ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title no. 93-S68

On the Behavior and Ductility of Reinforced

Concrete Coupling Beams of Shear Walls

by Theodosios P. Tassios, Marina Moretti, andAntonios Bezas

The results of an experimental program on coupling beams Table 1—Specimen characteristics

subjected to cyclic loading are presented. Ten specimens, scaled
down to 1:2, with five different reinforcement layouts and two
different shear ratios (equal to αs = 0.50 and 0.83) have been
tested. An attempt is made to classify the performance of the
specimens according to the ductility they exhibited. The validity of
various ductility criteria is also discussed.

Keywords: beams (supports); crack width and spacing; cyclic loads;

ductility; hysteresis; reinforced concrete; shear properties;stiffness.

RESEARCH SIGNIFICANCE
Coupling beams are an essential structural element in
aseismic design thanks to their ability to considerably reduce
the bending moment of a coupled wall as well as their energy-
dissipation capacity. Their detailing, however, is not always
easy when aiming for ductility. This paper offers quantitative
evidence for the selection of layouts of reinforcement other than
the conventional or biagonal ones, depending on the shear ratio
of the coupling beams. Moreover, the stiffness degradation of
coupling beams can be evaluated and possibly used for a more
pragmatic estimation of the reduction of bending moments of EXPERIMENTAL PROGRAM
the main walls. Test specimens
The characteristics of the specimens are shown in Table 1.
Numbers 1 through 5 in the notation of the specimens desig-
INTRODUCTION nate the layouts of reinforcement, while the letters A and B
Coupling beams with low shear ratios and conventional shear correspond to the shear ratios, 0.50 and 0.83, respectively.
reinforcement (stirrups) tend to fail in a brittle way: diagonal The compressive strength of concrete was measured on
failure, if αs < 0.75, when the stirrups are not sufficient to resist 150 x 300-mm cylinders (six for each test). Specimens CB-1
the whole shear, shear compression or sliding shear failure* at were conventionally reinforced. Specimens CB-2 were made
the boundaries when the stirrups are designed to undertake the with biagonal reinforcement consisting of four bars Ø10 mm,
whole shear force.1 provided along each main diagonal and confined by stirrups
This investigation intended to study the effect of various Ø6 at 50 mm. Specimens CB-3 contained the same rein-
layouts of reinforcement on the hysteretic response of “short” forcement as CB-1, with additional bent-up bars intersecting
(αs = 0.50, Specimens CB-A) and “medium” (αs = 0.83, Spec- at the midheight of the beam (thus contributing to the sliding
imens CB-B) coupling beams. This terminology is based on resistance without considerably increasing the flexural
experimental observations found in the literature,1,2 according capacity of the beams). Specimens CB-4 and CB-5
to which coupling beams may be classified as follows: contained long dowels (CB-4) and short dowels across the
αs = M/Vh = l/2h ≤ 0.50: short elements ends of the beams (CB-5); the dowels were located in the
0.50 < αs ≤ 1.00: medium elements
1.00 < αs: normal elements ACI Structural Journal, V. 93, No. 6, November-December 1996.
Received May 1, 1995, and reviewed under Institute publication policies. Copy-
l and h denote the length and height of the beam. right © 1996, American Concrete Institute. All rights reserved, including the making of
copies unless permission is obtained from the copyright proprietors. Pertinent discussion
will be published in the September-October 1997 ACI Structural Journal if received
*.Sliding shear failure is caused by the deterioration of the end parts of the beam, by May 1, 1997.
alternatively subjected to compression-tension and reversed shear.

