Performance-Based Seismic Design of Shape Memory

Alloy–Reinforced Concrete Bridge Piers. I: Development

of Performance-Based Damage States
A. H. M. Muntasir Billah, A.M.ASCE 1; and M. Shahria Alam, M.ASCE 2
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Performance-based seismic design aims to dictate the structural performance in a predetermined fashion given the possible
seismic hazard scenarios the structure is likely to experience. Identifying and assessing the probable performance is an integral part of
performance-based design. Before implementation, accurate and practical definitions of different performance levels and corresponding limit
states must be determined. This study aims to develop performance-based damage states for shape memory alloy (SMA)–reinforced concrete
(RC) bridge piers considering different types of SMAs and seismic hazard scenarios. Using incremental dynamic analysis (IDA), this study
develops quantitative damage states corresponding to different performance levels (cracking, yielding, and strength degradation) and specific
probabilistic distributions for RC bridge piers reinforced with different types of SMAs. Based on an extensive numerical study, this study
also proposes residual drift–based damage states for SMA-RC piers. Finally, an analytical expression is proposed to estimate the residual
drift of SMA-reinforced concrete elements as a function of the expected maximum drift and superelastic strain of SMA. Comparison with
experimental results revealed that the proposed equation could very well predict the residual drift obtained from the experimental results.
DOI: 10.1061/(ASCE)ST.1943-541X.0001458. © 2016 American Society of Civil Engineers.
Author keywords: Performance-based design; Shape memory alloy; Damage states; Incremental dynamic analysis; Residual drift; Special
design issues.

Introduction region), which are prone to experience more damage during an

In recent years, seismic design guidelines have focused on
performance-based design in order to predict and better manage
the postearthquake functionality and condition of structures.
Recent developments in performance-based seismic design and as-
sessment approaches have emphasized the importance of properly
assessing and limiting the residual (permanent) deformations that
are typically sustained by a structure after a seismic event (Pettinga
et al. 2006). If reinforced concrete (RC) structures can be designed
in such a way that they are not only capable of undergoing large
displacements with adequate energy dissipation capacity during a
seismic event, but can also eliminate the problem of permanent de-
formation, this will substantially scale down the repair and main-
tenance cost of structures and make them safer against earthquakes.
Superelastic shape memory alloy (SMA) possesses the distinct abil-
ity to experience large deformations and retrieve its original shape
upon load removal along with possessing a high resistance to cor-
rosion (Alam et al. 2009). This is a distinct property that makes
SMA a smart material and a strong contender for reinforcement
in RC structures particularly at critical locations (e.g., plastic hinge
1
Former Graduate Student, School of Engineering, Univ. of British
Columbia, Kelowna, BC, Canada V1V1V7; presently, Bridge Engineer,
Parsons, Burnaby, BC, Canada V5H4M2. E-mail: muntasir.billah@
alumni.ubc.ca
2
Associate Professor, School of Engineering, Univ. of British
Columbia, Kelowna, BC, Canada V1V1V7 (corresponding author).
E-mail: shahria.alam@ubc.ca
Note. This manuscript was submitted on December 16, 2014; approved
on October 20, 2015; published online on July 22, 2016. Discussion period
open until December 22, 2016; separate discussions must be submitted for
individual papers. This paper is part of the Journal of Structural Engineer-
ing, © ASCE, ISSN 0733-9445.
earthquake.
Numerous experimental and numerical studies proved the effi-
ciency of SMA reinforced structures in seismic regions. However,
there exists no proper design guideline for utilizing SMA in high-
way bridges. Moreover, most of the design guidelines are moving
towards performance-based design. AASHTO has already devel-
oped performance-based design guidelines for bridges referred
as AASHTO SGS (AASHTO 2011). Moreover, the recent edition
of the Canadian highway bridge design code (National Research
Council of Canada 2014) has also adapted performance-based
design and defined some performance levels and performance
criteria for different types of bridges. To successfully apply the
performance-based design concept to SMA-reinforced concrete
(SMA-RC) bridge piers, the performance objectives and their as-
sociated limit state criteria must be clearly defined first. Most of the
current researches on SMA-RC bridge piers are focused on seismic
performance assessment and comparison with regular RC bridge
piers (Billah and Alam 2014; Cruz and Saiidi 2012; Saiidi et al.
2009). Although a good number of studies exist on performance-
based damage states for steel RC bridge piers (Lehman et al. 2004;
Hose et al. 2000), no study so far has focused on the performance-
based damage states for SMA-RC piers. This is mainly due to lim-
ited number of experimental studies performed on SMA-RC piers
where the high cost of SMA was the main restraining factor. Since
the behavior of SMA-RC piers are significantly different from their
steel counterpart, using those damage states for SMA-RC piers
might lead to faulty design. Moreover, the mechanical properties
of SMAs (superelastic strain, elastic modulus, and yield strength)
vary widely where several compositions of SMAs have been
developed and used by different researchers in civil engineering
applications (Alam et al. 2007). In a recent study, Billah and
Alam (2014) assessed the comparative seismic vulnerability of
SMA-RC and steel-RC bridge piers in terms of drift limits based

© ASCE 04016140-1 J. Struct. Eng.

