Capacity Design of Coupled Composite Plate Shear

Wall–Concrete-Filled System
Morgan Broberg, S.M.ASCE 1; Soheil Shafaei, M.ASCE 2; Emre Kizilarslan, S.M.ASCE 3; Jungil Seo 4;
Amit H. Varma, M.ASCE 5; Michel Bruneau, F.ASCE 6; and Ron Klemencic, P.E., S.E., Dist.M.ASCE 7
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Composite plate shear walls–concrete-filled (C-PSW/CF) are a new and innovative lateral force–resisting system intended for
high-rise buildings. High-rise building applications of this system are particularly efficient in the coupled wall configuration, in which the
walls are C-PSW/CF and the coupling beams are concrete filled steel box sections. This paper presents a capacity design principle for the
seismic design of coupled composite plate shear wall–concrete filled (CC-PSW/CF) systems. The capacity design principle implements a
strong wall–weak coupling beam approach, in which flexural yielding occurs in the coupling beams before flexural yielding at the base of
walls. The coupling beams are sized to resist the calculated seismic lateral force level. The composite walls are sized to resist an amplified
seismic lateral force corresponding to the overall plastic mechanism for the structure, while accounting for the capacity-limited forces from the
coupling beams and the coupling action between the walls. The paper summarizes the recommendations and requirements for appropriate
sizing of the composite coupling beams and walls. These recommendations were used along with the capacity design principle to design four
example (8–22-story) structures and evaluate their seismic behavior. The structures were modeled using benchmarked finite-element models
and fiber-based models that accounted for the various limit states, including steel yielding, local buckling, fracture, concrete crushing, confine-
ment, and tension cracking. The numerical models were analyzed for monotonic pushover loading and scaled seismic ground motions. The
structural responses from the nonlinear pushover analysis and the nonlinear time history analyses were in accordance with the capacity limited
design philosophy, thus confirming its efficacy. DOI: 10.1061/(ASCE)ST.1943-541X.0003296. © 2022 American Society of Civil Engineers.
Author keywords: Composite; Steel; Concrete; Seismic; Earthquake engineering; Analysis; Design.

Introduction progresses at a much slower pace than the rest of the structure. With
the rise of modularity and prefabricated solutions, industry innova-
Tall buildings commonly are designed and constructed with rein- tors began looking for solutions to expedite the construction sched-
forced concrete (RC) core walls, steel gravity framing, and light- ule and improve overall project economy. Composite plate shear
weight composite steel deck floor systems. The schedule critical walls–concrete filled (C-PSW/CF) are a viable alternative to RC
aspect of this operation is the construction of RC core walls, which core walls (Morgen et al. 2018) because the construction schedule
can be expedited by the elimination of formwork (placement and
1
Graduate Research Assistant, Lyles School of Civil Engineering, removal), rebar cages (assembly and installation), and falsework
Purdue Univ., 1040 S. River Rd., West Lafayette, IN 47907 (corresponding (installation and removal). C-PSW/CF also offer opportunities to
author). ORCID: https://orcid.org/0000-0003-2406-8117. Email: mbroberg@ leverage modularization, prefabrication of steel modules, and opti-
purdue.edu
2 mize concrete casting and curing using self-consolidating concrete.
Postdoctoral Research Assistant, Lyles School of Civil Engineering,
Purdue Univ., 1040 S. River Rd., West Lafayette, IN 47907. ORCID: Construction and erection tolerance issues between the steel gravity
https://orcid.org/0000-0003-3475-8525. Email: sshafaei@purdue.edu framing and the C-PSW/CF core system can be minimized by co-
3
Graduate Research Assistant, Dept. of Civil Structural and Environ- ordination of assembly and erection activities. Despite their inno-
mental Engineering, Univ. at Buffalo, 212 Ketter Hall, Buffalo, NY 14260. vation and novelty, C-PSW/CF systems rely on standard composite
Email: emrekizi@buffalo.edu construction techniques already used in practice, reducing the need
4
Research Assistant Professor, Lyles School of Civil Engineering, for extensive worker training.
Purdue Univ., 1040 S. River Rd., West Lafayette, IN 47907. ORCID: C-PSW/CF can be used as uncoupled shear walls or as coupled
https://orcid.org/0000-0003-0861-4261. Email: seo2@purdue.edu core wall systems. Coupled systems consist of two major types of
5
Karl H. Kettelhut Professor, Lyles School of Civil Engineering, Purdue
Univ., 1040 S. River Rd., West Lafayette, IN 47907. ORCID: https://orcid
structural components: C-PSW/CF walls and composite coupling
.org/0000-0001-7153-4681. Email: ahvarma@purdue.edu beams. C-PSW/CF walls consist of two exterior steel faceplates
6
SUNY Distinguished Professor, Dept. of Civil, Structural, and Envir- that are connected to each other with steel tie bars. Steel-headed
onment Engineering, Univ. at Buffalo, 212 Ketter Hall, Buffalo, NY 14260. stud anchors also may be provided on the interior surfaces of
ORCID: https://orcid.org/0000-0003-1170-468X. Email: bruneau@buffalo the steel faceplates. These steel modules are prefabricated in the
.edu shop, transported to the field, assembled, and then filled with plain
7
Chairman and CEO, Magnusson Klemencic Associates, Magnusson concrete. No additional steel reinforcing bars are needed, thus
Klemencic Associates, 1301 Fifth Ave., Suite 3200, Seattle, WA 98101. eliminating the need for assembly and placement of rebar cages.
Email: rklemencic@mka.com
The steel modules serve as stay-in-place formwork for concrete
Note. This manuscript was submitted on July 15, 2021; approved on
November 9, 2021; published online on February 7, 2022. Discussion per- placement and also can be designed to serve as the falsework
iod open until July 7, 2022; separate discussions must be submitted for for construction loads, thus eliminating the need for assembly
individual papers. This paper is part of the Journal of Structural Engineer- and removal of formwork and falseworks. The composite coupling
ing, © ASCE, ISSN 0733-9445. beams are concrete-filled steel box sections, which also are referred

© ASCE 04022022-1 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 to as concrete-filled steel tubes (CFSTs) in the literature. A more experimental testing and numerical modeling approaches. The plas-
extensive review of the system is given in AISC Design Guide 37 tic moment is calculated using standard plastic stress distribution
(Varma et al. 2022). methods as described in Section I2 of AISC 360-22 (AISC 2016b).
Construction of the first building using the coupled C-PSW/CF The in-plane shear strength of C-PSW/CF walls was evaluated
(or CC-PSW/CF) system in seismic regions topped out in August experimentally and numerically by Varma et al. (2011), Seo et al.
2020. The Rainer Square Tower (About Rainier Square, n.d.) stands (2016), and Booth et al. (2020). Equations for conservatively
58 stories tall in downtown Seattle. This structure was constructed estimating the in-plane shear strength of C-PSW/CF walls were
in 10 months, about 40% faster than expected for a RC core build- developed based on those studies and included in AISC N690-18
ing (Post 2020). Over 500 steel modules were prefabricated off-site, (AISC 2018) and AISC 360-22 (AISC 2022b).
transported to the site, erected, and filled with concrete. This In addition to considering strength approaches, limits on the
structure offers a well-documented proof-of-concept for future ductility of the energy-dissipating components also are important
CC-PSW/CF buildings and proven construction schedule reduction in capacity design (Bruneau et al. 2011). For example, in AISC
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

