EARTHQUAKE ENGINEERING & STRUCTURAL DYNAMICS

Earthquake Engng Struct. Dyn. 2016

Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/eqe.2801

Seismic performance assessment of a hybrid coupled wall system

with replaceable steel coupling beams versus traditional RC
coupling beams

Xiaodong Ji1,*,†, Dan Liu1, Ya Sun2 and Carlos Molina Hutt3

1
Department of Civil Engineering, Key Laboratory of Civil Engineering Safety and Durability of China Education
Ministry, Tsinghua University, Beijing 100084, China
2
China Architecture Design and Research Group, Beijing 100044, China
3
Civil, Environmental and Geomatic Engineering, University College London (UCL), London WC1E6BT, UK

SUMMARY
This study assesses the seismic performance of a hybrid coupled wall (HCW) system with replaceable steel
coupling beams (RSCBs) at four intensities of ground motion shaking. The performance of the HCW system
is benchmarked against the traditional reinforced concrete coupled wall (RCW). Nonlinear numerical
models are developed in OpenSees for a representative wall elevation in a prototype 11-story building
designed per modern Chinese codes. Performance is assessed via nonlinear dynamic analysis. The results
indicate that both systems can adequately meet code defined objectives in terms of global and component
behavior. Behavior of the two systems is consistent under service level earthquakes, whereas under more
extreme events, the HCW system illustrates enhanced performance over the RCW system resulting in peak
interstory drifts up to 31% lower in the HCW than the RCW. Larger drifts in the RCW are because of
reduced coupling action induced by stiffness degradation of RC coupling beams, whereas the stable
hysteretic responses and overstrength of RSCBs benefit post-yield behavior of the HCW. Under extreme
events, the maximum beam rotations of the RSCBs are up to 42% smaller than those of the RC coupling
beams. Moderate to severe damage is expected in the RC coupling beams, whereas the RSCBs sustain
damage to the slab above the beam and possible web buckling of shear links. The assessment illustrates
the benefits of the HCW with RSCBs over the RCW system, because of easy replacement of the shear links
as opposed to costly and time-consuming repairs of RC coupling beams. Copyright © 2016 John Wiley &
Sons, Ltd.

Received 9 February 2016; Revised 12 July 2016; Accepted 29 July 2016

KEY WORDS: hybrid coupled wall (HCW); replaceable steel coupling beam (RSCB); RC coupling beam;
nonlinear dynamic analysis; seismic fragility; damage assessment

1. INTRODUCTION

Recent earthquakes, including the 2008 Wenchuan earthquake (China), 2010 Maule earthquake
(Chile), 2011 Tohoku earthquake (Japan) and 2011 Christchurch earthquake (New Zealand), have
demonstrated that modern buildings generally behave well in terms of life safety. However, because
of significant damage levels, post-earthquake repair of buildings is costly and time consuming, leading
to a long-lasting loss of occupancy and a slow recovery of the community. In order to ensure minimal
disruption in life and business in the urban society, prompt post-earthquake recovery of buildings is a
clear need. One possible solution to achieve this goal is to use easily replaceable components, used as
energy dissipation devices in which damage is concentrated (e.g. shear links, buckling restraint braces,
etc.), while the remainder of the structure is effectively undamaged.
*Correspondence to: JI, Xiaodong, Department of Civil Engineering, Tsinghua University, Beijing 100084, China.
†
E-mail: [email protected]

Copyright © 2016 John Wiley & Sons, Ltd.

 X. JI ET AL.

Coupled wall systems are often used in high-rise buildings because of their superior strength and
stiffness. In such a system, coupling beams distributed along the building height are designed as the
components that undergo inelastic deformation and dissipate seismic energy. While traditional
reinforced concrete (RC) coupling beams, if detailed appropriately, show adequate seismic
performance, once damaged, these components are expensive and time-consuming to repair.
Recently, various types of replaceable coupling beams have been proposed and recognized as an
alternative to traditional RC coupling beams (e.g. Fortney et al. [1], Christopoulos and Montgomery
[2] and Ji et al. [3]).
A new type of replaceable steel coupling beam (RSCB), which comprises of a central ‘fuse’ shear
link connected to steel beam segments at its two ends is used in this study. By appropriately
proportioning the beam segments and shear link, the inelastic deformation can concentrate in the
‘fuse’ shear links during severe earthquakes, while the steel beam segments remain elastic.
Extensive studies [4, 5] have indicated that a short shear link with proper detailing can provide very
stable, ductile and predictable behavior under cyclic shear loading. Recent tests by Ji et al. [3] have
further demonstrated that the damaged link can be readily replaced by using specialized connections
between the link and beam segments. Nevertheless, the seismic performance of a coupled wall
system using the novel RSCBs has yet to be estimated.
The objective of this paper is to assess the seismic behavior, expected damage and reparability of a
hybrid coupled wall (HCW) system consisting of RC wall piers and RSCBs. To illustrate the superior
performance and benefit in reparability for the innovative HCW system, it is compared against a
commonly used reinforced concrete wall (RCW) with conventional RC coupling beams. To this
end, Chapter 2 describes the design of coupled wall systems for assessment and the development of
nonlinear numerical models for all structural components in the HCW and the RCW systems.
Chapter 3 presents the intensities of ground shaking considered in the assessment and associated
ground motion records selected for use in the nonlinear dynamic analysis. Chapter 3 also
summarizes global and component responses of the HCW and the RCW systems subjected to
various levels of earthquakes. Chapter 4 estimates the possible damage and reparability based on the
seismic demands. The conclusions of this study are summarized in Chapter 5. Throughout this
paper, the term HCW refers to a coupled wall consisting of RC wall piers and RSCBs, whereas the
term RCW refers to a coupled wall with RC wall piers and RC coupling beams.

2. NONLINEAR NUMERICAL MODEL

2.1. Prototype structure and coupled wall system

2.1.1. Prototype structure. The prototype structure is an 11-story office building located in Beijing,
adopting a RC frame-shear wall interacting system. The total height of the structure is 48.5 m, and
the plan dimension is 48.6 m × 14.4 m. A representative floor plan is shown in Figure 1. The

Figure 1. Plan view of prototype structure (units: mm).

