Journal of Earthquake Engineering

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/ueqe20

Probabilistic Seismic Performance Analysis of RC

Bridges

Araliya Mosleh, Jose Jara, Mehran S. Razzaghi & Humberto Varum

To cite this article: Araliya Mosleh, Jose Jara, Mehran S. Razzaghi & Humberto Varum (2020)
Probabilistic Seismic Performance Analysis of RC Bridges, Journal of Earthquake Engineering,
24:11, 1704-1728, DOI: 10.1080/13632469.2018.1477637

To link to this article: https://doi.org/10.1080/13632469.2018.1477637

Published online: 05 Jun 2018.

Submit your article to this journal

Article views: 653

View related articles

View Crossmark data

Citing articles: 19 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ueqe20
JOURNAL OF EARTHQUAKE ENGINEERING
2020, VOL. 24, NO. 11, 1704–1728
https://doi.org/10.1080/13632469.2018.1477637

Probabilistic Seismic Performance Analysis of RC Bridges

a
Araliya Mosleh , Jose Jarab, Mehran S. Razzaghic, and Humberto Varuma
a
CONSTRUCT-LESE, Department of Civil Eng., Faculty of Engineering, University of Porto, Porto, Portugal;
b
Department of Civil Eng., Faculty of Engineering, Universidad Michoacana de San Nicolas de Hidalgo,
Morelia, Mexico; cDepartment of Civil Eng., Faculty of Engineering, Qazvin Branch, Islamic Azad University,
Qazvin, Iran

ABSTRACT ARTICLE HISTORY

This paper focuses on the evaluation of the seismic vulnerability of Received 31 July 2017
pre-1990 highway concrete bridges, based on the development of Accepted 10 May 2018
analytical fragility curves. The parameters considered are the span KEYWORDS
length, column height, lap splices, reinforcement yield strength, and Fragility Curves; Seismic
concrete compressive strength. Nonlinear response history analyses Vulnerability; Concrete
using 3D models were conducted for each set of bridge samples Bridge; Nonlinear Analyses
subjected to earthquake ground motions with different intensities.
Moreover, this study analyzes the influence of the earthquake fault
type on the seismic vulnerability of the bridges, using as a perfor-
mance parameter the displacement ductility demand of the piers.
Based on this parameter, to support future seismic-risk mitigation
efforts fragility curves are utilized to assesses the seismic vulnerability
of typical concrete bridge classes in Iran. The selected suite of seismic
records originates in strike-slip and reverse fault seismic sources. The
bridges subjected to reverse fault records present pier demands and
seismic vulnerability larger than the bridges subjected to strike-slip
accelerograms. The study shows that the span length, lap splice,
column height, and seismic fault type have significant effects on
the seismic vulnerability of bridge piers. Additionally, the results are
useful to identify and prioritize retrofit actions of the most seismically
vulnerable bridges.

1. Introduction
Past earthquakes around the world showed that bridges are one of the most vulnerable
components of highway transportation systems and exposed the need to survey the vulner-
ability of bridges, especially those designed according to old regulation codes [Jara et al.,
2015]. Seismic vulnerability assessment of highway bridges located in areas of high seismic
hazard plays an important role in the safety of transportation systems. A proper design tends
to reduce the amount of damage and ensures a better bridge performance. Strict recom-
mendations made by current guidelines and codes aimed to avoid collapses during earth-
quakes. Low ductility and inadequate resistance of the columns led to severe damage in past
earthquakes [Qiang et al., 2009; Kawashima and Unjoh, 1997]. In addition, insufficient
splice length, splices of the longitudinal reinforcement in a potential plastic hinge region,

CONTACT Araliya Mosleh [email protected]

Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/ueqe.
© 2018 Taylor & Francis Group, LLC
 JOURNAL OF EARTHQUAKE ENGINEERING 1705

and premature shear failures have also been investigated by several researchers [Basoz et al.,
1999; Eshghi and Ahari, 2005; Hsu and Fu, 2004; Kawashima, 2002; Yamazaki et al., 2000].
Many of existing highway bridges were designed with outdated codes and guidelines.
Such bridges usually do not meet the requirement of current seismic design codes.
[Caltrans, 2013]. The use of fragility curves to assess the seismic vulnerability of structures
can be traced back to 1975, when the seismic risk assessment method was formalized by
Whitman et al. [Whitman et al., 1975]. Veneziano et al. [1983] used the total probability
theorem to compute mean annual frequency of reaching or exceeding a particular limit
state. In the last two decades, fragility curves demonstrated to be an efficient and useful
tool to evaluate the safety of structures and infrastructures. In 1991, the Applied
Technology Council (ATC) introduced the concept of fragility function in the ATC 25
report [ATC25, 1991]. Subsequently, in 1997 a risk assessment software package [HAZUS,
1997] based on the geographical information system (GIS) was presented by the Federal
Emergency Management Agency (FEMA). A set of seismic fragility curves was presented
by HAZUS for conventional RC bridges [HAZUS, 2012].
In the published literature, there are several approaches to assess the seismic vulnerability
of bridges [Padgett et al., 2010; Ghosh et al., 2013; Seo and Linzell, 2013; Zelaschi et al., 2016;
Mangalathu, 2017; Monteiro et al., 2017; Simon and Vigh, 2017; Mosleh et al., 2018]. One of
the oldest and simplest methods of obtaining fragility curves is a judgmental or expert-based
fragility curve which considers expert opinions on the bridge response data [Rossetto and
Elnashai, 2003]. Other researchers developed empirical fragility curves based on the post-
earthquake damage data obtained from field investigations [Shinozuka et al., 2000b; Basoz
and Kiremidjian, 1997; Shinozuka et al., 2001]. In the absence of adequate damage data, an
analytical fragility curve is a good choice for probabilistic assessment of the seismic perfor-
mance of highway bridges. Development of fragility curves for a particular structure requires
calculation of capacity and demand of the structure. The bridge capacity and demand can be
estimated by a variety of analytical methods such as the elastic spectral method [Hwang et al.,
2000], nonlinear static analysis [Mander, 1999; Moschonas et al., 2009; Monteiro et al.,
2016a], nonlinear time history analysis (NLTHA) [Choi et al., 2004; Karim and Yamazaki,
2003; Mosleh et al., 2016a; Nielson and DesRoches, 2007a; Pan et al., 2010; Ramanathan
et al., 2012; Tavares et al., 2012; Mosleh et al., 2016b; Monteiro et al., 2016b], and incremental
dynamic analysis (IDA) [Billah et al., 2013]. Nonlinear response history analysis (NRHA) is
known as the most reliable method of developing fragility curves, although it is the most
computationally time-consuming method [Shinozuka et al., 2000b].
The main objective of this study is to assess the seismic vulnerability of conventional pre-
1990 RC bridges. A group of highway bridges representing the most common pre-1990 RC
bridges in Iran was selected. Nonlinear response history analysis was carried out, using
accelerograms of reverse and strike-slip seismic sources. Finally, analytical fragility curves of
typical Iran bridges were developed in terms of fault type of input motions, span length,
column height, lap splices, reinforcement yield strength, and concrete compressive strength.

2. Bridge Characteristics
Obtaining fragility curves for individual bridges is neither practical nor feasible when a
large number of bridges are considered. Bridge classification helps to select appropriate
type of bridges for vulnerability analysis. Even though there are many structural attributes
1706 A. MOSLEH ET AL.

that can contribute to the seismic response of bridges, different proposals (as this study
does) consider for practical reasons, limited number of parameters.
Previous studies classify bridges according to different structural properties. ATC-13
[ATC-13, 1985], for example, presents a bridge classification as a function of the total length.
In the classification developed by the Federal Highway Administration [FHWA, 1988],
bridges are grouped based on the superstructure type, material, and continuity at supports.
HAZUS [FEMA, 2003] considers a bridge classification in terms of the number of spans,
abutment and bearing types, pier type, and seismic design. Nielson [2005] proposes 11 bridge
classes considering construction type, construction material, and the number of spans. Avsar
et al. [2011] classify the bridges into four groups based on the skew angle of the superstructure.
They classified highway bridges into four groups by analyzing the characteristics of a family of
52 ordinary existing bridges in Turkey. Based on that information and engineering judge-
ment, the authors considered skew angle of the superstructure, number of columns per bent,
maximum span length and span number to create the groups. Latin hypercube sampling
(LHS) is used to generate bridge samples. The methodology utilizes a constrained sampling
approach instead of randomly selected samples. Following this method, a statistical distribu-
tion of the structural attributes was taken into account during selection.
Most of the traditional studies grouped bridge classes to study representative structures of
the highway infrastructure in a region. These classifications are normally based in subjective
parameters and they consider as an important parameter the engineering judgment. In
countries where complete bridge data are available it is possible to evaluate the traditional
methodologies to group bridges versus new proposals. Recently, Mangalathu et al. [2016]
proposed a performance-based groping methodology using Analysis of Covariance. Bridge
classification was carried out based on the bearing type (elastomeric versus rocker bearings,
column cross section (circular versus rectangular), abutment configuration (abutment on
piles versus spread footing), abutment backfill (clay versus sand), interior bent type (single-
column versus multi-column), and superstructure type (reinforced or prestressed concrete).
Subsequently, Mangalathu et al. [2017a] reviewed HAZUS bridge classes and proposed a
classification based on bridges with statistical similar performance. They adopted different
attributes of box-girder bridges based on the column cross–section, abutment type, pier type,
span continuity, number of spans, and seismic design. The authors concluded that they
improved bridge classification of HAZUS including additional structural attributes.
In this study, 56 highway bridges representative of the most common concrete bridges in
Iran are selected [Mosleh, 2016c]. The bridges are single or multi-span simply supported or
continuous bridges. Figure 1 presents statistics of structural characteristics of the bridges.
Most of the concrete bridges have less than seven spans, column heights are in the range of
6–12 m and 18–20 m and multicolumn bents are the most common substructure. Based on
the analysis of Iran bridge inventory, Mosleh [2016c] found that most of the bridges in Iran
have skew angles less than 5°, therefore the effect of skew angle is not considered. This study
classifies the bridges based on span length, superstructure type, column height, material, and
rebar lap splice.
The information collected during different surveys includes the compressive strength of
concrete and the yield strength of longitudinal bars. In addition to concrete test, Schmidt
hammer test was used to estimate the compression strength of concrete. Material properties
of the bridge elements were determined by testing concrete cores and bars specimens.
Table 1 displays the statistical analysis of the material properties Mosleh [2016c].
 JOURNAL OF EARTHQUAKE ENGINEERING 1707

