See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/283481169

An Investigation on the Seismic Behavior of Deep Reinforced Concrete

Coupling Beams

Article in ACI Structural Journal · January 2015

DOI: 10.14359/51687939

CITATIONS READS

50 1,647

4 authors, including:

Erwin Lim Shyh-Jiann Hwang

Bandung Institute of Technology National Taiwan University
30 PUBLICATIONS 180 CITATIONS 116 PUBLICATIONS 3,270 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Erwin Lim on 09 January 2020.

The user has requested enhancement of the downloaded file.

 ACI STRUCTURAL JOURNAL TECHNICAL PAPER
Title No. 113-S20

An Investigation on the Seismic Behavior of Deep

Reinforced Concrete Coupling Beams
by Erwin Lim, Shyh-Jiann Hwang, Ting-Wei Wang, and Yu-Hsuan Chang

The ACI 318-14 design procedure for a diagonally reinforced beams are subjected to very large displacement reversals
coupling beam is questioned because it proportions the amount and high force demands, especially for deep coupling beams
of diagonal reinforcement to meet shear demands by completely (ℓn/h < 2.0), the ACI 318-14 seismic design requires the use
neglecting the participation of concrete and failing to recognize of diagonal reinforcements. The ACI 318-14 design equa-
any flexural strength that may be developed. This ACI 318-14
tion for a diagonally reinforced coupling beam is based on
design procedure might introduce unnecessary additional forces
the recommendation from pioneering research2 and relies
to the entire coupled wall system. This study proposes that the
amount of diagonal reinforcement should be determined using solely on the shear strength provided by diagonal reinforce-
beam flexural theory to satisfy the design moment at the Design ment bars, as shown by Eq. (1)
Based Earthquake level. Test results of diagonally reinforced deep
coupling beam specimens showed that they can preserve ductile Vu
Vn _ ACI = 2 Avd f y sin α ≥ (1)
flexural behavior up to the Maximum Considered Earthquake level, φ
and part of the shear strength was contributed from concrete in
addition to diagonal bars. Finally, a strut-and-tie model to illus-
trate the involved mechanism is also proposed. where Avd is the total area of one group of diagonal bars; fy
represents the yielding strength of steel, α denotes the incli-
Keywords: design procedure; diagonally reinforced deep coupling beam; nation angle of the diagonal bars with respect to the longi-
seismic behavior; shear strength. tudinal axis of the beam; Vu represents the shear demand
at the DBE level; and ϕ is the shear reduction factor for a
INTRODUCTION diagonally reinforced coupling beam. By neglecting the
The ACI 318-141 design procedure determines the concrete contribution, previous research2-5 showed satisfac-
amount of longitudinal flexural reinforcement of a beam in a tory seismic performance, as sufficient shear strength at the
moment-resisting frame based on the code bending moment MCE level was ensured. However, no check on its flexural
at the Design Based Earthquake (DBE) level. According moment strength was performed.6 This design check on
to the capacity design approach, the shear strength is then flexure is important because the role of diagonal bars as the
increased accordingly by neglecting the concrete shear shear reinforcement is coupled with its role as the flexural
contribution to satisfy the plastic shear demand at the reinforcement. Hence, an overdesign of diagonal reinforce-
Maximum Considered Earthquake (MCE) level. However, ment may cause an increase in the flexural strength as well.
the design procedure for a diagonally reinforced coupling This study proposes that the amount of diagonal reinforce-
beam requires engineers to proportion the amount of diag- ment should be determined using beam flexural theory to
onal reinforcement according to the beam shear demand at satisfy the code moment at the DBE level. Four deep-beam
the DBE level and neglects the check of flexural strength. coupling beam specimens with clear span-depth ratios of 1.0
This different design approach for diagonally reinforced (ℓn/h = 1.0) and 2.0 (ℓn/h = 2.0) were tested. For each clear
coupling beams is questionable and requires evaluation. span-depth ratio, one specimen was reinforced traditionally
In the traditional ductile frame beam design, engineers while the other specimen was diagonally reinforced. Test
proportion a sufficient amount of longitudinal flexural results showed that the beam flexural theory, which assumes
reinforcement to satisfy the bending moment calculated at compatibility between concrete and steel (both longitudinal
the DBE level. Using the capacity design, adequate shear and diagonal) bars, gives good estimations of the flexural
strength must be provided at the MCE level. At beam ends strength. In addition, due to the strict confinement regulation
where plastic hinges and concrete degradation might occur, of ACI 318-14 for a coupling beam, the presence of stirrups
the plastic shear corresponding to the probable plastic and crossties assisted in maintaining the integrity of concrete
moment Mpr is assigned solely to the stirrups by neglecting under large displacement reversals. This research also shows
the concrete shear contribution. Because the roles of longi- that the shear strength can be estimated by properly consid-
tudinal flexural reinforcement bars and stirrups are clearly ering the shear resistance contributed by diagonal concrete
distinguished and independent of each other, the overdesign struts and diagonal reinforcement.
of shear reinforcement (in this case, stirrups) is acceptable
and does not affect the flexural strength of the beam. ACI Structural Journal, V. 113, No. 2, March-April 2016.
MS No. S-2014-274.R3, doi: 10.14359/51687939, was received April 22, 2015, and
In a coupled wall system, the yielding of coupling beams reviewed under Institute publication policies. Copyright © 2016, American Concrete
along the building height is also expected and becomes the Institute. All rights reserved, including the making of copies unless permission is
obtained from the copyright proprietors. Pertinent discussion including author’s
“fuse” to limit the input earthquake force. Because coupling closure, if any, will be published ten months from this journal’s date if the discussion
is received within four months of the paper’s print publication.

