Sei sulla pagina 1di 10

ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title No. 116-S46

Peak Strength of Shear-Critical Reinforced Concrete Walls


by Bismarck N. Luna and Andrew S. Whittaker

Twelve low-aspect-ratio rectangular reinforced concrete (RC)


walls were built and tested under reversed cyclic loading at the
University at Buffalo. One of the objectives of the research project
was to develop improved predictive equations for the peak shear
strength of low-aspect-ratio RC walls, suitable for inclusion in
design standards such as ACI 318 and ACI 349. Information on
load paths gathered from patterns of concrete cracking, strains on
reinforcement, and strain fields enabled the development of new
predictive equations for peak shear strength.

Keywords: load paths; peak shear strength; reinforced concrete; shear walls.

INTRODUCTION
Low-aspect-ratio reinforced concrete (RC) shear walls
(height-to-length ratio of 2 and less) are widely used in low-
and medium-rise buildings and in safety-related nuclear struc-
tures to resist lateral wind and seismic loadings. Code-based
design and seismic performance (risk) assessment require the
hysteretic response and peak shear strength to be accurately
characterized. Equations are available in the literature and in
standards of practice to predict the nominal shear strength of
low-aspect-ratio RC walls, but these equations are inaccurate
and insufficiently parameterized (refer to Gulec et al. [2008]
and Del Carpio Ramos et al. [2012] for details).
The U.S. National Science Foundation (NSF) funded a
research project on low-aspect-ratio shear walls of conven-
tional and composite construction to improve the profes-
sion’s understanding of their cyclic inelastic response.
Sixteen rectangular (or planar), low-aspect-ratio concrete Fig. 1—Photographs of SW8 and SW11 during testing.
shear walls were built and tested at the University at Buffalo PEAK STRENGTH OF SHEAR-CRITICAL RC
(UB): 12 conventionally reinforced concrete walls (Luna WALLS
et al. 2015, 2018) and four steel-plate concrete composite The data collected from the 12 RC walls tested at UB
walls (Epackachi et al. 2015). Two RC shear walls were built enabled the tracking of load paths from the point of force
and tested using hybrid simulation at the University of Cali- application to the foundation. Luna et al. (2018) describes
fornia, Berkeley (Whyte and Stojadinovic 2013). the experimental program in significant detail and only
The transducers used in the UB experiments on RC walls information relevant to calculations of peak shear strength
helped identify the flow of forces through the walls and thus is included herein. The 12 shear walls were identified as
determine contributions to peak shear strength. The identifica- SW1 through SW12. Figure 1 presents photographs of SW8
tion of load paths enabled the development of new predictive and SW11 during testing; the light-emitting diodes (LEDs)
equations for peak strength of rectangular, shear-critical RC that were part of the non-contact coordinate measurement
walls, without and with boundary elements contained within the system can be seen on the surface of the walls. The peak
web of the wall. Much additional information is documented in shear strength and corresponding average shear stress in
Luna (2015), Luna et al. (2018), and Rivera et al. (2018). the first and third quadrants of loading of the 12 UB walls
are reported in Table 1. Figure 2 presents the nominal shear
RESEARCH SIGNIFICANCE
Data obtained from tests of 12 low-aspect-ratio rectan-
ACI Structural Journal, V. 116, No. 2, March 2019.
gular RC walls, without and with boundary elements, were MS No. S-2018-145, doi: 10.14359/51712280, was received April 26, 2018, and
used to develop predictive equations for peak strength of reviewed under Institute publication policies. Copyright © 2019, American Concrete
Institute. All rights reserved, including the making of copies unless permission is
shear-critical walls, suitable for inclusion, after simplifica- obtained from the copyright proprietors. Pertinent discussion including author’s
closure, if any, will be published ten months from this journal’s date if the discussion
tion in design standards such as ACI 318 and ACI 349. is received within four months of the paper’s print publication.

ACI Structural Journal/March 2019 257


Table 1—Peak shear strength lated from non-contact coordinate measuring LEDs, and
First quadrant Third quadrant Average
cracking patterns were used to develop free-body diagrams
of different segments of the walls and to estimate the magni-
Peak Average Peak Average
tude of the forces in reinforcement and the strains in concrete.
force, shear stress force shear stress Peak force,
Wall kip (× √fc′) (kip) (× √fc′) kip
MOEHLE’S EQUATION FOR PEAK SHEAR
SW1 253 4.4 249 4.3 251
STRENGTH
SW2 563 7.0 490 6.1 526 Moehle (2015) presents an expression to estimate the peak
SW3 468 5.5 381 4.5 423 shear strength of a low-aspect-ratio RC wall. The derivation
SW4 226 3.6 216 3.5 221 assumes an idealized pattern of cracking and the free-body
diagrams of Fig. 3. The equation for peak shear strength is
SW5 726 11.5 547 8.7 633
SW6 571 9.6 411 6.9 491 1
Vn = ( N u + ρl f y Acv ) (1)
SW7 318 5.4 277 4.7 297 tan θ
SW8 623 11.0 546 9.6 584
SW9 622 9.9 633 10.1 627
where Vn is the peak shear strength; Nu is the total axial
load; ρl is the vertical reinforcement ratio; Acv is the gross
SW10 495 7.6 528 8.1 512
cross-sectional area of the wall; and θ is the angle of the crack
SW11 424 6.2 408 6.0 416 with respect to the horizontal. The stress in the horizontal
SW12 365 5.4 416 6.1 391 and vertical reinforcement at peak strength is assumed to be
the yield value fy. Horizontal reinforcement stressed to fy is
Note: 1 kip = 4.45 kN.
assumed in the derivation, although Eq. (1) does not include
it explicitly. Moehle recognized the role of distributed hori-
zontal reinforcement for equilibrium of segment c-d-e in
Fig. 3(c) and in transferring lateral forces from this segment
to the remainder of the wall. This equation is suitable for
walls with a very low aspect ratio. Additional analysis of the
Moehle derivation is provided in Luna et al. (2018).
The Moehle equation provides insight into load paths
in low-aspect-ratio walls at peak lateral strength. Here-
after, the derivation is extended to accommodate all shear-
critical, rectangular walls, without and with boundary
elements contained within the web of the wall, made possible
by analysis of cracking patterns, strain gauge data, and strain
fields derived from LED ata, all obtained from the testing of
12 RC shear walls at UB.