0.25 0.55
0.50
0.45
0.20
0.40
0.35

Frequency
0.15

Frequency
0.30
0.25
0.10 0.20
0.15

0.05 0.10
0.05
0.00
0.00
0-5 0 -15 -20 -25 -30 -35 re
2 3 4 5 6 7 8
More 5-1 10 15 20 25 30 Mo
Span number Skew angle (deg.)
0.50
0.24
0.22 0.45
0.20 0.40
0.18
0.35
0.16
Frequency

0.30

Frequency
0.14
0.12 0.25
0.10
0.20
0.08
0.15
0.06
0.04 0.10
0.02 0.05
0.00
0.00
s s 4-6 6-8 8-10 0-12 2-14 4-16 6-18 8-20 0-22 More 1 2 3 4 5 More
6
Le 1 1 1 1 1 2
Max. column height (m) Number of columns per bent

Figure 1 Statistical distributions of several structural attributes of selected highway concrete bridges.

Table 1 Statistical material properties of highway bridges in Iran (Mosleh [2016c])

f´c (col) MPa f´c (cap beam) MPa f´c (girder) MPa fsy (bar) MPa fsu(bar) MPa
Max 32 27 28 530 679
Min 19.5 24 16 252 344
Avg. 24.5 24.6 24.8 420 550

2.1. Span Length and Superstructure Type

Depending on the superstructure type, the bridges are classified into different categories
[Avsar et al., 2011; Choi et al., 2004; FEMA, 2003; Nielson and DesRoches, 2007b].
Figure 2 shows a 3D graph of the superstructure types and span lengths. The super-
structure types are: slabs, I girders and box girders. Slab bridges are usually short single-
span bridges supported on massive masonry abutments. Additionally, these bridges are
usually low-rise structures with low seismic vulnerability comparing to other two type of
structures. Furthermore, their rehabilitation time and cost in case of damage or failure are
smaller than the required time and cost for long-span bridges. Hence, the analyses exclude
the first type of superstructure. The other two types are simply supported superstructures.
The statistical analyses displayed in Figures 1 and 2 allow selecting the geometric para-
meters displayed in Table 2. Table 3 presents details of the selected RC bridges in terms of
column height, column cross-section, and longitudinal and transverse reinforcement.

2.2. Column Height

Avsar et al. [2011] classified bridge heights in the range of 4–10 m; Nielson [2005] selected
bridges with heights between 3 and 6 m. The histogram of maximum column height
(Figure 1) shows that pier heights of 4–12 m and 16–22 m are noticeably more frequent
comparing to other ranges of column heights. This study classifies the bridges in two
groups: short column bridges (SCB), with column height of 12 m or less, and high column
bridges (HCB) the others. Tables 4 and 5 show the structural attributes and geometric
characteristics for each of these categories.
1708 A. MOSLEH ET AL.

Figure 2 Distribution of superstructure type and span length.

Table 2 Structural attributes of the bridge samples in terms of span length (SL)
Abbreviation Deck type Max. span length (m) Longitudinal steel bar (%) Deck width (m) Seat length (cm)
20≤SL<30 I girder 20–24 0.91–1.2 12 95
SL≥30 I & Box girder 32–37 0.91–1.56 12–16 95

Table 3 Concrete reinforcing layout for each case study in terms of span length
Number of No. of columns Column section
Bridge classes Sample ID spans per bent Hcol (m) c (m) d (m) Longitudinal bar Transvers bar
20≤SL<30 1 15.6–24-24–15.6 3 6 1.1 4.5 30Φ22 Φ12@25
2 6@20 2 10.5 1.3 7 20Φ30 Φ18@20
3 7@20 2 16 1.4 7 20Φ30 Φ18@20
4 6@20 2 21 1.4 7 20Φ30 Φ18@20
SL≥30 1 6@32 3 18 1.4 6.4 20Φ30 Φ18@20
2 6@32 3 9 1.2 6.4 22Φ32 Φ14@5
3 6@35 3 8 1.2 4.0 26Φ26 Φ12@25
4 6@37 3 7 1.15 5.0 24Φ25 Φ12@25

Table 4 Structural attributes for the bridge samples in terms of column height
Column Column section Longitudinal Deck Span length
Abbreviation height (m) (circular) steel bar ratio (%) width (m) (m)
HCB 16–21 D = 1.4 0.91 12–16 20–32
SCB 6–10.5 D = 1.2–1.3 1.06–1.56 12–16 20–32
 JOURNAL OF EARTHQUAKE ENGINEERING 1709

Table 5 Geometric characteristics for each case study in terms of column height
Number No. of column Column section
Bridge classes Sample ID of spans per bent Hcol (m) c (m) d (m) Longitudinal bar Transvers bar
SCB 1 15.6–24-24–15.6 3 6 1.1 4.5 30Φ22 Φ12@25
2 6@32 3 9 1.2 6.4 22Φ32 Φ14@5
3 6@20 2 10.5 1.3 7.0 20Φ30 Φ18@20
HCB 1 7@20 2 16 1.4 7.0 20Φ30 Φ18@20
2 6@32 3 18 1.4 6.4 20Φ30 Φ18@20
3 6@20 2 21 1.4 7.0 20Φ30 Φ18@20

2.3. Material
Material strengths have effects on both the strength and stiffness of a bridge [Shinozuka et al.,
2000b]. Some studies use simulation methodologies to consider material uncertainties [Sudret
and Mai, ; Monteiro, 2016c]. To obtain the effect of material on fragility curves, three samples
of bridge component were collected. Based on Table 5, three-column bridge bent with column
heights of 10.5 m, 16 m, and 21 m were selected. Compressive strength of concrete ranges
between 20 and 30 MPa and steel yield strength of steel ranges between 300 and 400 MPa
(Table 1). Although the yield strength presented in Table 1 varies between 252 and 530 MPa,
most bridges have yield strength above 300 and below 400 MPa Based on statistical material
properties, three compressive strengths of concrete were selected: 20, 25, and 30 MPa.
Similarly, two yield strengths of the steel were considered: 300 and 400 MPa. Based on
three real cases, bridges with three column heights, 10.5, 16, and 21 m, were analyzed. For
each real case study, compressive strengths of 20, 25, and 30 MPa and yield strengths of 300
and 400 MPa were selected. The strength parameter combinations conducted to six analytical
models for each bridge. The combination of three bridges, three concrete strengths, and two
steel strengths generates 18 bridge samples to analyze the effect of material properties on
fragility curves.

2.4. Lap Splice of the Reinforcement

Hwang et al. [2001] proposed different limit states of damage in columns subjected to flexure.
According to that study, the presence of lap splice does not have influence on first and second
limit states; however, the third limit state (extensive) is highly dependent on the existence of
lap splice in the column. To consider the effect of column lap splices in fragility curves, Hwang
et al. [2001] proposed a methodology based on strain demands in plastic hinge regions. In
columns with lap splices, the column core starts disintegrating when the plastic hinge reaches
θp2 that corresponds to strain (εc) equal to 0.002. However, in columns without lap splices the
maximum strain to reach that limit state is equal to 0.004 [Hwang et al., 2001]. This study
assumes these strain values to assess damage limit states with and without lap splices.

3. Numerical Analysis of Representative Structures

Three-dimensional time history analyses were accomplished. The superstructure is com-
posed of cast-in-place reinforcement concrete slab over girders. The bent system and
abutments constitute the substructure of the bridge and elastomeric bearings are located
between the substructure and superstructure.
1710 A. MOSLEH ET AL.

