ACI STRUCTURAL JOURNAL TECHNICAL PAPER

MS No. S-2017-156.R2

Estimating Drift Capacity of Reinforced Concrete

Structural Walls
by Aishwarya Puranam, Ying Wang, and Santiago Pujol

Methods to proportion special reinforced concrete (RC) structural the field (in Chile: Wallace et al. 2012) and in laboratory
walls are based on the assumption that, unless shear failure, bond experiments (Escolano-Margarit et al. 2012) indicate that
failure, or out-of-plane wall buckling control, deformation capacity the height of the zone in which inelastic compression defor-
is limited by: 1) the ability of concrete to deform in compression; mations concentrate is radically different from the length
and 2) the height of the region in which plastic deformations
along which tension reinforcement yields.
concentrate near the base of the wall. Reliable means to estimate
The availability of a large quantity of recent experi-
these properties are not available. Additionally, failure in a wall is
not always controlled by compression in the concrete, as bar buck- mental programs, which used dense arrays of sensors, has
ling can also limit capacity. The height of the zone in which inelastic facilitated better understanding of the behavior of struc-
compression deformations concentrate is radically different from tural walls (Ghorbani-Renani 2009; Dazio 2009; Liu 2009;
the length along which tension reinforcement yields. Moreover, Johnson 2010; Tran 2012; Takahashi et al. 2013; Villalobos
estimates of drift capacity obtained on the basis of conventional Fernandez 2014; Wang 2014). This investigation evaluates
assumptions deviate by large margins from laboratory test results. four methods to estimate the deformation capacity of rein-
In this investigation, four methods to estimate the drift capacity of forced concrete (RC) structural walls controlled by flexure,
RC walls controlled by flexure were evaluated using results of 40 using experimental test results.
wall tests and a specific recommendation is provided. Drift capacity of walls was estimated using: 1) a formu-
Keywords: bar buckling; drift capacity; inelastic curvatures; plastic hinge;
lation based on concentration of inelastic curvatures in
structural walls. the “hinging” region (Blume et al. 1961); 2) a formula-
tion proposed by Takahashi et al. (2013); 3) a formulation
INTRODUCTION related to the susceptibility of longitudinal reinforcing bars
The conventional approach to estimate drift capacity is to buckle under load reversals proposed by Wang (2014);
based on the assumption that inelastic curvatures concentrate and 4) a formulation proposed by Berry and Eberhard (2005)
in a “hinging” region (Blume et al. 1961). This approach, to estimate drift capacity of columns at the onset of bar
often called the “plastic hinge method,” assumes that unless buckling. These four methods are used to compute flexural
the wall is vulnerable to shear failure or bond failure, the deformations, ignoring slip and shear deformations. More
limiting condition is associated with crushing of concrete involved methods to produce drift ratio estimates might be
in compression. It does not account for effects of buckling available, but the focus of this study is on methods that do
of longitudinal reinforcement and cyclic loading. It also not require help from iterative algorithms.
assumes uniform distribution of compressive strains over the Experimental test data of 40 structural walls with rectan-
“hinging” region. This approach has been widely accepted gular and I-shaped cross sections were used. The data were
and implemented to estimate drift capacity but there are no obtained from the ACI 445-B Shear Wall Database (Song
reliable means to estimate: 1) the limiting compressive strain et al. 2015), which is a compilation of experimental test
of concrete εcu at which the capacity is calculated; or 2) the data of RC structural walls subjected to lateral load. Exper-
length of the region lp in which plastic deformations concen- imental results such as the maximum base shear and drift
trate near the base of the wall. capacity were obtained from the load-deflection response
A limiting concrete strain of 0.003 has typically been used curves. Drift capacities of walls estimated using the four
for design and is known to have produced acceptable lower- methods listed were compared with the measured values.
bound estimates of drift capacity. But much larger values of The focus of this paper is on failures caused by: 1) crushing
εcu have been observed in experimental tests (Thomsen and of concrete in compression; and 2) buckling of longitudinal
Wallace 2004; Dazio 2009; Takahashi et al. 2013; Villalobos reinforcement. The experimental results used were selected
Fernandez 2014). There is also ambiguity involved in to exclude other types of failures. Conventional design
estimating lp. Initial calibrations using data from beams methodologies are based on failure caused by crushing of
suggested that lp was half of the effective depth (Corley concrete in compression. Estimating drift capacities of walls
1966). Wallace and Moehle (1992) suggested lp varies with other failure modes observed in the field (Sritharan et
between 0.5lw and lw. Takahashi et al. (2013) suggested that ACI Structural Journal, V. 115, No. 6, November 2018.
compressive strains in the boundary element concentrate in MS No. S-2017-156.R2, doi: 10.14359/51702444, was received November 9,
2017, and reviewed under Institute publication policies. Copyright © 2018, American
a small region near the wall base. The height of this region Concrete Institute. All rights reserved, including the making of copies unless
is estimated to be 2.5 times the smaller of wall thickness permission is 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
and neutral axis depth. Furthermore, observations made in discussion is received within four months of the paper’s print publication.

