22 and 1.02 for the undamped and damped
movements, respectively. Unlike the case of load combinations for
negative bending moments, the bridges with harp pattern show
a larger ratio. The layout of these bridges makes the cross sections
studied and the anchorage of the stay which fails coincide, while in
the case of the bridges with fan pattern, the distance between them
is 6.2 m. This small geometrical difference between fan and harp
patterns and the high rate of the bending moment in the proximity
of the anchorage of the stay that fails, see Fig. 4, lead to the larger
ratios encountered in the harp pattern. It has been verified that
when the deck cross section studied and the anchorage of the
stay coincide, the fan pattern leads to larger ratios as it happens
with load combinations of negative bending moments. In addition,
larger ratios have been encountered in the bridges with one plane
of stays and with a more flexible deck.

Fig. 7. Ratio of MAULSdyn to MULSp,t at the five deck cross sections of the ten bridges studied in the undamped (ξ = 0%) and damped (ξ = 2%) movements. Load combinations
for positive bending moments at the deck cross sections.
3294 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300
Fig. 8. Ratio of the dynamic bending moments obtained in the damped movement
to the undamped at the five deck cross sections studied of the ten bridges. Load
combinations for positive bending moments at the deck cross sections.
Fig. 8 shows that the influence of the damping is greater in the
deck cross sections which coincide with the pylon. In these cross
sections, the damped response is between 40% and 70% smaller
than the response of the undamped movement. This influence
is also greater in the bridges with the harp pattern and, it is in
these bridges, unlike those with the fan pattern, where the effect
of damping is greater when there is only one plane of stays. The
influence of the stiffness of the deck is only clear in the bridges with
the harp pattern. In these bridges, a stiffer deck leads to a larger
difference between the damped and undamped responses.
3.2. Dynamic amplification factor
In this section, the dynamic amplification factors, defined by
Eq. (3), corresponding to the bending moment at the deck cross
sections studied are detailed and compared.
3.2.1. Load combinations for negative bending moments
Table 11 summarizes the D.A.F. for each deck cross section
and bridge for the case of load combinations for negative bending
moments. The values are plotted in Fig. 9 for an easier comparison.
The obtained D.A.F. surpasses the value of 2.0 in most of the cases,
which is the maximum proposed in [10,11], reaching the extreme
Fig. 9. D.A.F. obtained at the five deck cross sections of the ten bridges studied in the undamped (ξ = 0%) and damped (ξ = 2%) movements. Load combinations for negative
bending moments at the deck cross sections.
values of 7.96 and 5.46 in the undamped and damped dynamic
responses, respectively, and 3.35 and 2.52 being the corresponding
average values.
Fig. 9 shows that the farther the distance of the cross section
from the pylon, the larger the D.A.F., especially at the lateral span
of the bridge. The greatest D.A.F. has been obtained at the cross
sections located at x = 24.80 m. The layout of the stays shows a
clear influence on the D.A.F., the fan pattern producing the largest
D.A.F. From the point of view of the stiffness of the deck, only
the bridges with the deck d0.40 have a similar response and the
obtained D.A.F. is close to 2.0. Nevertheless, there is a tendency for
the D.A.F. to increase in those bridges with a stiffer deck.
3.2.2. Load combinations for positive bending moments
The obtained D.A.F. with the load combinations for positive
bending moments are radically different; see Table 12 and Fig. 10.
With the exception of the deck cross section coinciding with the
pylon, where the D.A.F. reaches the extreme value of 42.53, the
D.A.F. is around 2.0. Thus, excluding these cross sections, the
average D.A.F. is 1.70 and 1.40 for the undamped and damped
movements, respectively. Once again, the layout of the stays, fan or
harp pattern, leads to clear differences in the obtained D.A.F., the
fan pattern generating the largest values. Otherwise, the stiffness
of the deck and the existence of one or two planes of stays do not
have an evident effect on the D.A.F.
3.3. Qualitative error incurred using the P.T.I. approach
In the simplified approach suggested in [10,11], the dynamic
analysis is substituted by a static analysis in which the dynamic
load produced by the rupture of the stay is approximated with
the force supported by the stay amplified with a D.A.F. equal to
2.0. Since the obtained D.A.F. is fairly different from this value, the
incurred error using the simplified approach is then investigated.
In order to quantify the magnitude of the error, the qualitative
error, Eq , Eq. (4), is obtained and expressed as a percentage of the
envelope of the ultimate limit states for permanent and transitory
loads, MULSp,t .
MALSdyn − MALSDAF =2
Eq = × 100. (4)
MULSp,t
 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300 3295

Fig. 10. D.A.F. obtained at the five deck cross sections of the ten bridges studied in the undamped (ξ = 0%) and damped (ξ = 2%) movements. Load combinations for
positive bending moments at the deck cross sections.