ACI Structural Journal/November-December 1996 711

 Table 2—Properties of reenforcement bars
Theodosios P. Tassios, FACI, is a professor at the National Technical University of
Athens. He is honorary president of CEB and a coauthor of the European Code No. 8 Bar diameter, mm Yield stress fy, MPa Tensile strength fst, MPa
on Aseismic Design. His research interests include reinforced concrete under seismic
conditions. 20 450 745
18 453 742
Marina Moretti is a graduate research assistant and PhD candidate in civil engineering
14 451 692
at the National Technical University of Athens. Her research interests include
reinforced concrete structures under seismic loading. 12 484 771
10 504 764
Antonios Bezas, a structural engineer, graduated from the National Technical University
of Athens in 1986. His research interests include the behavior and ductility of 8 296 407
reinforced concrete beams. 6 281 380
Note: 1 MPa = 145 psi; 1 mm = 0.0394 in.

rical with zero moment in the middle of the beam. The out-
of-plane displacements of the specimen were prevented by
adequate lateral arrangements. The loading conditions were
similar to those of a double fixed beam coupling two walls
subjected to seismic loading.
The loading history of the specimen is shown in Fig. 2. All
specimens were initially subjected to a displacement δ = δy
that led to yielding of the main reinforcement (longitudinal
or diagonal). Subsequently, three full cycles were performed
at each of the levels of displacement: δ = δy, 2δy, and 3δy
(μ = 1,2,3). Finally, if the specimens had not yet failed, they
were further loaded up to failure. Displacements were
measured by means of LDTs, whereas strains of the rein-
forcement were measured by means of strain gages.

Fig. 1—Elevation of testing rig. TEST RESULTS

The main results are listed in Table 3 together with analyt-
ically predicted values, according to the CEB-FIP Model
Code 78.

Load-displacement diagrams
Some of the load-displacement (V-δ) diagrams of the
beams tested are shown in Fig. 3. The analytically predicted
monotonic resistance “Vu,anal” of the specimens (see Table 3)
is shown on these diagrams with dotted lines. The fact that
Beams CB-2 and CB-3 developed strengths in excess of their
theoretical ultimate (see Table 3) may be attributed to a
Fig. 2—Standard loading history applied (imposed dis- possible underestimating of Vu,anal by the analytical methods
placements). used (i.e., by neglecting the contribution of concrete, in the
case of Beams CB-2, and of the bent-up bars in Beams CB-3.)
Intense pinching effect can be observed in the V-δ
middle vertical plane of the beams. Specimens CB-3, CB-4, diagrams of the specimens with conventional reinforcement
and CB-5 were planned to control the sliding failure at the CB-1 and with Dowels CB-4 and CB-5, especially for large
boundaries. displacements (δ ≥ 2δy). The pinching effect diminishes
In all but CB-2 specimens, stirrups Ø8 at 75 mm were when bent-up bars are provided (Specimens CB-3); it practi-
provided; Specimens CB-2 contained Ø6 at 120 mm, since cally disappears in the case of diagonally reinforced Speci-
their shear force was supposed to be undertaken by biagonal mens CB-2: the V-δ diagrams of CB-2 are somehow similar
reinforcement. The properties of the reinforcement are to those of a steel bar; this is further evidence that the
shown in Table 2. behavior of these specimens is mainly governed by steel.3
Test setup and loading procedure Crack pattern and mode of failure
The specimens were tested in vertical position and were The crack pattern as well as the mode of failure of the
subjected to cyclic shear displacements using the testing rig specimens differed according to the arrangement of rein-
shown in Fig. 1. One end of the specimen was free, while the forcement and the value of the shear ratio. During the first
other was fully fixed to the reaction frame by means of: a) cycles of loading, Specimens CB-1, CB-2, and CB-3 with αs =
wedged steel elements, and b) a hydraulic jack (500 kN 0.50 exhibited cracks along the main diagonals, while those
capacity) that imposed a constant compressive force to the with αs = 0.83 developed bending shear cracks near the ends
fixed part. (Fig. 4 and 5). Vertical sliding cracks appeared at the bound-
Displacements were induced to the free end of the spec- aries of all these specimens, but they did not expand (except
imen by means of an actuator (double action 1000/2000 kN). for Specimen CB-1A, in which the opening of such a crack
The axis of the actuator coincided with the axis of the spec- reached 2 mm at Cycle n = 5). Specimens CB-5 and CB-5,
imen; thus, the moment-diagram produced was antisymmet- which were reinforced with dowels, exhibited compressive

