Soil Dynamics and Earthquake Engineering 93 (2017) 135–146

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

journal homepage: www.elsevier.com/locate/soildyn

Seismic failure mode of coastal bridge piers considering the effects of MARK
corrosion-induced damage
⁎
Wei Yuan, Anxin Guo , Hui Li
Ministry-of-Education Key Laboratory of Structural Dynamic Behavior and Control, School of Civil Engineering, Harbin Institute of Technology, the 2nd
Campus, Harbin 150090, China

A R T I C L E I N F O A BS T RAC T

Keywords: Coastal bridges exposed to an aggressive environment are vulnerable to corrosion damage, which reduces the
Coastal bridge pier seismic resisting capacity of the structures. Focusing on the corrosion varying along the column height of the
Marine environment coastal bridge piers, this paper investigates the time-dependent failure mode and equivalent plastic hinge length
Corrosion of the aging bridge piers under seismic excitation in the whole life cycle. First, the possible seismic failure modes
Seismic failure mode
and the method to estimate the equivalent plastic hinge length are analyzed and discussed. Then, the corrosion
Hysteretic response
initiation time and performance deterioration of the reinforcement are presented. After that, the finite element
Equivalent plastic hinge length
models of the sound and aging bridge piers are introduced based on the OpenSees software package and the
time-dependent constitutive models of the reinforcement. According to the proposed flowchart of the
computational procedure, a numerical simulation is conducted to investigate the plastic hinge evolution
process of the coastal bridge pier. The analysis results indicate that the seismic failure mode of the continuously
corroded bridge pier varies with the service time, and the plastic hinge has the possibility of transferring from
the column end to the bottom of the splash and tidal zone.

1. Introduction
Coastal bridges exposed to a marine environment always suffer
significant chloride-induced corrosion. During the whole service life of
the bridge, the continuous penetration of chloride ions induces
cracking and spalling of the concrete cover [1], degradation of the
bond capacity [2] and decrease of the material strength [3,4]. In
seismic-prone regions, the corrosion damage of the bridges would
significantly reduce the lateral load-resisting capacity of the structure
in a seismic event. Understanding the seismic performance of the
corroded coastal bridges is an important issue for researchers and
engineers to assure the safety of the aging structures when sustaining
earthquake excitations.
In a marine environment, chloride penetration directly induces
material deterioration. To investigate the time-dependent performance
of the structures, many researchers have investigated the corrosion
penetration model and material deterioration model. Based on Fick's
law, the chloride diffusion models in saturated [5,6] and non-saturated
[7,8] concrete have been developed and employed for determining the
corrosion initiation time of the reinforcement. For the reinforced
concrete (RC) structures, the corrosion-induced performance degrada-
tion of reinforced steel is recognized to be the dominant cause of
premature failure of the structures [9,10]. Focusing on this issue, some
researchers have conducted experiments to investigate the time-
dependent corrosion damage of the reinforced bars, including the
mass loss of the steel and the reduction of strength, diameter and
ductility [11,12]. During the experiments, the corrosion of the speci-
mens was always achieved by exposure in the natural environment [13]
or through an accelerated corrosion method [14]. In addition, limited
efforts have been made to investigate the corrosion-induced cracking
[15], degradation of stiffness [16] and delamination [17] of the
concrete cover.
In the last decade, some researchers and engineers also devoted to
establish appropriate methods for the fragility analysis, life-cycle
reliability assessment and cost estimation of corroded bridges [18].
Based on currently available probabilistic models for pristine bridges,
Choe et al. [19,20] proposed novel probabilistic models for the seismic
demand and capacity (deformation and shear force) of corroded RC
bridges. Seismic fragility increment functions were also developed to
determine the fragility of corroded RC bridge columns in various
environment and material conditions [21,22]. Considering the reduc-
tion in cross-sectional area of the reinforcement and in stiffness due to
concrete cover spalling, Simon et al. [23] and Zhong et al. [24]
investigated the seismic responses and fragility of deteriorated RC

⁎
Corresponding author.
E-mail address: guoanxin@hit.edu.cn (A. Guo).

