Aging Considerations in The Development of Time Dependent Seismic Fragility Curves.

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/228373395
Aging Considerations in the Development of Time-Dependent Seismic Fragility

Curves
Article in Journal of Structural Engineering · December 2010

DOI: 10.1061/(ASCE)ST.1943-541X.0000260
CITATIONS READS
146 591
4 authors, including:
Jayadipta Ghosh Jamie Ellen Padgett

Indian Institute of Technology Bombay Rice University
38 PUBLICATIONS 543 CITATIONS 203 PUBLICATIONS 2,913 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
1st International Workshop on Risk and Resilience of Industrial installations Against Natural Threats and Mitigation Strategies View project
Intensity measure selection for probabilistic seismic risk assessment of complex structural systems View project
All content following this page was uploaded by Jamie Ellen Padgett on 09 October 2015.
The user has requested enhancement of the downloaded file.

Aging Considerations in the Development of
Time-Dependent Seismic Fragility Curves
Jayadipta Ghosh, S.M.ASCE1; and Jamie E. Padgett, A.M.ASCE2
Abstract: This paper presents the formulation of a time-dependent seismic fragility format for bridges, as well as new insights into the
potential effects of aging and deterioration on seismic vulnerability traditionally neglected in fragility modeling, including joint impacts
of multiple component deterioration not investigated to date. The study evaluates the impact of lifetime exposure to chlorides from deicing
salts on the seismic performance of multispan continuous highway bridges, considering corrosion of reinforced concrete columns and steel
bridge bearings. The components’ degradation and their influence on seismic response are illustrated through three-dimensional nonlinear
dynamic analysis. A full probabilistic analysis accounting for variation in bridge, ground motion, and corrosion parameters is conducted
to develop time-dependent seismic fragility curves. These fragility curves indicate the evolving potential for component and system
damage under seismic loading considering time-dependent corrosion-induced deterioration. The results indicate that while corrosion may
actually decrease the seismic vulnerability of some components, most critical components suffer an increase in vulnerability. Quadratic
models depicting the change in lognormal seismic fragility parameters are proposed to capture the time-dependent effect of aging on the
fragility of the bridge system. Overall, the seismic vulnerability significantly increases throughout the lifetime of the representative bridge
geometry, with a 32% shift in the median value of complete damage fragility near the end of the bridge’s life.
DOI: 10.1061/共ASCE兲ST.1943-541X.0000260
CE Database subject headings: Bridges; Seismic effects; Deterioration; Corrosion; Probability; Aging.
Author keywords: Bridges; Seismic; Deterioration; Corrosion; Probability; Fragility.
Introduction bearings, and corrosion of steel reinforcement in RC columns

