Sei sulla pagina 1di 13

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

net/publication/281830287

Seismic behavior of low-aspect-ratio reinforced concrete shear walls

Article  in  Aci Structural Journal · September 2015


DOI: 10.14359/51687709

CITATIONS READS

29 1,143

3 authors:

Bismarck Luna Jonathan Rivera


Praxair Inc., Tonawanda, NY University at Buffalo, The State University of New York
23 PUBLICATIONS   194 CITATIONS    25 PUBLICATIONS   199 CITATIONS   

SEE PROFILE SEE PROFILE

Andrew Whittaker
University at Buffalo, The State University of New York
501 PUBLICATIONS   6,957 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Seismic isolation of large light water reactors using elastomeric and sliding bearings View project

Verification and Validation of Numerical Models for Seismic Fluid-Structure-Interaction Analysis of Advanced Reactors View project

All content following this page was uploaded by Bismarck Luna on 17 September 2015.

The user has requested enhancement of the downloaded file.


ACI STRUCTURAL JOURNAL TECHNICAL PAPER
Title No. 112-S48

Seismic Behavior of Low-Aspect-Ratio Reinforced


Concrete Shear Walls
by Bismarck N. Luna, Jonathan P. Rivera, and Andrew S. Whittaker

Twelve low-aspect-ratio reinforced concrete walls were constructed sion of the global force-drift relationships are presented.
and tested at the University at Buffalo to develop validated equa- Measured peak in-plane shear strengths are compared
tions for peak shear strength and hysteretic rules for nonlinear with values predicted using equations from building codes
response-history analysis. The pretest analysis and construction and standards of practice. The importance of out-of-plane
of the walls are described. Global force-displacement relation-
loading and deformation on in-plane strength is identified.
ships are presented. Currently used equations for peak nominal
Values of initial stiffness are reported and the loss of stiff-
(in-plane) shear strength do not predict the measured resistance
of the walls. Out-of-plane forces and deformations affected peak ness and shear strength with repeated cycling is character-
in-plane shear strength. The resistance of the 12 walls degraded ized. Data presented in this report provide information to
quickly with repeated cycling at displacements greater than those enable the characterization of the hysteretic response of
associated with peak strength. The initial stiffnesses of the test low-aspect-ratio shear walls.
specimens were substantially lower than calculated using equa-
tions in design standards. TEST SPECIMENS AND SETUP
Twelve large-size, low-aspect-ratio, rectangular, RC shear
Keywords: reinforced concrete; shear strength; shear walls; stiffness;
walls (denoted SW1 to SW12) were designed, constructed,
strength degradation.
and tested at the NEES facility at UB. The length and
thickness of the test specimens were 10 ft (3.05 m) and
INTRODUCTION
8 in. (203 mm), respectively. The design variables consid-
Low-aspect-ratio structural walls (height-length ratio of 2
ered in developing the portfolio of test specimens included
or less) are widely used for low- and medium-rise buildings
wall aspect ratio (hw/lw), day-of-test concrete compressive
and safety-related nuclear structures. Standards of practice
strength (fc′), vertical and horizontal reinforcement ratio (ρl,
in the United States and abroad provide equations to esti-
ρt), yield and ultimate strengths of the reinforcement (fy, fu),
mate the maximum shear strength of a reinforced concrete
mechanical splices in vertical reinforcement, and the pres-
shear wall, but these equations are inaccurate and insuffi-
ence of boundary elements. The properties of the walls are
ciently parameterized (for example, Gulec et al. [2008] and
summarized in Table 1. A photograph of Specimen SW2 is
Del Carpio Ramos et al. [2012]). Equations for uncracked
presented in Fig. 1(a). The horizontal and vertical reinforce-
and cracked stiffness are available for design but have not
ments in the webs of the walls were spaced uniformly. The
been validated by large-scale testing.
vertical reinforcement in Specimens SW2 and SW3 was
To improve the profession’s understanding of the cyclic
joined mechanically using couplers at approximately 14 in.
response of low-aspect-ratio walls, the U.S. National Science
(356 mm) above the top of the foundation. The horizontal
Foundation (NSF) funded a Network for Earthquake Engi-
reinforcement in the walls was terminated with 90-degree
neering Simulation (NEES) research project on shear walls
hooks, as shown in Fig. 1(b) and 1(c). Boundary elements
of conventional and composite construction. Sixteen rect-
were constructed at the ends of Specimens SW11 and SW12
angular, low-aspect-ratio concrete shear walls were built
using: 1) crossties with seismic hooks (at the level of hori-
and tested at the NEES facility at the University at Buffalo
zontal web reinforcement, as shown in Fig. 1(c)); and 2) stir-
(UB): 12 conventionally reinforced concrete (RC) walls and
rups and crosstie with alternating seismic hooks (between
four steel-concrete (SC) walls. Two additional RC shear
the horizontal web reinforcement, as shown in Fig. 1(d)).
walls were built and tested using hybrid simulation at the
The foundation for each wall was cast approximately 7 days
University of California, Berkeley. This paper addresses the
before the wall; the wall-foundation interface was intention-
response of the 12 RC shear walls tested at UB. The experi-
ally roughened before the wall was cast.
mental data from the tests of the 12 RC walls are available on
The test program was executed in two phases: Phase I
NEES data repository (Luna et al. 2013). Information on the
(SW1-SW7) and Phase II (SW8-SW12). The Phase II walls
four SC walls is reported by Epackachi et al. (2014, 2015).
were designed after preliminary analysis of the test data
Whyte and Stojadinovic (2013) documented the results of
the hybrid simulation tests.
ACI Structural Journal, V. 112, No. 5, September-October 2015.
MS No. S-2014-161.R1, doi: 10.14359/51687709, received October 15, 2014 , and
RESEARCH SIGNIFICANCE reviewed under Institute publication policies. Copyright © 2015, American Concrete
Institute. All rights reserved, including the making of copies unless permission is
This paper presents data from the tests of 12 large-size obtained from the copyright proprietors. Pertinent discussion including author’s
RC shear walls built and tested at UB. Analysis and discus- 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/September-October 2015 593


