Drift Capacity of Reinforced Concrete

Columns with Light Transverse

Reinforcement

Kenneth J. Elwood,a) M.EERI, and Jack P. Moehle,b) M.EERI

Existing reinforced concrete columns with light transverse reinforcement

are vulnerable to shear failure during seismic response. Shear strength mod-
els, modeling the degradation of shear strength with increasing displacement
ductility demand, have been widely used to evaluate the interstory drift ca-
pacity of such columns. The application of a shear strength model to deter-
mine the drift capacities for a database of 50 shear-critical columns demon-
strates significant inaccuracies with such a method. An empirical drift
capacity model based on the shear-critical column database provides a better
estimate of the interstory drift at shear failure. The new drift capacity model
identifies the most critical parameters affecting the drift capacity of shear-
critical columns, namely, transverse reinforcement ratio, shear stress demand,
and axial load ratio. [DOI: 10.1193/1.1849774]

INTRODUCTION
It has been well established by experimental evidence that many existing reinforced
concrete columns are vulnerable to shear failure after flexural yielding (Masaya 1973,
Wight and Sozen 1973, Ohue et al. 1985, among others). Several models have been de-
veloped to represent the degradation of shear strength with increasing inelastic defor-
mations (Watanabe and Ichinose 1992, Aschheim and Moehle 1992, Priestley et al.
1994, Sezen 2002). While these shear strength models are useful for estimating the col-
umn strength as a function of deformation demands, it is unclear whether they are reli-
able for assessing the drift at shear failure. Drift capacity models are an essential ingre-
dient for displacement-based design and assessment methods for existing buildings
(ATC 1996, ASCE 2000).
Pujol et al. (1999) have proposed a drift capacity model for columns failing in shear.
The model makes an important contribution by focusing attention directly on displace-
ment capacity and by analyzing data for model development. Kato and Ohnishi (2002)
also have proposed a model for drift capacity at shear failure. The data sets used in these
studies include columns with circular cross sections and columns with transverse rein-
forcement ratios approaching those used in modern, special moment-resisting frames
(ACI 2002). Given the sensitivity of failure mechanisms to column details, it is impor-
tant to reexamine drift capacity models for conditions more representative of those oc-
curring in older reinforced concrete buildings. Of particular interest are reinforced con-

a)
Department of Civil Engineering, University of British Columbia, Vancouver, BC, Canada V6T 1Z4
b)
Pacific Earthquake Engineering Research Center, University of California, Berkeley, CA 94720

71
Earthquake Spectra, Volume 21, No. 1, pages 71–89, February 2005; © 2005, Earthquake Engineering Research Institute
72 K. J. ELWOOD AND J. P. MOEHLE

crete columns from pre-1970s construction in California that tend to have spacing of the
transverse reinforcement greater than half of the minimum cross-section dimension. (A
survey of reinforced concrete columns designed prior to the early 1970s is discussed in
Lynn [2001].)
The present study will use a database of 50 shear-critical reinforced concrete col-
umns having configurations representative of those used in older building construction.
For the database, the drift capacity at shear failure is calculated by the following models:
• The shear strength model by Sezen (2002)
• The statistical drift capacity model by Pujol et al. (1999)
• A plastic drift capacity model by Kato and Ohnishi (2002)
• An empirical model based on observations from the database
Results of the evaluation may be useful for the seismic evaluation of reinforced concrete
building columns having light transverse reinforcement.

EXPERIMENTAL DATABASE
Sezen (2002) compiled a database of 50 laboratory tests on reinforced concrete col-
umns representative of columns from older reinforced concrete buildings. The test col-
umns were selected based on a search of the literature for specimens tested under uni-
directional cyclic lateral load with low transverse reinforcement ratios (␳⬙⬍0.007),
yielding of longitudinal reinforcement prior to loss of lateral load capacity, and shear
distress observed at failure suggesting that loss in lateral load capacity was due to deg-
radation of the shear-transfer mechanism. All attempts were made to select large-scale
test specimens; however, due to the limited number of tests meeting the criteria listed
above, approximately twenty percent of the specimens selected for the database may be
considered to be one-third scale.
The database includes column specimens with the following range of properties:
• shear span to depth ratio: 2.0⭐a/d⭐4.0 (mean⫽3.0)
• transverse reinforcement spacing to depth ratio: 0.2⭐s/d⭐1.2 (mean⫽0.61)
• concrete compressive strength: 1900⭐f c⬘⭐6500 psi (mean⫽3600 psi)
• longitudinal-reinforcement yield stress: 47⭐f yl⭐76 ksi (mean⫽59 ksi)
• longitudinal reinforcement ratio: 0.01⭐␳l⭐0.04 (mean⫽0.023)
• transverse-reinforcement yield stress: 46⭐f yt⭐94 ksi (mean⫽62 ksi)
• transverse reinforcement ratio: 0.0010⭐␳⬙⭐0.0065 (mean⫽0.003)
• maximum shear stress: 2.8⭐ v/冑 f ⬘c ,psi ⭐8.6 (mean⫽5.6)
• axial load ratio: 0.0⭐ P/Ag f ⬘c ⭐0.6 (mean⫽0.2)
where a is the shear span, d is the depth to the centerline of the outermost tension re-
inforcement, s is the spacing of the transverse reinforcement, v is the maximum nominal
shear stress in psi, P is the axial load, and Ag is the gross cross-sectional area of the
column.
DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 73