among others. These components comprise the primary force re-
Across the United States, hundreds of thousands of existing sisting system of bridges under seismic loading thus having a
bridges are located in seismic zones 共FHWA 2009兲. Regional risk potentially significant impact on seismic fragility estimates,
assessment of this infrastructure has relied upon significant ad- which currently neglect these aging parameters and emphasize the
vances in the development of tools known as bridge fragility as-built or pristine condition of bridges. Moreover, the time-
curves 共Basoz and Mander 1999; Basoz and Kiremidjian 1999; dependent nature of bridge deterioration phenomena suggests that
Shinozuka et al. 2000; Mackie and Stojadinovic 2001; Nielson seismic fragility models of bridge components or systems should
and DesRoches 2007a; Straub and Der Kiureghian 2008兲, which also reflect this evolving vulnerability in order to increase the
offer statements of the probability bridge failure conditioned upon reliability of their application in risk assessment of aging bridge
intensity of ground motion. The integration of these conditional inventories.
reliability models into regional seismic risk assessments provides Despite the potential effect of deterioration across a large
an opportunity to screen seismically vulnerable bridges for retro- population of aged bridges on seismic performance, there has
fit, project anticipated damage and losses, or support postevent been a lack of historic consideration of the joint effect of seismic
inspection. However, according to the American Society of Civil and aging threats. Recent notable work by Choe et al. 共2008,
Engineers, over half of the 599,766 bridges in the United States 2009兲 has highlighted the potential reduction in capacity and in-
are approaching the end of their design life and nearly a quarter crease in fragility of a typical single-bent bridge in California
need significant retrofit or replacement to eliminate deficiencies considering nonlinear static analyses in the derivation of indi-
共ASCE 2009兲. The aging and deterioration of bridges manifest vidual component fragility models 共namely, RC columns in ma-
itself in a number of ways, such as spalling of reinforced concrete rine splash zone兲. This work illustrated the potential importance
共RC兲 members, buildup of debris leading to corrosion of steel of capturing the effects of aging on seismic fragility and identi-
fying the crucial material and corrosion parameters that most sig-
1
Graduate Research Assistant, Dept. of Civil and Environmental En- nificantly affect the bridge reliability. However, for other bridge
gineering, Rice Univ., 6100 Main St., MS-318, Houston, TX 77005. types, such as multiple span steel or concrete girder bridges, past
2
Assistant Professor, Dept. of Civil and Environmental Engineering, studies have illustrated the importance of capturing a number of
Rice Univ., 6100 Main St., MS-318, Houston, TX 77005 共corresponding vulnerable components in the fragility assessment of the bridge
author兲. E-mail: [email protected] system 共Nielson and DesRoches 2007b兲. Therefore, further re-
Note. This manuscript was submitted on July 31, 2009; approved on
search is required to evaluate the effect of aging on system re-
June 10, 2010; published online on June 15, 2010. Discussion period
open until May 1, 2011; separate discussions must be submitted for indi- sponse and fragility, considering not only the vulnerability of
vidual papers. This paper is part of the Journal of Structural Engineer- multiple components but also their simultaneous aging. Further-
ing, Vol. 136, No. 12, December 1, 2010. ©ASCE, ISSN 0733-9445/ more, the anticipated dynamic behavior of these systems under
2010/12-1497–1511/$25.00. various time-dependent deterioration mechanisms has yet to be
JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1497
Downloaded 15 Dec 2010 to 168.7.220.129. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
characterized to support the development of time-dependent seis- Case Study Bridge and Base Finite-Element Model
mic fragility curves.
A potential form of environmental degradation of bridges is To demonstrate the methodology for developing time-dependent
chloride induced corrosion of bridge components such as RC col- fragility curves and provide insight on the impacts of aging of
umns and steel bridge bearings 共Enright and Frangopol 1998; multiple components on seismic vulnerability, a sample MSC
Stewart and Rosowsky 1998; Montemor et al. 2002; Hoeke et al. steel girder bridge is used as a case study in this paper. The bridge
2009兲. Corrosion deterioration of bridges is predominantly found is representative of the median dimensions among all MSC steel
to occur at locations with close proximity to the sea coast and in girder bridges found in the CSUS bridge inventory based on past
regions where deicing salts are applied for snow and ice removal. statistical analyses 共Nielson 2005兲. The MSC steel girder bridge is
The study presented herein focuses on investigating the effects of adopted for the case study because of the prevalence of this
corrosion due to chloride-laden deicing salts because of 共1兲 their bridge class in regions of potential seismic hazard 共i.e., attributing
extensive use in bridges across the United States 共Broomfield 13.2% of CSUS bridges in seismic zones according to Nielson
1997兲 and 共2兲 the identification of chloride induced corrosion 2005兲, typical design details including multiple components that
from deicing salts as one of the most severe forms of corrosion are susceptible to corrosion, and concern regarding initial bridge
which causes significantly higher degradation than chlorides in a vulnerability even prior to considering aging. Past comparative
marine environment 共Stewart and Rosowsky 1998兲. The corrosion studies on classes of bridges, which are particularly common in
of critical components considered in this study includes corrosion the Central United States, indicate that MSC steel girder bridges
of the steel reinforcement in RC columns and the deterioration of are among the most vulnerable bridges 共Nielson and DesRoches
steel bridge bearings commonly found in existing inventories of 2007b兲. This can be attributed to the inadequate seismic detailing
of the columns having approximately 1% longitudinal reinforce-
steel girder bridges in seismically vulnerable regions such as the
ment ratio along with widely placed transverse ties, using the
Central and Southeastern United States 共CSUS兲. While bridge
vulnerable high type steel fixed and rocker bearings, short seat
columns are potentially subjected to traffic spray of chloride-
widths, and inadequately reinforced pile caps. Consequently, large
laden water under the bridge 共Weyers et al. 1994兲, steel bridge
inertial deck loads result in considerably high demands on the
bearings are affected by the leaking of chloride-laden water at
underreinforced columns, expansion bearings, and abutments dur-
deck joints 共Silano and Brinckerhoff 1993兲.
ing seismic events.
This paper offers the formulation of a time-dependent seismic
fragility format as well as new insights into the potential effects
of aging and deterioration on seismic vulnerability traditionally
Bridge Geometry
neglected in fragility modeling, including joint impacts of bearing
and column deterioration not investigated to date. The degrada-
The typical MSC steel bridge configuration used in this study is
tion mechanisms associated with corrosive chloride attack consid-
that identified by Nielson 共2005兲, as shown in Fig. 1, illustrating
ered in this study include reduction in strength of bridge columns
the continuity of the steel girders over the interior bents. Both the
due to decrease in diameter of the corroded reinforcement, in-
end spans and the middle span of this three span bridge are 22.30
creased bearing coefficient of friction due to corrosion debris ac- m long and 10.3 m wide consisting of five steel girders. Each bent
cumulation, and failure of the bridge bearing system due to consists of three circular columns having 645-mm2 nominal
corrosion of anchor bolts and keeper plates. The impact of these cross-sectional area, reinforced with 12 No. 29 longitudinal bars
multiple component deterioration mechanisms is assessed within 共metric size兲, and No. 13 共metric size兲 transverse stirrups spaced
the framework of bridge system fragility estimation for a typical at 300 mm. The bridge uses high type steel fixed bearings beneath
multispan continuous 共MSC兲 steel girder bridge throughout its each girder over the bent beam and high type steel expansion
service life. Analytical models for corroded RC columns and steel 共rocker兲 bearings at the abutments. These bearings are placed on
bridge bearings are introduced into the probabilistic seismic per- masonry plates and attached to the bridge pier and abutments
formance assessment, which assesses system reliability consider- using anchor bolts. Besides being highly prone to corrosion dete-
ing the contribution of multiple correlated component failures rioration, the nonductile nature of these bearings makes them
共e.g., columns, bearings, and abutments兲. Finite-element model- highly susceptible to seismic damage 共Mander et al. 1996兲.
ing and nonlinear dynamic analysis of the typical bridge with its
deteriorated components are conducted to provide new under-
standing of the effect of aging on seismic response at the indi- Finite-Element Modeling of Pristine Bridge
vidual component and bridge system level. A sampling-based Components
analytical fragility analysis is then conducted considering uncer-
tainty in hazard, bridge, and deterioration parameters to yield a A three-dimensional finite-element model for the chosen bridge
comparison of the anticipated seismic performance at different configuration is developed using the finite-element platform Op-
points in time along the bridge’s service life. New time-dependent enSees 共Mazzoni et al. 2009兲 following the recommendations by
fragility models of the MSC steel girder bridge are formulated Nielson and DesRoches 共2007a兲. An overview of the finite-
that capture the decrease in anticipated seismic performance as element model for the nondegraded bridge is presented herein for
the bridge ages and suffers continued exposure to the elements. completion prior to describing the influence of corrosion on the
The discussion of results offers insights on the potential joint bridge model. For the superstructure modeling, the composite ac-
effects of deterioration of multiple components on lifetime bridge tions of the steel girders and bridge deck are taken into account
vulnerability under seismic loads. In addition, opportunities for and modeled with linear elastic beam-column elements since
future work are discussed, such as accounting for additional de- damage is not expected in the superstructure. Analytical modeling
terioration mechanisms, conducting field-condition updating of of the bridge bearings relies heavily on the experimental and ana-
fragility estimates, or developing time-dependent fragility curves lytical suggestions following the reversed cyclic loading tests on
conditioned upon multiple parameters. steel bridge bearings by Mander et al. 共1996兲. Subsequently, bi-
1498 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010
22.3 m 22.3 m 22.3 m
1.2 m
Original Reinforcement
Corroded Reinforcement 4m
12 - # 29 bars
Expansion Bearing - Longitudinal ( 645mm 2 a r ea) Fixed Bearing - Longitudinal
Cl
50 mmCl Cl
Force Force Force k3
Fy_corr k2
Cl Cl Fy kp Fy k 2_corr
orr
Fy_corr = f( k3
corr kp
4_c
k4
ke ke k1
k
Fy = f(
Cl
Disp. 813 mm Cl Disp. u1
Disp.
u2
Cl Cl
#13 bars@ 300 mm o.c.
50 mm Cl Cl
Cl
Fig. 1. Elevation view of the three-day nonlinear analytical model of the MSC steel girder bridge. Modeling of components vulnerable to
corrosion degradation in the longitudinal direction is emphasized.
linear truss elements and link elements in OpenSees are used to the bridge. This correlation stems from the reasoning that the
model the prying 共of the masonry plate兲, rocking, and stiffness vulnerable bridge components are similarly exposed to chlorides
degradation of the bearing assembly along both longitudinal and within the same bridge, yet may lead to more highly correlated
transverse directions. Nonlinear beam-column elements with fiber failure probability between the bridge components. The increased
defined cross sections are used to model columns and bent beams, availability of field monitoring data can be used to refine this
while nonlinear springs represent the abutments with lumped assumption in the future using the same framework as presented
translational and rotational springs at the base of columns to in this paper. The sections below discuss the deterioration models
model the piles. adopted for the case study bridge, random variables considered in
The three-dimensional finite-element modeling approach de- the vulnerability assessment specific to the modeling of aged
scribed for the components corresponds to the bridge in its pris- components, and their influence on the finite-element model used
tine or nondegraded state. Hence, these models need to be in seismic response assessment previously detailed for the pristine
updated and refined with new deteriorated component models to bridge.
take into account the effects of aging and degradation. The
present study focuses on the degradation of the RC columns and
steel high type fixed and expansion bearings in a corrosive envi- RC Column Degradation due to Corrosion
ronment. The respective corrosion mechanisms and their influ-
ence on the finite-element modeling are discussed in the Corrosion of reinforcing steel can be mainly attributed to the
following sections. ingress of chloride ions from the concrete surface through the
concrete cover to the reinforcing steel. The chloride exposure
condition considered in this study is the traffic spray scenario,
Bridge Deterioration Modeling which refers to the case of deicing salt chlorides becoming air-
borne as a result of vehicles passing through chloride-laden water
The strength of the components of any structural system, in gen- beneath the bridge 共Enright and Frangopol 1998兲.
eral, is a time-dependent property which may decrease in resis- Chloride ion ingress into RC structures has been found to be
tance along the structure’s service life. Potential reasons for best represented by Fick’s second law of diffusion through a
structural strength degradation can be attributed to multiple fac- semiinfinite solid 共Stewart and Rosowsky 1998兲, represented as
tors such as corrosion, erosion, other forms of chemical deterio-
ration, and fatigue 共Melchers and Frangopol 2008兲. Highway ⳵ C共x,t兲 ⳵2C共x,t兲
bridges, often exposed to severe environmental conditions, are = − Dc 共1兲
⳵t ⳵ x2
particularly vulnerable to strength and resistance loss with time,
which may affect the response of the structure under dynamic where Dc = diffusion coefficient; C = chloride ion concentration;
loading. The present study considers the aging of typical highway x = depth of concrete from the surface; and t = time in years. As-
bridges by including probabilistic models of chloride induced cor- suming that the chloride ion concentration near the concrete sur-
rosion deterioration of the RC columns and steel bridge bearings face is constant as typically assumed for corrosion due to deicing
within the seismic fragility modeling framework. In this analysis, salts 共Hoffman and Weyers 1996; Vu and Stewart 2000兲, the cor-
the impacts of the environmental degradation in the form of corrosion initiation time is therefore taken as 共Thoft-Christensen et
rosion parameters are considered to be perfectly correlated across al. 1996兲
Table 1. Descriptors of Lognormal Random Variables Affecting the Cor- Normalized Time−Variant Area Reduction
rosion Deterioration of RC Columns
1 Mean Normalized Area
Descriptor Unit Mean COVa PDF of Normalized Area
Cover depth 共x兲 cm 3.81 0.20 0.8