ACI Structural Journal/November 2018 1

al. 2014), including wall out-of-plane buckling and rein-
 ε  ε 2
forcement fracture, are not addressed herein. The problem of f c = f c′′ 2 c −  c   for 0 < ε c < � ε 0 (3)
out-of-plane buckling was recently addressed by Parra and  ε 0  ε 0  
Moehle (2014). The problem of fracture is being addressed
by ACI 318 Building Code changes based on experiments by  ε − ε0 
Lu et al. (2015) and Puranam and Pujol (2016). f c = f c′′1 − c 0.15 for ε 0 ≤ � ε c ≤ ε cu (4)
This paper presents an alternative method to estimate drift  ε cu − ε 0 
capacity of structural walls. In the design of structures to where fc is the unit stress in concrete; εc is the unit strain in
resist earthquake demands, the designer can compare the concrete; fc″ is 0.85 times the reported cylinder compressive
estimated capacity with estimates of displacement demand, strength of concrete fc′; and ε0 = 2 fc″/Ec. The elastic modulus
which are related to ground motion intensity and structural of concrete was calculated as 57,000√fc′ for fc′ in psi. Tensile
stiffness (and go beyond the scope of the paper). strength of concrete was neglected.
The limiting compressive strain of concrete εcu was taken
RESEARCH SIGNIFICANCE as the strain at a decay of concrete resistance of 15% from
Predicting drift capacity of structural walls is challenging. the maximum value fc″ considering the effect of confinement
The conventional method to compute drift capacity, which
is the basis for ACI 318-14, Requirement 18.10.6.2, is based  0.15  for walls with confinnement
on the assumption that drift capacity in walls is limited by   ε 0 + ( ε 50 − ε 0 )
ε cu =   0.5 
crushing of concrete in the compression zone. Experimental
0.004
tests and observations from buildings damaged in earth-  for walls without confinement
quakes have shown that, among other factors, bar buckling (5)
can be as detrimental to wall response as concrete crushing.
The findings of this study are helpful in estimating lower- where ε50 is the strain corresponding to a decay of concrete
bound drift capacities of structural walls. resistance of 50% of the maximum value (Roy and Sozen
1965) and is estimated as
ANALYTICAL INVESTIGATION
Experimental test data 3ρ h 
Test results of 428 RC structural walls are summarized ε 50 = min  vb , 0.05 (6)
4 s 
in the ACI 445-B Structural Wall Database. The database is
available at https://datacenterhub.org/resources/142 (Song where ρvb is the volumetric boundary reinforcement ratio;
et al. 2015). Data are divided into three main categories: s is the spacing of transverse reinforcement; and h is the
1) overall geometry; 2) reinforcement details and material smaller of boundary element thickness and length.
properties; and 3) loading information and test results. Walls Neutral axis depth was estimated using equilibrium
for which information from all three categories were avail- and assuming strain is proportional to distance to neutral
able are considered herein. axis. Flexural capacity Mn was estimated by summing the
moments of the forces about the midpoint of the wall cross
Moment versus curvature relationships section (lw/2) and assuming the line of action of axial load
Moment versus curvature relationships were obtained passes through that point.
using the following expressions. A subgroup of 40 slender structural walls subjected to
A stress-strain relationship defined by Eq. (1) and (2) was lateral load reversals and inferred to have failed in flexure
used for the reinforcing bars in the longitudinal direction was selected for this analysis using the following criteria:
1) Height to length ratio of at least 2 (hw/lw > 2);
f s = E ε s for 0 < ε s < � ε y (1) 2) Reached at least 80% of the calculated flexural capacity
(Vmaxh/Mn ≥ 0.8);
3) Shear strength estimated using ACI 318-14 Provision
 fu − f y  11.5.4 (2014) was larger than the maximum shear force
fs = Eε y +  (ε s − ε sh ) for ε s > � ε y (2) applied during the test (Vn > Vmax);
 ε u − ε sh  4) Failure was not controlled by out-of-plane wall buck-
where fs is the unit stress in steel reinforcing bar; fu is the ling, shear, bond, and fracture of longitudinal reinforcement
ultimate unit stress in steel reinforcing bar; fy is the yield Load-deflection response curves of specimens meeting
stress of steel reinforcing bar; εs is the unit strain in steel the criteria mentioned previously were examined to exclude
reinforcing bar; εy is the yield strain of steel reinforcing bar; cases in which a sudden drop in load occurred prior to flex-
εsh is the strain at start of strain hardening, taken as 5εy; εu ural yielding (indicating the possibility of shear failure).
is the ultimate strain of steel reinforcing bar, taken as 0.15 if A summary of the properties for the 40 walls selected is
data were not available; and E is the modulus of elasticity of presented in Tables 1 to 3. Figure 1 shows the calculated
steel reinforcing bar taken as 29,000 ksi. versus measured flexural capacities for the 40 walls.
The stress-strain relationship for concrete was defined Measured drift capacity was extracted from measured
using expressions by Hognestad (1951) and Roy and Sozen load-deflection curves and is defined as the drift at the inter-
(1965) section of horizontal line at 0.8Vmax (Line A in Fig. 2) and the