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

coupling beams in the transverse direction were designed as RSCBs, while the coupling beams in the
longitudinal direction were designed as conventional RC coupling beams. The dead load of each story
including the self-weight of the floor slabs and the superimposed dead load varies from 5.5 kN/m2 to
6.5 kN/m2. The live load is 2.5 kN/m2.
The structure is designed according to the Chinese code for seismic design of buildings (GB 50011-
2010) [6] and Chinese technical specification for concrete structures of tall buildings (JGJ 3-2010) [7].
Per the Chinese code requirements, the structure is designed to satisfy the strength demand of the
service level earthquake (SLE, with a probability of exceedance of 63% in 50 years), which has a
peak ground acceleration of 0.07 g. Linear response spectrum analysis of a three-dimensional
structural model is performed to determine the design forces of structural components and the
deformation of the structure. In this analysis, a damping ratio of 5% is assumed for all modes. For
the RC coupling beams, their effective flexural stiffness EcIeff is taken as 15% of the stiffness value
EcIg based on gross section properties, to account for concrete cracking and slip and extension of
the flexural reinforcement at the beam–wall interface, as recommended by Naish et al. [8]. The
stiffness of the RSCBs is determined based on the gross section properties of shear links and beam
segments. In accordance with GB 50011-2010, the elastic stiffness EcIg is used for the RC wall piers
and columns, as their deformations are small under SLE. The stiffness of RC frame beams is taken
as 1.5 EcIg (for exterior beams) or 2.0 EcIg (for interior beams) as recommended by GB 50011-
2010, to account for the increased stiffness contributed by above RC slabs. Under SLE, the Chinese
code limits the interstory drift ratio to 1/800 for RC frame-wall structures. This strict drift limit
results in the design of relatively stiff buildings. The first three natural periods of the prototype
structure are 1.57 s, 1.52 s and 1.32 s, corresponding to the vibration modes of translation in the
transverse direction, translation in the longitudinal direction and the torsional mode, respectively.
Under the SLE, the base shear coefficient of the structure is 5.3% and 5.5% in the transverse and
longitudinal directions, respectively.
For a RC frame-wall interacting system, in general, seismic damage to RC frames is slighter than
that of RC walls, because the walls often carry a dominated portion of the base shear force and
overturning moment induced by seismic action and RC frames yield at a much larger drift than the
RC walls. Nonlinear dynamic analysis of the prototype structure indicates that the RC frames
sustain minor damage even under the maximum considered earthquakes (MCE), while seismic
damage is concentrated within the coupled walls. This paper thus considers the coupled walls only
for detailed analysis and performance assessment. The HCW on the left side of the prototype
structure is selected for this study. Figure 2 shows the geometry and detailing of the selected HCW.
Gravity loads are applied to the HCW according to its tributary area such that the gravity demands
on the isolated HCW model and the three-dimensional structural model are consistent. The seismic
masses in the isolated HCW are scaled such that the HCW has similar dynamic characteristics with
the prototype structure in the transverse direction, as shown in Table I. Note that the two-
dimensional analysis of the selected coupled wall does not include the additional seismic force
induced by torsional effects in the prototype structure.
An RCW with conventional RC coupling beams is also designed for comparison against the HCW
with RSCBs. The RC coupling beams are designed to have nearly identical nominal shear strengths
and effective stiffness as the RSCBs. The wall piers of the RCW are exactly the same as those of
the HCW. The same gravity loads and masses are applied to the RCW.

2.1.2. Design of wall piers and coupling beams. When a coupled wall is subjected to lateral loads, the
overturning moment is resisted by moment reactions developed at the base of the wall piers and
coupling action induced by the coupling beams. Coupling ratio (CR) is defined as the proportion of
overturning moment resisted by coupling action. In this paper, CR is calculated when all the
coupling beams and wall piers yield. A rational amount of CR should be considered in the design of
coupled walls. Harries [9] proposed a practical upper limit of 66% for the CR of coupled wall with
steel coupling beams. El-Tawil and Kuenzli [10] recommend that the CR ranges from 30% to 45%
for an efficient design. In this research, both the HCW and the RCW are designed to have a CR of 43%.
Figure 2 shows the dimensions and reinforcement layouts of the wall piers. The two boundary
columns of prototype walls (as see in Figure 1) are ignored for simplicity of analysis, which does

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 X. JI ET AL.

Figure 2. Wall dimensions and reinforcement details (dimensions are in mm).

not affect the comparison of the two systems. C45 concrete (nominal axial compressive strength
fck = 29.6 MPa) and HRB400 rebars (nominal yield strength fy = 400 MPa) are adopted for the wall
piers. The wall’s boundary elements and reinforcement are designed to satisfy strength demand
under SLE and the requirement of details specified by the GB 50011-2010 provisions. Note that,
longitudinal reinforcement ratio in boundary elements is governed by the minimum code
requirement, which results in flexural strength of the wall piers approximately twice the SLE
demands.
The RSCBs shown in Figure 2 consist of a ‘fuse’ shear link at the mid-span and two steel beam
segments, which connect the ends of the shear link to the walls. Both the shear link and the beam
segments adopt built-up I-shaped steel sections. Their cross-sectional dimensions are summarized in
Table II. The strength of RSCBs is governed by the link strength, and it is designed to satisfy the
strength demand under SLE. The link flanges are made of Q345 steel (fy = 345 MPa), and the link
webs of Q235 steel (fy = 235 MPa). Use of hybrid sections with lower yielding strength steel in web
is to promote early yielding in shear and to increase the inelastic rotation capacity of links. The
length of the shear link is assigned as 400 mm and the corresponding length ratio e/(Mp/Vp) ranges

Table I. Dynamic properties of coupled wall and prototype structure in transverse direction.