Fig. 11. Qualitative error, Eq , Eq. (4), incurred when using a static analysis with a D.A.F. equal to 2.0 instead of a dynamic analysis for evaluating the response of the bridge
to the sudden failure of one stay. Load combinations for negative bending moments at the deck cross section of the bridges studied.

3.3.1. Load combinations for negative bending moments
Fig. 11 shows the qualitative error, Eq , corresponding to the
load combinations for negative bending moments and Table 13 the
corresponding values. With the exception of those cross sections
coinciding with the pylon, most of the values have a negative sign.
The average of the error is −35.02% for the undamped movement
and −12.21 with damping. Extreme values, such as −126.60% and
−246.60%, are reached in some deck cross sections.
The layout of the stays, fan or harp pattern, and the number of
planes of stays have a great influence on the error. Thus, larger
qualitative errors can be found in the bridges with one plane of
stays and with the fan pattern. The stiffness of the deck has no clear
influence on the magnitude of the qualitative error.
Damping has an opposite effect depending on whether the cross
section coincides with the pylon or not. Thus, damping increases
the error in those deck cross sections which coincide with the
pylon and reduces it in the rest of the cases. In general, the effect of
damping is larger in bridges with the harp pattern and two planes
of stays.
Results show that the error when using a D.A.F. of 2.0 for eval-
uating the dynamic response to the loss of a stay has a relevant
magnitude in terms of the envelope of the bending moment for
permanent and transitory loads. Those cases in which the qualita-
tive error, Eq , has a negative sign and the magnitude of the dynamic
envelope, MAULSdyn , is larger than MULSp,t are of major importance
since the safety of the bridge is compromised in the case of the
rupture of the stay. Otherwise, if the qualitative error, Eq , is posi-
tive and the magnitude of the MAULSDAF =2 , is larger than MULSp,t , the
deck cross section will be unnecessarily overdimensioned. In order
to identify one or other situation, Table 5 details the ratio of MAULS
to MULSp,t , where MAULS is the largest of MAULSdyn and MAULSDAF =2 for
each case. Those cases in which MAULSdyn is larger than MAULSDAF =2
are underlined. At 24 cross sections, which means 48% of the total
50 cross sections studied, the undamped movement produces a dy-
namic bending moment which is larger than the static MAULSDAF =2
and also than the envelope of MULSp,t . Damping reduces this per-
centage to 20%, which continues being a relevant amount.
One of these cases is the cross section x = 24.80 m of the bridge
F 1P d 1.15, Fig. 4(a), where the distribution of normal forces per
unit length is shown in Fig. 12. This figure shows how the dynamic
normal forces per unit length, for both the undamped and damped
movements, exceed the values obtained in the simplified approach
suggested in [10,11], MAULSDAF =2 , and the envelope of permanent
and transitory loads, MULSp,t . Thus, the integrity of the cross section
will be compromised in the case that it has to face the internal
normal forces produced by the failure of the stay and they had been
evaluated with a static analysis and a D.A.F. of 2.0.
3296 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300

Table 5
Ratio of MAULS to MULSp,t , where MAULS is the largest of MAULSdyn and MAULSDAF =2 . Those cases in which MAULSdyn > MAULSDAF =2 are underlined. Load combinations for negative
bending moments at the deck cross sections.
Bridge MAULS /MULSp,t
ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