Diffusion coefficient 共Dc兲 cm2 / year 1.29 0.10
Surface chloride concentration 共C0兲 wt % concreteb 0.10 0.10
A(t)/A0
0.6
Critical chloride concentration 共Ccr兲 wt % concreteb 0.040 0.10
Rate of corrosion 共rcorr兲 mm/year 0.127 0.3
a
COV= coefficient of variation. 0.4
b
wt % concrete= percent by weight of concrete.
冋冉冊册
0.2
−2
x2 C0 − Ccr
Ti = erf−1 共2兲 0 20 40 60 80 100
4Dc C0 Time (years)
where Ti = corrosion initiation time; C0 = equilibrium chloride con-
Fig. 2. Distribution of normalized time-variant area of the column
centration at the concrete surface; Ccr = critical chloride concen-
reinforcement
tration that causes dissolution of the protective passive film
around the reinforcement and initiates corrosion; and erf
= Gaussian error function that can be mathematically represented
as 共Edwards 2006兲 bridge. The primary reason behind this assumption is the tradi-
冕
␪
tional lack of explicit data for time-dependent corrosion rate mod-
2 2 eling, which also prompted previous researchers 共Frangopol et al.
erf共␪兲 = e−p dt 共3兲
冑␲ 0
1997; Val et al. 2000; Akgül and Frangopol 2004; Liu 2005兲 to
conduct bridge reliability studies using models with average cor-
The corrosion initiation time depends on a number of parameters rosion rates. In this study, uncertainty in the corrosion rate is
that can vary considerably for different bridges depending on accounted for by adopting a probabilistic model for the corrosion
bridge location and environmental exposure condition. The rate rather than a single deterministic value, as shown in Table 1.
probabilistic models for the lognormally distributed parameters Due to the potential influence of the corrosion rate model
describing the corrosion initiation of the RC columns adopted for adopted, alternate models can be incorporated through the same
the present study are given in Table 1. These lognormally distrib- approach presented in this paper for bridge-specific analysis. On
uted random variables are identified based on in-field corrosion the basis of the estimated distribution for corrosion initiation time
related studies of existing bridge components in the United States presented above and anticipated corrosion rate, the area of rein-
exposed to deicing salts 共Whiting et al. 1990; Weyers et al. 1994; forcing steel is probabilistically assessed as a function of time.
Enright and Frangopol 1998兲. The distribution for the corrosion Fig. 2 shows the resulting time-dependent area reduction ratio,
initiation time is assessed through Monte Carlo simulation having which is the area of reinforcing steel at time t, A共t兲, normalized
a sample size of 50,000. A lognormal distribution with mean 8.85 by the initial area of reinforcement, A0. This figure illustrates the
years and standard deviation of 4.5 years is found to be a good fit reduction in steel cross-sectional area over time and the increase
to the simulated data for corrosion initiation time. This distribu- in variability or uncertainty about that estimate of reinforcement
tion will subsequently be used as a key input for probabilistic area due to the combined effect of the variability of initial rein-
modeling of rebar corrosion for bridge columns. forcement diameter, rate of corrosion, and corrosion initiation
Once the protective passive film around the reinforcement dis- time. The loss of area of steel due to corrosion of the RC columns
solves due to continued chloride ingress, corrosion initiates and is correspondingly modeled as a reduction in longitudinal rein-
the time-dependent loss of reinforcement cross-sectional area, forcing bar cross-sectional area in the fiber section model as com-
A共t兲, can be expressed as 共Thoft-Christensen et al. 1996; Enright pared to the pristine columns in the finite-element model.
and Frangopol 1998兲
冦冧
␲
nD2i for t ⱕ Ti Deterioration of Steel Bridge Bearings due to
4
Corrosion
A共t兲 = ␲ 共4兲
n关D共t兲兴2 for Ti ⬍ t ⬍ Ti + Di/rcorr
4 Degraded seismic performance of the bridge bearings results pri-
0 for t ⱖ Ti + Di/rcorr marily due to the corrosion debris accumulation resulting in “fro-
zen” or “locked” bearings, as well as the corrosion of anchor bolts
where n = number of reinforcement bars; Di = initial diameter of and keeper plates used in bearing assemblies. Typically, such an-
steel reinforcement; t = elapsed time in years; rcorr = rate of corro- chor bolts are used in both fixed and expansion 共rocker兲 bearings,
sion; and D共t兲 = reinforcement diameter at the end of 共t − Ti兲 years, while the keeper plates are typical details in the expansion bear-
which can be represented as ing assemblies. As identified by Mander et al. 共1996兲 bearing
anchor bolts often form a “weak link” in the chain of force trans-
D共t兲 = Di − rcorr · 共t − Ti兲共5兲
mission from the superstructure to the substructure during seismic
It is acknowledged that the corrosion rate is an environmen- events. Corrosion of these same elements may potentially result
tally dependent parameter with potential time dependence 共Vu in a shift in performance during seismic loading. The primary
and Stewart 2000兲. However, the rate of corrosion in this study is reasons for the corrosion of the bearing assembly are the leaking
considered to be constant on average along the service life of the of chloride-laden water from the deicing salts through the deck
5 Ultimate Lateral Strength Reduction
x 10
2.5
Mean Normalized Area
Ultimate Lateral Strength: Fult (N)

PDF of Normalized Area
2.4
2.3
2.2
2.1
2
0 20 40 60 80 100
Time (years)
Fig. 4. Reduction of ultimate lateral strength of fixed bearings along

the longitudinal direction due to corrosion of anchor bolts
Fig. 3. Rocking/force distribution mechanism in fixed bearings on
concrete pedestals along longitudinal direction
verse direction and f c is the concrete compressive strength.
The bond strength of the anchor bolt may be estimated as
joints 共Silano and Brinckerhoff 1993兲 and the traffic spray sce- 共Mander et al. 1996兲
narios which may further expose these components to airborne
chlorides resulting from the passage of vehicles beneath the B = bu共␲db兲ld 共9兲
bridge through chloride-laden water 共Enright and Frangopol where bu = kba = bond stress that is assumed to act uniformly on
1998兲. Recent in-field examples of severe anchor bolt corrosion the anchor bolt surface with diameter db over the embedment
exist in the literature for typical highway bridges in the state of length ld; ba = average bond stress over the length of the anchor
Georgia 共Hoeke et al. 2009兲. In regions where use of deicing salts bolt; and k = modification or judgment factor often imposed to
on bridge decks is more prevalent, the severity of anchor bolt account for reduced bearing capacity of the anchor bolts and ad-
corrosion may be even more critical. Additionally, accumulation verse effects of cyclic loading. From the above equations, the
of excessive corrosion products, dirt, and debris may potentially ultimate lateral strength, Fult, of the bearing maybe expressed as
冋冉冊冊册
result in frozen or locked bearings since it may restrict transla-
tional and rotational movements due to an increased coefficient of
friction 共Silano and Brinckerhoff 1993兲.
Fult 0.5wl
N
=
h
1+
␣B
N
−
N
0.85f cwlwt
冉
1+
␣B
N
2
共10兲
Anchor bolt corrosion leads to reduced ultimate lateral
strength of the bearing assembly as explained in the following While a similar relationship for estimating the ultimate
paragraphs. Fig. 3 shows the arrangement and distribution of strength of fixed bearing along the transverse direction can be
forces for a typical fixed bearing along the longitudinal direction. found in available literature 共Mander et al. 1996兲, it should be
This free-body diagram is used to derive the ultimate lateral noted that the calculated strengths along both directions corre-
strength for the fixed bearing in the longitudinal direction, which spond to that of the pristine bearings. Along the service life of the
changes over time due to corrosion. From the equilibrium of hori- bridge due to corrosion, the cross-sectional area of the bolt re-
zontal forces, the ultimate lateral strength of the bearing can be duces and consequently leads to reduced ultimate lateral strengths
obtained as for the deteriorated bearings. The corrosion parameters previously
defined in Table 1 are also assumed to affect the corrosion initia-
Fult = ␣S + ␮V 共6兲 tion time and subsequent area loss of steel in the concrete embed-
and from the equilibrium of vertical forces we have ded anchor bolts. This reduction in cross-sectional area impacts
the bond strength, B, of the embedded bolt as indicated in Eq. 共8兲.
V = N + ␣B 共7兲 Uncertainty in the corrosion parameters via reduction in bond
strength is then propagated through the assessment of ultimate
Also, from the equilibrium of moments about the center of the
lateral strength of the fixed bearing along the longitudinal and
concrete pedestal we obtain
transverse directions. This is shown for the longitudinal direction
Fulth = V 冉冊
wl − ␣
2
共8兲
in Fig. 4. Consequently, the time-evolving probability distribution
for ultimate lateral strength is accounted for in the finite-element
modeling for longitudinal and transverse fixed bearing responses.
where ␣ = number of anchor bolts; S = shear force on one anchor For the case of expansion 共rocker兲 bearings, the motion in the
bolt; ␮ = coefficient of friction between masonry plate and bed- longitudinal direction is primarily rocking, where the ultimate lat-
ding material; V = compression force on the concrete pedestal due eral strength is dependent on the coefficient of rocking friction of
to rocking; N = axial load on the bearing; B = bond strength of the the bearing. As per suggestions by Mander et al. 共1996兲, the co-
swedged anchor bolt in the concrete pedestal; h = height of the efficient of rocking friction varies from 0.04 for clean well worn
bearing from the concrete pedestal to the sole plate-rocker inter- rocker bearings to 0.12 for badly corroded bearings to take into
face; wl = width of masonry plate in the longitudinal direction; and account the “locking” effect as mentioned earlier. In the absence
a = depth of pedestal concrete stress block expressed as a of any further data to support time-dependent modeling of friction
= V / 0.85f cwt, where wt is the width of masonry plate in the trans- increase from corrosion product buildup, a linearly varying coef-
5 Shear Strength Degradation & Demand
x 10
2.5
Anchor Bolt Shear Strength Degradation
PDF of Shear Strength
Shear Strength Demand for Force
2 Transmission from Keeper Plate
Shear Strength (N)

1.5
0.5
0
Fig. 5. Force distribution mechanism through the anchor bolt when 0 20 40 60 80 100
the keeper plate strikes the rocker Time (years)
Fig. 6. Shear strength degradation of and demand on anchor bolt due