712 ACI Structural Journal/November-December 1996

 Fig. 3—Load-displacement (V-δ) diagrams for beams tested (1 mm =
0.0394 in.; 1 kN = 0.225 kips).

shear cracks along the main diagonals, irrespective of the observed. Specimens CB-4 failed at their central region, and
value of the shear ratio. The presence of dowels prevented an CB-5 at their end regions, all in shear compression.
extensive appearance of sliding cracks at the beam ends.
Specimen CB-1A failed in diagonal tension along the Observations on behavior of reinforcement
main diagonals, while Specimen CB-1B failed in shear Before the appearance of diagonal cracking, the distribu-
compression at the end regions, with indications of simulta- tion of the strains measured along the main reinforcement
neous bond failure. Specimens CB-2 failed due to buckling was very close to the one predicted by the flexural theory.
of the diagonal reinforcement near the ends of the beams; After diagonal cracking, the strain-distribution changes2-4
however, thanks to the good confinement (Ø6 at 50 mm) of and can no longer be described by the flexural theory. A
the diagonal bars, the failure was not brittle as in other much better estimator of the actual strain distribution after
similar tests reported in the literature.1 diagonal cracking was offered by Paulay.2 However, it has to
Specimens CB-3 developed cracks along the local diagonal be noted that the abrupt change in the distribution of the
bars, probably because of their inadequate anchorage. Bond- longitudinal strains was more evident in the case of the spec-
splitting failure along the main reinforcement and shear imens with the lower shear ratio αs = 0.50, in which diagonal
compression failure at the corners of the beams were also cracking occurred.

ACI Structural Journal/November-December 1996 713

Table 3—Experimental data and analytically predicted response of beams tested
Experimental Analytical V max *
--------------
-
Specimen n* δy,* mm Vd,cr ,* kN Vy,* kN Vmax,* kN Vwd,* kN VR,* kN V(MR),* kN V u, anal
CB-1A 7 7.8 179 179 212 170 221 229 0.96
CB-1B 10 6.4 63 100 124 100 124 122 1.02
CB-2A 10 7.4 136 214 283 — 212 †
1.33
CB-2B 10 8.5 101 115 170 — 142 † 1.20
CB-3A 9 8.5 150 183 302 170 221 229 1.37
CB-3B 10 6.4 99 121 151 100 130 122 1.24
CB-4A 7 10.5 162 282 282 170 342 448 0.82
CB-4B 6 10.8 88 162 162 100 206 170 0.95
CB-5A 5 12.0 152 237 237 170 221 229 1.07
CB-5B 7 11.2 100 130 130 100 130 122 1.07
*
See Notation.
†
Biagonally reinforced specimens can only fail by failure of their diagonals; “flexural” failure does not make sense in this case.
Note: 1 mm = 0.0394 in.; 1 kN = 0.225 kip.

Fig. 5—Crack patterns of several specimens at various levels

of imposed displacement ductilities.