Table 1—Properties of wall specimens
Web Boundary element
Wall hw/lw ρl, % ρt, % ρl, % ρt, % fc′, ksi fy, ksi fu, ksi
SW1 0.94 0.67 0.67 — — 3.6 67 102
SW2 1.0 1.0 — — 7.0 63 87
SW3 0.54 0.67 0.67 — — 7.8 63 87
SW4 0.33 0.33 — — 4.2 67 102
SW5 1.0 1.0 — — 4.3 67 102
SW6 0.33 0.67 0.67 — — 3.8 67 102
SW7 0.33 0.33 — — 3.8 67 102
SW8 1.5 1.5 — — 3.5 67 102
SW9 1.5 0.67 — — 4.3 67 102
SW10 0.54 1.5 0.33 — — 4.6 67 102
SW11 0.67 0.67 1.5 1.5 5.0 67 102
SW12 0.33 0.33 2.0 2.0 5.0 67 102

Note: 1 ksi = 6.9 MPa.

Fig. 1—Specimen details.


from the Phase I walls. Preconstruction analysis of the walls ment within each actuator clevis.) Load cells in the actuators
indicated that SW1 would fail in a combination of flexure measured the applied load.
and shear, and that SW2 through SW12 would fail in shear. No axial loads were applied to the specimens because: 1)
The foundation of each shear wall was post-tensioned axial stresses in low-aspect-ratio walls are typically small, as
to the strong floor in the laboratory using fourteen 1.5 in. measured by a fraction of the product of their cross-sectional
(38 mm) diameter bars. Lateral loads were applied to the area and the concrete compressive strength; and 2) applying
walls using two high-force-capacity actuators that were and maintaining even low axial loads on a wall with an
horizontally inclined by 9 degrees with respect to the longi- area of approximately 1000 in.2 (0.65 m2) at relatively large
tudinal axis of the walls (Fig. 2(a)). The actuators imposed lateral displacements is extremely difficult.
displacements on the shear walls via custom-made brackets
and thick steel plates that were post-tensioned to either side INSTRUMENTATION
of the specimen. Incremented cyclic displacements were Traditional displacement transducers were used to
imposed on the walls to indirectly simulate the effects of measure the in-plane and out-of-plane displacements of each
earthquake shaking. Displacement transducers in the actu- wall. An optical measurement system (Nikon Metrology Inc.
ators were used to control the lateral displacement of the 2013) was used to monitor displacements across one face
walls. (Discrepancies between the input and actual displace- of each wall, using LED sensors installed on a rectangular
ments are inevitable with this control strategy at very small grid. Data from this dense grid of sensors enabled calcula-
values of displacement because of nonlinear slip and move- tion of the contributions of flexure, shear, and base slip to

594 ACI Structural Journal/September-October 2015


Fig. 3—Predicted force-displacement relationships for SW5.
(Note: 1 kip = 4.45 kN.)
drift relationships for SW5 for: 1) a concrete compressive
strength of 6 ksi (41 MPa)—50% greater than the target
of 4 ksi (28 MPa); and 2) reinforcing bar yield strength of
75 ksi (517 MPa)—25% greater than the proposed minimum
specified value of 60 ksi (414 MPa). The predicted peak
shear strength for SW5 was 732 kip (3256 kN), which was
the greatest for the Phase I walls, and this resistance, with a
margin, was used to design the test fixture.