http://dx.doi.org/10.1016/j.soildyn.2016.12.002
Received 17 August 2016; Received in revised form 16 November 2016; Accepted 2 December 2016
0267-7261/ © 2016 Elsevier Ltd. All rights reserved.
W. Yuan et al.
bridges. The simulation results indicated that the losses in strength and
stiffness only marginally influenced the seismic fragility of the analyzed
bridge. Akiyama et al. [25] proposed a novel computational procedure
to evaluate the life-cycle seismic reliability of corroded RC bridges due
to the airborne chloride hazard in marine environment. In addition,
Kumar et al. [26] proposed a probabilistic approach to compute the
life-cycle cost (LCC) of corroded RC bridges considering the uncertain-
ties of the ground motions and the cumulative damage associated with
low-cycle fatigue and chloride- induced corrosion of reinforcement.
Ghosh and Padgett [27] developed a probabilistic approach to explicitly
incorporate time-dependent seismic vulnerability of aging bridges in
the seismic loss estimation via a framework based on a non-homo-
geneous Poisson process. Incorporating the visual inspection and the
seismic vulnerability analysis, Zanini et al. [28] established an inte-
grated procedure for assessing the maintenance state and related costs
of existing road bridges.
In the marine environment, the coastal bridge piers would suffer
corrosion varying along the column height in the submerged zone,
splash and tidal zone and atmospheric zone due to different types of
exposure. Many research efforts had been devoted to investigate the
time-dependent performance of the materials in one of the three zones
[4,29]. It is widely accepted that the materials in splash and tidal zone
has the highest risks of corrosion [30,31]. For conventional bridge
piers, the seismic-induced damage always occurs in the column bottom
due to large bending moment. However, if the load-resisting capacity at
specific region, such as the splash and tidal zone of the coastal bridge, is
more significantly reduced than the other regions due to corrosion, the
failure of the aging structures would deviate from the expected failure
status, and consequently affects the seismic performance of the
structures.
The failure mode is an important property affecting the seismic
performance of the civil infrastructures. For the sound bridge piers, the
seismic failure mode can be classified into three categories, i.e., flexure
failure, flexure-shear failure and shear failure, depending on the shear
span ratio and the shear strength of the components [32]. By testing 13
RC circular columns with different degrees of corrosion and different
axial load ratios, Ma et al. [33] experimentally investigated the time-
dependent failure phenomena of the uniform corroded components
and concluded that the failure pattern could change from flexural
failure to flexure-shear failure or shear failure when the stirrups were
subjected to serious corrosion.
The plastic hinge length of the columns has a direct correlation with
the seismic capacity of the structures. Due to the nonlinearity of the
material, the bond-slip behavior between the concrete and the re-
inforcement, and the shear-bending interactions, the experimental
method is a preferred way to estimate its value [34]. Park and
Priestley [35] conducted the experiment using four full-size reinforced
concrete columns with the application of different axial loads. The
conclusion indicated that the plastic hinge length was not sensitive to
Fig. 1. Flexural failure process of the bridge piers: (a) force-displacement curve, (b) stress distribution.
136
Soil Dynamics and Earthquake Engineering 93 (2017) 135–146
the axial load level. Considering different grades of reinforced bar,
Paulay and Priestley [36] modified the expression proposed by
Priestley and Park [37] and demonstrated that the plastic hinge length
can be calculated when the column height, the column diameter and
the strength of the flexural reinforcement of a column are determined.
This paper presents a methodology to investigate the seismic failure
mode of the corroded bridge piers in a marine environment. The main
contributions of this study include: (1) the possible time-dependent
seismic failure modes of the coastal bridge piers are investigated for the
structures with corrosion varying along the height in such specific
exposure; (2) a method is proposed for calculating the equivalent
plastic hinge length (EPHL) of the corroded bridge piers; (2) a
computational procedure is developed for distinguishing the seismic
failure mode and determining the EPHL of the corroded bridge piers to
obtain insight into the seismic performance of deteriorating bridges
piers. The main content of this study is organized as five parts. First,
the possible seismic failure modes of corroded bridge piers, the method
for estimating the EPHL and the corresponding flowchart for the
analysis are studied and discussed. Then, the method to determine the
corrosion initiation time and the performance deterioration of the steel
reinforcement are introduced, considering the corrosion varying along
the column height of the coastal bridge piers. Next, the finite element
models of the sound and aging bridge piers are established based on
the OpenSees software package and the time-dependent constitutive
models of the reinforcement. Using the proposed method, a numerical
example is conducted to investigate the time-dependent failure modes
of the coastal bridge piers and the corresponding EPHL. The simula-
tion results indicate that the plastic hinge would transfer from the
column end to the splash and tidal zone when the difference of the
corrosion damage between the submerged zone and the splash and
tidal zone reaches a certain degree.
2. Seismic failure mode of the flexural bridge piers with
corrosion varying along the column height
2.1. Failure process of the common bridge piers
Due to the water depth, the coastal bridges always have a relatively
large height. Therefore, this study is mainly focused on the cantilever
piers with a large shear-span ratio. For this type of structure without
corrosion damage, flexural failure is always sustained in a seismic
event. Under the combined action of constant axial load and mono-
tonically increasing lateral force, the failure process for the flexural
bridge piers can be divided into three stages separated by four critical
moments, as indicated with circles in the force-displacement curve
shown in Fig. 1(a). To explain the damage degree at those critical
moments more clearly, Fig. 1(b) depicts the corresponding stress
distribution of the reinforcement and the concrete [38]. As shown in
the figure, the compression stress of the two types of materials in the
W. Yuan et al.
whole cross section is identical when only the axial load is sustained at
point Ia. With the increase in the lateral force, both the concrete and
steel reinforcement experience continuously increasing damage from
linear to nonlinear, accompanied by concrete cracking at point IIa and
reinforcement yielding at point IIb. Finally, the load-carrying capacity
of the pier reaches the ultimate state at point IIIa with the crush of the
concrete in the compression zone.
2.2. Possible seismic failure modes of coastal bridge piers with
corrosion varying along the column height
For the flexural bridge piers, the plastic hinge of such sound
structures or structures with uniform corrosion would generally appear
at the column end because of the large bending moment and the
identical performance of all of the cross sections when subjected to
earthquake excitation. However, due to the exposure to different
environments, the coastal bridge piers always sustain corrosion varying
along the column height of the structures. Consequently, the corrosion
is generally divided into three regions, designated as the atmospheric
zone, the splash and tidal zone and the submerged zone. Considering
the effects of the corrosion varying along the column height under
different exposures, the position of the plastic hinge and the corre-
sponding seismic failure modes of the bridge pier become complicated.
Fig. 2 shows the possible corrosion zones along the elevation of the
piers with the increase of the water depth. For those cases, the possible
seismic failure modes of the coastal bridge pier are discussed as
follows:
(1) Case 1: bridge pier is located over the splash and tidal zone
In this situation, the whole pier is exposed to the atmospheric
environment, as shown in Fig. 2(a). With the airborne chloride, the
coastal bridges mainly sustain uniform corrosion along the whole
height of the piers. Consequently, the plastic hinge is easily
anticipated to be generated at the column end under the seismic
excitation.
(2) Case 2: splash and tidal zone is located in the column bottom
In the nearshore region, the water depth is relatively shallow.
As shown in Fig. 2(b), there is no obvious submerged zone, and the
splash and tidal zone is located at the bottom of the bridge piers. In
this case, the more serious corrosion in the splash and tidal zone
would further reduce the load-resisting capacity of the bottom
plastic hinge region. Obviously, the failure of the bridge pier will
occur at the bottom of the piers when an earthquake attacks the
structures.
(3) Case 3: splash and tidal zone is located over the common bottom
plastic hinge region
With a further increase in the water depth, the splash and tidal zone
Fig. 2. Possible cases of corrosion zones distributed along the whole height of the piers in the marine environment: (a) case 1; (b) case 2; and (c) case 3.
137
Soil Dynamics and Earthquake Engineering 93 (2017) 135–146
is extended from the bottom to the middle of the bridge piers, as shown
in Fig. 2(c). In this situation, the maximum bending moment induced
by the seismic excitation is generated at the bottom of the pier due to
the large arm. However, the cross sections in the splash and tidal zone
have weak load-resisting capacity because of the more serious perfor-
mance deterioration induced by the corrosion varying along the column
height. Considering the two factors, the eventual seismic failure mode
of the bridge piers shows several possibilities, depending upon the
corrosion level, location and length of the three different corrosion
zones.
In such a case, the failure mode of the degraded bridge pier can be
determined from the curvature distribution along the height of the
column when the component reaches the ultimate state. For a flexural
component, the plastic hinge is generally defined to have occurred
initially and to become fully formed when the bending moment first
exceeds the yield moment and reaches the ultimate moment of the
critical cross section, respectively. Consequently, the ultimate state can
be defined as the condition when a plastic hinge is fully formed.
The strain of reinforcement has a close relationship to the curvature
[39]. To investigate the effects of corrosion varying along the column
height on curvature distribution, the strain distribution of one corroded
reinforced bar is studied first and shown in Fig. 3. The strain of the
longitudinal reinforcement of the common flexural columns is well
known to decrease continuously from the bottom to the top due to the
corresponding reduction in the bending moment. However, because
the area and strength of the reinforced bars of the splash and tidal zone
are more seriously reduced than the other regions, the strain at the
splash and tidal zone is consequently increased, resulting in a sudden
change of the strain at the interface between the two different corrosion
zones.
Because the corrosion damage of the coastal bridge is time-
dependent, the relationship between strain ε2a and ε3 also varies with
the service time as follows:
⎧ <ε2y (t1), ε3 (t1) = ε3y (t1)
⎪
ε2a (ti ) ⎨ = ε2y (t2 ), ε3 (t2 ) = ε3y (t2 )
⎪ >ε (t ), ε (t ) = ε (t )
⎩ 2y 3 3 3 3y 3 (1)
where t1, t2 and t3 are the times elapsed since the column was exposed to
a corrosive marine environment, where t1 < t2 < t3; ε2a (ti ) and ε2y (ti )
(i =1,2,3) are the actual strain and yield strain, respectively, of the
reinforced bar at the bottom of the splash and tidal zone, and ε3 (ti ) and
ε3y (ti ) (i =1,2,3) are the actual strain and yield strain, respectively, of the
reinforced bar at the bottom of the pier. Eq. (1) shows that the
maximum damage location in the steel bar has the tendency of
transferring from the submerged zone to the splash and tidal zone.
A similar phenomenon can also be found in the curvature distribu-
tion. As shown in Fig. 4, after servicing for a certain period, the sudden
change also appears at the interface of the neighboring corrosion zones.
W. Yuan et al. Soil Dynamics and Earthquake Engineering 93 (2017) 135–146