Table 1. Database of shear-critical column tests* (double curvature specimens)

b h d a s ␳l ␳⬙ f yl f yt f ⬘c P ⌬y ⌬s Vtest
Specimen in. in. in. in. in. ksi ksi ksi kips in. in. kips
Sezen (2002)
2CLD12 18 18 15.5 58 12
0.025 0.0017 64 68 3.06 150 1.04 2.97 70.8
2CHD12 18 18 15.5 58 12
0.025 0.0017 64 68 3.06 600 0.57 1.02 80.7
2CVD12** 18 18 15.5 58 12
0.025 0.0017 64 68 3.03 500 0.76 2.23 67.6
2CLD12M 18 18 15.5 58 12
0.025 0.0017 64 68 3.16 150 1.11 3.33 66.2
Lynn (2001)
3CLH18 18 18 15 58 18 0.03 0.001 48 58 3.71 113 0.78 1.2 61.0
3SLH18 18 18 15 58 18 0.03 0.001 48 58 3.71 113 0.61 1.15 60.0
2CLH18 18 18 15 58 18 0.02 0.001 48 58 4.8 113 0.72 3 54.0
2SLH18 18 18 15 58 18 0.02 0.001 48 58 4.8 113 0.63 2.4 52.0
2CMH18 18 18 15 58 18 0.02 0.001 48 58 3.73 340 0.61 1.2 71.0
3CMH18 18 18 15 58 18 0.03 0.001 48 58 4.01 340 0.61 1.2 76.0
3CMD12 18 18 15 58 12 0.03 0.0017 48 58 4.01 340 0.74 1.8 80.0
3SMD12 18 18 15 58 12 0.03 0.0017 48 58 3.73 340 0.86 1.8 85.0
Ohue, Morimoto, Fujii, and Morita (1985)
2D16RS 7.87 7.87 6.89 15.7 1.97 0.02 0.0057 54 46 4.65 41.1 0.3 1.08 22.9
4D13RS 7.87 7.87 6.89 15.7 1.97 0.027 0.0057 54 46 4.34 41.1 0.26 0.58 24.9
Esaki (1996)
H-2-1/5 7.87 7.87 6.89 15.7 1.97 0.025 0.0052 52 53 3.34 36.2 0.16 0.79 23.2
HT-2-1/5 7.87 7.87 6.89 15.7 2.95 0.025 0.0052 52 53 2.93 31.8 0.19 0.82 22.9
H-2-1/3 7.87 7.87 6.89 15.7 1.57 0.025 0.0065 52 53 3.34 60.4 0.14 0.63 27.1
HT-2-1/3 7.87 7.87 6.89 15.7 2.36 0.025 0.0065 52 53 2.93 53 0.19 0.79 25.1
* Notation: b⫽column section width; h⫽column section depth; d⫽depth to centerline of tension reinforcement;
a⫽shear span; s⫽tie spacing; ␳l⫽longitudinal reinforcement ratio (Asl /bh); Asl⫽total area of the longitudinal
reinforcement; ␳⬙⫽transverse reinforcement ratio (Ast /bs); Ast⫽area of transverse reinforcement;
f yl⫽longitudinal steel yield strength; f yt⫽transverse steel yield strength; f c⬘⫽concrete compressive strength;
P⫽axial load (at time of shear failure for variable axial load test); ⌬y⫽yield displacement;
⌬s⫽displacement at shear failure (at 20% loss in peak shear); Vtest⫽peak shear recorded (see Figure 1 for
definition of ⌬y , ⌬s , and Vtest).
** Variable axial load test. All data given for compression cycles (that is, direction in which shear failure was
initiated).

The specimen properties and selected response quantities are presented in Table 1 for
columns tested in double curvature and Table 2 for columns tested in single curvature.
Note that all displacements are given for an equivalent column in double curvature (that
is, for those specimens tested in single curvature in Table 2 the tabulated displacements
are twice those recorded during the experiment). The yield displacement, ⌬y , and the
displacement at shear failure, ⌬s , were determined based on the idealized backbone
curve from the test data as shown in Figure 1. For this purpose, the test backbone curve
was defined to envelop the force-deformation response for the first cycles to displace-
ment amplitudes exceeding previous cycles. Shear failure nominally was defined by the
displacement, ⌬s , where shear resistance dropped below 80% of the maximum shear re-
corded. It should be noted that this definition of the displacement at shear failure is
somewhat arbitrary, and hence, variability in the results is expected. It may be preferred
74 K. J. ELWOOD AND J. P. MOEHLE