maximum shear capacity of the beams was achieved.2 The

only stirrups that did not yield, according to the measure-
ments of the strain gages, were those at the edges of Beams
CB-3, CB-4, and CB-5; probably, the dowels and the bend-
up bars transferred a great part of the shear force from the
end blocks to the beams.
In Fig. 7 and 8, the strains of the stirrups along Beams CB-1A
and CB-1B are shown for different values of the applied
load. The strain gages were situated on both legs of the stir-
rups along the main diagonals of the beams. Each measure-
ment is the average strain of both legs of the stirrups.
Fig. 4—Crack patterns of Specimens CB-1B (αs = 0.83) and
The differences in the strains measured (Fig. 7 and 8)
CB-2A (αs = 0.50) after three full reversals for imposed
should be interpreted in connection with the geometry of
ductility factor equal to 3 (n = 9).
cracking of the respective specimens, as portrayed in Fig. 5.
In most cases, small shear ratio (αs = 0.50) specimens (one
In Fig. 6, the strain distributions [εs] along the lower longi- main diagonal crack) exhibited large strains in the stirrups
tudinal bar of Specimen CB-1A, both experimentally situated in their midspan, whereas larger shear ratio (αs =
measured (exp) and analytically predicted (flexural theory), 0.83) specimens (inclined cracks, not necessarily corner-to-
are given for different values of applied shear force at Cycle n corner) showed a remarkable inactivity of midspan stirrups.
= 1+. The strain distribution predicted by the model of This is further evidence that, for specimens with shear ratio
Paulay2 is given for the load “Vd,cr” at which the diagonal αs = 0.50, the prevailing shear-carrying mechanism seems to
cracking appeared; the ratio η = Vw /Vd,cr used in Paulay’s be the diagonal compression strut, while for specimens with
equation is also noted, Vw being the measured shear force αs = 0.83 a broader truss is a more appropriate model for
resisted by all stirrups across the diagonal crack at V = Vd,cr. shear transfer. It can also be observed that in Fig. 7, stirrups
The general yielding of the stirrups in all specimens situated close to the beam ends are activated only at their
occurred, approximately, at Cycle n = 4+, during which the points where the diagonal strut is acting.

714 ACI Structural Journal/November-December 1996

 Fig. 7—Variation of stirrup-strains along Beam CB-1A as
function of applied load (n = 1+).

Fig. 6—Experimental and analytically predicted strain

distributions along longitudinal reinforcement of Specimens
CB-1A during first cycle: n = 1+ (1 kN = 0.225 kip).

In Fig. 9, the ratio of the average maximum shear force V n Fig. 8—Variation of stirrup-strains along Beam CB-1B as
attained at each cycle “n” to the shear capacity Vwd of the function of applied load (n = 1+).
stirrups (see Table 3) is represented for the specimens with
αs = 0.83. It is obvious that Specimen CB-3B undertakes
higher shear forces than the conventionally reinforced CB-1B
due to the contribution of the bent-up bars. The stirrups seem
to guarantee to Specimens CB-1B and CB-3B a shear
response equal to the shear capacity Vwd of the stirrups up to
Cycle n = 7 (δ = 3δy). The behavior of Specimens CB-4B and
CB-5B seems to be independent of the amount of shear rein-
forcement, since the shear transfer is mainly achieved
through dowels. The abrupt loss in shear response of Speci-
mens CB-4B and CB-5B with the increase of cycles is due to
the intense cracking of concrete. On the basis of available
experimental data, similar conclusions can be drawn
regarding specimens with αs = 0.50.

Role of layout of reinforcement

As was expected, the biagonally reinforced specimens
have demonstrated the most ductile behavior among all the Fig. 9—Variation of average normalized maximum shear
reinforcement layouts tested. In fact, after the diagonal force response V n with number of cycles.
cracking of concrete, the shear forces were directly under-
taken axially by the diagonal bars; thus, the behavior of these
specimens was finally governed by steel. onals could not be hindered, and hence, a premature force-
The reinforcement layout tested in Specimens CB-3 led to response degradation was observed. The large dowel diameters
an increased ultimate capacity (30 percent higher than used may also explain the ineffectiveness of this layout.
predicted) and to an improved overall behavior if compared The conventionally reinforced Specimens CB-1 behaved
to the conventional reinforcement, especially in the case of in a rather brittle manner, especially Specimen CB-1A with
the lower shear ratio. The bent-up bars restrained both the αs = 0.50. Their ultimate capacity was, nevertheless, almost
appearance of sliding cracks and the expansion of cracks in identical to the analytically predicted one. Despite their
the central regions of the beam (the latter being attributed to the intense cracking and considerable strength and stiffness
beneficial effect of in-plane compression forces “FA” developed degradation, they exhibited a more ductile behavior than the
along the curved parts of these bent-up bars [see Fig. 10]). specimens with dowels.
In Specimens CB-4 and CB-5, the dowels prevented the The average response degradation of Specimens CB-B at
appearance of sliding cracks at the boundaries. However, the the levels of imposed ductility μ = 2,3 is plotted against the
presence of dowels did not offer a better overall behavior; the number “n” of cycles in Fig. 11 and seems to be in accor-
appearance of compression shear cracks along the main diag- dance with the relative global behavior of the beams tested.