J. Struct. Eng., 2016, 142(12): 04016140

 Table 1. Properties of Different Types of SMA
εs εr E fy
SMA type Alloy (%) (%) (GPa)
SMA-1 NiTi45 6 0.5 62.5
SMA-2 NiTi45 8 0.5 68
SMA-3 FeNCATB 13.5 1.5 46.9
SMA-4 CuAlMn 9 0.4 28
SMA-5 FeMnAlNi 6.13 0.7 98.4
Note: f y = austenite to martensite starting stress; f P1 = austenite to martensite finishing stress; fT1 = martensite to austenite starting stress; f T2 = martensite to
austenite finishing stress, εs = superelastic plateau strain length; εr = recovery strain; and E = modulus of elasticity.
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
on performance criteria. The results showed that the performance
limits of SMA-reinforced piers vary considerably from their steel
counterparts. Hence, the current study aims at developing perfor-
mance-based damage states for SMA-RC bridge piers considering
five different SMAs with three different earthquake hazard levels.
The ultimate goal of this study is to provide a technical basis for the
development of performance-based seismic design, and other
evaluation methodologies, for SMA-RC bridge piers.
Using an incremental dynamic analysis (IDA)-based analytical
approach (Vamvatsikos and Cornell 2002), here performance-based
damage states (based on drift limits) have been developed for five
different SMA-RC bridge piers and validated against experimental
data. Application of such technique may palliate the burden of
gathering a large amount of test data and cost of experiments,
which were used in the past to develop different damage sates
for RC bridge piers. Past studies have demonstrated that it is nec-
essary to consider residual drifts to fully characterize the perfor-
mance of a structural system after a seismic excitation and the
potential damage that the system can experience (Christopoulos
et al. 2003; Erochko et al. 2011). Since SMA has the ability to re-
duce the residual drift significantly after unloading, the residual
drift of different SMA-RC bridge piers under varying intensity
earthquakes needs to be investigated. This study also developed
residual drift–based damage states for SMA-RC bridge piers and
proposed an analytical expression that can be used for predicting
the residual drift in SMA-reinforced concrete elements.
Shape Memory Alloys
Smart materials like SMAs have demonstrated a wide range of en-
gineering applications namely, biomedical, robotics, aerospace,
civil, and mechanical engineering. Two distinct properties such
as the shape memory effect (SME, ability to recover plastic strain
upon heating) and superelasticity (SE, ability to recover plastic
strain upon load removal) make SMA a strong contender against
conventional metals and alloys for the application in various sec-
tors. Several compositions of SMAs have been developed to date
such as Ni-Ti, Cu-Zn, Cu-Zn-Al, Cu-Al-Ni, Fe-Mn, Mn-Cu, Fe-Pd,
and Ti-Ni-Cu, etc. Numerous applications of SMAs in the civil en-
gineering field have been documented (Ocel et al. 2004; Saiidi and
Wang 2006; Lindt and Potts 2008; Alam et al. 2009; Araki et al.
2010; Billah and Alam 2012; Dezfuli and Alam 2013). Most ap-
plications have focused on the use of Ni-Ti alloy while very few
focused on the application of Cu-based SMAs (Araki et al. 2010;
Shrestha et al. 2013), or Fe-based SMAs (Dezfuli and Alam 2013;
Czaderski et al. 2014). Although Ni-Ti SMAs show large recover-
able strains, good superelasticity, and exceptional resistance to cor-
rosion, the high cost of Ni-Ti SMAs and their machinability restrict
their large-scale applications.
In an attempt to reduce the cost of SMAs, researchers have come
up with various Fe-based and Cu-based low cost SMAs such as
© ASCE
J. Struct. Eng., 2016, 142(12): 04016140
f p1 f T1 fT2
(MPa) (MPa) (MPa) (MPa) fy=E Reference
401.0 510 370 130 0.0065 Alam et al. (2008)
435.0 535.0 335 170 0.0063 Ghassemieh et al. (2012)
750 1,200 300 200 0.0159 Tanaka et al. (2010)
210.0 275.0 200 150 0.0075 Shrestha et al. (2013)
320.00 442.5 210.8 122 0.0033 Omori et al. (2011)
Fe-Mn-Si, Fe-Ni-Co-Ti, Fe-Ni-Nb, Cu-AL-Mn, etc. Fe-based
SMAs show good workability, machinability, weldability, and wide
transformation hysteresis as compared to Ni-Ti SMAs. Although
several compositions of Fe-based SMAs have been developed,
large-scale applications are still limited due to the poor shape
recovery limit and associated costly training treatment. Recently
Tanaka et al. (2010) developed a ferrous polycrystalline SMA
(Fe-Ni-Co-Al-Ta-B) which has a very high superelastic strain range
of over 13% at room temperature. This SMA has approximately 20
times higher SE than Fe-Ni-Co-Ti alloy and almost double that of
conventional Ni-Ti alloy. This Fe-based SMA has extremely high
ductility, greater strength, and also energy dissipation capacity sev-
eral times higher than that of commercially available Ni-Ti SMA.
More recently, Omori et al. (2011) developed another Fe-based
SMA (Fe-Mn-Al-Ni) which has superelasticity similar to the con-
ventional Ni-Ti SMA but much lower austenite finish temperature,
which allows this SMA to operate in superelastic range even at a
very low temperature. In order to improve the machinability and
reduce the cost, a Cu-based SMA (Cu-Mn-Al) has been developed
(Araki et al. 2010) that has comparable superelasticity to that of
NiTi SMAs. Moreover, these Cu-Al-Mn SMAs have comparatively
lower strain rate effects than Ni-Ti SMAs (Araki et al. 2012) and
also been reported to provide recentering and crack recovery capa-
bilities (Shrestha et al. 2013).
Many different types of SMAs have been developed that have
huge potential for smart structural applications, such as in bridge
piers. In this study, two Ni-Ti, one Cu-based, and two Fe-based
shape memory alloys have been selected for use in bridge piers.
The selected SMAs along with their mechanical properties such
as the elastic modulus (E), austenite to martensite starting stress
(fy ); austenite to martensite finishing stress (fP1 ); martensite to
austenite starting stress (fT1 ); martensite to austenite finishing
stress (fT2 ); superelastic strain (εs ); residual strain (εr ) are listed
in Table 1.
Design and Geometry of Bridge Piers
This section briefly describes the design and configurations of dif-
ferent SMA-RC bridge piers used in this study. Since SMA is a
costly material, it is only used in the bottom plastic hinge region
of the bridge piers. Five different SMAs are used in this study to
develop the performance-based damage states for SMA-RC bridge
piers. The bridge pier is assumed to be located in Vancouver,
British Columbia, Canada, and was seismically designed following
the Canadian highway bridge design code (CSA 2010). The con-
sidered bridge is a lifeline bridge as per CHBDC (CSA 2010). In an
event of the design earthquake (return period of 475 years), a life-
line bridge needs to remain open for immediate use to all traffic.
Therefore, the piers need to be designed to achieve this targeted
performance. According to CHBDC (CSA 2010), the importance
factor of I ¼ 3 and the response modification factor of R ¼ 3 were
04016140-2 J. Struct. Eng.
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
Fig. 1. Cross section and elevation of SMA reinforced concrete bridge pier
considered for this lifeline bridge. Fig. 1 shows the cross section
and elevation of the bridge pier. The diameter of all the columns
was 1.83 m; the piers were reinforced with 48 longitudinal reinforc-
ing bars of varying diameter (reinforcement ratio of 1.2–1.75%)
and 16-mm-diameter steel spirals at a 76 mm pitch (transverse
reinforcement ratio of 0.7%). The height of the pier is 9.14 m with
an aspect ratio of 5, which ensured the flexure dominated behavior.
A constant mass of 85t was applied at the top which represents the
weight of the superstructure. The design moment and axial force
were 11,265 kN · m and 9,500 kN, respectively. The shear capacity
of the pier was checked using modified compression field theory
(Vecchio and Collins 1986), which predicts the experimentally de-
termined shear failure within 1% error (Bentz et al. 2006). The
shear resistance of the pier was found to be 3,912 kN, which
was much higher than the shear force demand of 1,230 kN. The
moment-shear interaction and the axial load-moment interaction
of the bridge piers were investigated. However, for brevity, the re-
sults are not presented here. It was found that the maximum mo-
ment and shear force as well as the applied maximum axial load and
moment are within the safe boundary. Wang et al. (2008) recom-
mended that the shear capacity of the pier should be greater than 1.6
times the base shear corresponding to the design moment which
was also satisfied.
Different diameter bars were used for different SMAs since dif-
ferent SMAs have different elastic modulus and yield strength.
Although SMA does not have a yielding process, the term yield
18000
16000
14000
Moment (kN-m)
12000
10000
SMA-1
8000
SMA-2
6000 SMA-3
4000 SMA-4
SMA-5
2000
0 0
0 0.02 0.04
(a) Curvature (1/m)
Fig. 2. (a) Moment-curvature relationship of RC sections with different types of SMA; (b) static pushover curves for bridge piers reinforced with
different types of SMA
© ASCE
J. Struct. Eng., 2016, 142(12): 04016140
is being used to refer to the initiation of phase transformation of
SMA and the yield strain was calculated by defining the austenite
to martensite starting stress (fy ) by the elastic modulus (E).
Five different SMA rebars as presented in Table 1 were used to
design the different bridge piers. The bridge piers are designated
as SMA-RC-1 (reinforced with SMA-1), SMA-RC-2 (reinforced
with SMA-2), and so on. SMA-RC-1 and SMA-RC-2 are
reinforced with 48-28M SMA-1 and SMA-2 bars, SMA-RC-3 is
reinforced with 48-20M SMA-3 bars, SMA-RC-4 is reinforced
with 48-35M SMA-4 bars, and SMA-RC-5 is reinforced with
48-32M SMA-5 bars. Rebar sizes were selected in such a way
that the axial forces developed in the rebar are almost similar.
The bridge piers were designed in such a way that they have com-
parable moment capacities. Fig. 2(a) shows the moment-curvature
response of different SMA-RC sections. From this figure, it is evi-
dent that all the sections have similar initial stiffness and compa-
rable moment capacities. Since SMA-5 has higher elastic modulus,
SMA-RC-5 showed higher initial stiffness which is 1.78, 1.72,
2.21, and 3.87 times higher than that of SMA-RC-1, SMA-RC-2,
SMA-RC-3, and SMA-RC-4, respectively. Moment-curvature re-
sponse of all the sections revealed that this design process led
to comparable moment capacities for the five different SMA-
reinforced bridge piers. The elastic periods of the SMA-RC-1,
SMA-RC-2, SMA-RC-3, SMA-RC-4, and SMA-RC-5 were
calculated as 0.513 s, 0.513 s, 0.514 s, 0.515 s, and 0.511 s, respec-
tively which were close and expected to attract similar earthquake
2000
1600
Base Shear (kN)
1200
800
SMA-1 SMA-2
400 SMA-3 SMA-4
SMA-5
0.06 0 0.2 0.4 0.6 0.8 1 1.2
(b) Displacement (m)
04016140-3 J. Struct. Eng.
 Table 2. Material Properties for SMA-RC Bridge Pier the Paulay and Priestley (1992) equation provided different plastic
Material Property Value hinge lengths for different SMA-reinforced bridge piers. Therefore,
the Paulay and Priestley (1992) expression for plastic hinge length
Concrete Compressive strength (MPa) 42.4
calculation in SMA/RC elements can be considered with reasonable
Corresponding strain 0.0029
Tensile strength (MPa) 3.5
accuracy.
Elastic modulus (GPa) 23.1
Steel Elastic modulus (GPa) 200
Yield stress (MPa) 475 Analytical Modeling of Bridge Piers
Ultimate stress (MPa) 692
Ultimate strain 0.14 In this study, a fiber element–based nonlinear analysis pro-
Plateau strain 0.016 gram SeismoStruct version 6 has been employed to develop
Strain hardening parameter 0.0125 performance-based damage states for SMA-RC bridge piers. Incre-
Transition curve initial shape parameter 19.5 mental dynamic analyses (IDA) have been performed to determine
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.