Table 2. Database of shear-critical column tests (single curvature specimens)

b h d a s ␳l ␳⬙ f yl f yt f ⬘c P ⌬y ⌬s Vtest
Specimen in. in. in. in. in. ksi ksi ksi kips in. in. kips
Li, Park, and Tanaka (1991)
U-7 15.8 15.8 14.8 39.4 4.7 0.024 0.0047 64.7 55.4 4.21 104 0.7 2.8 73.7
U-8 15.8 15.8 14.8 39.4 4.7 0.024 0.0052 64.7 55.4 4.86 241 0.66 1.66 88.3
U-9 15.8 15.8 14.8 39.4 4.7 0.024 0.0057 64.7 55.4 4.95 368 0.6 2.4 96.6
Saatcioglu and Ozcebe (1989)
U1 13.8 13.8 12 39.4 5.9 0.033 0.003 62.4 68.2 6.32 0 1.34 4.18 61.8
U2 13.8 13.8 12 39.4 5.9 0.033 0.003 65.7 68.2 4.38 135 1.18 3.38 60.7
U3 13.8 13.8 12 39.4 3 0.033 0.006 62.4 68.2 5.05 135 1.26 3.54 60.3
Yalcin (1997)
BR-S1 21.7 21.7 19 58.5 11.8 0.02 0.001 64.5 61.6 6.5 469 0.64 1.82 130.0
Ikeda (1968)*
43 7.87 7.87 6.81 19.7 3.9 0.02 0.0028 63 81 2.84 18 0.26 1.04 16.6
44 7.87 7.87 6.81 19.7 3.9 0.02 0.0028 63 81 2.84 18 0.26 0.64 17.2
45 7.87 7.87 6.81 19.7 3.9 0.02 0.0028 63 81 2.84 35 0.38 0.64 18.5
46 7.87 7.87 6.81 19.7 3.9 0.02 0.0028 63 81 2.84 35 0.38 0.48 18.1
62 7.87 7.87 6.81 19.7 3.9 0.02 0.0028 50 69 2.84 18 0.24 1.46 13.0
63 7.87 7.87 6.81 19.7 3.9 0.02 0.0028 50 69 2.84 35 0.24 1.1 15.4
64 7.87 7.87 6.81 19.7 3.9 0.02 0.0028 50 69 2.84 35 0.28 1.32 15.4
Umemura and Endo (1970)*
205 7.87 7.87 7.09 23.6 3.9 0.02 0.0028 67 47 2.55 35 0.38 0.98 16.0
207 7.87 7.87 7.09 15.8 3.9 0.02 0.0028 67 47 2.55 35 0.32 0.5 23.8
214 7.87 7.87 7.09 23.6 7.9 0.02 0.0014 67 47 2.55 88 0.48 0.82 18.6
220 7.87 7.87 7.09 15.8 4.7 0.01 0.0011 55 94 4.77 35 0.12 0.94 17.6
231 7.87 7.87 7.09 15.8 3.9 0.01 0.0013 47 76 2.14 35 0.08 0.64 11.4
232 7.87 7.87 7.09 15.8 3.9 0.01 0.0013 47 76 1.9 35 0.1 0.64 13.1
233 7.87 7.87 7.09 15.8 3.9 0.01 0.0013 54 76 2.02 35 0.12 0.54 15.5
234 7.87 7.87 7.09 15.8 3.9 0.01 0.0013 54 76 1.9 35 0.12 0.64 15.1
Kokusho (1964)*
372 7.87 7.87 6.69 19.7 3.9 0.01 0.0031 76 51 2.88 35 0.2 0.84 16.7
373 7.87 7.87 6.69 19.7 3.9 0.02 0.0031 76 51 2.96 35 0.28 0.78 19.8
Kokusho and Fukuhara (1965)*
452 7.87 7.87 6.69 19.7 3.9 0.03 0.0031 52 88 3.18 88 0.24 0.6 24.8
454 7.87 7.87 6.69 19.7 3.9 0.04 0.0031 52 88 3.18 88 0.18 0.4 24.8
Wight and Sozen (1973)
40.033a 6 12 10.5 34.5 5 0.024 0.0033 72 50 5.03 42.5 0.6 2.5 22.3
40.033 6 12 10.5 34.5 5 0.024 0.0033 72 50 4.87 40 0.96 3.46 22.8
25.033 6 12 10.5 34.5 5 0.024 0.0033 72 50 4.88 25 0.94 2.48 23.5
0.033 6 12 10.5 34.5 5 0.024 0.0033 72 50 4.64 0 0.6 2.2 22.1
40.048 6 12 10.5 34.5 3.5 0.024 0.0048 72 50 3.78 40 1.14 3.82 21.2
0.048 6 12 10.5 34.5 3.5 0.024 0.0048 72 50 3.75 0 1.06 2.6 23.6
* Data given in Masaya (1973).
DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 75

Figure 1. Definition of displacements reported in database.