ACI Structural Journal/November-December 1996 715

Fig. 10—In-plane compression forces “FA” induced by
bent-up bars of Specimens CB-3.

Fig. 13—Secant stiffness degradation of all specimens

tested as function of number of cycles.

Table 4—Values of calculated ductility factors μd

Fig. 11—Average normalized values ΔV of response degra-
and μE for specimens tested
dation of Specimens CB-B for various cycles at levels of
imposed ductility μ = 2 and μ = 3. αs Specimen μd μE
CB-1A 3.1 2.1
CB-2A 5.6 4.7
0.50 CB-3A 2.8 2.9
CB-4A 4.6 1.8
CB-5A 1.6 0.7
CB-1B 2.9 3.7
CB-2B 5.2 5.8
0.83 CB-3B 4.1 3.9
CB-4B 2.7 1.3
CB-5B 1.6 1.5

of coupling beams to be considered in analysis of coupled

wall systems.

DUCTILITY CRITERIA
An attempt is made here to evaluate the performance of the
Fig. 12—Average values of normalized force response tested specimens in terms of ductility and energy dissipation
degradation ΔV as function of shear ratio. capacity using the normal ductility criteria as well as some
other criteria proposed in this paper.
Role of shear ratio 1. The displacement ductility factor “μd” is first calculated,
Different initial crack patterns were generally observed in as shown in Fig. 14. The envelope-curve of the third cycles
short specimens (CB-A) and in the medium ones (CB-B): at at every level of imposed ductility μ = 1,2,3 is drawn for each
the first stages of loading, cracks along the main diagonals specimen (Cycles n = 3, n = 6, and n = 9, respectively). The
appeared in CB-A beams, whereas inclined cracks devel- factor “μd” is calculated at the 85 percent level of Vmax exhib-
oped near the ends of Beams CB-B. Apparently, different ited by the conventionally reinforced specimen. The average
mechanisms of shear transfer are activated in each case: a μd for both directions of loading is reported.
diagonal compressive strut for αs = 0.50 and a more truss- The values of “μd” thus calculated (see Table 4) do not
like mechanism for αs = 0.83. seem to accurately describe the actual behavior of the speci-
Fig. 12 shows how the shear ratio “αs” affects the average mens; for instance, they do not depict the better performance
values of normalized response degradation ΔV9 = Vmax – V9 of the CB-B specimens compared to the CB-A ones. Further-
under the imposed level of ductility μ = 3 (δ = 3δy). It clearly more, they are, in certain cases, misleading (i.e., for CB-4,
appears that the response degradation of Specimens CB-1, which behaved in a brittle manner).
CB-2, and CB-3 tends to diminish as the shear ratio increases. 2. To better describe the performance of specimens in
terms of ductility, a method proposed by Japanese
Stiffness degradation researchers5 is used, appropriately adapted for relatively
As with all relatively brittle elements, the stiffness of brittle members such as coupling beams. A ductility factor
coupling beams is rapidly degrading1,4 with postelastic “μE” in terms of “stored energy” is calculated. To this end,
cycling. Fig. 13 shows the variation of the secant stiffness the following areas are calculated: that enclosed by the first
degradation with the number of cycles for the CB-B speci- half-cycles (Ay1 at yield) and the one enclosed by the enve-
mens. This may be useful in assessing the realistic stiffness lope-curve of the third cycles (Au3 at failure [see Fig. 15]). It

716 ACI Structural Journal/November-December 1996

Fig. 14—Definition of calculated displacement ductility
factor “μd,” accounting for cyclic behavior.
Fig. 17—Normalized shear response against imposed cyclic
displacement ductility factors for all reinforcement layouts
tested.