to incentivize widespread adoption of the system. Construction of 341-16 (AISC 2016a), moment frames are categorized largely
the first building using the CC-PSW/CF system in a high seismic based on the rotation capacity of their beams. No rotational capac-
region is ongoing in San Jose, California. The 200 Park Avenue ity criteria are applied to ordinary moment frame systems, whereas
building (Home, n.d.) stands 19 stories tall with a total of 89,692 m2 intermediate and special moment frame beams are required to dem-
(965,342 ft2 ) of residential area. onstrate rotational capacities of 0.02 and 0.04 rad, respectively,
The development of a robust design methodology and detailing while maintaining a load of at least 80% of their maximum flexural
requirements will enable designers to implement this CC-PSW/CF strength. ASCE 7 (ASCE 2022) applies different seismic response
system efficiently without having to perform computationally ex- coefficients depending on the classification of the system (ordinary,
pensive analyses to demonstrate system-level behavior. For consis- intermediate, or special moment frames).
tency with other steel lateral load–resisting systems, and to ensure The superior seismic performance of CC-PSW/CFs over un-
satisfactory seismic performance, such a robust design methodol- coupled walls was demonstrated in a set of FEMA P695 studies
ogy must be anchored in capacity design principles. The capacity performed by Agrawal et al. (2020) on uncoupled systems and
design concept was developed initially in New Zealand by Park and by Kizilarslan et al. (2021b) on coupled systems. Those studies
Paulay (1975). In this design approach, primary energy-dissipating employed capacity design principles to size the walls and coupling
elements of a given structural system are selected and detailed for beams, as applicable. Agrawal et al. (2020) confirmed that a seis-
ductility, whereas other structural elements are designed (per capac- mic force reduction factor (R factor) of 6.5 was appropriate for un-
ity design principles) with sufficient strength to ensure that the coupled C-PSW/CF systems, whereas Kizilarslan et al. (2021b)
target energy dissipating mechanism can be achieved. confirmed that an R factor of 8 was appropriate for CC-PSW/CF
systems. These values subsequently were implemented in FEMA
P-2082-1 (FEMA 2020) and ASCE 7 (ASCE 2022). This R factor
Background of 8 is consistent with recent recommendations for coupled concrete
shear walls (Tauberg et al. 2019). The coupled wall study enforced
Implementation of capacity-design principles for steel design in strong wall–weak coupling beam principles to spread plasticity
North America goes as far back as the 1980 edition of the SEAOC along the height of the structure. In the uncoupled wall study, yield-
seismic design provisions (SEAOC 1980), which were adopted in ing was limited to the base of wall elements (Agrawal et al. 2020),
the 1985 edition of the Uniform Building Code (ICBO 1985). whereas in the coupled study, yielding occurred in coupling beams
These provisions required beam-to-column connections in moment- before initiation of yielding in the walls (Kizilarslan et al. 2021b).
resisting frames to be capable of developing the expected strengths Much like moment frame systems, this additional ductility is
of the beam hinges, and connections of braces in braced frames to associated with the rotational capacity of the coupling beam ele-
be capable of developing the expected strengths of the bracing ments. Coupling beams designed for the FEMA P695 study in-
members. Uang and Bruneau (2018) provided a historical perspec- cluded provisions that coupling beams must be flexure-critical
tive on how this design approach evolved over the last decades into (i.e., have a span-to-depth ratio of at least 2.5) and the rotational
the complete system-based capacity design approach that is at the capacity of coupling beams must be 0.030 rad to use the higher R
core of the AISC Seismic Provisions for Steel Buildings, AISC factor of 8 (Kizilarslan et al. 2021b). Flexure-critical coupling
341-16 (AISC 2016a). beams connected with complete joint penetration welds to the steel
Capacity design principles already have been established for plates of C-PSW/CF have been able to develop this rotation capac-
uncoupled C-PSW/CF walls (Alzeni and Bruneau 2017; Kurt et al. ity while retaining 80% of their flexural strength (Nie et al. 2014;
2016). Tests of walls with boundary elements by Alzeni and Varma et al. 2021). Shear-critical coupling beams are not recom-
Bruneau (2017) indicated that the plastic moment capacity (M p ) mended because their ductility and rotation capacity are limited by
can be estimated conservatively using a plastic stress distribution the shear (compression strut) failure of the cracked concrete be-
model for walls with boundary elements. For walls with semicircular tween the ends of coupling beams (Nie et al. 2014). This paper
boundary elements, Alzeni and Bruneau found the ultimate capacity details the capacity design principles developed by the authors
to be 1.1M p , whereas for walls with circular boundary elements, for the design of CC-PSW/CF and demonstrates through numeri-
the ultimate capacity was 1.2M p considering the measured material cal modeling that the desired capacity design limits are imple-
properties and concrete strength. These results suggest that if wall mented successfully for CC-PSW/CF walls when applying
elements are limiting forces for other components, the expected plas- these provisions.
tic moment capacity (M p;exp ) of the section may need to consider an
amplification factor to capture this additional strength.
Kurt et al. (2016) showed that the capacity of walls without Basis of Design
boundary elements (flange plates) can be approximated reasonably
as the plastic moment of the section for wall aspect ratios (h=lw ) CC-PSW/CF systems use coupled walls to resist laterals loads such
greater than or equal to 1.5. This analysis is supported by as design basis and maximum considered earthquakes (MCEs).

© ASCE 04022022-2 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Fig. 1. Desired pushover behavior of the CC-PSW system.

In these events, the coupled wall system is expected to undergo Capacity Design
significant inelastic deformation.
Individual C-PSW/CFs have a plastic rotation capacity of A capacity design approach was adopted as the design basis. This
0.015–0.02 rad (Shafaei et al. 2021b). This rotation capacity is con- design approach is used to define the required strengths for com-
centrated at the base of the C-PSW/CF walls, whereas most of the ponents based on the expected strength of fuse elements. In this
wall remains essentially elastic along its height. As indicated pre- case, the plastic hinges at the ends of the coupling beams are
viously, a seismic response modification factor (R factor) of 6.5 the designated fuses. The expected strengths of these components
is assigned by ASCE-7 to this uncoupled system. To achieve a are used to define the demand loading (required strengths) for the
higher R factor of 8, the (coupled) CC-PSW/CF system must have walls. In other words, the walls are sized to have the capacity to
better inelastic behavior and dissipate more hysteretic energy than resist the forces imparted to them, including those imparted by all
uncoupled C-PSW/CFs. This additional energy dissipation is the coupling beams along the height of the structure that have
achieved by plastic hinging of coupling beams in all stories along formed flexural plastic hinges at their ends.
the height of the building, thereby altering the structure’s governing Implementing this principle should lead to the characteristic
plastic collapse mechanism. pushover behavior in Fig. 1, in which the initial branch represents
To ensure that the plasticity spreads along the height of the elastic behavior with a slope corresponding to the effective stiff-
structure and that hinges form in the coupling beams, the system ness. As the lateral load (and base shear) increases, the load level
must be proportioned such that the coupling beams yield before reaches Point A, corresponding to the equivalent lateral force (ELF)
yielding in the wall. This progression of failure is required so that level. This point corresponds to the required strengths (design
the coupling beams engage and dissipate energy—otherwise, as in demands) for the coupling beams, and thus initiation of plasticity
the uncoupled system, flexural hinging may be limited to the base in the coupling beams. As the lateral load (and base shear) is in-
of the wall and no additional ductility and energy dissipation will be creased, the response reaches Point B, at which all coupling beams
harnessed. In the context of CC-PSW/CF, the coupling beams are have formed plastic hinges at both ends. This load level corre-
concrete-filled box sections. Limited testing of these composite sponds to the required strengths (design demands) for the walls,
coupling beam-to-wall connections (Nie et al. 2014) suggested and thus potential initiation of plasticity at the base of the walls.
that the beam rotation capacity in flexural hinging is greater than As the lateral load is increased, the response reaches the next mile-
the rotation capacity associated with shear yielding. Additionally, stone, Point C, corresponding to the formation of plastic hinges at
concerns about the effects of shear yielding coupling beams on the base of the walls, and thus formation of the overall inelastic
wall behavior, for example, the spread of shear yielding into mechanism. Finally, Point D represents fracture failure of the cou-
the wall elements, have not been investigated. Therefore, the sys- pling beams or walls in a monotonic pushover behavior.
tem is required to be proportioned with flexure-controlled cou- To implement this design process, a linear elastic frame model is
pling beams. constructed and subjected to ELF level loads. This elastic model
The inelastic deformation in CC-PSW/CF systems has two sour- accounts for effective axial and flexure stiffness values for both
ces: (1) flexural plastic hinges at the ends of coupling beams, and wall and coupling beam elements. Coupling beam effective axial
(2) flexural yielding at the base of the walls. As mentioned previ- [ðEAÞeff ] and effective flexural [ðEIÞeff )] stiffness are calculated
ously, the preferred inelastic mechanism is to develop plastic hinges per AISC 360-22 (AISC 2016b), Section I1.5. C-PSW/CF wall
at the ends of coupling beams before the developing them at the effective stiffnesses can be estimated as the secant stiffness corre-
base of the walls. To achieve this inelastic mechanism the members sponding to 60% of the plastic moment. Alternatively, these stiff-
must be proportioned following a strong wall–weak coupling beam ness values can be estimated using Eqs. (1)–(3) in AISC 360-22
design philosophy. To ensure this type of behavior, a capacity de- (AISC 2022b) Section I1.6, which were developed by Agrawal et al.
sign principle is used. This principle ensures that the loading level (2020) by calibrating linear elastic models to nonlinear models
associated with the plastic capacity of the wall sections exceeds the (with concrete cracking) subjected to ELF level loads. These stiff-
loading level associated with the initiation of yielding in the cou- ness values also were shown to match closely the stiffness values
pling beams. By following the design approached discussed sub- calculated using the secant stiffness corresponding to 60% of the
sequently, the aforementioned objectives will be attained. plastic moment

