International Journal of Concrete Structures and Materials

Vol.6, No.1, pp.3~18, March 2012

DOI 10.1007/s40069-012-0001-4
ISSN 1976-0485 / eISSN 2234-1315

Behavior, Design, and Modeling of Structural Walls and

Coupling Beams
Lessons from Recent Laboratory Tests and Earthquakes
John W. Wallace
(Received February 13, 2012, Revised February 20, 2012, Accepted February 20, 2012)

Abstract: Observed wall damage in recent earthquakes in Chile and New Zealand, where modern building codes exist, exceeded
expectations. In these earthquakes, structural wall damage included boundary crushing, reinforcement fracture, and global wall
buckling. Recent laboratory tests also have demonstrated inadequate performance in some cases, indicating a need to review code
provisions, identify shortcomings and make necessary revisions. Current modeling approaches used for slender structural walls adequately capture nonlinear flexural behavior; however, strength loss due to buckling of reinforcement and nonlinear and shear-flexure interaction are not adequately captured. Additional research is needed to address these issues. Recent tests of reinforced
concrete coupling beams indicate that diagonally-reinforced beams detailed according to ACI 318-111 can sustain plastic rotations
of about 6% prior to significant strength loss and that relatively simple modeling approaches in commercially available computer
programs are capable of capturing the observed responses. Tests of conventionally-reinforced beams indicate less energy dissipation
capacity and strength loss at approximately 4% rotation.
Keywords: testing, structural wall, coupling beam, modeling, detailing.

1. Introduction
Design and construction practice for special structural walls
(ACI 318 designation) has evolved significantly since the system
was introduced in the 1970s. Throughout the 1970s and 1980s, it
was common to use so-called barbell-shaped wall cross sections,
where a column was used at each wall boundary to resist axial
load and overturning, along with a narrow wall web. In the late
1980s and early 1990s, use of rectangular wall cross sections
became common to produce more economical designs. Use of
walls with rectangular cross sections is common in many countries, including Chile and New Zealand. Although use of walls
with boundary columns is still common in Japan, based on information available in the literature, the AIJ Standard for Structural
Calculations of Reinforced Concrete Buildings was revised in
2010 to show RC walls with rectangular cross-sections. Engineers
around the world have pushed design limits in recent years, optimizing economy and design, and in many practices producing
walls with higher demands and more slender profiles than have
been verified in past laboratory testing or field experience. The
trend towards more slender profiles has been accelerated by use of
higher concrete strengths.
1)

Department of Civil and Environmental and Environmental

Engineering, University of California, Los Angeles, Los Angeles, CA 90095, USA, *Corresponding Author; E-mail: wallacej@
ucla.edu
Copyright 2012, Korea Concrete Institute. All rights reserved,
including the making of copies without the written permission of
the copyright proprietors.

Observed wall damage in recent earthquakes in Chile (2010)

and New Zealand (2011), where modern building codes exist,
exceeded expectations. In these earthquakes, structural wall damage included boundary crushing, reinforcement fracture, and global wall buckling. Recent tests of isolated structural walls in the
US and tests of two, full-scale, 4-story buildings with high-ductility structural walls at E-Defense in December 2010 provide vital
new data. A particularly noteworthy aspect of these recent tests is
the failure of relatively thin wall boundaries to develop ductile
behavior in compression, even though they complied with building code provisions and recommendations of ACI and AIJ.
The observed performance following recent earthquakes and in
recent laboratory tests suggests strongly that the problems
observed are not isolated and that analysis and design provisions
need to be reassessed. In particular, the quantity and configuration
of transverse reinforcement required at wall boundaries needs to
be reassessed to address issues associated with wall thickness,
slenderness, axial load, and configuration, as well as expected displacement demands and load history. Preliminary studies indicate
that greater amounts of transverse reinforcement may be required
for thin walls or walls with large cover and that tighter spacing of
transverse reinforcement may be required to suppress buckling of
vertical reinforcement, especially for walls with light axial load or
walls with flanges. These issues apply to both high ductility (ACI
Special) and moderate ductility (ACI Ordinary) walls.
The observed wall performance also raises important questions
and challenges related to nonlinear modeling of structural walls
and coupling beams, commonly accomplished using either beamcolumn models with plastic hinges or fiber models with uniaxial
3

material relationships. Beam-column element models with plastic

hinges are simple and provide reasonably good estimates of global
and average local responses; however, they have various drawbacks, such as accounting for migration of the neutral axis, incorporating in- and out-of-plane coupling, and accounting for
2
stiffness variation with axial load. Fiber and fiber-type models,
such as the multiple-vertical-line-element model, where flexural
response is simulated by a series of uniaxial elements (or macrofibers) along with the assumption that plane sections remain plane
after loading, address these shortcomings and provide a better
framework for incorporating more complex behaviors. However,
fiber models also have drawbacks, such as added complexity, convergence issues, and results that are sensitive to meshing. More
complex modeling approaches based on multi-axial material models are generally not used for design, and are not addressed here.
Fiber and beam-column models have been incorporated into
research oriented programs such as opensees (2009) and wall as
practice-oriented programs used for performance-based design
such as CSI Perform 3D76. Considerable effort has focused on val2-6
idating and calibrating these models for axial-flexural behavior,
7
8
9-11
shear behavior, anchorage/splice behavior, and axial failure.
More Recent research has focused on accounting for interaction
12-15
(or coupling) between axial-flexural and shear responses,
with
various modeling approaches proposed, e.g., fiber/section based
models,7,16-18 strut models,19 and simplified models using analyti20
21
cal or experimental results. Wall test programs focused on providing data for validation of shear-flexure interaction models for
intermediate wall aspect ratios have recently been completed.22
Various testing programs have been carried out to assess the
load deformation behavior of coupling beams.23-30 Primary test
variables in these studies were the ratio of the beam clear span to
the beam total depth (commonly referred to as the beam aspect
ratio) and the arrangement of the beam reinforcement. In a majority of these studies, the load deformation behavior of low-aspect
ratio beams (1.0 to 1.5) constructed with beam top and bottom
longitudinal reinforcement were compared with beams constructed with diagonal reinforcement. Concrete compressive
strengths for most tests were around 4 ksi (~25 to 30 MPa).
Although these tests provided valuable information, they do not
address issues for current tall building construction, where beam
aspect ratios are typically between 2.0 and 3.5 and concrete
strengths are in the range of 6 to 8 ksi (~40 to 55 MPa). In addition, in none of the prior studies was a slab included as part of the
test specimen; whereas the slab might restrain axial elongations
and impact stiffness, strength, and deformation capacity. Recent
studies31-33 address many of these issues.
Nonlinear modeling of coupling beams has become important
as the use of coupled core wall systems have become more common.34,35 For coupling beams, important modeling parameters
include effective bending stiffness EcIeff , allowable plastic rotation
prior to significant lateral strength degradation, and residual
strength. The effective bending stiffness for beams in ASCE 41-06
Table 6-4 was reduced to 0.3EcIg to account for the added flexibility due to reinforcement slip/extension;36 however, modeling
parameters in Table 6-18 for RC coupling beams were not
changed. Verifying that the relatively simple modeling approaches
commonly used for design adequately capture coupling beam load
- deformation responses, as well as recommending parameters

associated with unloading/reloading and pinching behavior, are

important issues that have not been adequately investigated.
Given this background, the objectives of this paper are to review
current wall and coupling beam test results and to identify issues
that are not adequately addressed, both in terms of code design
provisions and nonlinear modeling.