Fig. 15—Definition of factor “μE.”

Fig. 18—Definition of dissipated energy values in case of

cyclic loading.

Fig. 16—Definition of cyclic displacement ductility factor “μd,P.”

is noted that the so-called “stored energy” differs from the

dissipated energy (calculated from the area enclosed by the Fig. 19—Correlation between average normalized shear
hysteresis loops). The area Ay1 in Fig. 15 expresses the stored response V n :Vu,anal and normalized energy Wpl,n:Wpl,y
“yield energy” Ey, while the area Au3 corresponds to the dissipated during nth cycle of loading.
stored “ultimate energy” Eu of the system. The factors μE =
Eu/Ey thus calculated are shown in Table 4 and seem to be tation was not able to differentiate the behavior of specimens
better estimators of the overall performance of the specimens, with different shear ratios.
since they take into account both the resistance degradation 4. Another evaluation could be based on the relationship
and the plastic deformation capacity of the system. between a “normalized shear response” V n : Vu,anal and the
“normalized dissipated energy” Wpl,n:Wpl,y calculated for the
3. We also attempted to describe the performance of the
relevant n-cycles (Fig. 18). The resulting ranking (Fig. 19) of
specimens by means of a cyclic displacement ductility factor the tested layouts seems to be in accordance with the actual
“μd,P” defined in Fig. 166 (δy being the observed yield behavior of the specimens.
displacement). On the other hand, the ratio of the shear 5. For a good seismic performance of a coupling beam, the
force-response ⏐Vn⏐ at the nth-half cycle to the analytically following characteristics are required: a) a large amount of
predicted resistance Vu,anal of each specimen (Table 3) was dissipated energy, and b) a capacity for “sustaining” as many
calculated. These “normalized shear force-response” values cycles “n” of induced postelastic displacements as possible.
versus the imposed cyclic displacement ductility factors Having observed that only the very first cycles influence the
“μd,P” are represented in Fig. 17; only mean curves are plotted behavior of the specimen (there is a practical stabilization of
for both shear ratios. The curves accurately express the higher force-response after, say, four full cycles at each postelastic
response and the larger plastic deformation capacity of Spec- imposed displacement), the square root of the total number
imens CB-2, as well as the better behavior of Specimens CB-3 of cycles “nt” seems to be more significant in predicting the
if compared to Specimens CB-1. However, such a represen- force-response degradation. The total hysteretic dissipated

ACI Structural Journal/November-December 1996 717

Fig. 20—Classification of reinforcement layouts tested
according to criterion n t * ∑ Wpl up to failure.
Fig. 22(a)—Geometry and detailing of specimens with
“conventional” reinforcement, shear ratios αs = 0.50
and αs = 0.83.

Fig. 21—(a) Values of various ductility criteria for each

reinforcement layout tested; (b) force response degradation
of specimens at μ = 2 for reinforcement layouts tested.

energy ∑Wpl during the entire loading history up to failure,

multiplied by n t , is illustrated in Fig. 20. The overall
ranking seems to be correct.
6. For the sake of comparison, Fig. 21(a) shows for all the
specimens the response degradation and the calculated values
of the ductility criteria μE, n t * ∑ Wpl, as well as the “cumu-
lative” cyclic displacement ductility factor ∑ μd,P, where
μd,P denotes the factor defined in Fig. 16 (summation covers
all cycles up to failure). Fig. 21(b) shows the normalized
response degradation ΔV 6 : Vmax = ( V max – V 6 ) : Vmax (the
average for both directions of loading) for the imposed level of Fig. 22(b)—Geometry and detailing of other specimens with
ductility μ = 2 (δ = 2δy) that was attained by all the specimens. shear ratios αs = 0.50.