Transition curve shape calibrating coefficient (A1) 18.5
Transition curve shape calibrating coefficient (A2) 0.15
forces. Fig. 2(b) shows the pushover response curves for these five
different SMA-RC bridge piers. From this figure, it can be observed
that all the bridge piers had similar stiffness and load-carrying
capacities.
The material properties of the concrete and steel rebar used
in the bridge piers are summarized in Table 2. In the SMA-RC
bridge piers, SMA was used as longitudinal reinforcement only
at the plastic hinge region. In the remaining part, steel rebars were
used as reinforcement. The plastic hinge length, Lp was calculated
according to the Paulay and Priestley (1992) equation
Lp ¼ 0.08L þ 0.022db f y ð1Þ
where L = length of the member in mm; db = bar diameter in mm; and
f y = yield strength of the rebar in MPa. Previously, Alam et al. (2008)
and O’Brien et al. (2007) showed that the Paulay and Priestley
(1992) equation can reasonably estimate the plastic hinge length
of SMA-reinforced concrete element. Moreover, Saiidi and Wang
(2006), Saiidi et al. (2009), and Cruz and Saiidi (2012) also used
the Paulay and Priestley (1992) equation to calculate the plastic
hinge length for their experimental studies where SMA rebars were
placed in the bottom plastic hinge region of bridge piers. AASHTO
SGS (AASHTO 2011) also suggests using the same equation for
calculating the plastic hinge length of concrete bridge piers.
However, CHBDC (2010) recommends using the greatest of the
maximum cross sectional dimension of column, one-sixth of clear
height of the column or 450 mm. However, according to CHBDC
(2010), the plastic hinge length does not depend on strength and
diameter of longitudinal rebar. Since this study utilized different
types of SMA rebars of varying diameter and yield strength, using
the various damage states of the bridge piers. The program has
the ability to determine the large displacement behavior and the
collapse load of framed structures accurately under either static
or dynamic loading, while taking into account both geometric non-
linearities and material inelasticity (Pinho et al. 2007). The bridge
piers were modeled with three-dimensional (3D) inelastic beam–
column elements (force-based elements), with circular sections
for the piers; the constitutive laws of the reinforcing steel and con-
crete were as per the Menegotto and Pinto (1973) and Mander
et al. (1988) models, respectively. This steel model takes into ac-
count the Bauschinger effect, which is relevant for the representa-
tion of the columns’ stiffness degradation under cyclic loading. The
superelastic SMA model developed by Auricchio and Sacco (1997)
has been employed for modeling SMA using the parameters
provided in Table 1. The material models were calibrated (values
provided in Table 1) to capture the strength and stiffness degrada-
tion under cyclic loading.
The accuracy of the program in predicting the seismic response of
bridge structures has been demonstrated by several researchers
through comparisons with experimental results (Casarotti and Pinho
2006; Alam et al. 2008, 2009; Billah and Alam 2013). Fig. 3 shows
the comparison of experimental and analytical results from two
different studies using two different SMAs. Fig. 3(a) shows the
comparison of shake table test results and analytical results of a
SMA-steel RC bridge pier where SMA was particularly used in
the plastic hinge region. The numerical results obtained from
SeismoStruct could predict the experimental result of Saiidi and
Wang (2006) accurately where the variations were only 5.6%,
6.1%, and 9.4% for base shear, tip displacement, and amount of en-
ergy dissipation, respectively. Fig. 3(b) shows the load-rotation re-
sponse of concrete beams reinforced with Cu-Al-Mn SMA (SMA-4)
in the midspan under four-point reverse cyclic loading (Shrestha
et al. 2013). From this figure, it is evident that the adopted analytical
model was capable of predicting the experimental response very well

Fig. 3. Comparison of experimental and numerical results: (a) SMA-RC (SMA-1) bridge pier; (b) SMA-RC (SMA-4) beam

© ASCE 04016140-4 J. Struct. Eng.

J. Struct. Eng., 2016, 142(12): 04016140


Selection of Shape Memory Alloy
Design of bridge piers with different SMA
Moment-curvature analysis of different
SMA-RC section
Selection of suitable ground motions and
hazard spectra
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
Extensive experimental investigations on bridge piers performed in
the past were utilized to develop the damage states for reinforced
concrete bridge piers (Berry and Eberhard 2003; Hose et al. 2000).
Due to the fact that very limited experimental results are available
for SMA-reinforced bridge piers, an IDA-based approach, as illus-
trated in Fig. 4, was developed in this study to generate the neces-
sary data for developing performance-based damage states for
bridge piers reinforced with different types of SMAs. First, five
different types of SMAs were selected and five different SMA-RC
bridge piers were designed. In the next step, moment-curvature
analyses of the designed piers were carried out to compare their
moment capacities. After that, a suite of suitable ground motions

IDA of different SMA-RC bridge pier
Capture the different performance limits
for each EQ at different intensity
Develop the dynamic pushover curves for
each EQ and compute the median,
5 percentile and 95 percentile curves
Obtain the drift limit at different
performance levels and determine the
suitable distribution
Define drift limits at different performance
levels for different SMAs
was selected and IDA was carried out. Based on the strain limits,
different performance criteria were captured from IDA. Using the
IDA results, dynamic pushover curves were generated and the per-
formance limits were identified. In the final step, suitable probabi-
listic distributions corresponding to different limit states were
identified and drift limits corresponding to different limit states
were defined.
IDA (Vamvatsikos and Cornell 2002) was employed to deter-
mine the performance limit states of different bridge piers using
an ensemble of 10 selected ground motions. IDA is a useful method
for more-detailed seismic performance predictions of structures
subjected to different seismic excitation levels. In IDA, the finite-
element model is subjected to numerous inelastic time history
analyses using one or a set of ground motion record(s), each scaled
(up and/or down) to different seismic intensity levels while tracking
the response of the structure (e.g., displacements, accelerations,
etc.). This procedure of scaling and time history analysis is repeated
until dynamic instability in the form of large drifts occurs, indicat-
ing structural collapse.

Fig. 4. Steps for the development of performance-based damage states
for SMA-RC bridge pier
along with strength and stiffness degradation where the variations
were only 3.4% and 5.9% for maximum force and beam rotation,
respectively.
IDA-Based Approach for Developing
Performance-Based Damage States
For successful implementation of the performance-based design
concept in SMA-RC bridge pier, the performance objectives and
their corresponding damage state criteria need to be clearly defined.
Selection of Ground Motions
The incremental dynamic analyses were carried out using the 10
selected ground motions as shown in Table 3. These ground motion
records were obtained from the Pacific Earthquake Engineering
Research (PEER) Center ground motion database (PEER 2011).
These accelerograms were chosen such that they represent the seis-
mic characteristics of the site of the structure. The ratio between the
peak ground acceleration (PGA) and peak ground velocity (PGV)
is an indicator of the frequency content of seismic motion. The
characteristic seismic motions for the western part of Canada have
a PGA=PGV ratio of around 1.0 (Naumoski et al. 1988). The se-
lected ensemble of earthquake records is presented in Table 3 where
the PGA=PGV ratio varies between 0.8 and 1.3. These 10 ground
motion records were obtained from the PEER strong motion data-
base. The recent edition of Canadian highway bridge design code

Table 3. Selected Earthquake Ground Motion Records (Data from PEER 2011)
Number Event Year Record station M a R b (km) PGA (g) PGA/PGV
1 Imperial Valley 1979 El Centro array#11 6.5 21.9 0.36 0.8
2 Imperial Valley 1979 Chihuahua 6.5 28.7 0.254 0.84
3 Kobe 1995 Takatori 6.9 4.3 0.56 0.9
4 Kobe 1995 JMA 6.9 3.4 0.77 1.02
5 Loma Prieta 1989 Holister South and Pine 6.9 28.8 0.371 0.97
6 Loma Prieta 1989 16 LGPC 6.9 16.9 0.605 1.19
7 Nothridge 1994 Rinaldi 6.7 7.5 0.87 0.93
8 Nothridge 1979 Olive View 6.7 6.4 0.721 0.95
9 Superstition Hill 1987 Wildlife liquefaction array 6.7 24.4 0.134 1.0
10 Superstition Hill 1987 Wildlife liquefaction array 6.7 24.4 0.132 1.03
a
Moment magnitudes.
b
Closest distances to fault rupture.