Vibration period (s) Mass participation factor

Mode HCW RCW Prototype HCW RCW Prototype

1 1.60 1.60 1.57 71.9% 71.9% 75.5%

2 0.38 0.38 0.42 16.1% 16.4% 13.5%
3 0.16 0.16 0.19 5.1% 4.7% 4.8%

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

Table II. Design parameters of RSCBs.

hw bf
Component Story l (mm) h (mm) bf (mm) tf (mm) tw (mm) tw 2tf e/(Mp/Vp) Vp (kN)
Shear link 1 400 310 130 10 6 48.3 6.5 0.61 224
2 to 3 400 370 130 10 8 43.8 6.5 0.73 361
4 to 5 400 340 130 10 8 40.0 6.5 0.75 330
6 to 7 400 320 130 10 8 37.5 6.5 0.76 310
8 to 9 400 300 130 10 8 35.0 6.5 0.77 289
10 to 11 400 350 130 10 6 55.0 6.5 0.59 255
Beam segment 1 to 3 600 550 150 14 8 65.3 5.4 — —
4 to11 600 520 150 14 8 61.5 5.4 — —
Notes: l denotes length of the component; h denotes height of the section; hw denotes height of the web; bf denotes
width of the flange; tw denotes thickness of the web; tf denotes thickness of the flange and Vp denotes the plastic
shear strength of the component.

from 0.59 to 0.77. Note that e denotes the length of shear link, and Mp and Vp denote the plastic flexural
strength and shear strength of the link, respectively. Both the width-to-thickness ratios for link flange
and web satisfy the requirement for highly ductile members by the AISC 341-10 [11] and GB 50011-
2010 [6] provisions. According to AISC 341-10 [11], a shear link yields in shear if the length ratio is
smaller than 1.6, and the plastic shear strength of the link is given by Vn = 0.6fy,wAw, where fy,w denotes
the yield strength of link web steel, and Aw denotes the cross-sectional area of the link web. To ensure
that the beam segments remain elastic when the shear link is fully yielded and strain-hardened, their
strength is designed to exceed the strength demand corresponding to the overstrength of the shear
link. The overstrength factor Ω of the shear link with a length ratio less than 1.0 is taken as 1.9 as
suggested by Ji et al. [5]. Both the flanges and webs of steel beam segments are made of Q345 steel.
The shear link is connected to the beam segments using the end-plate connection with high-strength
bolts and shear keys, as shown in Figure 2. The link-to-beam connection is designed such that the shear
keys transfer the shear force and the high-strength bolts resist the bending moment. The strength of the
connection is designed to exceed the overstrength capacity of the shear link. Ji et al. [3] demonstrated
that this specialized link-to-beam connection can ensure favorable seismic behavior of the RSCBs and
easy replacement of the shear links after being damaged in severe earthquakes. The embedded beam–
wall connection design complies with the requirements for steel coupling beams in AISC 341-10 [11].
The connection strength is also designed to exceed the overstrength capacity of the shear link in order
to ensure that the joint would remain elastic even under severe earthquakes. The embedment length of
coupling beams is determined using the design formulas specified in AISC 341-10 [11] to satisfy the
strength of beam–wall connection.
In China, RC coupling beams are usually conventionally, rather than diagonally, reinforced. There-
fore, conventional RC coupling beams are used in the RCW. The thicknesses of the RC coupling
beams are the same as the thicknesses of the connected wall piers. The depth of the RC coupling beams
is determined to provide the similar effective stiffness with the RSCBs. Table III summarizes the
design parameters of the RC coupling beams. The Chinese code GB 50011-2010 recommends that

Table III. Design parameters of RC coupling beams.

qffiffiffiffiffi
Story b (mm) h (mm) L/h Longitudinal reinforcement Hoops (mm) Vn (kN) Vn/Vcs V n =bh f ′c
1 400 600 2.67 3D16 + 1D18 D10@100 223 0.42 1.9
2 to 3 400 600 2.67 2D20 + 2D22 D10@100 361 0.68 3.0
4 to 5 350 600 2.67 5D18 D10@100 331 0.66 3.1
6 to 7 350 600 2.67 6D16 D10@100 314 0.62 3.0
8 to 9 350 600 2.67 3D16 + 2D18 D10@100 289 0.39 2.8
10 to 11 350 600 2.67 3D14 + 2D18 D10@100 252 0.29 2.4
Notes: b and h denote width and depth of the section, respectively; L denotes length of the coupling beam; Vn and
Vcs denote the nominal shear strength of the RC coupling beams and their shear strength capacity calculated per
GB 50010-2010; fc′ denotes the axial compressive strength of the concrete in psi; the top and bottom longitudinal
rebars in the coupling beams are symmetrical, and this table presents the rebars in either side.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 X. JI ET AL.

RC coupling beams are designed to be governed by flexure to ensure adequate ductility, and therefore
those beams in this study are designed to satisfy the ‘strong shear and weak bending mechanism’. The
longitudinal reinforcement of the RC coupling beams is designed such that their nominal shear
strength, Vn = Mn/L, is comparable to that of the RSCBs, where Mn denotes the yield flexural strength
of RC coupling beams and L denotes the length of coupling beams.

2.2. RC wall pier modeling

The multi-layer shell element is adopted for modeling of the RC wall piers in OpenSees [12]. In the
multi-layer shell element, the concrete cover and inside concrete are represented by a number of
concrete layers, and the distributed reinforcements are represented by the smeared rebar layers in
vertical and horizontal directions, respectively, as illustrated in Figure 3. The longitudinal rebars in
the boundary elements are modeled with truss elements, and they are coupled with the surrounding
shell elements. Lu et al. [13] have implemented the multi-layer shell element in the computation
platform OpenSees for modeling RC walls. Ji et al. [14] have validated the models by comparison
with test results.
In the modeling, the concrete is assumed to behave in a plane-stress manner, and the cracking of the
concrete is modeled by the fixed smeared crack approach. Well-calibrated models are used to represent
the uniaxial compressive stress–strain relationship of the concrete. The concrete cover is represented by
the Kent–Park model [15] and the stirrup-confined concrete by the Saatcioglu–Razvi model [16]. The
uniaxial tensile stress–strain relationship of concrete is represented by a bilinear curve which takes into
account the tension softening. The Giuffré–Menegotto–Pinto material model [17] is adopted to
represent the uniaxial stress–strain relationship of the steel reinforcement. More details of parameter
setting of the models can be found in [14].

2.3. RSCB modeling

2.3.1. Analytical RSCB model. Figure 4 shows the simplified numerical model for the RSCB. The
shear link is simulated by a nonlinear link element. The mechanical behavior on each degree of
freedom is modeled by a user-defined spring. In this case, the shear link is designed to yield in shear.
Therefore the axial spring and flexural spring are elastic, while the shear spring is nonlinear. The
properties of the axial, shear and flexural springs of the link element are listed in Table IV.
Experiments [4, 5] report that shear links present similar hysteretic performance with steel under
cyclic loading. Therefore, the hysteretic behavior of the nonlinear shear spring is characterized by a
uniaxial material model based on Giuffré–Menegotto–Pinto hysteretic model [17]. The parameters R0,
cR1 and cR2 that reflect the Bauschinger effect, the parameter b that reflects kinematic hardening
effect and the parameters a1 through a4 that represent isotropic hardening effect are calibrated against
experiment data. The values of these parameters are listed in Table IV as well. It should be mentioned
that axial deformation and associated axial force develop in shear links when they experience large
inelastic shear deformation [3, 5], which may cause redistribution of shear forces in the wall piers that
are connected to the RSCBs. However, the mechanism of the axial forces in shear links has yet to be
clearly understood. In addition, the presence of RC slabs may suppress the influence of axial forces in
shear links. Therefore, the study does not include the possible influence of axial forces in shear links.