PEAK SHEAR STRENGTH OF SHEAR-CRITICAL


WALLS
Fringes of shear strain and patterns of cracking at
Fig. 2—Measured and predicted peak shear strength of 12 peak strength
UB walls (Luna et al. 2015). (Note: 1 kip = 4.45 kN.) Figure 4 presents fringes of shear strain and patterns of
cracking in SW8 and SW11 at peak strength, which are
strength of the 12 walls calculated using equations from representative of fringes and patterns of walls without and
Chapters 11 and 18 of ACI 318-14, Barda et al. (1977), with boundary elements, respectively. The fringes of shear
Wood (1990), Gulec and Whittaker (2011), and the measured strain and patterns of cracking in all 12 walls are presented
peak shear strength from Luna et al. (2015). (The measured in Chapter 4 of Luna et al. (2018).
peak strength reported in the figure is the greater of the peak
strengths in the first and third quadrants.) Luna et al. (2015) Walls without boundary elements
described the accuracy of these predictive equations and that Figure 5(a) identifies three segments of the walls without
discussion is not repeated herein. As noted in Luna et al. boundary elements with significantly different strains at
(2015): 1) there is a significant scatter in the predictions of peak strength and the corresponding idealized patterns of
peak shear strength; and 2) none of the equations are partic- cracking and forces. The shear strains at peak shear strength
ularly suitable for either design or performance assessment. in Segment B are substantially greater than those in Segments
Data from the tests of these 12 RC walls, as described in A and C. The toe of the wall shown in the bottom left corner
Chapters 3 and 4 of Luna et al. (2018), were used to derive of Segment A (open red circle) can resist compression and
predictive equations for peak strength of shear-critical RC shearing forces because the inclined cracks do not propagate
walls without and with boundary elements. Data from strain to the corner. The distance from the bottom end of the wall
gauges, strain distributions on the faces of the walls calcu- to the bottom of the diagonal crack in Segment A is denoted

258 ACI Structural Journal/March 2019


Fig. 3—Idealized forces on a low-aspect-ratio RC wall (Moehle [2015], reproduced with permission from McGraw-Hill
Education).

Fig. 4—Shear strain fields at peak strength.


c. The patterns of concrete cracking at peak shear strength 3. The strain in the horizontal reinforcement of Walls SW2
in SW3 and SW8 are shown in Fig. 6, with length c high- through SW10 was greatest in the midsection of the walls
lighted with an open red circle. Segment B is divided into and relatively small at the ends. The strains in Segments A
three representative concrete compression struts (or simply of SW2 through SW10, as calculated using LED data, were
struts) in Fig. 5(a). (The number can be less or more than relatively small.
3.) Variables p and v represent the axial load per unit length The free-body diagram of Segment B of the wall is shown
and shear force per unit length at the centerline of horizontal in Fig. 7. Actions Fv and Fh are the total forces carried by
loading, respectively. No axial load was applied to the 12 the vertical and horizontal reinforcement, respectively, that
walls tested at UB and their self-weight is insignificant for cross a diagonal crack. Action Fs is the force associated
the purpose of calculations of shear strength. Cracks are with aggregate interlock. Variable x is the horizontal projec-
assumed to propagate at a constant angle θ with respect to tion of the width of the strut. Summing moments about the
the horizontal. Variables hw and lw are the height (distance mid-bottom of a strut
between foundation and centerline of loading) and length of
the wall, respectively. hw h
vxhw = Fy + px w − Fs ( x sin θ) (2)
Information from Sections 4.5, 4.6, and 6.5 of Luna et al. tan θ tan θ
(2018) are summarized here because they are used to formu-
late an equation for the peak shear strength of walls without where Fy is the axial force in the vertical reinforcement
boundary elements. over distance x. The strains in the vertical reinforcement in
1. The idealized cracking pattern presented in Fig. 5(a) segment B of SW2 through SW7 at peak shear strength were
is applicable for walls SW2 through SW10. The cracking between 6 millistrain (≃εy) and 20 millistrain (≃10εy), where
patterns observed on SW1, SW11, and SW12 were different εy is the yield strain. For walls SW8, SW9, and SW10, the
from that shown in Fig. 5(a). Cracks in SW1 were due to both corresponding strains were between one millistrain (≃0.5εy)
shear and flexure, and the cracks in the boundary elements of and six millistrain (≃3εy). The force in the vertical reinforce-
walls SW11 and SW12 were horizontal. ment, Fy, shown in Fig. 7(b) is
2. All vertical reinforcement yielded in tension (over the
height of the wall) for walls SW2 through SW7 (walls with Fy = ρltwxfy (3)
vertical reinforcement ratio of 1% and less). Many vertical
bars in walls SW8 through SW10 (walls with vertical rein- where ρl is the vertical reinforcement ratio; tw is the thickness
forcement ratio of 1.5%) did not yield in tension. (The of the wall; and fy is the stress in the vertical reinforcement,
strains in the horizontal and vertical reinforcement of the 12 and equal to: a) the average of the measured yield strength
UB walls are presented in Chapter 4 of Luna et al. [2018]). fy and measured ultimate strength fu of the reinforcement to

ACI Structural Journal/March 2019 259


Fig. 5—Different sections and idealized cracking pattern of low-aspect-ratio rectangular RC wall.