ACI Structural Journal/March-April 2016 217

Fig. 1—Test specimens. (Note: Cover thickness is 40 mm [1.5 in.] for top and bottom; 20 mm [0.75 in.] for left and right of
section; 1 mm = 0.0394 in.)
RESEARCH SIGNIFICANCE forcement, as shown in Fig. 1(b). These two specimens were
This study investigates the ACI 318-14 design procedure confined using D13 (No. 4) bars that were equally spaced at
of a diagonally reinforced coupling beam, which neglects the 100 mm (3.94 in.) intervals as the stirrups. Ten D10 (No. 3) and
check of flexural moment strength that might develop at the four D13 (No. 4) bars were used as horizontal shear rein-
DBE level. Using the test results of four deep coupling beams, forcement in CB10-1 and CB10-2, respectively.
this study shows the importance of considering both concrete The second test series of coupling beams with ℓn/h = 2.0
and diagonal reinforcement when predicting bending and (CB20 series) was carried out in 2012.8 The cross-sectional
shear capacities at the DBE and MCE levels, respectively. dimensions were 300 x 500 x 1000 mm (11.8 x 19.7 x 39.4 in).
This research also shows that a properly confined coupling Specimen CB20-1 was designed using four D29 (No. 9) bars
beam would maintain the integrity of concrete and allow for each group of diagonal bars, as shown in Fig. 1(c). The
the applicability of a strut-and-tie model for shear strength nominal shear strength calculated using Eq. (1) was 599 kN
prediction at the MCE level. Finally, a design concept based (134.7 kip); meanwhile, the shear corresponding to nominal
on the beam flexural demand to proportion the amount of flexural strength calculated using a conventional bending
diagonal reinforcement and a shear strength model based on analysis was 982 kN (220.8 kip). The counterpart of Spec-
strut-and-tie are proposed accordingly. imen CB20-1, Specimen CB20-2, used a traditional beam
layout with two D32 (No. 10) bars and one D36 (No. 11) bar
TEST PROGRAM as the main longitudinal flexural reinforcement to maintain
Two test series using coupling beams with clear span- a similar flexural strength (Fig. 1(d)). Similar to the CB10
depth ratios of 1.0 (ℓn/h = 1.0) and 2.0 (ℓn/h = 2.0) were series, these two specimens were also confined using the
carried out in two different time periods. In each test series, ACI 318-14 confinement requirement. Ten D13 (No. 4) and
one specimen was detailed according to the ACI 318-14 four D13 (No. 4) bars were used as horizontal shear rein-
diagonal reinforcement layout while another specimen was forcement in CB20-1 and CB20-2, respectively.
detailed using the traditional layout of a beam. The first test It is noteworthy that for specimens with a diagonal rein-
series of coupling beams with ℓn/h = 1.0 (CB10 series) was forcement layout (CB10-1 and CB20-1), the diagonal bars
carried out in 2011.7 The cross-sectional dimensions were were bent at a distance of 50 mm (1.97 in.) from the beam-
250 x 500 x 500 mm (9.8 x 19.7 x 19.7 in.) (width x depth reaction block interface. This construction detailing was
x length). Specimen CB10-1 was designed using four D25 introduced in 20059 for an easier handling process. All
(No. 8) bars for each group of diagonal bars, as shown the reinforcement bars, including the horizontal rein-
in Fig. 1(a). Using Eq. (1), the calculated nominal shear forcement bars (D13 [No. 4]), were extended to the far
strength was 746 kN (167.7 kip). Meanwhile, using conven- end of the concrete block and enough development length
tional bending analysis and considering a concrete strain was provided.
at the extreme compression fiber equals 0.003, the shear In the laboratory, the specimens were oriented vertically
corresponding to the nominal flexural strength was 1220 kN and tested using a stiff L-shaped steel frame and four actua-
(274.3 kip). To provide a direct comparison, a counterpart tors, as illustrated in Fig. 2(a). The lateral load was applied
specimen labeled CB10-2 was designed using a traditional through two horizontal actuators fixed on the reaction wall.
beam layout (without diagonal bars). To maintain a similar One of these two actuators was displacement-controlled
nominal flexural strength, two D25 (No. 8) and one D29 using the loading protocol recommended by ACI 374.1-0510
(No. 9) bars were used as both tension and compression rein- (Fig. 2(b)). Meanwhile, the other actuator was force-controlled