Figure 3. Sketch of multi-layer element for RC wall [13].

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

Figure 4. Nonlinear model for RSCB.

The beam segment is designed to remain elastic under seismic action, and thereby it is
modeled by an elastic beam element in OpenSees. This beam element does not include shear
deformation, but shear deformation of short-span beam segments is not negligible. Therefore, a
zero-length shear spring element is set between the beam element and the link element, for
which the stiffness equals to the shear stiffness of beam segment, that is k = GAb,w / lb, where
Ab,w denotes the sectional area of the web of beam segment and lb denotes the length of the
beam segment.

2.3.2. Verification of the RSCB model. Both shear link experiments and RSCB experiments by Ji
et al. [3, 5] are modeled and analyzed cyclically to validate the numerical model developed for
the RSCB. The analytical results compare well to the experimental results. Figure 5 shows the
analytical and experimental force–displacement relationship of a shear link specimen and a
RSCB specimen. More details of verification studies are reported in reference [18].

2.4. RC coupling beam modeling

RC coupling beams are modeled with a nonlinear link element. The skeleton of the force–displacement
relationship of RC coupling beams is defined following the ASCE/SEI 41-13 provisions [19], as shown
in Figure 6(a). θ denotes beam rotation, and V/Vn denotes normalized shear force of the coupling beam.
The effective bending stiffness EcIeff is taken as 15% of the stiffness value EcIg based on gross section
properties, as recommended by Naish et al. [8]. The ultimate strength (point C of the skeleton curve in
Figure 6(a)) is taken as 1.15 times the yield flexural strength as reported in Kwan and Zhao [20].
qffiffiffiffiffi
Values of a, b and c depend on the failure mode and the nominal shear stress level V n =bh f ′c of
qffiffiffiffiffi
the RC coupling beams. Given the flexural failure mode and the values of V n =bh f ′c listed in
Table III, values a, b and c are calculated as 0.025 rad, 0.05 rad and 0.75, respectively, according to
the ASCE/SEI 41-13 provisions [19].
The hysteretic behavior, including the pinching effect, of conventional RC coupling beams is related
to the following three design parameters: the span-to-depth ratio, flexural-to-shear strength ratio Vn/Vcs

Table IV. Properties of the link element for shear links.

Springs Parameters Formulas or values
Elastic axial spring Axial stiffness Kl,a K l,a = EAl / e
Elastic flexural spring Flexural stiffness Kl,f Kl,f = EIl / e
Nonlinear shear spring Yield force Vl,y Vl,y = 0.6fy,wAl,w
Elastic stiffness Kl,s Kl,s = [e3 / (12EIl) + e / (GAl,w)] 1

Parameters that control the transition R0 = 18.5;

from elastic to plastic branch cR1 = 0.9;
cR2 = 0.1
Kinematic hardening ratio b = 0.003
Isotropic hardening parameter a1 = a3 = 0.14; a2 = a4 = 1.0
Notes: E and G denote Young’s modulus and shear modulus of the steel, respectively; e denotes length of the shear
link; Al denotes the cross-sectional area of the shear link; Il denotes moment of inertia of the link section; Al,w
denotes the cross-sectional area of the link web and fy,w denotes the yield strength of the link web.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 X. JI ET AL.

Figure 5. Comparison of experimental and analytical results of shear link and RSCB.
qffiffiffiffiffi
and nominal shear stress level V n =bh f ′c [21]. The test data of experimental specimens by Naish et al.
[8] and Kwan et al. [20], which are similar to the RC coupling beams to be analyzed, are selected to
calibrate the parameters of the hysteretic model that controls the unloading and reloading stiffness,
pinching effect, etc. In general, the analytical results can track well the hysteretic curves obtained
from the tests, of which one example is shown in Figure 6(b).

2.5. Beam–wall connection and slabs

The rigid connection between coupling beams and wall piers is modeled with additional rigid beam
elements which transfer the forces developed at the coupling beam end to all wall shell elements at
the coupling beam height. For the HCW, in fact, opening and closure of the gap between the
embedded steel beams and RC wall piers may lead to additional nonlinearity. Pushover analysis
with a sophisticated model that includes contact springs to reflect this nonlinearity at beam–wall
connection indicates that the overall effect of gap opening and closure on the system behavior is
small. Similar observation was found in Hassan and El-Tawil [22]. Considering the computational
cost, the possible nonlinearity induced by gap opening and closure in beam–wall connection of the
HCW is not considered in this study.
Ji et al. [23] recommended that the RC slabs are elevated from the RSCBs and reported that
the elevated slab has limited influence on the initial stiffness, shear strength and hysteretic
performance of the RSCB. Naish et al. [8] reported that the RC slabs could increase the
nominal strength of RC coupling beams by 17%, while the slabs did not influence the
ductility and hysteretic performance of RC coupling beams. As the coupled wall analyzed in
this paper is an exterior wall, the effect of the RC slabs would be further decreased because
the slabs are on one side of the RSCBs only. Therefore, the effect of RC slabs on coupling
beams is ignored in the analysis.

Figure 6. RC coupling beam modeling.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

3. NONLINEAR RESPONSE TIME HISTORY ANALYSIS

3.1. Ground motions and dynamic analysis parameters

3.1.1. Intensities of ground motion shaking. Four intensities of ground motion shaking are considered
for the performance assessment. These ground motion intensities are specified in the seismic ground
motion parameter zonation map of China (GB 18306-2015) [24]: SLE, design basis earthquake
(DBE), MCE and very rare earthquake (VRE). The probabilities of exceedance and return periods
for these intensities of shaking and the associated design response spectra are shown in Figure 7.
The amplitudes of the SLE, MCE and VRE design spectra are 0.35, 2 and 3 times that of the DBE de-
sign spectrum, respectively. According to GB 18306-2015, the peak ground acceleration of DBE for
the site of the prototype building is 0.2 g. The site of the prototype building falls into Site Class III, with
an average shear wave velocity in the top 30 m of soil, VS30, between 150 m/s and 250 m/s. The char-
acteristic period of seismic spectra, Tg, is 0.45 s.