LOADING PROTOCOL
The test specimens were subjected to reversed cyclic,
displacement-controlled loadings to indirectly simulate
the effect of earthquake shaking. The guidelines provided
by ATC-24 (1992), ACI 374.1-05 (ACI Committee
Fig. 2—Specimen setup.
374 2005) and a 2009 working draft of ACI 374.2R-13
the total lateral displacement, which are reported elsewhere. (ACI Committee 374 2013) were considered in the develop-
Traditional transducers and LED sensors were installed to ment of the loading protocol.
monitor lateral displacement and rotation of the foundation. The protocol was developed on the basis of results of
A rectangular grid was marked on the opposite face of analysis using VecTor2. A reference (yield) displacement
the wall to the LED sensors to monitor cracks and to facil- (∆r) was determined for each wall using the predicted
itate interpretation of high-resolution panoramic (GigaPan load-displacement relationship and an equal energy assump-
Systems 2013) images of the cracked wall. Gauges capable tion. Figure 4 shows the derived reference displacement
of measuring 15% strain were installed on the horizontal for SW5. The different displacement increments, labelled
and vertical reinforcement in each wall. An average of 52 LS to refer to load steps, were anchored to fractions/multi-
and 84 strain gauges were installed on Phase I and Phase II ples of ∆r. Load Step LS0 was the first load step for Walls
walls, respectively. Figure 2(b) is a photograph of SW8 after SW1 and SW5-SW12; LS1 was the first load step for SW2,
instrumentation, showing the grid of LED sensors. SW3, and SW4. Load Step LS0 was 0.25∆r and was used
for the calculation of initial stiffness. Load Step LS1 was
PRE-TEST NUMERICAL ANALYSIS 0.5∆r. Succeeding load steps were increments of either
Numerical simulations of the cyclic response of the walls 0.5∆r or ∆r. The first load step for each wall comprised three
were performed to help design and detail the test speci- displacement cycles, and subsequent load steps comprised
mens and the test fixture. Upper-bound estimates of mate- two displacement cycles. The loading protocol for SW5 is
rial strength were used to establish the maximum size of presented in Fig. 5. The displacement loading rate for all
the test specimens, which was dictated by: 1) the capacity the specimens ranged from approximately 0.005 to 0.01 in./s
of the NEES actuators; and 2) the need to construct walls (0.13 to 0.25 mm/s).
of different aspect ratios with identical plan dimensions.
VecTor2 ver. 2.8 (Vecchio and Wong 2002), a nonlinear GLOBAL RESPONSE
finite element program for the analysis of RC structures, Figures 6 to 10 present the measured force-drift relation-
was used to simulate the response of each test specimen to ships for the 12 walls. A backbone curve is shown for each
monotonic and reversed cyclic loading using the Modified wall. The values inside the parenthesis in the title of each
Compression Field Theory (Vecchio and Collins 1986), figure correspond to: 1) aspect ratio; 2) horizontal rein-
and Popovics (1973) model to represent the pre- and post- forcement ratio; 3) vertical reinforcement ratio; 4) concrete
peak response of concrete in compression. Figure 3 shows compressive strength; and 5) boundary elements reinforce-
the predicted monotonic and cyclic base shear versus story ment ratio (for SW11 and SW12). The in-plane force was

ACI Structural Journal/September-October 2015 595


Fig. 4—Reference displacement for SW5. (Note: 1 kip = 4.45 kN; 1 in. = 25.4 mm.)

Fig. 5—Loading protocol for SW5. (Note: 1 in. = 25.4 mm.)

Fig. 6—Global force-displacement relationships of SW8, SW9, and SW10. (Note: 1 kip = 4.45 kN; 1000 psi = 6.9 MPa.)
calculated using load cells in the actuators. The lateral was calculated by dividing the lateral displacement at the
displacement was calculated from in-plane displacement centerline of loading by the distance between the centerline
transducers attached at the centerline of loading. Drift ratio of loading and the top of the foundation.

596 ACI Structural Journal/September-October 2015


Fig. 7—Global force-displacement relationships of SW11 and SW12. (Note: 1 kip = 4.45 kN; 1000 psi = 6.9 MPa.)
Figure 6 presents the measured force-drift relationships (667 kN) greater than SW6. The force-drift relationships of
for Walls SW8, SW9, and SW10. The three walls have an SW2, SW3, and SW4 (walls with an aspect ratio of 0.54 and
aspect ratio of 0.54 and a vertical reinforcement ratio of reinforcement ratios of 1.0%, 0.67%, and 0.33%, respec-
1.5%; the horizontal reinforcement ratio varies from 0.33% tively) are shown in Fig. 9. The peak shear strength of SW2
to 1.5%. Some of the key observations from the response is greater than SW3, although its concrete compressive
of these walls are: 1) the hysteretic response of SW8 and strength is smaller. The peak shear strength of SW4 is much
SW9 are qualitatively similar, measured herein in terms of less than SW3, which was expected because the concrete
peak strength and loss of stiffness and strength with repeated strength and reinforcement ratios are substantially smaller
cycling beyond the displacement corresponding to peak in SW4.
strength; 2) the pinching of the hysteresis loops is due to Specimens SW1, SW3, and SW6 have the same horizontal
sliding at the base, which occurs after the peak strength is and vertical reinforcement ratios and aspect ratios of 0.94,
attained; 3) the peak shear strengths of SW8 and SW9 are 0.54 and 0.33, respectively. The force-drift relationship of
similar, although SW9 has much less horizontal reinforce- SW1, which was nominally flexure-critical, is shown in
ment, which suggests that above a threshold value, the Fig. 10. The measured peak strength in the first quadrant
effect of the horizontal reinforcement ratio on peak strength of loading is greater for SW6 than SW3, even though SW6
is small; and 4) the ability of the wall to sustain signifi- has a significantly lower concrete compressive strength than
cant lateral load at displacements greater than that at peak SW3. SW6 and SW1 have a similar concrete compressive
strength is affected by the horizontal reinforcement ratio. strength, but the measured peak shear strength is greater for
The force-drift relationships of Walls SW11 and SW12 SW6 than SW1. These results support past observations that
(walls with boundary elements; refer to Fig. 1(c) and 1(d)) peak shear strength is greater for walls with lower aspect
are shown in Fig. 7. A plateau can be observed in the hyster- ratio, with all other parameters being equal.
etic response of the two walls, which suggests that boundary Specimens SW8 and SW10 have the same vertical rein-
elements help maintain peak shear strength for cycles at forcement ratio (1.5%), and horizontal reinforcement ratios
displacements beyond peak strength. The two walls have of 1.5% and 0.33%, respectively. The concrete compressive
approximately the same average vertical reinforcement ratio strength of SW10 is approximately 1.1 ksi (7.6 MPa) greater
(0.84% for SW11 and 0.77% for SW12), same day-of-test than SW8. Photographs of the damage to Walls SW8 and
concrete compressive strength, and comparable peak shear SW10 are presented in Fig. 11 at drift ratios of 1.3% (SW8)
strengths (slightly higher for SW11 on the first quadrant of and 1.4% (SW10). The red arrows identify the direction of
loading). These results suggest that the distribution of the loading at the instant the photograph was taken. Although
total vertical reinforcement along the length of the wall does the drift ratios are similar, the damage to the two walls is
not have a significant effect on peak shear strength, noting very different: SW10 failed in diagonal tension whereas
that the amount and distribution of horizontal reinforcement SW8 failed in diagonal compression. Wall SW10 had fewer
is expected to have little influence on the shear strength of but wider diagonal cracks than SW8: the amount and distri-
low-aspect-ratio walls (Gulec and Whittaker 2011; Barda et bution of horizontal reinforcement affects the distribution
al. 1977). and size of cracks in RC shear walls. Although both types
Figure 8 presents the force-drift relationships of SW5, of damage might be considered acceptable in building
SW6, and SW7. These walls have an aspect ratio of 0.33 and structures after a design-basis earthquake, the wide crack
reinforcement ratios of 1.0%, 0.67%, and 0.33%, respec- in SW10 would compromise a confinement or containment
tively. The concrete compressive strength of SW6 and SW7 boundary in a safety-related nuclear structure.
are identical, but the peak shear strength of SW6 is signifi- The peak forces, corresponding average shear stress, and
cantly greater than SW7. Although the concrete compres- drift ratios in the first and third quadrants of loading are
sive strength of SW5 is only marginally greater than SW6, reported in Table 2. The last column in the table identifies the
the peak shear strength of SW5 is approximately 150 kip mode of failure that is defined herein as a reduction in shear