ficient of friction is assumed for this study starting with 0.04 for to force transmission from keeper plate
the pristine bridge and 0.12 for a 100-year-old bridge near the end
of its service life.
The most interesting case observed in this study is the effect of 0.20 for P and 0.6 and 0.4 for Q, with a correlation coefficient of
aging on the transverse model of the expansion bearing response. ⫺0.55 between the parameters. Steel bearing keeper plates, typi-
The transverse motion of the expansion bearing initially consists cally made of carbon steel, are assumed to undergo similar cor-
of a sliding frictional component. Once the horizontal frictional rosion penetration.
force exceeds the frictional resistance of the sole plate-rocker Fig. 6 shows the degradation in the shear strength of the an-
interface, the sole plate slides on the rocker until the rocker bear- chor bolt over time due to corrosion. Also plotted is the maximum
ing strikes the keeper plate provided to prevent excess transverse shear force that can be transmitted to each anchor bolt when the
motion. With additional horizontal loading, the keeper plate bends expansion bearing strikes the keeper plate in the transverse direc-
significantly and fails by tearing of the fillet weld securing the tion. Beyond this force level in the anchor bolt, the keeper plate
plate 共Mander et al. 1996兲. The free-body diagram in Fig. 5 shows yields and does not transmit further loads to the anchor bolt.
the distribution of forces for this phenomenon. Along with the mean values presented in the figure, the full
As it can be seen, the force P with which the rocker strikes the probabilistic analysis conducted herein considers variation about
keeper plate gets transmitted in the form of shear forces S1 and S2 both quantities. As illustrated in the figure, before approximately
through the anchor bolts. For intact bearings with no deteriora- 70 years, the failure of the bearing assembly is dictated by failure
tion, the shear strength of the 25-mm diameter anchor bolt is of the keeper plate since the maximum force in the bolt is much
found to be sufficient to transmit the forces, with the failure of the less than its shear strength capacity. However, beyond 70 years,
bearing governed solely by the tearing failure of the keeper plate. the capacity of the anchor bolt decreases below the level of force
However, with corrosion deterioration, there is significant de- transmitted. Hence, the failure of the bearing is governed by shear
crease in the shear strength of the anchor bolts, such that they are failure of the anchor bolt. This phenomenon is captured by ad-
no longer capable to transmit the forces when the rocker strikes justing the ultimate strength of the keeper plate assembly in the
the keeper plate. This refers to the phase where the failure of the expansion bearing model, depending on the point in time along
bearing assembly is determined by the shear failure of the anchor the bridge’s service life.
bolts rather than the failure of the keeper plate.
In addition to the corrosion of the concrete embedded anchor
bolts, the transverse failure mechanism and ultimate strength of Scope of Deterioration Modeling and Future
the expansion bearings as previously described are also a function Opportunities
of the chloride exposed steel keeper plates. The corrosion degra-
dation of the keeper plate is modeled similar to past studies on It is acknowledged that other possible bridge deterioration mecha-
corrosion of steel bridge girders. The corrosion of these exposed nism and permutations of component aging are feasible, which
steel elements is assumed to follow an empirical model following are outside of the scope of the present study and modeling. Em-
the form of a power law 共Komp 1987兲 phasis is placed upon corrosion of steel reinforcement in RC col-
umns, corrosion of keeper plates and anchor bolts in bearing
y共t兲 = PtQ 共11兲
assemblies, and increased coefficient of friction for the steel bear-
where y共t兲 = average corrosion penetration in micrometers; t ings. These mechanisms are among the most severe phenomena
= time in years; and P and Q = parameters determined from regres- associated with corrosion of aging bridges exposed to deicing
sion analysis of field experimental data. Parameters P and Q are salts identified in a series of past studies 共Silano and Brinckerhoff
assumed to follow a truncated correlated bivariate lognormal dis- 1993; Pantazopoulou et al. 2001; Lindquist 2008; Li et al. 2009兲.
tribution and are determined based on field tests by Albrecht and Furthermore these phenomena affect the components of the MSC
Naeemi 共1984兲. The parameter means, coefficients of variation, steel bridge class identified as the most significant and vulnerable
and correlation coefficient used in this study are obtained from components under seismic loads and hence deemed a priority
regression analysis of corrosion penetration field tests on carbon when incorporating effects of aging phenomena within the seis-
steel girders due to deicing salt corrosion. Estimated values of mic reliability analysis. However, it is noted that the corrosion of
mean and coefficient of variation are found to be 53.5 ␮m and reinforcing steel within the concrete columns may lead to second-
ary effects such as cracking and spalling of concrete and loss of MSC bridge. For the specific bridge type and geometry consid-
bond strength, among others, in addition to the loss of reinforce- ered, the first two fundamental modes are longitudinal and trans-
ment cross-sectional area considered herein. These additional verse modes, with respective periods of 0.34 and 0.25 s. The
phenomena highlight the opportunity for future study, although dynamic response is illustrated through nonlinear time history
preliminary analysis conducted as a part of this research indicates analysis of the bridge with median values for all variable param-
that some of these degradation effects can be neglected owing to eters using a ground motion from the synthetic suite by Rix and
their limited impact on the bridge fragility. Fernandez 共2004兲. This motion has peak ground acceleration
Corrosion of steel girders in the bridge superstructure has, for 共PGA兲 of 0.5 g and duration of 29 s. In addition to the dynamic
example, received significant attention in the literature with re- response of the structure in both the longitudinal and transverse
gard to corrosion process modeling and reliability assessment of directions, the influence of corrosion on the load resisting capac-
bridges under live loading 共Kayser and Nowak 1989; Czarnecki ity is also presented. For brevity, comparisons are made between
and Nowak 2008兲. However, a preliminary sensitivity study con- the pristine bridge at time zero and the bridge at 50 years into its
ducted by the writers revealed the negligible impact of this pro-
service life. The following sections illustrate the seismic demands
cess on the bridge’s seismic vulnerability. Even in the event that
due to individual consideration of deterioration mechanisms of
additional girder or superstructure components are added to the
the components as well as the demands due joint consideration of
system fragility definition, the demands placed on a 100-year cor-
both column and bearing deterioration mechanisms.
roded bridge girder under seismic loading are not anticipated to
exceed the yield capacity of the component and hence remain
elastic under seismic loading and negligible in the fragility assess-
ment. Additionally, concrete cover loss due to spalling and bond Impact of Corrosion on the Seismic Response of
strength loss are not modeled in this study for the RC columns or RC Columns
other components. A preliminary sensitivity analysis revealed that
the effect of column concrete cover loss on bridge vulnerability is The effect of time-dependent corrosion of the RC column is first
insignificant, resulting in only a 2–6% shift in the PGA value for assessed. Due to corrosion and subsequent area loss of reinforcing
the corroded bridge column with complete cover loss when com- steel, the load carrying capacity and yield curvature of the RC
pared to the corroded bridge column with intact cover across the columns undergo a significant reduction. This phenomenon is il-
different damage states. Moreover, explicit time-dependent mod- lustrated in Fig. 7共a兲, which shows a 16.6% reduction in yield
eling of concrete spalling leading to cover loss is not a trivial task curvature and a 21% reduction in the yield moment of a 50-year-
and is expected to introduce additional uncertainty in the fragility old corroded column as compared to that of a pristine column.
estimation. Traditionally, studies on concrete cover loss due to Subsequently, when the bridge is subjected to the sample ground
spalling have assumed uniform loss of cover along the whole motion, the demands placed on the corroded RC column increase
length of the column which is seldom the case observed in the relative to the pristine column. This finding is illustrated in Fig.
field. The results of this sensitivity study, however, do not dimin- 7共b兲. The seismic demand placed on the columns is quantified by
ish the potential influence that early aged cracking could have on the curvature ductility demand ratio, which can be expressed as
fragility due to its likely impact on corrosion initiation time, not
explored herein. Furthermore, loss of bond strength is acknowl-
edged to be a potential additional mechanism requiring future kmax
research. Previous studies have shown that for RC members with ␮␸ = 共12兲
kyield
insufficient splice lengths and no confinement, the loss of bond
strength is significant 共Fang et al. 2004; Aquino and Hawkins
2007兲, whereas for members with reinforcement confinement, the
loss of bond strength due to corrosion is not substantial 共Fang et where kmax corresponds to the maximum curvature demanded on
al. 2004兲. The modestly confined columns considered in the case the column throughout the seismic loading and kyield = curvature in
study bridge, typical of preseismic design detailing, fall in be- the column which causes first yield of the outermost reinforcing
tween the two categories noted. The limited availability of experi- bar. While the peak curvature ductility demand of 3.3 for the
mental data for model validation and capacity estimation of bond pristine bridge already indicates significant damage in the form of
strength that affects the seismic response of corroded columns, cracking and spalling, after 50 years of exposure to deicing salts
including poorly confined columns, further highlights the need of the bridge subjected to the same motion suffers peak curvature
future experimental and analytical research. Additionally, some ductility demands of 5.4, which signifies a more severe damage in
other forms of deterioration of the RC members, such as concrete the form of column reinforcement buckling 共Hwang et al. 2001;
strength degradation due to freeze-thaw cycles and chemical at- Buckle 2006兲.
tacks along with the aforementioned factors, provide opportuni- An increased seismic demand on the corroded RC columns is
ties for future study. found to correspond to a negligible increase in demands on cer-
tain components such as expansion bearings and abutments,
which only show approximately 3 and 1% increases, respectively,
Impact of Corrosion-Induced Component in peak displacement in the corroded bridge relative to the pris-
Deterioration on Seismic Response and Capacity of tine bridge. In some components, however, there is a reduction in
Bridge Components the peak seismic demands when column corrosion is modeled.
For instance, compared to the pristine bridge, there is an approxi-
Before conducting a full probabilistic analysis of the impact of mately 13% reduction in the peak longitudinal displacement of
corrosion on the fragility of the representative MSC bridge, a the fixed bearings in the 50-year-old bridge. This reduction in
sample deterministic simulation is presented to illustrate the in- bearing demands is attributed to the concentration of damage in
fluence of time-dependent aging on the seismic response of the the corroded columns as compared to the pristine bridge.
Reduced Moment Resisting Capacity Longitudinal Loading & Response
[Columns]
2500 [Fixed Bearings]
200
2000
100
Moment (kN−m)
Force (kN)
1500
0
Kyield = 0.0025 (1/m)
1000
Kyield = 0.003 (1/m) −100

500 Pristine
50 Year Old
Pristine −200
50 year Old −0.03 −0.02 −0.01 0 0.01 0.02 0.03
0
0 0.005 0.01 0.015 0.02 0.025 0.03 Displacement (m)
Curvature (1/m)
Longitudinal Loading & Response