© ASCE 04016140-5 J. Struct. Eng.

J. Struct. Eng., 2016, 142(12): 04016140

 1
0.9 2%/50 Year (Target)
Spectral Acceleration (g)
0.8 5%/50 Year (Target)
10%/50Year (Target)
0.7
2%/50 Year (Mean)
0.6
5%/50 Year (Mean)
0.5
10%/50 Year (Mean)
0.4
0.3
0.2
0.1
0
0 1 2 3
Time (sec)
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
Fig. 5. Design and mean response spectrum of 10 records used for IDA
analysis matching the three different CHBDC spectrum (2, 5, and 10 in
50 years)
(CHBDC) (National Research Council of Canada 2014) requires
that highway bridges should meet target performance levels
under seismic ground motions with different return periods. In this
study, three different levels of seismic ground motions were
considered according to CHBDC 2014 [CSA S6-14 (National
Research Council of Canada 2014)]. These records correspond
to three different hazard levels with a 2, 5, and 10% probability
of exceedance in 50 years. The respective return periods are
2,475, 975, and 475 years. For each hazard level, 10 ground mo-
tions given in Table 3 were used. The selected ground motions
were scaled to specific hazard levels using SeismoMatch. This soft-
ware is able to adjust any ground motion accelerogram to match a
specific design response spectrum using wavelet algorithm pro-
posed by Abrahamson (1992) and Hancock et al. (2006). Matching
was done within the period range of interest, which was 0.05–4 s
as suggested by Baker et al. (2011). The mean spectra and the
target spectra corresponding to different hazard levels are shown
in Fig. 5.
Performance-Based Damage States Criterion
Performance-based seismic design largely relies on the identifica-
tion and selection of proper limit/damage states. Often damage
states are defined in terms of drift or displacement. Damages are
usually defined as discrete observable damage states (e.g., rebar
yielding, concrete spalling, longitudinal bar buckling, and bar frac-
ture) (Marsh and Stringer 2013). For example, the recent edition of
the Canadian highway bridge design code (National Research
Council of Canada 2014) defined four different performance levels,
such as, immediate service, limited service, service disruption, and
life safety. Based on the material strain limit (concrete and steel),
four different damage states corresponding to the four performance
levels are defined. For example, limited service indicates a repair-
able damage, which corresponds to a limiting reinforcing steel
strain of 0.015 for concrete structures, and no buckling of primary
members are allowed in steel structures.
In this study, four quantitative performance limit states were de-
fined for the SMA-RC bridge piers based on the performance levels
and damage states proposed by Hose et al. (2000). Table 4 shows
the four performance limit states and their associated functional
level definition adopted in this study. The performance limit states
considered here are, the drift (%) at the onset of hairline cracks,
longitudinal rebar yielding, cover concrete spalling, and crushing
of core concrete. In this study, a strain-based damage detection ap-
proach was used for defining the drift levels at different damage
© ASCE
J. Struct. Eng., 2016, 142(12): 04016140
Table 4. Proposed Damage State Framework
Damage Damage Functional
parameter state level Description
Cracking DS-1 Immediate Onset of hairline cracks
Yielding DS-2 Limited Theoretical first yield
of longitudinal rebar
Spalling DS-3 Service disruption Onset of concrete spalling
Core crushing DS-4 Life safety Crushing of core concrete
4 states. The yielding of SMA rebar was monitored by defining
the yield strain of SMA bar and tracking the occurrence of first
yield in SMA rebar. The spalling strain was assumed to be 0.004
as suggested by Priestley et al. (1996). Paulay and Priestley (1992)
found that the crushing strain of confined concrete ranges between
0.015 and 0.05. In this study, the crushing strain of confined con-
crete for different SMA-RC bridge piers was calculated using the
Paulay and Priestley (1992) equation
εcu ¼ 0.004 þ 1.4ρs f yh εsm =f c0 ð2Þ
where εcu = ultimate compression strain (0.019); εsm = steel strain
at maximum tensile stress (0.14); f c0 = concrete compressive
strength in MPa (42.4 MPa); f yh = yield strength of transverse steel
in MPa (450 MPa); and ρs = volumetric ratio of confining
steel (0.7%).
Most of the damage states available in literature are discrete in
nature and quantify the damage deterministically (Marsh and
Stringer 2013). Practically, the drift level corresponding to certain
damage is not a discrete deterministic quantity and each damage
level is associated with a distribution of values. The drift limits de-
fined at different damage states should clearly indicate whether it
represents the lower bound, median, or some intermediate value for
the onset of a particular damage type. In order to develop compre-
hensive performance-based damage states, in this study, the prob-
abilistic distribution of each damage state was also identified and
the median of the distribution was defined as the drift limit corre-
sponding to each damage state.
In order to determine the limit state drift values for different per-
formance levels, the drift limits corresponding to the strain values
were determined using IDA results for different hazard levels for
the five different SMA-RC bridge piers. The drift limits at various
performance levels were identified using the dynamic pushover
curves obtained from IDA. Dynamic pushover curves represent
the relation between the maximum drift and its corresponding base
shear obtained from IDA while being subjected to an earthquake
record (Elnashai and Luigi 2008). These curves represent the struc-
tural capacity under specific earthquake loading. Dynamic push-
over curves, obtained from IDA, take into account progressive
structural stiffness degradation, change of modal characteristics,
and period elongation of the structure for increasing values of ex-
ternal action, which is not achievable through static pushover
analysis. Inelastic characteristics such as strength degradation
and energy dissipation largely affect the seismic performance of
structures, which are also required for developing performance-
based damage states for performance-based design. The drift levels
for different performance levels obtained from IDA were used to
find a suitable distribution for each damage state that describes
the statistical distribution of the developed damage states. Statisti-
cal analyses were carried out to find the most suitable probability
density function (PDF) to represent the data related to each damage
state. Using statistical tools and analyses, a suitable distribution
for each damage state was determined using goodness-of-fit tests.
04016140-6 J. Struct. Eng.
 The following section discusses the development of performance-
based damage states for five different SMA-reinforced bridge piers.
Probabilistic Distribution of Drift-Based Damage States
Using the results obtained from IDA, the probabilistic distribution
of each damage state corresponding to different hazard levels was
determined to represent the statistical variability of damage states
at different hazard levels. The probabilistic distribution of each
damage state, i.e., yielding, spalling, and crushing, was superim-
posed on the dynamic pushover curves obtained from IDA and
are shown in Figs. 6–10. The expected (median) drift level at a
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
particular damage state is represented by the vertical solid line.
From Figs. 6–10 it can be observed that the uncertainty of each
damage state is unique, as indicated by the dispersion or width
of the distribution. The median drift level of each damage state
is defined as the limiting drift value for each performance level.
The drift levels corresponding to different damage states for
different hazard levels are shown in Table 5. Table 5 also shows
the probabilistic distribution of each damage state. From IDA,
measurements of drift levels corresponding to each damage state
were obtained and statistically processed to find out the most
suitable distribution. The suitability of the selected distributions
for representing each damage state was evaluated using a

Fig. 6. Dynamic pushover response and different damage states with distribution for SMA-RC-1 for (a) 2% in 50 years; (b) 5% in 50 years; (c) 10% in
50 years probability of exceedance

Fig. 7. Dynamic pushover response and different damage states with distribution for SMA-RC-2 for (a) 2% in 50 years; (b) 5% in 50 years; (c) 10% in
50 years probability of exceedance

Fig. 8. Dynamic pushover response and different damage states with distribution for SMA-RC-3 for (a) 2% in 50 years; (b) 5% in 50 years; (c) 10% in
50 years probability of exceedance

© ASCE 04016140-7 J. Struct. Eng.

J. Struct. Eng., 2016, 142(12): 04016140

Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.