ACI Structural Journal/September-October 2015 597


Fig. 8—Global force-displacement relationships of SW5, SW6, and SW7. (Note: 1 kip = 4.45 kN; 1000 psi = 6.9 MPa.)

Fig. 9—Global force-displacement relationships of SW2, SW3, and SW4. (Note: 1 kip = 4.45 kN; 1000 psi = 6.9 MPa.)

598 ACI Structural Journal/September-October 2015


Fig. 10—Global force-displacement relationships of SW1.
(Note: 1 kip = 4.45 kN; 1000 psi = 6.9 MPa.)
capacity post-peak strength. The shear-critical walls failed
either by diagonal tension (DT) or diagonal compression
(DC), where DT was chosen to describe a failure associated
with wide cracks and measured horizontal reinforcing bar
strains substantially greater than yield, and DC was assumed
otherwise. SW1 failed in a combination of flexure and shear.
Walls that failed in DC slid at their base after achieving peak
shear strength, but the significant drop in shear strength was
due to crushing of concrete at the toes of the walls. Shear-
critical Walls SW2, SW3, SW4, SW6, SW7, and SW11 failed
in DC but had peak average shear stress less than 10√fc′,
which calls into question the ACI 318-11 (ACI Committee
318 2011) limit on average shear stress that is associated
with crushing of concrete compression struts. Walls SW4
and SW7 (reinforcement ratios of 0.33%) had lower peak
average shear stress than the other walls that failed in DC,
which is attributed to the lower reinforcement ratio in the
Fig. 11—Photos of damage on SW8 and SW10.
webs of the two walls. Web reinforcement ratios and DC
failures are coupled because this reinforcement serves to moments, and the corresponding curvatures, were calculated
confine the concrete compression struts. using the concrete compressive strength on the day of testing
Shear-critical Walls SW2, SW3, SW4, and SW12 and a measured reinforcing bar stress-strain relationship.
achieved their peak shear strength at drift ratios greater than The maximum OOP displacements were estimated by inte-
1%, which was an unexpected result because RC walls are grating the moment-curvature relationship over the height
often used because they can achieve significant strength at of the wall. Interaction with in-plane moments and shears
low levels of story drift. The use of 90-degree hooks at the was ignored. The calculated maximum OOP displacement
ends of SW1 through SW10 (refer to Fig. 1(b)) did not affect was normalized by the height of the wall and reported as
the response of these walls: none of the hooks straightened drift in the second-to-last column of Table 3. The ratios of
before the walls slid. The boundary elements in SW11 and the first quadrant OOP drift (calculated using the average of
SW12 (refer to Fig. 1(c) and 1(d)) changed the orientation of the OOP displacements at the ends of the wall) to the calcu-
the shear cracks to horizontal at the ends of the walls. lated maximum OOP drift are presented in the last column of
Significant differences in the measured first and third Table 3. Five of the six walls (SW2, SW3, SW5, SW6, and
quadrant peak shear strength were observed in Walls SW2, SW7) with significant differences in first and third quadrant
SW3, SW5, SW6, SW7, and SW8. The loading apparatus peak strengths had the greatest ratios of the first quadrant
did not prevent the walls from out-of-plane (OOP) twisting, OOP drift to the calculated maximum OOP drift, indicating
which should be anticipated in the field during earthquake that OOP drift has a significant influence on peak in-plane
shaking. The OOP displacement and twisting of the walls, shear strength.
normalized by their height, in the first and third quadrants of Figure 12 shows the typical OOP displacement and twisting
loading to peak strength are presented in Table 3. Data from of Walls SW2, SW3, SW5, SW6, and SW7 during the first
the OOP string potentiometers attached at the two corners on and third quadrants of loading to peak in-plane strength.
the top of the wall—approximately 16 in. (406 mm) above The red arrow shows the direction of loading, and the red
the centerline of loading—were used for these calculations. box shows the final position of the top section of the wall at
A maximum value of OOP displacement was estimated for peak strength. Tensile yielding of the vertical reinforcement
each wall by cross section analysis. The yield and ultimate in these walls was measured at the ends of the walls in the