2 ACI Structural Journal/November 2018

Table 1—Overall geometry
ID Author-Specimen ID Height to load pts, mm Wall length, mm Aspect ratio Web thickness, mm Shape of section
1 Wang (2014)-W1 *
1067 254 4.2 51 I
2 Wang (2014)-W2 *
1067 254 4.2 51 I
3 Wang (2014)-W3 *
1067 254 4.2 51 I
4 Wang (2014)-W4* 1067 254 4.2 51 I
5 Wang (2014)-W5* 1067 254 4.2 51 I
6 Wang (2014)-W6 *
1067 254 4.2 51 I
7 Thomsen et al. (1995)-RW1 *
3810 1219 3.1 102 R
8 Thomsen et al. (1995)-RW2* 3810 1219 3.1 102 R
9 Ji (2002)-Ji_SW2 3000 1000 3 60 I
10 Zhou (2004)-Zhou_SW1 2350 900 2.6 75 R
11 Zhou (2004)-Zhou_SW2 2350 900 2.6 75 R
12 Alarcon et al. (2014)-W1 *
1750 700 2.5 100 R
13 Alarcon et al. (2014)-W2* 1750 700 2.5 100 R
14 Alarcon et al. (2014)-W3 *
1750 700 2.5 100 R
15 Hube et al. (2014)-W4 *
1750 700 2.5 75 R
16 Hube et al. (2014)-W6 *
1750 700 2.5 100 R
17 Hube et al. (2014)-W7* 1750 700 2.5 100 R
18 Hube et al. (2014)-W8 *
1750 700 2.5 100 R
19 Hube et al. (2014)-W9 *
1750 700 2.5 100 R
20 Oesterle et al. (1976)-R1 *
4572 1905 2.4 102 R
21 Oesterle et al. (1976)-B1* 4572 1905 2.4 102 I
22 Oesterle et al. (1976)-B2 4572 1905 2.4 102 I
23 Oesterle et al. (1979)-B10 4572 1905 2.4 102 I
24 Dazio et al. (2009)-WSH1 4560 2000 2.3 150 R
25 Dazio et al. (2009)-WSH2* 4560 2000 2.3 150 R
26 Dazio et al. (2009)-WSH3* 4560 2000 2.3 150 R
27 Dazio et al. (2009)-WSH4 *
4560 2000 2.3 150 R
28 Dazio et al. (2009)-WSH5 *
4560 2000 2.3 150 R
29 Dazio et al. (2009)-WSH6* 4520 2000 2.3 150 R
30 Villalobos (2014)-WMCC* 3315 1524 2.2 203 R
31 Villalobos (2014)-WMCN *
3315 1524 2.2 203 R
32 Villalobos (2014)-W60N *
3315 1524 2.2 203 R
33 Villalobos (2014)-W60C *
3315 1524 2.2 203 R
34 Villalobos (2014)-W40C* 3315 1524 2.2 203 R
35 Villalobos (2014)-W60N2 *
3315 1524 2.2 203 R
36 Zhang et al. (2000)-SW7 1500 700 2.1 100 R
37 Ghorbani-Renani et al. (2009)-B2C 1140 548 2.1 84 R
38 Ghorbani-Renani et al. (2009)-A2C 2700 1300 2.1 200 R
39 Tran (2012)-RW-A20-P10-S38 2438 1219 2 152 R
40 Tran (2012)-RW-A20-P10-S63 2438 1219 2 152 R
Maxima 4572 3327 4 203 —
Minima 1067 254 2 51 —
*
Specimens in which bar buckling was observed.
Note: 1 in. = 25.4 mm.

ACI Structural Journal/November 2018 3

Table 2—Reinforcement and concrete properties
Web longitudinal Boundary longitudinal Web transverse Boundary transverse Concrete compres-
ID Author-Specimen label reinforcement ratio reinforcement ratio reinforcement ratio reinforcement ratio (vol.) sive strength, MPa
1 Wang (2014)-W1* 0.79% 2.80% 0.48% 0.00% 34.7
2 Wang (2014)-W2* 0.79% 2.80% 0.48% 0.00% 34.7
3 Wang (2014)-W3 *
0.79% 2.80% 0.48% 0.00% 32.1
4 Wang (2014)-W4 *
0.79% 2.80% 0.48% 0.00% 30.8
5 Wang (2014)-W5 *
0.79% 2.80% 0.48% 0.00% 30.1
6 Wang (2014)-W6* 0.79% 2.80% 0.48% 0.00% 28.8
7 Thomsen et al. (1995)-RW1* 0.30% 2.93% 0.33% 0.90% 42.1
8 Thomsen et al. (1995)-RW2 *
0.30% 2.93% 0.33% 1.00% 39.1
9 Ji (2002)-Ji_SW2 0.40% 4.70% 0.40% 3.30% 9
10 Zhou (2004)-Zhou_SW1 0.70% 3.72% 0.84% 1.70% 25.4
11 Zhou (2004)-Zhou_SW2 0.70% 3.72% 0.84% 1.70% 25.4
12 Alarcon et al. (2014)-W1 *
0.72% 0.45% 0.44% 0.00% 26
13 Alarcon et al. (2014)-W2 *
0.72% 0.45% 0.44% 0.00% 26
14 Alarcon et al. (2014)-W3 *
0.72% 0.45% 0.44% 0.00% 26
15 Hube et al. (2014)-W4* 0.67% 0.49% 0.46% 0.00% 26
16 Hube et al. (2014)-W6 *
1.34% 0.00% 0.44% 0.00% 26
17 Hube et al. (2014)-W7 *
0.72% 0.45% 0.44% 0.00% 26
18 Hube et al. (2014)-W8 *
0.72% 0.45% 0.64% 1.16% 26
19 Hube et al. (2014)-W9* 0.72% 0.45% 0.56% 0.00% 26
20 Oesterle et al. (1976)-R1 *
0.25% 1.47% 0.31% 0.00% 42.5
21 Oesterle et al. (1976)-B1 *
0.29% 1.11% 0.31% 0.00% 50.4
22 Oesterle et al. (1976)-B2 0.29% 3.67% 0.63% 0.00% 50.9
23 Oesterle et al. (1979)-B10 0.29% 1.97% 0.63% 1.35% 43.3
24 Dazio et al. (2009)-WSH1 0.30% 1.32% 0.25% 1.04% 42.8
25 Dazio et al. (2009)-WSH2 *
0.30% 1.32% 0.25% 1.04% 38.5
26 Dazio et al. (2009)-WSH3 *
0.54% 1.54% 0.25% 0.95% 37.2
27 Dazio et al. (2009)-WSH4* 0.54% 1.54% 0.25% 0.00% 38.9
28 Dazio et al. (2009)-WSH5* 0.27% 0.67% 0.25% 1.22% 36.4
29 Dazio et al. (2009)-WSH6 *
0.54% 1.54% 0.25% 1.58% 43.3
30 Villalobos (2014)-WMCC *
0.40% 4.90% 0.55% 1.23% 29.2
31 Villalobos (2014)-WMCN *
0.40% 4.90% 0.55% 0.00% 28.8
32 Villalobos (2014)-W60N* 0.40% 4.90% 0.55% 0.00% 32.1
33 Villalobos (2014)-W60C *
0.40% 4.90% 0.55% 1.23% 29.8
34 Villalobos (2014)-W40C *
0.40% 4.90% 0.55% 1.23% 29.5
35 Villalobos (2014)-W60N2 *
0.40% 4.90% 0.55% 0.00% 30.1
36 Zhang et al. (2000)-SW7 0.67% 0.88% 1.01% 2.30% 27.6
Ghorbani-Renani et al.
37 0.71% 3.56% 0.67% 2.60% 39.5
(2009)-B2C
Ghorbani-Renani et al.
38 0.59% 4.03% 0.59% 2.02% 23.8
(2009)-A2C
39 Tran (2012)-RW-A20-P10-S38 0.27% 3.23% 0.27% 1.70% 44.7
40 Tran (2012)-RW-A20-P10-S63 0.61% 7.11% 0.61% 1.60% 46.2
Maxima 1.34% 7.11% 1.01% 3.30% 51
Minima 0.25% 0.00% 0.25% 0.00% 9.00
*
Specimens in which bar buckling was observed.
Note: 1 psi = 0.006895 MPa.