F 2P d0.40 0.76 1.15 0.80 0.88 0.73 0.75 1.15 0.80 0.84 0.71
F 1P d1.15 1.69 2.69 1.40 1.24 3.36 1.40 2.33 1.40 1.01 2.69
F 2P d1.15 1.26 1.50 0.80 1.16 2.05 0.86 1.25 0.80 0.75 1.03
F 1P d2.25 1.17 1.80 1.06 1.34 3.87 0.90 1.61 1.06 0.95 2.37
F 2P d2.25 1.17 1.06 0.68 0.86 1.55 0.94 0.92 0.68 0.58 0.89
F 2P d0.40 0.86 1.12 1.15 0.86 0.76 0.79 1.01 1.15 0.84 0.72
H 1P d1.15 1.25 1.78 1.70 1.24 1.36 1.01 1.78 1.70 0.97 1.21
H 2P d1.15 0.83 1.10 0.95 0.90 0.91 0.71 1.08 0.95 0.82 0.80
H 1P d2.25 1.08 1.24 1.05 1.12 1.58 0.91 1.17 1.05 0.89 1.26
H 2P d2.25 0.82 0.76 0.65 0.78 0.99 0.70 0.75 0.65 0.63 0.81

Envelope A.U.L.S.dyn ξ =0% A.U.L.S.DAF=2

Envelope A.U.L.S.dyn ξ =2% Envelope U.L.S. p,t

Fig. 12. Normal force per unit length (kN/m) of web and slab acting at the deck cross section x = 24.80 m of the bridge F 1P d1.15. Envelopes of the undamped and damped
dynamic analyses, AULSdyn , static analysis with a D.A.F. equal to 2.0, ALSDAF =2 and envelope of the ULS for permanent and transitory loads, ULSp,t . Load combination for
negative bending moment.

Fig. 13. Bending moment (My) and normal force (Nx) throughout the first twenty seconds after the failure of the stay. Deck cross section x = 24.80 m of the bridge F 1P
d1.15. Load combination for negative bending moment at the deck cross section.

Table 6
Average values of the maxima and minima ratios of the dynamic internal force, normal force (Nx) and bending moment (My), to its initial values at time t = 0 s obtained in
the ten bridges studied. Load combinations for negative bending moments at the deck cross sections.
Load ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

Max 1.14 1.10 1.03 1.40 2.93 1.10 1.04 1.01 1.22 2.34
Nx
Min 0.91 0.85 0.92 0.63 −0.78 0.95 0.88 0.94 0.73 0.07
Max 3.09 11.28 2.58 3.56 3.00 2.52 9.65 2.34 2.69 2.34
My
Min −0.32 −2.39 0.99 −0.28 −1.10 0.32 0.30 1.00 0.66 −0.39

The values plotted in Fig. 12 are produced by the combination
of the total internal normal force and bending moment acting at
the cross section, whose values during the first twenty seconds
of movement are shown in Fig. 13 for the dynamic cases. This
figure shows how the amplitude of the interval where the dynamic
bending moment and the dynamic normal force take values is
very different. Thus, in the undamped movement, the dynamic
bending moment varies between 1.70 and 5.00 times its initial
value at time t = 0 s. Otherwise, the dynamic normal force
takes values in a smaller interval which varies from 0.93 to 1.08
times its initial value. Similar results are observed in the rest of
the bridges studied. Table 6 summarizes the average values of the
maxima and minima ratios of the dynamic internal force to its
initial value at time t = 0 s obtained in the ten bridges studied.
Except for the central cross section, the dynamic response of the
bridge produces a larger variation in dynamic bending moments
than in the dynamic normal forces acting on the deck cross section.
3.3.2. Load combinations for positive bending moments
In the case of load combinations for positive bending moments,
the qualitative error, Eq , defined by Eq. (4), is positive in those deck
cross sections placed at the lateral and central spans of the bridges
and its average value is 20.86% and 35.14% for the undamped and
damped movements, respectively. Opposite results are obtained at
 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300 3297

Table 7
Ratio of MAULS to MULSp,t , where MAULS is the largest of MAULSdyn and MAULSDAF =2 . Those cases in which MAULSdyn > MAULSDAF =2 are underlined. Load combinations for positive
bending moments at the deck cross sections.
Bridge MAULS /MULSp,t
ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