Fig. 9. Dynamic pushover response and different damage states with distribution for SMA-RC-4 for (a) 2% in 50 years; (b) 5% in 50 years; (c) 10% in
50 years probability of exceedance

Fig. 10. Dynamic pushover response and different damage states with distribution for SMA-RC-5 for (a) 2% in 50 years; (b) 5% in 50 years; (c) 10%
in 50 years probability of exceedance

Table 5. Damage States of Different SMA-RC Bridge Pier and Their Associated Distribution
Drift (%)
Probability of exceedance
SMA-1 SMA-2 SMA-3 SMA-4 SMA-5
Damage Damage 2% 5% 10% 2% 5% 10% 2% 5% 10% 2% 5% 10% 2% 5% 10%
parameter state in 50 in 50 in 50 in 50 in 50 in 50 in 50 in 50 in 50 in 50 in 50 in 50 in 50 in 50 in 50 Distribution
Cracking DS-1 0.28 0.28 0.28 0.30 0.28 0.28 0.28 0.28 0.28 0.28 0.28 0.28 0.28 0.28 0.28 Uniform
Yielding DS-2 1.68 1.76 1.86 1.66 1.72 1.80 2.28 2.42 2.58 1.74 1.83 1.95 1.10 1.16 1.21 Lognormal
Spalling DS-3 2.66 2.79 2.88 2.69 2.77 2.87 1.64 1.72 1.80 2.52 2.61 2.68 1.97 2.02 2.10 Normal
Crushing DS-4 5.05 5.68 5.94 5.51 5.91 6.05 7.65 7.81 7.94 5.56 5.63 5.72 4.73 4.79 4.84 Gamma

Kolmogorov-Smirnov (K-S) goodness-of-fit test. The following
conclusions are derived from the distribution of different damage
states:
• Irrespective of the type of SMA and earthquake hazard level,
cracking occurs at a drift of 0.28%, and it can be represented
better with a uniform distribution. Since the cracking strain
of concrete depends only on the tensile strength of concrete,
small variations in concrete cracking drift were observed. Uni-
form distribution is a preferable choice when all of the outcomes
have an equal probability of occurring. Since the cracking drift
of all the SMA-RC bridge piers ranged between 0.28 and 0.30%
and have equal probability of occurrence, the cracking drift is
assumed to follow a uniform probability distribution. Results of
the K-S goodness-of-fit test also confirmed the suitability of a
uniform distribution for representing the distribution of cracking
drift. This drift value of 0.28% can be identified as Damage
State-1(DS-1);
• From statistical analyses, it was found that the lognormal
distribution better represents the uncertainty in drift limits for
DS-2 (yielding). Usually the variation in metal strength, such
as yield strength of steel, is better represented by a lognormal
distribution (Ellingwood 1977; Ghobarah et al. 1998). Similar
distributions for yield drift limits for SMA-RC bridge piers were
obtained, which is largely dependent on the yield strength
of SMA;
• A normal distribution was found to be the best fit for repre-
senting the variability in drift limits corresponding to DS-3
(spalling) based on the K-S goodness-of-fit test. A normal dis-
tribution is better suited for representing the spalling drift since
all the SMA-RC bridge piers showed a strong tendency towards

© ASCE 04016140-8 J. Struct. Eng.

J. Struct. Eng., 2016, 142(12): 04016140

 the central value of spalling drift as well as the positive and ne-
gative deviations from this central value are equally likely. The
selected distribution seems reasonable since concrete strength
can be better represented by a normal distribution (Ellingwood
1977; Mirza et al. 1979); and
• A K-S goodness-of-fit test was performed to identify the most
suitable distribution for defining the variation of DS-4 (crush-
ing). The K-S goodness-of-fit test indicated that the gamma dis-
tribution, which usually indicates an extreme event, provides
best fit to the data and was the most suitable for representing
the crushing drift.
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
Maximum Drift–Based Damage States
Figs. 6–10 show the dynamic pushover curves for SMA-RC-1
through SMA-RC-5 under different levels of earthquakes, respec-
tively. The dynamic pushover curves derived from 10 earthquakes
(for each bridge pier) were statistically processed to obtain the
median, 5th percentile, and 95th percentile capacity curves.
Comparisons of Figs. 6–10 reveal that:
• SMA-RC-3 (Fig. 8) has higher deformation and strength capa-
city as compared to the other SMA-RC bridge piers;
• For seismic hazard level of 2% in 50 years, the median capacity
of SMA-RC-3 was 2,743 kN which was 16%, 15%, 20%, and
17% higher than that of SMA-RC-1, SMA-RC-2, SMA-RC-4,
and SMA-RC-5, respectively; and
• Maximum base shear demand is also significantly influenced by
the earthquake hazard level. For example, the median maximum
base shear of SMA-RC-1, for 2% in 50 years is 2,305 kN which
is 5% and 7% higher than that of 5% and 10% in 50-years re-
cords, respectively.
Evaluation of the results presented in Table 5 provides a valu-
able insight on the damage states developed for different SMA-RC
bridge piers. The damage states are defined for different hazard
levels. From Table 5 it can be observed that
• Damage State-2 or yielding occurs at a drift level below 2%
except for SMA-RC-3. At DS-2, there is significant variation
in drift limits for different SMA-RC bridge piers. For SMA-
RC-1 and SMA-RC-2, the drift limit is quite similar irrespective
of the earthquake hazard levels which ranges between 1.68–
1.86% and 1.66–1.80%, respectively. Since SMA-RC-1 and
SMA-RC-2 are reinforced with Ni-Ti SMAs (different compo-
sitions) with similar mechanical properties, they tend to have
similar drift limits at DS-2;
• Before yielding, SMA-RC-3 sustained higher drift compared to
the other SMA-RC piers. At DS-2, under 2% in 50 years hazard
level, the drift limit for SMA-RC-3 was 2.28%, which was sig-
nificantly higher than the drift limits obtained for the other
SMA-RC bridge piers. This is expected since SMA-3 has higher
yield strength and post yield stiffness as compared to the other
SMAs considered in this study;
• Although SMA-4 has a low elastic modulus, its yield strength
is very high, which eventually increased the yield strain and
resulted in higher drift values. At DS-2, the drift limits for
SMA-RC-4 were 3.4%, 3.8%, and 4.6% higher than that of
SMA-RC-1 for 2%, 5%, and 10% in 50-years probability of ex-
ceedance, respectively;
• Exceptional performance was observed for SMA-RC-5, which
yielded at a very low drift level (1.1–1.2%) as compared to the
other SMA-RC piers. This is due to SMA-5’s very low yield
strength to elastic modulus ratio (0.0033), which reduced the
drift capacity of SMA-RC-5;
• Although both SMA-3 and SMA-5 are Fe-based, due to the var-
iation in their yield strength and elastic modulus, the drift limits
© ASCE 04016140-9
J. Struct. Eng., 2016, 142(12): 04016140
for SMA-RC-3 at DS-2 were 52% higher than that of SMA-
RC-5, irrespective of the hazard level;
• From Table 5, it can be observed that except for SMA-RC-3,
yielding occurred in all the bridge piers before the initiation
of cover spalling. The delayed rebar yielding of SMA-RC-3
can be attributed to its higher yield strength. A similar observa-
tion of the SMA-RC column has been reported by Saiidi and
Wang (2006) where spalling of cover concrete took place before
the initiation of SMA yielding;
• DS-3 is considered at the onset of cover concrete spalling. All
the piers experienced yielding before spalling where the only
exception was SMA-RC-3. For SMA-RC-1, DS-3 occurred
at a drift level of 2.66%, 2.79%, and 2.88% for 2%, 5%, and
10% in a 50-years hazard level, respectively. This is expected
since a hazard level with lower probability indicates a more da-
maging earthquake. A similar trend is also observed for the
other SMA-RC bridge piers where the limiting drift value in-
creased with decreased return period. In terms of drift limit,
SMA-RC-1 and SMA-RC-2 performed better than the other
three SMA-RC piers as they could sustain more drift before
entering into DS-3;
• At DS-4 (crushing of concrete), all the SMA-RC bridge piers
sustained more than 5% drift under various hazard levels
whereas the SMA-RC-3 exceeded 7.5%. For a hazard level with
a 2% of probability of exceedance in 50 years, SMA-RC-3
sustained a drift of 7.65% before crushing, which was 34%,
28%, 27%, and 38% higher than that of SMA-RC-1, SMA-
RC-2, SMA-RC-4, and SMA-RC-5, respectively; and
• For a particular SMA-RC bridge pier, crushing drift also varied
significantly at different hazard levels. In the case of SMA-RC-
1, the crushing drift corresponding to 2% in 50-years hazard
level is 11.5% and 15% lower than the crushing drift at 5%
and 10% in 50-years hazard level, respectively. However, in
the case of SMA-RC-5, the crushing drift at 2% in 50-years ha-
zard level was 2.3% and 1.25% lower than the crushing drift at
5% and 10% in 50-years hazard level, respectively.
The drift limits presented in Table 5 can be used for perfor-
mance-based design of SMA-RC bridge pier. Based on the design
earthquake scenario, the designer can define the target performance
level and associated drift limits. Since, performance-based damage
states are proposed for different types of SMA, the designer can
select any particular SMA and design the bridge pier according
to the owners expected performance level.
Residual Drift–Based Damage States for SMA-RC
Bridge Piers
Residual drift has been considered as one of the significant perfor-
mance indicators in judging a structure’s postearthquake safety and
the economic feasibility for repairing (Ramirez and Miranda 2012).
Although residual drift dictates the postearthquake functionality of
highway bridges, no other design guidelines except the Japanese
code for highway bridge design (JRA 2006) provide any residual
drift limit of bridge piers. In a recent study, Saiidi and Ardakani
(2012) found that bridge piers meeting current seismic require-
ments can withstand larger traffic loads even when the residual
drift is 1.2% or more. Lee and Billington (2011) considered a
1% residual drift to be large enough for a bridge replacement.
In order to develop the damage states (DS) for SMA-RC bridge
piers, a probabilistic approach has been adopted in this study.
Based on the existing literature (O’Brien et al. 2007; Billah and
Alam 2014), four different damage states have been identified
and a range of limiting residual drifts were considered. It was
J. Struct. Eng.
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.