ACI Structural Journal/September-October 2015 599


Table 2—Peak shear strength and corresponding drift ratio
First quadrant Third quadrant
Average shear stress Drift ratio, Average shear stress Drift ratio,
Wall Peak force, kip (× √fc′) % Peak force, kip (× √fc′) % Mode of failure*
SW1 253 4.4 1.30 249 4.3 1.28 Flexure/DC
SW2 563 7.0 1.25 490 6.1 0.94 DC
SW3 468 5.5 2.09 381 4.5 0.76 DC
SW4 226 3.6 1.08 216 3.5 0.33 DC
SW5 726 11.5 0.89 547 8.7 1.31 DC
SW6 571 9.6 0.81 411 6.9 0.81 DC
SW7 318 5.4 0.45 277 4.7 0.41 DC
SW8 623 11.0 0.70 546 9.6 0.65 DC
SW9 622 9.9 0.60 633 10.1 0.78 DT
SW10 495 7.6 0.52 528 8.1 0.58 DT
SW11 424 6.2 0.56 408 6.0 0.78 DC
SW12 365 5.4 0.89 416 6.1 1.24 DT
*
DT – diagonal tension, DC – diagonal compression.
Note: 1 kip = 4.45 kN.

Table 3—Out-of-plane displacement and twisting (normalized by the height of wall)


Measured drifts (× 10–3) Measured twist (rad/in. × 10–6)
First quadrant Third quadrant First quadrant Third quadrant Maximum drift (× 10–3)
(2) (3) (4) (5) (6) |(2)/(6)|
SW1 7.5 –1.9 53.1 –5.3 161.9 0.05
SW2 –21.5 –2.9 –133.8 43.1 100.0 0.22
SW3 –7.1 –1.1 –43.1 15.4 80.0 0.09
SW4 –10.3 –1.1 –69.2 –1.5 24.6 0.42
SW5 –10.5 -0.2 –48.8 22.0 70.7 0.15
SW6 –5.1 0.5 –34.1 4.9 58.5 0.09
SW7 –4.4 –1.7 –4.9 17.1 17.1 0.26
SW8 –0.3 –0.5 3.1 18.5 126.2 0.00
SW9 1.8 0.6 23.1 –0.3 126.2 0.01
SW10 0.2 2.6 –4.6 24.6 120.0 0.00
SW11 0.9 0.6 –6.2 10.8 92.3 0.01
SW12 –1.5 0.9 –6.2 16.9 81.5 0.02

Note: 1 in = 25.4 mm.

load steps corresponding to peak in-plane strength. The OOP ments and twisting resulted in significant differences in peak
stiffness at peak in-plane shear strength will vary consider- strength in the first and third quadrants of loading.
ably over the length of a wall, from a small value at the end
in tension under in-plane loading to a relatively large value SHEAR STRENGTH
at the end in compression. The OOP displacements of the Equations are provided in standards of practice such as
top of the walls were larger at the end of the wall in tension ACI 318-11 (and ACI 349-13 [ACI Committee 349 2013])
than at the end of the wall in compression, as indicated in to predict the nominal shear strength of RC walls. Each
Fig. 12. The variation in OOP stiffness along the length of equation uses design variables such as compressive strength
the walls led to twisting of the walls. OOP displacement and of concrete, yield strength of reinforcement, reinforce-
twisting of RC shear walls in buildings and safety-related ment ratio, and aspect ratio. The nominal shear strength of
nuclear structures are inevitable in the event of earthquake the 12 walls calculated using equations from Chapters 11
shaking. In the absence of OOP displacement and twisting, and 21 of ACI 318-11, Barda et al. (1977), Wood (1990),
the peak shear strengths in the first and third quadrants of and Gulec and Whittaker (2011) are presented together with
loading were expected to be similar. Large OOP displace- the measured peak shear strength in Fig. 13. (The measured
peak strength reported herein is the maximum of the peak

600 ACI Structural Journal/September-October 2015


Fig. 12—Typical out-of-plane displacement and twisting of
walls at peak strength.
strengths in the first and third quadrants.) It is evident from
the figure that: 1) there is a significant scatter in the predic-
tions of peak shear strength; 2) none of the equations are
particularly suitable for either design or performance assess-
ment in their current form; and 3) the variables used in some
of the predictive equations are incomplete.
The Barda equation produces unconservative predictions Fig. 13—Measured and predicted peak shear strength.
for the rectangular walls, with over-predictions by more than (Note: 1 kip = 4.45 kN.)
a factor of 2 in three of the walls. The empirical equation
derived by Gulec and Whittaker provides conservative (but
biased) estimates of peak shear strength. Of the two ACI 318
equations, the Chapter 11 equation provided a mean value of
the ratio of predicted to measured strength closer to 1.0 than
the Chapter 21 equation.