Fig. 8. Increase in seismic demand on fixed bearings along longitu-
[Columns] dinal direction
2500
µ = 3.3
pristine
2000
reduced ultimate strength of the bearing assembly results in an
1500
18% increase in peak deformation for the 50-year-old bridge
1000 bearing as compared to the pristine bearing.
Moment (kN−m)
500
µ50 = 5.4
0
Joint Consideration of the Impact of Corrosion on
−500
the Seismic Response of the MSC Steel Girder
−1000 Bridge
−1500
To reflect field conditions for corroded bridges throughout their
−2000 Pristine
50 Year Old
lifetime, the joint occurrence of column and bearing corrosion are
−2500
−0.015 −0.01 −0.005 0 0.005 0.01 0.015 0.02 evaluated in this section. Consideration of the simultaneous ef-
Curvature (1/m) fects of corrosion degradation of reinforcing bars in the RC col-
umns and steel bridge bearing assembly reveals several
Fig. 7. 共a兲 Reduction in load resisting capacity and yield curvature;
interesting trends in the seismic response of the MSC steel girder
共b兲 increase in curvature ductility demand placed on the 50-year-old
bridge. Fig. 9 illustrates the seismic demands on the RC columns,
corroded bridge column
expansion bearings in the longitudinal direction, and fixed bear-
ings in the transverse direction due to joint consideration of the
corrosion deterioration mechanisms of the columns, fixed and ex-
Impact of Corrosion on the Seismic Response of pansion bearing assembly. The 50- and 100-year corroded bridge
Steel Bridge Bearings responses are compared to the pristine, or time zero, bridge re-
sponse using column moment-curvature and bearing force-
Corrosion deterioration and subsequent strength reduction of high displacement plots for the sample ground motion.
type steel fixed and expansion bridge bearing assemblies are ad- The corroded RC columns are found to show a consistent in-
dressed in this section. Recalling the deterioration model pre- crease in the curvature ductility demand, which increases by 63
sented in the previous section, corrosion of the steel bearings and 115% for the 50- and 100-year-old column, respectively, rela-
results in reduced stiffness and reduced ultimate strength in the tive to the nondeteriorated column. It is interesting to note that as
fixed bearing assemblies. While this is also true for expansion opposed to earlier findings when single component deterioration
bearings along the transverse direction, increased coefficient of was considered, a joint consideration of column and bearing cor-
friction along the longitudinal direction however results in in- rosion reveals a decrease in peak deformation demand on the
creased expansion bearing stiffness. Analogous to the response of fixed bearings in the longitudinal direction. Quantitatively, 11 and
columns in the previous section, considering only bearing degra- 44% reductions in peak deformations for the 50- and 100-year-old
dation results in an increase in the peak displacement of the fixed steel fixed bearings are observed along the longitudinal direction.
bearing assembly by 16 and 11% in the longitudinal and trans- This reduction in fixed bearing deformations is primarily attrib-
verse directions, respectively. As expected, the reduced post yield uted to the concentration of damage in the corroded columns and
stiffness and ultimate strength due to corrosion shift the hysteretic the dynamic response of the bridge deck and columns as a nearly
characteristics of the bearing and result in larger peak deforma- single degree of freedom system, with little deformation occur-
tions under seismic loading. Fig. 8 shows the comparative re- ring over the columns at the location of the fixed bearings. Addi-
sponse of longitudinal loading response of the fixed bearings for tionally, there is a reduction in the expansion bearing
the pristine bridge and 50-year-old bridge. deformations in the longitudinal direction due to continued in-
For expansion bearings, the increase in coefficient of friction crease in coefficient of friction due to debris accumulation along
due to debris accumulation increases the yield force by 19% and the service life of the bridge 共Fig. 9兲. Consequently, reduced dis-
reduces deformation of the expansion bearings in the longitudinal placements in the longitudinal direction result in reduced pound-
direction by 21%. Additionally, in the transverse direction, the ing forces upon the closure of the 71-mm gap between the deck
longitudinal direction and a reduced post yield stiffness of the
corroded fixed and expansion bearings leads to higher peak bear-
ing displacements as the bridge nears the end of its service life.
For example, the fixed bearing deformations increase by approxi-
mately 15 and 110% 关Fig. 9共c兲兴, and the expansion bearings de-
formations increase by 13 and 69% in the transverse direction for
the 50- and 100-year-old bridge. The impacts of these findings on
the seismic fragility of the bridge system along its service life are
discussed in the next section, considering uncertainty in the
bridge and ground motion realizations.
Impact of Corrosion-Induced Component

Deterioration on Seismic Fragility
The deterministic analysis for the corroded case study bridge

served to help illustrate the bridge response characteristics under
combined corrosion-induced deterioration and seismic loading.
However, given the aleatoric and epistemic uncertainties inherent
in the modeling parameters of the structure, corrosion process,
and ground motion realizations, a full probabilistic analysis is
required to ascertain the effect of corrosion on the probability of
seismic bridge damage. Additionally, since the deterioration
mechanisms are time-dependent processes and the corrosion pa-
rameters tend to increase in variability toward the end of the
bridge’s service life, as previously shown in Figs. 2 and 4, the
evolution of bridge fragility in time should be quantified in the
probabilistic assessment.
Methodology for Development of Fragility Curves
Seismic fragility curves represent the probability of structural

damage conditioned upon ground motion intensity and can pro-
vide insight on bridge seismic vulnerability at both the component
and system levels. The generic expression of seismic fragility
follows the form
p f = P共demand ⬎ capacity兩IM兲共13兲
where IM= intensity measure of the ground motion, taken as PGA
in this study, and the capacity varies for different damage states.
Changes in structural performance due to corrosion inevitably
result in variations in seismic vulnerability. Therefore, fragility
curves are developed herein using a simulation-based analytical
approach rooted in nonlinear time history analysis to capture the
effect of aging on the seismic vulnerability of the MSC steel
girder bridge. A total of 96 two component ground motions from
the synthetic ground motion suites of Wen and Wu 共2001兲 and
Rix and Fernandez 共2004兲 are used in the analysis, representative
of a range of potential ground motions for the CSUS region. An
Fig. 9. 共a兲 Increase in curvature ductility demand on corroded RC equal number of three-dimensional bridge samples are generated
columns; 共b兲 increase in expansion bearing force and decrease in through the Latin hypercube sampling, considering the potential
bearing deformation in longitudinal direction; and 共c兲 increase in uncertainty in structural, material, and corrosion related param-
peak displacement of corroded fixed bearings in transverse direction eters at each point in time along the service life of the bridge. In
addition to the corrosion related parameters presented above in
Table 1, the probabilistic models for the random variables consid-
and the abutment. The decrease in pounding results in a respec- ered for the bridge structure include those previously identified
tive 11 and 27% decreases in the passive deformation of the abut- for MSC steel bridges in the CSUS 共Nielson and DesRoches
ments for the 50- and 100-year-old corroded bridges. 2007a兲, namely, for concrete compressive strength, steel strength,
Both fixed and expansion bearings are found to experience stiffness of foundation piles, damping ratios, and gap between the
large demands in the transverse direction for both pristine and deck and abutment. As previously noted, probability density func-
corroded bridge. Additionally, the increase in demand on the col- tions 共PDFs兲 for corrosion parameters that affect both the cross-
umns in the transverse direction is not as dramatic as that in the sectional area of longitudinal reinforcement and column yield
Table 2. Increase 共 ↑ 兲 or Decrease 共 ↓ 兲 in Component Fragilities and Example Percentage Changes in Median PGA Values for the Complete Damage State
at 75 Years into the Bridge’s Service Life
Percent change in median PGA for complete damage state at
Component Trend in vulnerability with aging 75 years
RC columns ↑ ⫺35%
Fixed bearing—longitudinal ↓a N/A
Fixed bearing—transverse ↑ ⫺38%
Expansion bearings—longitudinal ↓ +10%
Expansion bearings—transverse ↑ ⫺47%
Abutment passive—longitudinal ↓a N/A
Abutment—transverse ↓a N/A
a
Insignificant value of component fragilities at extensive and complete damage states.
curvature, and as well as the yield and ultimate strength of bear- 1998; Cimellaro et al. 2010兲. This system definition is consistent
ing components, are also integrated in the bridge sample genera- with the capacity estimates adopted in the paper, which offer the
tion. limit of component response upon which system level function-
Following a similar approach as presented in the component of ality inhibition is anticipated and implies that failure of any one of
Nielson and DesRoches 共2007b兲 and Padgett and DesRoches the components is indicative of overall bridge system failure.
共2008兲, system level fragility analysis is conducted. For the Under the series system assumption, the probability of the bridge
probabilistic analysis along the service life of the bridge, the 96 system is at or beyond a particular failure limit state that is the
bridge samples generated for each point in time 共e.g., 0, 25, 50, union of the probabilities of each of the components being in the
75, and 100 years兲 are each subjected a seismic ground motion same limit state. This can be mathematically shown as
from the suite in a nonlinear time history analysis. Probabilistic
n
seismic demand models are developed, which reflect the relation-
P关failuresystem兴 = 艛 P关failurecomponenti兴共15兲
ship between peak component demands and ground motion inten- i=1
sity. The responses considered in this research include such
seismically vulnerable components as columns, fixed bearings, where P关failuresystem兴 = probability of failure of the bridge system;
expansion bearings, and abutments. Consequently, the median P关failurecomponenti兴 = probability of failure of the ith component;
value and logarithmic standard deviation of seismic demand at a and n = total number of vulnerable bridge components. As previ-
particular point in time are calculated through regression analysis, ously noted, the estimate of correlation between peak component
where the median value is assumed to follow a power law model responses enables construction of a joint PDF for component de-
共Cornell et al. 2002兲 mand. The bridge system fragility is then evaluated by comparing
the joint PDF of demand with the component capacities for each
D共t兲median = r1共t兲IMr2共t兲共14兲 damage state via the Monte Carlo analysis to derive system fra-
gility estimates that account for component correlations at differ-
where r1共t兲 and r2共t兲 = regression parameters for the point in time ent points in time along the service life of the bridge. The overall
under consideration 共e.g., year兲 and IM= ground motion intensity. corroded bridge system fragility can thus be mathematically rep-
The correlation between component demands is also evaluated for resented as
再冎
use in the system fragility analysis.
In addition to the probabilistic model of seismic demand, the ln共PGA兲 − ln关m共t兲兴
fragility analysis requires estimates of the structural capacities of P关DS兩PGA兴 = ⌽ 共16兲
␨共t兲
the different bridge components. The limit state capacities used in
this study are the lognormal capacity estimates presented by Niel- where m共t兲 and ␨共t兲 = median values 共in units of g PGA兲 and
son and DesRoches 共2007a兲 for typical MSC steel girder bridges. logarithmic standard deviations of the system fragilities at differ-
The limit state capacities for each damage state 共slight, moderate, ent points in time 共t兲 along the service life and DS= damage state.
extensive, and complete兲 are presented in Table 5 in the Appendix
for reference. The use of curvature ductility demand for the col-
umn capacity limit state, as a maximum curvature normalized by Bridge Component and System Fragilities
the yield curvature, averts the need to explicitly change the ca-
pacity limit state models. However, the change in yield curvature Bridge component and system fragilities are evaluated at different
and the reduced moment capacity for the corroded columns trans- points in time along the service life of the bridge to assess the
late into a change in curvature ductility demands placed on the effect of corrosion on the seismic vulnerability. It is observed that
columns, as previously illustrated. at the component level, while there is a steady increase in the
With the lognormal distributions for the structural demand and fragility of certain elements, some other components show a re-
capacity, the lognormal distribution for the seismic fragility of the duced vulnerability with time. This contradicting trend in the
components can be found in closed form 共Melchers 1999兲. As- component seismic fragilities are consistent with the findings
sessment of the bridge system reliability is carried out by assum- from the deterministic analysis which revealed that increased de-
ing the bridge as a series system, wherein failure of a single mands on typical components 共such as deteriorated RC columns兲
component is representative of bridge failure, similar to the sys- result in decreased demands on certain other bridge elements
tem representation adopted in structural reliability studies for 共such as fixed bearings in the longitudinal direction兲. Due to space
bridges 共e.g., Nielson and DesRoches 2007b; Nowak and Cho limitations, changes in fragilities of other bridge components,
2007兲 or other structural systems 共Der Kiureghian and Dakessian whether increasing or decreasing, for the aged bridge relative to
System Fragility − Slight Damage System Fragility − Moderate Damage
1 1
0.8 0.8
P[Moderate|PGA]
P[Slight|PGA]
0.6 0.6
Pristine Pristine
0.4 25 years 0.4 25 years
50years 50years
0.2 75years 0.2 75years
100 years 100 years
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
PGA(g) PGA(g)
System Fragility − Extensive Damage System Fragility − Complete Damage