Fig. 11. Fragility curves in terms of residual drift at (a) 10% in 50 years; (b) 5% in 50 years; (c) 2% in 50 years probability of exceedance

assumed that a residual drift below 0.25% would meet the service-
ability requirement (DS-1) while a residual drift larger than 1%
would be characterized as a collapse damage state (DS-4). The in-
termediate Damage States DS-2 and DS-3 are assumed to take
place at a residual drift larger than 0.5% and 0.75%, respectively.
DS-1 requires that no structural realignment is necessary and the
bridge is fully operational. DS-2 consists of minor structural repair-
ing and requires the bridge to be operational without requiring
bridge closure. A pier experiencing DS-3 will require major repair
and may require bridge closure but should be usable for restricted
emergency traffic after inspection. DS-4 corresponds to the case
when the residual drift is sufficiently large that the structure is
in danger of collapse from earthquake aftershocks.
Once the damage states were identified, fragility curves for
residual drifts were developed using the IDA results for three differ-
ent seismic hazard levels. In this study, fragility functions were
developed using Eq. (3), which takes the form of lognormal cumu-
lative distribution functions having a median value of θ and loga-
rithmic standard deviation or dispersion of β
 
lnðRD=θÞ
FðRDÞ ¼ ϕ ð3Þ
β difference goes down to 6.5 and 13.1% for DS-4.
the limiting value for the corresponding damage state. For example,
in Fig. 11(a) (10% in 50 years), the 50% probability of occurrence
of DS-2 corresponds to a RD of 0.48% while the limiting RD values
for DS-2 for hazard levels of 5% and 2% in 50 years correspond to
0.55% and 0.62%, respectively. It can be observed that the limiting
RD value for DS-2 was assumed to be 0.5% and the values obtained
from the median probability of exceedance are quite close. Simi-
larly, the limiting RD values with a 50% probability of occurrence
at different damage states and hazard levels were developed as out-
lined in Table 6. From Table 6, it can be observed that as the ground
motion return period decreases (probability of occurrence in-
creases), the limiting residual drift corresponding to different DS
decreases. For example, at DS-4, the limiting drift value for an
earthquake with a 2,475 years return period is 1.22%, which is
6.5% and 13.1% higher than an earthquake with 975 years and
475 years return periods, respectively. Observations from Table 6
indicate that, as the damage level increases (DS-1 to DS-4) the dif-
ference in limiting RD values at different hazard levels decreases.
For instance, at DS-2, the limiting RD value corresponding to a
2,475 years return period is 11% and 22.5% higher than that of
975 years and 475 years return period, respectively. However, this

where FðRDÞ = conditional probability that the bridge pier will

be damaged to a given DS as a function of the residual drift
(RD); F = standard normal cumulative distribution function; and
θ and β = median value of the probability distribution and the log-
arithmic standard deviation corresponding to the DS, respectively.
Fig. 11 shows the fragility curves for SMA-RC bridge piers for
different damage states at three different hazard levels. Here, the
fragility curves are plotted irrespective of the SMA types to general-
ize the associated damage states. Using these fragility curves, the
residual drift–based damage states for SMA-RC bridge piers were
developed. From the fragility curves corresponding to each damage
state, the RD value with a 50% probability of occurrence indicates
Prediction of Residual Drift
For performance-based design, prediction of residual drift as
a function of the target or maximum drift would be very useful.
Previous research has shown that residual drift predictions using
nonlinear analysis are highly variable and subjected to different
modeling features (Applied Technology Council, ATC-58) (ATC
2012). Recently ATC-58 (ATC 2012) recommended some general
equations for predicting residual drift using peak transient drift
and yield drift. ATC-58 (ATC 2012) suggested that prediction of
residual drift requires advanced nonlinear simulation with careful

Table 6. Residual Drift Damage States of SMA-RC Bridge Pier

Residual drift, RΔ (%)
Probability of exceedance
Damage state Functional level Description 10% in 50 5% in 50 2% in 50
Slight (DS ¼ 1) Fully operational No structural realignment is necessary 0.24 0.28 0.33
Moderate (DS ¼ 2) Operational Minor structural repairing is necessary 0.48 0.55 0.62
Extensive (DS ¼ 3) Life safety Major structural realignment is required to restore safety margin 0.73 0.82 0.87
for lateral stability
Collapse (DS ¼ 4) Collapse Residual drift is sufficiently large that the structure is in danger of 1.04 1.16 1.22
collapse from earthquake aftershocks

© ASCE 04016140-10 J. Struct. Eng.

J. Struct. Eng., 2016, 142(12): 04016140

 3 0.5
2.5
Predicted RD (%)
2
1.5
1
0.5
0 0
0 0.5 1 1.5 2 2.5
(a) Experimental RD (%)
Fig. 12. Comparison of residual drift prediction with experimental results: (a) O’Brien et al. (2007); (b) Youssef et al. (2008)
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
attention to cyclic hysteretic response of the models and numerical
accuracy of the solution. In this study, the residual drift responses
were obtained using IDA, which is one of the most advanced non-
linear analysis techniques and the models were validated against
experimental results. From the residual drift responses of the
SMA-RC bridge piers, it was found that the residual drift in
SMA-RC bridge piers is a function of maximum drift and supere-
lastic strain of the SMA. Using the residual drift responses obtained
from different SMA-RC bridge piers under a wide range of ground
motions, a nonlinear regression analysis was conducted to inves-
tigate the effect of maximum drift and superelastic strain on the
residual drift response. Using a nonlinear regression analysis,
the following equation was developed for predicting the residual
drift of SMA-RC bridge piers:
     