© ASCE 04022022-3 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 ðEIÞeff ¼ Es I s þ 0.35Ec I c ð1Þ The portion of the total overturning moment resisted by cou-
pling action between the walls is calculated as the equal and op-
ðEAÞeff ¼ Es As þ 0.45Ec Ac ð2Þ posite axial forces (Pw;CA ) times the distance between them.
The remaining portion of the overturning moment is resisted by
ðGAÞeff ¼ Es As þ Ec Ac ð3Þ the composite walls and distributed to them in accordance with
their relative flexural stiffnesses [Eqs. (9)–(11)]. Compression
This elastic model is used to calculate the required strengths and tension walls have different flexural stiffnesses due to the dif-
(or design demands) for coupling beams and the story drifts for ferences in the extent of (uncracked) concrete in compression
the structure. These calculated story drifts are compared with the contributing to the section secant stiffness. The section secant stiff-
corresponding drift limits specified by the applicable building nesses of the walls subjected to axial tension (EI T Wall ) and axial
code. After designing the coupling beam sections to provide avail- compression (EI C Wall ) can be estimated as the secant stiffness cor-
able (design) strength greater than or equal to the required responding to 60% of the flexural capacity accounting for the
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

strength, the walls are designed as capacity-protected members. effects of the axial forces (Shafaei et al. 2021b). For example, the
At the core of this analysis is the determination of the maximum section secant stiffness can be obtained from a moment-curvature
force that the coupling beams will transfer to the walls due to the analysis of the wall considering axial force
formation of plastic hinges at the ends of all coupling beams. The
axial force in the walls is calculated as the sum of the capacity- Mwalls ¼ γ 1 OTMA − Pw;CA Leff ð9Þ
limited coupling beam shear capacity along the height of the
EI T Wall
structure (added to gravity forces). To determine this axial force, MU T Wall ¼ × M walls ð10Þ
expected flexural capacities of selected and sized coupling beams EI C Wallþ EI T Wall
are calculated, accounting for the material expected strength (Ry
EI C Wall
and Rc factors). These expected capacities are amplified by 1.2 to MU C Wall ¼ × M walls ð11Þ
account for strain hardening in the steel, a biaxial stress state in EI C Wall þ EI T Wall
the steel tension flange (Shafaei et al. 2021a), and concrete
where Leff = distance between geometric elastic centroid of tension
confinement.
and compression walls; EI T Wall and EI C Wall = effective section
Conceptualizing this calculation on the theoretical pushover
stiffnesses of tension and compression walls, respectively; and
curve, the amplified expected flexural capacity of the coupling
M U T Wall and M U C Wall = required flexural strengths for tension
beams is associated with the loading at Point B, at which all the
and compression walls, respectively.
coupling beams have formed plastic hinges. Experimental results
Finally, the shear forces resisted by the walls for ELF level loads,
indicate that the plastic hinge forms over a length equal to approx-
calculated from elastic analysis, are amplified by a factor of 4 as a
imately one-half the coupling beam depth (Nie et al. 2014). For
conservative approach to account for higher-mode effects and over-
design, the plastic hinges are considered to form at the ends of the
strength in the walls from the difference in the design demand (Point
clear span of the coupling beams. The total overturning moment at
B) and the expected flexural capacity (Point C). Applying such a
Point B can be estimated using the total overturning moment at
shear amplification factor is consistent with the approach and
Point A according to
findings from coupled concrete shear walls (Tauberg et al. 2019).
P
1.2M cb
p;exp
γ 1 ¼ nP cb ð4Þ
n Mu
Additional Design and Detailing Requirements
OTMB ¼ γ 1 OTMA ð5Þ
In addition to the preceding capacity design approach, several design
P and detailing requirements are needed to ensure that the composite
where n 1.2M cb p;exp = sum of expected
P flexural capacities of cou-
pling beams along structure height; n M cb sections develop their full plastic strength, have adequate ductility,
u = sum of flexural de-
sign demands for coupling beams along structure height; n = and are sized appropriately for construction concerns such as trans-
number of coupling beams along structure height; and OTMA and portation and standard concrete casting pressures. Such require-
OTMB = overturning moments at Point A and B, respectively, on ments effectively are those prescribed in ASCE 7 (ASCE 2022),
theoretical pushover curve. and AISC 341-22 (AISC 2022a). These include limits on the
Returning to calculating the axial force in the wall, the amplified reinforcement ratio, plate slenderness, and tie-bar spacing based
expected moment is converted into the capacity-limited expected on the research of Zhang et al. (2014, 2020) and Varma et al.
shear at the ends of the coupling beam following (2019).
Coupling beam requirements similarly are specified to ensure
2 × 1.2M CB
p;exp that they behave as intended. First, the existing AISC 360-22
amp;exp ¼
V CB ð6Þ (AISC 2016b) provisions and the upcoming AISC 360-22 (AISC
LCB
2022b) and AISC 341-22 (AISC 2022a) provisions apply in this
where LCB = clear span length of coupling beam. case, including those for the minimum area of steel, compactness
This beam end shear is summed along the height of the structure criteria, and flexural and shear stiffness values. Second, in addition
and added to the gravity load in the wall to calculate the capacity- to these existing requirements, coupling beams must be designed to
limited axial force in the wall, following Eqs. (7) and (8). One wall be flexure-critical. In other words, the ultimate behavior of the
will be subjected to axial compression and the other wall will be beams must be governed by flexural yielding rather than shear
subjected to axial tension strength. As previously discussed, this is because existing test data
P on composite coupling beams (Nie et al. 2014) indicated greater
2.4 Mcb p;exp ductility in flexure-controlled sections. Flexure-controlled coupling
Pw;CA ¼  ð7Þ
Lcb beams can be obtained by requiring their length-to-depth ratio to be
greater than or equal to 3. Therefore, a range of 3–5 was specified
Pw;exp ¼ Pw;CA þ Pgravity ð8Þ for the archetype structures evaluated in the FEMA P695 study

© ASCE 04022022-4 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 performed to validate the R factor for these walls (Kizilarslan et al. Planar coupled walls generally are appropriate for low- to midrise
2021b), with the upper limit of 5 imposed to ensure meaningful (up to 15-story) buildings, whereas C-shaped walls are better suited
contributions of the coupling beams to the structural system. for midrise to high-rise (12–24-story) buildings because the greater
More rigorously, to ensure that the coupling beams are flexure- seismic loads necessitate additional strength and stiffness. Standard
controlled, it is specified that the shear strength of the composite materials were used, namely 41.4 MPa (6,000-psi) normal-weight
coupling beam must be higher than the capacity-limited shear in the concrete and A992 steel [345 MPa (50 ksi)].
coupling beams due to forming plastic hinges at both ends. This Structural analysis followed the equivalent lateral force pro-
limit is checked using the following equation: cedure outlined in ASCE 7-16, using design spectra based on
the site-specific seismic response parameters for a Seismic design
2.4M p;exp category D earthquake (SDS ¼ 1.0 g and SD1 ¼ 0.6 g). The funda-
V n;exp ≥ ð12Þ
Lcb mental period was estimated based on the structure type and geom-
etry, with the upper-limit period (Cu T a ) used as the initial-period
where V n;exp = expected shear strength of composite coupling beam
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

estimate. In these examples, all structural periods obtained from

calculated per AISC 360-22 (AISC 2022b) Section I4.2 using ex- eigenvalue analysis performed using commercially available soft-
pected yield strength, Ry Fy for steel and expected compressive ware (SAP2000 17) were greater than the upper-limit periods. The
strength Rc f c0 for concrete [the shear strength equations in AISC seismic response coefficient, Cs , was calculated and used to distrib-
360-22 (AISC 2022b) Section I4.2 are based on Kenarangi et al. ute the seismic forces vertically along the height of the structure.
(2021a)]; and M p;exp = expected flexural capacity of composite
The amplified base shear and overturning moment were calculated
coupling beam calculated using expected yield strength, Ry Fy , for
based on this distribution. The amplified base shear was taken as
steel and expected compressive strength Rc f c0 for concrete.
the equivalent lateral force base shear multiplied by a factor of 4 to
Generally, when the coupling beam forms plastic hinges at its
account for higher-mode effects (for reasons mentioned previously).
ends, the capacity-limited shear in the beam (experienced at the
An assumed coupling ratio was used to distribute the total over-
ends) is equivalent to 2 times the plastic moment divided by the
turning moment into contributions due to the (1) axial force couple,
member length, but Eq. (12) is more conservative to consider addi-
and (2) individual walls. Initially, the coupling ratio was assumed to
tional capacity from a biaxial (tensile) stress effect in the steel, steel
be 60%; this value was adjusted, as appropriate, during design iter-
strain hardening, and confinement of the concrete. This additional
capacity was discussed by Bruneau et al. (2019), and is similar to ations. Initial wall and coupling beam dimensions were assumed,
that discussed by Shafaei (2020). The shear and flexural strength and the shear, flexural, and axial stiffnesses were calculated for
can be calculated according to the methods prescribed in both wall and coupling beam elements. No recommendations cur-
AISC 360-22 (AISC 2022b) Sections I4.2 and I1.2a, respectively. rently exist for initial wall sizes but, through iterations, reasonable
Finally, the only coupling beam-to-wall connection details per- wall sizes were determined. The stiffness values then were used in
mitted are those able to develop a coupling beam chord rotation an elastic model with equivalent lateral force procedure loads. As
capacity of 0.030 rad before flexural strength decreases to 80% mentioned previously, these stiffness values were developed by
of the beam’s flexural plastic strength. This requirement is consis- Agrawal et al. (2020) by calibrating linear elastic models (with
tent with the models used in the FEMA P695 study (Kizilarslan the effective stiffness values) with nonlinear models (with concrete
et al. 2021b) and the rotation capacity seen in tests (Nie et al. 2014; cracking modeled explicitly) subjected to ELF level loads. This
Varma et al. 2021). model was used to calculate the interstory drift ratio. Finally, the
maximum amplified inter-story drift was compared to the code-
specified drift limit, in this case, 2% interstory drift per ASCE7-16.
Design Examples If the system met the drift limit, the design process continued; if the
system did not meet the drift requirements, the system was resized
The capacity design procedure and corresponding requirements to bring drift below the acceptable limit.
were demonstrated using a series of design examples. These struc- Next, the coupling beams were designed. The design checks for
tures represent typical buildings that could use the CC-PSW/CF coupling beams were outlined earlier. The calculated strengths
system. Building geometry, floor loading, and material properties were compared with the required coupling beam strengths (shear
for the buildings considered are presented in Table 1. These in- and flexure) obtained from the analysis. Coupling-beam geometry
cluded a typical story height of 4.3 m (14 ft) and floor plan dimen- requirements including limits on the length-to-depth ratio, flange
sions of 61.0 × 36.6 m (200 × 120 ft). Both planar and C-shaped plate slenderness, and web plate slenderness also were checked.
walls were included in the design examples. The building floor lay- If necessary, the coupling beams were resized, and the preceding
outs for planar and C-shaped coupled walls are shown in Fig. 2. analysis was repeated with the new cross sections. Then the