Fig. 6—Cracking patterns at peak strength (Luna et al. 2018).


account for cyclic hardening of the vertical reinforcement in Force Fs is associated with aggregate interlock (or friction
Walls SW2 through SW7; and b) the measured yield strength forces). There is considerable uncertainty in the magnitude of
fy for Walls SW8, SW9, and SW10. this force (Gulec and Whittaker 2009). Moehle (2015) conser-
For Walls SW2 through SW10, the strains in the horizontal vatively assumed that the normal and shear stresses along a
reinforcement are the greatest in the midsection (midlength) diagonal crack were resisted only by the reinforcement, and
of the wall and small at the ends. Based on the gauge data that assumption is adopted herein because it was not possible
at peak strength, 0.25εy is a conservative (low) estimate of to assess the width of the cracks away from the surfaces of the
the strain in the horizontal reinforcement at the boundaries tested walls. Accordingly, the term that includes Fs in Eq. (5)
of Segments A and B, and of Segments B and C. The net is set equal to zero. Equation (5) can be rewritten as
effect of the forces in the horizontal reinforcement on the
two inclined boundaries of Segment B, on the moment at the  h  1
mid-bottom of a strut is assumed to be negligible. Assuming Vnb = (ρl tw f y + p )  lw − w − c (6)
 tan θ  tan θ
a uniform distribution of applied shear along the horizontal
projection of Segment B, and using Eq. (3) for Fy, the shear Segments A and C contribute to the peak shear resistance of
resistance of Segment B can be estimated as low-aspect-ratio walls. Consider first Fig. 8(a) that presents a
free-body diagram of Segment A. Actions Fcy and Fcx are the
 h   h  h normal and shear forces acting at the bottom of segment A.
v  lw − w − c hw = ρl tw f y  lw − w − c w + ...
 tan θ   tan θ  tan θ The other variables were defined previously. Summing forces
(4)
 h  h  h  in the vertical direction and ignoring the vertical component
... + p  lw − w − c w − Fs  lw − w − c sin θ of aggregate interlock for the reason given previously
 tan θ  tan θ  tan θ 

 h 
The total lateral resistance provided by Segment B of the Fcy = Fv + p  w + c (7)
wall, Vnb, is given by  tan θ 

 h   h  1 where Fv is equal to
Vnb = v  lw − w − c = ρl tw f y  lw − w − c + ...
 tan θ   tan θ  tan θ
(5)  h 
 h  1  h  sin θ Fv = ρl tw  w  f y (8)
... + p  lw − w − c − Fs  lw − w − c  tan θ 
 tan θ  tan θ  tan θ  hw

260 ACI Structural Journal/March 2019


Fig. 7—Free-body diagram of Segment B of wall without boundary elements.

where f y is set equal to the average of measured yield


and ultimate strength of the reinforcement for walls SW2
through SW7. The strains in the vertical reinforcement at
the boundary of segments A and C of walls SW8, SW9, and
SW10 ranged between 0.1 and 1.0 millistrain, and so f y is
set equal to 0.25fy in Eq. (8) for these walls. Denoting fc as the
axial stress in the concrete at the bottom of segment A, Fcy is

Fcy = fcctw (9)

and using Eq. (7), (8), and (9), c can be written as

 h 
(ρl tw f y + p )  w 
 tan θ 
c= (10)
f c tw − p Fig. 8—Free-body diagram of Segments A and C of wall
without boundary elements.
Equation (13) can be simplified by setting θ equal to 40
Shear force Fcx can be estimated by multiplying Fcy by a degrees, which is likely appropriate for walls with aspect
coefficient of friction μ. If Vna is the contribution to the shear ratios between 0.25 and 0.75 and low axial compres-
resistance of segment A sive stresses. (The average crack angle for all the walls is
discussed in Chapter 4 of Luna et al. [2018].)
  h  
Vna = Fcx = µFcy = µ (ρl tw f y + p )  w  + pc  (11)
  tan θ   Vn = µ (ρl tw f y + p )(1.2hw ) + pc 
(14)
+ 1.2(ρl tw f y + p )(lw − 1.2hw − c) + 0.25ρt hw tw f y
Consider now Fig. 8(b) that presents a free-body diagram
of segment C. The sum of forces in the horizontal rein-
To judge the accuracy of this equation for the shear-crit-
forcement in segment C at peak strength, which is equal to
ical walls SW2 through SW10, p is set to zero and Eq. (14)
the shear resistance of the segment, Vnc, can be estimated
can be further simplified to
using strain gauge data from the horizontal reinforcement.
A conservative (low) estimate of the strain in the horizontal
hw
reinforcement at the boundary of segments B and C (and A Vn = 1.2µρl Acv fy
and B) is 0.25fy. Accordingly, Vnc can be taken as lw
(15)
 h c h
Vnc = ρthwtw(0.25fy) (12) + 1.2ρl Acv f y 1 − 1.2 w −  + 0.25ρt Acv w f y
 lw lw  lw
The total lateral resistance (peak shear strength) of a wall
where Acv is the gross cross-sectional area of the wall. To esti-
without boundary elements, Vn, is the sum of Vna, Vnb, and Vnc
mate the length c, a value of 0.8fc′ is assumed: less than the
measured uniaxial compressive strength because there was no
  h   evidence of spalling of concrete at the compression toes of
Vn = µ (ρl tw f y + p )  w  + pc 
  tan θ   walls at peak shear strength. The values of c calculated using
(13)
 hw  1 Eq. (10) are listed in the second column of Table 2.
+ (ρl tw f y + p )  lw − −c + 0.25ρt hwtw f y
 tan θ  tan θ
ACI Structural Journal/March 2019 261
Table 2—Compressive stress on struts in Segment B and shear strength of Walls SW2 through SW10
Wall c, in. fc, ksi fc/fc′ Vna, kip Vnb, kip Vnc, kip Vn, kip Vn/Vpeak
SW2 10.4 1.82 0.26 234 228 82 544 1.03
SW3 6.2 1.22 0.16 157 173 55 384 0.91
SW4 6.4 0.67 0.16 87 95 29 211 0.95
SW5 12 2.05 0.48 166 477 55 698 1.10
SW6 9.1 1.37 0.36 111 335 37 484 0.98
SW7 4.5 0.67 0.18 55 178 18 251 0.84
SW8 7 2.43 0.69 78 338 131 547 0.94
SW9 5.7 2.43 0.57 78 351 58 487 0.78
SW10 5.3 2.43 0.53 78 354 29 461 0.9
Average 0.94