εs εs 1 spalling, and core crushing. For all SMA-RC bridge piers, the
RD ¼ 0.5 MD2 − MD þ ð4Þ
100 100 εs
where RD = residual drift (%); εs = superelastic strain; and MD =
maximum drift (%).
In order to investigate the accuracy of the proposed residual drift
prediction equation, a comparison was carried out with experimen-
tal results. Fig. 12 shows the comparison of residual drifts obtained
from experimental investigations and the prediction equation.
Fig. 12(a) shows the comparison with experimental results of
O’Brien et al. (2007), where a SMA-RC bridge pier was tested
under reverse cyclic loading. The bridge pier was constructed using
Ni-Ti SMA in the plastic hinge region and the superelastic strain of
SMA rebar was 6%. Maximum drift values and the corresponding
residual drifts were obtained from experimental results. Using the
maximum drift value and the superelastic strain of the SMA rebar,
the residual drifts were predicted. Fig. 12(a) shows that the pro-
posed equation predicted the residual drift very well with an aver-
age absolute error (AAE) of 4.65% and average standard deviation
of 0.03. Fig. 12(b) shows the comparison of residual drift predic-
tion with the experimental results of Youssef et al. (2008) where
Ni-Ti SMA with superelastic strain of 6% was used as reinforce-
ment in the beam-column joint. From Fig. 12(b) it is evident that
the proposed equation is capable of predicting the residual drift
with reasonable accuracy with an AAE of 2.06% and average stan-
dard deviation of 0.015.
Conclusions
In order to develop a performance-based design guideline for
SMA-RC bridge piers, structural performance objectives and their
corresponding limit state criteria must be clearly defined. Due to
significant differences in the behavior of SMA-reinforced bridge
piers as compared to conventional piers, damage states for typical
© ASCE
J. Struct. Eng., 2016, 142(12): 04016140
0.4
Predicted RD (%)
0.3
0.2
0.1
3 0 0.1 0.2 0.3 0.4 0.5
(b) Experimental RD (%)
bridge piers may not be applicable for SMA-RC bridge piers.
In this study, a set of performance-based damage states for
bridge piers reinforced with five different types of SMAs has been
developed in terms of both maximum and residual drift. Using an
IDA-based approach, this study proposed performance-based drift
levels for different damage states considering three different hazard
levels. To predict the residual deformation of SMA-RC bridge
piers, a relationship between maximum drift, residual drift, and
superelastic strain of SMA has been developed. The prediction
equation was validated against experimental observations from
SMA reinforced concrete members. Based on the results of this
study, the following major conclusions can be drawn:
• The progression of damage was similar for all the RC bridge
piers reinforced with different SMAs (except for SMA-RC-3):
concrete cracking, longitudinal reinforcement yielding, cover
cracking occurred at the same level of drift (due to same cross
section) while the drift at other performance levels varied based
on the mechanical properties of SMA used;
• Different performance-based drift limits, i.e., cracking, spal-
ling, yielding, and crushing of SMA-RC bridge piers, strongly
follow uniform, normal, lognormal, and gamma distributions,
respectively. These distributions can be used for selecting the
target drift levels for performance-based design of SMA-RC
piers;
• Except for DS-1(cracking), other three damage states are signif-
icantly influenced by the type of SMA used. For DS-2 (yield-
ing), the limiting maximum drift varies from 1.18% (SMA-5) to
2.28% (SMA-3) and for DS-3 (spalling), the limiting maximum
drift varies from 1.64% (SMA-3) to 2.69% (SMA-2) for hazard
level corresponding to a 2,475 years return period;
• In terms of maximum drift, consideration of different hazard
levels does not have any significant impact on DS-1 and DS-2.
On the other hand, different hazard levels have substantial im-
pacts on DS-3 and DS-4;
• The proposed residual drift limit states tend to decrease with
increased probability of occurrence (decreases return period).
The damage states developed in terms of residual drifts correlate
well with damages observed from experimental studies;
• Residual drift can be expressed as a function of maximum drift
and the superelastic strain of SMA rebar;
• Based on the residual drift responses of all the SMA-RC piers
under different levels of ground motions, a prediction equation
was developed to predict the residual drift response of an SMA-
RC bridge pier. The proposed equation can correlate very well
with experimental observations; and
• The proposed equation can be used for predicting the residual
drift of SMA-RC bridge pier when designing the pier for a target
residual drift. Based on the maximum drift and residual drift, the
designer would be able to select an SMA with the required
04016140-11 J. Struct. Eng.
 superelastic strain, thereby ensuring the safety of bridges under
extreme earthquake events.
The current study considers one particular type of bridge pier
model without considering uncertainty in geometry and material
properties such as concrete and steel properties. A further study
using bridge piers with various aspect ratios, axial load levels,
and with different sets of geometry/material properties should be
conducted for better understanding the contributions of other
parameters to the performance-based damage states for SMA-
reinforced concrete bridge piers. However, the results of this study
will work as a benchmark for evaluating future studies, developing
more-refined damage states, and improving the capacity limit states
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
for SMA-RC bridge piers.
Acknowledgments
The financial contributions of Natural Sciences and Engineering
Research Council of Canada (NSERC) through Discovery Grant
and Industrial Postgraduate Scholarship Program were critical to
conduct this study and are gratefully acknowledged.
References
AASHTO. (2011). AASHTO guide specifications for LRFD seismic bridge
design, LRFDSEIS-2-M, 2nd ed., Washington, DC.
Abrahamson, N. A. (1992). “Non-stationary spectral matching.” Seismol.
Res. Lett., 63(1), 30.
Alam, M. S., Youssef, M. A., and Nehdi, M. (2007). “Utilizing shape
memory alloys to enhance the performance and safety of civil infra-
structure: A review.” Can. J. Civ. Eng., 34(9), 1075–1086.
Alam, M. S., Youssef, M. A., and Nehdi, M. (2008). “Analytical prediction
of the seismic behaviour of superelastic shape memory alloy reinforced
concrete elements.” Engr. Struct., 30(12), 3399–3411.
Alam, M. S., Youssef, M. A., and Nehdi, M. (2009). “Seismic performance
of concrete frame structures reinforced with superelastic shape memory
alloys.” Smart Struct. Syst., 5(5), 565–585.
Applied Technology Council. (2012). “Seismic performance assessment of
buildings.” ATC-58, Redwood City, CA.
Araki, Y., et al. (2010). “Potential of superelastic Cu-Al–Mn alloy bars for
seismic applications.” Earthquake Eng. Struct. Dyn., 40(1), 107–115.
Araki, Y., Maekawa, N., Omori, T., Sutou, Y., Kainuma, R., and Ishida, K.
(2012). “Rate-dependent response of superelastic Cu–Al–Mn alloy rods
to tensile cyclic loads.” Smart Mater. Struct., 21, 032002.
Auricchio, F. and Sacco, E. (1997). “Superelastic shape-memory-alloy
beam model.” J. Intell. Mater. Syst. Struct., 8(6), 489–501.
Baker, J. W., Lin, T., Shahi, S. K., and Jayaram, N. (2011). “New ground
motion selection procedures and selected motions for the PEER trans-
portation research program.” PEER Technical Rep. 2011/03, PEER
Center, Univ. of California, Berkeley, CA.
Bentz, E. C., Vecchio, F. J., and Collins, M. P. (2006). “Simplified modified
compression field theory for calculating shear strength of reinforced
concrete elements.” ACI Struct. J., 103(4), 614–624.
Berry, M., and Eberhard, M. (2003). “Performance models for flexural
damage in reinforced concrete columns.” PEER Technical Rep.
2003/18, PEER Center, Univ. of California, Berkeley, CA.
Billah, A. H. M. M., and Alam, M. S. (2012). “Seismic performance of
concrete columns reinforced with hybrid shape memory alloy
(SMA) and fiber reinforced polymer (FRP) bars.” Constr. Build. Mater.,
28(1), 730–742.