Table 1. Input parameters for structure design

Parameter Values Reasoning
Coupling beam aspect ratio (L=d) 3, 4, 5 Typical coupling beam aspect ratios
Story height First story: 5.2 m (17 ft), typical story: 4.3 m (14 ft) Typical story heights
Seismic weight Floor load of 5.75 kPa (120 psf) Estimated from components: steel framing [0.58 kPa
(12 psf)]; 2.5-in. normal-weight concrete on 3-in.
steel deck [2.39 kPa (50 psf)]; curtain wall [0.72 kPa
(15 psf)] on facade area]; superimposed dead load
[0.72 kPa (15 psf)]; and partitions [0.72 kPa (15 psf)]
Coupled wall length 9.1 m (30 ft) Typical bay length is 9.1 m (30 ft), set of walls would
span the length of a bay
Floor dimensions 36.3 × 61.0 m (120 × 200 ft) Typical floor geometry

© ASCE 04022022-5 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Fig. 2. Layout of (a) planar walls; and (b) coupled walls.

composite walls were designed. The shear and axial capacities were such as the wall length, wall thickness, and steel plate thickness.
compared with the corresponding demands determined from the However, because the wall length often is constrained by architec-
analysis. The flexure demand was determined based on the capacity tural requirements, the geometric parameters that were adjusted to
design process detailed earlier. This step considered the relative redesign the walls were limited to the wall thickness and/or steel
flexural stiffnesses of the tension or compression walls. The flexu- plate thickness. Increasing the wall thickness increases stiffness,
ral demands obtained from capacity design were compared with the whereas increasing the steel plate thickness increases strength.
flexural capacity of the wall, calculated using the plastic stress dis- Connection design and detailing issues were beyond the scope
tribution method, while accounting for the effects of axial forces of this paper, but would be performed at this point. This would
(compression or tension). Walls were redesigned if their capacity include the connections between the tie bars and steel plates and
was less than demand, and reanalyzed for a new iteration. Typi- between C-PSW/CF components, coupling beam-to-wall connec-
cally, walls can be redesigned by adjusting geometric parameters tions, and wall-to-foundation connections. The requirements for

© ASCE 04022022-6 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 Table 2. Section sizes for coupled planar walls
Stories L=d Lw [mm (in.)] tsc [[mm (in.)] tp [mm (in.)] Lcb [mm (in.)] Coupling beam section [mm (in.)]
8 4 3,350 (132) 610 (24) 14.3 (9/16) 2,440 (96) 24 × 24, 1/2(f), 3/8(w) [610 × 610, 12.7(f), 9.5(w)]
12 3 5,180 (204) 457 (18) 14.3 (9/16) 1,830 (72) 18 × 24, 5/16(f), 3/8(w) [457 × 610, 7.9 (f), 9.5(w)]

Table 3. Section sizes for CC-shape walls

Hw Lf tsc;f tsc;w tp Lcb Coupling beam section
Stories L=d [mm (in.)] [mm (in.)] [mm (in.)] [mm (in.)] [mm (in.)] [mm (in.)] [mm (in.)]
18 5 9,140 (360)) 3,960 (156)) 660 (26)) 406 (16)) 14.3 (9=16) () 3,050 (120) 26 × 24, 1/2(f), 3/8(w) [610 × 610, 12.7(f), 9.5(w)]
22 4 9,140 (360)) 4,880 (192)) 610 (24)) 356 (14)) 12.7 (½)) 2,440 (96) 24 × 24, 7/16(f), 3/8(w) [610 × 610, 11.1(f), 9.5(w)]
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Table 4. Summary of important structure design to demand ratios

Ratio of V n;exp to
Ratio of ϕM n to Mu , Ratio of ϕM n to ð2.4M p;exp Þ=Lcb
Stories Cu T a (s) T n (s) compression wall M u , tension wall ϕV n =V base for coupling beams IDRmax (%)
8 0.98 1.42 1.3 1.2 3.6 1.1 1.4
12 1.32 1.83 1.7 1.8 4.7 1.2 1.3
18 1.78 2.98 6.0 6.6 2.9 1.3 1.9
22 2.07 3.60 1.8 1.9 2.6 1.2 2.4

these connections are specified in AISC 341-22 (AISC 2022a). 2021a). These benchmarking studies were presented in detail by
Tie-to-plate connections are required to be detailed to develop Bruneau et al. (2019), and are not repeated here for brevity.
the full tensile strength of the tie bar. C-PSW/CF components
(between steel web plate and flange plates) are required to be
complete joint penetration welds. Coupling beam-to-wall connec- 2D Abaqus Model
tions need to develop the rotation capacity and strength described The 2D Abaqus model consisted of finite-element components for
in the coupling beam requirements section. Wall-to-foundation all beam and wall elements. The walls were modeled with four-
connections are to be designed to resist 1.1 times the expected node composite shell sections with reduced integration (S4R) to
plastic composite flexural strength of the wall. This requirement represent the infill concrete core and web plates, and with two-node
for wall-to-foundation connections is in accordance with AISC truss element (T3D2) to represent the steel flange plates. The cou-
341-16 Section H7. This 1.1 factor is included to prevent failure pling beams similarly were modeled with layered composite shell
of the foundation connection prior to development of the wall elements (S4R) for infill concrete and web plate and truss element
plastic hinge. (T3D2) for coupling beam flange plates. Additionally, a leaning
The properties of four resulting example structures designed per column with gravity loads was modeled using truss elements
the preceding procedure are presented in Tables 2 and 3. These (T3D2) to represent the P-delta effects on the structure. The steel
properties include the length of the planar wall, Lw ; length of the and concrete material models used in this analysis were based on
C-shaped wall web, H w ; length of the C-shaped wall flange, Lf ; effective stress–strain curves developed by Shafaei et al. (2021a)
total wall thickness, tsc ; plate thickness, tp ; and coupling beam from a three-dimensional (3D) finite-element model benchmarked
length, Lcb . These properties were chosen to be in the practical to C-PSW/CF testing. These effective stress–strain curves were
range for building construction while meeting the controlling limit used to implicitly account for the effects of the biaxial stress state
states (drift or strength) without introducing significant over- in the steel and confinement in the concrete.
strength (plastic strength-to-required strength ratio). Table 4 sum- The behavior of the 2D Abaqus wall model was benchmarked to
marizes the relevant structural performance markers. tests performed by Shafaei et al. (2021a). The detailed 3D finite-
element, 2D finite-element, and experimental results for one of the
benchmarking experiments for the walls are shown in Fig. 3. This
Finite-Element Modeling analysis showed that the effective stress–strain curve implementa-
tion in the 2D model can capture very similar behavior to the more
These example structures were modeled using multiple approaches: robust 3D model. This 2D finite-element model is computationally
a two-dimensional (2D) finite-element model in Abaqus version efficient for analyzing multistory buildings. The 3D and 2D finite-
2017, and two independently calibrated fiber-based models in element models also accurately can simulate the cyclic behavior of
OpenSees version 2.5.0. This approach was used because the the experimental test including the lateral stiffness, capacity, post-
finite-element model directly calculated and illustrates numerous peak degradation, and fracture failure of the steel plates. However,
failure modes (including local buckling, shear yielding, and effects due to limitations in the concrete damage model, the pinching ob-
of confinement), whereas the fiber-based model was computation- served in the experiment was not captured accurately; the cyclic
ally efficient and able to simulate expected failure modes (such comparisons were reported in full by Shafaei et al. (2021a).
as coupling beam facture, cyclic degradation of steel, and steel A similar process was repeated for coupling beam elements,
fracture). All these models were benchmarked independently to and associated stress–strain curves for these elements also were ex-
planar and C-shaped C-PSW/CF walls and coupling beams tests tracted and implemented in the 2D finite-element model. This
(Kizilarslan et al. 2021a; Kenarangi et al. 2021b; Shafaei et al. analysis was benchmarked to the flexural critical tests discussed by