Notes: 1 kip = 4.45 kN; 1 ksi = 6.9 MPa; 1 in. = 25.4 mm.

The shear force transfer across the wall-to-foundation reasonable given that the toes of the walls had experienced
interface in segment A will be limited by: a) principal tensile tensile and compressive loadings prior to the cycle to peak
strength of concrete; and b) shear friction. To calculate strength. The ratio of the predicted peak shear strength to
capacity based on a), the concrete is assumed not to be rein- measured peak shear strength (listed in the last column of
forced, to have a tensile strength of 0.15fc′, and loaded in Table 2) of Walls SW2 through SW10 range between 0.78
uniaxial plane stress: a normal stress corresponding to Fcy (SW9) and 1.1 (SW5), with an average of 0.94. Excluding
and a shear stress. If the normal stress is 0.8fc′ and the corre- the walls that failed in diagonal compression, SW8 and SW9,
sponding shear stress is 0.4fc′, the corresponding principal the average is 0.96. The contributions of the horizontal rein-
stresses are 0.96fc′ (compression) and –0.16fc′ (tension): the forcement, Vnc, to the peak strength of the walls is relatively
ratio of shear stress to normal stress is 0.5. small, which supports the findings of Luna et al. (2015) and
Shear-friction calculations can be informed by the studies observations by Barda et al. (1977), Gulec and Whittaker
of Mattock (1976, 1977), which form the basis of the provi- (2011), and Moehle (2015).
sions in ACI 318. Test series B of Mattock (1976) shows Although the contribution of the horizontal reinforce-
that for normal stress between 226 and 1576 psi (1.56 and ment to the peak shear strength of a low-rise wall is rela-
10.87 MPa), the ultimate shear stress ranges between 487 tively small, a sufficient amount of horizontal reinforcement
and 1700 psi (3.36 and 11.72 MPa), with the ratio of shear is needed to transfer the lateral load from the centerline of
stress to normal stress of between 2.15 and 1.07 (greater than loading of the wall to the different segments of the wall and
1.0). Test series M of Mattock (1977) shows that the average to confine the concrete in the compression struts, where
ratio of measured cyclic to monotonic shears strengths was confinement here relates to maintaining the integrity of the
0.90. Note that the incremented cyclic loading for test series compression struts (refer to Section 23.5 of ACI 318-14). A
M was applied for more than 45 cycles—many more cycles portion of the lateral load in segment A (refer to Fig. 8(a))
than that expected in design basis earthquake shaking and is transmitted to Segments B and C by the horizontal rein-
imposed in the tested walls. A substantial reduction in the forcement. Lateral force in Segment C is transferred to the
number of reversed cycles to failure would increase the ratio foundation via compression struts that form in in Segment C
from the average of 0.90 to close to 1.0 (that is, the mono- (refer to Fig. 8(b)).
tonic shear strength). It is common practice to assume that walls are uniformly
Accordingly, a coefficient of friction μ equal to 0.5 is loaded in shear. On the basis of θ equal to 40 degrees,
assumed here for the calculation of Vna, which caps the shear Segments A (and C) and B should resist 39% and 61% of the
stress at 0.4fc′ and satisfies both the aforementioned prin- total lateral force, respectively, for an aspect ratio of 0.33,
cipal tensile stress-based calculation and is substantially where the percentage assigned to Segment A (and C) is given
less that the shear-friction strength supported by Mattock’s by the ratio of the projected horizontal length of the crack,
monotonic and cyclic tests. (Note the maximum value of the hw/tanθ, to the length of the wall, lw. For an aspect ratio of
normal stress [product of the reinforcement ratio and yield 0.54, Segments A (and C) and B should resist 64% and 36%
strength] of 1576 psi [10.87 MPa] in the 1976 monotonic of the total lateral force, respectively. On the basis of the
tests was equal to approximately 0.25fc′ and smaller than the values of Vna, Vnb, and Vnc in Table 2, Segments A (and C)
normal stress assumed here of 0.8fc′.) and B resist approximately 30% and 70% of the total lateral
The first, second, and third terms of Eq. (15) correspond force, respectively, for the walls with an aspect ratio of 0.33
to Vna, Vnb, and Vnc, respectively. The calculated values of the (SW5, SW6, and SW7). For walls with aspect ratio of 0.54
three terms in Eq. (15) and the predicted peak shear strength for (SW2, SW3, SW4), Segments A (and C) and B resist approx-
walls SW2 through SW10 are listed in Table 2. A coefficient of imately 60% and 40% of the total lateral force, respectively.
friction μ equal to 0.5 is assumed—less than the value of 1.0 (Walls SW8 and SW9 failed in diagonal compression and so
given in ACI 318-14 for a concrete-to-concrete interface but are excluded from this discussion.) These results indicate the