2.3. Equivalent plastic hinge length

In the seismic design code, the EPHL is a critical factor to evaluate

Fig. 3. Strain distribution of one corroded reinforced bar along the elevation of the pier.
Fig. 4. Curvature distribution at the ultimate state of the corroded column.
The curvature of the bottom cross section of the splash and tidal zone
φ2a is between its yield value φ2y and ultimate value φ2u when the cross
section at the column end reaches the ultimate state φ3 = φ3u . This
phenomenon indicates that another plastic hinge is forming at the
splash and tidal zone when the bottom plastic hinge is just fully
completed. With the increase in the performance difference of different
zones, it is reasonable that the critical cross section would transfer to
the bottom of the splash and tidal zone.
The above analysis shows that a total of five possible failure modes
of the coastal bridges are anticipated to occur at the ultimate states, as
summarized in Table 1. For the ultimate states I and II, the corrosion of
the splash and tidal zone is not significant, and the plastic hinge still
forms at the bottom of the submerged (BS) zone similar to the
conventional bridge piers. When the corrosion degree in the splash
and tidal zone becomes more serious than in the other two regions, two
plastic hinges will be generated at the BS and the bottom of splash and
tidal (BST) zone. Furthermore, if the corrosion degree at the splash and
tidal zone is very significant, the plastic hinge is formed at the BST, and
the number is further reduced to one.
Table 1
Possible failure modes of the corroded pier in the ultimate states.
Ultimate state Description Number of Location of
plastic hinge plastic hinge
the ultimate deformation of the bridge piers. The curvature distribution
in the plastic hinge region of the common bridge piers is generally
simplified into elastic and constant inelastic curvature. For cantilever
columns, the relationship between the ultimate displacement and the
EPHL can be described as [37]:
Δu = φy l 2 /3 + (φu − φy ) lp (l − 0.5lp ) (2)
where lp and l are the EPHL and column height, respectively; φy and φu
are the yield curvature and ultimate curvature, respectively; and Δu is
the ultimate displacement.
Eq. (2) has commonly been used to estimate the lp of columns.
Based on the force-displacement and moment-curvature data obtained
from tests [36], the lp used for seismic design and performance
assessment can be expressed as
lp = 0.08l + 0.022db fy (3)
where db is the diameter of the longitudinal bar in mm; and fy is the
yield strength of the longitudinal bar in MPa.
Although Eqs. (2) and (3) can be used for estimating the EPHL of
structures without damage, they cannot be employed directly for the
aging coastal bridges with corrosion varying along the column height in
the marine environment. However, the equivalence principle can be
referenced herein for this problem. Fig. 5 depicts the equivalent
curvature distribution at the ultimate states of the five possible seismic
failure modes listed in Table 1. The curvature distribution of the
corroded bridge pier is also divided into two parts consisting of elastic
and constant inelastic, while the difference here is that the equivalent
curvature is individually considered in each corrosion zone.
As shown in the figure, the EPHL at the splash and tidal zone and
the EPHL at the submerged zone are assumed to be lp2 and lp3,
respectively. According to the relationship between the curvature
distribution and the lateral displacement, they can be calculated by:
[l32 /2 − l33 /(6l )] φ3y + (φ3 − φ3y ) lp3 (l3 − 0.5 × lp3) = Δ2b (4)
Δy + (φ2a − φ2y ) lp2 (l1 + l2 − 0.5lp2 ) + (φ3 − φ3y ) lp3 (l − 0.5lp3) = Δu (5)
where l is the total height of the column; l1, l2 and l3 are the height of the
atmospheric zone, the splash and tidal zone and the submerged zone,
respectively; Δy and Δu are the yield and ultimate displacement of the
column, respectively; Δ2b is the lateral displacement at the top of the
submerged zone when the displacement at the pier top reaches Δu .
Depending on the actual corrosion status, either φ2a or φ3 is equal to its
ultimate curvature in the above two equations. For the two equations,
Eq. (5) is used to predict the EPHL of case I, II and V in Fig. 5. In those
cases, either lp2 or lp3 equals zero because one of the corrosion zones
does not reach its yield curvature. For cases III and IV with two plastic
hinges, the combination of Eqs. (4) and (5) can be employed to
estimate the values of lp2 and lp3. Actually, the two plastic hinges co-
exist in some cases. However, one of the plastic hinges is not fully
formed. Considering this problem, it is better to use the nominal EPHL
l′pi to express the EPHL as follows:
′
φip lpi = (φiu − φiy ) lpi (6)

I φ2b = φ2a < φ2y , φ3 = φ3u 1 BS where φip is the equivalent plastic curvature of the cross section at the
II φ2b < φ2a < φ2y , φ3 = φ3u 1 BS plastic hinge zone. For the splash and tidal zone, the φip equals
III φ2y < φ2a < φ2u , φ3 = φ3u 2 BS+BST φ2a − φ2y , while the φip equals φ3 − φ3y for the submerged zone.
IV φ2a = φ2u , φ3y < φ3 < φ3u 2 BS+BST The above analysis shows that the determination of the yield and
V φ2a = φ2u , φ3 < φ3y 1 BST ultimate displacement of the column, as well as the yield and ultimate
curvature of the cross section, are fundamental for calculating the
Notes: the φ2a , φ2b and φ3 are the actual curvatures of the cross section located at the EPHL of the corroded structures. For the five possible failure modes,
BST, the top of submerged (TS) zone and the BS, respectively; φ2y and φ2u are the yield the yield displacement of the column is taken as the integral result of
and ultimate curvature of the cross section at the BST; and φ3y and φ3u are the yield and the equivalent elastic curvature depicted in Fig. 5. As illustrated in
ultimate curvature of the cross section at the BS. Fig. 6(a), the ultimate displacement can be obtained from the force-