© ASCE 04022022-7 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 Table 5. Steel material properties used for planar wall benchmarking
Parameter Value
Elastic modulus, Es [ksi (MPa)] 29,000 (200,000)
Strain hardening ratio, b 0.01
Tangent at strain hardening, Esh [ksi (MPa)] 290 (2,000)
Yield stress, Fy [ksi (MPa)] 50 (345)
Ultimate stress, Fu [ksi (MPa)] 65 (448)
Strain at initiation of strain hardening, εsh 2Fy =Es
Strain at ultimate stress, εult 0.1
Slenderness ratio, Lsr 10
Buckled stress amplification factor, β 1
Buckling reduction factor, r 0.65
Buckling constant, γ 0.5
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Coffin–Manson constant C, Cf 0.6

Coffin–Manson constant a, α 0.5
Fig. 3. Finite-element results from 3D and 2D models. Cyclic strength reduction constant, Cd 0.35

Nie et al. (2014) and Varma et al. (2021). Further test details on this
benchmarking process were presented by Bruneau et al. (2019).
Table 6. Concrete material parameters for planar wall benchmarking

OpenSees Models Parameter Value

Compressive strength, f c0[ksi (MPa)] 6.0 (41)
The OpenSees models consisted of fiber elements representing
Strain at maximum strength, epsc0 0.0022
beam and wall sections. The first OpenSees model relied on imple- Crushing strength, Fpcu 0.6f c0
menting the effective stress–strain curves in OpenSees and model- Strain at crushing strength, epsu 0.008
ing coupling beams with concentrated plasticity elements, whereas Tension softening stiffness, Ets 0.1f c0
the second model used material models natively available in Strain at tensile strength, et 8 × 10−5
OpenSees and modeled coupling beams with distributed plasticity Ratio between unloading slope at epsc0 and initial slope, λ 0.1
elements. Analyzing the example structures using differing fiber-
based modeling approaches adds additional certainty to the behav-
ior if the analytical models produce similar results.
In the first model, referred to herein as the concentrated plas-
ticity (CP) model, walls were modeled with nonlinear fiber ele- Table 7. Parameters used to model behavior of coupling beams with
ments from the base of the wall to a height equal to the length of concentrated plasticity elements
the wall. These nonlinear fiber elements used the same steel and Parameter Value
concrete effective stress–strain curves as the Abaqus walls. These
Elastic stiffness, K 0 [kN-m/rad (kip-in./rad)]
stress–strain curves were implemented in OpenSees using Reinfor- Planar walls (8 and 12 stories) 2.8 × 106
cingSteel and Concrete02 to model the steel and concrete behavior, (2.5 × 107 )
respectively. ReinforcingSteel was chosen because it could simu- C-shaped walls (18 and 22 stories) 5.6 × 106
late cyclic degradation and nonsymmetric tension and compression (5.0 × 107 )
behavior (to simulate local buckling of steel plates in compression). Strain hardening ratio, as 0.005
Concrete02 was chosen because it adequately captured crack closure Yield moment, M y [kN-m (kip-in.)]
and had a higher rate of convergence than other alternatives tried. 8-story 2,710 (24,000)
Wall elements above the nonlinear section were assigned cracked 12-story 1,810 (16,000)
18-story (two coupling beams) 5,620 (49,700)
transformed elastic properties. Nonlinear properties used for wall
22-story (two coupling beams) 4,980 (44,100)
steel and concrete are presented in Tables 5 and 6, respectively. Precapping rotation, θp (rad) 0.025
Coupling beam elements were modeled with nonlinear concen- Post-capping rotation, θpc (rad) 0.04
trated plasticity elements at both ends representing the flexural Ultimate rotation capacity, θu (rad) 0.05
plastic hinge behavior. This concentrated plasticity element used Cyclic deterioration parameter for 0.5
a Modified Ibarra–Medina–Krawinkler deterioration model with strength deterioration, λS
bilinear hysteretic response material to capture the cyclic behavior Cyclic deterioration parameter for 1.0
of coupling beam elements. Although this model has the capability postcapping strength deterioration, λC
of capturing nonsymmetric behavior in the push and pull cycles, the Cyclic deterioration parameter for accelerated 1.0
reloading stiffness deterioration, λA
coupling beams’ behavior was considered to be symmetric because
Cyclic deterioration parameter for unloading 1.0
the sections were compact, resulting in uniform properties in the stiffness deterioration, λK
push and pull directions. The concentrated plasticity elements Residual strength ratio, Res 0.25
were connected with elastic elements assigned elastic properties. Rate of cyclic deterioration, D 1
Concentrated plasticity parameters are presented in Table 7. Rate of strength deterioration, cS 1.0
For dynamics analysis, the structure was assigned a damping Rate of postcapping strength deterioration, cC 0.0
ratio of 5%. This damping was implemented using Rayleigh damping Rate of accelerated reloading deterioration, cA 1.0
and enforcing mass and stiffness coefficients according to the first Rate of unloading stiffness deterioration, cK 1.0
two natural periods of the structure. This was implemented as mass Elastic stiffness amplification factor, nFactor 0
Ratio of reloading stiffness, Apinch 1
and stiffness–proportional damping based on the first and second

© ASCE 04022022-8 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 Table 8. First and second structural periods for CP and DP models
First Second First Second
period, period, period, period,
T 1 (s) T 2 (s), T 1 (s), T 2 (s),
Stories CP model CP model DP model DP model
8 1.1 0.26 1.05 0.26
12 1.4 0.32 1.38 0.31
18 2.7 0.59 2.05 0.48
22 3.3 0.71 2.56 0.58

periods of the structure. The periods for each example structure are
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

reported in Table 8. Parameters used in the benchmarking models

were similar to these but accounted for the actual yield and ultimate
strengths of the steel and the nominal strength of the concrete. Fig. 4. Representative finite element (Abaqus) results for 8-story
Further information about the benchmarking of these models structure.
and comparisons with the remainder of the tests were given by
Bruneau et al. (2019).
In the second model, referred to herein as the distributed plas- which all coupling beams yielded, and fracture of coupling beams
ticity (DP) model, for walls, the nonlinear beam column elements preceding fracture in the walls. For example, Fig. 4 shows the char-
were assigned only to the first story of the walls, and the rest of the acteristic base shear–roof displacement response obtained from con-
stories were modeled using elastic beam–column elements with ducting a static pushover analysis of a nonlinear inelastic finite
effective stiffness per AISC 341-22 (AISC 2022a) mentioned element model of an 8-story structure design. Behavior milestones
previously. The coupling beams were all modeled using distributed- are marked along the pushover response, and the corresponding stress
plasticity nonlinear beam–column elements. The nonlinear elements states (emphasizing the extent of yielding) are included in Fig. 5.
were assigned to the centroid of composite wall cross sections as Similar results were obtained from the fiber analysis models.
they were calibrated. To simplify analysis, only half of each build- The milestones identified from these analyses included (1) first
ing (and thus half of the C-shaped walls) was modeled due to sym- coupling beam yields, (2) last coupling beam yields, (3) wall yields,
metry, but all results are presented for full buildings. Leaning and (4) coupling beam fracture. For the CP model, coupling beam
columns were added to the structural model to capture the P-Δ yielding was identified as the point at which the coupling beam
effects due to the story gravity loads that were not acting on the moment exceeded the specified yield moment (M p ) value. Cou-
CC-PSW/CF system itself. These columns were modeled using pling beam yielding then spread along the height of the structure.
elastic beam–column elements. The moments of inertia and cross- When all coupling beams reached M p , this event was marked as the
section area of the elastic beam–column elements should be multi- last coupling beam yields milestone. For the DP model, coupling
plied to represent the number of leaning columns assumed to exist beam yielding was identified as when the extreme fiber strain ex-
in the archetype structure. Because there was no definitive infor- ceeded the specified yield strain value. For both models, wall yield-
mation on the number of leaning columns in the archetype design, ing was identified when the stress in the extreme fibers in the wall
these values were chosen arbitrarily to provide insignificant flexu- section exceeded the yield stress (Fy ). A coupling beam fracture
ral stiffness. Tributary loads coming to the C-PSW/CF walls were occurred when the first coupling beam began to maintain a lower
applied to the walls on each floor. Rigid links were assigned be- load than the ultimate load.
tween the C-PSW/CF wall center of gravity and the point at which Fig. 6 shows the base shear versus roof displacement responses
the coupling beams framed into the walls, and rigid beams were obtained from the nonlinear static pushover analyses of the 8-, 12-,
used to connect the leaning column and C-PSW/CF wall at every 18-, and 22-story structures conducted using both fiber-based
floor. These rigid beams were modeled using elastic (almost rigid) (CP and DP) models. These plots consistently show the equivalent
truss elements. No seismic mass was assigned to the leaning lateral force load level and yielding of the first coupling beam close
columns. Seismic masses were applied to the C-PSW/CF walls and to one another. Next, the last coupling beams yielded. This mile-
distributed equally to their left and right joints at every story. stone was followed by yielding in the walls, and finally coupling
Rayleigh damping was used with a value of 5% damping specified beam fracture. These pushover analysis curves from both the CP
for the first and second periods of vibration. The sensitivity of and the DP models were similar for all four example structures, and
results to other damping ratios was considered and discussed by the initial milestones were shown to occur at similar displacements,
Kizilarslan et al. (2021b). except the DP model exhibited a more ductile postpeak behavior
with a longer and gentler descending branch. This extended duc-
tility occurred because the material models used were benchmarked
Seismic Behavior to cyclic (hysteretic) component behavior and used cumulative
plastic strains to model fracture failure (Kizilarslan et al. 2021a).
The fracture-initiating cumulative plastic strain values therefore
Pushover Behavior were reached only at large displacements during the monotonic
The finite-element and fiber models were subjected to pushover pushover analyses. On the other hand, the CP model used the
analysis. Following the procedure defined in ASCE 41 (ASCE envelope of the cyclic hysteretic behavior to define the moment-
2017), the structures were subjected to an increasing lateral load rotation behavior of the concentrated plastic hinges at the ends of
distributed according to their first mode response. The behavior in the coupling beam elements. Consequently, the concentrated plastic
the pushover response followed the capacity design principles: hinges were programmed to undergo fracture failure at a rotation of
namely yielding of the coupling beams around the equivalent lat- 0.03 rad, irrespective of cumulative inelastic strain history. This
eral force level loads, yielding of the walls near the loading level at event (indicated with a solid triangle on the pushover response