F 2P d0.40 1.60 2.20 0.82 1.64 1.00 1.60 2.20 0.41 1.64 1.00
F 1P d1.15 2.18 1.64 2.08 1.50 1.41 2.18 1.64 1.51 1.50 1.39
F 2P d1.15 1.09 0.96 1.00 0.88 0.94 1.09 0.92 0.57 0.88 0.94
F 1P d2.25 1.60 1.50 1.21 1.27 0.86 1.60 1.50 0.72 1.27 0.86
F 2P d2.25 0.91 0.86 1.02 0.77 0.88 0.91 0.86 0.68 0.77 0.88
H 2P d0.40 1.60 2.10 1.16 1.74 0.99 1.60 2.10 1.16 1.74 0.99
H 1P d1.15 2.29 3.79 1.69 2.60 0.88 2.29 3.79 1.20 2.60 0.88
H 2P d1.15 1.17 1.82 1.04 1.31 0.98 1.17 1.82 0.43 1.31 0.98
H 1P d2.25 1.58 2.47 1.31 1.72 0.87 1.58 2.47 0.72 1.72 0.78
H 2P d2.25 0.90 1.21 0.76 0.95 0.76 0.90 1.21 0.40 0.95 0.76

Fig. 14. Qualitative error, Eq , Eq. (4), incurred when using a static analysis with a D.A.F. equal to 2.0 instead of a dynamic analysis for evaluating the response of the bridge
to the sudden failure of one stay. Load combinations for positive bending moments at the deck cross sections of the bridges studied.

Fig. 15. Normal force per unit length (kN/m) of web and slab acting at the deck cross section x = 173.60 m of the bridge H 1P d2.25. Envelopes of the undamped and
damped dynamic analysis, AULSdyn , static analysis with a D.A.F. equal to 2.0, ALSDAF =2 and envelope of the ULS for permanent and transitory loads, ULSp,t . Load combinations
for positive bending moments.

the cross sections which coincide with the pylon. In these cases,
the qualitative error is negative and it has an average magnitude of
−71.18% and −37.40% for the undamped and damped movements,
respectively. Fig. 14 shows the qualitative error, Eq , for each deck
cross section and bridge and the values are detailed in Table 14.
In the case of load combinations for positive bending moments, a
stiffer deck and only one plane of stays are characteristics which
lead to a greater error.
Damping has a larger effect on those bridges with fan pattern
and on the cross sections which are resting on the pylon.
Table 7 summarizes the ratio of MAULS to MULSp,t as was defined
for the load combinations for negative bending moments. The sim-
plified analysis proposed in [10,11], AULSDAF =2 , leads to bending
moments which are larger than those obtained in the dynamic
analysis, AULSdyn , and than the envelope of permanent and tran-
sitory loads, ULSp,t in 46% and 50% for the undamped and damped
movements, respectively. In these cases, the deck will be unneces-
sarily overdimensioned if the dynamic analysis is replaced by the
simplified analysis AULSDAF =2 .
Fig. 15 shows the normal forces per unit length of one of these
cases which correspond with the cross section x = 173.60 m of the
bridge H 1P d2.25. This figure illustrates how using the simplified
static analysis AULSDAF =2 leads to an overdimensioned deck cross
section.
Fig. 16 depicts the dynamic bending moments and normal
forces acting on the cross section x = 173.60 m of the bridge H
1P d1.15 during the first twenty seconds of movement. Table 8
summarizes the average ratios of the dynamic internal force to its
initial value at time t = 0 s obtained in the ten bridges studied. As
in the case of load combinations for negative bending moments,
the dynamic bending moment varies in an interval with a larger
amplitude than the dynamic normal force.
3298 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300

Fig. 16. Bending moment (My) and normal force (Nx) during the first twenty seconds after the failure of the stay. Deck cross section x = 173.60 m of the bridge H 1P d2.25.
Load combination for positive bending moments.

Table 8
Average values of the maxima and minima ratios of the dynamic internal force, normal force (Nx) and bending moment (My), to its initial values at time t = 0 s obtained in
the ten bridges studied. Load combinations for positive bending moments.
Load ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

Max 1.15 1.07 1.12 1.37 2.80 1.11 1.05 1.07 1.18 2.35
Nx
Min 0.86 0.96 0.88 0.54 −0.75 0.88 0.98 0.93 0.64 0.09
Max 4.48 7.56 22.68 4.05 2.57 3.94 6.76 15.33 3.53 2.20
My
Min 0.98 0.99 −17.83 0.99 0.92 0.99 0.99 −8.82 1.00 0.95