2. Observed performance of structual walls &

coupling beams
2.1 Recent earthquake reconnaissance
Recent earthquakes in Chile (Mw 8.8, February 2010), New
Zealand (February 2011, ML=6.3), and Japan (Mw 9.0, March
2011) have provided a wealth of new data on the performance of
modern buildings that utilize structural walls for the primary lateral-force-resisting system. Although complete building collapse
was rarely observed, damage was widespread and generally
exceeded expectations.
In 1996, Chile adopted a new code (NCh 433. of 96)37 based on
38
ACI 318-95 and produced an immense inventory of progressively more slender buildings corresponding essentially to the US
reinforced concrete code provisions, except boundary element
confinement was not required. The 2010 Mw 8.8 earthquake
caused serious damage to many of these buildings, including
crushing/spalling of concrete and buckling of vertical reinforcement, often over a large horizontal extent of the wall (Fig. 1).
Damage tended to concentrate over a relatively short height of one
to three times the wall thickness, apparently because buckling of
vertical bars led to concentration of damage. Closer inspection of
the wall boundary regions (Fig. 1) revealed the relatively large
spacing of hoops (20 cm) and horizontal web reinforcement
(20 cm), as well as the 90-degree hooks used on hoops and horizontal web reinforcement, which may have opened due to concrete crushing and/or buckling of vertical reinforcement (Fig.
1(d)). Some of the failures are attributable to lack of closelyspaced transverse reinforcement at wall boundaries, which was
not required by the Chilean code based on the good performance
of buildings in the 1985 M7.8 earthquake; however, many of the
failures are not yet understood, and many suggest that there are
deficiencies in current US design provisions.39,40 In some cases,
lateral instability (buckling) of a large portion of a wall section
was observed (Fig. 2); prior to the Chile and New Zealand earthquakes, this global buckling failure had been primarily observed in
41
42
laboratory tests. Detailed surveys conducted as part of ATC-94
indicate that global wall buckling was not driven by prior yielding
in tension (as had originally been suspected based on past
43-45
research ) but instead was the result of lateral instability of previously crushed boundary zones. Furthermore, the ATC-9442
study has been unable to establish through analysis the role of preemptive longitudinal bar buckling as a trigger for compression
failure of lightly confined boundary zones. Laboratory testing is
required to understand these behaviors; preliminary studies are
underway in Chile and the US to investigate these issues.
The 2011 Christchurch earthquake46,47 shows many similar wall
failures, suggesting the deficiencies observed in the 2010 Chile
earthquake are not isolated (Fig. 3(a)). All of the walls depicted in
Figs. 2 and 3 have either T-shaped (Figs. 2, 3(b)) or L-shaped (Fig.
3(a)) cross sections, which lead to large cyclic tension and com-

4International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)

Fig. 1 Typical wall damage in Chile earthquake.

Fig. 2 Wall lateral instability.

48

pressive demands at the wall web boundary. The wall web

boundaries are susceptible to out-of-plane buckling following
1
cover concrete spalling. Although current ACI 318-11 provisions
require consideration of an effective flange width, the provisions
do not restrict use of narrow walls and do not address this out-ofplane failure mode, i.e., there are no restrictions on wall thickness
or wall slenderness. Failures of diaphragm-to-wall connections
were observed in Christchurch, potentially contributing to the col49
lapse of the several buildings. In Chile, typical buildings have a
large number of walls that well-distributed in plane; therefore, diaphragm failures were not observed.

2.2 Recent laboratory studies of conventional

walls
Recent laboratory testing of structural walls in the US has
focused on addressing concerns related to behavior of walls with
rectangular and T-shaped cross sections subjected to uniaxial and
50-52
biaxial loading,
walls with couplers and splices in the plastic
53,54
hinge region,
walls with higher shear demands,54-56 and walls
with coupling beams.32,33,57 All of these studies involved quasi-

Fig. 3(a) Wall failure in 2011 Christchurch earthquake.49

Fig. 3(b) Specimen TW2 web boundary failure.41

static testing. Shake table testing of walls has been limited, except
for 7-story building slice tests of walls with rectangular and Tshaped cross sections conducted by Panagiotou and Restrepo.58
The overwhelming majority of quasi-static and shake table tests
conducted in Japan have been conducted on barbell-shaped walls
and low-rise buildings with wing walls,59-61 which are not common in the US. Only recently have the Japanese Building Standard Law and Architectural Institute of Japan recommendations
been modified to allow the use of rectangular walls with boundary
elements, but their use is not widespread.
Johnson53 reports test results of isolated, slender (hw/lw and Mu/
Vulw=2.67) cantilever walls to investigate the behavior of anchorage details for flexural reinforcement. Three walls were tested, one
each with continuous (RWN), coupled (RWC), and spliced
(RWS) vertical reinforcement. The wall cross sections were 6
in. 90 in. (152.4 mm 2.29 m), and the walls were subjected to
horizontal lateral load approximately 20 ft or 6.1m above the base.
Although the wall cross-sections were rectangular, different
amounts of boundary vertical reinforcement were used to simulate
the behavior of T-shaped wall cross sections; 4-#6 (db=19 mm)
and 2-#5 (db=15.9 mm) at one boundary and 8-#9 (db=28.7 mm)
at the other boundary. Horizontal wall web reinforcement, of #3
@7.5 in. or t=0.0049 (db=9.5 mm @ 19 cm), was selected to
resist the shear associated with the expected moment strength
(including overstrength). Wall web vertical reinforcement consisted of #4 @18 in. or v=0.0037 (db=12.7 mm @ 45.7 cm). It is
noted that the 18 in. (45.7 cm) spacing of vertical web reinforcement is the maximum spacing allowed by ACI 318-11 21.9.2.1. It
is questionable whether such a large spacing (45.7 cm) in such a
thin wall (15.2 cm), satisfies the intent of R21.9.4, which states
that wall we reinforcement should be appropriately distributed
along the length and height of the wall... should be uniform and at
a small spacing. Lateral load versus top lateral displacement relations for RWC and RWS are plotted in Fig. 4(a); since results for
RWC and RWN are very similar. For RWC, the wall reached rotations exceeding +0.035 (#5 in tension) and 0.02 (#9 in tension),
whereas for RWS, the wall reached rotations of approximately
+0.02 (#5 in tension) and 0.012 (#9 in tension). Damage was
concentrated at a single, large crack at the foundation-wall inter-

International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)5

Fig. 4(a) Load-displacement relations.

Fig. 5(a) NEESR UW wall tests.

Fig. 5(b) Base shear vs drift.

Fig. 4(b) Wall damage at end of test (RWS).

face, which accounted for about 0.015 of the top rotation of 0.02.
It is noted that the applied shear is close to or exceeds the web
shear friction capacity Vn of the walls, depending on the direction
of the applied load and the value assumed for the coefficient of
friction. Significant horizontal cracking also was observed for
specimens RWN and RWC, suggesting that the quantity (and
large spacing) vertical web reinforcement was insufficient to
restrain sliding between the wall boundaries. Damage concentrated at the foundation-wall interface for specimen RWS (Fig.
4(b)). However, the test results do indicate adequate performance
in the case of the coupler and that the presence of the splice significantly reduced the wall lateral deformation capacity.
Tests of walls with splices also were conducted by Birely et al.54
The test specimens were roughly one-half scale replicas of the bottom three stories of a ten-story wall (Fig. 5(a)). Base shear versus
3rd story (top) displacement plots are shown in Fig. 5(b) for three
of the tests, PW1 (splice, Mb=0.71hwVb), W2 (splice, Mb=
0.50hwVb), and W4 (no splice, Mb=0.50hwVb). Design wall shear
stresses were 0.23, 0.33, and 0.33 f c MPa MPa for W1, W2, and
W4, respectively (equivalent to 0.7, 0.9, and 0.9Vn). The #4
(db=12.7 mm) boundary bars were lapped 0.61m, with spacing of
boundary transverse reinforcement of 51 mm (s/db=4). The test

with lower shear stress was reasonably ductile, achieving 1.08Mn

and a 3rd story lateral drift of 1.5% prior to strength loss; however,
test PW4, with no splice, reached only 1.0% lateral drift at the
third story (top) prior to strength loss. For all tests with splices,
damage initiated with buckling of the interior bar at the wall edge
(Fig. 6(a)) and then concentrated at the top of the splices (Fig.
6(b)), whereas damage was concentrated at the foundation-wall
interface for test PW4 with no splice (Fig. 6(c)). Even without
consideration of the elastic deformations over the top seven stories
not included in the test, deformation capacities of the walls are less
than expected, especially for PW4, with no splice.
Nagae et al.62 summaries important details for NIED (EDefense) tests on two 4-story buildings, one conventionally reinforced and the other using high-performance RC construction,
both with rectangular wall cross sections (Fig. 7a). The conventionally reinforced wall had confinement exceeding US requirements, with axial load of approximately 0.03Agf'c, yet the
compression boundary zone sustained localized crushing and lateral buckling (Fig. 7(b), following Kobe 100% motion). The base
overturning moment versus roof displacement responses are plotted in Fig. 8; base rotations are slightly less than the roof drift ratio
(e.g., for Kobe 100%, the base rotation measured over 0.27lw is a
little more than 0.02). Following crushing of boundary regions,
sliding shear responses increased substantially during the Kobe

6International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)

Fig. 6 Wall damage: (a) PW2 @ 1.0% drift; (b) PW2 end of test; (c) PW4 @ 1.0% drift.

Fig. 7(a) RC conventional wall.62

100% test (Fig. 8). Sliding displacements in the Takatori 60% test
reached the limits of the sensor, +45 mm and 60 mm with peak
shear of +/ 2000 kN. It is noted that the relatively large clear
cover over the boundary longitudinal bars was used (~40 mm) and
the boundary transverse reinforcement was insufficient to maintain the boundary compressive load following cover spalling. It is
noted that the crushing/spalling of the boundary region was
accompanied by lateral buckling of the compression zone, as was
observed in Chile and New Zealand (Fig. 2). It is yet unclear what
role biaxial loading had on the observed wall damage, this issue is
still being studied; however, it is plausible that the susceptibility of
the wall to lateral instability was impacted by biaxial loading.
The pre-NEESR tests conducted at NEES@Minnesota51,52,63
studied the role of biaxial loading by subjecting cantilever walls
with T-shaped cross sections to biaxial loading and comparing

Fig. 7(b) Wall damage.