218 ACI Structural Journal/March-April 2016

 Table 1—Material properties
Yielding strength of bars, MPa
Specimen ID Concrete fc′, MPa D13 D25 D29 D32 D36
(1) (2) (3) (4) (5) (6) (7)
CB10-1 34.5
468.4 485.7 439.6 — —
CB10-2 36.1
CB20-1 52.1
502.0 — 466.3 450.3 447.8
CB20-2 52.2

Note: 1 MPa = 145 psi.

displacement transducer (LVDT) divided by the beam clear

span ℓn. To better illustrate the force transfer mechanism
within a coupling beam, Fig. 4 shows the crack patterns for
all the specimens along with the maximum width of the shear
cracks at different drift ratios. Overall, specimens with diag-
onal reinforcement bars possessed better seismic behavior,
indicated by a robust hysteretic loop and good toughness. A
summary of peak points of the load-deflection curves, shear
stress corresponding to the maximum lateral force attained,
and failure modes are presented in Table 2.

CB10 series
In Specimen CB10-1, which was detailed using a diag-
onal reinforcement layout, no major diagonal cracks were
observed at early drift ratios (DR < 2.1%), as indicated
in Fig. 4(a). At a DR of 2.1%, the shear crack width was
measured to be 0.35 mm (0.01 in.). The formation of a mature
Fig. 2—Test arrangement. diagonal strut was observed at a DR of 5.8%, where a peak
lateral strength Vmax of 1443.8 kN (324.6 kip) was reached,
in such a manner that the resulting force passed through the
as shown in Fig. 3(a). At this drift level, although the width
midheight of the beam. Two actuators were also used in the
of the shear crack widened up to 1.3 mm (0.05 in.), no signif-
vertical direction. The first actuator was force-controlled
icant concrete crushing was apparent. The inclination angle
to ensure that no axial load was exerted on the specimens.
of the major cracks was approximately coincident with the
Meanwhile, the second one was set as a displacement-
inclination of the diagonal bars. The specimen began to lose
controlled actuator to ensure that double-curvature bending
its strength at the second cycle and the test was terminated at
with no rotation at the top concrete block was produced.
the third cycle when the drift level reached 7.2%. At the final
Strain gauges were also attached at the main reinforcement
stage, crushing of concrete occurred near the end of the strut
(both diagonal and longitudinal bars) and stirrups. Details of
and buckling of the diagonal reinforcement was observed.
the measurement can be found elsewhere.7,8
The failure mode for this specimen (CB10-1) was flexural-
shear failure. Figure 3(a) also shows that, at a DR of 2%, the
TEST RESULTS
lateral load corresponding to the first yielding strength of the
Material properties
diagonal bar (that is, 1308 kN [294.1 kip]) can be preserved
The concrete and steel material properties for all four
up to a DR of 5.8%, where Vmax was reached. This force level
specimens are presented in Table 1. The concrete strength
can be reached because the shear capacity is contributed by
fc′ of each specimen was obtained from the average value of
concrete in addition to the diagonal bars.
three concrete cylinders tested on the same day as the test
The shear failure of CB10-2 was clearly observed through
of the coupling beam specimens. Meanwhile, for the steel
its load-deflection curve in Fig. 3(b). At an early DR of
properties, because these four specimens were constructed
0.8%, the width of the diagonal shear crack was 0.55 mm
and tested at two different time periods, for the same rein-
(0.02 in.). The crack then continued to widen up to 2.50 mm
forcing sizes there were two different values of steel yielding
(0.1 in.) as its peak lateral strength Vmax of 873.6 kN (196.4 kip)
strength. For each reinforcing bar size, three samples of rein-
at a DR of 1.7% was reached. At that corresponding DR, the
forcing bar were tested under a tensile strength test and the
crack pattern in Fig. 4(b) indicates the occurrence of major
average yielding strengths were reported.
diagonal cracks. Crushing of the concrete, which was also
observed, caused a drastic drop of the lateral load to less than
Hysteretic behaviors and crack patterns
50% of the peak lateral strength at the third cycle. At the next
The load-deflection curves for all specimens are presented
DR of 3.6%, even more concrete crushing and spalling of the
in Fig. 3, where the drift ratio (DR) is defined as the lateral
concrete cover were obvious.
deflection of a specimen measured from the linear variable

ACI Structural Journal/March-April 2016 219

Fig. 3—Load-deflection curves.

Fig. 4—Crack patterns at different drift ratios (DR). (Note: 1 mm = 0.0394 in.)