4. Conclusions
A parametric study on the response of cable stayed bridges to
the sudden rupture of one stay has been developed in this paper.
The bridges studied were defined by combining characteristics
such as the layout of the stays, fan or harp pattern, the stiffness of
the deck and the arrangement of the stays in one or two planes.
The Accidental Ultimate Limit State of failure of a stay, A.U.L.S.,
was investigated following two different approaches: by means of
a dynamic analysis with time-step integration of the equation of
motion, A.U .L. S .dyn , and by a simplified static analysis in which a
dynamic amplification factor (D.A.F.) equal to 2.0 is applied to the
static load, A.U .L. S .DAF =2 , as is suggested in [11,10]. The A.U.L.S.
was also compared with the envelope of the ultimate limit state
of permanent and transitory loads, U .L. S .p,t . The results presented
and discussed in this paper focus on the bending moments acting
on five cross sections of the deck of the bridges studied. The
objectives of the study are to determine the qualitative importance
of the accidental limit state of failure of a stay in the design of the
deck and to define the safety level reached when using a static
analysis with a D.A.F. equal to 2.0 for determining the response of
a cable stayed bridge to the sudden failure of a stay.
The main conclusions reached are:
• Cable stayed bridges are redundant structures in which a static
load has a relatively small zone of influence, which is placed
around the point where the load is applied, and where the
internal loads acting on the bridge are modified by the load.
Otherwise, if the same load is dynamically applied, its effects
produce a relevant modification of the internal loads acting on
cross sections which are located far away from the load. Thus,
using the same D.A.F. for all of the cross sections and elements of
the structure is not a rational approach to study the Accidental
Limit State of failure of a stay.
• The accidental failure of one stay produces large bending mo-
ments on the deck becoming a ULS critical for the design of
the deck in many cases. The bending moment corresponding to
A.U .L. S .dyn overcomes the envelope of U .L. S .p,t in 54% of the
cases studied for the undamped movement. Damping reduces
this percentage to about 40%. A fan pattern, stays arranged in
one central plane and a stiffer deck leads to a larger ratio of
A.U .L. S .dyn to U .L. S .p,t .
• In the case of load combinations for negative bending moments,
the D.A.F. greatly exceeds the value of 2.0. The average of the ob-
tained D.A.F. in the undamped and damped movements is 3.35
and 2.52, respectively, and an extreme value of 7.96 was ob-
served. A fan pattern and a stiffer deck lead to a larger D.A.F.
which also increases with the distance of the cross section to
the pylon.
• In the case of load combinations for positive bending moments,
the average of the D.A.F. was 1.70 and 1.40 for the undamped
and damped movements, respectively. It was observed that a
fan pattern leads to a larger D.A.F.
• The simplified approach consisting of using a constant D.A.F
equal to 2.0 for evaluating the accidental ultimate limit state of
failure of a stay leads us to incur in a non-negligible error. In case
of load combinations for negative bending moments, the error is
negative (safe) in most of the cases, and with average values, ex-
pressed as percentages of the envelope of U .L. S .p,t , of −35.02%
and −12.21% for the undamped and damped movements, re-
spectively. Thus, the simplified approach becomes very unsafe
in these cases not only because of the sign of the error but also
because of its magnitude. On the other hand, for the load com-
binations for positive bending moments, the error is positive in
most of the cases. The simplified approach does not compro-
mise the safety of the bridge but leads to an unnecessary overdi-
mensioned deck.
• From the design of the deck point of view, it is necessary to per-
form a dynamic analysis for evaluating the response of the deck
to the accidental limit state of the sudden loss of a stay.
Acknowledgements
The authors would like to thank the Spanish Junta de Comu-
nidades de Castilla-La Mancha for the economical support given
for the research project PBI-05-031.
Appendix
See Tables 9–14.
 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300 3299

Table 9
Ratio of MAULSdyn to MULSp,t at the five deck cross sections of the ten bridges studied in the undamped (ξ = 0%) and damped (ξ = 2%) movements. Load combinations for
negative bending moments at the cross sections.
Bridge MAULSdin /MULSp,t
ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