262 ACI Structural Journal/March 2019


loading plates distributed shear uniformly over the length of tion). Based on the values of the ratio of fc to fc′ for walls
the tested walls, which would suggest that floor diaphragms SW8 and SW9 that failed in diagonal compression, it is
in buildings should be reinforced for this purpose. reasonable to limit fc to 0.5fc′. Compression struts also form
The shear resistance of the wall provided by Segment in segment C. Assuming an axial stress in these struts of
B will be limited by the compressive strength of its struts. 0.5fc′, the diagonal compression strength of the wall can be
Consider again the free-body diagram of Segment B estimated as
presented in Fig. 7. Assuming the force in the horizontal
reinforcement on the two inclined boundaries of a strut are not Vn = fcsinθcosθtwlw (22)
significantly different, the compressive force at the bottom of a
strut, Fc, can be estimated by summing horizontal forces along For a strut angle of 40 degrees with respect to the hori-
the horizontal projection of Segment B. That is zontal, Eq. (22) corresponds to a horizontal shearing stress
of 0.25fc′, which in turn maps to 15.8√fc′ for concrete with
vx = Fccosθ (16) a uniaxial compressive strength of 4000 psi (27.59 MPa).
Given that the maximum average shear stress at peak
If fc is the compressive stress in a strut, Fc can be expressed as strength for SW8 and SW9 ranged between 10√fc′ and 11√fc′
(refer to Table 1), it is reasonable to increase the ACI 318-14
Fc = fc(xsinθ)tw (17) limit on nominal shear stress to 10√fc′.
The nominal shear strength of a rectangular, low aspect
Considering all the struts in Segment B of the wall, the ratio RC wall without boundary elements can be estimated as
lateral force in Segment B, Vnb, can be expressed in terms
of fc as hw
Vn = 0.6 (ρl Acv f y + plw ) + 0.5 pc + ...
lw
 h 
Vnb = f c sin θ cos θtw  lw − w − c (18)  h c
 tan θ  ...1.2(ρl Acv f y + plw ) 1 − 1.2 w −  + ... (23)
 lw lw 
Using Eq. (6) and (18), an approximate value of the hw
...0.25ρt Acv f y ≤ 10 f c′Acv
compressive stress of the struts in Segment B, fc, can be lw
calculated as
or
 h 
f c sin θ cos θtw  lw − w − c  h c
 tan θ  Vn = 1.2(ρl Acv f y + plw ) 1 − 0.7 w − 
(19)  lw lw 
 h  1 (24)
= (ρl tw f y + p )  lw − w − c h
 tan θ  tan θ + 0.25ρt w Acv f y + 0.5 pc ≤ 10 f c′Acv
lw
Using a value of 40 degrees for θ and simplifying,
Eq. (19) becomes where all terms have been defined previously—θ = 40
degrees and μ = 0.5. In Eq. (24), f y is set equal to 1.25fy,
 p which requires that the vertical reinforcing bars be devel-
f c = 2.4  ρl f y +  (20) oped for this stress or greater (perhaps 1.5fy based on the test
 tw 
data) at the upper ends of the compression struts.
For Walls SW2 through SW10, Eq. (20) can be written as
Walls with boundary elements
Walls SW11 and SW12 included boundary elements
f c = 2.4ρl f y (21)
contained within the web. The reinforcement in the
boundary elements changed the orientation of the diagonal
Values of fc for Walls SW2 through SW10 per Eq. (21)
cracks to horizontal at the web-boundary element junctions.
are listed in third column of Table 2. The ratios of fc to fc′
The fringes of shear strain and pattern of cracking at peak
are listed in the fourth column of Table 2. For Walls SW2
strength in SW11 are shown in Fig. 4(b). The segments of
through SW7, walls with vertical reinforcement ratio of
SW11 and SW12 with significantly different strains at peak
1% and less, the ratio of fc to fc′ ranged between 0.16 (SW3
strength, the idealized pattern of cracking and corresponding
and SW4) and 0.48 (SW5). For Walls SW8, SW9, and
forces are identified in Fig. 5(b), where lbe is the length of
SW10 (walls with vertical reinforcement ratio of 1.5%),
each boundary element (equal to 16 in. [406 mm] for SW11
the ratios are 0.69, 0.57, and 0.53, respectively. Walls that
and SW12), lweb is the length of the web of the wall between
are heavily reinforced—assumed here to be a few multiples
the two boundary elements, and ρbe is the boundary element
of the minimum reinforcement ratio specified in ACI 318
reinforcement ratio defined as the area of vertical reinforce-
(= 0.0025 for Grade 60 reinforcing bar)—may fail in diag-
ment (As,be) in one boundary element divided by the area of
onal compression (excessive axial stress in the diagonal
that boundary element (lbe × tw), and all other variables were
struts that transfer horizontal shearing force to the founda-
defined previously. Herein, Eq. (13) is extended to accom-