220 ACI Structural Journal/March-April 2016

CB20 series still preserved its flexural behavior for the next two drift
In the CB20 series, the load-deflection curves of both ratios (Fig. 3(c)). A more distinct drop of lateral strength of
CB20-1 and CB20-2 showed elastic behavior up to a DR Specimen CB20-1 began to occur at a DR of 7.3%, where a
of 1%. The strain gauge measurement of longitudinal rein- larger portion of concrete crushing and spalling occurred at
forcement indicated that the first yielding occurred at a both ends. Buckling of diagonal bars was also observed at
lateral load Vy of approximately 1000 kN (224.8 kip). The this deformation level. The test was terminated at the first
crack width measurement suggested that shear cracks were cycle at a DR of 9.4% and was determined to have failed
not dominant at this deformation level. Specimens CB20-1 from flexural failure. Figure 3(c) shows that the Vmax attained
and CB20-2 reached a maximum lateral load Vmax of –1073.0 at a DR of –2.2% can be preserved up to a DR of –7.3%.
and 1098.0 kN (–241.2 and 246.8 kip), respectively, at a DR Similar to the argument in CB10-1, this force level can be
of approximately 2%. At this DR, the maximum width of maintained because the shear capacity is a contribution of
the shear crack of CB20-1 was only 0.25 mm (0.01 in.), as both the concrete and the diagonal bars.
shown in Fig. 4(c), while that of CB20-2 was 0.5 mm The shear cracks that occurred in CB20-2 widened up
(Fig. 4(d)). Up to this deformation level, both specimens quickly and caused a large crushing and spalling of concrete
were able to develop their flexural behavior. at the first cycle of DR of 4.5% (Fig. 4(d)). Because no
With the presence of diagonal bars, Specimen CB20-1 diagonal reinforcing bars were provided for this specimen,
developed and maintained its flexural behavior well. At a concrete crushing caused a significant loss of shear resis-
DR of 4.3%, few regions with crushing of concrete, espe- tance. This was reflected in the significant drop of lateral
cially at the corner of the beam, were noted. This crushing load at the second and third cycle of the same DR (Fig. 3(d)).
was most likely due to flexural compression. No significant As a result, this specimen could not maintain its flexural
drop of lateral strength was observed, and Specimen CB20-1 strength to a higher drift level. The test was terminated after
the completion of the third cycle of DR of 5.7%. This mech-
anism was recognized as flexural shear failure.
Table 2—Test results
Peak point Vmax Strain gauge measurement
, Strain gauge readings of diagonal bars could provide
Load DR f c′bd
Specimen (Vmax), (θmax), helpful reference values to identify the roles of diagonal
MPa ( psi)
ID kN % Failure mode bars. Figure 5 plots the strain gauge measurements of the
(1) (2) (3) (4) (5) diagonal bars in Specimen CB10-1. The strain gauge read-
CB10-1 1443.8 5.8 2.6 (31.8) Flexure-shear (FS)
ings of D1 and D2 (in tension for positive displacement) as
well as D3 and D4 (in tension for negative displacement)
CB10-2 873.6 1.7 1.3 (16.2) Shear (S)
indicated yielding of the diagonal bars during tension. Strain
CB20-1 –1073.0 –2.2 1.3 (16.2) Flexure (F) gauge readings for the diagonal bars in compression also
CB20-2 1098.0 2.3 1.2 (14.2) Flexure-shear (FS) suggested that they reached yielding strain as indicated by
Note: 1 kN = 0.2248 kip.

Fig. 5—Strain gauge measurements of diagonal bars for CB10-1.

ACI Structural Journal/March-April 2016 221

Fig. 6—Strain gauge measurements of stirrups for CB10-1.
(Note: Diagonal bars not shown.)
D3 and D4. Meanwhile, readings from D1 and D2 suggested
that the diagonal bars barely yielded.
In addition to the observation of the crack pattern, the
elastic behavior of the stirrups is one of the essential param-
eters to justify the integrity of concrete. Three of four strain
gauges attached on the CB10-1 stirrups (Fig. 6) indicated that
they were in the elastic range along the loading history. This
observation shows that although abundant cracks developed
in Specimen CB10-1 at Vmax (Fig. 4(a)), the concrete still
maintained its integrity. Identical strain gauge readings for
stirrups were also obtained from a test of Specimen CB20-1.8