Bridge superstructures are usually considered to perform well during earthquakes. The
reason is that they remain in the elastic range of behavior with very little or no nonlinear
displacement demands. The superstructure damages are most of the times related to insuffi-
cient seat length [Ramanathan, 2012]. Many studies consider the superstructure to be in the
elastic range of behavior [Mander et al., 2007; Mangalathu et al., 2016; Mangalathu et al.,
2017a; Mangalathu et al., 2017b]. Simply supported superstructures of medium span bridges
supported on flexible elastomeric bearings, move closely as a rigid body when subjected to
lateral loads and concentrate seismic demands in piers. Consequently, this study assumes the
superstructure remains in the elastic range of behavior.
The mass of the bridges is derived from the reinforced concrete girders, slab, diaphragms,
parapets, and weight of asphalt and sidewalks. Recent regulation codes [AASHTO LRFD,
2012; AASHTO LRFD, 2017] suggest including a fraction of live load in the bridge seismic
response calculations. In previous versions [AASHTO LRFD, 2007] it was not recommended
to include live loads. Wibowo et al. [2013] conducted an extensive study of the effect of live
load in the seismic response of bridges. They found that live loads might have in some cases
beneficial effects and in others can have the opposite effect, depending on several parameters
of the vehicles. In this study truck loads are not taken into account in the numerical models.
Frame elements with six degrees of freedom at each node are used to model the columns,
bent caps, and girders; the deck and diaphragms are modeled with shell finite elements. To
denote the mass distribution through the element length, the superstructure and the piers
are divided into a sufficient number of small elements.
The abutments and backfill soil were modeled as elastic springs in the longitudinal and
transverse directions according to the Caltrans recommendation [Caltrans, 2013].
Elastomeric bearings are located between the superstructure and substructure compo-
nents, without any dowel or connecting device. The lateral and vertical stiffness of the
elastomeric bearing, modeled as linear elastic springs, are [Priestley and Park, 1987]:

Kv ¼ G hA
2
(1)
Ks ¼ ð6GS
6GAKS
2 þ K Þh

where Kv and Ks are the shear and vertical stiffness of the bearings, respectively. G is the
shear modulus of rubber (1 MPa), A is the gross rubber area, and h is the total rubber
height. K is the rubber bulk modulus and S is the shape factor.
For the confined concrete, previous researchers developed different constitutive models [Kent
and Park, 1971; Bazant and Bhat, 1976; Sheikh and Uzumeri, 1980; Mander et al., 1988]. Some of
the proposed methods have limitations (e.g., circular or rectangular sections). The method
suggested by Mander et al. [1988] neglects tensile strength of concrete members and it is
applicable to all section shapes and all levels of confinement according to Eqs. (2)–(6).
Figure 3 displays the models for confined and unconfined concrete strength parameters
[Mander et al., 1988].
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi !
0 0
0 0 7:94fl 2fl
fcc ¼ fc 2:254 1 þ  0  1:254 (2)
fc0 fc
 JOURNAL OF EARTHQUAKE ENGINEERING 1711

Figure 3 Material models for confined and unconfined concrete.

 0 
fcc
εcc ¼ 0:002 1 þ 5 0  1 (3)
fc

1:4 ρs fyh εsu

εcu ¼ 0:004 þ (4)
fcc0

4 Ah
ρs ¼ (5)
D0 s
0 1
fl ¼ ke ρs fyh (6)
2
where f´cc and ɛcc are concrete stress and strain at peak stress, f´l is the effective lateral
confining stress and ɛcu is the ultimate compression strain respectively. fyh is the yield
strength of the transverse reinforcement, εsu is the steel strain at the maximum tensile
stress, f′cc is the compressive strength of the confined concrete, ρs is the volumetric ratio of
confining steel, Ah is the cross-sectional area of transverse reinforcement, D′ is the
diameter of the confined concrete core, ke is a confinement coefficient, and s is the
longitudinal spacing of hoops or spirals.
The model includes P-delta effects to consider the action of gravity loads on laterally
displaced columns. Newmark’s beta method is applied to solve the dynamic equations at
each time step of the analysis. The mass and stiffness proportional Rayleigh damping
coefficients are determined by considering the first two modal periods. Direct integration
and nonlinear time history analyses in two orthogonal directions are conducted to
evaluate the seismic vulnerability of the bridges. Multi-column bents in the longitudinal
direction behave as a cantilever structural system, and thus plastic hinges can only form at
the bottom part of the columns. However, in the transverse direction, the columns and
cap beam form a frame type system. In this case, plastic hinges may develop at both
column ends. Figure 4 shows a typical moment-curvature diagram and an elastic—
1712 A. MOSLEH ET AL.

Figure 4 Moment-curvature diagram of columns (HCB).

perfectly plastic idealization generated with the SAP2000 program (Computers and
Structures Inc, CSI, 2009].
The equivalent yield curvature (φy) corresponds to the relative displacement of the column
when the vertical reinforcing bars at the bottom of the column reach the yield point. φy is
obtained by extrapolating the line joining the origin and the point corresponding to the first
yielding point of a reinforcing bar up to the nominal moment capacity Mn. Mn is the bending
moment when εc = 0.005, where εc is the compressive strain of the concrete column [Priestley
and Park, 1987; Priestley et al., 1996]. φy ¼ MMy φy1 gives the curvature φy, where My and φy1
n

are the moment and curvature at first yielding of a vertical reinforcing bar.
Assuming a constant curvature along the length of the plastic hinge, the rotation angle
can be calculated by: θ ¼ φ  Lp . Different expressions exist to estimate the hinge length
(LP). In this study, the expressions proposed by Priestley [Priestley et al., 1996], given by
Lp ¼ 0:08L þ 0:022fye dbl  0:044fye dbl were used. fye is the yield strength of the reinforcing
bars and dbl is the diameter of the longitudinal reinforcing bars. The nonlinear behavior of
the columns is considered with a concentrated plasticity model by assigning plastic hinges
at both column ends, as recommended in Caltrans code [Caltrans, 2013]. Outside the
plastic hinge length, the behavior of the column is assumed to be linear. Figure 5 shows
the analytical model of one of the bridges created with SAP2000 program [Computers and
Structures Inc, CSI, 2009]. Link elements, modeled as fixed springs, are used to locate the
exact position of bearings.

4. Ground Motion Selection

One of the components needed to develop the fragility curves is a family of earthquake
ground motions. The intensity measure selected has important effect on fragility curves.
There are different methods of selecting the appropriate intensity measure for a fragility
analysis. Previous researchers used common intensity measurements (IMs) such as: PGA,
peak ground velocity (PGV) or peak ground displacement (PGD). The selection of an IM
 JOURNAL OF EARTHQUAKE ENGINEERING 1713

Figure 5 Three-dimensional finite-element model.

is the topic of several studies [Kwon and Elnashai, 2006; Baker and Cornell, 2006; Padgett
et al., 2008; Bradley, 2010a; Buratti and Tavano, 2014; Bradley et al., 2015; Zelaschi et al.,
2015; Borzi et al., 2015; Kwong and Chopra, 2015; Franchin et al., 2016; Zelaschi et al.
2016, Zelaschi et al., 2017; Monteiro et al., 2017]. PGA is one of the most common IMs in
fragility analysis of the RC bridges [Avsar, 2009; Seo and Linzell, 2013; Park and
Towashiraporn, 2014; Jeon et al., 2017]. Baker [2011] proposed the use of a Conditional
Mean Spectrum (CMS) that match a specific spectral acceleration (Sa) to select a family of
ground motions. The procedure requires selecting a Sa value, for a target period, to
construct the CMS that involves the use of predicted mean and standard deviation of
logarithmic spectral acceleration (usually calculated from attenuation laws), to select the
range periods of interest and finally choose a family of seismic records that match the
CMS. Baker [2011] shows that the use of CMS to select ground motions is a less
conservative procedure than the use of Uniform Hazard Spectrum. Baker and Lee
[2017] suggested an algorithm for selecting seismic records based in matching a family
of ground motions to a response spectrum distribution, which it could be a CMS. The
authors remark that their proposal could be very helpful tool to select a suite of seismic
records, especially when large ground motion databases are available.
In order to generate reliable fragility curves, appropriate earthquake records
should be selected. The records should reflect a wide range of seismic hazard levels.
In this study, real earthquake ground motions that show the seismic potential of the
investigated region are considered. Most of the recorded ground motions in Iran,
with high potential of destruction, are produced in reverse and strike-slip faults. All
bridges were assumed located on hard soil, thus soil flexibility at the bridge founda-
tions are not taken into account in the analytical models. 104 earthquake ground
1714 A. MOSLEH ET AL.

motions (60 for reverse fault and 44 for strike-slip fault) satisfying the following
conditions were initially selected:

– Earthquake ground motions from strike-slip and reverse faults recorded in Iran and
other regions
– Accelerograms recorded on soil sites having Vs ≥ 360 m/s
– Ground motions having PGA ≥ 0.05 g
– Earthquakes with moment magnitude Mw ≥ 5.2
– Seismic records from far-fault earthquakes.