ACI Structural Journal/March 2019 263


modate boundary elements, each of which is assumed to be
loaded by an axial force P.
The free-body diagram of Segment A of the wall with
boundary elements is shown in Fig. 9(a). The shearing force
at the bottom of Segment A, Vna,be, is calculated similarly
to Eq. (11). The distance c is set equal to the length of the
boundary element, lbe, based on the patterns of cracking
observed in the tests of SW11 and SW12. The normal force
Fcy in Eq. (11) will increase by an amount equal to the
concentrated load on the boundary element P. The strains in
the vertical reinforcement in the webs of SW11 and SW12
were mostly between 1 and 6 millistrain and so the stress in Fig. 9—Free-body diagram of Segments A and B of wall
this reinforcing bar is assumed to be fy. The shear force Vna,be with boundary elements.
can be written as
Vn ,be = 0.6hw (ρl tw f y + p ) + 1.7( plbe + P )
  h   + 1.2(ρl tw f y + p )(lw − 1.2hw − 2lbe ) (29)
Vna ,be = µ (ρl tw f y + p )  w  + plbe + P  (25)
  tan θ   + 1.2(ρbe As ,be f y ) + 0.25ρt hw tw f y

Similar to Eq. (12), the contribution of the horizontal rein-


For Walls SW11 and SW12 for which p and P are zero,
forcement to the lateral resistance of the wall with boundary
Eq. (28) can be further simplified to
elements, Vnc,be, is given by

Vnc,be = ρthwtw(0.25fy) (26) Vn ,be = 0.6ρl tw hw f y


+ 1.2ρl tw f y (lw − 1.2hw − 2lbe ) (30)
where all variables have been defined previously. + 1.2ρbe lbe tw f y + 0.25ρt hwtw f y
The free body diagram of Segment B is presented in
Fig. 9(b). Following the formulation for Eq. (6), the total
lateral resistance provided by Segment B for the walls with In Eq. (30), the first term on the right side corresponds to
boundary elements, Vnb,be, is Vna,be, the sum of the second and third terms correspond to
Vnb,be, and the fourth term corresponds to Vnc,be. The calcu-
 h  1 lated values of Vna,be, Vnb,be, and Vnc,be, and the predicted peak
Vnb ,be = (ρl tw f y + p )  lw − w − 2lbe  shear strength for Walls SW11 and SW12 are listed in Table
 tan θ  tan θ
(27) 3. Similar to the walls without boundary elements, the contri-
1 1
+ (ρbe tw f y + p )(lbe ) +P bution of the horizontal reinforcement to peak strength, Vnc,be,
tan θ tan θ is small. The average of the ratios of the predicted peak shear
strength to the measured peak shear strength (listed in the last
where the axial force on the right boundary element is column of Table 3) for SW11 and SW12 is 1.01. The strengths
assigned to Segment B, identically to the boundary element of SW11 and SW12 associated with a diagonal compression
vertical reinforcement. The strains in the vertical reinforce- failure is 679 kip (3020 kN) for both, and greater than those
ment of the boundary elements of SW11 and SW12 ranged associated with a diagonal tension failure.
between 2 and 20 millistrain and the stress in this vertical The nominal shear strength of a rectangular, low-
reinforcement is taken as f y (that is, the average of the yield aspect-ratio RC wall with boundary elements contained within
and ultimate strengths) to account for cyclic strain hardening. the web of the wall can be estimated by simplifying Eq. (28) as
The diagonal tension shear strength of a wall with
boundary elements, Vn,be, is the sum of Vna,be, Vnb,be, and Vnc,be Vn ,be = 1.2(ρl tw f y + p )(lw − 0.7 hw − 2lbe ) + 1.7( plbe + P)
(31)
  h   + 1.2ρbe As ,be f y + 0.25ρt hwtw f y ≤ 10 f c′Acv
Vn ,be = µ (ρl tw f y + p )  w  + plbe + P 
  tan θ   or
 h  1
+ (ρl tw f y + p )  lw − w − 2lbe 
 tan θ  tan θ  h l 
Vn ,be = 1.2(ρl Acv f y + p ) 1 − 0.7 w − 2 be  + 1.7( plbe + P)
1 1  lw lw 
+ (ρbe tw f y + p )(lbe ) +P + ρt hw tw (0.25 f y ) (32)
tan θ tan θ h
+ 1.2ρbe As ,be f y + 0.25ρt Acv w f y ≤ 10 f c′Acv
(28) lw
where all terms have been defined previously—θ = 40
Setting θ = 40 degrees, μ = 0.5, and simplifying, Eq. (28) degrees and μ = 0.5. In Eq. (32), f y is set equal to 1.25fy,
becomes which requires that the vertical reinforcing bars be developed

264 ACI Structural Journal/March 2019


Table 3—Shear strength of Walls SW11 and SW12
Wall Vna,be, kip Vnb,be, kip Vnc,be, kip Vn,be, kip Vn,be/Vpeak
SW11 140 238 58 436 1.05
SW12 69 281 29 379 0.97
Average 1.01

Note: 1 kip = 4.45 kN.