138
W. Yuan et al.
Fig. 5. Equivalent curvature distributions of the corroded piers.
displacement envelope curve of the bridge piers. The ultimate dis-
placement is reached when the load-resisting capacity of the compo-
nent is reduced to 85% of its maximum value or the hoop fracture
starts. According to Priestley and Park [37], the onset of hoop fracture
can be taken as the moment when the confined concrete strain reaches
the ultimate compression strain εc, ls , which can be expressed as
ρv fyh εsu
εc, ls = 0.004 + 1.4
fcc′
where εsu and fyh are the ultimate strain and yield strength of the
stirrups, respectively; ρv is the volumetric confinement ratio; and fcc′ is
the compressive strength of confined concrete. The yield curvature and
ultimate curvature can be determined from the moment-curvature
envelope curve of the corresponding cross sections. As shown in
Fig. 6(b), the yield curvature and ultimate curvature can be obtained
from the moment-curvature relationship according to the equivalent
bilinear method and the similar principle of determining ultimate
displacement.
2.4. Flowchart for distinguishing the seismic failure mode
According to the above discussions, the flowchart for distinguishing
the seismic failure mode and determining the EPHL of the corroded
bridge pier is illustrated in Fig. 7. From the figure, the procedure can be
summarized as follows:
(1) The finite element models of the bridge pier are first built by the
finite element analysis software, such as OpenSees [40], consider-
ing the time-dependent characteristics of the material properties
with the service time.
(2) Through the hysteresis analysis of the structures, the force-
displacement curve of the structure, the moment-curvature curve
of the cross section at the BST (M2, φ2 ), the BS (M3, φ3) and the
stress-strain curves of the outermost confined concrete of the two
cross sections (σc2, εc2 , σc3, εc3) are obtained.
(3) Based on the definitions in Fig. 6(b) and Eq. (7), the ultimate and
Fig. 6. Definition of: (a) ultimate displacement, and (b) yield curvature and ultimate curvature.
Soil Dynamics and Earthquake Engineering 93 (2017) 135–146
yield curvature and the corresponding occurrence sequence of the
failure and the yield of the two cross sections (φ2u , s3), (φ3u , s4 ),
(φ2y , s1) and (φ3y , s2 ) are determined from the hysteretic curve. The
smaller values of si represent the earlier occurrence of the
corresponding events.
(4) Sorting out the occurrence sequence of the four events in the
loading process, the number, location and status of the plastic
hinge, which govern the final failure mode of the corroded bridge
(7) pier, can be confirmed.
(5) From the obtained failure mode, the equivalent curvature distribu-
tion and the yield displacement of the corroded pier are deter-
mined. Furthermore, according to the force-displacement curve,
the ultimate displacement of the structure can also be calculated.
(6) Finally, the EPHL can be calculated using Eqs. (4)–(6).
3. Performance degradation of reinforcement due to
chloride corrosion
The failure mode of the corroded bridge pier under seismic
excitation is strongly dependent on the degree of corrosion of the
structures. Understanding the corrosion mechanism and establishing
the model to represent the material corrosion are the important basis
for the seismic analysis of the deteriorating structures. Generally, the
performance deterioration of concrete and reinforcements and the
reduction of bond-slip capacity between the two materials all con-
tribute to the structural degradation phenomenon. The reinforcement
corrosion is acknowledged to be the main cause among the those
factors [9,10]. Consequently, only the reinforcement degradation
model is considered in this study.
3.1. Corrosion initiation time
Chloride-induced corrosion is identified as one of the major reasons
for the performance deterioration of the coastal bridges. Under the
actions of capillary, diffusion, electrochemical migration and permea-
tion, chloride ions continuously penetrate into the concrete. For the
structures located in a marine environment, the diffusion process can
139
W. Yuan et al. Soil Dynamics and Earthquake Engineering 93 (2017) 135–146

Fig. 7. Flowchart for distinguishing the seismic failure mode and determining the EPHL of the corroded bridge pier.

be modeled by Fick's second law [41], where Cb is the bound chloride concentration; we is the evaporable
water content; DCl and DCl,ref are the effective diffusion coefficient and
∂Cf ∂ ⎛ * ∂Cf ⎞
= ⎜DCl ⎟ reference diffusion coefficient, respectively; F1 (h ), F2 (T ), F3 (t ) and
∂t ∂x ⎝ ∂x ⎠ (8) F4 (Cf ) are the modified factors that account for the influence of the
where Cf is the free chloride concentration; x and t are the depth relative humidity, the ambient temperature, the age of the concrete and
measured from the concrete surface and the time elapsed since the the free chloride concentration, respectively. The mathematical expres-
concrete has been exposed to the corrosive environment, respectively; sions of the parameters used in Eq. (10) are summarized in Table 2[5].
* is the apparent diffusion coefficient of the chloride ion and is
and DCl The corrosion is initiated when the total chloride concentration (Ct )
given by around the reinforced bars increases to a threshold value. In concrete,
the total chloride concentration can be obtained from the free and
*= DCl bound chloride concentration as follows:
DCl 1 ∂Cb
1+ we ∂Cf (9) Ct = Cb + we Cf (11)
with with
DCl = DCl,ref F1 (h ) F2 (T ) F3 (t ) F4 (Cf ) (10)