41
their results with similar tests subjected to in-plane loading. The
6 in (152.4 mm) thick walls exhibited rotations over the first story
(hs=0.8lw) of approximately 0.02 prior to lateral strength degradation. Their findings suggest that analytical models validated previously for in-plane loading of walls adequately captured the
measured responses for combined in- and out-of-plane loading.
However, based on video and post-test observations, damage at
wall boundaries of the conventional reinforced concrete building
tested on the E-Defense shaking table may have been influenced

Fig. 8 RC conventional building responses (structural wall direction).

International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)7

by simultaneous in-plane and out-of-plane responses. The New

Zealand Royal Commission report47 raises the issue of biaxial
loading as a possible contributing factor to the unexpected wall
damage in the February 2011 earthquake. This issue has not been
adequately studied, and the issue is complicated by the observation that out-of-plane failures are observed at wall boundaries for
in-plane loads alone.

2.3 Recorded ground motions

Response Spectra computed using ground motions recorded in
recent earthquakes have significantly exceeded values used for
64
design. For example, spectra for records in Chile and
49
Christchurch significantly exceed values used for design (Fig.
11). For Chile, many buildings are designed for the Soil II spectrum, whereas spectral ordinates are generally 2 to 6 times the values for Soil II over a broad period range. Given such large
demands it is important to re-evaluate how displacement demands
influence design requirements for structural walls.

2.4 Coupling beam tests

Recent tests of eight one-half scale coupling beams focused on
assessing detailing and modeling parameters for coupling beam
configurations common for taller buildings, including the influence of reinforced and post-tensioned slabs. A brief summary of
these studies is presented here, with more information available in
31
65
Naish and Naish et al. Beams with transverse reinforcement
provided around the bundles of diagonal bars (referred to as diagonal confinement) were designed according to ACI 318-05
S21.7.7.4, whereas beams with transverse reinforcement provided
around the entire beam cross section (referred to as full section
confinement) were designed according to ACI 318-08 S21.9.7.4

(d). Three test specimens with aspect ratio of 2.4 were constructed
with 4 (101.6 mm)-thick slabs. CB24F-RC contained a slab reinforced with #3 bars @12 spacing (db=9.5 mm @ 304.8 mm), on
the top and bottom in the transverse direction, and on the top only
in the longitudinal direction, without post-tensioning strands.
CB24F-PT and CB24F-1/2-PT both contained a similar reinforced-concrete slab, but also were reinforced with 3/8" (9.5 mm)
7-wire strands.
Load-deformation responses of CB24F and CB24D are very
similar over the full range of applied rotations (Fig. 12(a)); similar
results were obtained for 3.33 aspect ratio tests. Notably, both
beams achieve large rotation (~8%) without significant degradation in the lateral load carrying capacity, and the beams achieve
shear strengths of 1.25 and 1.17 times the ACI nominal strength.
The shear strength of CB24D degraded rapidly at around 8% rotation, whereas CB24F degraded more gradually, maintaining a
residual shear capacity of ~80% at rotations exceeding 10%. The
test results indicate that the full section confinement option of ACI
318-08 provides equivalent, if not improved performance, compared to confinement around the diagonals per ACI 318-05. Diagonal crack widths for the full section confinement were generally
less than for diagonal confinement.
Four beams with aspect ratio of 2.4 were tested to assess the
impact of a slab on load-deformation responses. CB24F did not
include a slab, whereas CB24F-RC included an RC slab, and
CB24F-PT and CB24F-1/2-PT included PT slabs (with 150 psi
(1.03 MPa) of prestress). Load-displacement responses of CB24FRC vs. CB24F-PT are compared in Fig. 12(b). The plots reveal
that the slab increases the shear strength; however, this strength
increase can be accounted for by considering the increase in nominal moment strength due to the presence of the slab and the pre-

Fig. 10 Load vs displacement relations: (a) web direction; (b) Flange direction.

Fig. 11 Spectra from recent large earthquakes.

8International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)

63

Fig. 12 Load displacement relations for coupling beams without (a) and with (b) slabs.

stress. The peak loads for beams CB24F-RC vs. CB24F-PT correspond to shear stresses of 13.0 f c A cw psi (1.08 f c Acw MPa)
and CB24F-RC 11.8 f c Acw psi ( 0.98 f c Acw MPa ) , respectively. The presence of a slab (RC or PT) restrains axial growth
prior to yield, leading to modestly higher stiffness; however, the
secant stiffness values following yield for beams with and without
slabs are very similar and significant strength degradation for all
beams occurs at approximately the same rotation (8%). This
increase in strength is primarily due to the axial force applied to
the specimen by the tensioned strands, and increased the nominal
moment strength. Between 8% and 10% rotation, strength degradation is more pronounced for CB24F-PT than CB24F-RC, with
30% reduction for CB24F-PT vs. 10% for CB24F-RC, possibly
due to the presence of pre-compression.
A 3.33 aspect ratio beam with longitudinal beam reinforcement,
referred to as a Frame Beam or FB33, was tested to assess the
impact of providing straight bars as flexural reinforcement instead
of diagonal bars in beams with relatively low shear stress demand
(< 4.0 f c psi; 0.33 f c MPa). A plot of load vs. deformation for
FB33 (Fig. 13(a)) indicates that plastic rotations greater than 4%
can be reached prior to strength degradation. These results correspond well with prior test results27 (Fig. 13(b)) on similarly sized
beams, which achieved maximum shear stresses of about 4.7 f c
(0.39 f c MPa) and plastic chord rotations greater than 3.5%.
Compared to a similar beam with diagonal reinforcement and fullsection confinement (CB33F), or diagonal confinement (CB33D),
FB33 experiences more pinching in the load-deformation plot,
indicating that less energy is dissipated. As well, the beams with
diagonal reinforcement exhibited higher ductility, reaching plastic
rotations exceeding 7% prior to strength degradation, versus
approximately 4% for frame beams. The results indicate that use
of longitudinal reinforcement for coupling beams, which are much
easier to construct, is appropriate provided shear stress demands
are less than approximately 5.0 f c (0.42 f c MPa) and total
rotation demands are less than approximately 4%.

2.5 Summary
Wall performance in recent earthquakes and laboratory tests
raises a number of design concerns. In Chile, brittle failures at wall
boundaries were likely influenced by the level of axial stress (possibly leading to compression failures), the larger than expected displacement demands, the use of unsymmetric (or flanged) wall
cross sections, and the lack of closely-spaced transverse reinforcement at wall boundaries. A particularly noteworthy aspect of

57,62,66

recent tests
is the failure of relatively thin wall boundaries to
develop ductile behavior in compression, even though they complied with ACI 318 special boundary element requirements, as
well as Japan Standard Building Law and AIJ (2010) requirements. Recent tests to investigate the role of splices within the
plastic hinge region of structural walls suggest that splices will
substantially reduce wall inelastic deformation capacity. Given
these observations, current ACI 318-111 code provisions for Special Structural Walls are reviewed to identify possible concerns
and to suggest changes that could be implemented to address these
concerns.
Results from recent tests on diagonally- and longitudinally-reinforced coupling beams provide valuable new data to assess stiffness, detailing, and modeling requirements. The tests indicate that
full section confinement is as effective as diagonal confinement,
slab impacts on stiffness and nominal strength are modest, and
beams with longitudinal reinforcement exhibit less energy dissipation and total rotation capacity compared to beams with diagonal
reinforcement. New detailing provisions in ACI 318-08 were
introduced based, in-part, on these test results.

3. ACI 318 Chapter 21 provisions for special

structural walls & coupling beams
Provisions for Special Structural Walls are contained in ACI
318-11 21.9 and include provisions for Reinforcement (21.9.2),
Shear Strength (21.9.4), Design for Flexural and Axial Loads
(21.9.5), and Boundary Elements of Special Structural Walls
(21.9.6). In light of the preceeding discussion, key aspects of these
provisions are reviewed and areas of concern are noted. In many
cases, insufficient information is available to develop comprehensive requirements and comments provided here are meant to
inform.

3.1 Reinforcement and splices

A single curtain of web reinforcement is allowed if wall shear
stress is less than 0.17 f c M Pa MPa. This provision is acceptable
for squat walls with low shear stress (e.g., walls with aspect ratio
less than 1.5); however, for slender walls where buckling of
boundary vertical reinforcement and lateral instability are more
likely due to significant tensile yielding of reinforcement under
cyclic loading, two curtains should always be used. This recommendation applies to both Special Structural Walls (high ductility)
and Ordinary Structural Walls (moderate ductility).