© ASCE 04022022-9 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Fig. 5. Finite-element milestones for 8-story structure: (a) ELF; (b) all coupling beams yield; (c) plastic mechanism with all coupling beams and wall
yielding; and (d) fracture of beams and walls.

Fig. 6. Fiber analysis (CP and DP model) pushover results for 8-, 12-, 18-, and 22-story example structures.

plots) demarcates the deviation in the postpeak responses from the modeling approach. In addition to exploring the overstrength of
two models. the system, the coupling ratio and axial load in the walls also were
From the pushover response, values for γ 1, γ 2 , and Ω0 were cal- investigated. These two parameters measure the behavior of the
culated, and are presented in Tables 9 and 10 for the CP and DP coupling beams relative to that of the rest of the structure. The cou-
models, respectively. These values were close to each other and to pling ratio represents the portion of the total overturning moment
the calculated design values [Eq. (4)] despite differences in that is resisted by the axial load in the walls applied by the coupling

© ASCE 04022022-10 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 Table 9. Capacity design overstrength values from analysis and calculations from CP model
Stories γ 1 , Eq. (4) γ 1 , analysis γ 2 , analysis Ω0 , analysis γ 1 ðanalysisÞ=γ 1 [Eq. (4)]
8 1.91 1.77 1.16 2.04 0.93
12 1.81 2.06 1.11 2.29 1.14
18 1.76 1.84 1.09 2.00 1.05
22 1.46 1.88 1.09 2.05 1.29

Table 10. Capacity design overstrength values from analysis and calculations from DP model
Stories γ 1 , Eq. (4) γ 1 , analysis γ 2 , analysis Ω0 , analysis γ 1 ðanalysisÞ=γ 1 [Eq. (4)]
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

8 1.91 1.57 1.36 2.14 0.82

12 1.81 1.71 1.45 2.49 0.93
18 1.76 1.77 1.26 2.24 1.01
22 1.46 2.05 1.13 2.31 1.38

Fig. 7. CP model pushover response for 8-story structure (a) moment versus roof displacement including axial couple contribution, tension wall, and
compression wall; (b) coupling ratio versus roof displacement; and (c) axial load versus average story drift considering Pw;nom and Pw;exp limit.

beam end shears. This behavior was investigated by examining the to 55% as the coupling beam contribution to the axial load leveled
results from the 8-story structure. Figs. 7(a and b) divides the total out and the moment in the individual walls continued to increase.
overturning moment into portions resisted by the individual walls, The capacity-limited axial load in the axial walls was evaluated
and the axial force couple caused the coupling beams. In Fig. 7(b), by comparing the maximum axial load in the wall (from the analy-
the contribution from the axial force couple is normalized by the sis) with the capacity limited load from the initial design. For the
total overturning moment to calculate the coupling ratio. Together, 8-story structure, the axial load in the walls varied as the displace-
these plots show that the coupling ratio varied with additional drift. ment increased [Fig. 7(c)]. This change occurs as all the coupling
Initially, the coupling ratio was about 65%, but this value decreased beams yield and undergo strain hardening. According to the capacity

© ASCE 04022022-11 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 Table 11. Comparison of capacity limited axial load and axial load seen in pushover analysis
Pw;nom [kN (kip)], Pw;exp [kN (kip)], Pw , analysis [kN (kip)], Pw , analysis [kN (kip)], Pw =Pw;nom , Pw =Pw;nom ,
Stories Eq. (13) Eq. (8) CP model DP model CP model DP model
8 19,700 (4,420) 22,000 (4,940) 19,900 (4,470) 20,600 (4,640) 1.01 1.05
12 25,900 (5,820) 29,200 (6,570) 27,700 (6,230) 29,500 (6,630) 1.07 1.14
18 74,800 (16,800) 83,500 (18,800) 72,500 (16,300) 78,300 (17,600) 0.97 1.05
22 100,000 (22,600) 112,000 (25,200) 99,200 (22,300) 108,000 (24,200) 0.99 1.07

Table 12. Ground motions used to analyze time-history response of structures

Earthquake Year Magnitude Recording station PEER NGA database information
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Superstition Hills 1987 6.5 El Centro Imp. Co. SUPERST/B-ICC090

Loma Prieta 1989 6.9 Gilroy Array #3 LOMAP/G03000
Northridge 1994 6.7 Canyon Country-WLC NORTHR/LOS270
Kobe 1995 6.9 Nishi-Akashi KOBE/NIS090

Fig. 8. (a) Superstition Hills acceleration versus time ground motion; and (b) finite-element (Abaqus) results for the 8-story structure subjected to the
design basis earthquake.

design philosophy, the maximum axial load in the wall can be esti- interstory drift ratio. The corresponding ground motion intensity
mated according to Eq. (8). However, because the analysis models was referred to as the failure-level earthquake. Table 12 lists the
used nominal material properties instead of expected values, the ground motions for which each structure was analyzed. These
Eq. (8) estimate for the axial load in the wall exceeded the actual ground motions were selected because they all represented strong,
load level from the analysis. If instead the axial load in the walls far-field motions of various durations, magnitudes, and locations.
was estimated using the nominal capacity of the coupling beams Although each structure was analyzed for all the ground motions,
[Eq. (13)], then it closely matched the maximum axial load in the following discussion focuses on the response of each structure
the walls. Table 11 compares this axial load level in the walls for for only one ground motion, because the observed responses were
both the CP and DP analysis models consistent.
First, the behavior of the 8-story structure was investigated using
2.4Mcb
p;nom the Abaqus finite-element model. This model was subjected to the
Pw;nom ¼ þ Pgravity ð13Þ
Lcb ground motions scaled to the spectral acceleration levels corre-
sponding to the design basis earthquake, maximum considered
earthquake, and failure-level earthquake. The analysis results for
Nonlinear Time-History Response the DBE-level earthquake are shown in Fig. 8, and a breakdown
Nonlinear time-history analysis was performed for the design basis of the milestones observed is shown in Fig. 9. Under the
earthquake (DBE), maximum considered earthquake, and failure- design-level earthquake, the structure experienced yielding in
level earthquake (FLE). This evaluation was performed to verify the coupling beams first, followed by yielding at the base of the
that the nonlinear dynamic response of the designed archetype gen- wall. Yielding spread along the height of the structure, with all cou-
erally was in accordance with the capacity design principle imple- pling beams undergoing some plastification. No fracture failures
mented while proportioning and designing the structure. Ground were observed. Similarly, for the maximum considered earthquake
motions were considered in one direction, and the record intensity (Figs. 10 and 11), the first milestone was coupling beam yielding,
was scaled up until failure occurred. Failure was defined conserva- followed by yielding in the walls. The propagation of yielding
tively as the point at which the structure reached and exceeded 5% (Fig. 11) was greater than the yielding observed for the DBE

© ASCE 04022022-12 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Fig. 9. Milestones observed in the DBE finite element analysis of the 8-story structure.

earthquake, but did not include any fracture failure. Finally, the
failure-level earthquake (with ground accelerations over 4 times
the DBE ground accelerations) caused the structure to experience
all anticipated milestones. These results are shown in Figs. 12
and 13. The coupling beams yielded, followed by yielding in
the wall, as observed in the DBE and MCE analysis. Next, fracture
initiated in the coupling beams, and finally, fracture occurred in the
walls. The progression of events generally was in accordance with
the capacity design principle, indicating that the structure was pro-
portioned properly to follow the prescribed mechanisms.
The CP and DP models in OpenSees were used to simulate the
seismic responses of all structures to the DBE, MCE, and FLE.
Similar to the analyses conducted using the Abaqus finite-element
models, the focus was on identifying performance milestones and
observing whether the nonlinear dynamic response was in accor-
dance with the capacity design principle. The results for the 8-story
Fig. 10. Finite-element (Abaqus) results for the 8-story structure sub-
structure are shown in Fig. 14 for both models. For the DBE,
jected to the maximum considered earthquake.
the first milestone was yielding of the first coupling beam.