F 2P d0.40 0.76 1.04 0.78 0.88 0.73 0.73 0.96 0.76 0.78 0.68
F 1P d1.15 1.69 2.69 1.26 1.24 3.36 1.40 2.33 1.14 1.01 2.69
F 2P d1.15 1.26 1.50 0.79 1.16 2.05 0.86 1.25 0.74 0.75 1.03
F 1P d2.25 1.17 1.80 0.98 1.34 3.87 0.90 1.61 0.88 0.95 2.37
F 2P d2.25 1.17 1.06 0.65 0.86 1.55 0.94 0.92 0.59 0.58 0.89
H 2P d0.40 0.86 1.12 1.11 0.86 0.76 0.79 0.92 1.08 0.79 0.69
H 1P d1.15 1.25 1.76 1.50 1.24 1.36 1.01 1.66 1.25 0.97 1.14
H 2P d1.15 0.83 1.10 0.83 0.90 0.91 0.71 1.01 0.76 0.77 0.80
H 1P d2.25 1.08 1.24 0.94 1.12 1.58 0.91 1.11 0.85 0.89 1.26
H 2P d2.25 0.82 0.76 0.62 0.78 0.99 0.70 0.72 0.56 0.63 0.81

Table 10
Ratio of MAULSdyn to MULSp,t at the five deck cross sections of the ten bridges studied in the undamped (ξ = 0%) and damped (ξ = 2%) movements. Load combinations for
positive bending moments at the cross sections.
Bridge MAULSdin /MULSp,t
ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

F 2P d0.40 1.17 1.68 0.82 1.22 0.93 1.08 1.56 0.41 1.15 0.82
F 1P d1.15 2.00 1.61 2.08 1.31 1.41 1.81 1.42 1.51 1.13 1.26
F 2P d1.15 1.04 0.96 1.00 0.81 0.92 0.95 0.87 0.57 0.71 0.82
F 1P d2.25 1.60 1.31 1.21 1.14 0.86 1.25 1.20 0.72 0.93 0.72
F 2P d2.25 0.84 0.82 1.02 0.68 0.87 0.77 0.75 0.68 0.62 0.73
H 2P d0.40 1.27 1.58 1.04 1.37 0.87 1.07 1.50 1.12 1.21 0.81
H 1P d1.15 1.78 3.05 1.69 1.91 0.75 1.68 2.64 1.20 1.66 0.66
H 2P d1.15 1.04 1.44 1.04 1.03 0.91 0.90 1.30 0.43 0.91 0.76
H 1P d2.25 1.34 1.89 1.31 1.34 0.87 1.14 1.69 0.72 1.18 0.69
H 2P d2.25 0.81 0.98 0.76 0.80 0.73 0.70 0.87 0.40 0.68 0.60

Table 11
D.A.F. obtained at the five deck cross sections of the ten bridges studied in the undamped (ξ = 0%) and damped (ξ = 2%) movements. Load combinations for negative
bending moments at the cross sections.
Bridge D.A.F.
ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

F 2P d0.40 2.11 1.66 1.89 2.19 2.05 1.89 1.43 1.79 1.74 1.85
F 1P d1.15 5.28 4.07 1.75 3.86 2.89 4.14 3.49 1.50 2.96 2.43
F 2P d1.15 7.96 4.75 1.92 3.76 3.69 4.70 3.96 1.72 2.22 2.80
F 1P d2.25 7.83 2.57 1.78 3.71 4.51 5.35 2.24 1.50 2.38 2.99
F 2P d2.25 7.40 2.77 1.82 4.43 4.54 5.46 2.30 1.44 2.51 3.27
H 2P d0.40 2.56 2.40 1.88 2.09 2.17 2.17 1.71 1.77 1.75 1.88
H 1P d1.15 5.38 1.97 1.71 4.06 2.36 3.77 1.82 1.33 2.66 1.84
H 2P d1.15 5.18 2.07 1.61 2.41 2.62 3.65 1.80 1.36 1.72 2.07
H 1P d2.25 6.27 2.18 1.70 3.45 3.58 4.77 1.87 1.45 2.31 2.65
H 2P d2.25 7.52 2.05 1.83 3.76 3.81 5.45 1.88 1.43 2.32 2.71

Table 12
D.A.F. obtained at the five deck cross sections of the ten bridges studied in the undamped (ξ = 0%) and damped (ξ = 2%) movements. Load combinations for positive
bending moments at the cross sections.
Bridge D.A.F.
ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