1 0.8
Pristine Pristine
0.8 25 years 25 years
0.6
P[Extensive|PGA]
P[Complete|PGA]
50years 50years
0.6 75years 75years
100 years 0.4 100 years
0.4
0.2
0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
PGA(g) PGA(g)
Fig. 10. System level seismic fragility curves for different damage states
the pristine bridge are presented qualitatively in Table 2. For the age for the pristine, or time zero, bridge, but after 75 years of
RC columns, fixed and expansion bearings, these trends are found exposure to deicing salts, the chance of complete damage for the
to be consistent for at all damage states at different points in time same level of earthquake is 49%. A complete list of the median
along the service life of the bridge. The percent change in the and dispersion values of the system fragility at all damage states
median value PGA for the complete damage state after 75 years is is provided in Table 3 for different points in time. The decrease in
also listed for each component. This quantity provides insight on median values of fragility for the different damage states along
the relative magnitude of change in fragility due to corrosion for the service life of the deteriorated bridge is further a direct indi-
each bridge component. The abutment fragilities of the pristine cation of the increased bridge system vulnerability due to corro-
along both directions are insignificant beyond the moderate dam- sion of its critical structural components. In general, there is also
age state and correspondingly the changes in fragility in the cor- a slight change in the dispersion over time, indicating reduced
roded bridge are found to be negligible. uncertainty in estimating the PGA value corresponding to exceed-
Although consideration of joint degradation of RC columns ance of each damage state when the bridge is corroded. The over-
and steel bridge bearings results in increasing fragility of some all increase in seismic fragility of the bridge can be attributed to
components 共for example, columns兲 and decreasing fragility of the dominance of the columns, transverse and longitudinal expan-
others 共for example, fixed bearings兲, the overall seismic fragility sion bearings, followed by the transverse fixed bearings, indicat-
at the system level increases in time as the bridge continues to ing the bridge system vulnerability. On the whole, the seismic
corrode 共Fig. 10兲. For example, for a ground motion having vulnerability of these components tends to be negatively affected
PGA= 0.6 g, there is a 30% chance of achieving complete dam- by the continued corrosion of the bridge.
Table 3. Median and Dispersion Values of System Fragilities for All Damage States at Different Points in Time
Damage state
Slight Moderate Extensive Complete
Time 共years兲 m ␨ m ␨ m ␨ m ␨
0 0.269 0.701 0.517 0.621 0.657 0.648 0.888 0.7
25 0.266 0.646 0.48 0.582 0.608 0.621 0.789 0.64
50 0.261 0.607 0.467 0.544 0.596 0.586 0.788 0.617
75 0.235 0.542 0.395 0.53 0.508 0.563 0.674 0.596
100 0.208 0.526 0.35 0.492 0.455 0.537 0.634 0.582
Note: m = median 共PGA兲 and ␨ = dispersion.
 for the fragility parameters for all of the damage states are given
m(t)= -1.8e-5 t2 - 0.0009 t +0.67 in Table 4. These coefficients provide time-dependent models for

the median and dispersion that can be substituted in the following

equation to evaluate the bridge system fragility at any point in
time considering column and bearing corrosion
冉冊

ln共PGA兲 − ln共p1_mt2 + p2_mt + p3_m兲
 P关DS兩PGA兴共t兲 = ⌽
p1_␨t2 + p2_␨t + p3_␨
 共18兲

      where the subscripts on the coefficients indicate median, m, or
  dispersion, ␨.