International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)9

Fig. 13 Load - displacement relations for frame beams.

0.003l w
lw
lw
---------------------------c limit = ---------------------- = ---------------------------2 ( u hw ) 667 ( u h w ) 600 ( u h w )

(2)

In this approach, it is obvious that the result is sensitive to the

values used for the design displacement and the plastic hinge
length. Revised formulations, using a detailed displacement-based
67
design approach and a plastic hinge length that varies with wall
thickness (lp=atw as suggested by Wallace,39 produces the following more comprehensive relation;
u
sy
t l
tw +
11 hw t w + 2 tw t w
------ = cu ---w- ---w- 1 --- ----- ---------------------- ------ ------------- ----
l c
hw
c
2 hw ( 1 /l w) 40 lw
lw
hw lw
w

Fig. 14 ACI 318-11 21.9.6.2 model.

Recent laboratory tests have identified that wall deformation

capacity may be compromised in cases where splices exist within
the wall critical section (plastic hinge) because nonlinear deformations are concentrated outside of the splice region, either at the
wall-foundation interface (large moment gradient)53 or above the
splice (nearly uniform wall moment).54 Given these results, it is
questionable whether boundary vertical reinforcement should be
lapped spliced within the plastic hinge region. Test results did indicate that use of ACI 318-11 Type II couplers performed adequately. The option of staggering splices is not addressed here.

3.2 Design displacement and plastic hinge

length
The model used to develop ACI 318-11 21.9.6.2 provisions is
shown in Fig. 14. Given this model, the design displacement
u(ACI) x=Cde /I(ASCE 7) is related to local plastic hinge rotation p and extreme fiber compressive strain c as:
u

c
p = ------ ; p = u = ----c lp = ---w- c = 2 -----u- ---hw
c
2
hw lw

(1)

Where lp is the plastic hinge length, hw is the wall height, c is the

neutral axis depth for (Mn, Pu,max ), and lw is the wall length. If the
compressive strain exceeds a limiting value, typically taken as
0.003, then special transverse reinforcement is required. In ACI
318-11 Equation (21-8), Equation (1) is rearranged to define a limiting neutral axis depth versus a limiting concrete compressive
strain as:

(3)
where tw is the wall thickness, and sy is the tensile reinforcement
yield strain. The constant 11/40 results based on the assumed dis68
tribution of lateral force over the height of the wall. Using Eq.
(3), the relationship between the wall neutral axis depth, concrete
compressive strain, and drift is computed for various ratios of lw/tw
and hw/lw and plastic hinge length. For this preliminary study, wall
aspect ratio hw/lw is set to 3.0 and the ratio of lw/tw is set to 13.3,
which is fairly typical for U.S. construction. Concrete compressive
strain is set to 0.003; results presented in Fig. 15 for three values of
(2, 6, 12). For the ratio of lw/tw selected (13.33), =6 is equivalent to lp=0.45lw, or about the same value of 0.5lw assumed in the
development of ACI 318-11 relations in Eq. (2). Special transverse reinforcement is required at wall boundaries for values
above and to the right of the lines.
According to Fig. 15, if the drift ratio is 0.01, the neutral axis
must exceed 0.17lw before SBEs are required by ACI 318-11.
However, for the same neutral axis depth of 0.17lw, if inelastic
deformations are concentrated over a short height (lp=(=2)tw),
only less than one-half of this drift ratio (0.005), can be tolerated
before SBEs are required. The sensitivity of the results suggests
that measures are needed to ensure appropriate spread of plasticity
by requiring walls to be tension-controlled or by ductile yielding
of concrete in compression for compression-controlled walls.
1
These issues are not currently addressed in ACI 318-11 .
In current US codes the intent is to provide 90% confidence of
non-collapse for MCE shaking. In contrast, the current ACI confinement trigger (Eq. 2) is based on 50% confidence of not
exceeding the concrete crushing limit in the Design Basis Earthquake (which is much lower shaking intensity than the MCE). To
address this issue, it is necessary to adjust ACI Equation (21-8),
also Eq. (2) in this paper, to be more consistent with the building
code performance intent. Three factors need to be considered: 1)

10International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)

prudent to limit the nonlinear deformations. This objective can be

accomplished by calculating a limiting drift ratio for a given limit
on concrete compressive strain. For an assumed neutral axis depth
c=0.6lw (for balanced failure), a limiting compression strain of
0.01, Eq. (1) gives: u/hw < 0.010/(2)(0.6)=0.0083. Given the simplifying assumptions associated with Eq. (1), a slightly higher drift
limit might be appropriate (e.g., u/hw < 0.01).

3.4 Boundary element detailing

Fig. 15 Influence of plastic hinge length on need for SBEs.

MCE exceeds DBE. 2) There is dispersion about the median

response. 3) Damping is likely to be lower than the 5% value
assumed in the ACI provisions. To address these issues, the coeffi1
cient of 600 in the denominator of Equation (21-8) in ACI 318-11
should be increased by a factor of approximately 1.5 to adjust to
MCE level shaking and to consider dispersion, and by approximately 1.2 to 1.3 to account for potential lower damping ratios;
therefore, a coefficient of 1000 to 1200 should be used as recently
recommended in the NIST Technical Brief No. 6.69

3.3 Axial load and compression-controlled walls

As noted above, the provisions of 318-11 21.9.6.2 assume that
nonlinear deformations within the critical (plastic hinge) region of
the wall will spread out over a distance equal to one half the member depth. ACI 318-11 9.4 defines tension- and compressioncontrolled sections; however, no guidance is provided on how
these requirements should be applied to special (or ordinary) structural walls. In addition, ACI 318 and ASCE 7 do not place limits
on wall axial stress. The performance of walls in Chile suggests
that higher axial stresses and wall cross section shape (e.g., Tshaped) may lead to cases where concrete compressive strain
reaches 0.003 prior to yield of tension steel.
Various approaches could be used to address this issue, such as
placing limit on axial stress or requiring wall critical sections to be
tension-controlled. In the 1997 version of the Uniform Building
Code,70 wall axial load was limited to 0.35P0; for higher axial
loads the lateral strength and stiffness of the wall could not be considered. An alternative to neglecting the lateral-force-resistance of
compression-controlled walls would be to impose more stringent
design requirements, such as always requiring Special Boundary
Elements (SBEs) for wall critical sections that are not tension-controlled according to ACI 318-11 9.4, where a section is tensioncontrolled if the reinforcement tensile strain exceeds 0.005. In
addition, it also might be necessary to impose a larger minimum
wall thickness (tw) and a smaller wall slenderness ratio (hs/tw) for
compression-controlled walls. The objective of these requirements
would be to maintain a stable compressive zone as the concrete
yields in compression.
Even with more stringent design requirements for compressioncontrolled wall sections, it may not be reasonable to expect significant inelastic deformation capacity (rotation) can be achieved
through compression yielding of concrete; therefore, it might be

ACI 318-111 detailing requirements for SBEs are based on

requirements that were developed for columns; these provisions
may be insufficient for thin walls. The review of recent wall damage in earthquakes and laboratory tests provides sufficient evidence to raise concerns related to detailing of thin walls. For
example, although the quantity of transverse reinforcement provided at the boundaries of the conventional RC wall tested at EDefense were 1.4 and 2.1 times that required by ACI 318-11
21.9.6.4 (for the larger spacing of 100 mm used at Axis C), concrete crushing and lateral instability (Fig. 7(b)) occurred earlier in
the Kobe 100% test, followed by substantial sliding (Fig. 8).
Inspection of the damaged boundary zone revealed that relatively
large clear cover was used, on the order of 40 mm (larger than the
code minimum in ACI 318, which is 19 mm), suggesting that the
confined core was incapable of maintaining stability of the compression zone following loss of concrete cover. For columns, ACI
318-11 Equation (21-4), which is based on maintaining column
axial load capacity after cover concrete spalling, typically governs
the selection of transverse reinforcement for smaller columns
where cover makes up a larger percentage of the gross concrete
section. This equation also was required for wall SBEs prior to
ACI 318-9971; it was dropped because it rarely controlled for the
thicker walls that were commonly used at that time. For the EDefense conventional RC wall, the provided transverse reinforcement is only 0.34 and 0.45 times that required by ACI 318-11
Equation (21-4), suggesting that improved performance may have
resulted had this relation been required. Additional testing is
needed to determine if reinstating (21-4) is sufficient to ensure
ductile behavior of thin boundary zones.
ACI 318-11 21.6.6.2 allows a distance of 14 (356 mm)
between adjacent hoops or ties. Use of such a large spacing for
thin SBEs is unlikely to provide sufficient confinement (Fig. 16)
and use of such a large horizontal dimension is incompatible with
use of a vertical spacing one-third the wall thickness. For example,
for a 10 in. (254 mm) thick wall, such as used in the E-Defense
test, SBE vertical spacing is limited to 3.33" (84.6 mm); however,
the horizontal spacing along the wall can reach 14 in. (356 mm);
therefore, the ratio of vertical to horizontal spacing can reach 14/
3.33=4.2. An additional limit should be considered for wall SBEs,
similar to that used for vertical spacing, where the horizontal spacing between legs of hoops or ties along the length of the wall is
limited to a fraction of the wall thickness, e.g., 0.67tw. As well, use
of unsupported bars at the wall edge, which initiated the section
failure for test PW2 (Fig. 6(a)), should not be allowed until more
information is available to justify this detail.
Most of the issues raised in the preceding paragraphs are currently under study by ACI Committee 318 with potential changes
being introduced in ACI 318-14.