Earthquake ground motions were downloaded from the strong motion databases of
PEER (http://peer.berkeley.edu/smcat/) and COSMOS (http://db.cosmoseq.org/scripts/
default.plx). To select and reduce the number of earthquake records, the acceleration
spectral intensity (ASI) was used. The area under the 5% damped response spectrum,
defined by Eq. (7), between the boundary periods, defined by Ti and Tf, is defined as
acceleration spectrum intensity (ASI); Ti and Tf are the initial and final periods used to
calculate ASI. Several authors found ASI as a better indicator to predict seismic demands
in structures [Bradley, 2010b; Avsar et al., 2011; Von Thun et al., 1998; Yakut and
Yılmaz, 2008]
Tf
ASI ¼ ò SAðT; ÞdT (7)
Ti

The bridges have fundamental period values in the range of 0.99–2.06 s. Based on this
range, Ti and Tf periods of 0.9 and 2.12 s were selected to determine ASI measure.
Nonlinear response history analysis is particularly time consuming. Hence, 20 of the
above-mentioned 104 records were selected for each type of seismic fault, based on the
highest ASIs. Therefore, a total of 40 unscaled ground motion records from Iran and other
regions with similar seismic fault types were selected to represent the record-to-record
variability (20 records for reverse and 20 records for strike-slip faults respectively).
The response spectrum of each ground motion is determined by taking the SRSS of the
response spectrum of the two horizontal ground motion components. Figure 6 presents
the response spectra for 5% damping ratio of the earthquake ground motions; thicker lines
represent the mean values of the response spectra: (a) strike-slip and (b) reverse faults
records. Mean values show that earthquakes from reverse faults present a larger amplitude

Figure 6 Response spectra of the selected ground motions for (a): reverse and (b): strike slip faults.
 JOURNAL OF EARTHQUAKE ENGINEERING 1715

Table 6 Some important parameters of the selected earthquake ground motions (reverse)
Earthquake Year Mw R (Km) PGA (g) ASI (g*s)
Chi-Chi-0.1125g 1999 7.62 101.62 0.1125 0.491401
Chi-Chi-0.1348g 1999 7.62 63.29 0.1348 0.579302
Chi-Chi-0.1748g 1999 7.62 37.83 0.1748 0.814761
Chi-Chi-0.2058g 1999 7.62 47.86 0.2058 0.967536
Chi-Chi-0.2595g 1999 7.62 39.70 0.2595 0.949184
Chi-Chi-0.3643g 1999 7.62 86.39 0.3643 1.118649
Chi-Chi-0.5283g 1999 7.62 95.70 0.5283 1.066492
Chi-Chi-0.0823g 1999 7.62 59.80 0.0823 0.393901
Northridge-0.2148g 1994 6.69 14.92 0.2148 0.760153
Northridge-0.3908g 1994 6.69 18.62 0.3908 0.647099
Northridge-0.4673g 1994 6.69 39.39 0.4673 0.567823
Northridge-0.4898 1994 6.69 40.68 0.4898 1.692820
Northridge-0.5102g 1994 6.69 16.27 0.5102 0.890333
Northridge-0.5908g 1994 6.69 22.45 0.5908 0.976386
Sanfernando-0.2994g 1971 6.61 25.36 0.2994 0.618939
Whittier Narrows-0.3408g 1987 5.99 21.26 0.3408 0.516234
Capemendocino-0.1668g 1992 7.01 53.34 0.1668 0.568378
Capemendocino-0.4244g 1992 7.01 22.64 0.4244 1.278070
Tabas-0.3505g 1978 7.40 20.63 0.3505 0.766511
Tabas-0.8128g 1978 7.40 55.24 0.8128 2.280662

Table 7 Some important parameters of the selected earthquake ground motions (strike-slip)
Earthquake Year Mw R (Km) PGA (g) ASI (g*s)
Morgan Hill-0.0983g 1984 6.19 30.05 0.0983 0.284824
Parkfield-0.469g 2004 6.00 14.50 0.4690 0.447411
Parkfield-0.602g 2004 6.00 14.80 0.6020 0.624203
Manjil-0.5051g 1990 7.40 40.43 0.5051 0.986322
Morgan Hill-0.2814g 1984 6.19 36.34 0.2814 0.789886
Morgan Hill-0.3426g 1984 6.19 16.67 0.3426 0.847620
Kobe-0.7105g 1995 6.90 18.27 0.7105 2.794062
Imperial Valley-0.176g 1979 6.53 24.82 0.1760 0.579679
Duzce-0.1445g 1999 7.14 27.74 0.1445 0.356142
Victoria-0.5722g 1980 6.33 33.73 0.5722 0.997996
Parkfield-0.2934g 1966 6.19 40.26 0.2934 0.480003
Landers-0.1407g 1992 7.28 27.33 0.1407 0.560219
Landers-0.3733g 1992 7.28 82.12 0.3733 1.100193
Kobe-0.0765g 1995 6.90 123.33 0.0765 0.166192
Duzce-0.2101g 1999 7.14 29.27 0.2101 0.479420
Duzce-0.7367g 1999 7.14 24.05 0.7367 0.806259
Parkfield-0.271g 2004 6.00 32.10 0.2710 1.179797
Imperial Valley-0.1661g 1979 6.53 48.62 0.1661 0.329791
Duzce-0.1174g 1999 7.14 31.56 0.1174 0.297492
Kojaeli-0.1387g 1999 7.51 77.63 0.1387 0.682200

(at period = 0.24s), than strike slip faults (at 0.27 s). Tables 6 and 7 report some of the
important features of the selected earthquakes and the associated intensity measure
parameters of the ground motions.

5. Damage Limit States for Fragility Analysis

Performance evaluation of bridges during past earthquakes revealed that the overall response
is highly dependent to the columns behavior. The limit states of the bridges may be defined in
terms of ductility demands. Ductility demands can be expressed in terms of displacement,
rotation, or curvature [Mosleh et al., 2015]. Zhu et al. [2007] analyzed a 125-column database
1716 A. MOSLEH ET AL.

and showed that aspect ratios (length/depth) smaller than two, conducted to shear failures
whereas columns with an aspect ratio greater than four, failed in flexure mode. Ying and Jin-
Xin [2018] found similar results. All the (length/depth) ratios in this study are greater than 4;
thus, it was assumed that the predominant failure mode is bending and therefore the shear
behavior was ignored. Eqs. (8)–(14) define the displacement ductility.
Δyi
μci ¼ (8)
Δy1
1
Δy1 ¼ φy1 L2 (9)
3

1 φy L
2
Δ2 Δy
μc2 ¼ ¼ ¼ (10)
Δy1 Δy1 3 Δy1

Δ3
μc3 ¼ (11)
Δy1

Lp
Δ3 ¼ Δ2 þ θp ðL  Þ (12)
2
θp ¼ ðφ3  φy ÞLp (13)

μc4 ¼ μc3 þ 3 (14)

where µci is the ductility demand at the ith damage state, Δyi is the relative displacement of
the limit state i in the top of a column, and Δy1 is the relative displacement of a column
when the longitudinal reinforcing bars reach the first yield. L is the distance from the
plastic hinge to the point of contra-flexure.
Four damage limit states (µc1, µc2, µc3, and µc4) are identified in this research: slight
(LS1), moderate (LS2), extensive (LS3), and collapse (LS4). These limit states are similar
to those proposed by [Hwang et al., 2001]. φy1 is the curvature corresponding to the
relative displacement of a column when the vertical reinforcing bars at the bottom of
the column reach the first yield (Figure 4). Hence, µc1 denotes the first limit state
corresponding to a first yield displacement ductility ratio equal to 1. The second
damage state, µc2, represents the yield displacement ductility ratio. φy is the curvature
associated to the relative displacement of a column when the vertical reinforcing bars at
the bottom of the column reach the yield strength (Figure 4). The displacement
ductility of the third damage state (µc3) corresponds to εc = 0.002 or εc = 0.004 for
the columns with or without lap splices, respectively, and εc is the compressive strain in
the column. θp and Lp are the rotation and the plastic hinge length, respectively. The
plastic hinge length can be calculated with Equation which proposed by [Priestley et al.,
1996] as explained in section 3.