3.1.2. Selection and scaling of ground motions. Ground motion records are selected such that their
spectral shape is similar to the target spectra considered in the assessment. The target spectrum for
record selection is the DBE design response spectrum. In order to select ground motions to match the
target spectrum, the NGA West 2 Ground Motion Database [25] is used for record selection. Record
characteristics with magnitudes greater than 6, average shear wave velocity consistent with Site Class
III and no restriction on fault type and fault distance are used to search the database. The records are
linearly scaled to match the target spectrum, and they are selected such that the computed mean
squared error (MSE) of their response spectra (assuming 5% damping) of the suite average is
minimized with respect to the target spectrum over the period range of interest. As recommended by
Qu et al. [26], the period range of interest is selected to span from 0.1 s to Tg, the characteristic site
period, and from T1 minus 0.2 s to 2 T1, where T1 is the fundamental period of the structure. The
selected ground motion records, individual record spectra, the mean spectrum of the selected records
and the target spectrum plotted against the period range of interest are shown in Figure 7. The suite
contains seven ground motion records, which are sufficient when the fit between the ground motion
spectra and the target spectra is good [27] and the goal of the assessment is to estimate mean values
of response [28].

3.1.3. Analysis assumption and dynamic parameters. The selected ground motions are input at the
base of the coupled wall models, which are assumed to be fixed at their base. Following the
recommendations in GB 50011-2010 [6], a damping ratio of 5% is assumed in the analysis,
implemented using the Rayleigh damping model for the first and third vibration modes of the
coupled walls.

Figure 7. Selected ground motion records and response spectra.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 X. JI ET AL.

3.2. Dynamic analysis results

3.2.1. Yield mechanism. Figure 8 compares yield mechanisms of the HCW and the RCW. The
coupling beams are designed to satisfy strength demands under SLE. Therefore, very few RSCBs
and RC coupling beams yield when subjected to the SLE records, while all RSCBs and RC
coupling beams yield when subjected to the DBE, MCE and VRE records. The minimum
requirement of Chinese code on boundary longitudinal reinforcement results in a flexural strength of
these wall piers twice their SLE strength demand. As a result, for both the HCW and the RCW, no
wall piers yield when subjected to all SLE motions as well as most of DBE motions. Both wall
piers yield under combined axial forces and bending moments when subjected to most of the MCE
motions and all VRE motions.
3.2.2. Strength demands. Figure 9 provides a comparison between the shear forces of coupling beams
distributed along the height of the HCW and the RCW. Both mean values and 84th percentile values
(i.e., mean value plus one standard deviation) of the responses are provided in this figure. When
subjected to SLE records, the mean values of beam shear forces are below the nominal shear
strength of the coupling beams. When subjected to DBE records, all coupling beams yield, slight
overstrength is observed in the RC coupling beams, while RSCBs develop considerable
overstrength. When subjected to MCE and VRE, the RSCBs show significantly greater overstrength
than the RC coupling beams.
Figure 10 shows the hysteretic responses of coupling beams at the 11th story when subjected to the
Codroipo motion at VRE. Besides the large overstrength factor, the RSCB also shows full hysteretic
loops and large energy dissipation, while the RC coupling beam shows hysteresis properties with

Figure 8. Yield mechanism of the HCW and the RCW.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

Figure 9. Beam shear force demands for the HCW and the RCW.

evident pinching effects and stiffness degradation. Therefore, the coupling action of the RCW is
reduced under severe earthquakes, which would lead to the degradation of entire lateral stiffness and
larger lateral drifts than the HCW.
Figure 11 shows the variation of base shear versus axial force in the right wall pier when subjected
to the Codroipo motion at MCE and VRE. The shear strength of the wall is calculated following GB
50010-2010 [6]. For both the HCW and the RCW, the base shear is far less than shear strength of
the wall pier, and no shear failure occurs. Figure 12 shows the variation of moment versus axial
force at the base of the right wall pier when subjected to the Codroipo motion at MCE and VRE.
The wall pier of the HCW bears larger variation of axial force, while the wall pier of the RCW
bears larger variation of bending moments. The larger variation of axial force in the wall pier of the
HCW than that of the RCW is mainly because of the stable hysteretic responses and large
overstrength of the RSCBs. In the HCW, the stable responses and overstrength of RSCBs post
yielding of shear links increase the coupling action of the system which in turn increases the tensile
and compressive forces acted on the wall piers, whereas in the RCW, stiffness degradation of the
RC coupling beams decreases the coupling action and thus the wall piers have to carry increased
bending moments.

3.2.3. Deformation. Figure 13 compares the peak transient interstory drift distribution of the HCW
and the RCW. The maximum interstory drift ratio appears at the top stories for both systems. When
subjected to SLE records, the interstory drifts are almost identical for both systems with maximum
values around 1/800, that is the drift limit required by GB 50011-2010 [6]. When subjected to DBE,
MCE and VRE, the maximum interstory drift ratios of the HCW are 16%, 29% and 31% smaller
than those of RCW, respectively. This is attributed to the fact that under large deformation, the

Figure 10. Hysteretic response of a sample RSCB and RC coupling beam at VRE.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 X. JI ET AL.

Figure 11. N–V relationship of wall piers in the HCW and the RCW subjected to Codroipo motion at MCE
and VRE.

Figure 12. N–M relationship of wall piers in the HCW and the RCW subjected to Codroipo motion at MCE
and VRE.

Figure 13. Mean interstory drift ratios of the HCW and the RCW.

RSCBs show stable hysteretic response with large energy dissipation and overstrength, while the RC
coupling beams present little overstrength and significant stiffness degradation.
Figure 14 compares the beam rotation distribution of the HCW and the RCW. When subjected to
SLE records, the beam rotations are nearly identical for both the RSCBs and the RC coupling
beams. When subjected to DBE, MCE and VRE records, the maximum beam rotations of the
RSCBs are 24%, 36% and 42% smaller than those of the RC coupling beams, respectively, because

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

Figure 14. Mean beam rotations of the HCW and the RCW.

of larger energy dissipation and overstrength of the RSCBs. Even under VRE, the maximum chord
rotations of the shear links in the RSCBs are 0.086 rad, which are less than the inelastic rotation
capacity γp = 0.14 rad suggested by Ji et al. [5] for shear links with a length ratio less than 1.0.