4 ACI Structural Journal/November 2018

Table 3—Loading conditions and test results
ID Author-Specimen label Height to load point, mm Axial load ratio Maximum base shear, kN Measured roof drift capacity, mm
1 Wang (2014)-W1 *
1067 0 19 53
2 Wang (2014)-W2 *
1067 0 17 43
3 Wang (2014)-W3 *
1067 0 19 43
4 Wang (2014)-W4* 1067 0 16 43
5 Wang (2014)-W5* 1067 0 19 43
6 Wang (2014)-W6 *
1067 0 15 32
7 Thomsen et al. (1995)-RW1 *
3810 0.08 149 70
8 Thomsen et al. (1995)-RW2* 3810 0.08 158 85
9 Ji (2002)-Ji_SW2 3000 0.27 117 56
10 Zhou (2004)-Zhou_SW1 2350 0.23 140 46
11 Zhou (2004)-Zhou_SW2 2350 0 155 47
12 Alarcon et al. (2014)-W1 *
1750 0.16 144 48
13 Alarcon et al. (2014)-W2* 1750 0.26 166 31
14 Alarcon et al. (2014)-W3 *
1750 0.37 186 27
15 Hube et al. (2014)-W4 *
1750 0.16 113 24
16 Hube et al. (2014)-W6 *
1750 0.16 138 34
17 Hube et al. (2014)-W7* 1750 0.16 149 38
18 Hube et al. (2014)-W8 *
1750 0.16 156 44
19 Hube et al. (2014)-W9 *
1750 0.16 145 43
20 Oesterle et al. (1976)-R1 *
4572 0 118 108
21 Oesterle et al. (1976)-B1* 4572 0 271 152
22 Oesterle et al. (1976)-B2 4572 0 680 127
23 Oesterle et al. (1979)-B10 4572 0.09 707 133
24 Dazio et al. (2009)-WSH1 4560 0.05 336 42
25 Dazio et al. (2009)-WSH2* 4560 0.06 359 65
26 Dazio et al. (2009)-WSH3* 4560 0.06 454 93
27 Dazio et al. (2009)-WSH4 *
4560 0.06 443 60
28 Dazio et al. (2009)-WSH5 *
4560 0.14 439 60
29 Dazio et al. (2009)-WSH6* 4520 0.11 597 95
30 Villalobos (2014)-WMCC* 3315 0.1 703 100
31 Villalobos (2014)-WMCN *
3315 0.1 694 83
32 Villalobos (2014)-W60N *
3315 0.09 707 68
33 Villalobos (2014)-W60C *
3315 0.1 721 85
34 Villalobos (2014)-W40C* 3315 0.1 721 85
35 Villalobos (2014)-W60N2 *
3315 0.1 756 67
36 Zhang et al. (2000)-SW7 1500 0.26 201 32
37 Ghorbani-Renani et al. (2009)-B2C 1140 0 82 25
38 Ghorbani-Renani et al. (2009)-A2C 2700 0 429 81
39 Tran (2012)-RW-A20-P10-S38 2438 0.08 481 76
40 Tran (2012)-RW-A20-P10-S63 2438 0.08 742 74
Maxima 4572 0.37 756 152
Minima 1067 0 15 25
*
Specimens in which bar buckling was observed.
Notes: 1 in. = 25.4 mm; 1 kip = 4.448 kN.

ACI Structural Journal/November 2018 5

Fig. 1—Estimated versus measured flexural capacity for
subgroup of 40 walls selected. (Note: 1 kip-ft = 1.356 kN-m.) Fig. 3—Definition of curvature at flexural yielding.

Fig. 2—Definitions of measured drift ratio at yield and drift

ratio capacity.
envelope to the load-deflection curve. Vmax is the absolute
maximum lateral force. Drift ratio capacity is defined herein
as the ratio of drift capacity to wall height to loading point.