to define shear failure based on the magnitude of the shear deformations; however, such
data are not available for all columns in the database. The adopted definition of shear
failure has also been used by others, including Sezen (2002), Pujol et al. (1999), and
Kato and Ohnishi (2002). For approximately 20 percent of the specimens, the shear
strength did not drop below the 80% criterion. In such cases, ⌬s was taken as the maxi-
mum displacement reported, and, as such, represents a lower bound to the true displace-
ment at shear failure. Response histories and specimen details are provided in Sezen
(2002).
Although all of the specimens in Tables 1 and 2 were intended to be representative of
building columns, the general observations noted below may also apply to bridge col-
umns with similar cross sections and reinforcement details. Bridge columns, however,
often have larger cross sections, requiring more longitudinal reinforcement. Further-
more, the database does not include circular cross sections commonly used for bridge
columns. It is not known how these differences will impact the shear degradation
mechanisms.
Figure 2 compares the drift ratio at shear failure (that is, the displacement at shear
failure, ⌬s , divided by the clear height of the column, L⫽2a) with several key param-
eters. It is apparent from the plots that there is considerable variability in the results and
no single parameter that appears directly correlated with the drift at shear failure. The
data in Figure 2 suggest that the maximum nominal shear stress (in psi) recorded during
the tests (v⫽Vtest /(bd)), expressed as a fraction of 冑 f ⬘c (psi), is not strongly correlated
with the drift at shear failure. The median drift ratio at shear failure for columns with
high axial loads tends to be less than the median for columns with low axial loads. Fur-
thermore, drift at shear failure appears to increase with increasing transverse reinforce-
ment ratios, ␳⬙, though the relationship is barely discernible. The correlation is not im-
proved by comparing drift ratio and the transverse reinforcement index ␳⬙ f yt /f c⬘ . There is
an apparent inverse relationship between the spacing of the hoops (s/d) and the drift at
shear failure. The aspect ratio, a/d, has no clear relationship with the drift ratio at shear
failure.
76 K. J. ELWOOD AND J. P. MOEHLE

Figure 2. Effect of key parameters on drift ratio at shear failure.

DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 77

Figure 3. Variation of normalized shear stress with drift ratio at shear failure. (normalized
shear stress⫽Vtest /bd/(2冑 f c⬘⫹␳⬙ f yt) in psi units)

When the maximum shear stress is normalized by a nominal shear strength

2冑 f ⬘c ⫹␳⬙ f yt , in psi units (ACI 2002), as shown in Figure 3, a slight degradation of the
shear strength with increasing drift ratio can be observed. Observations like this have led
to the degrading shear-strength models proposed in the literature (Watanabe and Ichi-
nose 1992, Aschheim and Moehle 1992, Priestley et al. 1994, Sezen 2002). Given the
scatter in the data, and the large variation in the drift at shear failure for a modest change
in normalized shear stress, it seems unlikely that this relation would be useful for deter-
mining the displacement capacity for a given shear demand. The difficulties of using
shear strength models to predict the drift at shear failure will be discussed further in the
next section.

MODELS FOR DRIFT RATIO AT SHEAR FAILURE

Most models for estimating the drift capacity of reinforced concrete columns are
based on the performance of columns with seismic detailing that promotes ductile be-
havior. Such models assume the response is dominated by flexural deformations and use
estimates of the ultimate concrete and steel strains to determine the ultimate curvatures
the section can withstand. These models may not be applicable to older reinforced con-
crete columns with light transverse reinforcement because the degradation of shear-
resisting mechanisms begins before the flexural deformation capacity can be achieved.
Hence, models based on flexural mechanics will not be considered here. Instead, empiri-
cal models emphasizing shear failure will be pursued.

SHEAR STRENGTH MODEL (SEZEN 2002)

It is well established (e.g., Wight and Sozen 1973, Watanabe and Ichinose 1992,
Aschheim and Moehle 1992, Priestley et al. 1994) that shear strength decreases with in-
creasing deformation demand under applied deformation reversals. Sezen (2002) pre-
sented a model in which shear strength of columns with light transverse reinforcement
78 K. J. ELWOOD AND J. P. MOEHLE

Figure 4. Use of Sezen model to estimate (a) shear capacity and (b) displacement ductility ca-
pacity.

degrades with increasing displacement ductility demand. As indicated in Figure 4a, the
model is intended to estimate the shear strength given knowledge of the displacement
ductility demand. It seems a natural extension to use the Sezen shear-strength model (or
similar models) to estimate the displacement capacity associated with a given shear de-
mand, as suggested in Figure 4b. This latter practice has become widespread in the seis-
mic assessment of existing buildings and bridges, and will be examined in the following
paragraphs.
The Sezen shear-strength model divides the shear strength, Vn , into two components:
the shear nominally carried by concrete, Vc , and the shear nominally carried by rein-
forcement through a 45⬚ truss model, Vs , as follows.

Vn⫽k共Vc⫹Vs兲⫽k 冉 冑 冊冑
6 f c⬘
a/d
1⫹
P
6冑 f ⬘c Ag
0.8Ag⫹k
Ast f ytd
s
(1)

The coefficient k defines the degradation of shear strength with increasing displacement
ductility as shown in Figure 5. The degradation coefficient is applied to both Vc and Vs
under the assumption that the concrete component degrades due to increased cracking
and degradation of the aggregate interlock mechanism, while the steel component like-
wise degrades due to hoop opening and bond degradation.
Results of the Sezen shear-strength model can be compared with the test results sum-
marized in Tables 1 and 2. The displacement ductility at shear failure was defined as
⌬s /⌬y , where the quantities are defined in Figure 1 and tabulated in Tables 1 and 2. A

Figure 5. Variation of degradation coefficient k with displacement ductility.

DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 79

Figure 6. Comparison of calculated and measured drifts for the Sezen (2002) shear strength
model.

strength ratio, defined as the ratio of the measured strength to the calculated strength
(Equation 1), was calculated for each column. The mean strength ratio is 1.06 with a
coefficient of variation of 0.15.
Applicability of the Sezen shear-strength model for defining the displacement at
shear failure likewise can be studied using the data in Tables 1 and 2. For this purpose,
the shear demand was defined as the maximum shear measured during a test, and dis-
placement ductility was then calculated using Equation 1 (this approach is illustrated in
Figure 4b). Figure 6 compares the results with the measured drift ratios at failure. (Data
for two specimens are not shown because the shear demand passed below the shear fail-
ure surface, indicating that the shear strength model would not predict shear failure for
those specimens.) The results show relatively high bias and dispersion (mean of the mea-
sured drift ratio at shear failure divided by the calculated drift ratio is 1.78 and the co-
efficient of variation is 0.63).
Sezen (2002) recognized the aforementioned shortcoming in estimating the drift ra-
tio at shear failure, noting that a small variation in shear strength or shear demand cor-
responds to a large change in the estimated drift ratio at shear failure. If the variability in
the shear strength is assumed to be adequately represented by a normal distribution with
a mean, ␮, given by Equation 1 and a standard deviation, ␴, and the response of the
column is assumed to be elastic-perfectly-plastic prior to shear failure, then for the case
illustrated in Figure 7, the variation in the displacement ductility demand required to
cause shear failure for columns within ⫾␴ of the mean shear strength (i.e., ⌬␮␦ in Fig-
ure 7) is (80/3)␴. For example, if ␴⫽0.16 (determined using the shear-critical column
database discussed above), then ⌬␮␦ is 4.2.
A further shortcoming of using the shear strength model to estimate the drift ratio at
shear failure can be observed by considering the influence of axial load. According to
Equation 1, the nominal shear strength of a column is expected to increase with an in-
crease in axial load. Hence, if the shear strength model is used to estimate the drift ratio
80 K. J. ELWOOD AND J. P. MOEHLE

Figure 7. Displacement at shear failure as a function of model variability. (Vo⫽Vn from Equa-
tion 1 with k⫽1)

at shear failure, as shown in Figure 8, an increase in axial load will result in an increase
in the predicted drift at shear failure. In contrast, the experimental database (Figure 2)
suggests that an increase in the axial load may reduce the drift ratio at shear failure.

PUJOL et al. (1999)

Pujol et al. (1999) used a database of 92 columns to establish a conservative estimate
of the maximum drift ratio at shear failure (defined the same as ⌬s in Figure 1). The
database included columns with both circular and rectangular cross sections with the fol-
lowing ranges of experimental parameters:
• shear span to depth ratio: 1.3⭐a/d⭐5.0
• concrete compressive strength: 3000⭐f c⬘⭐12500 psi
• longitudinal reinforcement ratio: 0.005⭐␳l⭐0.051
• transverse reinforcement ratio: 0.0⭐␳⬙⭐0.0164
• maximum shear stress: 2.0⭐ v/冑 f c⬘ (psi units)
• axial load ratio: 0.0⭐ P/Ag f ⬘c ⭐0.2

Figure 8. Change in drift ratio at shear failure due change in axial load according to the shear
strength model.
DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 81

Figure 9. The Pujol et al. (1999) model compared with the shear-critical column database.

The most significant differences with the shear-critical column database introduced
in this paper include the consideration of columns with transverse reinforcement ratios
greater than 0.0065, the relatively low limit placed on the axial load ratio, and the in-
clusion of circular-cross-section columns representative of bridge columns.
Pujol et al. (1999) observed that the ratio of the maximum drift ratio to the column
aspect ratio, a/d, tended to increase with an increase in the reinforcement index, ␳⬙ f yt /v.
Based on statistical evaluation of the database results, and in an effort to establish a con-
servative estimate of the maximum drift ratio, Pujol et al. recommended the following
relationship:

⌬s ␳⬙ f yt a a
100 ⫽ ⭐ (2)
L v d d
where the calculated maximum drift ratio, ⌬s /L, is limited to less than or equal to 4%.
As shown in Figures 9 and 10, the maximum drift ratio model (Equation 2) is not
conservative for six of the columns in the shear-critical column database. Three of those
columns were subjected to axial loads in excess of the axial loads considered in the de-
velopment of the model.
The mean of the measured drift ratio at shear failure divided by the drift ratio cal-
culated according to Equation 2 is 1.71; the coefficient of variation is 0.42. The disper-
sion in results shows marked improvement relative to the use of the Sezen shear-strength
model, and the bias is consistent with the stated intent to provide a relatively conserva-
tive result.