140
W. Yuan et al.
Table 2
Summary of parameters used for estimation of effective diffusion coefficient.
No. Function Assigned values
1 DCl.ref = 10 aw / c − b a = 2.4 ; b = 12.06
2 F1 (h ) = 1/{1 + [(1 − h )/(1 − h c )]4 } h c = 0.75
3 F2 (T ) = exp[U /R (1/Tref − 1/T )] U = 41800 (J /mol) ;
Tref = 293 (o K) R = 8.314 (J /mol.o K )
4 F3 (t ) = (tref /t )m m = 0.35; tref = 28 (day)
5 F4 (Cf ) = 1 − k ion (Cf )n k ion = 8.366 ; n = 0.5
Note: a and b in this table are empirical coefficients; w /c is the water-cement ratio; h and
h c are the relative humidity and the critical humidity level, respectively; U , R , Tref and T
are the activation energy for the diffusion process, the gas constant, the reference
temperature and the ambient temperature, respectively; tref and m are the reference time
and the empirical age coefficient, respectively; and k ion and n are empirical coefficients.
Cb = αCfβ
where α and β are binding constants with the assigned values of 1.05
and 0.36, respectively. With the given values of the parameters in Eqs.
(8)–(12) at the bridge site, the corrosion initiation time can be
calculated by solving the partial differential equation for the condition
of the surface chloride concentration of the reinforced bars reaching a
threshold value.
3.2. Performance degradation of reinforcement
After the initiation of corrosion, the corrosion effect is reflected
mainly in the decrease of both the diameter and the yield strength of
the steel bars [9]. In this study, uniform corrosion is assumed to occur
in each corrosion zone, and the time-dependent diameter of the
reinforced bars D (t ) can be expressed as [1]
⎧ D0 t ≤ Ti
⎪
D (t ) = ⎨ D0 − 0.023icorr (t − Ti ) Ti < t ≤ Tf
⎪0
⎩ Tf < t
where Ti and t are the corrosion initiation time and the time elapsed
since the concrete was exposed to the corrosive environment, respec-
tively; D0 is the initial diameter of the sound reinforced bar, and icorr is
the corrosion current density. Considering the different corrosion
environments, the value of icorr can be considered to be 3 μA /cm2 for
the atmospheric zone and 6 μA /cm2 for the splash and tidal zone [42].
During the corrosion process, the yield strength of the steel material
is also reduced. After t years, the residual yield strength fy (t ) can be
estimated by
fy (t ) = [1 − 0.005Qcorr (t )] fy0
with
Qcorr (t ) = {1 − [D (t )/ D0]2 } × 100
where fy0 is the initial yield strength of the reinforcement, and Qcorr (t ) is
the percentage of mass loss of the reinforcement after being in service
for t years.
4. Finite element model of the corroded piers
According to the flowchart in Fig. 7, finite element analysis is one
major content for determining the seismic failure mode of the corroded
bridge piers in a marine environment. In this study, the OpenSees
software package [40] is adopted for the structural analysis. Fig. 8
shows the schematic diagram of the finite element model. As shown in
Fig. 8(a), the uniaxial material models of Concrete01 and Reinforcing
Steel were employed for modeling the concrete and steel material. To
consider the strain penetration effects, the Bond_SP01 model was used
at the column end. Furthermore, to capture the spread of the plasticity
along the column and the rotation phenomenon at the column end due
Soil Dynamics and Earthquake Engineering 93 (2017) 135–146
to strain penetration effects, the nonlinear beam-column element and
zero-length element were adopted in this simulation. As shown in
Fig. 8(c), the zero-length element was set at the column end, between
the Node i and Node j. In the simulation, the translational degree-of-
freedom of the two nodes was constrained with each other to prevent
sliding between them.
To effectively represent the mechanical properties of the steel
material under cyclic loading, the fatigue parameters in the constitutive
model of Reinforcing Steel, i.e., damage strain range constant (α ),
material constant ( β ), strength reduction constant (Cd ) and ductility
constant (Cf ) are calibrated. The relationship between the four para-
meters is [43–45]:
Δεp = Cf (2Nf )−α (16)
and
Δεp = Cd ( fSR) β (17)
(12)
where Δεp is the total half cycle plastic strain amplitude, 2Nf is the
number of the half cycles until failure and fSR is the strength loss factor
per cycle.
By analyzing the 70 stress-strain curves of reinforcements obtained
from the low-cycle fatigue experiments [46–48], the values of Δεp and
2Nf for each test and the values of fSR for 20 tests were obtained.
According to the assumption of α = β [49] and the nonlinear fitting
procedure, the value of the fatigue parameters is Cf = 0.24 ,
α = β = 0.49 and Cd = 0.41, as shown in Fig. 9.
Based on the above method, the finite element model of the sound
bridge piers can be established by using the selected elements,
constitutive models and the corresponding parameters. To consider
the performance deterioration of damaged structures with corrosion
varying along the column height, the finite element model of the sound
structure must be modified by adjusting the area and strength of the
reinforcements separately for the three different corrosion zones.
5. Illustrative example
(13)
5.1. Bridge pier model
A reinforced concrete bridge pier subjected to continuous chloride-
induced corrosion in the marine environment is selected as the object
of the illustrative example. As shown in Fig. 10, the clearance height
and diameter of the cylindrical column are 10.5 m and 1.6 m,
respectively. Reinforced rebars (46#32) are used as the longitudinal
steel with the reinforcement ratio of 1.84%. The thickness of the
concrete cover measured from the concrete surface to the outside edge
(14) of the longitudinal steel is 50 mm. Circular stirrups (#16) are set at
both ends of the column with a pitch of 100 mm within a range of
1.8 m, while a pitch of 200 mm is used for the other end. The
(15) compressive strength of the concrete and the yield strength of the
stirrups and longitudinal bars are assumed to be 34.8 MPa, 345 MPa
and 400.0 MPa, respectively. Gravity loads applied to each column are
700 t with the corresponding axial compression ratio of 0.1. According
to the surrounding environment, the heights of the atmospheric zone,
the splash and tidal zone and the submerged zone are assumed to be
3.0, 4.0 and 3.5 m, respectively. In this study, the service life of the
structure is assumed to be 100 years, and 11 finite element models of
the deterioration structures are built with the time step of 10 years.
5.2. Performance deterioration of reinforcements
In this study, the corrosion effect in the submerged zone is not
considered because of the lack of oxygen and light [50]. Recommended
by Life-365 [51], the surface chloride concentration in the atmospheric
zone is increased with the accumulation rate of 0.10% per year until the
maximum value of 1.0% is reached. In the splash and tidal zone, the
process to the maximum value 0.8% is instantaneous. To obtain the
141
W. Yuan et al. Soil Dynamics and Earthquake Engineering 93 (2017) 135–146

Fig. 8. Fiber-based modeling method of the bridge pier.

apparent diffusion coefficient DCl * in Eq. (8), the values of we , w / c , h and

T should be determined. In this study, the we and w / c are assumed to be
0.08 and 0.40, respectively. By applying a nonlinear fit on the statistical
data for a coastal region, the sinusoidal function of the humidity and
the temperature are given as

⎛ 2π ( j − 159.043) ⎞
h = 0.731 + 0.016 sin ⎜ ⎟ (0 < j ≤ 365)
⎝ 365 ⎠ (18)

⎛ 2π ( j − 111.098) ⎞
T = 290.468 + 11.844 sin ⎜ ⎟ (0 < j ≤ 365)
⎝ 365 ⎠ (19) Fig. 10. Schematic diagram of the bridge pier.

where j is the j−th day in one year.