Estimating drift capacity

Drift capacity for the subgroup of 40 walls was estimated
using four methods.
Method 1: The conventional approach to estimate drift
capacity is based on the assumption that inelastic curvatures Fig. 4—Measured DR versus DR calculated using Eq. 7(a).
concentrate in a “hinging” region (Blume et al. 1961). This to loading point; εcu is limiting concrete compressive strain
model has been used to estimate drift capacity of walls by estimated using Eq. (5), and c is the calculated neutral axis
previous researchers (for example, Wallace and Moehle depth (as described in the “Moment versus curvature rela-
1992; Thomsen and Wallace 2004; Takahashi et al. 2013). tionship” section).
The length over which inelastic curvature is assumed to Curvature at flexural yielding φy was obtained from
accumulate is commonly referred to as the length of the calculated moment-curvature relationships as the curva-
idealized plastic hinge lp. Drift ratio capacity DRu, defined ture at the intersection of the tangent to the cracked section
as drift capacity divided by wall height, was estimated using and the tangent at the point of maximum moment (Fig. 3).
Eq. 7(a) Plastic hinge lengths ranging from 0.5lw to lw were consid-
ered herein. Figure 4 shows drift ratio capacities calculated
1 1  lp   using Eq. 7(a) plotted against measured drift ratio capacities.
DRu =
h 3
2
( )
 ϕ y h + ϕu − ϕ y l p  h −  
 2
(7a) Equation 7(b), a simplified version of Eq. 7(a), was also used
to estimate drift ratio capacity.
where φy is the curvature at flexural yielding; φu is limiting
curvature (εcu/c); lp is plastic hinge length; h is wall height DRu = φulp (7b)

6 ACI Structural Journal/November 2018

 1
DRy = ϕ y h (9)
3
where φy is curvature at flexural yielding used in Method 1;
and h is wall height to loading point.
The second component of Eq. (8), related to plastic defor-
mations, is calculated using Eq. (10)

εp
DR p = l p (10)
c
εp is defined as the plastic component of ultimate strain by
Takahashi et al. (2013). In this study, the value of curva-
ture at a concrete compressive strain of εcu (estimated using
Eq. (5)) is used as the ratio of εp to c. Plastic hinge length
lp is taken as 2.5 times the minimum of wall thickness t and
neutral axis depth c. This recommendation for lp by Taka-
hashi et al. (2013) was based on uniaxial compression tests
done by Markeset and Hillerborg in 1995. Figure 6 shows
drift ratio capacities calculated using Eq. (8) plotted against
measured DR capacities.
Method 3: A method to estimate limiting drift ratio of
structural walls with longitudinal reinforcement prone
to buckling was proposed by Wang (2014). It is based on
the idea that buckling occurs when the ratio of tangential
modulus of elasticity to compressive stress reaches a critical
Fig. 5—Measured DR versus DR calculated using Eq. 7(b).
value in longitudinal reinforcing bars. From Euler’s work on
buckling in 1757

EI
Pcr = C0 (11)
l2
where Pcr is critical axial load; C0 is coefficient related to
the restraint at bar ends; l is buckling length; I is moment of
inertia of the bar; and E is initial modulus of the bar.
Engesser extended Euler’s expression to the nonlinear
range of response in 1889. Initial modulus E was replaced
π
with tangential modulus Et. For E = Et and I = rb4
4
Pcr Et
= σ cr = C0 (12)
Ab (l / rb )2
where σcr is critical stress; Ab is cross-sectional area of bar;
Fig. 6—Measured DR versus DR calculated using formula- Et is tangential modulus; rb is bar radius (db/2); and l is buck-
tion (Eq. (8)) by Takahashi et al. (2013). ling length.
Rearranging Eq. (12) gives ratio of tangential modulus
Figure 5 shows drift ratio capacities calculated using to critical stress expressed as a function of the unsupported
Eq. 7(b) plotted against measured values. length, bar radius, and coefficient C0
Method 2: Takahashi et al. (2013) proposed a method
to calculate drift capacity of structural walls based on test Et
= (l / rb ) / C0 (13)
2
results of 10 RC wall specimens. Flexural drift ratio capacity
σ cr
was calculated using Eq. (8), which is similar to Eq. 7(a)
Bresler and Gilbert (1961) studied the variables influencing
DRu = DRy + DRp (8) longitudinal bar buckling in columns and proposed an
expression for minimum tie spacing in columns to prevent
where DRy and DRp are the linear and plastic components of longitudinal bar buckling at a stress smaller than the yield
drift ratio capacity, respectively. stress. Rodriguez et al. (1999) tried to determine a relation-
For a cantilever wall with one concentrated load at its ship between tensile strain attained in a previous reversal
free end, the linear component of Eq. (8) is calculated using and the onset of bar buckling. Wang (2014) suggested
Eq. (9) that bar buckling can be assumed to occur when the ratio

ACI Structural Journal/November 2018 7

 assumed the height of the region in which concrete may
crush and therefore bars buckle (l) to be the minimum
of: 1) neutral axis depth, c; and 2) width of the boundary
element in compression, t. Neutral axis depth was estimated
at a concrete compressive strain εcu estimated using Eq. (5)
and using the assumptions described in the “Moment versus
curvature relationship” section. Figure 7 shows drift ratio
capacities (DR) calculated using Eq. (16) plotted against
measured drift ratio capacities.
Method 4: A model to compute drift capacity at the onset
of bar buckling in RC columns was proposed by Berry and
Eberhard (2005). This model was calibrated using cyclic
tests of 62 columns with rectangular sections and 42 columns
with circular sections. Although the method was conceived
and calibrated for columns, it is used to estimate the drift
capacity of walls here to save time others may spend testing
Fig. 7—Measured DR versus DR calculated using formula- it on the basis of similarities in the mechanics of columns
tion (Eq. (16)) by Wang (2014). and walls controlled by flexure. Drift ratio at the onset of bar
bulking is expressed as
of tangential modulus of elasticity of the reinforcing steel
bars to compressive stress (Et/σmin) reaches a critical value.  f y db  
−1
DR = ∆ bb λ L  P   L
The Et/σmin ratio was inferred to decrease as the peak tensile
L
=
3 Es D 
( 
)
f y + C0  1 + C1ρeff 1 + C2 
f c′Ag    1 + C3 + C4
D D  