International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)11

3.5 Wall slenderness and lateral stability

Limits on wall slenderness should be considered to address
instability failures, similar to what was done in the UBC (1997),
which imposed a slenderness limit of tw hs/16, where hs is the
unsupported height (typically, one story). Based on observations in
recent earthquakes and tests, a lower limit should probably be
used within plastic hinge zone, a ratio of tw hs/10 was recently
recommended in Moehle et al.66 This issue is currently under
42
study by ATC 94.

4. Wall and coupling beam modeling

Use of beam-column models with rigid-plastic hinges and fiber
models with uniaxial material relations for concrete and reinforcement have become very common for analysis and design of buildings. For coupling beams, a beam-column model is common used,
since the added complexity of using a fiber model is generally not
warranted, especially for diagonally-reinforced coupling beams.
For a fiber model, the cross section geometry is prescribed with
concrete and steel fibers and elements are stacked to enable modeling of an element (e.g., planar wall). For fiber models, it is
important to use sufficient fibers to define the strain gradient at
equilibrium for a given loading and sufficient elements over the
wall height to capture the overall wall behavior; however, use of
too many fibers and elements may substantially increase computer
run time and lead to convergence issues. Although axial-bending
(P-M) interaction can be accounted for with beam-column models, typically a discrete bending stiffness must be specified;
whereas, for a fiber model, the flexural stiffness and section axialbending strength are derived from the specified material relations
and vary depending on the magnitude of axial load. Monitored
response quantities are plastic rotations for beam-column models
and average strain, curvature, or rotation over a specified element
or gage length for fiber models, since use of small element lengths
may lead to strain concentration and spurious results. Element or
gage lengths are typically selected based on assumed spread of
plasticity; use of half the member depth for structural walls is
common, although this value may not be appropriate for some
cases, as noted in the review of recent test results. Acceptance criteria are typically based on rotation or strain limits derived from
test results or engineering judgment, e.g., as given in ASCE 410672 Tables 6-18 and 6-19, and 6.4.3.1 sets the maximum permissible strain limits.
Comparisons between analytical and experimental results for
structural walls using simple beam-column and fiber models have
been reported by various researchers, including Thomsen and
41
73
36
6
Wallace, Wallace1, Elwood et al., Orakcal and Wallace, and
PEER/ATC-72.74 The focus here is on the comparisons for fiber

Fig. 16 Confinement of thin wall sections.

models, such as given in Fig. 17,6 which reveal that fiber models
using fairly sophisticated uniaxial material models are capable of
capturing load versus top displacement measured for flexural
deformations in laboratory tests for low-to-moderate axial stress
levels P= 0.10Agf'c. It is noted that the model is not capable of capturing strength degradation due to rebar buckling and rebar fracture; therefore, the strength degradation that initiates under
positive load at the end of the test is not captured by the model.
Comparisons between model and test results for a wall with a Tshaped cross-section (Fig. 17(b)) indicate that the overall load-displacement response is reasonably captured, although the model
slightly over-predicts the wall strength for the flange in tension.
The likely reason for this discrepancy is the inability of the model
to capture the nonlinear tensile strain variation in the flange,74
since the model assumes the same strain gradient (plane sections
remain plane) for the web and the flange. Waugh and Sritharan51
investigated the use of a modified fiber model to address this
issue, and report moderately improved comparisons, although the
model is limited to two-dimensional analysis. Orakcal and
6
Wallace also report that fiber models are capable of capturing
local responses, such as base rotation, average curvature, and average strains. Given that fiber models use uniaxial material models
for assumed plane sections, the results indicate that moment curvature analysis is an appropriate tool for assessing the stiffness and
strength, and to a lesser degree, deformation capacity, of slender
walls. This observation is supported by findings reported in
PEER/ATC-7274 and Johnson.53
The results presented in Fig. 17 compare nonlinear flexural
deformations obtained from the test and from the model, i.e., the
test data were processed to separate deformations due to flexure
and shear using the procedure recommended by Massone and
15
Wallace. Analysis results for wall RW2 using a coupled model,
or shear-flexure interaction model17 are shown in Fig. 18 for two
monotonic (pushover) analyses. For the first analysis, a monotonic
steel stress - strain relation was used, whereas in the second analysis, the steel stress - strain relation was manipulated to approximate the impact of cyclic loading (since the coupled model used
did not have cyclic material models). It is noted that the manipulated cyclic analysis results more closely match the test results and
are consistent with results presented in Fig. 14(a). Strain profiles
for the coupled model at three drift ratios are compared with test
results (Fig. 18(b)) and indicate that larger compressive strains are
predicted with the model compared with an uncoupled model.6
53
Johnson reports similar observations. The findings suggest that
coupling (shear-flexure interaction) leads to significantly larger
concrete compressive strains than would be predicted using an
uncoupled model. Although the results presented here are preliminary, they indicate that the larger compressive strains measured in
the tests are likely related to physical phenomena; therefore, they
cannot be discounted. An alternative (uncoupled) modeling
approach, where the shear force-deformation behavior is softened
to account for nonlinear shear deformations, is presented in ATC75
76 ; however, this modeling approach does not account for the
impact of shear-flexure interaction on concrete compressive strain,
it only addresses the underestimation of lateral deformations.
Since the approach used in ACI 318-11 21.9.6.2 to assess detailing requirements (presented earlier) is based on estimating the
concrete compressive strain, the likely under-estimation of con-

12International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)

Fig. 17 Comparison of model and test results6.

Fig. 18 Shear-flexure interaction model: (a) Load-displacement; (b) curvature.

crete compressive strain due to shear-flexure interaction should be

considered (indirectly, probably via the coefficient in ACI 318-11
Equation 21-8). Clearly, this is an area that requires additional
research.
The results presented here and the other studies noted do indicate that fiber models (and beam-column models) are valuable
design tools provided that the one understands that the results
obtained are not precise, i.e., the sensitivity of the results are considered. For example, local responses are more likely to be sensitive to model (e.g., mesh) and material (e.g., reinforcement strain
hardening) parameters,2,74,75 and studies indicate that concrete
compressive strains are generally under-estimated (unless shearflexure interaction is considered).
It also is important to note that the studies summarized here do
not address modeling of splice behavior (anchorage slip/extension,
sometimes referred to as strain penetration, has been studied) and
sliding shear behavior. As discussed in the review of recent tests,
splice behavior significantly impacted wall deformations capacity,
focusing inelastic deformations either below (Fig. 4(b)) or above
(Fig. 5(b), Fig. 6) the splice region, whereas concrete crushing and
rebar buckling at the wall boundary for the E-Defense test led to
large sliding shear deformations (Fig. 8(b)). Although it is possible
to incorporate these behaviors into fiber models, insufficient test
data exist to calibrate and validate these models. As well, even
with test data, it is questionable whether modeling these behaviors
is recommended. At least for new design, it is probably advisable
to avoid these problems, although additional testing is needed to
better determine how to accomplish this goal.

4.1 Coupling beams

Nonlinear modeling approaches commonly used by practicing
engineers are investigated to assess how well they are able to represent the measured test results presented earlier. Two models are
considered, one utilizing a rotational spring at the ends of the
beam to account for both nonlinear flexural and shear deformations (Mn hinge) and one utilizing a nonlinear shear-displacement
spring at beam mid-span to account for both shear and shear
deformations (Vn hinge). Both models were subjected to the same
loading protocol used in the tests.31 In this study, CSI Perform 3D
was used.76 Naish31 provides detailed information on modeling
parameters used to generate analysis results. Backbone relations
for the models were derived from test results, described below.