6. Development of Fragility Curves

Fragility curves relate strong ground motion severity to the probability of reaching or
exceeding a certain limit state [Razzaghi and Eshghi, 2015]. The main objective of this
 JOURNAL OF EARTHQUAKE ENGINEERING 1717

study is to develop fragility curves for ordinary highway RC bridges piers designed and
constructed according to pre-1990 code regulations. For this purpose, the following
specific methodology is proposed [Hwang et al., 2001]:

(1) Categorization of bridges based on statistical analyses of their primary structural

characteristics.
(2) Selection of earthquake ground motions from important seismic fault types.
(3) Definition of damage limit states for each bridge classification. In this study,
displacement ductility of columns is selected as engineering demand parameter.
(4) Creation of numerical 3D bridge classes and seismic analysis accounting for
uncertainty in bridge properties. Bridge uncertainties for the selected case studies
include column heights, lap splice and span lengths. Additionally, the influence of
material properties (e.g. the compressive strength of concrete and the yield strength
of steel) was considered.
(5) Determination of performance levels of each bridge by comparing demands
obtained from nonlinear time history analysis with the damage limits as described
in step 3.
(6) Evaluation of the performance level of each bridge, for a given ground motion
record, and determination of the damage limit state.
(7) Plotting calculated probabilities versus PGAs for each damage state.
(8) Performing regression analysis to fit an appropriate lognormal curve to the plotted
points. (Figure 7). In this graph the x-axis is the seismic intensity measure of the
ground motion and y-axis is the probability of reaching or exceeding certain
damage limit state.

A thorough discussion of the underlying concepts and step-by-step procedure used to

obtain fragility curves is given in the following sections.
Each bridge was subjected to two orthogonal horizontal components of the ground
motions. The maximum absolute ductility demand, defined as the maximum value of
the ductility demand in the longitudinal and transverse direction, for each seismic
record was determined. The number of bridges that reach or exceed a specified damage
limit state is obtained by subjecting the models to the seismic records with a certain

Figure 7 Schematic representation of a fragility curve.

1718 A. MOSLEH ET AL.

intensity measure (PGA). The intensity measure is defined as the root of the sum of the
squares (SRSS) of the two horizontal components of the ground motions. For each
PGA, the probability of exceeding a limit state is calculated as the ratio of the number
of bridges which reach or exceed the specified limit state to the total number of sample
bridges.
The distribution of the probability of exceedance as a function of the selected intensity
measure is shown schematically in Figure 7. Then, an expression is used to characterize
the sharply varying points to achieve smooth fragility curves for a specific damage limit
state and bridge model, which is the best fit for the points of the probability of exceedance.
Several researches, and this study as well, adopt the lognormal distribution to obtain
fragility curves [Banerjee and Shinozuka, 2007; Elnashai et al., 2004; FEMA, 2003; Karim
and Yamazaki, 2003; Nielson and DesRoches, 2007b]. Each fragility function depends on a
median value and a dispersion parameter (standard deviation of the log values) of ground
motion. To select the parameters of the lognormal probability density function, the least
squares method is applied. The coefficient of correlation (R2) determines how well the
fitted curve relates to the data of the fragility functions.

7. Fragility Curves
In the following sections, the effect of the span length (SL), column height, presence of lap
splices in columns and material properties is analyzed. Utilizing the procedure described
above, the fragility curves are calculated.

7.1. Span Length

Figure 8 shows fragility curves of the bridges in terms of span length and superstructure
type subjected to reverse and strike-slip fault movements. As an example of the influence
of the superstructure type and seismic source on the probability of reaching the limit
states, Figure 8 (a) shows that, for PGA = 0.4g, the probabilities of reaching or exceeding
LS1, LS2, and LS3 of the I girder superstructure type subjected to the reverse fault records
are 95, 85, and 42%, respectively. In contrast, Figure 8 (c) indicates that the probabilities of
reaching or exceeding LS1, LS2, LS3, and LS4 for the box-girder superstructure type are
90, 82, 66, and 11%, respectively. The comparison between these two types of super-
structures shows similar exceedance probabilities of the limit states LS1 and LS2. Bridges
with span lengths larger than 30 m are more vulnerable than the bridges with span lengths
in the range of 20–30 m. Figure 8 (b) and (d) show similar differences between the two
superstructure types when subjected to accelerograms of the strike-slip fault.
The bridges subjected to reverse fault records were more vulnerable than those
subjected to the strike-slip fault signals. Table 8 presents fragility curve parameters
for different damage limit states in terms of the span length. In general, the results
indicate that, to some extent, bridges are more susceptible to reverse fault ground
motions.
There is not a general trend in data dispersion and in most of the cases the standard
deviation is higher than 0.5. Furthermore, R2 does not show clear tendency on span length
or damage limit states.
 JOURNAL OF EARTHQUAKE ENGINEERING 1719

Figure 8 Fragility curves for different damage limit states (PGA) in terms of span length (SL) (a): Reverse &
20 ≤ SL < 30, (b): Strike slip & SL 20 ≤ SL < 30, (c): Reverse & SL ≥ 30, (d): Strike slip & SL ≥ 30.

Table 8 Fragility curve parameters for different damage limit states in terms of span length
Slight (LS1) Moderate (LS2) Extensive (LS3) Collapse (LS4)
Median Disp. R2 Median Disp. R2 Median Disp. R2 Median Disp. R2
20 ≤ SL < 30
reverse −2.13 0.73 0.68 −1.69 0.73 0.60 −0.89 0.52 0.41 −0.25 0.16 0.43
strike slip −1.38 0.77 0.61 −1.12 0.69 0.44 −0.65 0.43 0.80 −0.11 0.21 0.42
SL≥30
reverse −1.75 0.67 0.54 −1.53 0.70 0.61 −1.22 0.75 0.50 −0.27 0.55 0.44
strike slip −1.32 0.78 0.40 −1.12 0.73 0.59 −0.62 0.47 0.61 −0.13 0.64 0.45

7.2. Column Height

Figures 9 and 10 present fragility curves of the bridges in terms of column height
subjected to reverse and strike-slip fault records using PGA as an intensity measure.
Figure 11 illustrates the four damage limit states to compare the effect of different seismic
sources on the fragility curves. HCB bridges are more vulnerable to seismic effects than
SCB models. One reason for this behavior is the insufficient longitudinal steel ratio and
the larger displacement demands of the HCB bridges (as shown in Table 3 and Table 5).
The longitudinal steel ratio has a significant effect on the moment strength and ductility
capacity of the column Jara et al. [Jara et al., 2013]. The fragility curves show that both
bridge heights are more vulnerable to reverse fault records (Table 9).
1720 A. MOSLEH ET AL.

Figure 9 Fragility curves for different damage limit states (PGA) in terms of column height (SCB)
without lap splice, (a): Reverse, (b): Strike slip.

Figure 10 Fragility curves for different damage limit states (PGA) in terms of column height (HCB)
without lap splice, (a): Reverse, (b): Strike slip.

Figure 11 Fragility curves subjected to reverse and strike slip fault for different damage limit states
(PGA) in terms of column height without lap splice, (a): SCB, (b): HCB.
 JOURNAL OF EARTHQUAKE ENGINEERING 1721

TABLE 9 Fragility curve parameters for different damage limit states in terms of column height
Slight (LS1) Moderate (LS2) Extensive (LS3) Collapse (LS4)
Median Disp. R2 Median Disp. R2 Median Disp. R2 Median Disp. R2
SCB
reverse −1.42 0.44 0.66 −1.18 0.45 0.69 −0.82 0.28 0.72 −0.16 0.25 0.68
strike slip −1.19 0.62 0.53 −0.97 0.50 0.54 −0.71 0.35 0.69 −0.11 0.2 0.71
HCB
reverse −2.36 0.88 0.45 −1.60 0.51 0.76 −0.87 0.47 0.70 −0.13 0.26 0.51
strike slip −1.63 0.69 0.53 −1.23 0.45 0.41 −0.51 0.27 0.44 −0.12 0.17 0.70

Figure 6 shows that the maximum amplitude of the mean SA response spectrum of a
reverse fault is greater than the mean SA value of a strike-slip fault. Therefore, the bridges
subjected to the reverse fault records displayed larger demands than the bridges subjected to
the strike-slip fault accelerograms. In addition, these results reveal that the failure probability
of the Iranian bridges for PGA of less than 0.4g is low. This outcome is consistent with the
bridge responses observed during past earthquakes that occurred in Manjil and Bam in Iran
[Astaneh-Asl, 1994; Manafpour, 2008; Zahrai and Heidarzadeh, 2007].
In these cases, the standard deviations of fragility curves corresponding to SCB models, were
in the range of 0.2–0.62. However, for HCB models this range expanded to 0.17–0.88. In other
words, the data dispersion seems to be susceptible to the column height. R2 does not noticeably
changes in SCB and HCB fragility curves.

7.3. Lap Splice

Figures 12 and 13 show fragility curves for bridge columns with and without a lap splices
in (HCB) models subjected to both groups of ground motions. As indicated in the figures
lap splice has a considerable effect on the fragility curves. Columns with lap splices are
more seismically vulnerable comparing to other ones. Hence, lap splices of longitudinal
reinforcements should not be used in critical locations of ductile elements [AASHTO,
2017; AASHTO, 2012; Caltrans, 2013; FHWA, 2006].
These results show that extensive and consequently collapse-limit states are completely
susceptible to the presence of column lap splices. However, it is a common practice in old

Figure 12 Fragility curves for different damage limit state s (PGA) in terms of column height (HCB) with
lap splice, (a): Reverse, (b): Strike slip.
1722 A. MOSLEH ET AL.