Pinching mechanism
The presence of diagonal bars also contributed signifi-
cantly to the energy dissipation capacity of a coupling beam, Fig. 7—Illustration of pinching behavior.
generally indicated by the absence of pinching. Specimens
CB10-2 and CB20-2 displayed severe pinching in their shear element of Fig. 7(b). In the principle directions, the
hysteretic loops, while Specimens CB10-1 and CB20-1 resistance from these steel bars did not cancel one another,
showed a well-rounded shape (Fig. 3). This pinching resulting in the absence of pinching.
behavior was explained well by Mansour et al.11 and is
briefly discussed and illustrated in Fig. 7. Figures 7(a) and DISCUSSION
7(b) plot the load-deflection curve of the first cycle at a DR In general, the test results of this study reconfirm the supe-
of 3.3% for Specimens CB20-2 and CB20-1, respectively. riority of coupling beams with a diagonal reinforcement
As more displacement was imposed on Specimen CB20-2, layout. The presence of diagonal bars ensures a good defor-
the cracks continued to grow and open up. During reversed mation capacity and energy absorption through its robust
loading, the cracks that had opened, closed up, but were not hysteretic loop.
necessarily perfectly in contact and left gaps. Due to these
gaps, little resistance was expected from concrete under the Evaluation of ACI 318-14 shear design equation
reloading (from Point B to C). Examination of a small shear The shear strengths of CB10-1 and CB20-1 calculated
element in its principal direction also indicated a small resis- using Eq. (1) are presented in Column (4) of Table 3. It
tance contribution from the orthogonally placed steel bars. shows that there is a large discrepancy between the shear
Due to the principal compression stress, the orthogonally strength predicted using Eq. (1) and the maximum attained
placed steel bars were subjected to compression; mean- lateral force Vmax. In other words, if Eq. (1) is used as a design
while, in the perpendicular principal direction, the steel bars tool to proportion the required amount of diagonal reinforce-
were subjected to tensile stress and, therefore, cancelled ment, overstrength factors of 1.67 and 1.61 for CB10-1 and
one another (Fig. 7(a)). Because neither the steel bars nor CB20-1 can be expected, respectively. The main reason
concrete provided shear resistance, severe pinching was for this overstrength is due to the fact that Eq. (1) neglects
observed near the origin. the concrete and leaves out the examination of any flexural
When a coupling beam with diagonal bars (CB20-1) strength that may have developed at the DBE level. Because
was reloaded (point B to C), although little resistance was the roles of diagonal bars as flexural and shear reinforce-
contributed by the concrete due to the gap, the diagonally ment are coupled, this overly provided shear strength would
placed steel bars did provide resistance, as illustrated in the further increase the flexural strength, which might cause the

222 ACI Structural Journal/March-April 2016

Table 3—Strength prediction of test specimens
Shear strength at high deformation level
Specimen ID d, mm Vmax, kN Vn_ACI, kN Vmn, kN a, mm K VSTM, kN VSTM/Vmax FM*
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
CB10-1 372.0 1443.8 862.7 1411.6 142.1 1.27 1376.6 0.95 FS
CB10-2 433.0 873.6 — 1614.4 114.2 1.24 405.1 0.46 S
CB20-1 371.0 –1073.0 664.8 1087.6 111.3 1.67 1185.3 1.10 F
CB20-2 429.0 1098.0 — 1009.2 97.7 1.67 441.0 0.40 FS
*
FM indicates failure mode predicted by strut-and-tie model.
Notes: 1 MPa = 145 psi; 1 mm = 0.0394 in.; 1 kN = 0.2248 kip.