Table 4—Predicted peak shear strength of


rectangular walls without boundary elements
Researcher Wall Vpeak, kip Vn (Eq. (36)), kip Vn/Vpeak
SW-8 128.1 127.6 1.00
Cardenas
SW-13 142.1 135.7 0.95
M1 45.5 47.3 1.04
Greifenhagen
M2 45.7 44.3 0.97
SWN-1D 67.2 69.2 1.03
SWN-5D 55.1 52.7 0.96
SW-5 54.4 63.4 1.17
SW9 55.8 56.2 1.01
Sheu SW11 49.8 56.2 1.13
SW12 55.8 57.2 1.03
SW13 54.7 54.6 1.00
Fig. 10—Free-body diagram for modified Gulec and Whit-
SW17 40.5 42.0 1.04 taker (2011) equation for peak strength.
SW19 35.2 37.4 1.06
Luna made clear that the free-body diagram assumed by
SW2 526.0 438.0 0.83 Gulec applies best to walls without boundary elements.
SW3 423.0 349.9 0.83 The orientation of the cracks in the webs of the walls with
SW4 221.0 215.8 0.98 boundary elements changed from diagonal to horizontal at
the junctions of web and the boundary elements: challenging
SW5 633.0 592.5 0.94
the assumed free-body diagram used for the derivation.
Luna SW6 491.0 442.0 0.90 The free-body diagram of Gulec was therefore modified
SW7 297.0 297.3 1.00 based on the pattern of cracking pattern observed from the
SW8 584.0 568.6 0.97 tests of Luna’s walls. The modified free-body diagram is
shown in Fig. 10(a). Summing moments at Point A
SW9 627.0 541.8 0.86
SW10 512.0 530.2 1.04 l 
 h 
Average 1.00
P  a1 − w  + Fvwt (a2 ) + Fvwc (a3 ) + Fhw  w  + Fcy (a4 )
 2  2
V=
Note: 1 kip = 4.45 kN. hw

for this stress or greater at the upper ends of the boundary (33)
elements.
where Fvwt and Fvwc are the forces in the vertical reinforce-
AN EMPIRICAL EQUATION FOR PEAK LATERAL ment in tension and compression, respectively; and Fcy is the
STRENGTH OF SHEAR-CRITICAL WALLS vertical component of the force in the concrete in compres-
WITHOUT BOUNDARY ELEMENTS sion. The areas of vertical reinforcement in compression at
Equations (24) and (32), which estimate the peak lateral the base of the walls at peak lateral strength in SW2 through
strength of a rectangular shear critical wall, without and with SW10 were small and the corresponding vertical forces Fvwc
boundary elements, respectively, would have to be simpli- can be set to 0. Forces Fvwt can then be replaced by Fvw,
fied for inclusion in a design standard. Gulec and Whittaker which acts at a distance a5 from point A (refer to Fig. 10(b)).
(2011) developed a simple empirical equation for peak shear Assuming strain hardening in some of the vertical reinforce-
strength using nonlinear regression analysis of data from ment in tension, force Fvw can be approximated as the product
tests of 74 rectangular shear-critical walls (with and without of the total area of vertical reinforcement and its yield stress.
boundary elements) conducted by others. The patterns of Force Fhw can also be approximated using the total horizontal
cracking at peak lateral strength in the 12 walls tested by reinforcement area and yield stress of the horizontal reinforce-
ment. Equation (33) can then be written as