CONCLUSIONS 3. As expected, the performance of Specimens CB-1 was

The following ranking of the efficiency of the reinforce- unsatisfactory; small plastic deformation ability and consid-
ment layouts can be established: erable strength and rigidity degradation was observed.
1. The higher overall performance of the biagonally rein- 4. The specimens with added short dowels (CB-5), although
forced beams CB-2 among all the layouts tested was once they have exhibited no sliding at their ends, have shown the
again confirmed. most brittle behavior among all the specimens tested.
2. The introduction of bent-up bars in Specimens CB-3 led 5. However, the specimens with long dowels (CB-4)
to an increased ultimate capacity (30 percent higher than behaved slightly better than the specimens with short dowels.
predicted) and to an improved overall behavior if compared 6. The performance superiority of Reinforcement Layouts 2
to conventionally reinforced Specimens CB-1. and 3 (compared to the conventional Layout 1) becomes

718 ACI Structural Journal/November-December 1996

more evident for the shear ratio αs = 0.50 and tends to fst = tensile strength of steel
diminish as the shear ratio increases. fy = yield stress of steel
h = height of beam
As expected, the specimens with higher shear ratio (αs = l = length of beam
0.83) behaved in a more ductile manner than the specimens M = maximum moment acting at end parts of beam
with the smaller one (αs = 0.50). They were, in fact, less n = number of cycles performed
cracked, had shown greater plastic deformation ability, and nt = number of full cycles up to failure
exhibited smaller resistance and rigidity degradations. Based t = width of beam
V = shear force
on crack patterns and other experimental evidence, one can Vd,cr = shear force at diagonal cracking
conclude that the predominant mechanism of shear transfer Vmax = maximum shear force-response attained
in the specimens with αs = 0.50 is the diagonal compressive V(MR) = shear resistance corresponding to flexural capacity of beam for
strut, while in specimens with αs = 0.83 a truss-like mechanism doubly fixed condition: V(MR) = 2MR/l
seems to be a better model. Vn = maximum shear force-response attained at Cycle “n”
Vn = average maximum shear force-response at Cycles n+, n–
Among the ductility criteria considered, the proposed VR = analytical shear capacity including contribution of: 1) con-
criterion “μE” seems to be the most representative of the crete, Specimens CB-1, CB-3, CB-4, and CB-5; 2) biagonal
observed behavior of the beams tested. The calculation of the reinforcement, Specimens CB-2; and 3) long dowels, Speci-
factor “μE” is based on the shape of the envelopes of hyster- mens CB-4 (assuming 50 percent activation of dowels)
esis loops both for n = 1 at yield and for n = 3; thus, it Vu,anal = min [VR, V(MR)]
Vw = measured total shear force taken by stirrups at diagonal cracking
accounts for both the response degradation and energy dissi- Vwd = shear resistance of stirrups
pation. It is also worth mentioning that, for members Vy = shear force at yielding of main reinforcement
subjected to the same loading history, one of the simplest Wpl,n = energy dissipated at nth cycle of loading
criteria is the rate of response-degradation at the same level Wpl,y = energy dissipated at cycle that leads to yielding of main
of imposed ductility. However, when comparing alternative reinforcement
αs = shear ratio = M/Vh
structural solutions, it is recommended to evaluate the δ = imposed displacement
ductility of members by a combination of ductility criteria δu = ultimate displacement of specimen
rather than by means of a single criterion. δy = displacement leading to yielding of main reinforcement
Some design recommendations may also be formulated: ΔV n = average force response degradation during cycles of loading at
1. The present experimental research seems to confirm given level of imposed ductility
εs = strain of main reinforcement
that for coupling beams with shear ratios “αs” lower than εw = strain of stirrups