LATERAL STIFFNESS
Linear and nonlinear seismic analysis requires the analyst
to calculate elastic stiffness. Mechanics-of-materials formu-
lations are often used for this purpose. Reliable estimates
Fig. 14—Force-deformation relationship (ASCE 41-06).
of elastic stiffness are needed to estimate seismic response
in the linear range of behavior and to generate in-structure and SW4 were not included in the analysis due to high
floor response spectra for the design and qualification of uncertainty caused by noise in the collected data from the
secondary systems and nonstructural components. displacement transducers during the first load step. The
Figure 14 presents the generic load-deformation relation- secant stiffness (Column 3) was calculated at the point of
ship specified in ASCE 41-06. Point B represents the yield peak shear strength in the first quadrant. The average of
point, Point C represents the nominal strength, and Point D the ratios of the calculated initial stiffness to the effective
represents the loss of significant strength to a residual value. stiffness recommended by ASCE 43-05 (ASCE 41-06) for
For walls whose response is dominated by shear, the strength cracked wall sections is 0.64 (0.42). The secant stiffness at
associated with Points B and C are taken to be the same. To peak resistance is as low as 3% (2%) of the effective stiffness
calculate effective stiffness of uncracked walls (the slope of recommended by ASCE 43-05 (ASCE 41-06) for cracked
Line AB), ASCE 43-05 recommends the use of 100% of the wall sections.
shear and flexural rigidities, whereas ASCE 41-06 recom- Sozen and Moehle (1993) observed that measured initial
mends the use of 80% of the flexural rigidity and 100% of stiffness was less than theoretical initial stiffness and noted
the shear rigidity. For cracked walls, ASCE 43-05 recom- that the difference was due to: 1) cracks invisible to the eye
mends a 50% reduction in both shear and flexural rigidities, present before the test, and 2) deformability of the base
and ASCE 41-06 recommends a reduction of 50% in flexural girder. No foundation (base girder) rotation was observed
rigidity and no reduction in shear rigidity. Table 4 presents in the 12 aforementioned tests during the measurement of
the lateral stiffness of the 12 walls. Day-of-test concrete initial stiffness. The presence of cracks in the concrete near
compressive strength, Eq. (8.5.1) of ACI 318-11 for Young’s the base of the wall due to: 1) restrained shrinkage associ-
modulus, and a Poisson’s ratio equal to 0.2 were used in the ated with the large foundation and the reinforcement; and
calculations. The initial stiffness (Column 2) was calculated 2) tensile strain in the concrete at low levels of drift, is the
from the first load step in each test, which involved force cause of the increased flexibility. Concrete near the base of
less than 15% of peak strength and a drift ratio of less than the walls cracked during the first load step of testing. Drift
0.025%. Data for the initial stiffness of Walls SW2, SW3, ratios of 0.025% or less were sufficient to crack the concrete

ACI Structural Journal/September-October 2015 601


Table 4—Lateral stiffness
Initial Secant at peak ASCE 41-06 ASCE 43-05
Wall Ki, kip/in. strength, Ks, kip/in. K41-06, kip/in. K43-05, kip/in. Ki/K41-06 Ki/K43-05
SW1 1840 170 3060 2445 0.60 0.75
SW2 — 690 14,835 9855 — —
SW3 — 340 15,660 10,405 — —
SW4 — 320 11,495 7635 — —
SW5 8900 1990 26,250 15,265 0.34 0.58
SW6 7570 1720 24,680 14,350 0.31 0.53
SW7 12260 1720 24,680 14,350 0.50 0.85
SW8 5370 1380 10,490 6970 0.51 0.77
SW9 4980 1250 11,630 7725 0.43 0.64
SW10 5250 1400 12,025 7990 0.44 0.66
SW11 4090 1170 12,540 8330 0.33 0.49
SW12 3850 520 12,540 8330 0.31 0.46

Note: 1 kip/in. = 0.18 kN/mm.