Figure 13 Fragility curves for different damage limit states (PGA) as a function of the presence or not of
lap splice in columns for HCB, (a): LS3-reverse fault, (b): LS3-strike slip fault, (c): LS4-reverse fault, (d):
LS4-strike slip fault.

bridges to have lap splices near the base of columns. Figure 13 displays the fragility curves
of the bridges subjected to the seismic records of both seismic sources; the graphs present
only the LS3 and LS4 limit states because the difference between LS1 and LS2 is negligible.
As indicated in Figure 13, for PGA = 0.4g (for example), the probabilities of reaching or
exceeding extensive damages (LS3) with and without lap splices are 48 and 80%, respectively,
for a reverse fault. However, these probabilities change to 8 and 58% for a strike-slip source,
respectively. In other words, as it presented in Table 10, the fragility curves indicate that the
selected typical pier pre-1990 RC bridges are sensitive to the seismic source and the bridges
are more vulnerable to the effect of reverse fault ground motions.
Dispersion and coefficient of determination in all cases are similar, showing, for each
damage limit state, to be independent of lap spice variable.

Table 10 Fragility curve parameters for different damage limit states in terms of present of lap splice for
HCB classes
Slight (LS1) Moderate (LS2) Extensive (LS3) Collapse (LS4)
Median Disp. R2 Median Disp. R2 Median Disp. R2 Median Disp. R2
HCB (without lap splice)
reverse −2.36 0.88 0.45 −1.60 0.51 0.76 −0.87 0.47 0.70 −0.13 0.26 0.51
strike slip −1.63 0.69 0.53 −1.23 0.45 0.41 −0.51 0.27 0.44 −0.12 0.17 0.70
HCB (with lap splice)
reverse −2.36 0.88 0.44 −1.60 0.51 0.77 −1.32 0.48 0.77 −0.35 0.18 0.52
strike slip −1.63 0.69 0.52 −1.22 0.45 0.41 −0.96 0.36 0.40 −0.17 0.28 0.44
 JOURNAL OF EARTHQUAKE ENGINEERING 1723

20-S 300-S
25-S 400-S
1.0 30-S 1.0 300-M
20-M 400-M
25-M 300-E
30-M 400-E
0.8 0.8
20-E 300-C

Probability of Exceedance
25-E
Probability of Exceedance

400-C
30-E
0.6 20-C 0.6
25-C
30-C
0.4 0.4

0.2 0.2

0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
PGA (g) PGA (g)

(a) (b)

Figure 14 Fragility curves based on different (a): compressive concrete strength, (b): steel yield strength.

7.4. Material
Figure 14 shows the fragility curves for bridges with column heights varying between 10.5 and
21 m in terms of material properties: concrete compressive strength and steel yield strength.
The compressive concrete strength is in the range of 20–30 MPa and the yield strength of steel
is between 300 and 400 MPa. The bridge response is significantly less sensitive to the concrete
compressive strength compared to the steel yield strength. This outcome is consistent with the
bridge responses observed by previous researches [Pan et al., 2010). Table 11 presents fragility
curve parameters for the material properties in terms of compressive concrete strength and
steel yield strength.
Data dispersion is clearly dependent on damage limit states when material variability is
included. Slight damage presents more disperse results than collapse limit state, which is a
logical result considering the less variation of steel and concrete strength when the
columns are in regions of high local demands. In most of the cases the coefficient R2 is
greater than 0.5. After analyzing the influence of seismic source, span length, pier height,
lap splices, and material strength on fragility curves for several damage limit states, the
following general results are obtained.

(a) The reverse seismic source produces consistently higher bridge vulnerability than
the vulnerability of the bridges subjected to seismic records from strike slip fault.
(b) Bridge vulnerability rises with the increase of span length and column height.

Table 11 Fragility curve parameters for the material properties in terms of compressive concrete
strength and steel yield strength
Slight (LS1) Moderate (LS2) Extensive (LS3) Collapse (LS4)
compressive concrete
strength (MPa) Median Disp. R2 Median Disp. R2 Median Disp. R2 Median Disp. R2
Compressive concrete strength
20 (MPa) −1.74 0.71 0.53 −1.2 0.54 0.51 −0.49 0.46 0.51 −0.15 0.25 0.86
25 (MPa) −1.60 0.71 0.45 −1.13 0.63 0.53 −0.46 0.43 0.53 −0.11 0.24 0.85
30 (MPa) −1.46 0.55 0.46 −0.96 0.54 0.55 −0.50 0.41 0.64 −0.08 0.22 0.48
Steel yield strength
300 (MPa) −2.08 0.83 0.53 −1.38 0.63 0.59 −0.77 0.37 0.77 −0.29 0.25 0.91
400 (MPa) −1.66 0.82 0.43 −0.96 0.46 0.64 −0.48 0.33 0.56 −0.16 0.16 0.92
1724 A. MOSLEH ET AL.

(c) Among the studied parameters, lap splice had the most important effect on the
expected bridge damages. For example for a PGA = 0.6 g, the probabilities of reaching
extensive damage limit state were of 0.50–0.80, 0.78–0.82, 0.78–0.88 and 0.80–0.96 for
changes in material properties, span length, bridge height and lap splices, respectively.

8. Conclusions
This research aims to study the seismic vulnerability of existing multi-column RC bridges
located in Iran, by developing seismic fragility curves for representative structures of the design
and construction practices. Based on the results, the following conclusions can be made:

● The results showed that column height, span length, and lap splices in piers had a
significant effect on the seismic vulnerability of the bridges. Meanwhile material
properties, particularly concrete compressive strength, did not have a remarkable
impact on the seismic vulnerability of the structures.
● Seismic records must be carefully selected. The study showed that seismic source
could play considerable role on the seismic vulnerability of bridges.
● Bridges built in the 80’s with lap splices in columns showed more vulnerability when
subjected to reverse fault records than those subjected to strike-slip fault ground
motions.
● Based on the results of this study, Pre-1990 RC bridges with lap splices in columns
are considerably vulnerable. Hence, a retrofitting priority should be given to them, in
order to reduce the seismic risk of highway transportation system.
● In terms of the span length, the probabilities of reaching LS1 and LS2 limit states
were similar for the two types of superstructures analyzed. However, for LS3 & LS4
limit states, the bridges with span lengths larger than 30 m were more seismically
vulnerable.
● The developed fragility curves can be the basis of loss estimation models as well as
the framework of retrofit prioritization strategies.

References
AASHTO. [2012] AASHTO LRFD Bridge Design Specifications, American Association of State
Highway and Transportation Officials, Washington, D.C.
AASHTO. [2017] AASHTO LRFD Bridge Design Specifications, 8th Ed, American Association of State
Highway and Transportation Officials, Washington, D.C. ISBN Number: 978-1-56051-654-5.
AASHTO LRFD. [2007] Standard Specifications for Highway Bridges, Subcommittee for Seismic
Effects on Bridges, Imbsen Consulting, Imbsen.
Astaneh-Asl, A. [1994] “Lessons of the 1990 manjil-iran erathquake,” Paper presented at the Tenth
world conference, Balkema, Rotterdam.
ATC. [1985] “Earthquake damage evaluation data for California,” Report No. ATC-13. Applied
Technology Council, Redwood City, California.
ATC. [1991] “Seismic vulnerability and impact of disruption of lifelines in the coterminous United
States,” Report No. ATC-25. Applied Technology Council, Redwood City, CA.
Avsar, O. [2009]. “Fragility based seismic vulnerability assessment of ordinary highway bridges in
Turkey,” (Doctor of Philosophy), Middle East Technical University, Ankara, Turkey.
 JOURNAL OF EARTHQUAKE ENGINEERING 1725