4.2 Test backbone relations

Backbone relations derived from the test data (solid line) are
compared with the original unmodified test backbone relations
72
(broken lines) and ASCE 41 relations (wide line) in Fig. 19. The
test relations were modified because slip/extension deformations,
which were significant for the one-half scale tests, produce less
72
beam chord rotation for full-scale beams. The ASCE 41 relation,
primarily based on test results for coupling beams with aspect
ratio less than 1.5, is too stiff. Naish31 reassessed the relation used
for low aspect ratio coupling beams using fragility relations, and
recommends new, slightly modified relations.

4.3 Diagonally-reinforced coupling beams (2.0 <

ln/h < 4.0)
The Mn-hinge model consists of an elastic beam cross-section

International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)13

with EcIeff =0.5EcIg, elastic-rotation springs (hinges) at each beamend to simulate the effects of reinforcement slip/extension deformations, and rigid plastic rotational springs (hinges) at each beamend to simulate the effects of nonlinear deformations. The stiffness
of the slip/extension hinges are defined using the Alsiwat and
77
Saatcioglu model, whereas the plastic rotations of the nonlinear
flexural hinges are modeled using the backbone relations derived
from test results (Fig. 19, for original test data, but excluding the
elastic deformation) with nominal shear strength defined using
ACI 318-08 Equation (21-9). The Vn-hinge model also consists of
an elastic beam cross-section and slip/extension hinges; however,
instead of using flexural hinges at the beam ends, a shear force
versus displacement hinge (spring) is used at beam mid-span to
simulate the effects of nonlinear deformations. The shear hinge
properties are defined using the backbone relations derived from
the test results (Fig. 19, for original test data).
Figure 20 shows cyclic load-deformation plots for the two models and the test results for CB24F, which are representative of
results obtained for other specimens. Both models accurately capture the overall load-displacement response of the member; however, the Mn-hinge model (Fig. 20(a)) captures the unloading
characteristics better than the Vn-hinge model (Fig. 20(b)), due to
the fact that unloading stiffness modeling parameters, which help
to adjust the slope of the unloading curve, are available for the
flexural hinges in the commercial computer program used, but not
31
for the shear hinges (see Naish for a complete description of the
modeling parameters and assigned values).
Model results for two frame beam tests are shown in Fig. 13 for
76
the Mn hinge model, again using the CSI Perform 3D program.
The models accurately capture the measured responses, specifi-

Fig. 19 Coupling beam test backbone curves.

cally in the slope of the loading and unloading curves, and in the
pronounced pinching of the cyclic load-deformation plot. The
commercial computer program used allowed the shape of the
load-deformation loops to be manipulated through specifying
energy dissipation parameters to simulate the pinching of the load31
deformation plots of the test beams. Naish includes detailed
information on the model parameters used in the comparisons.

5. Conclusions
Wall performance in recent earthquakes and laboratory tests is
reviewed and American Concrete Institute 318 provisions are
reassessed to identify possible shortcomings. The findings suggest
a number of issues require more in-depth study, particularly for
thin walls. Approaches that could be implemented within ACI 318
to address these issues also are presented. In particular, changes
are needed to increase the design displacement used in ACI 31811 Equation (21-8), changing the value of the denominator from
600 to 1200 is recommended. To ensure spread of plasticity consistent with the derivation of Equation (21-8), walls should be tension-controlled or be designed and detailed to maintain a stable
compressive zone as the concrete yields in compression. Limits on
wall thickness and slenderness are suggested as one way of
addressing this latter issue. Limiting wall compression strain for
compression-controlled walls also might be prudent; this can be
accomplished by limiting the drift ratio to about 0.01.
Recent tests of 2.4 and 3.33 aspect ratio coupling beams are presented and reveal that beams detailed according to the new provision in ACI 318-08, which allow for full section confinement,
have performance, in terms of strength and ductility, that is slightly
better than beams detailed according to the old provision in ACI
318-05, which requires confinement of the diagonal bar groups.
Including a reinforced concrete slab increases the beam shear
strength approximately 15-20%, whereas adding post-tensioning
increases the beam shear strength an additional 10%. However,
the strength increase was directly related to the increase in beam
moment strength, as the beam shear force was limited by flexural
yielding.
Modeling approaches used for structural walls adequately capture the nonlinear axial-bending responses, but are unable to capture strength loss, which typically results for buckling of vertical
boundary reinforcement or lateral instability of the flexural-compression zone. Additional experimental studies are required to better characterize these types of failures, particularly for thin walls.
Recent research related to wall modeling has focused on capturing

Fig. 20 Model and test results: (a) Mn hinge model; (b) Vn hinge model.

14International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)

observed shear-flexure interaction, where nonlinear shear deformations are observed for slender walls where behavior is dominated by flexural responses. A variety of modeling approaches
have recently been proposed using biaxial material models, truss
models, and empirical approaches. Available information strongly
suggests that shear-flexure interaction leads to large concrete compressive strains than would be predicted with an uncoupled model,
suggesting that current ACI 318 provisions that base wall boundary detailing requirements on concrete compressive strain should
include a measure of conservatism until this behavior is better
understood. Additional research, including detailed experimental
measurements of global and local responses, is needed to validate
and calibrate models for cyclic loads and for cases where nonlinear shear deformations are more significant (typically aspect ratio
1.5 to 3.0 walls).
Simple nonlinear model approaches for coupling beams, either
moment-hinge or shear-hinge, accurately represent the load-deformation behavior of test beams. The flexural hinge model better
matches the test results in the unloading and reloading range, due
to the specific modeling parameters available in the computer software used (unloading stiffness modeling parameters), although
both models produce acceptable results up to 3% total rotation for
beams with ln/h between 2.0 and 4.0. Therefore, depending on the
computer program used, the influence of modeling parameters on
the load versus deformation responses should be compared with
test results to ensure that they adequately represent observed
behavior.

Acknowledgements
This research described in this paper was carried out with funding from various sources, including the EERI Learning from
Earthquakes program (NSF CMMI-0758529), NSF RAPID
projects to enhance US-Japan collaboration related to the EDefense tests in December 2010 (CMMI-1110860 and CMMI1000268; Program Director Joy Pauschke), NSF NEES REU
(CMMI-0927178), Charles Pankow Foundation (CPF Grant No.
4-06), as well as support provided to the first author by the Japan
Society for the Promotion of Science (JSPS) Invitation Fellowship
Program during the fall 2010. This support is gratefully acknowledged. The author would like to thank those researchers who have
contributed their research results to NEEShub, which provides an
invaluable resource, as well as other researchers for their comments and input, including C French (U Minnesota), S. Sritharan
(Iowa State), L Lowes and D Lehman (U Washington), K Elwood
(UBC), and J Moehle (UC Berkeley). And finally, the author
would like to express his deep appreciation to the Japanese
researchers involved with the December 2010 E-Defense tests for
sharing their research ideas and results, including: T Nagae
(NIED), K Tahara (NIED), T Matsumori (NIED), H Shiohara (U
Tokyo), T Kabeyasawa (U Tokyo ERI), S Kono (Kyoto U), M
Nishiyama (Kyoto U); and M. Nakashima (NIED, Kyoto U).
Opinions, findings, conclusions, and recommendations in this
paper are those of the author, and do not necessarily represent
those of the sponsors or other individuals mentioned here.