Fig. 11. Milestones observed in MCE finite-element analysis of the 8-story structure.

© ASCE 04022022-13 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 This milestone was followed by the yielding of all coupling at slightly different times, except that initiation of wall yielding
beams. The final milestone was yielding of the wall. For both was not observed with the DP model. Rather, wall yielding with
models, the milestones observed in this analysis were similar to the DP model was observed first during the MCE analysis. The
those observed from the Abaqus analysis, although they occurred MCE analysis included an additional milestone—fracture in the
coupling beams for the CP model. This fracture was not seen in
the Abaqus model, which reiterates that the CP model is more
conservative than other models, including the finite-element
model. The FLE added failure of all coupling beams to the mile-
stones for both models. After the failure of the last coupling beam,
the structure appeared to have a significantly different effective
period, as indicated by the long period response during the second
half of the ground motion. This change in the effective period is
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

most evident by comparing the analysis results from the DBE and
MCE earthquakes after the first 20 s of the ground motion. The
FLE had a few high-amplitude cycles with longer duration than
the numerous lower amplitude cycles observed in the DBE and
MCE analysis.
The analysis results for the 12-, 18-, and 22-story structures
showed similar performance milestones and confirmed that the
nonlinear dynamic responses generally were in accordance with the
capacity design principle. These results are shown in Figs. 15–17
Fig. 12. Finite-element (Abaqus) results for the 8-story structure sub-
for both the CP and DP models. For the sake of brevity, only the
jected to the failure-level earthquake.
analysis results that deviated slightly from the expected progression

Fig. 13. Milestones observed in failure-level finite-element analysis of the 8-story structure.

© ASCE 04022022-14 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Fig. 14. (a) Superstition Hills acceleration versus time ground motion; (b) CP and DP model results for the 8-story structure subjected to the DBE; (c) CP
and DP model results for the 8-story structure subjected to the MCE; and (d) CP and DP model results for the 8-story structure subjected to the FLE.

Fig. 15. (a) Loma Prieta acceleration versus time ground motion; (b) CP and DP model results for the 12-story structure subjected to the DBE; (c) CP and
DP model results for the 12-story structure subjected to the MCE; and (d) CP and DP model results for the 12-story structure subjected to the FLE.

© ASCE 04022022-15 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

Fig. 16. (a) Northridge acceleration versus time ground motion; (b) CP and DP model results for the 18-story structure subjected to CP and DP model
results for the 18-story structure subjected to the DBE; (c) CP and DP model results for the 18-story structure subjected to the MCE; and (d) CP and
DP model results for the 18-story structure subjected to the FLE.

Fig. 17. (a) Kobe acceleration versus time ground motion; (b) CP and DP model results for 22-story structure subjected to the DBE; (c) CP and DP
model results for 22-story structure subjected to the MCE; and (d) CP and DP model results for 22-story structure subjected to the FLE.

© ASCE 04022022-16 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 of the milestones (first coupling beam yielding, final coupling beam AISC 341-16 (AISC 2016a). The capacity design philosophy for
yielding, wall yielding, first coupling beam fractures, and final cou- CC-PSW/CF system relies on a strong wall–weak coupling beam
pling beam fractures) are discussed here for these example struc- approach in which wall elements are sized to resist the capacity
tures. Of all the analyses, the MCE response for the 18-story limited loads from coupling beam elements.
structure, and the MCE and FLE responses for the 22-story struc- Nonlinear push-over analyses and time-history analyses of
ture deviated slightly from the expected progression of milestones. CC-PSW/CFs, using three different modeling approaches, consis-
For the CP model, the 18-story MCE response [Fig. 16(c)] tently showed a progression of milestones that met the intent of the
showed wall yielding preceding yielding in all coupling beams. capacity design philosophy. The preferred sequence of yielding in
Although the difference is observable on the plot due to the sharp coupling beams along the structure height, yielding in walls at the
increase in roof displacement between these two events, the time base, fracture in coupling beams, and fracture in walls generally
between the two events was a fraction of a second. Moreover, the was observed in a range of example structures considered, with pro-
DBE and FLE responses exhibited the expected progression of gressively more of these events happening up to the failure-level
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

milestone. Therefore, this variation from the anticipated failure earthquakes. These failure-level earthquakes, which are signifi-
mechanism is considered to be marginal. Nevertheless, as the height cantly more severe than the design basis or maximum considered
of the building increases, the milestone corresponding to initiation earthquakes, were defined here as those resulting in maximum in-
of yielding at the base of the wall may occur after formation of most terstory drifts of 5% for the selected ground motions considered,
plastic hinges in the coupling beams along the height of the struc- which conservatively was assumed to be failure for the purpose
ture, but not necessarily before the formation of plastic hinges in all of these evaluations. This performance confirms that the capacity
coupling beams along this height. The reasons for this change in design philosophy presented in this paper can lead to overall struc-
progression are described subsequently. ture or system design exhibiting a desirable ductile seismic perfor-
For the 22-story structure subjected to the MCE, coupling beam mance. As the structure height increases beyond the permitted
fracture occurred before all coupling beams yielded [Fig. 17(c)] in value 64 m (210 ft), yielding (not plastic hinging) may occur at
the CP model. This behavior occurred because of the conservative the base of the C-PSW/CF walls before the formation of plastic
concentrated plasticity hinge model (with rotation capacity of hinges in all coupling beams along the height of the structure. This
0.03 rad at fracture) for all the coupling beams, further exacerbated does not hamper the overall seismic performance or ductility of the
by higher-mode effects associated with the 22-story structure. The system.
assumption that all coupling beams along the height of the structure
engage equally and form plastic hinges at the ends becomes invalid
when the rotational demand on the critical coupling beam (i.e., the Data Availability Statement
first coupling beam to yield) exceeds the plastic rotation capacity
(0.03 rad) before hinging of all other coupling beams. Simply put, Some or all data, models, or code that support the findings of this
the plastic rotation capacity of the coupling beams is finite, and as study are available from the corresponding author upon reasonable
the height (and the number of coupling beams) of the structure in- request.
creases, at some point the critical coupling beams do not have
enough ductility for all other coupling beams to hinge before the
initiation of fracture in the critical coupling beam. For the 22-story Acknowledgments
structure subjected to the FLE [Fig. 17(d)] in the CP model, the
22-story structure experienced wall yielding slightly before yield- The project was supported by the Charles Pankow Foundation and
ing of all the coupling beams. However, the first four milestones for the American Institute of Steel Construction, through CPF research
this record occurred within less than 1 s. The wall yielding and Grant #05-17 awarded to Michel Bruneau from the University at
yielding of all coupling beam occurred less than 0.25 s apart. This Buffalo and Amit H. Varma from Purdue University. All opinions,
variation from the anticipated failure sequence is considered to be findings, conclusions, and recommendations presented in this pa-
marginal. per are those of the authors and do not necessarily reflect the view
For the DP models, responses followed the expected progres- of the sponsors. The researchers are grateful to members of the
sion of milestones and confirm that the nonlinear dynamic re- FEMA P695 peer review panel (Gregory G. Deierlein, Professor,
sponses generally were in accordance with the capacity design Stanford University; Ron Klemencic, Chairman and CEO, Mag-
principle. Differences in the time-history response between the two nusson Klemencic and Associates; and Rafael Sabelli, Principal
models became more accentuated for the taller buildings, which is and Director of Seismic Design, Walter P. Moore), and members
logical given that the models largely differed in how the coupling of the project advisory team (Larry Kruth, Vice President, AISC;
beams were modeled (i.e., the more stories, the larger was the num- John D. Hooper, Senor Principal/Director of Earthquake Engineer-
ber of coupling beams, and the greater was the divergence). In ing, MKA; Jim Malley, Senior Principal, Degenkolb Engineers;
addition, after the last coupling beam fractured, the DP model in- Bonnie Manley, Regional Director of Construction Codes and
dicated more displacement and permanent deformation than that in Standards, American Iron and Steel Institute; and Tom Sabol,
the CP model. Despite these differences, both models (CP and DP) Principal, Englekirk Institution) for their technical guidance.
showed that the structures could develop the maximum interstory
drift ratio of 5% for the FLE without collapse.
References
Conclusions About Rainier Square. n.d. “Rainier square.” Accessed July 1, 2021. https://
www.rainiersquare.com/about/.
CC-PSW/CF systems have emerged as a viable seismic force Agrawal, S., M. Broberg, and A. H. Varma. 2020. Seismic design coeffi-
resisting system for tall building construction. The capacity design cients for SpeedCore or composite plate shear walls—Concrete filled
procedure outlined in this paper is consistent with the philosophy (C-PSW/CF). Bowen Laboratory Research Reports. Paper 1. West Lafa-
adopted for other steel seismic force resisting systems included in yette, IN: Lyles School of Civil Engineering, Purdue Univ.