Billah, A. H. M. M., and Alam, M. S. (2013). “Statistical distribution of
seismic performance criteria of retrofitted multi-column bridge bents
using incremental dynamic analysis: A case study.” Bull. Earthquake
Eng., 11(6), 2333–2362.
Billah, A. H. M. M., and Alam, M. S. (2014). “Seismic fragility assessment
of concrete bridge piers reinforced with superelastic shape memory
alloy.” Earthquake Spectra, 31(3), 1515–1541.
© ASCE 04016140-12
J. Struct. Eng., 2016, 142(12): 04016140
Casarotti, C., and Pinho, R. (2006). “Seismic response of continuous span
bridges through fibre-based finite element analysis.” Earthquake Eng.
Eng. Vibr., 5(1), 119–131.
Christopoulos, C., Pampanin, S., and Priestley, M. J. N. (2003).
“Performance-based seismic response of frame structures including
residual deformations. I: Single-degree-of-freedom systems.” J. Earth-
quake Eng., 7(1), 97–118.
Cruz, N. C., and Saiidi, M. S. (2012). “Shake-table studies of a four-span
bridge model with advanced materials.” J. Struct. Eng., 10.1061/
(ASCE)ST.1943-541X.0000457, 183–192.
CSA (Canadian Standards Association). (2010). “CSA S6-10, CHBDC
2010, Canadian highway bridge design code.” CAN/CSA S6-1,
Canadian Standards Association, Rexdale, ON.
Czaderski, C., Shahverdi, M., Brönnimann, R., Leinenbach, C., and
Motavalli, M. (2014). “Feasibility of iron-based shape memory alloy
strips for prestressed strengthening of concrete structures.” Constr.
Build. Mater., 56, 94–105.
Dezfuli, F. H., and Alam, M. S. (2013). “Shape memory alloy wire-
based smart natural rubber bearing.” Smart Mater. Struct., 22(4),
45013–45029.
Ellingwood, B. (1977). “Statistical analysis of RC Beam column interac-
tion.” J. Struct. Eng., 103, 1377–1388.
Elnashai, A. S., and Luigi, D. S. (2008). Fundamentals of earthquake
engineering, Wiley, New York.
Erochko, J., Christopoulos, C., Tremblay, R., and Choi, H. (2011).
“Residual drift response of SMRFs and BRB frames in steel buildings
designed according to ASCE 7-05.” J. Struct. Eng., 10.1061/(ASCE)ST
.1943-541X.0000296, 589–599.
Ghassemieh, M., Mostafazadeh, M., and Sadeh, M. S. (2012). “Seismic
control of concrete shear wall using shape memory alloys.” J. Intell.
Mater. Syst. Struct., 23(5), 535–543.
Ghobarah, A., Aly, N. M., and El-Attar, M. (1998). “Seismic reliability
assessment of existing reinforced concrete buildings.” J. Earthquake
Eng., 2(4), 569–592.
Hancock, J., et al. (2006). “An improved method of matching response
spectra of recorded earthquake ground motion using wavelets.”
J. Earthquake Eng., 10(1), 67–89.
Hose, Y., Silva, P., and Seible, F. (2000). “Development of a performance
evaluation database for concrete bridge components and systems under
simulated seismic loads.” Earthquake Spectra, 16(2), 413–442.
JRA (Japan Road Association). (2006). “Specifications for highway
bridges.” Tokyo.
Lee, W. K., and Billington, S. K. (2011). “Performance-based earthquake
engineering assessment of a self-centering, post-tensioned concrete
bridge system.” Earthquake Engr. Struct. Dyn., 40(8), 887–902.
Lehman, D., Moehle, J., Mahin, S., Calderone, A., and Henry, L. (2004).
“Experimental evaluation of the seismic performance of reinforced con-
crete bridge columns.” J. Struct. Eng., 10.1061/(ASCE)0733-9445
(2004)130:6(869), 869–879.
Lindt, J. W., and Potts, A. (2008). “Shake table testing of a superelastic
shape memory alloy response modification device in a wood shear
wall.” J. Struct. Eng., 10.1061/(ASCE)0733-9445(2008)134:8(1343),
1343–1352.
Mander, J. B., Priestley, M. J. N., and Park, R. (1988). “Theoretical stress-
strain model for confined concrete.” J. Struct. Eng., 10.1061/(ASCE)
0733-9445(1988)114:8(1804), 1804–1826.
Marsh, L. K., and Stringer, S. J. (2013). “Performance-based seismic bridge
design, a synthesis of highway practice.” NCHRP Synthesis-440, TRB,
Washington, DC.
Menegotto, M., and Pinto, P. E. (1973). “Method of analysis for cyclically
loaded R.C. plane frames including changes in geometry and non-
elastic behaviour of elements under combined normal force and
bending.” Symp. on the Resistance and Ultimate Deformability of
Structures Acted on By Well-Defined Repeated Loads, International
Association for Bridge and Structural Engineering, Zurich, Switzerland,
15–22.
Mirza, S. A., Hatzinikolas, M., and MacGregor, J. G. (1979). “Statistical
descriptions of strength of concrete.” J. Struct. Eng., 105(ST6),
1021–1036.
J. Struct. Eng.
 National Research Council of Canada. (2014). “Canadian highway bridge
design code.” CAN/CSA-S6-14, Ottawa, ON.
Naumoski, N., Tso, W. K., and Heidebrecht, A. C. (1988). “A selection of
representative strong motion earthquake records having different A/V
ratios.” EERG Rep. 88-01, Dept. of Civil Engineering, McMaster Univ.,
Hamilton, ON, Canada.
O’Brien, M., Saiidi, M. S., and Zadeh, M. S. (2007). “A study of concrete
bridge columns using innovative materials subjected to cyclic loading.”
Rep. No. CCEER-07-01, CCEER, Dept. of Civil Engineering, Univ. of
Nevada, Reno, NV.
Ocel, J., et al. (2004). “Steel beam-column connections using shape
memory alloys.” J. Struct. Eng., 10.1061/(ASCE)0733-9445(2004)
130:5(732), 732–740 .
Omori, T., et al. (2011). “Superelastic effect in polycrystalline ferrous
Downloaded from ascelibrary.org by University Of British Columbia on 04/12/19. Copyright ASCE. For personal use only; all rights reserved.
alloys.” Science, 333(6038), 68–71.
Paulay, T., and Priestley, M. N. J. (1992). Seismic design of reinforced
concrete and masonry buildings, Willey, New York.
PEER (Pacific Earthquake Engineering Research). (2011). “New ground
motion selection procedures and selected motions for the PEER trans-
portation research program.” PEER Rep. 2011/03, Univ. of California,
Berkeley, CA.
Pettinga, D., Pampanin, S., Christopoulos, C., Priestley, N. (2006).
“Accounting for residual deformations and simple approaches to their
mitigation.” 1st European Conf. on Earthquake Engineering and
Seismology, Geneva, Switzerland.
Pinho, R., Casarotti, C., and Antoniou, S. (2007). “A Comparison of single-
run pushover analysis techniques for seismic assessment of bridges.”
Earthquake Eng. Struct. Dyn., 36(10), 1347–1362.
Priestley, M. J. N., Seible, F., and Calvi, G. M. (1996). Seismic design and
retrofit of bridges, Wiley, New York.
Ramirez, C. M., and Miranda, E. (2012). “Significance of residual drifts
in building earthquake loss estimation.” Earthquake Eng Struct Dyn.,
41(11), 1477–1493.
© ASCE 04016140-13
J. Struct. Eng., 2016, 142(12): 04016140
Saiidi, M. S., and Ardakani, S. M. S. (2012). “An analytical study of
residual displacements in RC bridge columns subjected to near-fault
earthquakes.” Bridge Struct., 8(1), 35–45.
Saiidi, M. S., O’Brien, M., and Zadeh, M. S. (2009). “Cyclic response
of concrete bridge columns using superelastic nitinol and bendable
concrete.” ACI Struct J., 106(1), 69–77.
Saiidi, M. S., and Wang, H. (2006). “Exploratory study of seismic response
of concrete columns with shape memory alloys reinforcement.” ACI
Struct. J., 103(3), 435–442.
SeismoMatch version 2.1 [Computer software]. SeismoMatch Earthquake
Engineering Software Solutions, Pavia, Italy.
Seismostruct version 6.0 [Computer software]. Pavia, Italy.
Shrestha, K. C., et al. (2013). “Feasibility of Cu–Al–Mn superelastic alloy
bars as reinforcement elements in concrete beams.” Smart Mater.
Struct., 22(2), 025025.
Tanaka, Y., Himuro, Y., Kainuma, R., Sutou, Y., Omori, T., and Ishida, K.
(2010). “Ferrous polycrystalline shape-memory alloy showing huge
superelasticity.” Science, 327(5972), 1488–1490.
Vamvatsikos, D., and Cornell, C. A. (2002). “Incremental dynamic
analysis.” Earthquake Engr. Struct. Dyn., 31(3), 491–514.
Vecchio, F. J., and Collins, M. P. (1986). “The modified compression-filed
theory for reinforced concrete elements subjected to shear.” ACI Struct.
J., 83(2), 219–231.
Wang, D. S., Ai, Q. H., Li, H. N., Si, B. J., and Sun, Z. G. (2008).
“Displacement based seismic design of RC bridge piers: Method
and experimental evaluation.” 14th World Conf. on Earthquake Engi-
neering, Beijing, China.
Youssef, M. A., Alam, M. S., and Nehdi, M. (2008). “Experimental
investigation on the seismic behavior of beam-column joints reinforced
with superelastic shape memory alloys.” J. Earthquake Eng., 12(7),
1205–1222.
J. Struct. Eng.