      The format of time-dependent functions presented in this
  paper offers an efficient approach to capture the impact of aging
and deterioration on bridge system fragility, thus aiding bridge
Fig. 11. Time-dependent quadratic fit of median values for the ex- owners and managers in assessing evolving seismic risk at vari-
tensive damage state ous stages of bridge service life. Furthermore, by identifying a
tractable time-varying quadratic format for fragility parameters
this method can be easily integrated into regional risk assessment
Time Variant Quadratic Model for the Fragility and loss estimation software. The fragility parameters and corre-
Parameters sponding fragility curves derived in this paper reflect the time-
dependent seismic fragility of the representative MSC steel bridge
To continuously evaluate and predict the vulnerability of the cor- with corrosion deterioration models as discussed in the previous
roded bridge, fragility parameters are required at different points sections and are not intended for application to portfolios of struc-
in time along its service life. Tractable functions for time- tures. However, the approach presented herein illustrates the im-
dependent fragility models provide an efficient way to achieve portance of integrating the deterioration of key structural
this by adopting curve fitting techniques on the fragility param- components, such as columns and bearings, influence on dynamic
eters as a function of time. While for ease of presentation, Fig. 10 behavior, and propagation to affect bridge component and system
illustrates the effect of corrosion at five different points in time, fragility. Furthermore, the methodology for analysis and time-
the median PGA and dispersion of the lognormal fragility curves dependent seismic fragility format, including assessment of qua-
for each damage state are assessed for the case study bridge at a dratic models for time-variant lognormal fragility parameters, can
total of 10 different points in time along its service life. Analytical be applied in the future to additional bridge types and classes, or
functions are assessed to represent the change in median and dis- portfolios, of bridges with gross geometric variation.
persion values in time due to corrosion. Such time-dependent Additionally, as outlined by Gardoni et al. 共2002兲, the ability
models offer the advantage of estimating bridge fragility param- to incorporate experimental or field data as it emerges can offer a
eters at any point in time for the given bridge and corrosion valuable opportunity to improve the accuracy of bridge vulner-
parameters without the need to conduct complete fragility analy- ability estimates, particularly for models targeted at bridge-
ses. With respect to curve fitting techniques, a quadratic model of specific fragility analysis. Over the years, the Bayesian
the form framework has emerged as a potential tool for fragility estimation
of structures 共Singhal and Kiremidjian 1998; Der Kiureghian
parameter共t兲 = p1t2 + p2t + p3 共17兲
2002; Gardoni et al. 2002; Berahman and Behnamfar 2007; Graf
is found to best fit the time-dependent variation in fragility pa- et al. 2009兲. Future opportunities exist for exploiting such ap-
rameters, where parameter共t兲 is either the value of median, m共t兲, proaches as the Bayesian techniques for updating the time-
or dispersion, ␨共t兲, at t years and p1, p2, and p3 = quadratic coef- dependent component and system fragility models with field data,
ficients from regression analysis. Such proposed quadratic models such as improved estimates of corrosion modeling parameters.
have also been adopted in the past to proficiently demonstrate the Similarly, multidimensional fragility curves or fragility surfaces
loss of bridge resistance over time 共Enright and Frangopol 1999兲. 共Gardoni et al. 2002; Koutsourelakis 2010兲 can also be developed
Fig. 11 shows an example of the quadratic coefficients estimated as a part of future research endeavors exploiting the time-
for the time-dependent median of the bridge fragility at the ex- dependent fragility format and propagation of joint component
tensive damage state. The coefficients of the quadratic equation aging phenomena presented in this study.
Table 4. Coefficients of Quadratic Interpolation for the Median and Dispersion Values at Different Damage States
Quadratic coefficients
Median Dispersion
Damage state p1_m p2_m p3_m p1_␨ p2_␨ p3_␨
Slight −9.8⫻ 10 −6
⫺0.0002 0.28 −7.3⫻ 10 −6
⫺0.002 0.70
Moderate −1.5⫻ 10−5 ⫺0.0006 0.53 −7.0⫻ 10−6 ⫺0.0009 0.63
Extensive −1.8⫻ 10−5 ⫺0.0009 0.67 −8.6⫻ 10−6 ⫺0.0007 0.65
Complete −9.7⫻ 10−6 ⫺0.0023 0.89 −1.097⫻ 106 ⫺0.00169 0.72
Table 5. Capacity Limit States for Different Bridge Components for the Chosen Bridge 共Adapted from Nielson and DesRoches 2007a兲
Slight Moderate Extensive Complete
Component Median Dispersion Median Dispersion Median Dispersion Median Dispersion
Columns 1.29 0.59 2.10 0.51 3.52 0.64 5.24 0.65
Fixed bearings—longitudinal 6.0 0.25 20.0 0.25 40.0 0.47 187 0.65
Fixed bearings—transverse 6.0 0.25 20.0 0.25 40.0 0.47 186.6 0.65
Expansion bearings—longitudinal 37.4 0.60 104.2 0.55 136.1 0.59 186.6 0.65
Expansion bearings—transverse 6.0 0.25 20.0 0.25 40.0 0.47 187 0.65
Abutment—passive 37.0 0.46 146.0 0.46 N/A N/A N/A N/A
Abutment—active 9.8 0.70 37.9 0.90 77.2 0.85 N/A N/A
Abutment—transverse 9.8 0.70 37.9 0.90 77.2 0.85 N/A N/A
Conclusions seismic vulnerability and enable more accurate estimates of po-