KATO AND OHNISHI (2002)

Kato and Ohnishi proposed that the plastic drift capacity can be estimated based on
the maximum edge strain in the core concrete, the axial load ratio, and the cross section
82 K. J. ELWOOD AND J. P. MOEHLE

Figure 10. Comparison of calculated and measured drifts for the Pujol et al. (1999) model.

dimensions. The total drift ratio is given by the sum of the drift ratio at yielding of the
longitudinal reinforcement, ⌬y /L, and the calculated plastic drift ratio, ⌬p /L:

⌬s ⌬ y ⌬ p
⫽ ⫹ (3)
L L L

where
⌬p
L
⫽
再 D
D冉 冊冉 冒 冊 冉
m␧cp 2
je 3

冉 冊冉 冒 冉 冊冊 冉
m␧cp 2
je 3
e␩

5e␩⫺
4
3
0⬍e␩⬍

1
3

⭐e ⬍
3 ␩ 3
1

2
冊
冊
(4)

where D is the full depth of the gross cross section, je is the depth of the core, ␧cp is the
strain at maximum stress for the core concrete, m is the ratio of the concrete strain at the
edge of the core concrete to ␧cp , and e␩ is an equivalent axial load ratio (see Figure 11).

Figure 11. Equivalent axial load ratio (from Kato and Ohnishi 2002).
(␥⫽min. axial load/max. axial load⭓0; ␩p⫽max. axial load/Ag f ⬘c )
DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 83

Figure 12. Comparison of calculated and measured drifts for the Kato and Ohnishi model.
Shaded data points denote columns with high axial load (P/Ag f c⬘⬎0.25).

The equivalent axial load ratio is used to account for the influence of variable axial load
on the drift at shear failure. The coefficient m was selected to achieve a good agreement
between Equation 4 and measured drifts ratios from 36 pseudo-static column tests. For
the drift at shear failure (also defined the same as ⌬s in Figure 1), Kato and Ohnishi
(2002) recommend m⫽2.3.
Figure 12 compares the calculated drift ratios at shear failure based on Equation 3
with the measured drift ratios for the specimens in Tables 1 and 2. To avoid introducing
additional errors into the model, the displacement at yielding of the longitudinal rein-
forcement (⌬y in Equation 3) was taken as equal to ⌬y given in Tables 1 and 2. The depth
of the core, je , was estimated as the distance between the tension and compression re-
inforcement. The core concrete was assumed to be unconfined for each of the column
specimens; hence, ␧cp was set equal to 0.002. Since Equation 4 is undefined for e␩ equal
to zero, columns with zero axial load could not be evaluated using the model, and do not
appear in Figure 12. The mean of the measured drift ratio divided by the calculated drift
ratio is 0.84; the coefficient of variation is 0.44. Although the model provides a better
estimate of the measured drift at shear failure compared with the models by Pujol and
Sezen, Equation 3 relies on an accurate estimate of the drift at yielding of the longitu-
dinal reinforcement and significantly overestimates the drift at shear failure for many of
the columns with low axial loads.

PROPOSED EMPIRICAL DRIFT CAPACITY MODEL

The models for the drift ratio at shear failure presented in the previous sections do
not adequately model the behavior of the shear-critical columns considered in this study.
This section will introduce a drift capacity model based on observations from the ex-
perimental database. Due to its empirical nature, the model may not be applicable to
columns with parameters outside the ranges included in the database.
84 K. J. ELWOOD AND J. P. MOEHLE

Figure 13. Shear response of columns from database sorted by transverse reinforcement ratio.
Shaded data points denote columns with high axial load (P/Ag f c⬘⬎0.25).

Figure 13 plots the maximum nominal shear stress carried by a column versus the
drift ratio at failure. Taken as a whole, there is no apparent trend. If, however, the data
are differentiated according to transverse steel ratio and axial load ratio, some trends
emerge. Ranges of transverse reinforcement ratio and axial load ratio are shown by dif-
ferent symbols in Figure 13. For a given level of transverse reinforcement, the drift ratio
at failure increases with decreasing nominal shear stress. Furthermore, as shown in Fig-
ure 13, and the upper right plot of Figure 2, drift ratio at shear failure tends to be lower
for columns with high axial load ratio. These observations suggest that the maximum
shear stress, the transverse reinforcement ratio, and the axial load ratio should be in-
cluded in the proposed drift capacity model (an alternative model without the axial load
ratio is discussed in Elwood and Moehle [2003]). Based on a least-squares fit to the data,
the following empirical expression is proposed to estimate the drift ratio at shear failure:

⌬s 3 1 v 1 P 1
⫽ ⫹4␳⬙⫺ ⫺ ⭓ 共psi units兲 (5)
L 100 500 冑 f c⬘ 40 Ag f c⬘ 100
Figure 14 compares Equation 5 with the results from the database. The mean of the
measured drift ratio divided by the calculated drift ratio is 0.97, the coefficient of varia-
tion is 0.34. Note that the calculated drift ratio was determined using the maximum mea-
sured shear stress.
Figure 15 compares the accuracy of the shear strength model (Equation 1) and drift
ratio capacity model (Equation 5) for the column database. The plots illustrate that the
shear strength can be estimated more accurately than the drift ratio at shear failure. In-
terestingly, the even distribution of data about the datum (1.0, 1.0) indicates that under-
estimation or overestimation of the column strength is no indicator of whether the dis-
placement capacity will be underestimated or overestimated.
Figure 16 demonstrates how Equation 5 can be used for the seismic assessment of
existing reinforced concrete columns. Using a column similar to specimen 2CLD12
from the database as an example, the figure shows the idealized elastic, perfectly plastic
DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 85