4. ASSESSMENT OF SEISMIC PERFORMANCE

This section assesses the expected seismic damage and suggests associated repair methods for both
coupled wall systems based on the estimated responses under four intensities of ground motion
shaking. The results can provide an indicator for enhanced performance and reparability of the new
HCW system against earthquake hazards when compared with the traditional RCW system. The
method for seismic damage assessment refers to a recently developed guideline of FEMA P-58 [27].

4.1. Structural components fragility data

Seismic performance assessment requires basic knowledge of the fragility data of building
components. A component fragility function is a statistical distribution that indicates the conditional
probability of incurring damage at a given value of demand parameter, which is typically assumed
to be lognormal distribution. Component fragility functions contain unique fragilities for each
possible damage state in the component. As described below, fragility functions for the RSCBs are
proposed based on recent experiments conducted by Ji et al. [3, 5]. For RC wall piers and RC
coupling beams, the fragility curves recommended in FEMA P-58 are used as such data is not yet
available in Chinese codes and standards. However, it is noted that the GB 50011-2010 provisions
for conventional RC coupling beams are similar to those in ACI 318-14 [29], and therefore the RC
coupling beams designed in accordance with both codes are expected to have similar fragility
curves. The RC walls designed by GB 50011-2010 appear to have less boundary transverse
reinforcement than the special walls designed by ACI 318-14. Preliminary analysis on a large
volume of slender RC wall test data from China indicates that their fragility curves of DS1 and DS2
(minor to moderate damage) are nearly identical to FEMA P-58 curves, while the median wall drift cor-
responding to DS3 and DS4 (severe damage) is somewhat smaller than the values suggested in FEMA
P-58.
4.1.1. RC walls. FEMA P-58 [27] provides fragility functions for slender RC shear walls (aspect
ratio ≥ 2.0) (Fragilities B1044.091 to B1044.113). Because of its simplicity of calculation and good
correlation to damage observed in the tests, in this study effective wall drifts (EWD) are selected as
the demand parameter in the fragility functions to estimate different damage levels to the walls.
Note that, interstory drift can also be used as the demand parameter for the fragility functions.
However, directly using individual story drift data to estimate wall damage at every story is
inaccurate, as often times in shear wall systems, the large drifts obtained at higher stories are a result
of rigid body rotation of the top stories induced by damage to the wall near its base. Table V

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 X. JI ET AL.

Table V. Summary of slender RC wall fragility data. Source: FEMA P-58 [27].
Damage Fragility
state data Damage description Repair method Extent of damage

DS1 Median: Formation of initial cracking. Cosmetic repair of surface finish

0.0012 (EWD) Crack width <0.02 in to maintain first resistance and
Dispersion: (0.5 mm). prevent water infiltration.
0.76

DS2 Median: Spalling of concrete cover. Epoxy injection of cracks and

0.0093 (EWD) Appearance of vertical patching of spalled concrete.
Dispersion: cracks. Crack width >1/16
0.48 in (1.6 mm).

DS3 Median: Exposure of longitudinal Shoring of the wall and removing

0.013 (EWD) wall reinforcement. and replacing all concrete in
Dispersion: the damaged region.
0.34

DS4 Median: Crushing of concrete. Replacement of the wall or

0.019 (EWD) Buckling or fracture of concrete jacketing.
Dispersion: reinforcement.
0.44

provides a summary of the median values and dispersions associated with the slender shear wall
fragility, damage state descriptions and associated repair measures, as well as a visual illustration of
the damage level for each damage state.
4.1.2. RC coupling beams. FEMA P-58 [27] provides fragility functions for RC coupling beams for a
variety of aspect ratios, thicknesses and reinforcement layouts (diagonally and conventionally
reinforced). Beam rotation is the demand parameter utilized in the fragility functions. In the
structure considered in this study, RC coupling beams are conventionally reinforced, with a
thickness of 350–400 mm and an aspect ratio of 2.4. Consistent with those beam properties, FEMA
P-58’s fragility B1042.002b is selected to best represent these structural components. Table VI
provides a summary of the median values and dispersions associated with the RC coupling beam
fragility, damage state descriptions and associated repair measures, as well as a visual illustration of
the damage level for each damage state.
The fragilities here presented make no consideration of damage to the slabs above the RC coupling
beams. While researchers have studied the impact of the slab on the load–deformation response of RC
coupling beams [8], these findings have not been incorporated into the fragility functions available in
the literature.
4.1.3. RSCBs. Similar as eccentrically braced frames (EBFs), link rotation is taken as the demand
parameter utilized in the fragility functions of RSCBs. Three distinct damage states are observed in
the specimens of past tests [23]. Figure 15 illustrates the experimental and fitted fragilities proposed
for the RSCBs. The method of maximum likelihood is used to develop the fragility curves for the

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

Table VI. Summary of RC coupling beam fragility data. Source: Naish et al. [8].
Damage Fragility
state data Damage description Repair method Extent of damage

DS1 Median: Appearance of residual Epoxy injection of

0.014 (rad) cracks no greater than cracks.
Dispersion: 1/16 in (1.6 mm).
0.21 Limited flexural
cracking.

S2 Median: Appearance of residual Epoxy injection of

0.026 (rad) cracks greater than cracks and replacement
Dispersion: 1/8 in (3.2 mm) and of spalled concrete.
0.33 minor spalling of
concrete.

DS3 Median: Significant strength Removing and replacing

0.041 (rad) degradation. Buckling damaged concrete and
Dispersion: and fracture of reinforcement.
0.75 reinforcement.
Crushing of concrete.

RSCBs. This method finds the parameters such that the resulting distribution has the highest likelihood
of having produced the observed data [27]. Table VII provides a summary of the median values and
dispersions associated with the RSCB fragility, damage state descriptions and associated repair
measures, as well as a visual illustration of the damage level for each damage state. It is notable that
the yielding of link is regarded as DS0 (no damage), as the structural repair is not necessary and the
shear strength and stiffness of the component are not affected. Note that test data associated with
DS1 is limited to a handful of samples with different connection details of the RC slab to the
RSCB. Median rotation demands at which the slab requires replacement are obtained from the
limited dataset. The estimated mean is close to the values reported in FEMA P-58 [27] for slab
replacement above shear links in EBFs.