yielding mechanisms of the coupling beam to occur at a much reinforced section. Based on previous experimental studies,2-5
higher force level. This overstrength would later contribute engineers need not worry about the shear strength at the
as a larger axial load to the wall and, ultimately, increase the MCE level if a diagonal layout is adopted. Moreover, this
wall flexural capacity and its corresponding shear demand. study also shows that, although the acting shear of the two
This phenomenon needs further investigation. diagonally reinforced specimens (CB10-1 and CB20-1) far
The maximum lateral load Vmax attained by the test spec- exceeded the ACI 318-14 maximum shear limitation (that is,
imens can be more reasonably estimated using flexural 0.83 f c′ (MPa)bd [10 f c′ (psi)bd ]), these two specimens
strength. Column (5) of Table 3 shows the shear corre- still possessed good deformation behavior. Column (4) of
sponding to the nominal flexural strength of a beam (Vmn), Table 2 indicates that, although the acting shear of CB10-1 was
which is given by 2.6 f c′ (MPa)bd ( 31.8 f c′ (psi)bd ), it still could maintain a
good deformation capacity up to a DR of 6%. Meanwhile,
2 × Mn CB20-1 could develop its flexural behavior up to a DR of 7%
Vmn = (2)
n despite its acting shear reaching as high as 1.3 f c′ (MPa)bd
(16.2 f c′ (psi)bd ).
where Mn denotes the flexural moment strength defined when
the extreme compression fiber of concrete reaches 0.003. Shear strength at MCE level
The strength ratios Vmax/Vmn were 1.02 and 0.99 for CB10-1 As discussed in the previous sections, a large difference
and CB20-1, respectively. A similar analysis can also be was observed when comparing the shear strengths calculated
easily carried out for conventionally reinforced coupling based solely on the contribution of diagonal bars using Eq. (1)
beams that failed from flexural shear (CB20-2). This result and Vmax (862.7 kN [193.9 kip] compared to 1443.8 kN
implies that a conventional bending analysis that considers [324.6 kip] for CB10-1, and 664.8 kN [149.4 kip] compared
the strain compatibility between concrete and flexural (both to –1073.0 kN [–241.2 kip] for CB20-1). However, the
diagonal and longitudinal) reinforcement bars provides a hysteretic loops of CB10-1 and CB20-1 (Fig. 3(a) and 3(c))
good strength estimation for flexure. indicated that they can preserve those force levels up to a
Based on the aforementioned discussion and test results, very large displacement level. This observation suggests that
this study recommends that the design procedure of a diago- the shear capacity was not only contributed by the diagonal bars.
nally reinforced coupling beam should adopt the traditional In addition, CB10-2 and CB20-2 also showed that a certain
beam seismic design approach. The required amount of amount of shear capacity is reserved at a high displacement
diagonal reinforcement should be proportioned based on the level in the absence of diagonal bars (Fig. 3(b) and 3(d)). It
flexural analysis to satisfy the design moment at the DBE implies that the shear strength is contributed from both the
level. It should be noted that because the bars are in a diag- diagonal bars and concrete. In the following section, a shear
onal layout, only the horizontal component of the diagonal strength model at a high displacement level is developed.
bars contributes to the flexural strength. For example, as an
illustration purpose, the flexural design equation for a singly ANALYTICAL FORMULATION FOR
reinforced section is given by Eq. (3) as follows SHEAR STRENGTH
The previous discussion addressed that the total shear
 a M strength of a coupling beam is contributed by both concrete
M n = Avd f y cos α  d −  ≥ u (3)
 2 φ and diagonal bars (Fig. 8(a)). A close examination of the
failure conditions of each specimen suggested that the
major distress arises from the compressive failure of a diag-
where d represents the effective beam depth; a denotes the
onal concrete strut. Hence, the concrete shear resistance is
depth of plastic compression zone calculated using Whitney’s
assumed to be contributed by the diagonal concrete strut.
stress block of a singly reinforced beam section; Mu is the
This study proposes a strut-and-tie model as a tool to esti-
moment demand at the DBE level; and ϕ denotes the strength
mate the compressive strength of a diagonal concrete strut
reduction factor for flexure. In the design practice, the amount
of a coupling beam. Under an elastic shear reinforcement
of diagonal reinforcement can be determined using a doubly

ACI Structural Journal/March-April 2016 223

Fig. 8—Macro model for coupling beam.
condition at a high displacement level, the strut-and-tie model compressive strength of a diagonal strut through yielding
allows the consideration of both concrete and diagonal bars strength. Yielding of diagonal bars in compression was also
(if any). confirmed through the strain gauge measurement of CB10-1
Coupling beams as well as beam-column joint subas- (Fig. 5). Meanwhile, the resistance from diagonal bars under
semblages are structural members in which the interactions tension needs to include the strain hardening effect due to
between bending- and shear-controlled failures dominate. the flexural deformation. In this paper, because the actual
In the research of a beam-column joint subassemblage, the yielding strength of steel was used, the strain hardening of
softened strut-and-tie (SST) model was shown to provide a steel was taken as 1.1fy, as shown in Eq. (4).
conservative shear strength estimation.12,13 Similarly, this At the MCE level, the shear capacity would decay due to
study also adopts an SST model to estimate the shear strength the cracking of concrete, especially in regions with plastic
of a coupling beam. The SST model, which satisfies force hinges. Hence, the nonlinear behavior of concrete must be
equilibrium, strain compatibility, and softened behavior properly considered in the macro model of the strut and tie.
of cracked reinforced concrete, is modified to include the One simple yet rational assumption to represent the nonlin-
structural behavior observed in the coupling beam tests at earity of concrete is to model the depth of concrete compres-
the MCE level. This model assumes that the main diagonal sion zone using the plastic compression zone a (Fig. 8(b)).
compression strut was formed from the end where a loading Consequently, the inclination angle of the diagonal strut is
was applied to the other end of beam.
The shear capacity at the MCE level can be estimated  h − 2 × a/2
from the compressive strength of a diagonal concrete strut θ = tan −1   (5)
 n
and diagonal bars, as given by Eq. (4)

VSTM = Cdsinθ + (1.0 + 1.1)Avd fysinα (4) where h denotes the depth of a coupling beam.
The SST model assumes that the contribution of a concrete
where Cd and θ represent the compressive strength of a strut arises not only from the main diagonal strut, but also
diagonal concrete strut and the inclination angle of the from the concrete sub-struts. These additional sub-struts are
main diagonal strut at the MCE level calculated from the provided by the presence of vertical and horizontal shear
simplified SST algorithm,12 respectively, by considering the reinforcement, represented as the vertical and horizontal ties
nonlinearity of concrete behavior. (Fig. 8(b)). Hence, the compressive strength of a diagonal
Because it is assumed that the main diagonal strut is concrete strut at the high ductility demand Cd is estimated as
formed from one end of the beam to the other end, the
inclination angle of the concrete strut and diagonal bars Cd = Kζ fc′Astr (6)
would not highly deviate. Hence, based on the argument
from ACI 318-14, compression reinforcement bars that are where K represents the strut-and-tie index; ζ indicates the
properly anchored, parallel to and located within a diagonal softening coefficient of cracked reinforced concrete taken
strut, and enclosed in ties shall be permitted to increase the as 3.35/ f c′ (MPa) ≤ 0.52 , and fc′ denotes the compressive