Considering the actual surrounding conditions, the local tempera-
ture function T is applied to both the splash and tidal zone and the
atmospheric zone, while the relative humidity function h is used just for
the atmospheric zone. Furthermore, due to the dry-wet cycle, the
concrete in the splash and tidal zone is assumed to be saturated. In this
study, the corrosion is assumed to be initiated when the total chloride
concentration around the reinforced bars increases to the threshold
value of 0.05% of the concrete weight.
The partial differential equation of Eq. (8) is calculated by the
COMSOL software. Fig. 11 shows the total chloride concentration
around the reinforced rebar versus the service time. From the
comparisons with the threshold value, the corrosion is seen to be
initiated at 8.2 years for the longitudinal reinforcement in the splash
and tidal zone and at 25.5 years for the atmospheric zone. The
corresponding corrosion initiation time for the stirrups is 5.5 and
20.6 years, respectively. As the distance between the stirrup surface to
the outside concrete surface is smaller than the distance between the
longitudinal reinforcement surface to the outside concrete surface, the
corrosion initiation time of the stirrup is earlier than the corrosion
initiation time for the longitudinal reinforcement.
Based on the deteriorating models of the steel material, Fig. 12
shows the time-dependent diameter and the yield strength of the
longitudinal rebar and stirrup considering the corrosion varying along
the column height. The performance deterioration of the reinforcement
in the splash and tidal zone is more obvious than the performance
deterioration of the reinforcement in the atmospheric zone. Compared
to the sound state, the diameter of the longitudinal bar and stirrup in
the splash and tidal zone has a decrease of 39.6% and 81.5%,
respectively, after providing service for 100 years, while the corre-
sponding decrease in yield strength is 31.7% and 48.3%, respectively.
5.3. Seismic failure mode and equivalent plastic hinge length
Based on the finite element model, the hysteresis analysis of the
bridge pier was conducted for the structure under the axial and cyclic
lateral loads. To investigate the effects of the time-dependent corrosion
damage, the numerical simulation is performed on the 11 finite
element models in different states of deterioration with the time
interval of 10 years. Due to the limited space, only the simulation
results for the bridge pier in service for 40 years are given in Fig. 13 as
an example.
From the moment-curvature curves shown in Fig. 13(a), the load-
resisting capacity of the BS and BST does not drop after the yield of the
structure. However, from the stress-strain curves of the confined
concrete, the concrete in the two cross sections has reached its ultimate
compression strain before the load is ended. According to Fig. 6(b), the
hoop fracture can be used to define the ultimate state, and the mean
values of the ultimate curvatures in the positive and negative directions

Fig. 9. Calibration results of the values of the fatigue parameters: (a) Cf and α ; and (b) Cd and β .

142
W. Yuan et al.
Fig. 11. Corrosion initiation time for the splash and tidal zone and the atmospheric zone: (a) longitudinal bars, (b) stirrups.
are 0.034 and 0.026 1/m, respectively, for the cross sections located at
the BS and the BST. The mean yield curvatures can be obtained by the
equivalent bilinear method described in Fig. 6(b), and the values for
the cross sections at the BS and the BST are 0.00275 and 0.00227 1/m,
respectively. The force-displacement curve of the structure is shown in
Fig. 13(b). In this figure, the circle stands for the displacement
corresponding to the hoop fracture moment, and the star is the
displacement when the base shear dropped to 85% of the peak force.
Compared to others, the confined concrete of the cross section at the BS
first reaches its ultimate compression strain. As the figure shows, the
mean value of the ultimate displacement is 495 mm.
From Fig. 12(a), it can be seen that the area reduction of the
stirrups is more severe than the longitudinal rebar due to the relative
small penetration distance of chloride ions measured from the surface
of the concrete cover to the stirrup surface. The corrosion of the
stirrups would consequently result in the reduction of the shear
capacity of the corroded structure. However, due to the large shear
span ratio and the small axial compression ratio of the selected bridge
pier, the shear capacity of the corroded structure is still strong enough
to resist the lateral shear force, and the shear failure mode is not found
in the corroded bridge pier during the analysis process. For the shear
failure of the coastal bridge piers with corrosion varying along the
column height in condition of small shear span ratio, it should be
further investigated in the future.
After analyzing the moment-curvature curves of all of the 11 finite
element models, the yield and ultimate curvatures of the two cross
sections versus the service time are shown in Fig. 14. As Fig. 14(a)
shows, the φ3u maintains a constant value of 0.034 1/m due to the
neglect of the corrosion, while the φ2u decreases from 0.0273 1/m to
0.0229 1/m in the whole service life. Fig. 14(b) indicates that φ3y
maintains the same value of 0.00275 1/m, while the φ2y approximately
linearly decreases from 0.00267 1/m to 0.00166 1/m with the increase
in the service time. Consequently, the difference in the ultimate
curvature and the yield curvature between the two sections becomes
Fig. 12. Deterioration of the reinforcement: (a) diameter; and (b) yield strength.
143
Soil Dynamics and Earthquake Engineering 93 (2017) 135–146
more and more significant.
Based on the numerical simulation results of the sound and
corroded structures, it is recognized that the φ2u and φ3u do not occur
simultaneously. The earlier parameter represents the full formation of
the corresponding plastic hinge and the attainment of the ultimate
state of the bridge pier. In the early 40 years, the performance
deterioration at the splash and tidal zone is not significant. Under
the large moment, the prior fully formed plastic hinge is still located at
the column end. Then, the fully formed plastic hinge is transferred
from the column end to the bottom of splash and tidal zone due to the
more serious corrosion damage in this region.
To find the final failure mode of the column at a certain corrosion
level, it is important to understand the occurrence sequence of φiy and
φju (i =2, j =3 or i =3, j =2) in the whole load process. Based on the above
analysis, j equals 3 in the early 40 years, and it equals 2 in the
remaining service time. To clearly illustrate this problem, the evolu-
tionary process of the curvature distributions from the initial state to
the failure state of the four types of failure modes are shown in Fig. 15
for the sound structure and the corroded structure after service for 30,
60 and 90 years, respectively. The four cases correspond to the failure
modes I, III, IV and V depicted in Fig. 5.
In those figures, the first three curvature distributions are redrawn
and depicted at the upper right of the figure to more clearly observe the
curves with small values. Fig. 15(a) shows that the curvature of the
bottom cross section of the submerged zone equals 0.072 1/m when φ2y
is reached. In this case, only one plastic hinge zone appears in the
submerged zone at the failure state. For Figs. 15(b) and (c), two plastic
hinges are observed for the two cases. For the structure in service for
30 years, the plastic hinge in the submerged zone is fully formed. At the
same time, the cross section in the splash and tidal zone is found to
slightly exceed the yield curvature to generate a plastic hinge that is not
fully formed. For case 3, this phenomenon is inverse. When the
structure has been in service for 90 years, the curvatures of the cross
sections in the submerged zone are all less than φ3y when φ2u is
W. Yuan et al. Soil Dynamics and Earthquake Engineering 93 (2017) 135–146

Fig. 13. Hysteretic curve of the 40-year-old bridge pier: (a) moment-curvature curve; and (b) force-displacement curve.

Fig. 14. Time-dependent curvature of the cross section at the BST and the BS: (a) ultimate curvature; and (b) yield curvature.