F 2P d0.40 1.37 1.48 7.45 1.40 1.78 1.24 1.36 5.46 1.30 1.44
F 1P d1.15 1.81 1.95 7.63 1.67 2.05 1.60 1.65 5.43 1.36 1.59
F 2P d1.15 1.86 2.14 8.89 1.73 1.94 1.63 1.81 5.78 1.36 1.57
F 1P d2.25 2.01 1.67 15.96 1.74 2.03 1.46 1.49 8.80 1.29 1.44
F 2P d2.25 1.76 1.81 19.01 1.57 1.95 1.51 1.55 12.26 1.31 1.42
H 2P d0.40 1.51 1.46 4.48 1.51 1.64 1.22 1.38 2.78 1.30 1.43
H 1P d1.15 1.50 1.60 12.98 1.42 1.60 1.40 1.38 9.20 1.21 1.29
H 2P d1.15 1.69 1.55 20.93 1.47 1.80 1.40 1.38 9.85 1.24 1.39
H 1P d2.25 1.63 1.49 23.71 1.47 2.39 1.33 1.32 11.96 1.26 1.64
H 2P d2.25 1.70 1.52 42.53 1.55 1.86 1.34 1.30 18.51 1.20 1.33
3300 C.M. Mozos, A.C. Aparicio / Engineering Structures 32 (2010) 3288–3300

Table 13
Qualitative error, Eq , see Eq. (4), incurred using a static analysis with a D.A.F. equal to 2.0 instead of a dynamic analysis for evaluating the response of the bridge to the sudden
failure of one stay. Load combinations for negative bending moments at each deck cross section of the bridges studied.
Bridge Eq (%)
ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

F 2P d0.40 −1.79 11.25 2.38 −4.22 −1.27 1.86 18.80 4.49 5.71 3.67
F 1P d1.15 −84.03 −126.60 13.40 −49.08 −128.64 −54.91 −90.99 26.04 −25.42 −62.13
F 2P d1.15 −74.27 −85.74 1.60 −46.60 −192.78 −33.63 −61.03 5.84 −5.73 −91.61
F 1P d2.25 −63.27 −33.92 7.99 −50.06 −246.60 −36.39 −14.38 18.11 −11.02 −97.17
F 2P d2.25 −63.72 −23.00 3.04 −35.16 −132.72 −40.84 −8.93 9.48 −7.34 −66.30
H 2P d0.40 −9.97 −11.35 3.63 −1.86 −4.09 −3.08 8.13 6.88 5.47 2.91
H 1P d1.15 −50.15 1.92 19.84 −38.91 −15.24 −26.22 11.97 45.38 −12.41 6.87
H 2P d1.15 −24.77 −2.14 11.35 −7.81 −12.99 −12.84 6.54 18.34 5.34 −1.57
H 1P d2.25 −48.38 −6.95 10.79 −29.76 −53.17 −31.43 5.25 19.56 −6.43 −21.88
H 2P d2.25 −31.54 −0.87 2.61 −17.55 −29.59 −19.71 2.30 8.96 −3.20 −11.60

Table 14
Qualitative error, Eq , see Eq. (4), incurred using a static analysis with a D.A.F. equal to 2.0 instead of a dynamic analysis for evaluating the response of the bridge to the sudden
failure of one stay. Load combinations for positive bending moments at each deck cross section of the bridges studied.
Bridge Eq (%)
ξ = 0% ξ = 2%
24.40 74.40 Pylon 173.60 201.50 24.40 74.40 Pylon 173.60 201.50

F 2P d0.40 43.54 52.28 112.16 42.16 7.06 51.95 64.50 71.28 49.35 17.59
F 1P d1.15 18.12 3.37 −145.79 19.35 −1.69 37.16 22.65 −88.89 37.19 13.58
F 2P d1.15 5.18 −3.80 −94.47 6.98 1.50 14.15 5.01 −51.77 16.76 11.45
F 1P d2.25 −0.40 19.30 −95.56 12.29 −0.73 34.56 29.68 −46.52 33.40 13.94
F 2P d2.25 6.92 4.71 −84.57 9.34 1.19 14.14 11.38 −51.02 15.00 15.02
H 2P d0.40 33.64 51.95 12.68 37.28 11.24 53.57 59.78 4.00 53.27 17.83
H 1P d1.15 51.45 73.45 −141.59 68.38 12.43 61.25 114.46 −92.82 93.16 22.15
H 2P d1.15 13.88 37.51 −103.93 27.91 7.21 27.06 51.90 −43.10 40.17 21.96
H 1P d2.25 24.03 58.01 −109.29 38.17 −9.45 43.68 77.26 −50.13 53.80 8.82
H 2P d2.25 9.17 23.06 −61.40 15.05 3.31 20.11 33.92 −25.02 26.88 15.78