Table 5—Strength degradation


Wall V2/V1 V3/V1
SW1 0.84 0.79
SW2 0.79 0.70
SW3 0.60 0.42
SW4 0.72 0.67
SW5 0.92 0.40
SW6 0.86 0.79
SW7 0.84 0.75
SW8 0.80 0.68
Fig. 15—First quadrant hysteresis loops of SW1. (Note:
SW9 0.86 0.80
1 kip = 4.45 kN.)
SW10 0.92 0.90
at the base of the walls based on measured strains in the
SW11 0.82 0.75
vertical reinforcement at the base of the wall.
SW12 0.77 0.73
STRENGTH DEGRADATION
Walls in buildings and safety-related nuclear structures at multiple cycles of loading, it is evident from the ratios
are expected to maintain their strength for multiple cycles presented in Table 5 that the 12 walls tested show a rapid
of loading at and beyond the displacement corresponding to loss of strength after achieving their peak shear strength.
peak shear strength: Line BC in the force-deformation curve
is shown in Fig. 14. SUMMARY AND CONCLUSIONS
Figure 15 shows the loss of strength in SW1 with repeated Twelve large-size, low-aspect-ratio, rectangular RC shear
cycling to the displacement corresponding to peak strength. walls were built and tested at the University at Buffalo. The
To quantify the loss of strength in the twelve walls, shear key findings of this study are:
strengths V1, V2, and V3 (marked with open red circles in 1. The effect of horizontal reinforcement ratio on the peak
Fig.  15) were extracted from the first quadrant hysteresis shear strength of a shear-critical wall is small above a certain
loops of each specimen. (Force V1 corresponds to the peak threshold value that is sufficient to maintain the integrity of
shear strength at a displacement D; forces V2 and V3 corre- the diagonal compression struts.
spond to the shear strength at the second and third excursions 2. Boundary elements help maintain peak shear strength at
to the displacement D. Table 5 presents the ratios V2/V1 and cycles of displacements beyond peak strength.
V3/V1. On the second (third) excursion to D, the shear strength 3. The upper limit on average shear stress of 10√fc′ in
is as low as 60% (40%) of the peak shear strength. Walls Chapters 11 and 21 of ACI 318-11 may be unconservative,
SW9 and SW10 (vertical reinforcement ratio of 1.5% and except for heavily reinforced walls.
horizontal reinforcement ratio of 0.67% and 0.33%, respec- 4. Low-aspect-ratio walls are prone to base sliding after
tively) had the smallest loss of strength after two excursions the peak shear strength is attained. The significant pinching
to D. Although the walls are expected to sustain strength of the hysteresis loops is attributed to sliding.