References
1. American Concrete Institute. Building Code Requirements

for Structural Concrete (ACI 318-11) and Commentary (ACI

318R-11), Farmington Hills, Michigan, 2011.
2. Orakcal, K., Conte, J. P., and Wallace, J. W., Flexural
Modeling of Reinforced Concrete Walls - Model Attributes,
ACI Structural Journal, Vol. 101, No. 5, 2004, pp. 688~698.
3. Kabeyasawa, T., Shiohara, H., Otani, S., and Aoyama, H.,
Analysis of the Full-Scale Seven-Story Reinforced Concrete
Test Structure, Journal of the Faculty of Engineering, The University of Tokyo (B), Vol. 37, No. 2, 1983, pp. 431~478.
4. Fischinger, M., Vidic, T., Selih, J., Fajfar, P., Zhang. H. Y.,
and Damjanic, F. B., Validation of a Macroscopic Model for
Cyclic Response Prediction of RC Walls, in N.B. Bicanic and
H. Mang (eds.), Computer Aided Analysis and Design of Concrete Structures, Pineridge Press, Swansea, Vol. 2, 1990, pp.
1131~1142.
5. Colotti, V., Shear Behavior of RC Structural Walls, Journal of Structural Engineering, ASCE, Vol. 119, No. 3, 1993, pp.
728~746.
6. Orakcal, K. and Wallace, J. W., Flexural Modeling of
Reinforced Concrete Walls - Model Calibration, ACI Structural
Journal, Vol. 103, No. 2, 2006, pp. 196~206.
7. Massone, L. M., Orakcal, K., and Wallace, J. W., Modeling of Squat Structural Walls Controlled by Shear, ACI Structural Journal, Vol. 106, No. 5, 2009, pp 646~655.
8. Limkatanyu, S. and Spacone, E., Reinforced Concrete
Frame Element with Bond Interfaces. I: Displacement-Based,
Force-Based, and Mixed Formulations, J. Struct. Eng., Vol. 128,
2002, 346 pp.
9. Elwood, K. J. and Moehle, J. P., Axial Capacity Model for
Shear-Damaged Columns, Structural Journal, ACI, Vol. 102,
No. 4, 2005, pp. 578~587.
10. Elwood, K. J. and Moehle, J. P., Dynamic collapse analysis for a reinforced concrete frame sustaining shear and axial
failures, Earthquake Engineering and Structural Dynamics,
Vol. 37, No. 7, 2008, pp. 991~1012.
11. Wallace, J. W., Elwood, K. J., and Massone, L. M., An
Axial Load Capacity Model for Shear Critical RC Wall Piers,
Journal of Structural Engineering, ASCE, Vol. 134, No. 9, 2008,
pp. 1548~1557.
12. Oesterle, R. G., Fiorato, A. E., Johal, L. S., Carpenter, J.
E., Russell, H. G., and Corley, W. G., Earthquake Resistant
Structural Walls - Test of Isolated Walls, Report No. GI-43880/
RA-760815, National Science Foundation, Arlington, Va., 1976,
315 pp.
13. Oesterle, R. G., Aristizabal-Ochoa, J. D., Fiorato, A. E.,
Russell, H. G., and Corley, W. G., Earthquake Resistant Structural Walls - Test of Isolated Walls - Phase II, Report No.
ENV77-15333, National Science Foundation, Arlington, Va.,
1979, 327 pp.
14. Hiraishi, H., Evaluation of Shear and Flexural Deformations of Flexural Type Shear Walls, Bulletin of the New
Zealand National Society for Earthquake Engineering, Vol. 17,
No. 2, 1984, pp. 135-144.
15. Massone, L. M. and Wallace, J. W., Load Deformation
responses of Slender Reinforced Concrete Walls, ACI Structural
Journal, Vol. 101, No. 1, 2004, pp. 103~113.
16. Petrangeli, M., Pinto, P. E., and Ciampi, V., Fiber Element for Cyclic Bending and Shear of RC Structures. I: The-

International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)15

ory, Journal of Engineering Mechanics, ASCE, Vol. 125, No.

9, 1999, pp. 994~1001.
17. Massone, L. M., Orakcal, K., and Wallace, J. W., Modeling flexural/shear interaction in RC walls, ACI-SP-236, Deformation Capacity and Shear Strength of Reinforced Concrete
Members under Cyclic Loadings, American Concrete Institute,
Farmington Hills, MI, Paper 7, 2006, pp. 127~150.
18. Jiang, H. and Kurama, Y. C., Analytical Modeling of
Medium-Rise Reinforced Concrete Shear Walls, ACI Structural
Journal, Vol. 107, No. 4, 2010, pp. 400~410.
19. Panagiotou, M. and Restrepo, J. I., (2011), Nonlinear
Cyclic Truss Model for Strength Degrading Reinforced Concrete
Plane Stress Elements, Report No. UCB/SEMM-2011/01,
Structural Engineering, Mechanics and Materials, Department of
Civil and Environmental Engineering University of California,
Berkeley, 37 pp., February 2011.
20. Xu, S.-Y. and Zhang, J., Hysteretic shearflexure interaction model of reinforced concrete columns for seismic response
assessment of bridges, Earthquake Engng Struct. Dyn., Vol. 40,
No.3, 2011, pp. 315~337.
21. Beyer, K., Dazio, A., and Priestley, M. J. N., Shear deformations of slender reinforced concrete walls undet seismic loading, ACI Structural Journal, Vol. 108, No. 2, 2011, pp. 167~177.
22. Trna, T. A. and Wallace, J. W. Lateral Load Behavior of
Moderate Aspect Ratio RC Structural Walls, PhA Prospectus,
University of California, Los Angeles, Department of Civil Engineering, 2010, 62 pp.
23. Paulay, T., Coupling Beams of Reinforced Concrete Shear
Walls, Journal of Structural Division, ASCE, 1971, pp. 843~862.
24. Paulay, T. and Binney, J. R., Diagonally Reinforced Coupling Beams of Shear Walls, Shear in Reinforced Concrete, SP42, American Concrete Institute, Farmington Hills, Mich., 1974,
pp. 579~598.
25. Barney, G. B., Shiu, K. N., Rabbit, B. G., Fiorato, A. E.,
Russell, H. G., and Corley, W. G., Behavior of Coupling Beams
under Load Reversals (RD068.01B), Portland Cement Association, Skokie, IL, 1980.
26. Tassios, T. P., Moretti, M., and Bezas, A., On the Coupling
Behavior and Ductility of Reinforced Concrete Coupling Beams
of Shear Walls, ACI Structural Journal, Vol. 93, No. 6, 1996, pp.
711~720.
27. Xiao, Y., Esmaeily-Ghasemabadi, A., and Wu, H., HighStrength Concrete Beams Subjected to Cyclic Shear, ACI Structural Journal, Vol. 96 No.3, 1999, pp.392~399.
28. Galano, L. and Vignoli, A., Seismic Behavior of Short
Coupling Beams with Different Reinforcement Layouts, ACI
Structural Journal, Vol. 97, No. 6, 2000, pp. 876~885.
29. Kwan, A. K. H. and Zhao, Z. Z., Testing of coupling
beams with equal end rotations maintained and local joint deformation allowed, Structures and Buildings, Thomas Telford, London, Vol. 152, No. 1, 2001, pp. 67~78.
30. Fortney, P., The Next Generation of Coupling Beams,
PhD Dissertation, University of Cincinnati, 2005, 370 pp.
31. Naish, D., Testing and Modeling of Reinforced Concrete
Coupling Beams, Ph.D. Dissertation, Department of Civil &
Environmental Engineering, University of California, Los Angeles, CA, 2010, 251 pp.
32. Naish, D. and Wallace, J. W., Testing and Modeling of

Diagonally-Reinforced Reinforced Concrete Coupling Beams,

Proceedings Special Session, 9th U.S. National Conference on
Earthquake Engineering, Toronto, Canada, 2010 (Paper 1575).
33. Parra-Montesinos, G., Wight, J. K., Lequesne, R. D., and
Seekit, M., A summary of ten years of research on HPFRC coupling beams, High Performance Fiber Reinforced Cement Composites 6, Parra-Montesinos, Gustavo J., Reinhardt, Hans W.,
Naaman, Antoine E. (Ed.), 2012, 560 pp.
34. Wallace, J. W., Modeling Issues for Tall Reinforced
Concrete Core Wall Buildings, The Structural Design of Tall
and Special Buildings, Wiley InterScience, Vol. 16, 2007, pp.
615~632.
35. Wallace, J. W., Performance-Based Design of Tall Core
Wall Buildings, Earthquake Engineering in Europe, Garevski
M and Ansal A, editors, Springer, 2010, pp. 279~307.
36. Elwood, K. J., Matamoros, A. B., Wallace, J. W., Lehman,
D. E., Heintz, J. A., Mitchell, A. D., Moore, M. A., Valley, M. T.,
Lowes, L. N., Comartin, C. D., and Moehle, J. P., Update to
ASCE/SEI 41 Concrete Provisions, Earthquake Spectra, Vol. 23,
No. 3, 2007, pp. 493~523.
37. NCh 433.Of 96, Diseo Sismico de Edificios (Chilean
Building Code), Chile, 2006.
38. American Concrete Institute. Building Code Requirements for Structural Concrete (ACI 318-95) and Commentary
(ACI 318R-95), Farmington Hills, MI, 1995.
39. Wallace, J. W., February 27, 2010 Chile Earthquake:
Preliminary Observations on Structural Performance and Implications for U.S. Building Codes and Standards, ASCE Structures Congress, Paper 1159, Las Vegas, 2011.
40. Massone, L. M. and Wallace, J. W., Lessons from Chile:
Impacts of Earthquake Engineering of RC Buildings in the US,
EERI/NEES Webinar, 2011. http://nees.org/resources/3192
41. Thomsen, I. V., Wallace, J. H., and Wallace, J. W., Displacement-Based Design of Slender RC Structural Walls - Experimental Verification, J. Struct. Eng., ASCE, Vol. 130, No. 4,
2004, pp 618~630.
42. ATC-94, Analysis of Seismic Performance of Reinforced
Concrete Buildings in the 2010 Chile Earthquake, Including
Effects of Non-Seismic-Force-Resisting Building Structural Elements - Task Order 21, 2011. https://www.atcouncil.org/Projects/
nehrp-jv.html
43. Corley, W. G., Fiorato, A. E., and Oesterle, R. G., Structural Walls, Publication SP-72, ACI, 1981, pp. 77~131.
44. Paulay, T. and Priestley, M. J. N., Stability of Ductile
Structural Walls, ACI Structural Journal, American Concrete
Institute, Vol. 90, No. 4, 1993, pp. 385~392.
45. Chai, Y. H. and Elayer, D. T., Lateral Stability of Reinforced Concrete Columns under Axial Reversed Cyclic Tension
and Compression, ACI Structural Journal, ACI, Vol. 96, No. 5,
1999, pp. 780~789.
46. EERI, The M 6.3 Christchurch, New Zealand, Earthquake of February 22, 2011, EERI Special Earthquake Report,
2011.
47. NZRC, Canterbury Earthquakes Royal Commission Interim Report, 2011. http://canterbury.royalcommission.govt.nz/
Interim-Report.
48. Wallace, J. W., Evaluation of UBC-94 Provisions for
Seismic Design of RC Structural Walls, Earthquake Spectra,

16International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)

Vol. 12, No. 2, 1996, pp. 327~348.