Figure 14. Comparison of calculated drift ratio at shear failure using Equation 5 with results
from the shear-critical column database. (Dashed lines are +/− one standard deviation from the
mean.)

flexural response of the column, where the plastic shear capacity, Vp , is calculated as-
suming the column is fixed against rotation at both ends. The figure also includes lines
denoting the Sezen shear-strength model (Equation 1) and the proposed drift capacity
model (Equation 5). Comparing the Sezen shear-strength model with the plastic shear, it
is apparent that the shear demands are in a range where shear distress should be ex-
pected. With this knowledge, the expected drift ratio at shear failure (Equation 5) is cal-
culated to be 0.024, using v⫽Vp /(bd). The measured drift at shear failure for specimen
2CLD12 was 0.026.

Figure 15. Measured shear strength versus the drift at shear failure normalized by the calcu-
lated shear strength (Vn from Equation 1) and the calculated drift at shear failure (⌬s /L from
Equation 5), respectively.
86 K. J. ELWOOD AND J. P. MOEHLE

Figure 16. Comparison of the Sezen shear-strength model (Equation 1) and the proposed drift
capacity model (Equation 5) for specimen 2CLD12.

An important point of the preceding paragraph is that the drift capacity model should
be used only for columns that are expected to sustain shear failure following flexural
yielding. Therefore, as shown in Figure 16, the shear corresponding to inelastic flexural
response should fall in the range between the initial shear strength, Vo , and the final
shear strength, 0.7Vo . Because there is some dispersion between actual and calculated
shear strength, some columns with shear demand less than the calculated shear may still
experience shear failure after flexural yielding. Engineering judgment may be required
to define a lower-bound shear below which shear failure will be deemed acceptably im-
probable.
Figure 17 illustrates the influence of axial load and transverse reinforcement on the

Figure 17. Influence of (a) axial load and (b) transverse reinforcement ratio on drift capacity at
shear failure for specimen 2CLD12.
DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 87

drift capacity at shear failure according to the proposed model. The base curve shown in
both Figure 17a and 17b is for specimen 2CLD12 from the database with ␳⬙⫽0.0017
and P⫽0.15Ag f ⬘c . Each of the other curves is for the same column with only the noted
change in axial load or transverse reinforcement ratio. Note that shear strength (accord-
ing to Equation 1) varies with axial load and transverse reinforcement, and, hence, the
limits within which Equation 5 should be used are seen to shift with changes in axial
load and transverse reinforcement. For columns yielding prior to shear failure, the pro-
posed drift capacity model indicates that the drift at shear failure is expected to increase
with decreasing axial load and increasing transverse reinforcement ratio.

CONCLUSIONS
This study has demonstrated that shear strength models, developed to provide an es-
timate of the strength of columns susceptible to shear failure during seismic response,
may not be appropriate for estimating the interstory drift at shear failure. Furthermore, it
has been shown that other drift capacity models based on data from well confined col-
umns do not provide an adequate estimate of the mean drift at shear failure for shear-
critical columns with low transverse reinforcement ratios. An empirical model (Equation
5) is proposed to estimate the drift capacity of shear-critical columns with low transverse
reinforcement ratios typical of those occurring in older building construction. Applica-
tions and limitations of the model define the applicability of the proposed model for
seismic evaluation of existing building construction.

ACKNOWLEDGMENTS
This work was supported in part by the Pacific Earthquake Engineering Research
Center (PEER) through the Earthquake Engineering Research Centers Program of the
National Science Foundation under Award Number EEC-9701568. This support is grate-
fully acknowledged. The experimental studies were conducted in the research laborato-
ries of PEER at the University of California, Berkeley. The experimental database used
in this study was assembled by Professors Halil Sezen and Abraham Lynn. Their help is
gratefully acknowledged.

NOTATION
a shear span
Ag gross cross-sectional area of the column (bh)
Asl total area of the longitudinal reinforcement
Ast area of transverse reinforcement
b column section width
d depth to centerline of tension reinforcement
D full depth of the gross cross section
e␩ equivalent axial load ratio (see Figure 13)
f ⬘c concrete strength
88 K. J. ELWOOD AND J. P. MOEHLE

f yl longitudinal steel yield strength

f yt transverse steel yield strength
h column section height
je depth of the core
L clear height of the column
m ratio of the concrete strain at the edge of the core concrete to ␧cp
P axial load
s tie spacing
v maximum nominal shear stress (Vtest /bd)
Vc concrete contribution to the nominal shear strength
Vn nominal shear strength (Equation 1)
Vo nominal shear strength, Vn , from Equation 1 with k⫽1
Vp plastic shear capacity
Vs transverse reinforcement contribution to the nominal shear strength
Vtest peak shear recorded (see Figure 1)
⌬s displacement at shear failure (at 20% loss in peak shear, see Figure 1)
⌬y yield displacement (see Figure 1)
⌬␮␦ change in displacement ductility
␧cp strain at maximum stress for the core concrete
␥ ratio of minimum axial load to maximum axial load⭓0
␩p maximum axial load ratio, Pmax /Ag f c⬘ (see Figure 13)
␮ mean shear strength
␮␦ displacement ductility
␳l longitudinal reinforcement ratio (Asl /bh)
␳⬙ transverse reinforcement ratio (Ast /bs)
␴ standard deviation for shear strength