224 ACI Structural Journal/March-April 2016

strength of the concrete cylinder specimen. Meanwhile, the a capacity-to-demand ratio of CB10-1 equal to 0.95, which is
strut area accounting for concrete nonlinearity, Astr, is given as slightly less than unity.

Astr = a × b (7) CONCLUSIONS

Based on the cyclic loading tests of four deep coupling
where b denotes the width of a coupling beam. beam specimens with ℓn/h ≤ 2.0, several conclusions can be
The strut-and-tie index K represents the beneficial effect drawn from this study, and are given as follows:
of the presence of vertical and horizontal shear reinforce- 1. Test results confirm the superiority of the seismic
ment. The assumed force transfer mechanism in the SST behavior of a diagonally reinforced coupling beam. The
model regards these shear reinforcement bars as tension presence of diagonal bars ensures sufficient shear strength
ties, which allow an additional load path, known as substruts at the maximum considered earthquake level, which would
(Fig. 8(b)) in addition to the main diagonal strut. In the allow for a large deformation capacity. In addition, they
design of a coupling beam, ACI 318-14 requires full confine- also provide good energy absorption properties through the
ment on the entire beam section. This code provision would absence of pinching.
ensure that the tension ties remain elastic in the SST formu- 2. By neglecting the examination of the flexural strength,
lation. Therefore, the strut-and-tie index can be determined the current ACI 318-14 design procedure for a diagonally
mainly from the geometry of a beam as follows12 reinforced coupling beam might result in a larger flexural
strength. This larger flexural strength would cause unex-
1 1 pected forces to the coupled wall system and need to be
K≈ + −1 (8)
(
1 − 0.2 γ h + γ 2
h ) (
1 − 0.2 γ v + γ v2 ) investigated further.
3. Because the flexural strength can accurately capture
the maximum force developed at a diagonally reinforced
where γh and γv denote the fraction of diagonal compression coupling beam, this study proposes that the amount of
transferred by the horizontal or vertical tie in the absence diagonal reinforcement be proportioned using the conven-
of a vertical or horizontal tie, respectively, as suggested in tional flexural analysis based on the flexural demand at the
the CEB-FIP Model Code 1993.14 The terms γh and γv are Design Based Earthquake (DBE) level. By adopting a diag-
expressed as follows onal reinforcement layout, experimental results showed that
a coupling beam would have a good deformation capacity
2 cot θ − 1 during the Maximum Considered Earthquake (MCE) level.
γv = for 0 ≤ γ v ≤ 1 (9)
3 4. Test observations showed that the shear strength is
contributed by the compressive strength of diagonal concrete
2 tan θ − 1 struts and diagonal bars. The compressive strength of a diag-
γh = for 0 ≤ γ h ≤ 1 (10) onal concrete strut can be evaluated using the strut-and-tie
3
model, provided that the shear reinforcement is maintained
in an elastic stage. The current ACI 318-14 confinement
A summary of the strength calculations is presented in requirement for a coupling beam is sufficient to ensure
Column (8) of Table 3 with the shear capacity-to-demand elastic behavior of stirrups in this study.
ratios (VSTM/Vmax) in Column (9). A coupling beam with a
shear capacity-to-demand ratio larger than unity implies AUTHOR BIOS
that its shear capacity can fully support the development Erwin Lim is an Academic Assistant at Institut Teknologi Bandung, Indo-
of flexural behavior at a high displacement level and there- nesia, where he received his BEng in 2005. He received his MS and PhD
from the Department of Civil Engineering at National Taiwan University,
fore possesses a good deformation capacity. Meanwhile, a Taipei, Taiwan, in 2009 and 2015, respectively. His research interests
ratio less than unity suggests that the coupling beam cannot include seismic behavior of reinforced concrete (RC) beam-column joints,
preserve its shear strength and might have a limited defor- shear strength of deep RC members, and RC coupling beams.
mation capacity. Shyh-Jiann Hwang, FACI, is a Professor of civil engineering at National
The shear strength calculation shows that, without the pres- Taiwan University. He received his PhD from the University of California
ence of diagonal bars and solely depending on the diagonal at Berkeley, Berkeley, CA, in 1989. He is a member of Joint ACI-ASCE
Committee 352, Joints and Connections in Monolithic Concrete Structures.
concrete strut, the shear capacity-to-demand ratios (VSTM/Vmax) His research interests include the seismic behavior of beam-column joints,
for CB10-2 and CB20-2 are 0.46 and 0.40, respectively. These shear strength of reinforced concrete members, and seismic retrofitting of
two values strongly suggest that the deformation capacities reinforced concrete structures.
of these two specimens are very poor, which agree well with Ting-Wei Wang is an Engineer at CECI Engineering Consultants, Inc.,
the test results. The proposed shear strength model can also Taiwan. He received his MS from National Taiwan University in 2011.
reasonably capture the enhanced deformation capacity due
Yu-Hsuan Chang received his MS from National Taiwan University
to the presence of diagonal bars by estimating shear capaci- in 2012.
ty-to-demand ratios approaching unity. The predicted capac-
ity-to-demand ratio for CB20-1 is 1.10, indicating ductile ACKNOWLEDGMENTS
flexural failure and agreeing well with the test results. Further- The authors would like to extend their gratitude to the Ministry of
Science and Technology of Taiwan and the National Center for Research on
more, the flexural shear failure mode of Specimen CB10-1 Earthquake Engineering of Taiwan for the funding and the testing facilities.
can be reasonably well predicted by the proposed model with