strain reached in a preceding cycle increased
(17)
Et 1
= − 1 (14)
σ min 4ε where λ is the parameter for yield curvature approximation;
Es is the elastic modulus of longitudinal reinforcement; fy
A relationship between maximum tensile strain and
is the yield stress of longitudinal reinforcement; L is the
maximum drift ratio was obtained by Wang (2014) using
distance from the column base to the point of contraflexure,
strain measurements
taken as wall height to loading point herein; D is column
depth, taken as wall length herein; ρeff is effective confine-
 ε
for walls with continuous longitudinal reinforcementt ment ratio; P is applied axial load; Ag is gross area of column
DR =  1.5
cross section, taken as gross area of wall cross section herein;
 ε for walls with discontinuous longitudinal reinforcement
 2.5 and fc′ is compressive strength of concrete. C0 = 1.47, C1 =
1.33, C2 = 1.88, and C4 = 0.08 are parameters used in defor-
(15) mation approximations of rectangular reinforced columns as
in Table 3 of Berry and Eberhard (2005).
DR is drift ratio (drift divided by wall height) and ε is Figure 8 shows drift ratio capacities calculated using
maximum tensile strain in the longitudinal reinforcing bar. Eq. (17) plotted against measured drift ratio capacities.
Discontinuous longitudinal reinforcement refers to longitu- In all figures showing measured versus calculated drift
dinal reinforcement with lap splices or bar cutoffs within the capacities: 1) the solid line represents a line with 1:1 slope;
zone in which the reinforcement is expected to yield. 2) the dotted line represents the best-fit straight line through
Combining Eq. (13) through (15) the origin; and 3) the values of R2 correspond to the shown
best-fit straight line through the origin (that is, the dotted
 1
  2

for walls with continuous longitudinal reinforcement line).
  1  
   rb  
 6 + 1
RESULTS AND DISCUSSION
  35 
   Table 4 lists means and standard deviations of the ratio
 
DR =  of calculated to measured drift ratio capacity for the four
 1
  l  2 
for walls with discontinuous longitudinal reinforcement formulations described in the previous section and all the
   specimens listed in Table 1. Figures 4 through Fig. 8 show
   rb  
  35
10 + 1  comparisons of calculated and measured drift ratio capaci-
  
   ties. The most salient conclusion from these comparisons is
  
that the quality of the calculated drift capacities is not very
(16) good. The scatter in all these comparisons makes it difficult
to pick one formulation over the others.
where rb is bar radius (db/2); and l is buckling length. It is interesting to note that the results from Eq. 7(a) and (b)
Using the recommendations and data reported in Taka- are similar for all “hinge” lengths considered. Equation 7(b)
hashi et al. (2013) and Villalobos (2014), Wang (2014) is preferred over Eq. 7(a) for estimating drift capacity, as it

8 ACI Structural Journal/November 2018

involves less computational effort while producing results could be due to: 1) concentration of inelastic deformations in
of similar quality. The mismatch between measured drift concrete; and 2) bar buckling. In approximately 70% of the
ratio capacity and estimates obtained using Eq. 7(a) and (b) studied walls, failure was reported to have been controlled
by bar buckling and not by concrete crushing. The extent to
which confinement and shear deformations contribute to the
said mismatch between measured and calculated drift was
also investigated:
a) Limiting concrete strength was varied (Eq. (5)) to
account for confinement in the boundary elements. The
effect of confinement was considered by using the expres-
sions proposed by Roy and Sozen (1965). Results from this
analysis and tests conducted at UC Berkeley (Arteta 2016)
and CERL show that the deformation capacity of boundary
elements in compression is not as sensitive to confinement
from rectangular hoops and crossties as previously believed.
b) Conventionally shear deformations are estimated as

6 V
γ= (18)
5 G × Acv
Fig. 8—Measured DR versus DR calculated using formula- For shear modulus (G) = 0.4 × 4000 ksi = 1600 ksi and shear
tion (Eq. (17)) by Berry and Eberhard (2005). stress (V/Acv) as high as 300 psi, this expression produces
Table 4—Comparison of results
Calculated/measured DR
Assumed plastic hinge Standard Coefficient of
Equation No. Formulation/source length Mean deviation variation
All specimens
Lw/2 0.87 0.57 0.65
2Lw/3 1.06 0.72 0.68
(7a) Conventional/Blume et al. (1961)
3Lw/4 1.15 0.79 0.68
Lw 1.40 0.98 0.70
Lw/2 0.88 0.64 0.73
2Lw/3 1.17 0.85 0.73
(7b) Conventional/Blume et al. (1961)
3Lw/4 1.32 0.96 0.73
Lw 1.76 1.28 0.73
(8) Takahashi et al. (2013) NA 0.67 0.38 0.56
(16) Wang (2014) NA 0.70 0.30 0.43
(17) Berry and Eberhard (2005) NA 1.52 0.56 0.37
Min. of (7b) and (16) Blume et al. (1961) and Wang (2014) 3Lw/4 0.55 0.17 0.31
Specimens for which bar buckling was not reported
Lw/2 1.37 0.65 0.48
2Lw/3 1.71 0.82 0.48
(7a) Conventional/Blume et al. (1961)
3Lw/4 1.86 0.90 0.48
Lw 2.28 1.11 0.49
Lw/2 1.46 0.72 0.49
2Lw/3 1.95 0.96 0.49
(7b) Conventional/Blume et al. (1961)
3Lw/4 2.19 1.08 0.49
Lw 2.92 1.44 0.49
(8) Takahashi et al. (2013) NA 1.00 0.50 0.50
Specimens for which bar buckling was reported
(16) Wang (2014) NA 0.72 0.3 0.44
(17) Berry and Eberhard (2005) NA 1.41 0.3 0.22

ACI Structural Journal/November 2018 9

Fig. 9—Measured DR versus DR calculated using Eq. (7a) Fig. 11—Measured DR versus calculated DR (using Methods
and with Eq. (18) and without contribution from shear 3 and 4) for walls in which bar buckling was observed.
deformations.