4.2. Performance assessment

The seismic performance of the HCW and the RCW are assessed based on the residual deformation
acceptance criteria and the fragility functions for the structural components. The comparison is
conducted based on mean estimates of response under at DBE, MCE and VRE. An evaluation at

Figure 15. Fragility functions for RSCBs

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 X. JI ET AL.

Table VII. Summary of RSCB fragility data.

Damage Fragility
state data Damage description Repair method Extent of damage

DS1 Median: Damage to the slab above Replacement of the

0.05 (rad) link the RSCB, with no damage portion of the slab
rotation to the RSCB itself. above the RSCB.
Dispersion:
0.3

DS2 Median: Buckling of the web or Heat straightening

0.09 (rad) link flanges in the shear link. buckled elements or
rotation replacement of shear
Dispersion: link.
0.19

DS3 Median: Facture of the web in the Replacement of the

0.11 (rad) link shear link or fracture of shear link.
rotation the link flange-to-end plate
Dispersion: welds.
0.15

SLE is not considered as all structural components remain elastic at this intensity of shaking, i.e. there
is no damage. Residual drifts are an important indicator of reparability given that in the presence of
large residual drifts, a structure may be deemed irreparable because of the technical and economic
feasibility of repair. In accordance with FEMA P-58 [27], the median residual drifts of 0.2% in
buildings imply that no structural realignment is necessary (although repairs may be required for
nonstructural components). Table VIII summarizes the mean values of the maximum residual drifts
of all stories in both the HCW and the RCW system. Both system comply with the
recommendations of FEMA P-58 [27] to ensure no structural realignment is required.
Regarding the performance of structural components, damage to the RC wall piers is very limited in
both the HCW and the RCW, even under VRE events. In this study, damage to the shear walls is
concentrated in the lower two stories. Therefore, the effective wall drift was computed using an
effective wall height equal to the first two stories. Table VIII summarizes the effective mean wall
drifts in the HCW and the RCW. Figure 16 illustrates the probabilities of observing different
damage levels in the wall piers under DBE, MCE and VRE where DS0 denotes no damage. At
DBE and MCE, expected damage in both systems is effectively identical, with large probabilities of
observing DS1 (55–60% and 88–90%, respectively) requiring surface finish repairs. Under VRE,
expected damage between the two systems is still comparable, although greater probability of
observing DS2 is observed for the RCW (14%) than the HCW (8%), which would require epoxy
injecting of cracks and patching spalled concrete. Overall, expected damage is very low because of
the high structural stiffness required by Chinese codes.

Table VIII. Mean residual story drifts and effective wall drifts for the HCW and the RCW.
Earthquake level Residual story drifts (%) Effective wall drifts (%)
HCW RCW HCW RCW
DBE 0.013 0.007 0.13 0.15
MCE 0.025 0.013 0.30 0.37
VRE 0.070 0.039 0.48 0.56

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

Figure 16. Expected damage in the wall piers of the HCW and the RCW.

Figure 17 illustrates the probability of observing different damage states for the RSCBs and the RC
coupling beams along the building height, where DS0 denotes no damage. Under DBE, though
yielded, the beams are effectively undamaged. Under MCE, the RSCBs show high probabilities of
no damage (54–99% probability of DS0), whereas the RC coupling beams illustrate significant
probabilities of limited damage, particularly in the upper stories (66–72% probability of DS1) and
probabilities of moderate damage (DS2) and severe damage (DS3) in the order of 16–21% and 10–
13%, respectively, in the top 5 stories. Under VRE, RSCBs have high probabilities of incurring
DS1, with values of up to 78%, and probabilities of moderate damage (DS2) in the order to 21–36%
in the top 5 stories. Ji et al. [3] report that with residual beam rotation less than 0.45%, damaged
links can be readily replaced if the RSCBs adopt specialized link-to-beam connections. The mean
value of the maximum residual beam rotation under VRE is 0.23%, and therefore the shear link can
be easily replaced if necessary. On the other hand, the RC coupling beams, particularly in the upper
stories, illustrate significant probability of moderate to severe damage (38–42% and 45–55%
probabilities of DS2 and DS3, respectively). These results illustrate how the RSCBs have enhanced

Figure 17. Expected damage in the RSCBs and the RC coupling beams.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 X. JI ET AL.

performance over the RC coupling beams under the strong earthquake motions. Under MCE and VRE,
the RSCB repairs are limited to replacement of the portion of the slabs above RSCBs and possible
replacement of shear links, whereas RC coupling beam repairs would require removing and
replacing damaged concrete and reinforcement.

5. CONCLUSIONS

This study assesses the seismic performance of a HCW system with RSCBs at four intensities of
ground motion shaking as defined in the Chinese code: SLE, DBE, MCE and VRE. The
performance of the HCW system is benchmarked against the traditional RCW with RC coupling
beams. Nonlinear numerical models are developed in OpenSees for a representative wall elevation
in a prototype 11-story building located in Beijing and designed per modern Chinese standards.
The results indicate that behavior of the two systems is consistent at SLE, where both systems
remain elastic and DBE, where all coupling beams yielded, yet the walls remained elastic. At MCE
and VRE, because of significant stiffness degradation of RC coupling beams, the coupling action of
the RCW is reduced and a relatively larger portion of overturning moment is resisted by the wall
piers, resulting in larger drifts in RCW than in the HCW. The maximum interstory drifts are 29%
smaller at MCE and 31% smaller at VRE in the HCW than in the RCW. At MCE and VRE, the
maximum beam rotations of the RSCBs are 36 to 42% smaller than those of the RC coupling beams.
Expected damage to the walls, even under extreme events, is limited to cracks and slight spalling of
concrete, which only require surface repairs for both the HCW and the RCW. Whereas no damage is
expected in the beams under SLE and DBE, some damage is expected in both the RSCBs and the RC
coupling beams under MCE and VRE. At these intensities, expected damage to the RSCBs is limited to
the slab above the RSCB and possible web buckling of the shear link, whereas damage to the
conventional RC coupling beam would result in cracks, spalling and crushing of concrete, buckling
and fracture of reinforcement, particularly in coupling beams in the upper stories. Overall, under
extreme earthquake events, the HCW with RSCBs illustrates enhanced performance over the
conventional RCW with RC coupling beams, as it can quickly recover by easily replacing the
damaged links as opposed to conducting costly and time-consuming repairs of RC coupling beams.