Fig. 15. Curvature distribution evolution along the column with the service time: (a) 0 years; (b) 30 years; (c) 60 years; and (d) 90 years.

reached, and the plastic hinge is concentrated only in the splash and
tidal zone, as shown in Fig. 15(d).
To calculate the EPHL, the yield displacement, Δy , and the ultimate
displacement, Δu , at the top piers are also analyzed. As shown in
Fig. 16, the yield displacement is slightly decreased over the whole
service period. However, the ultimate displacement is first increased in
the early 40 years and then significantly reduced. This phenomenon is
induced by the following reasons. In the early 40 years, the fully formed
plastic hinge is found to be located in the submerged zone. The failure
modes of the sound structure, the aging structure with 10 years and
[20–40] years of service periods correspond to the cases I, II and III in
Fig. 5, respectively. Until the failure of the structure at the column end,

144
W. Yuan et al.
Fig. 16. Yield and ultimate displacement of the aging column.
Fig. 17. Variation of EPHL.
the contribution to the lateral displacement of the splash and tidal zone
becomes larger and larger. Consequently, the ultimate displacement of
the structure is also increased. When the service period increases
further, the serious deterioration of the structure reduces the seismic
resistance capacity. Therefore, the ultimate displacement of the aging
structure shows the tendency to decrease.
Fig. 17 shows the analyzed results for the dimensionless EPHL. In
this figure, the unit “D” refers to the diameter of cross section. From the
figure, it is observed that the fully formed plastic hinge is located in the
submerged zone in the early 40 years. Because the corrosion in this
region is neglected, the curvature distribution in the ultimate state also
remains unchanged for the submerged zone. Therefore, the EPHL, lp3,
is also constant with the value of 0.63D. In the remaining 60 years, the
fully formed plastic hinge is transferred from the submerged zone to
the splash and tidal zone. The EPHL is within the scope of [0.30–0.37]
D.
From the analysis results, the co-existence of two plastic hinges,
with one plastic hinge not fully formed, appears at 20 years and ends at
80 years. To satisfy the computational formula described in Eqs. (4)
and (5), the lp3 increase sharply with time after 40 years of service,
especially obvious at 70 years and 80 years. However, the contribution
of this plastic hinge to the top displacement is quite small at these two
time points. The same phenomenon is also observed for the lp2 at 20
and 30 years. To solve this problem, the calculated nominal EPHL l′pi is
also depicted in Fig. 17 according to Eq. (6). The equivalent plastic
length obtained using this method is more reasonable for this
phenomenon.
6. Conclusions
This paper presents a study on the time-dependent seismic failure
mode and EPHL of coastal bridge piers with corrosion varying along
the column height in the marine environment. The five possible seismic
failure modes of such structures are discussed, considering the effects
of the specific corrosion phenomenon in the long-term service period.
145
Soil Dynamics and Earthquake Engineering 93 (2017) 135–146
Through the analysis of the constitutive models of the deteriorated
material and the finite element model of the corroded structure, the
flowchart for determining the seismic failure mode and calculating the
EPHL is given. A numerical simulation is conducted to demonstrate the
proposed method. The conclusions of this study are summarized as
follows:
(1) The seismic failure mode of the corroded bridge piers in the
marine environment is correlated with the location of the splash
and tidal zone and the degree of the performance deterioration of
the structure, which is dependent upon the corrosion level and the
time of service.
(2) Due to the different exposures, the performance deterioration of
the bridge pier is more serious in the splash and tidal zone. During
the whole service period, the coastal bridge piers have the
possibility of sustaining the seismic damage with one fully formed
plastic hinge at the column end in the early service stage, two
plastic hinges at the two different regions in the middle service
stage and one plastic hinge at the splash and tidal zone in the final
service stage.
(3) For the case with two plastic hinges, the nominal EPHL is more
reasonable for representing the plastic hinge that is not fully
formed. From the numerical simulation results, the nominal EPHL
at the column end remains constant in the early service stage,
while decreasing continuously to zero in the remaining time.
However, the nominal EPHL at the splash and tidal zone shows
a trend of increasing from zero in the early service stage while
decreasing slowly in the remaining time.
Acknowledgments
The financial support from the National Natural Science
Foundation of China with grant No. 51678197, the Major State Basic
Research Development Program of China (973 Program) with Grant
No. 2011CB013604, and the Fundamental Research Funds for the
Central Universities of China with grant No. HIT.BRETIV.201320 is
greatly appreciated by the authors.
References
[1] Andrade C, Alonso C, Molina F. Cover cracking as a function of bar corrosion: part
I-Experimental test. Mater Struct 1993;26(8):453–64.
[2] Fu X, Chung DDL. Effect of corrosion on the bond between concrete and steel rebar.
Cem Concr Res 1997;27(12):1811–5.
[3] Coronelli D, Gambarova P. Structural assessment of corroded reinforced concrete
beams: modeling guidelines. J Struct Eng -ASCE 2004;130(8):1214–24.
[4] Ou YC, Fan HD, Nguyen ND. Long-term seismic performance of reinforced
concrete bridges under steel reinforcement corrosion due to chloride attack. Earthq
Eng Struct D 2013;42(14):2113–27.
[5] Xi YP, Bazant ZP. Modeling chloride penetration in saturated concrete. J Mater Civ
Eng 1999;11(1):58–65.
[6] Xi YP, William K, Frangopol DM. Multiscale modeling of interactive diffusion
processes in concrete. J Eng Mech-ASCE 2000;126(3):258–65.
[7] Ababneh A, Benboudjema F, Xi YP. Chloride penetration in nonsaturated concrete.
J Mater Civ Eng 2003;15(2):183–91.
[8] Saetta AV, Scotta RV, Vitaliani RV. Analysis of chloride diffusion into partially
saturated concrete. ACI Mater J 1993;90(5):441–51.
[9] Du YG, Clark LA, Chan AHC. Residual capacity of corroded reinforcing bars. Mag
Concr Res 2005;57(3):135–47.
[10] Zanini MA, Pellegrino C, Morbin R, Modena C. Seismic vulnerability of bridges in