Avsar, O., Yakut, A. and Caner, A. [2011] “Analytical fragility curves for ordinary highway bridges
in Turkey,” Earthquake Spectra 27(4), 971–996.
Baker, J. W. and Cornell, C. A. [2006] Vector-Valued Ground Motion Intensity Measures for
Probabilistic Seismic Demand Analysis, Pacific Earthquake Engineering Research Center:
Department of Civil and Environmental Engineering, Stanford University, Stanford, California.
Baker, J. W. [2011] “Ground motion selection for performance-based engineering, and the
Conditional Mean Spectrum as a selection tool”, Proceedings of the Tenth Pacific Conference
on Earthquake Engineering Building an Earthquake-Resilient Pacific 6–8 November 2015,
Sydney, Australia
Baker, J. W. and Lee, C. [2017] “An improved algorithm for selecting ground motions to match a
conditional spectrum,” Journal of Earthquake Engineering 22,1–16.
Banerjee, S. and Shinozuka, M. [2007] “Nonlinear static procedure for seismic vulnerability assess-
ment of bridges,” Computer-Aided Civil and Infrastructure 22, 293–305.
Basoz, N., Kiremidjian, A., King, S. A. and Law, K. H. [1999] “Statistical analysis of bridge damage
data from the 1994 Northridge, CA, earthquake,” Earthquake Spectra 15(1), 25–54.
Basoz, N. and Kiremidjian, A. S. [1997] “Evaluation of bridge damage data from the Loma Prieta and
Northridge CA earthquakes (Report No. MCEER-98-0004),” Retrieved from. University of Buffalo,
Buffalo, New York.
Bazant, Z. P. and Bhat, P. D. [1976] “Endochronic theory of inelasticity and failure of concrete,”
Journal Engineering Mechanisms Division 102, 701–722.
Billah, A. H. M. M., Alam, M. S. and Bhuiyan, A. R. [2013] “Fragility analysis of retrofitted multi-
column bridge bent subjected to near fault and far field ground motion,” ASCE Journal of Bridge
Engineering 18(10), 992–1004.
Borzi, B., Ceresa, P., Franchin, P., Noto, F., Calvi, G. M. and Pinto, P. E. [2015] “Seismic
vulnerability of the Italian roadway bridge stock,” Earthquake Spectra 31(4), 2137–2161.
Bradley, B. A. [2010a] “A generalized conditional intensity measure approach and holistic ground-
motion selection,” Earthquake Engineering and Structural Dynamics 39(12), 1321–1342.
doi:10.1002/eqe.995.
Bradley, B. A. [2010b] “Site-specific and spatially distributed ground-motion prediction of accel-
eration spectrum intensity,” Bulletin of the Seismological Society of America 100(2), 792–801.
doi:10.1785/0120090157.
Bradley, B. A., Burks, L. S. and Baker, J. W. [2015] “Ground motion selection for simulation-based
seismic hazard and structural reliability assessment,” Earthquake Engineering and Structural
Dynamics 44(13), 2321–2340.
Buratti, N. and Tavano, M. [2014] “Dynamic buckling and seismic fragility of anchored steel tanks
by the added mass method,” Earthquake Engng Structureal Dynamics 43, 1–21.
Caltrans. [2013] Seismic Design Criteria, Version 1.7.Sacramento, California Department of
Transportation, CA.
Choi, E., DesRoches, R. and Nielson, B. [2004] “Seismic fragility of typical bridges in moderate
seismic zones,” Engineering Structures 26(2), 187–199. doi:10.1016/j.engstruct.2003.09.006.
Computers and Structures Inc., CSI. Sap2000 V-14 [2009] Integrated finite element analysis and
design of structures basic analysis reference manual, Berkeley, California.
Elnashai, A., Borzi, B. and Vlachos, S. [2004] “Deformation-based vulnerability functions for RC
bridges,” Structural Engineering and Mechanics 17, 215–244.
Eshghi, S. and Ahari, M. N. [2005] “Performance of transportation systems in the 2003 Bam, Iran,
earthquake,” Earthquake Spectra 21(S1), 455–468.
FEMA. [2003] HAZUS-MH MR1: Technical Manual, Vol. Earthquake Model, Federal Emergency
Management Agency, Washington DC.
FHWA. [1988] “Federal highway administration,recording and coding guide for the structure inven-
tory and appraisal of the nation’s bridges,” Rep. No. FHWA-ED-89-044, Ofc. of Engrg., Bridge
Division, Washington, DC.
FHWA. [2006] “Seismic retrofitting manual for highway structures federal highway administration,”
Report No. FHWA-HRT-06-032, Federal Highway Administration Turner-Fairbank Highway
Research Center, Washington, DC.
1726 A. MOSLEH ET AL.

Franchin, P., Lupoi, A., Noto, F. and Tesfamariam, S. [2016] “Seismic fragility of reinforced
concrete girder bridges using Bayesian belief network,” Earthquake Engineering and Structural
Dynamics 45(1), 29–44.
Ghosh, J., Padgett, J. E. and Dueñas-Osorio, L. [2013] “Surrogate modeling and failure surface
visualization for efficient seismic vulnerability assessment of highway bridges,” Probabilistic
Engineering Mechanics 34, 189–199.
HAZUS. [1997] Technical Manual, Federal Emergency Management Agency, Washington, DC.
HAZUS. [2012] MH 2.1. Multi-Hazard Loss Estimation Methodology: Earthquake Model HAZUS-
MH 2.1 Technical Manual, Federal Emergency Management Agency, Washington, DC.
Hsu, Y. T. and Fu, C. C. [2004] “Seismic effect on highway bridges in Chi Chi earthquake,” Journal
of Performance of Constructed Facilities 18(1), 47–53.
Hwang, H., Jernigan, J. B. and Lin, Y. W. [2000] “Evaluation of seismic damage to Memphis bridges
and highway systems,” ASCE, Journal of Bridge Engineering 5, 322–330.
Hwang, H., Liu, J. and Chiu, Y. [2001] Seismic fragility analysis of highway bridges, Center for
Earthquake Research and Information, The University of Memphis, Memphis, Tennessee.
(MAEC RR-4 Project).
Jara, J. M., López, M. G., Olmos, B. A. and Jara, M. [2015] “Engineering demand functions for RC
medium length span bridges,” Bulletin of Earthquake Engineering 13, 679–702.
Jara, M., Jara, J. M. and Olmos, B. A. [2013] “Seismic energy dissipation and local concentration of
damage in bridge bents,” Structure and Infrastructure Engineering 9(8), 794–805. doi:10.1080/
15732479.2011.615330.
Jeon, J. S., Mangalathu, S., Song, J. and DesRoches, R. [2017] “Parameterized seismic fragility curves
for curved multi-frame concrete box-girder bridges using Bayesian parameter estimation,”
Journal of Earthquake Engineering. doi:10.1080/13632469.2017.1342291.
Karim, K. R. and Yamazaki, F. [2003] “A simplified method of constructing fragility curves for
highway bridges,” Earthquake Engineering and Structural Dynamics 32, 1603–1626.
Kawashima, K. [2002] “Damage of bridges resulting from fault rupture in the 1999 Kocaeli and
Duzce, Turkey earthquakes and the 1999 Chi-Chi, Taiwan earthquake,” Structural Engineering
and Earthquake Engineering, JSCE 19(2), 179s–197s.
Kawashima, K. and Unjoh, S. [1997] “The damage of highway bridges in the 1995 Hyogo-Ken Nanbu
earthquake and its impact on Japanese seismic design,” Journal of Earthquake Engineering 1, 504–541.
Kent, D. C. and Park, R. [1971] “Flexural members with confined concrete,” Journal of the
Structural Division 97, 1969–1990.
Kwon, O. S. and Elnashai, A. S. [2006] “The effect of material and ground motion uncertainty on
the seismic vulnerability curves of RC structure,” Engineering Structures 28, 289–303.
Kwong, N. S. and Chopra, A. K. [2015] “Evaluation of the exact conditional spectrum and general-
ized conditional intensity measure methods for ground motion selection,” Earthquake
Engineering and Structural Dynamics 45(5), 757–777.
Manafpour, A. R. [2008] “Bam earthquake, Iran: lessons on the seismic behavior of building structures,”.
Paper presented at the The 14th World Conference on Earthquake Engineering, Beijing, China.
Mander, J., Dhakal, R., Mashiko, N. and Solberg, K. [2007] “Incremental dynamic analysis applied
to seismic financial risk assessment of bridges,” Engineering Structures 29, 2662–2672.
Mander, J. B. [1999] Fragility Curve Development for Assessing the Seismic Vulnerability of Highway
Bridges, University at Buffalo, State University of New York, New York.
Mander, J. B., Priestley, M. J. N. and Park, R. [1988] “Observed stress strain behavior of confined
concrete,” Journal Structureal Engineering 114(8), 1827–1849.
Mangalathu, S. [2017] “Performance based grouping and fragility analysis of box girder bridges in
California,” Ph.D. Thesis, Georgia Tech, USA.
Mangalathu, S., Jeon, J. and DesRoches, R. [2017b] “Critical uncertainty parameters influencing seismic
performance of bridges using Lasso regression,” Earthquake Engng Structureal Dynamics 47(3) 784–
801.
Mangalathu, S., Jeon, J., Padgett, J. and DesRoches, R. [2016] “ANCOVA-based grouping of bridge
classes for seismic fragility assessment,” Engineering Structures 123, 379–394.
 JOURNAL OF EARTHQUAKE ENGINEERING 1727