ACKNOWLEDGEMENTS
The work presented in this paper was sponsored by the International Science and Technology Cooperation
Program of China (Grant No. 2014DFA70950), the National Natural Science Foundation of China (Grants
No. 51261120377) and by Tsinghua University Initiative Scientific Research Program (Grant No.
2012THZ02-1). The authors wish to express their sincere gratitude to the sponsors.

REFERENCES
1. Fortney PJ, Shahrooz BM, Rassati GA. Large-scale testing of a replaceable “fuse” steel coupling beam. Journal of
Structural Engineering 2007; 133(12):1801–1807.
2. Christopoulos C, Montgomery MS. Viscoelastic coupling dampers (VCDs) for enhanced wind and seismic perfor-
mance of high-rise buildings. Earthquake Engineering & Structural Dynamics 2013; 42(15):2217–2233.
3. Ji X, Wang Y, Ma Q, Taichiro O. Cyclic behavior of replaceable steel coupling beams. Journal of Structural Engi-
neering, ASCE 2016. DOI: 10.1061/(ASCE)ST.1943-541X.0001661.
4. Okazaki T, Arce G, Ryu HC, Engelhardt MD. Experimental study of local buckling, overstrength, and fracture of
links in eccentrically braced frames. Journal of Structural Engineering, ASCE 2005; 131(10):1526–1535.
5. Ji X, Wang Y, Ma Q, Okazaki T. Cyclic behavior of very short steel shear links. Journal of Structural Engineering,
ASCE 2016; 142(2 04015114).
6. CMC (Ministry of Construction). Code for Seismic Design of Buildings (GB 50011-2010). China Architecture &
Building Press: Beijing, 2010.
7. CMC (Ministry of Construction). Technical Specification for Concrete Structures of Tall Building (JGJ 3-2010).
China Ministry of Construction: Beijing, 2010 (in Chinese).
8. Naish D, Fry A, Klemencic R, Wallace J. Reinforced concrete coupling beams—Part I: testing. ACI Structural
Journal 2013; 110(6):1057.
9. Harries KA. Ductility and deformability of coupling beams in reinforced concrete coupled walls. Earthquake Spectra
2001; 17(3):457–478.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe
 SEISMIC ASSESSMENT OF A HCW SYSTEM WITH RSCBS VS RC COUPLING BEAMS

10. El-Tawil S, Kuenzli CM. Pushover of hybrid coupled walls. II: Analysis and behavior. Journal of Structural Engi-
neering, ASCE 2002; 128(10):1282–1289.
11. AISC. Seismic Provisions for Structural Steel Buildings (ANSI/AISC 341-10). American Institute of Steel Construc-
tion: Chicago, 2010.
12. OpenSEES. Open System for Earthquake Engineering Simulation. Pacific Earthquake Engineering Research Center.
(Available from http://opensees.berkeley.edu/) [accessed on January 15, 2015].
13. Lu X, Xie L, Guan H, Huang Y, Lu X. A shear wall element for nonlinear seismic analysis of super-tall buildings
using OpenSees. Finite Elements in Analysis and Design 2015; 98:14–25.
14. Ji X, Sun Y, Qian J, Lu X. Seismic behavior and modeling of steel reinforced concrete (SRC) walls. Earthquake En-
gineering & Structural Dynamics 2015; 44(6):955–972.
15. Kent DC, Park R. Flexural members with confined concrete. Journal of the Structural Division 1971; 97
(7):1969–1990.
16. Saatcioglu M, Razvi SR. Strength and ductility of confined concrete. Journal of Structural Engineering, ASCE 1992;
118(6):1590–1607.
17. OpenSees Wiki. Steel02 Material—Giuffré–Menegotto–PintoModel with Isotropic Strain Hardening. http://
opensees.berkeley.edu/wiki/index.php/Steel02_Material_--_Giuffr%C3%A9-Menegotto-
Pinto_Model_with_Isotropic_Strain_Hardening [Accessed on January 15, 2015].
18. Sun Y. Master’s thesis. Study on Seismic Behavior of Hybrid Coupled Wall with Replaceable Steel Coupling Beams.
Tsinghua University: Beijing, 2015 (in Chinese).
19. ASCE/SEI. Seismic Evaluation and Retrofit of Existing Buildings (ASCE/SEI 41-13). Structural Engineering Insti-
tute: Reston, VA, 2014.
20. Kwan AKH, Zhao ZZ. Cyclic behaviour of deep reinforced concrete coupling beams. Structures and Buildings 2002;
152(3):283–293.
21. Gong B, Fang E. Behavior of reinforced concrete coupling beams between shear walls under cyclic loading. Journal
of Building Structures 1988; 01:34–41 (in Chinese).
22. Hassan M, El-Tawil S. Inelastic dynamic behavior of hybrid coupled walls. Journal of Structural Engineering 2004;
130(2):285–296.
23. Ji X, Wang Y, Zhang J, Okazaki T. Seismic behavior and fragility curves of replaceable steel coupling beams with
slabs. Journal of Constructional Steel Research 2016 under review.
24. Standardization Administration of China (SAC). Seismic Ground Motion Parameters Zonation Map of China (GB
18306-2015). Beijing: Standards Press of China (in Chinese).
25. PEER. NGA-West2 Database. Pacific Earthquake Engineering Research Center, Report No. 2013/03. University of
California, Berkeley: Berkeley, CA, 2013.
26. Qu Z, Ye L, Pan P. Comparative study on methods of selecting earthquake ground motions for nonlinear time history
analyses of building structures. China Civil Engineering Journal 2011; 44(7):10–21 (in Chinese).
27. ATC. Seismic Performance Assessment of Buildings Volume 1—Methodology (FEMA P-58-1). Applied Technology
Council: Redwood City, CA, 2012.
28. NIST. Selecting and Scaling Earthquake Ground Motions for Performing Response History Analyses (NIST GCR 11-
917-15). National Institute of Standards and Technology, US Department of Commerce: Washington, D.C., 2011.
29. ACI (American Concrete Institute). Code Requirements for Structural Concrete (ACI 318-14) and Commentary.
American Concrete Institute: Farmington Hills, MI, 2014.

Copyright © 2016 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2016)
DOI: 10.1002/eqe