Fig. 12—Measured DR versus minimum of DR calculated

using Eq. (7b) and formulation (Eq. (16)) by Wang (2014).
than 20% on average. This change is exceeded by a large
margin by the spread in the data (as shown in Fig. 9). There-
fore, shear deformations are ignored, as including them does
not help improve results.
The quality of the results does not improve by a large
margin when Eq. 7(a) and the expression proposed by Taka-
hashi et al. (2013, Eq. (8)), which were calibrated to consider
the effects of crushing only, are used exclusively for walls in
which bar buckling was not observed (Fig. 10). The quality
of the results improves (Fig. 11(a)) when the formulation
by Wang (2014, Eq. (16)), which was calibrated to take bar
buckling into account, is used exclusively for walls in which
bar buckling was observed (walls marked with asterisk in
Tables 1 through 3).
The error in the ratio of calculated to measured drift ratio was
observed to be nearly insensitive to variations in axial load ratio
and longitudinal and transverse reinforcement ratios.
After a careful examination of the available data and the
Fig. 10—Measured DR versus calculated DR (using strengths of the formulations studied (Table 4), a reasonable
Methods 1 and 2) for walls in which bar buckling was not lower bound to wall drift ratio capacity was obtained by
observed. using the minimum between Eq. 7(b) (with lp = 3lw/4 ) and
Eq. (16) (Fig. 12). This compromise also happens to produce
a negligible estimate of 0.02%. Assuming that shear defor-
the smallest standard deviation in the ratio of calculated to
mation can be as high as 30 × γ (Matamoros 1999) does
measured drift ratio capacity and a mean of approximately
not help reduce the scatter in the measured versus calculated
3/5 (Table 4). The inferences made apply only within the
drift capacities and does not change drift capacity by more
ranges defined by the maxima and minima listed in Tables 1

10 ACI Structural Journal/November 2018

through 3. It should also be noted that: 1) roof drift ratio is As = cross-sectional area of longitudinal reinforcement
Asc = gross sectional area of all longitudinal bars in compression
smaller than maximum interstory drift ratio, especially in the Ast = gross sectional area of all longitudinal bars in tension
linear range of response; and 2) roof drift ratio exceeding 2% C = compressive force of concrete
is undesirable for many reasons, including P-delta effects C0 = coefficient related to restraint of bar ends (Wang 2014)
c = neutral axis depth
and safety of contents. D = column depth (taken as length of wall in this analysis)
DR = drift ratio capacity
SUMMARY AND CONCLUSIONS DRy = elastic component of drift ratio capacity
DRp = plastic component of drift ratio capacity
A subgroup of 40 RC walls was selected from a structural d = effective depth
wall database with data from 428 tests. The selected walls E = modulus of elasticity of steel reinforcing bar
were inferred to have failed in flexure. Their measured drift Et = tangential modulus
fc = concrete stress
ratio capacities (DR, ratio of roof drift to wall height) were fc′ = 28-day concrete compressive stress
compared with estimates from four methods. fc″ = 0.85fc′
Method 1 was the conventional approach based on fs = unit stress in steel reinforcing bar
fu = ultimate stress of longitudinal reinforcement
assumed plastic hinge length lp and limiting compres- fy = yield stress of longitudinal reinforcement
sive concrete strain εcu. Plastic hinge lengths ranging from G = shear modulus
0.5lw to lw were considered. No discernible correlation was h = wall height
I = moment of intertia
observed between measured DR and DR calculated with l = buckling length
Method 1. This mismatch may be a consequence of diffi- lp = plastic hinge length
culties associated with estimating εcu and lp. It may also be lw = wall length
L = distance from column base to point of contraflexure (Berry and
related to the fact that, in approximately 70% of the studied Eberhard 2005) taken as wall height to load point herein
walls, failure may have been controlled by bar buckling Mn = nominal flexural capacity
and not by crushing as assumed in Method 1. Drift capac- P, N = applied axial load
Pcr = critical axial load
ities were also estimated using formulations by Takahashi rb = radius of bar (db/2)
et al. (2013), Wang (2014), and Berry and Eberhard (2005) t = wall width (thickness)
(referred to as Methods 2, 3, and 4). Method 3 is based on Vmax = maximum total lateral force
Vn = shear strength
the assumption that failure in a structural wall is controlled α = wall aspect ratio
by longitudinal bar buckling. Buckling is assumed to occur β1 = factor relating depth of equivalent rectangular compressive
when the ratio of compressive stress to tangential modulus stress (ACI Committee 318 2014)
ε = maximum tensile strain
reaches a threshold. This ratio was inferred to decrease as εc = concrete strain
peak tensile strain increased, implying the likelihood of bar εcu = limiting concrete strain
buckling increases as drift ratio increases. εp = plastic component of ultimate strain
εs = unit strain in steel reinforcing bar
It was concluded that a reasonable lower bound to drift εy = yield strain in steel reinforcing bar
capacity (DR) of a wall governed by flexure and restrained ε0 = concrete strain at peak stress
against out-of-plane buckling may be estimated as the ε50 = strain corresponding to decay of concrete resistance of 50% of
maximum value (Roy and Sozen 1965)
smaller of: 1) the product of limiting curvature and 3/4 of γ = shear strain
wall length; and 2) the limit by Wang (2014) that is related φu = limiting curvature
to susceptibility of longitudinal bars to buckling. This φy = curvature at flexural yielding
λ = parameter for yield curvature approximation
conclusion applies only within the ranges defined in Tables ρeff = effective confinement ratio (Berry and Eberhard 2005)
1 through 3. σcr = critical stress