ACI Structural Journal/March 2019 265


h   l  after simplification, could be used for design and seismic
Fcy (a3 ) + Fvw (a5 ) + Fhw  w  + P  a1 − w  performance assessment of buildings constructed with low-
 2  2
Vn = (34) aspect-ratio RC walls. The empirical equation of Gulec and
hw Whittaker (2011) for peak lateral strength of low-aspect-ratio
walls without boundary elements was updated using data
The functional form of a modified equation for the peak collected from the testing of the 12 large-size walls.
lateral strength of a rectangular shear-critical wall without
boundary elements, based on the equilibrium expression in AUTHOR BIOS
Eq. (34), can be expressed as ACI member Bismarck N. Luna is a Senior Development Specialist at
Linde, plc. He received his BS from University of the Philippines, the Phil-
ippines; his MS degree from Purdue University, West Lafayette, IN; and his
β1 f c′Aeff + β 2 Fvw + β3 Fhw + β 4 P PhD degree from the University at Buffalo, Buffalo, NY. His research inter-
Vn = (35) ests include earthquake engineering and reinforced concrete structures.
(hw /lw )β5
Andrew S. Whittaker, FACI, is a SUNY Distinguished Professor in the
Department of Civil, Structural and Environmental Engineering at the
where the concrete contribution is expressed in terms of √fc′. University at Buffalo. He received his BS from the University of Melbourne,
Melbourne, Australia, and his MS and PhD from the University of Cali-
Regression analysis was performed using MATLAB fornia, Berkeley, Berkeley, CA, in 1977, 1985, and 1988, respectively.
R2012b (MathWorks, Inc. 2012) to determine the values of He is a member of ACI Committee 349, Concrete Nuclear Structures.
the coefficients in Eq. (35). The nonlinear solver fmincon in His research interests include earthquake and blast engineering, and
performance-based seismic design and assessment.
MATLAB R2012b, a function that uses Sequential Quadratic
Programming to find the minimum of a constrained nonlinear
function with multiple variables, was used. Thirteen rectan- REFERENCES
ACI Committee 318, 2014, “Building Code Requirements for Struc-
gular walls without boundary elements were chosen from tural Concrete (ACI 318-14) and Commentary (ACI 318R-14),” American
the database assembled by Gulec and Whittaker (2011): two Concrete Institute, Farmington Hills, MI, 520 pp.
Barda, F.; Hanson, J. M.; and Corley, W. G., 1977, “Shear Strength of
from Cardenas, two from Greifenhagen, and nine from Sheu. Low-Rise Walls with Boundary Elements,” Reinforced Concrete Structures
Nine of Luna’s walls, SW2 through SW10, were included in Seismic Zones, SP-53, American Concrete Institute, Farmington Hills,
in the regression analysis, where their peak lateral strengths MI, pp. 149-202.
Del Carpio Ramos, M.; Whittaker, A. S.; and Gulec, C. K., 2012, “Predic-
were taken as the average of the measured peak strengths tive Equations for the Peak Shear Strength of Low-Aspect Ratio Rein-
in the first and third quadrants of loading. The coefficients forced Concrete Walls,” Journal of Earthquake Engineering, V. 16, No. 2,
determined from the regression analysis using the 22 walls pp. 159-187. doi: 10.1080/13632469.2011.613529
Epackachi, S.; Nguyen, N. H.; Kurt, E. G.; Whittaker, A. S.; and Varma,
without boundary elements are provided in Eq. (36) A. H., 2015, “In-Plane Seismic Behavior of Rectangular Steel-Plate
Composite Wall Piers,” Journal of Structural Engineering, ASCE, V. 141,
1.11 f c′Aeff + 0.25 Fvw + 0.08 Fhw + 0.21P No. 7, p. 04014176 doi: 10.1061/(ASCE)ST.1943-541X.0001148
Vn = (36) Gulec, C. K., and Whittaker, A. S., 2009, “Performance-Based Assess-
(hw /lw )0.81 ment and Design of Squat Reinforced Concrete Shear Walls,” Technical
Report MCEER-09-0010, University at Buffalo, Buffalo, NY, 326 pp.
Gulec, C. K., and Whittaker, A. S., 2011, “Empirical Equations for Peak
The predicted peak shear strength of the 22 walls using Shear Strength of Low Aspect Ratio Reinforced Concrete Walls,” ACI
Structural Journal, V. 108, No. 1, Jan.-Feb., pp. 80-89.
Eq. (36) is shown in Table 4. The ratios of the predicted Gulec, C. K.; Whittaker, A. S.; and Stojadinovic, B., 2008, “Shear
to experimental peak strength, shown in the last column Strength of Squat Rectangular Reinforced Concrete Walls,” ACI Structural
of Table 4, had a mean value of 1.0 (the input constraint) Journal, V. 105, No. 4, July-Aug., pp. 488-497.
Luna, B. N., 2015, “Seismic Response of Low Aspect Ratio Reinforced
and a coefficient of variation of 0.08. The small value of Concrete Walls for Buildings and Safety-Related Nuclear Application,”
the coefficient on the horizontal reinforcement indicates, PhD dissertation, University at Buffalo, Buffalo, NY, 435 pp.
as others have also made clear, that the contribution of the Luna, B. N.; Rivera, J. P.; Epackachi, S.; and Whittaker, A. S., 2018,
“Seismic Response of Low Aspect Ratio Reinforced Concrete Walls,” Tech-
horizontal reinforcement to the peak shear strength is small, nical Report MCEER-18-0002, University at Buffalo, Buffalo, NY, 456 pp.
although this reinforcement is needed for the reasons given Luna, B. N.; Rivera, J. P.; and Whittaker, A. S., 2015, “Seismic Behavior
previously. Equation (36) can therefore be generalized and of Low Aspect Ratio Reinforced Concrete Shear Walls,” ACI Structural
Journal, V. 112, No. 5, Sept.-Oct., pp. 593-603. doi: 10.14359/51687709
simplified as follows MathWorks, Inc., 2012, “MATLAB R2012b,” Natick, MA.
Mattock, A. H., 1976, “Shear Transfer under Monotonic Loading, across
an Interface between Concretes Cast at Different Times, Final Report, Part
1.1 f c′Aeff + 0.25 Fvw + 0.2 P 1,” Report No. SM 76-3, University of Washington, Seattle, WA, 67 pp.
Vn = (37)
(hw /lw )0.8 Mattock, A. H., 1977, “Shear Transfer under Monotonic Loading, across
an Interface between Concretes Cast at Different Times, Final Report, Part
2,” Report No. SM 77-1, University of Washington, Seattle, WA, 101 pp.
SUMMARY AND CONCLUSIONS Moehle, J., 2015, Seismic Design of Reinforced Concrete Buildings,
McGraw-Hill Education, New York, 760 pp.
New equations for the peak lateral strength of shear- Rivera, J. P.; Luna, B. N.; and Whittaker, A. S., 2018, “Seismic Damage
critical RC walls, without and with boundary elements, Assessment of Low Aspect Ratio Reinforced Concrete Shear Walls,” Tech-
have been derived on the basis of internal-force-resisting nical Report MCEER-18-0003, University at Buffalo, Buffalo, NY, 306 pp.
Whyte, C. A., and Stojadinovic, B., 2013, “Hybrid Simulation of the
mechanisms calculated from measured strains and defor- Seismic Response of Squat Reinforced Concrete Shear Walls,” PEER
mations, and patterns of concrete cracking observed during Report 2013/02, Pacific Earthquake Engineering Research Center, Univer-
reversed cyclic testing of 12 large-size rectangular RC walls sity of California, Berkeley, Berkeley, CA, 227 pp.
Wood, S. L., 1990, “Shear Strength of Low-Rise Reinforced Concrete
with aspect ratios of 0.33, 0.54, and 0.94. These equations, Walls,” ACI Structural Journal, V. 87, No. 1, Jan.-Feb., pp. 99-107.

266 ACI Structural Journal/March 2019