tential damage, life cycle cost, and needed rehabilitation. The
This paper provides a framework for time-dependent fragility paper has also highlighted opportunities for future work, includ-
analysis of corroded bridge performance in seismic events, illus- ing field-condition informed updating of the fragility curves or
trating the impact of aging on component and system reliability. A multidimensional fragility surfaces conditioned upon various
case study MSC steel girder bridge representative of median geo- structural or aging parameters to improve the analysis for bridge-
metric characteristics of the CSUS bridge class is considered in specific fragility parameters. Given the significance of this study’s
the seismic performance assessment though variability in struc- findings, future work should also address time-dependent fragility
tural component, material, and corrosion modeling parameters of additional bridge types and geometries, generalized classes of
that are explicitly modeled. Susceptibility to corrosion attacks bridges, as well as deterioration mechanisms.
from the application of chloride-laden deicing salts commonly
used across the United States is considered in the simulation of
bridge aging and deterioration. The degradation mechanisms con- Acknowledgments
sidered include the corrosion deterioration of RC columns due to
reinforcement area loss, the deterioration of steel bridge bearing The writers gratefully acknowledge the support from the National
assemblies due to corrosion of steel anchor bolts, the transverse Science Foundation through grant CMMI-0928493.
keeper plates, and the buildup of corrosion debris. The deteriora-
tion of these components affects the lateral force resisting system
of bridges under seismic loading, resulting in reduced moment Appendix
capacity and yield curvature of the columns, reduced ultimate
lateral strength of the fixed and expansion bearings, and increased The capacity limit states adopted for each bridge component are
coefficient of friction in the bearings due to debris accumulation. listed in Table 5.
The nonlinear time history analysis illustrates that when the de-
terioration of bridge components is considered individually, there
is a significant shift in the dynamic response of the bridge pro- References
ducing an increase in the seismic demand on the individual com-
ponents. However, joint consideration of the component corrosion
Akgül, F., and Frangopol, D. M. 共2004兲. “Lifetime performance analysis
effects reveals that while the seismic demand on some compo- of existing prestressed concrete bridge superstructures.” J. Struct.
nents 共e.g., RC columns and expansion bearings兲 shows a steady Eng., 130共12兲, 1889–1903.
increase along the service life of the bridge, there is a decrease in Albrecht, P., and Naeemi, A. 共1984兲. Performance of weathering steel in
the demand on some components such as the fixed bearings in the bridges, National Cooperative Highway Research Program, Washing-
longitudinal direction. ton, D.C.
A full probabilistic analysis is conducted to evaluate the seis- Aquino, W., and Hawkins, N. M. 共2007兲. “Seismic retrofitting of corroded
mic fragility of the MSC steel bridge given uncertainty in bridge, reinforced concrete columns using carbon composites.” ACI Struct. J.,
ground motion, and corrosion parameters. The bridge system fra- 104共3兲, 348–356.
gility curves for each damage state reveal a significant increase in ASCE. 共2009兲. “ASCE: Infrastructure fact sheet.” 具http://www.
the bridge system vulnerability over time due to aging. For ex- infrastructurereportcard.org/sites/default/files/RC2009_bridges.pdf典
ample, after 75 years of exposure to chlorides, the median value 共May 22, 2009兲.
Basoz, N., and Kiremidjian, A. S. 共1999兲. “Development of empirical
PGA for the complete damage state decreases by 27%. Simple
fragility curves for bridges.” ASCE Technical Council on Lifeline
quadratic functions are introduced to enable easy assessment of
Earthquake Engineering Monograph, ASCE, Reston, Va.
the time-dependent shift in median and dispersion values of seis-
Basoz, N., and Mander, J. B. 共1999兲. Enhancement of the highway trans-
mic fragility for each damage state due to corrosion. The findings portation module in HAZUS, National Institute of Building Sciences,
highlight the importance of considering the effects of aging and Washington, D.C.
deterioration on the seismic vulnerability of bridges. Moreover, Berahman, F., and Behnamfar, F. 共2007兲. “Seismic fragility curves for
the time-dependent fragility format presented provides a straight- un-anchored on-grade steel storage tanks: Bayesian approach.” J.
forward approach to incorporate the influence of bridge age and Earthquake Eng., 11共2兲, 166–192.
exposure condition within seismic risk assessment packages. Broomfield, J. P. 共1997兲. Corrosion of steel in concrete, Taylor & Francis,
Such models offer more realistic estimates of corroded bridge London.
Buckle, G. 共2006兲. Seismic retrofitting manual for highway structures, Lindquist, L. 共2008兲. “Corrosion of steel bridge girder anchor bolts.” MS
MCEER, Buffalo. thesis, Georgia Institute of Technology, Atlanta, Georgia.
Choe, D., Gardoni, P., Rosowsky, D., and Haukaas, T. 共2008兲. “Probabi- Liu, C. 共2005兲. “Evolutionary multiobjective optimization in engineering
listic capacity models and seismic fragility estimates for RC columns management: An empirical study on bridge deck rehabilitation.”
subject to corrosion.” Reliab. Eng. Syst. Saf., 93共3兲, 383–393. Proc., 6th Int. Conf. on Parallel and Distributed Computing, Applica-
Choe, D., Gardoni, P., Rosowsky, D., and Haukaas, T. 共2009兲. “Seismic tions and Technologies, PDCAT 2005, Institute of Electrical and Elec-
fragility estimates for reinforced concrete bridges subject to corro- tronics Engineers Computer Society, Dalian, China, 773–777.
sion.” Struct. Safety, 31共4兲, 275–283. Mackie, K., and Stojadinovic, B. 共2001兲. “Probabilistic seismic demand
Cimellaro, G., Reinhorn, A. M., and Bruneau, M. 共2010兲. “Seismic resil- model for California highway bridges.” J. Bridge Eng., 6共6兲, 468–
ience of a hospital system.” Struct. Infrastruct. Eng., 6共1兲, 127–144. 481.
Cornell, C. A., Jalayer, F., Hamburger, R. O., and Foutch, D. A. 共2002兲. Mander, J. B., Kim, D., Chen, S., and Premus, G. 共1996兲. Response of
“Probabilistic basis for 2000 SAC Federal Emergency Management steel bridge bearings to reversed cyclic loading, NCEER, Buffalo.
Agency steel moment frame guidelines.” J. Struct. Eng., 128共4兲, 526– Mazzoni, S., McKenna, F., Scott, M. H., and Fenves, G. L. 共2009兲. Op-
533. enSees command language manual, Univ. of California, Berkeley,
Czarnecki, A. A., and Nowak, A. S. 共2008兲. “Time-variant reliability Calif.
profiles for steel girder bridges.” Struct. Safety, 30共1兲, 49–64. Melchers, R. E. 共1999兲. Structural reliability analysis and prediction,
Der Kiureghian, A. 共2002兲. “Bayesian methods for seismic fragility as- Wiley, New York 共Import兲.
sessment of lifeline components.” ASCE Council on Disaster Reduc- Melchers, R. E., and Frangopol, D. M. 共2008兲. “Probabilistic modelling
tion and Technical Council on Lifeline Earthquake Engineering of structural degradation.” Reliab. Eng. Syst. Saf., 93共3兲, 363–500.
Monograph 21, ASCE, Reston, Va., 61–77. Montemor, M. F., Simões, A. M. P., and Ferreira, M. G. S. 共2002兲.
Der Kiureghian, A., and Dakessian, T. 共1998兲. “Multiple design points in “Chloride-induced corrosion on reinforcing steel: From the funda-
first and second-order reliability.” Struct. Safety, 20共1兲, 37–49. mentals to the monitoring techniques.” Cem. Concr. Compos., 25共4–
Edwards, R. V. 共2006兲. Processing random data, World Scientific, Sin- 5兲, 491–502.
gapore. Nielson, B. 共2005兲. “Analytical fragility curves for highway bridges in
Enright, M. P., and Frangopol, D. M. 共1998兲. “Probabilistic analysis of moderate seismic zones.” Ph.D. thesis, Georgia Institute of Technol-
resistance degradation of reinforced concrete bridge beams under cor- ogy, Atlanta, Ga.
rosion.” Eng. Struct., 20共11兲, 960–971. Nielson, B. G., and DesRoches, R. 共2007a兲. “Analytical seismic fragility
Enright, M. P., and Frangopol, D. M. 共1999兲. “Condition prediction of curves for typical bridges in the Central and Southeastern United
deteriorating concrete bridges using Bayesian updating.” J. Struct. States.” Earthquake Spectra, 23共3兲, 615–633.
Eng., 125共10兲, 1118–1125. Nielson, B. G., and DesRoches, R. 共2007b兲. “Seismic fragility methodol-
Fang, C., Lundgren, K., Chen, L., and Zhu, C. 共2004兲. “Corrosion influ- ogy for highway bridges using a component level approach.” Earth-
ence on bond in reinforced concrete.” Cem. Concr. Res., 34共11兲, quake Eng. Struct. Dyn., 36共6兲, 823–839.
2159–2167. Nowak, A. S., and Cho, T. 共2007兲. “Prediction of the combination of
FHWA. 共2009兲. “FHWA bridge programs NBI data.” 具http:// failure modes for an arch bridge system.” J. Constr. Steel Res.,
www.fhwa.dot.gov/bridge/britab.cfm典共March 29, 2009兲. 63共12兲, 1561–1569.
Frangopol, D. M., Lin, K.-Y., and Estes, A. 共1997兲. “Reliability of rein- Padgett, J. E., and DesRoches, R. 共2008兲. “Methodology for the develop-
forced concrete girders under corrosion attack.” J. Struct. Eng., ment of analytical fragility curves for retrofitted bridges.” Earthquake
123共3兲, 286–297. Eng. Struct. Dyn., 37共8兲, 1157–1174.
Gardoni, P., Der Kiureghian, A., and Mosalam, K. M. 共2002兲. “Probabi- Pantazopoulou, S., Bonacci, J., Sheikh, S., Thomas, M., and Hearn, N.
listic capacity models and fragility estimates for reinforced concrete 共2001兲. “Repair of corrosion-damaged columns with FRP wraps.” J.
columns based on experimental observations.” J. Eng. Mech., Compos. Constr., 5共1兲, 3–11.
128共10兲, 1024–1038. Rix, G., and Fernandez, J. 共2004兲. “Earthquake ground motion simula-
Graf, M., Nishijima, K., and Faber, M. 共2009兲. “Bayesian updating in tion.” 具http://geosystems.ce.gatech.edu/soil_dynamics/research/
natural hazard risk assessment.” Australian Journal of Structural En- groundmotionsembay/典共Oct. 19, 2009兲.
gineering, 9共1兲, 35–44. Shinozuka, M., Feng, M. Q., Kim, H., and Kim, S. 共2000兲. “Nonlinear
Hoeke, L. J. L., Singh, P. M., Moser, R. D., Kahn, L. F., and Kurtis, K. E. static procedure for fragility curve development.” J. Eng. Mech.,
共2009兲. “Degradation of steel girder bridge bearing systems by corro- 126共12兲, 1287–1295.
sion.” Corrosion 2009, NACE—Int. Corrosion Conf. Series, National Silano, L. G., and Brinckerhoff, P. 共1993兲. Bridge inspection and reha-
Association of Corrosion Engineers International, Atlanta, Ga. bilitation, Wiley-IEEE, New York.
Hoffman, P. C., and Weyers, R. E. 共1996兲. “Probabilistic durability analy- Singhal, A., and Kiremidjian, A. S. 共1998兲. “Bayesian updating of fragili-
sis of reinforced concrete bridge decks.” Proc., 1996 7th Specialty ties with application to RC frames.” J. Struct. Eng., 124共8兲, 922–929.
Conf. on Probabilistic Mechanics and Structural Reliability, ASCE, Stewart, M. G., and Rosowsky, D. V. 共1998兲. “Time-dependent reliability
Worcester, Mass., 290–293. of deteriorating reinforced concrete bridge decks.” Struct. Safety,
Hwang, H., Liu, J. B., and Chiu, Y. 共2001兲. “Seismic fragility analysis of 20共1兲, 91–109.
highway bridges.” 具http://hdl.handle.net/2142/9267典共Mar. 12, 2010兲. Straub, D., and Der Kiureghian, A. 共2008兲. “Improved seismic fragility
Kayser, J. R., and Nowak, A. S. 共1989兲. “Capacity loss due to corrosion modeling from empirical data.” Struct. Safety, 30共4兲, 320–336.
in steel-girder bridges.” J. Struct. Eng., 115共6兲, 1525–1537. Thoft-Christensen, P., Jensen, F. M., Middleton, C. R., and Blackmore, A.
Komp, M. 共1987兲. “Atmospheric corrosion ratings of weathering steels- 共1996兲. “Assessment of the reliability of concrete slab bridges.” Conf.
calculation and significance.” Materials Performance, 26共7兲, 42–44. or Workshop Item, Univ. of Cambridge, Cambridge, U.K., 具http://
Koutsourelakis, P. 共2010兲. “Assessing structural vulnerability against publications.eng.cam.ac.uk/6361/典共Mar. 5, 2010兲.
earthquakes using multi-dimensional fragility surfaces: A Bayesian Val, D., Stewart, M., and Melchers, R. 共2000兲. “Life-cycle performance
framework.” Probab. Eng. Mech., 25共1兲, 49–60. of RC bridges: Probabilistic approach.” Comput. Aided Civ. Infra-
Li, J., Gong, J., and Wang, L. 共2009兲. “Seismic behavior of corrosion- struct. Eng., 15共1兲, 14–25.
damaged reinforced concrete columns strengthened using combined Vu, K. A. T., and Stewart, M. G. 共2000兲. “Structural reliability of concrete
carbon fiber-reinforced polymer and steel jacket.” Constr. Build. bridges including improved chloride-induced corrosion models.”
Mater., 23共7兲, 2653–2663. Struct. Safety, 22共4兲, 313–333.
Wen, Y. K., and Wu, C. L. 共2001兲. “Uniform hazard ground motions for gram, Washington D.C., 357.
Mid-America cities.” Earthquake Spectra, 17共2兲, 359–384. Whiting, D., Stejskal, B., and Nagi, M. 共1990兲. Condition of Prestressed
Weyers, R. E., Fitch, M., Larsen, E., Al-Qadi, I., and Hoffman, P. 共1994兲. concrete bridge components: Technology review and field surveys,
“Concrete bridge protection and rehabilitation: Chemical and physical Federal Highway Administration, Washington, D.C.
techniques.” Service life estimates, Strategic Highway Research Pro-
View publication stats Downloaded 15 Dec 2010 to 168.7.220.129. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org

Aging Considerations in The Development of Time Dependent Seismic Fragility Curves.

Uploaded by

Copyright:

Available Formats

Aging Considerations in The Development of Time Dependent Seismic Fragility Curves.

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Aging Considerations in The Development of Time Dependent Seismic Fragility Curves.

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Aging Considerations in the Development of Time-Dependent Seismic Fragility

Article in Journal of Structural Engineering · December 2010

Jayadipta Ghosh Jamie Ellen Padgett

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

Introduction bearings, and corrosion of steel reinforcement in RC columns

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1497

1498 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1499

Cover depth 共x兲 cm 3.81 0.20 0.8

1500 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010

Ultimate Lateral Strength: Fult (N)

Fig. 4. Reduction of ultimate lateral strength of fixed bearings along

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1501

Shear Strength (N)

Fig. 6. Shear strength degradation of and demand on anchor bolt due

1502 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1503

Kyield = 0.003 (1/m) −100

Longitudinal Loading & Response

1504 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010

Impact of Corrosion-Induced Component

The deterministic analysis for the corroded case study bridge

Methodology for Development of Fragility Curves

Seismic fragility curves represent the probability of structural

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1505

1506 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010

System Fragility − Extensive Damage System Fragility − Complete Damage

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1507

1508 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010

Conclusions seismic vulnerability and enable more accurate estimates of po-

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1509

1510 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1511

You might also like