© ASCE 04022022-17 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022

 AISC. 2016a. Seismic provisions for structural steel buildings. ANSI/AISC Nie, J.-G., H.-S. Hu, and M. R. Eatherton. 2014. “Concrete filled steel plate
341-16. Chicago: AISC. composite coupling beams: Experimental study.” J. Constr. Steel Res.
AISC. 2016b. Specification for structural steel buildings. ANSI/AISC 94 (Mar): 49–63. https://doi.org/10.1016/j.jcsr.2013.10.024.
360-22. Chicago: AISC. Park, R., and T. Paulay. 1975. “Ductile reinforced concrete frames.”
AISC. 2018. Specification for safety-related steel structures for nuclear Bulletin of the New Zealand Society for Earthquake Engineering 8 (1):
facilities. ANSI/AISC N690-18. Chicago: AISC. 70–90.
AISC. 2022a. Seismic provisions for structural steel buildings: Public Post, N. M. 2020. “Rainier Square erector says speed core could rise even
review ballot. ANSI/AISC 341-22. Chicago: AISC. faster.” Accessed March 4, 2020. www.enr.com/articles/48817-rainier
AISC. 2022b. Specification for structural steel buildings: Public review -square-erector-says-speed-core-could-rise-even-faster.
ballot. ANSI/AISC 360-22. Chicago: AISC. SEAOC (Structural Engineers Association of California). 1980. Recom-
Alzeni, Y., and M. Bruneau. 2017. “In-plane cyclic testing of concrete filled mended lateral force requirements and commentary. Sacramento,
sandwich steel panel walls with and without boundary elements.” CA: SEAOC.
J. Struct. Eng. 143 (9): 04017115. https://doi.org/10.1061/(ASCE)ST Seo, J., A. H. Varma, K. Sener, and D. Ayhan. 2016. “Steel-plate composite
.1943-541X.0001791.
Downloaded from ascelibrary.org by University Of British Columbia on 02/12/22. Copyright ASCE. For personal use only; all rights reserved.

(SC) walls: In-plane shear behavior, database, and design.” J. Constr.

ASCE. 2017. Seismic rehabilitation of existing buildings. ASCE/SEI Steel Res. 119 (Mar): 202–215. https://doi.org/10.1016/j.jcsr.2015
41-17. Reston, VA: ASCE. .12.013.
ASCE. 2022. Minimum design loads for buildings and other structures: Shafaei, S. 2020. “Seismic and wind behavior of planar composite plate
Public review ballot. ASCE 7. Reston, VA: ASCE. shear walls concrete filled (C-PSWCF).” Ph.D. dissertation, Lyles
Booth, P. N., S. R. Bhardwaj, T.-C. Tseng, J. Seo, and A. H. Varma. 2020. School of Civil Engineering, Purdue Univ.
“Ultimate shear strength of steel-plate composite (SC) walls with boun- Shafaei, S., A. H. Varma, M. Broberg, and R. Klemencic. 2021a. “Modeling
dary elements.” J. Constr. Steel Res. 165 (Feb): 105810. https://doi.org the cyclic behavior of composite plate shear walls/concrete filled (C-
/10.1016/j.jcsr.2019.105810. PSW/CF).” J. Constr. Steel Res. 184 (2021): 106810. https://doi.org/10
Bruneau, M., C. Uang, and A. Whittaker. 2011. Ductile design of steel .1016/j.jcsr.2021.106810.
structures. 2nd ed. New York: McGraw-Hill.
Shafaei, S., A. H. Varma, J. Seo, and R. Klemencic. 2021b. “Cyclic lateral
Bruneau, M., A. H. Varma, E. Kizilarslan, M. Broberg, S. Shafaei, and
loading behavior of composite plate shear walls/concrete filled.” J.
J. Seo. 2019. R-factors for coupled composite plate shear walls—
Struct. Eng. 147 (10): 04021145. https://doi.org/10.1061/(ASCE)ST
Concrete filled (coupled-C-PSW/CF). Final Project Report, Charles
.1943-541X.0003091.
Pankow Foundation, CPF #06-16. McLean, VA: Charles Pankow
Tauberg, N., K. Kolozvari, and J. Wallace. 2019. Ductile reinforced concrete
Foundation.
coupled walls: FEMA P695 Study. Final Project Report, Submitted to the
FEMA. 2020. NEHRP recommended seismic provisions for new buildings
Charles Pankow Foundation for CPF research grant #06-17. McLean,
and other structures. Volume 1: Part 1 provisions, part 2 commentary.
VA: Charles Pankow Foundation.
FEMA P-2082-1. Washington, DC: Building Seismic Safety Council,
National Institute of Building Sciences. Uang, C.-M., and M. Bruneau. 2018. “State-of-the-art on seismic design of
Home. n.d. “200 Park Avenue San Jose.” Accessed July 1, 2021. https:// steel structures.” J. Struct. Eng. 144 (4): 03118002. https://doi.org/10
200parkavesanjose.com/. .1061/(ASCE)ST.1943-541X.0001973.
ICBO (International Conference of Building Officials). 1985. Uniform Varma, A. H., M. Ahmad, S. Shafaei, and T. Bradt. 2021. Seismic design
building code. Whittier, CA: ICBO. and behavior of composite coupling beam-to-C-PSW/CF connections.
Kenarangi, H., M. Bruneau, A. Varma, and M. Ahmad. 2021a. “Simplified Final Project Rep. for Research Grant #06-16. McLean, VA: Charles
equations for shear strength of composite concrete filled steel tubes.” Pankow Foundation and American Institute of Steel Construction.
Eng. J. 58 (3): 197–221. Varma, A. H., M. Broberg, S. Shafaei, and A. A. Taghipour. 2022. Design
Kenarangi, H., E. Kizilarslan, and M. Bruneau. 2021b. “Cyclic behavior of guide 37. Chicago: AISC.
C-shaped composite plate shear walls—Concrete filled.” Eng. Struct. J. Varma, A. H., S. Shafaei, and R. Klemencic. 2019. “Steel modules of
226 (1): 111306. https://doi.org/10.1016/j.engstruct.2020.111306. composite plate shear walls: Behavior, stability and design.” Thin-
Kizilarslan, E., M. Broberg, S. Shafaei, A. H. Varma, and M. Bruneau. Walled Struct. 145 (Dec): 106384. https://doi.org/10.1016/j.tws.2019
2021a. “Non-linear analysis models for composite plate shear walls- .106384.
concrete filled (C-PSW/CF).” J. Constr. Steel Res. 184 (Sep): 106803. Varma, A. H., K. Zhang, H. Chi, P. N. Booth, and T. Baker. 2011. “In-plane
https://doi.org/10.1016/j.jcsr.2021.106803. shear behavior of SC composite walls: Theory vs. experiment.”
Kizilarslan, E., M. Broberg, S. Shafaei, A. H. Varma, and M. Bruneau. Accessed July 1, 2021. http://engineering.purdue.edu/∼ahvarma
2021b. “Seismic design coefficients and factors for coupled composite /Publications/p764.pdf.
plate shear walls/ concrete filled (CC-PSW/CF).” Eng. Struct. 244 (Oct): Zhang, K., J. Seo, and A. H. Varma. 2020. “Steel-plate composite (SC)
112766. https://doi.org/10.1016/j.engstruct.2021.112766. walls: Local buckling and design for axial compression.” J. Struct. Eng.
Kurt, E. G., A. H. Varma, P. Booth, and A. S. Whittaker. 2016. “In-plane 146 (4): 04020044. https://doi.org/10.1061/(ASCE)ST.1943-541X
behavior and design of rectangular SC wall piers without boundary .0002545.
elements.” J. Struct. Eng. 142 (6): 04016026. https://doi.org/10.1061 Zhang, K., A. H. Varma, S. R. Malushte, and S. Gallocher. 2014. “Effect of
/(ASCE)ST.1943-541X.0001481. shear connectors on local buckling and composite action in steel con-
Morgen, B., R. Klemencic, and A. H. Varma. 2018. Core solution— crete composite walls.” Nucl. Eng. Des. 269 (Apr): 231–239. https://
Modern steel construction. Chicago: AISC. doi.org/10.1016/j.nucengdes.2013.08.035.

© ASCE 04022022-18 J. Struct. Eng.

J. Struct. Eng., 2022, 148(4): 04022022