49. Elwood, K. J., Personal Communication. [see Also EERI
Christchurch Earthquake Clearing House, 2011: http://www.eq
clearinghouse.org/2011-02-22-christchurch/
50. Waugh, J., Aaleti, S., Sritharan, S., and Zhao, J., Nonlinear Analysis of Rectangular and T-Shaped Concrete Walls,
ISU-ERI-Ames Report ERI-09327, Iowa State University,
Department of Civil, Construction and Environmental Engineering, 2008, 351 pp.
51. Waugh, J. D. and Sritharan, S., Nonlinear Analysis of TShaped Concrete Walls Subjected to Multi-Directional Loading,
Proceedings Paper 1506, 9th U.S. National Conference and
10th Canadian Conference on Earthquake Engineering, Ontario,
Canada, 2010.
52. Brueggen, B. L. and French, C. W., Simplified Modeling
of Non-Rectangular RC Structural Walls, Proceedings Paper
1713, 9th U.S. National Conference and 10th Canadian Conference on Earthquake Engineering, Ontario, Canada, 2010.
53. Open System for Earthquake Engineering Simulation
(OpenSees), Pacific Earthquake Engineering Research Center,
University of California, Berkeley, 2009. http://opensees.berkeley.edu/
54. Birely, A., Lehman, D., Lowes, L., Kuchma, D., Hart, C.,
and Marley, K., Investigation of the Seismic Response of Planar Concrete Walls, Proceedings Paper 773, 9th U.S. National
Conference and 10th Canadian Conference on Earthquake Engineering, Ontario, Canada, 2010.
55. Birely, A., Lehman, D., Lowes, L., Kuchma, D., Hart, C.,
and Marley, K., Investigation of the Seismic Behavior and
Analysis of Reinforced Concrete Structural Walls, Proceedings
14th World Conference on Earthquake Engineering, Beijing,
China, 2008.
56. Sriram, A. and Sritharan, S., Nonlinear Fiber-Based Analysis of Rectangualr Concrete Walls Designed with Different
Anchorage Details,, Proceedings Paper 123, 9th U.S. National
Conference and 10th Canadian Conference on Earthquake Engineering, Ontario, Canada, 2010.
57. Lehman, D. E. and Lowes, L. N., Personal communication
about laboratory test results for NEESR project Seismic Behavior,
Analysis, and Design of Complex Wall Systems, 2011. see also
nees.org/warehouse/project/104
58. Panagiotou, M. and Restrepo, J. I., Practical Lessons
Learned from the Full-Scale 7-Story Building Shake Table Test
at UC San Diego, 2007 SEAOC Convention, Squaw Creek,
CA, 2007, pp. 57~74.
59. Kabeyasawa, T., Kabeyasawa, T., Matcumori, T., Kabeyasawa, T. and Kim, Y., Full-scale dynamic collapse tests of
three-story reinforced concrete buildings on flexible foundation
at E-defense, Proceedings of the 14th World Conference on
Earthquake Engineering, Beijing, China, Paper ID: S15-002,
2008.
60. Kabeyasawa, T., Kabeyasawa, T., and Kim Y., Collapse
Simulation of Reinforced Concrete Buildings with ASFI
Approach, Proceedings Paper 816, 9th U.S. National Conference
and 10th Canadian Conference on Earthquake Engineering,
Ontario, Canada, 2010.
61. Kabeyasawa, T., Kabeyasawa, T., Kim, Y., Kabeyasawa,
T., Bae, K., and Quang, P. V., Strength and Deformability of

Reinforced Concrete Columns with Wing Walls, Proceedings

Paper 813, 9th U.S. National Conference and 10th Canadian
Conference on Earthquake Engineering, Ontario, Canada, 2010b.
62. Nagae, T., Tahara, K., Matsumori, T., Shiohara, H., Kabeyasawa, T., Kono, S., Nishiyama, M. Wallace, J. W., Ghannoum,
W., Moehle, J. P., Sause, R., Keller, W., and Tuna, Z., Design
and Instrumentation of the 2010 E-Defense Four-Story Reinforced Concrete and Post-Tensioned Concrete Buildings, Report
No 2011-XX, Pacific Earthquake Engineering Research Center,
Department of Civil & Environmental Engineering, University of
Califoria, Berkeley, 2011, 235 pp.
63. Brueggen, B. L., Performance of T-shaped Reinforced
Concrete Structural Walls under Multi-Directional Loading,
PhD Dissertation, University of Minnesota, Department of Civil
Engineering, 2009, 498 pp.
64. Boroschek, R., Soto, P., and Leon, R., Registros del Terremoto del Maule Mw=8.8 27 de Febrero de 2010, Red Nacional de Acelergrafos del Departamento de Ingeniera Civil,
Facultad de Ciencias Fsicas y Matemticas, Universidad de
Chile, Informe RENADIC 10/05, 2010, 100 pp. (http://www.terremotosuchile.cl/)
65. Naish, D., Wallace, J., Fry, J. A., and Klemencic, R.,
Experimental Evaluation and Analytical Modeling of ACI 31805/08 Reinforced Concrete Coupling Beams Subjected to
Reversed Cyclic Loading, Report No. UCLA-SGEL 2009/06,
2009.
66. Moehle, J. P., Acevedo, C., and Creagh, A., Exploratory
tests of wall boundary elements subjected to alternating tensile
and compressive loadings, Poster and oral presentations at
2010 PEER Annual Meeting, 2011.
67. Wallace, J. W. and Orakcal, K., ACI 318-99 Provisions
for Seismic Design of Structural Walls, ACI Structural Journal,
American Concrete Institute, Vol. 99, No. 4, 2002, pp. 499~508.
68. Wallace, J. W. and Moehle, J. P., Ductility and Detailing
Requirements of Bearing Wall Buildings, J. Struct. Eng.,
ASCE, Vol. 118, No. 6, 1992, pp.1625~1644.
69. Moehle, J. P., Ghodsi, T., Hooper, J. D., Fields, D. C., and
Gedhada, R., Seismic Design of Cast-in-Place Concrete Special
Structural Walls and Coupling Beams: A guide for practicing
engineers, NEHRP Seismic Design Technical Brief No. 6,
National Institute of Standards and Technology, Gaithersburg,
MD, 2011.
70. UBC, Uniform Building Code, International Council of
Building Code Officials, Whittier, CA, 1997.
71. American Concrete Institute. Building Code Requirements for Structural Concrete (ACI 318-99) and Commentary
(ACI 318R-99), Farmington Hills, MI, 1999.
72. American Society of Civil Engineers, ASCE/SEI Standard 41-06, Seismic Rehabilitation of Existing Buildings,
Reston, VA, 2007, 428 pp.
73. Wallace, J. W., Slender Wall Behavior and Modeling,
PEER/EERI Technical Seminar Series, New Information on the
Seismic Performance of Existing Concrete Buildings, 2006. see
www.eeri.org/
74. PEER/ATC-72-1, Modeling and acceptance criteria for
seismic design and analysis of tall buildings, Applied Technology Council, 2010, 242 pp.
75. ATC-76, NIST GCR 10-917-08 (2012), Evaluation of

International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)17

the FEMA P-695 Methodology for Quantification of Seismic

Performance Factors, 2011. www.nehrp.gov/pdf/nistgcr10-9178.pdf
76. CSI, PERFORM-3D, Computers & Structures, Inc., Ber-

keley, CA, 2009.

77. Alsiwat, J. and Saatcioglu, M. Reinforcement Anchorage Slip under Monotonic Loading, Journal of Structural Engineering ASCE, Vol. 118, No. 9, 1992, pp. 2421~2438.

18International Journal of Concrete Structures and Materials (Vol.6 No.1, March 2012)