ACI Structural Journal/March-April 2016 225

 REFERENCES 9. Canbolat, B. A.; Parra-Montesinos, G. J.; and Wight, J. K., “Experi-
1. ACI Committee 318, “Building Code Requirements for Structural mental Study on Seismic Behavior of High-Performance Fiber-Reinforced
Concrete (ACI 318-14) and Commentary (ACI 318R-14),” American Cement Composite Coupling Beams,” ACI Structural Journal, V. 102, No. 1,
Concrete Institute, Farmington Hills, MI, 2014, 519 pp. Jan.-Feb. 2005, pp. 159-166.
2. Paulay, T., and Binney, J. R., “Diagonally Reinforced Coupling Beams 10. ACI Committee 374, “Acceptance Criteria for Moment Frames
of Shear Walls,” Shear in Reinforced Concrete, SP-42, American Concrete Based on Structural Testing (ACI 374.1-05) and Commentary,” American
Institute, Farmington Hills, MI, 1974, pp. 579-598. Concrete Institute, Farmington Hills, MI, 2005, 9 pp.
3. Tassios, T. P.; Moretti, M.; and Bezas, A., “On the Behavior and 11. Mansour, M.; Hsu, T. T. C.; and Lee, J. Y., “Pinching Effect in Hyster-
Ductility of Reinforced Concrete Coupling Beams of Shear Walls,” ACI etic Loops of R/C Shear Elements,” Finite Element Analysis of Reinforced
Structural Journal, V. 93, No. 6, Nov.-Dec. 1996, pp. 711-720. Concrete Structures, SP-205, K. Willam and T. Tanabe, eds., American
4. Tegos, I. A., and Penelis, G., “Seismic Resistance of Short Columns Concrete Institute, Farmington Hills, MI, 2001, pp. 293-321.
and Coupling Beams Reinforced with Inclined Bars,” ACI Structural 12. Hwang, S. J., and Lee, H. J., “Strength Prediction for Disconti-
Journal, V. 85, No. 1, Jan.-Feb. 1988, pp. 82-88. nuity Regions by Softened Strut-and-Tie Model,” Journal of Structural
5. Naish, D.; Fry, A.; Klemencic, R.; and Wallace, J., “Reinforced Engineering, ASCE, V. 128, No. 12, 2002, pp. 1519-1526. doi: 10.1061/
Concrete Coupling Beams—Part I: Testing,” ACI Structural Journal, V. 110, (ASCE)0733-9445(2002)128:12(1519)
No. 6, Nov.-Dec. 2013, pp. 1057-1066. 13. Lima, C.; Martinelli, E.; and Faella, C., “Capacity Models for Shear
6. Moehle, J., Seismic Design of Reinforced Concrete Buildings, Strength of Exterior Joints in RC Frames: Experimental Assessment and
McGraw-Hill Education, New York, 2015, 782 pp. Recalibration,” Bulletin of Earthquake Engineering, V. 10, No. 3, 2012,
7. Wang, T. W., “Seismic Details of Reinforced Concrete Coupling pp. 985-1007. doi: 10.1007/s10518-012-9342-2
Beams for Shear Walls,” master’s thesis, Department of Civil Engineering, 14. CEB-FIP, “CEB-FIP Model Code for Concrete Structures,” Comite
National Taiwan University, Taipei, Taiwan, 2011, 234 pp. (in Chinese) Euro-International du Beton, Federation International de la Precontrainte,
8. Chang, Y. H., “Study on Detailing for Reinforced Concrete Coupling 1993.
Beams of Shear Walls,” master’s thesis, Department of Civil Engineering,
National Taiwan University, Taipei, Taiwan, 2012, 246 pp. (in Chinese)

226 ACI Structural Journal/March-April 2016

View publication stats