References
[1] Walther R, Houriet B, Isler W, Moïa P, Klein JF. Cable stayed bridges. 2nd ed.
Thomas Telford Publishing; 1999.
[2] Podolny W, Muller JM. Construction and design of prestressed concrete
segmental bridges. John Wiley and Sons, Inc.; 1982.
[3] Gimsing NJ. Cable supported bridges, concept and design. 2nd ed. John Wiley
and Sons, Inc.; 1997.
[4] I.A.B.S.E. Proceedings of IABSE conference cable-stayed bridges. past, present
and future. IABSE; 1999.
[5] I.A.B.S.E. Proceedings of IABSE conference on cable-supported bridges in Seoul.
IABSE; 2001.
[6] Saul R, Humpf K. Inspection and maintenance of cable stayed bridges —
German experiences. In: Proceedings of IABMAS international conference on
bridge maintenance, safety and management. 2002.
[7] Lecroq P, Virlogeux M, Foucriat JC, Biétry J. Replacement of the suspension
cables of the tancarville bridge. In: Proceedings of IABSE conference on cable-
supported bridges in Seoul. IABSE; 2001.
[8] Andersen H, Hommel DL, Veje EM. Emergency rehabilitation of the Zárate-
Brazo Largo Bridges, Argentina. In: Proceedings of IABSE conference cable-
stayed bridges. Past, present and future. 1999.
[9] Hyttinen E, Välimaki J, Jävenpää E. Cable-stayed bridges, effect of breaking of a
cable. Proceedings of international conference A.I.P.C. F.I.P.: cable-stayed and
suspension bridges, vol. II. 1994.
[10] P.T.I. Recommendations for stay cable design, testing and installation. USA:
Post-Tensioning Institute; 2007.
[11] S.E.T.R.A. Haubans. Recommandations de la Commission Interministérielle
de la Précontrainte, Service d’Etudes Techniques des Routes et Autoroutes.
Francia; 2001.
[12] Clough RW, Penzien J. Dynamics of structures. 2nd ed. McGraw-Hill, Inc.; 1993.
[13] Chopra AK. Dynamics of structures. In: Theory and applications to earthquake
engineering. 2nd ed. Prentice-Hall, Inc.; 2001.
[14] Mozos CM, Aparicio AC. Static strain energy and dynamic amplifica-
tion factor on multiple degree of freedom systems. Eng Struct, on line.
doi:10.1016/j.engstruct.2009.07.003.
[15] Mozos CM, Aparicio AC. Proceedings of IABMAS conference on bridge
maintenance, safety, management, life-cycle performance and cost. IABMAS;
2006.
[16] Mozos CM. Theoretical and experimental study on the structural response of
cable stayed bridges to a stay failure. Ph.D. thesis. Advisor: Aparicio, A.C. Spain:
Universidad de Castilla-La Mancha; 2007 [in Spanish].
[17] Wolff M, Starossek U. Robustness assessment of a cable-stayed bridge. In:
Proceedings, 4th international conference on bridge maintenance, safety, and
management. 2008.
[18] Wolff M, Starossek U. Cable loss and progressive collapse in cable-stayed
bridges. Bridge Struct 2009;5/2009:17–28.
[19] Starossek U. Avoiding disproportional collapse of major bridges. Struct Eng
Internat 2009;3/2009:289–97.
[20] Ruiz–Teran AM, Aparicio AC. Dynamic amplification factors in
cable-stayed structures. J Sound Vibration 2007;300:197–216.
doi:10.1016/j.jsv.2006.07.028.
[21] Mozos CM, Aparicio AC. Parametric study on the dynamic response of cable
stayed bridges to the sudden failure of a stay, Part II: Bending moment
acting on the pylons and stress on the stays. Eng Struct 2010;32:3301–12.
doi:10.1016/j.engstruct.2010.07.002.
[22] I.A.P. IAP: Instrucción sobre las Acciones a Considerar en el Proyecto de Puentes
de Carretera. Ministerio de Fomento. 1998.