602 ACI Structural Journal/September-October 2015


5. The shear resistance of walls without boundary Angeles, CA); C. Goksu (Istanbul Technical University); L. Lowes and
J. Pugh (University of Washington); A. Birely (Texas A&M University);
elements degrades rapidly with repeated cycling at lateral and B. Stojadinovic and C. Whyte (ETH in Zurich). The staff of the NEES
displacements equal to or greater than that associated with Equipment Site at the University at Buffalo enabled the tests described in
peak shear strength. this paper, and their advice and support are gratefully acknowledged.
6. Web reinforcement confines concrete compression
struts, and its volume and distribution are coupled with diag- REFERENCES
ACI Committee 318, 2011, “Building Code Requirements for Structural
onal compression failure. For those walls that failed in diag- Concrete (ACI 318-11) and Commentary,” American Concrete Institute,
onal compression, the walls with lower web reinforcement Farmington Hills, MI, 503 pp.
ratios achieved lower peak average shear stress than walls ACI Committee 349, 2013, “Code Requirements for Nuclear Safety-
Related Concrete Structures (ACI 349-13) and Commentary,” American
with higher reinforcement ratios. Concrete Institute, Farmington Hills, MI, 153 pp.
7. In the absence of out-of-plane displacement and ACI Committee 374, 2005, “Acceptance Criteria for Moment Frames
twisting, the peak shear strength of a wall in the first and Based on Structural Testing (ACI 374.1-05) and Commentary,” American
Concrete Institute, Farmington Hills, MI, 11 pp.
third quadrants of loading are expected to be similar. Large ACI Committee 374, 2013, “Guide for Testing Reinforced Concrete
out-of-plane displacements and twisting result in significant Structural Elements under Slowly Applied Simulated Seismic Loads (ACI
differences in first and third quadrant peak strengths. 374.2R-13),” American Concrete Institute, Farmington Hills, MI, 18 pp.
ASCE/SEI 43-05, 2005, “Seismic Design Criteria for Structures,
8. Type 2 mechanical couplers (defined in ACI 318-11) on Systems, and Components in Nuclear Facilities,” American Society of Civil
vertical reinforcement appear to have no significant effect on Engineers, Reston, VA.
the cyclic response of low-aspect-ratio walls. ASCE/SEI 41-06, 2006, “Seismic Rehabilitation of Buildings,” Amer-
ican Society of Civil Engineers, Reston, VA.
9. There is a significant scatter in the shear strength of ATC-24, 1992, “Guidelines for Cyclic Seismic Testing of Component of
the 12 walls predicted using equations from Chapters 11 and Steel Structures,” Applied Technology Council, Redwood City, CA.
21 of ACI 318-11, Barda et al. (1977), and Wood (1990), Barda, F.; Hanson, J. M.; and Corley, W. G., 1977, “Shear Strength of
Low-rise Walls with Boundary Elements,” Reinforced Concrete Structures
confirming the prior observation of Gulec et al. (2008). in Seismic Zones, SP-53, American Concrete Institute, Farmington Hills,
10. There is a significant difference between the measured MI, pp. 149-202.
initial stiffness of each wall and the corresponding initial Del Carpio Ramos, M.; Whittaker, A. S.; and Gulec, C. K., 2012,
“Predictive Equations for the Peak Shear Strength of Low-Aspect Ratio
stiffness computed using recommendations by ASCE 43-05 Reinforced Concrete Walls,” Journal of Earthquake Engineering, V. 16,
and ASCE 41-06 for uncracked wall sections. The recom- No. 2, pp. 159-187. doi: 10.1080/13632469.2011.613529
mended reductions in shear and flexural rigidities used in Epackachi, S.; Nguyen, N. H.; Kurt, E. G.; Whittaker, A. S.; and
Varma, A. H., 2014, “In-Plane Seismic Behavior of Rectangular Steel-
calculating the initial stiffness should be reexamined. Plate Composite Wall Piers,” Journal of Structural Engineering, ASCE, p.
11. The secant stiffness at peak resistance was signifi- 04014176 doi: 10.1061/(ASCE)ST.1943-541X.0001148
cantly smaller than the effective stiffness recommended by Epackachi, S.; Whittaker, A. S.; and Huang, Y. N., 2015, “Analytical
Modeling of Rectangular SC Wall Panels,” Journal of Constructional Steel
ASCE 43-05 and ASCE 41-06 for cracked wall sections. The Research, V. 105, pp. 49-59. doi: 10.1016/j.jcsr.2014.10.016
impact of a substantial reduction in effective stiffness from GigaPan Systems, 2013, www.gigapansystems.com.
that which is traditionally assumed for analysis on the design Gulec, C. K., and Whittaker, A. S., 2011, “Empirical Equations for Peak
Shear Strength of Low Aspect Ratio Reinforced Concrete Walls,” ACI
of primary and secondary systems in buildings and safety- Structural Journal, V. 108, No. 1, Jan.-Feb., pp. 80-89.
related nuclear structures should be carefully assessed. Gulec, C. K.; Whittaker, A. S.; and Stojadinovic, B., 2008, “Shear
Strength of Squat Rectangular Reinforced Concrete Walls,” ACI Structural
Journal, V. 105, No. 4, July-Aug., pp. 488-497.
AUTHOR BIOS Luna, B. N.; Rivera, J. P.; Rocks, J. F.; Goksu, C.; Weinreber, S.; and
ACI member Bismarck N. Luna is a PhD Candidate at the University at
Whittaker, A. S., 2013, “University at Buffalo—Low Aspect Ratio Rect-
Buffalo, Buffalo, NY. He received his BS from University of the Philippines,
angular Reinforced Concrete Shear Wall—Specimen SW1, SW2, SW3,
Los Baños Laguna, Philippines, and his MS from Purdue University, West
SW4, SW5, SW6, SW7, SW8, SW9, SW10, SW11, SW12,” Network for
Lafayette, IN. His research interests include earthquake engineering and
Earthquake Engineering Simulation (distributor). doi: 10.4231/D3542J820,
reinforced concrete structures.
10.4231/D38W3825B, 10.4231/D32B8VB8X, 10.4231/D3639K516,
10.4231/D3RX93D28, 10.4231/D3WP9T661, 10.4231/D31G0HV0P,
ACI member Jonathan P. Rivera is a PhD Candidate at the University at
10.4231/D3XP6V35S, 10.4231/D3SX6492H, 10.4231/D3P55DG82,
Buffalo, where he received his BS and MS in 2011 and 2013, respectively.
10.4231/D3JD4PP4X, 10.4231/D3DN3ZW0S
His research interests include the behavior of low-aspect-ratio shear walls
Nikon Metrology Inc., 2013, Krypton K600 Optical Measurement
and earthquake engineering.
System, www.nikonmetrology.com.
Popovics, S., 1973, “A Numerical Approach to the Complete Stress-
Andrew S. Whittaker, FACI, is Professor and Chair of the Department
Strain Curve of Concrete,” Cement and Concrete Research, V. 3, No. 5, pp.
of Civil, Structural and Environmental Engineering at the University at
583-599. doi: 10.1016/0008-8846(73)90096-3
Buffalo. He received his BS from the University of Melbourne, Melbourne,
Sozen, M. A., and Moehle, J. P., 1993, “Stiffness of Reinforced Concrete
Australia, and his MS and PhD from the University of California, Berkeley,
Walls Resisting In-Plane Shear,” Report No. EPRI TR-102731, Electrical
Berkeley, CA, in 1977, 1985, and 1988, respectively. He is a member of ACI
Power Research Institute, Palo Alto, CA.
Committee 349, Concrete Nuclear Structures. His research interests include
Vecchio, F. J., and Collins, M. P., 1986, “The Modified Compression
earthquake and blast engineering, and performance-based seismic design
Field Theory for Reinforced Concrete Elements Subjected to Shear,” ACI
and assessment.
Journal Proceedings, V. 83, No. 2, Mar.-Apr., pp. 219-231.
Vecchio, F. J., and Wong, P. S., 2002, VecTor2 and FormWorks User’s
ACKNOWLEDGMENTS Manual, University of Toronto, Toronto, ON, Canada.
The National Science Foundation (NEESR Program) provided financial Whyte, C. A., and Stojadinovic, B., 2013, “Hybrid Simulation of the
support under Grant No. CMMI-0829978 for the studies described in this Seismic Response of Squat Reinforced Concrete Shear Walls,” PEER
paper. The authors thank the following individuals for their contributions Report 2013/02, Pacific Earthquake Engineering Research Center, Univer-
to the work at the University at Buffalo: J. Rocks (Constellation Energy sity of California, Berkeley, Berkeley, CA, 227 pp.
Nuclear Group, LLC, Lycoming); K. Gulec (Thornton-Tomasetti, Los Wood, S. L., 1990, “Shear Strength of Low-Rise Reinforced Concrete
Walls,” ACI Structural Journal, V. 87, No. 1, Jan.-Feb., pp. 99-107.

ACI Structural Journal/September-October 2015 603


NOTES:

604 ACI Structural Journal/September-October 2015


View publication stats