REFERENCES
American Concrete Institute (ACI) Committee 318, 2002. Building Code Requirements for
Structural Concrete (318-02) and Commentary (318R-02), American Concrete Institute,
Farmington Hills, MI.
American Society of Civil Engineers (ASCE), 2000. Prestandard and Commentary for the Seis-
mic Rehabilitation of Buildings, FEMA-356, Federal Emergency Management Agency,
Washington D.C., November.
Applied Technology Council (ATC), 1996. Seismic Evaluation and Retrofit of Concrete Build-
DRIFT CAPACITY OF REINFORCED CONCRETE COLUMNS WITH LIGHT TRANSVERSE REINFORCEMENT 89

ings, ATC-40, California Seismic Safety Commission (SSC 96-01), Sacramento, CA.
Aschheim, M., and Moehle, J. P., 1992. Shear Strength and Deformability of RC Bridge Col-
umns Subjected To Inelastic Displacements, UCB/EERC 92/04, University of California,
Berkeley.
Elwood, K., and Moehle, J. P., 2003. Shake Table Tests and Analytical Studies on the Gravity
Load Collapse of Reinforced Concrete Frames, PEER Report 2003/01, Pacific Earthquake
Engineering Research Center, University of California.
Esaki, F., 1996. Reinforcing effect of steel plate hoops on ductility of R/C square columns, Pro-
ceedings of 11th World Conference on Earthquake Engineering, Pergamon, Elsevier Science
Ltd., Oxford, England, Disc 3, Paper No. 196.
Kato, D., and Ohnishi, K., 2002. Axial load carrying capacity of R/C columns under lateral load
reversals, Third U.S.-Japan Workshop on Performance-Based Earthquake Engineering Meth-
odology for Reinforced Concrete Building Structures, Seattle, WA, PEER Report 2002/02,
Pacific Earthquake Engineering Research Center, University of California, Berkeley, CA, pp.
247–255.
Li, X., Park, R., and Tanaka, H., 1991. Effects of variations in axial load level on the strength
and ductility of reinforced concrete columns, Proceedings of Pacific Conference on Earth-
quake Engineering, New Zealand, Vol. 1, pp. 147–158.
Lynn, A. C., 2001. Seismic Evaluation of Existing Reinforced Concrete Building Columns,
Ph.D. dissertation, Department of Civil and Environmental Engineering, University of Cali-
fornia, Berkeley.
Masaya, H., (ed.), 1973. A List of Past Experimental Results of Reinforced Concrete Columns,
Building Research Institute, Ministry of Construction, Japan.
Ohue, M., Morimoto, H., Fujii, S., and Morita, S., 1985. The behavior of RC short columns
failing in splitting bond-shear under dynamic lateral loading, Transactions of the Japan Con-
crete Institute, Vol. 7, pp. 293–300.
Priestley, M. J. N., Verma, R., and Xiao, Y., 1994. Seismic shear strength of reinforced concrete
columns, J. Struct. Eng. 120 (8), 2310–2329.
Pujol, S., Ramirez, J. A., and Sozen, M. A., 1999. Drift capacity of reinforced concrete columns
subjected to cyclic shear reversals, Seismic Response of Concrete Bridges, SP-187, Ameri-
can Concrete Institute, Farmington Hills, MI, pp. 255–274.
Saatcioglu, M., and Ozcebe, G., 1989. Response of reinforced concrete columns to simulated
seismic load, ACI Struct. J. 86 (1), 3–12.
Sezen, H., 2002. Seismic Response and Modeling of Lightly Reinforced Concrete Building
Columns, Ph.D. dissertation, Department of Civil and Environmental Engineering, Univer-
sity of California, Berkeley.
Watanabe, F., and Ichinose, T., 1992. Strength and ductility of RC members subjected to com-
bined bending and shear, Concrete Shear in Earthquake, Elsevier Applied Science, New
York, pp. 429–438.
Wight, J. K., and Sozen, M. A., 1973. Shear strength decay in reinforced concrete columns sub-
jected to large deflection reversals, Structural Research Series No. 403, University of Illi-
nois, Urbana.
Yalcin, C., 1997. Seismic Evaluation and Retrofit of Existing Reinforced Concrete Bridge Col-
umns, Ph.D. dissertation, Department of Civil Engineering, University of Ottawa.
(Received 6 November 2003; accepted 8 June 2004)