Mangalathu, S., Soleimani, F. and Jeon, J. [2017a] “Bridge classes for regional seismic risk assess-
ment: improving HAZUS models,” Engineering Structures 148, 755–766.
Monteiro, R. [2016c] “Sampling based numerical seismic assessment of continuous span RC
bridges,” Engineering Structures 118, 407–420.
Monteiro, R., Delgado, R. and Pinho, R. [2016a] “Probabilistic seismic assessment of RC bridges:
Part II - nonlinear demand prediction,” Structures 5, 274–283.
Monteiro, R., Delgado, R. and Pinho, R. [2016b] “Probabilistic seismic assessment of RC bridges:
part I - uncertainty models,” Structures 5, 258–273.
Monteiro, R., Zelaschi, C., Silva, A. and Pinho, R. [2017] “Derivation of fragility functions for
seismic assessment of RC bridge portfolios using different intensity measures,” Journal of
Earthquake Engineering 1–17.doi:10.1080/13632469.2017.1387188.
Moschonas, I. F., Kappos, A. J., Panetsos, P., Papadopoulos, V., Makarios, T. and Thanopoulos, P.
[2009] “Seismic fragility curves for Greek bridges: methodology and case studies,” Bulletin of
Earthquake Engineering 7(2), 439–468. doi:10.1007/s10518-008-9077-2.
Mosleh, A. [2016c] “Seismic vulnerability assessment of existing concrete highway Iranian bridges,”
Doctor of Philosophy, Faculty of civil engineering department, Aveiro University, Aveiro,
Portugal.
Mosleh, A., Razzaghi, M., Jara, J. and Varum, H. [2016a] “Seismic fragility analysis of typical pre-
1990 bridges due to near and far-field ground motions,” International Journal of Advanced
Structural Engineering and Mechanics (IJASE) 8(1), 1–9. doi:10.1007/s40091-016-0108-y.
Mosleh, A., Razzaghi, M., Jara, J. and Varum, H. [2016b] “Development of fragility curvesfor RC
bridges subjected to reverse and strike-slip seismic sources,” Earthquakes and Structures 11(3),
517–538. doi:10.12989/eas.2016.11.3.517.
Mosleh, A., Sepahvand, S., Varum, H., Jara, J., Razzaghi, M. S. and Marburg, S. [2018] “Stochastic
collocation-based nonlinear analysis of concrete bridges with uncertain parameters,” Structure
and Infrastructure Engineering Journal. doi:10.1080/15732479.2018.1434209.
Mosleh, A., Varum, H. and Jara, J. [2015] “A methodology for determining the seismic vulnerability
of old concrete highway bridges by using fragility curves,” Structural Engineering and Geotechnics
5(1), 1–7.
Nielson, B. G. [2005] “Analytical fragility curves for highway bridges in moderate seismic zones,”
(PhD Thesis), School of Civil and Environmental Engineering, Georgia Institute of Technology,
Atlanta, Georgia.
Nielson, B. G. and DesRoches, R. [2007a] “Analytical seismic fragility curves for typical bridges in
the central and southeastern United States,” Earthquake Spectra 23, 615–633.
Nielson, B. G. and DesRoches, R. [2007b] “Seismic fragility methodology for highway bridges using
a component level approach,” Earthquake Engineering and Structural Dynamics 36, 823–839.
Padgett, J. E., DesRoches, R. and Nielson, E. [2010] “Regional seismic risk assessment of bridge
network in Charleston, South Carolina,” Journal of Earthquake Engineering 14(6), 918–933.
Padgett, J. E., Nielson, B. G. and DesRoches, R. [2008] “Selection of optimal intensity measures in
probabilistic seismic demand models of highway bridge portfolios,” Earthquake Engineering and
Structural Dynamics 37, 711–725.
Pan, Y., Agrawal, A. K., Ghosn, M. and Alampalli, S. [2010] “Seismic fragility of multi-span simply
supported steel highway bridges in New York State. I: bridge modeling, parametric analysis, and
retrofit design,” ASCE, Journal of Bridge Engineering 15, 448–461.
Park, J. and Towashiraporn, P. [2014] “Rapid seismic damage assessment of railway bridges using
the response-surface statistical model,” Structural Safety 47, 1–12.
Priestley, M. J. N. and Park, R. [1987] “Strength and ductility of concrete bridge columns under
seismic loading,” ACI Structural Journal 84(1), 61–76.
Priestley, M. J. N., Seible, F. and Calvi, G. M. [1996] Seismic Design and Retrofitting of Bridges, John
Wiley & Sons, New York.
Qiang, H., Xiuli, D., Jingbo, L., Zhongxian, L., Liyun, L. and Jianfeng, Z. [2009] “Seismic damage of
highway bridges during the 2008 Wenchuan earthquake,” Earthquake Engineering & Engineering
Vibration 8, 263–273.
1728 A. MOSLEH ET AL.

Ramanathan, K., DesRoches, R. and Padgett, J. E. [2012] “A comparison of pre- and post-seismic
design considerations in moderate seismic zones through the fragility assessment of multi-span
bridge classes,” Engineering Structures 45, 559–573.
Ramanathan, K. N. [2012] “Next generation seismic fragility curves for California bridges incorpor-
ating the evolution in seismic design philosophy,” [Ph.D. thesis], School of Civil and
Environmental Engineering, Georgia Institute of Technology, Atlanta, GA.
Razzaghi, M. and Eshghi, S. [2015] “Probabilistic seismic safety evaluation of precode cylindrical oil tanks,”
Journal of Performance of Constructed Facilities 29(6). doi:10.1061/(ASCE)CF.1943-5509.0000669.
Rossetto, T. and Elnashai, A. S. [2003] “Derivation of vulnerability functions for European type RC
structures based on observational data,” Engineering Structures 25, 1241–1263.
Seo, J. and Linzell, D. G. [2013] “Use of response surface metamodels to generate system level
fragilities for existing curved steel bridges,” Engineering Structure 52(2013), 642–653.
Sheikh, S. and Uzumeri, M. [1980] “Strength and ductility of confined concrete columns,”
Proceeding, ASCE 106, 1079–1102.
Shinozuka, M., Feng, M. Q., Kim, H., Uzawa, T. and Ueda, T. [2001] “Statistical analysis of fragility
curves, (Report No. MCEER-03-0002),” Retrieved from MCEER, University at Buffalo, The State
University of New York., Buffalo, NY.
Shinozuka, M., Feng, M. Q., Kim, H. K. and Kim, S. H. [2000b] “Nonlinear static procedure for
fragility curve development,” ASCE Journal of Engineering Mechanics 126, 1287–1296.
Simon, J. and Vigh, L. [2017] “Seismic reliability assessment of typical road bridges in Hungary,”
Journal of Earthquake Engineering 1–29.doi:10.1080/13632469.2017.1297270.
Tavares, D.H., Padgett, J.E. and Paultre, P. [2012] “Fragility curves of typical as-built highway
bridges in eastern Canada,” Engineering Structures 40, 107–118.
Veneziano, D., Casciati, F. and Faravelli, L. [1983] Method of Seismic Fragility for Complicated
Systems, International Nuclear Information System, Livermore, California. Report Number:
CONF-8305220–.
Von Thun, J. L., Rochim, L. H., Scott, G. A. and Wilson, J. A. [1998] “Earthquake ground motions
for design and analysis of dams, earthquake engineering and soil dynamics II recent advances in
ground-motion evaluation,” Geotechnical Special Publication 20, 463–481.
Whitman, R. V., Biggs, J. M., Brennan, J. E., Cornell, A. C., Neufville, R. L. and Vanmarcke, E. H.
[1975] “Seismic design decision analysis,” ASCE Journal of Structural Division 101, 1067–1084.
Wibowo, H., Sanford, D. M., Buckle, I. G. and Sanders, D. H. [2013] “The effect of live load on the
seismic response of bridges. Technical Report,” Center for Civil Engineering Earthquake
Research, University of Nevada, Reno.
Yakut, A. and Yılmaz, H. [2008] “Correlation of deformation demands with ground motion
intensity,” ASCE Journal of Structural Engineering 134(12), 1818–1828.
Yamazaki, F., Hamada, T., Motoyama, H., and Yamauchi, H. [2000] Earthquake damage assessment
of expressway bridges in japan. Technical Council on Lifeline Earthquake Engineering (TCLEE).
361–370.
Ying, M. and Jin-Xin, G. [2018] “Seismic failure modes and deformation capacity of reinforced
concrete columns under cyclic loads,” Periodica Polytechnica Civil Engineering 62, 80–91.
Zahrai, S. M. and Heidarzadeh, M. [2007] “Destructive effects of the 2003 bam earthquake on
structures,” Asian Journal Of Civil Engineering (Building And Housing) 8(3), 329–342.
Zelaschi, C., Monteiro, R. and Pinho, R. [2015] “Improved fragility functions for RC bridge popula-
tions,” Paper presented at the 5th ECCOMAS Thematic Conference on Computational Methods
in Structural Dynamics and Earthquake Engineering (COMPDYN 2015), Crete Island, Greece.
Zelaschi, C., Monteiro, R. and Pinho, R. [2016] “Parametric characterization of RC bridges for
seismic assessment purposes,” Structures 7, 14–24.
Zelaschi, C., Monteiro, R. and Pinho, R. [2017] “Critical assessment of intensity measures for
seismic response of Italian RC bridges portfolios,” Journal of Earthquake Engineering 1–21.
doi:10.1080/13632469.2017.1342293.
Zhu, L., Elwood, K. J. and Haukaas, T. [2007] “Classification and seismic safety evaluation of
existing reinforced concrete columns,” Journal of Structural Engineering and Mechanics 133(9),
1316. doi:10.1061/(ASCE)0733-9445.