AUTHOR BIOS REFERENCES

ACI member Aishwarya Puranam is a Graduate Student in the School ACI Committee 318, 2014, “Building Code Requirements for Struc-
of Civil Engineering at Purdue University, West Lafayette, IN, where she tural Concrete (ACI 318-14) and Commentary (ACI 318R-14),” American
received her BS and MS in civil engineering. Concrete Institute, Farmington Hills, MI, 519 pp.
Arteta, C. A., 2016, “Seismic Response Assessment of Thin Boundary
Ying Wang is a Senior Structural Engineer at WSP Global, New York, Elements of Special Concrete Shear Walls,” PhD thesis, School of Civil and
NY. She received her MS in civil engineering and her PhD from Purdue Environmental Engineering, University of California, Berkeley, Berkeley, CA.
University. Berry, M. P., and Eberhard, M. O., 2005, “Practical Performance Model
for Bar Buckling,” Journal of Structural Engineering, ASCE, V. 131, No. 7,
Santiago Pujol, FACI, is a Professor in the School of Civil Engineering pp. 1060-1070. doi: 10.1061/(ASCE)0733-9445(2005)131:7(1060)
at Purdue University. He is a member of ACI Committees 133, Disaster Blume, J. A.; Newmark, N. M.; and Corning, L. H., 1961, “Design of
Reconnaissance; 314, Simplified Design of Concrete Buildings; ACI Multistory Reinforced Concrete Buildings for Earthquake Motions,” Port-
Subcommittee 318-R, High-Strength Reinforcement; and Joint ACI-ASCE land Cement Association, Skokie, IL.
Committee 445, Shear and Torsion. Bresler, B., and Gilbert, P. H., 1961, “Tie Requirements for Rein-
forced Concrete Columns,” ACI Structural Journal, V. 58, No. 11, Nov.,
ACKNOWLEDGMENTS pp. 555-570.
The authors would like to thank J. Rautenberg, X. Lu, Y. Zhou, J. Yang, Escolano-Margarit, D.; Klenke, A.; Pujol, S.; and Benavent-Climent, A.,
J. Qian, C. Song, and M. Usta for helping compile and process the exper- 2012, “Failure Mechanism of Reinforced Concrete Structural Walls with
imental data. and without Confinement,” Proceedings of the Fifteenth World Conference
on Earthquake Engineering, Lisboa, Portugal.
Lu, Y.; Henry, R. S.; and Ma, Q. T., 2015, “Experimental Testing and
NOTATION Modelling of Reinforced Concrete Walls with Minimum Vertical Rein-
Ab = cross-sectional area of bar forcement,” Proceedings of the 2015 NZSEE Annual Conference, Rotorua,
Acv = gross area of concrete section bounded by web thickness and New Zealand.
length of section

ACI Structural Journal/November 2018 11

 Matamoros, A. B., 1999, “Study of Drift Limits for High-Strength Walls and Design Implications,” Earthquake Spectra, V. 30, No. 1,
Concrete Columns,” PhD thesis, University of Illinois at Urbana-Cham- pp. 307-334. doi: 10.1193/021713EQS036M
paign, Urbana, IL, 435 pp. Takahashi, S.; Yoshida, K.; Ichinose, T.; Sanada, Y.; Matsumoto, K.;
Parra, P. F., and Moehle, J. P., 2014, “Lateral Buckling in Reinforced Fukuyama, H.; and Suwada, H., 2013, “Flexural Drift Capacity of Rein-
Concrete Walls,” Tenth National Conference on Earthquake Engineering, forced Concrete Wall with Limited Confinement,” ACI Structural Journal,
Anchorage, AK. V. 101, No. 1, Jan.-Feb., pp. 95-104.
Puranam, A., and Pujol, S., 2017, “Minimum Flexural Reinforcement in Thomsen, J. H. IV, and Wallace, J. W., 2004, “Displacement-Based
Reinforced Concrete Walls,” Proceedings of the 16th World Conference on Design of Slender Reinforced Concrete Structural Walls—Experimental
Earthquake Engineering, Santiago, Chile. Verification,” Journal of Structural Engineering, ASCE, V. 130, No. 4,
Rodriguez, M. E.; Botero, J. C.; and Villa, J., 1999, “Cyclic Stress-Strain pp. 618-630. doi: 10.1061/(ASCE)0733-9445(2004)130:4(618)
Behavior of Reinforcing Steel Including Effect of Buckling,” Journal of Wallace, J. W.; Massone, L. M.; Bonelli, P.; Dragovich, J.; Lagos, R.;
Structural Engineering, ASCE, V. 125, No. 6, pp. 605-612. doi: 10.1061/ Lüders, C.; and Moehle, J., 2012, “Damage and Implications for Seismic
(ASCE)0733-9445(1999)125:6(605) Design of RC Structural Wall Buildings,” Earthquake Spectra, V. 28,
Roy, H. E. H., and Sozen, M. A., 1965, “Ductility of Concrete,” Flexural pp. S281-S299. doi: 10.1193/1.4000047
Mechanics of Reinforced Concrete, pp. 213-235. Wallace, J. W., and Moehle, J. P., 1992, “Ductility and Detailing
Song, C.; Wang, Y.; Puranam, A.; and Pujol, S., ACI Subcommittee Requirements of Bearing Wall Buildings,” Journal of Structural
445-B, Usta, M., 2015, “ACI 445B Shear Wall Database,” https:// Engineering, ASCE, V. 118, No. 6, pp. 1625-1644. doi: 10.1061/
datacenterhub.org/resources/142. (last accessed Sept. 17, 2018) (ASCE)0733-9445(1992)118:6(1625)
Sritharan, S.; Beyer, K.; Henry, R. S.; Chai, Y. H.; Kowalsky, M.; and Wang, Y., 2014, “Effects of Web Reinforcement Discontinuities on the
Bull, D., 2014, “Understanding Poor Seismic Performance of Concrete Seismic Response of Structural Walls,” PhD thesis, School of Civil Engi-
neering, Purdue University, West Lafayette, IN.

12 ACI Structural Journal/November 2018