approximately 3/4 (a limit value under which the diagonal εwy = strain of stirrups at yielding
strut seems to be activated), biagonal reinforcement under- ∑ Wpl = total dissipated energy
taking the entire shear force and bending moment is the most ∑ μ d,P= cumulative cyclic displacement ductility factor
recommended layout. The good confinement of the biagonal μ = level of imposed displacement ductility = δ/δy
μd = displacement ductility factor = δu/δy
reinforcement, to avoid a brittle failure after possible buck- μd,P = cyclic displacement ductility factor (as shown in Fig. 16)
ling of diagonal bars, is indispensable. μE = ductility factor in terms of stored energy (as shown in Fig. 15)
2. For shear ratios “αs” higher than approximately 3/4, ρd = reinforcement ratio of biagonal bars or bent-up bars or dowels
long dowels in midheight and bent-up bars at the ends of the ρl = reinforcement ratio of “conventional” longitudinal bars
beam may profitably be used in addition to the conventional Ød = diameter of diagonal bars, bent-up bars, or dowels
Øl = diameter of “conventional” longitudinal bars
reinforcement. This arrangement seems to considerably
improve the behavior of conventionally reinforced coupling
beams, and although it may not lead to as satisfactory results REFERENCES
1. Paulay, T., “Simulated Seismic Loading of Spandrel Beams,” Journal
as the biagonal layout, this arrangement seems to be imple- of the Structural Division, American Society of Civil Engineers, V. 97,
mented more easily. No. ST9, Sept. 1971, pp. 2407-2419.
The bent-up bars and dowels should be calculated to 2. Paulay, T., “Coupling Beams of Reinforced Concrete Shear Walls,”
undertake 150 percent of the end shear force, so as to prevent Journal of the Structural Division, American Society of Civil Engineers,
V. 97, No. ST3, Mar. 1971, pp. 843-862.
shear sliding; however, all conventionally needed stirrup 3. Paulay, T., and Binney, J. R., “Diagonally Reinforced Coupling Beams of
reinforcement should also be provided. In addition, the in- Shear Walls,” Shear in Reinforced Concrete, SP-42, V. 2, American Concrete
plane local compression induced to concrete by the curved Institute, Detroit, 1974, pp. 579-598.
parts of the bent-up bars should be checked. Nevertheless, 4. Paulay, T., “Coupling of Reinforced Concrete Shear Walls,” Proceed-
ings, 4WCEE, Santiago, Chile, 1969, B-2, pp. 75-90.
further research is needed to reach more quantitative
5. Maruyama, K.; Sugita, A.; and Shimizu, K., “Some Considerations on
evidence on this layout of reinforcement. Ductility of Reinforced Concrete Members,” Transactions of the Japan
3. For shear ratios higher than approximately 4/3, normal Concrete Institute, V. 11, 1989, pp. 293-300.
(“conventional”) detailing may be used. 6. Paulay, T., and Santhakumar, A. R., “Ductile Behavior of Coupled
Shear Walls,” Journal of the Structural Division, American Society of Civil
Engineers, V. 102, No. ST1, Jan. 1976, pp. 93-108.
ACKNOWLEDGMENTS 7. CEB-FIP Model Code 78.
The authors gratefully acknowledge the financing of this investigation by 8. Yu, G. R.; Wang, W. H.; Yao, C. D.; Ni, V.; and Liu, J. C., “Design of
the Organization for Aseismic Protection (OASP) and the contracting Sintered Fly Ash Aggregate Concrete Coupling Beams Subjected to
company Technodomiki. Reversed Cyclic Loadings,” International Symposium of Fundamental
Theory of Reinforced and Prestressed Concrete, Sept. 1986, NIT, Nanjing,
CONVERSION FACTORS China, pp. 727-734.
1 mm = 0.0394 in. 9. Chongghang, W.; Zongzhe, W.; Ping, C.; and Zhiren, D., “Ductility of
1 kN = 0.225 kip Reinforced Concrete Shear Walls,” International Symposium of Funda-
mental Theory of Reinforced and Prestressed Concrete, Sept. 1986, NIT,
1 N/mm2 = 1 MPa = 145 psi
Nanjing, China, pp. 751-758.
10. Ding Dajun et al., “Experimental Investigation on High-Ductility
NOTATION Reinforced Concrete Shear Walls,” Scuola di Specializzazione in Costruz-
fc′ = cylinder compressive strength of concrete ioni in C.A., Studi e Ricerche, V. 13, 1992, pp. 395-419.

ACI Structural Journal/November-December 1996 719