transport networks subjected to environmental deterioration. Bull Earthq Eng
2013;11(2):561–79.
[11] Caprili S, Salvatore W. Cyclic behaviour of uncorroded and corroded steel
reinforcing bars. Constr Build Mater 2015;76:168–86.
[12] Palsson R, Mirza MS. Mechanical response of corroded steel reinforcement of
abandoned concrete bridge. ACI Struct J 2002;99(2):157–62.
[13] Maslehuddin M, Allam IM, Alsulaimani GJ, Almana AI, Abduljauwad SN. Effect of
rusting of reinforcing steel on its mechanical-properties and bond with concrete.
ACI Mater J 1990;87(5):496–502.
[14] Qiao D, Nakamura H, Yamamoto Y, Miura T. Evaluation method of tensile behavior
of corroded reinforcing bars considering radius loss. J Adv Concr Technol
2015;13(3):135–46.
[15] Molina FJ, Alonso C, Andrade C. Cover cracking as a function of rebar corrosion. 2.
W. Yuan et al.
Numerical-model. Mater Struct 1993;26(163):532–48.
[16] Zhong J, Gardoni P, Rosowsky D. Stiffness degradation and time to cracking of
cover concrete in reinforced concrete structures subject to corrosion. J Eng Mech-
ASCE 2010;136(2):209–19.
[17] Li CQ, Zheng JJ, Lawanwisut W, Melchers RE. Concrete delamination caused by
steel reinforcement corrosion. J Mater Civ Eng 2007;19(7):591–600.
[18] Gardoni P, Trejo D. Handbook of seismic risk analysis and management of civil
infrastructure systems. Cambridge, UK: Woodhead Publishing Ltd; 2013.
[19] Choe DE, Gardoni P, Rosowsky D, Haukaas T. Probabilistic capacity models and
seismic fragility estimates for RC columns subject to corrosion. Reliab Eng Syst
Safe 2008;93(3):383–93.
[20] Choe DE, Gardoni P, Rosowsky D, Haukaas T. Seismic fragility estimates for
reinforced concrete bridges subject to corrosion. Struct Saf 2009;31(4):275–83.
[21] Choe DE, Gardoni P, Rosowsky D. Fragility increment functions for deteriorating
reinforced concrete bridge columns. J Eng Mech-ASCE 2010;136(8):969–78.
[22] Gardoni P, Rosowsky D. Seismic fragility increment functions for deteriorating
reinforced concrete bridges. Struct Infrastruct E 2011;7(11):869–79.
[23] Simon J, Bracci JM, Gardoni P. Seismic response and fragility of deteriorated
reinforced concrete bridges. J Struct Eng-ASCE 2010;136(10):1273–81.
[24] Zhong JQ, Gardoni P, Rosowsky D. Seismic fragility estimates for corroding
reinforced concrete bridges. Struct Infrastruct E 2012;8(1):55–69.
[25] Akiyama M, Frangopol DM, Matsuzaki H. Life-cycle reliability of RC bridge piers
under seismic and airborne chloride hazards. Earthq Eng Struct D
2011;40(15):1671–87.
[26] Kumar R, Gardoni P, Sanchez-Silva M. Effect of cumulative seismic damage and
corrosion on the life-cycle cost of reinforced concrete bridges. Earthq Eng Struct D
2009;38(7):887–905.
[27] Ghosh J, Padgett JE. Probabilistic seismic loss assessment of aging bridges using a
component-level cost estimation approach. Earthq Eng Struct D
2011;40(15):1743–61.
[28] Zanini MA, Faleschini F, Pellegrino C. Cost analysis for maintenance and seismic
retrofit of existing bridges. Struct Infrastruct E 2016;12(11):1411–27.
[29] Alipour A, Shafei B, Shinozuka M. Performance evaluation of deteriorating highway
bridges located in high seismic areas. J Bridge Eng 2011;16(5):597–611.
[30] Bastidas-Arteaga E, Sanchez-Silva M, Chateauneuf A, Silva MR. Coupled reliability
model of biodeterioration, chloride ingress and cracking for reinforced concrete
structures. Struct Saf 2008;30(2):110–29.
[31] Lu CH, Jin WL, Liu RG. Probabilistic lifetime assessment of marine reinforced
concrete with steel corrosion and cover cracking. China Ocean Eng
2011;25(2):305–18.
[32] Qi YL, Han XL, Ji J. Failure mode classification of reinforced concrete column
using Fisher method. J Cent South Univ 2013;20(10):2863–9.
146
Soil Dynamics and Earthquake Engineering 93 (2017) 135–146
[33] Ma Y, Che Y, Gong J. Behavior of corrosion damaged circular reinforced concrete
columns under cyclic loading. Constr Build Mater 2012;29:548–56.
[34] Bae S, Bayrak O. Plastic hinge length of reinforced concrete columns. ACI Struct J
2008;105(3):290–300.
[35] Park R, Priestley MJN. Ductility of square-confined concrete columns. J Struct Div
1982;108(4):929–50.
[36] Paulay T, Priestley MJN. Seismic design of reinforced concrete and mansory
buildings. New York: John Wiley and Sons; 1992.
[37] Priestley MJN, Park R. Strength and ductility of concrete bridge columns under
seismic loading. ACI Struct J 1987;84(1):61–76.
[38] Winter G, Nilson AH. Design of concrete structures. United States of America:
McGraw Hill Higher Education; 1979.
[39] Priestley MJN, Calvi GM, Kowalsky MJ. Displacement-based seismic design of
structures. Pavia: IUSS press; 2007.
[40] Mazzoni S, McKenna F. OpenSees command language manual. Berkeley, USA:
Pacific Earthquake Engineering Research Center, University of California; 2006.
[41] Martin-Perez B, Zibara H, Hooton RD, Thomas MDA. A study of the effect of
chloride binding on service life predictions. Cem Concr Res 2000;30(8):1215–23.
[42] Rodriguez J, Andrade C. A validated users manual for assessing the residual service
life of concrete structures. Madrid: DG Enterprise; 2001.
[43] Coffin LF. A study of the effect of cyclic thermal stresses on a ductile metal. Trans
Am Soc Mech Eng 1954;76:931–50.
[44] Kumar R, Gardoni P, Sanchez-Silva M. Effect of cumulative seismic damage and
corrosion on the life-cycle cost of reinforced concrete bridges. Earthq Eng Struct D
2009;38(7):887–905.
[45] Manson SS. Fatigue: a complex subject-some simple approximations. ExM
1965;5(7):193–226.
[46] Brown J, Kunnath SK. Low-cycle fatigue behavior of reinforcing steel bars. ACI
Mater J 2004;101(6):457–66.
[47] Lv P. Reasearch on low cycle fatigue behavior of high strength steel HRB500, in:
structural engineering, Dalian University of Technology, China National Knowledge
Infrastructure; 2011.
[48] Mander JB, Panthaki FD, Kasalanati A. Low-cycle fatigue behavior of reinforcing
steel. J Mater Civ Eng 1994;6(4):453–68.
[49] Kunnath SK, Heo Y, Mohle JF. Nonlinear uniaxial material model for reinforcing
steel bars. J Struct Eng-ASCE 2009;135(4):335–43.
[50] Bastidas-Arteaga E, Sánchez-Silva M, Chateauneuf A, Silva MR. Coupled reliability
model of biodeterioration, chloride ingress and cracking for reinforced concrete
structures. Struct Saf 2008;30(2):110–29.
[51] Bentz EC, Thomas MDA. Life-365 service life prediction model and computer
program for predicting the service life and life-cycle cost of reinforced concrete
exposed to chlorides, in, Life-365 Consortium II; 2012.