Professional Documents
Culture Documents
ACI SP 297 Nonlinear Modeling Parameters and Acceptance Criteria
ACI SP 297 Nonlinear Modeling Parameters and Acceptance Criteria
ACI SP 297 Nonlinear Modeling Parameters and Acceptance Criteria
Abstract
A database of 490 pseudo-static tests of reinforced concrete columns subjected to load reversals
was used to evaluate nonlinear modeling parameters that define the lateral force versus lateral
deformation envelope relation of columns under seismic excitations. Based on the modeling
parameters, criteria that identify acceptable deformation levels at various performance objectives
are proposed. The effects of bi-directional loading and number-of-cycles of the displacement
history on the drift ratio at axial failure are discussed, and recommendations are given to account
for such effects. Modeling parameters and acceptance criteria are provided in a format that is
consistent with provisions of the ASCE 41-06 Standard entitled “Seismic Rehabilitation of
Existing Structures”.
1. Introduction
When estimating the performance of existing structures subjected to seismic events, it is
often necessary to conduct nonlinear dynamic analyses that require the definition of the lateral
force versus lateral deformation relation of frame members. Based on a database comprised of
490 reinforced concrete column tests [1, 2], relations are proposed to define key limiting
deformations in the lateral force-deformation relation of reinforced concrete columns subjected
to seismic demands. More specifically, relations for evaluating plastic rotations at incipient
lateral-strength degradation and incipient axial degradation are proposed. Furthermore,
recommendations are provided for selecting acceptable deformations or Acceptance Criteria
(AC) below which the performance of reinforced concrete columns is deemed acceptable for
selected target performance objectives.
The proposed relations and recommendations are given in a format that is compatible with
the standard for seismic rehabilitation ASCE 41-06 Supplement 1 [3, 4]; hereafter referred to as
ASCE 41. The standard defines the nonlinear force-deformation backbone relations of concrete
columns and other members as illustrated in Fig.1a. In the Figure, the plastic rotation at
incipient lateral-strength degradation is given through the Modeling Parameter (MP) a. The
plastic rotation in a concrete column at incipient axial degradation is given through the MP b.
The residual lateral strength of a column is given by the MP c. MP a and b are given in the
current standard as conservative lower-bound estimates of experimental values (Fig. 1b) [5]. To
avoid skewing the results of nonlinear simulations, the proposed MP a and b will target median
experimental values. AC in the standard were based on fixed percentages of the MP for various
performance objectives. Because the error on the estimates of MP exhibit different dispersions
for various parameters and members, selecting a fixed fraction of those MP values for AC
results in varying probabilities of exceedance for the AC (e.g., 75% of a MP does not always
provide the same probabilities of exceedance because that probability depends on the standard
1.1
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
deviation on the MP estimate) . Thus, proposed AC are defined through fixed probabilities of
exceedance for various performance objectives and the corresponding factions of MP values
that achieve those probabilities are given.
Rotation
a) b)
Figure 1: a) ASCE 41-06 backbone response for nonlinear modeling of RC components (from [3]);
b) illustration of MP deviation from median values and resulting backbone for modeling RC elements.
2. Column Database
The database used to support the development of the proposed MP and AC contains 319
rectangular column tests and 171 circular column tests for a total of 490 tests [1, 2]. Much of
the data was derived from the PEER column database [6]. All tests in the database were
conducted quasi-statically. The new database is webcast and accessible to the public ([1, 2])
and additional information about the database can be found in Sivaramakrishnan [7]. The
distributions of key parameters of the columns in the data sets are illustrated in the bar charts of
Fig. 2. Out of the rectangular columns in the database, 37 can be considered to satisfy the
requirements of ACI 318-11 [8] for Special Moment Resisting Frames (SMRF). Out of the
circular columns, 24 can be considered to satisfy these requirements. A limited number of
rectangular columns (25 out of 171) have reported ties with 90o hooks. The tie hook-angle is
unknown for 26 rectangular columns, while 269 rectangular columns have ties with 135o hooks
or welded ends. A limited number of circular columns had ties with lapped ends (13 out of
171). Tie details were unknown for one circular column, while the remainder of the circular
columns had spirals, ties with welded ends, or hooks anchored into the core. Limited
information on hook extensions was available for column tests of the database. Splices and
anchorage deficiencies are not within the scope of the work presented, and as such no tests on
columns containing anchorage deficiencies were used.
1.2
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
90 140
Rectangular Rectangular
80 Circular Circular
120
70
100
Number of Columns
Number of Columns
60
50 80
40 60
30
40
20
20
10
0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.25 0.5 0.75 1 1.25 1.5 1.75
Axial Load Ratio s/d
90
Rectangular
Circular
80
90
Rectangular
70 80 Circular
60 70
Number of Columns
Number of Columns
60
50
50
40
40
30
30
20
20
10 10
0 0
0 0.0025 0.005 0.0075 0.01 0.0125 0.015 0.0175 0.02 0.0225 0.025 0.0275 0.03 0.0325 0.035 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Transverse Reinforcement Ratio Vy/Vo
Figure 2: Distribution of key parameters in column database – see Notation section for term definitions
3. Data Extraction
The values of a and b were extracted for all column tests. The a values were taken as: a =
(Δ0.8-Δy)/L with Δ0.8 = lateral drift at which the lateral strength of an element degrades by 20%
from peak, Δy = lateral drift at onset of significant inelastic deformations, and L = column clear
length. The drift at onset of inelastic deformation, Δy, was obtained as recommended by Sezen
and Moehle [9]. A secant line was extended on the lateral force-deformation plot of a column
test from the origin to the point on the backbone curve at 70% of the maximum shear (0.7 Vmax).
Δy was then taken as the drift at the intersection of the secant line with the horizontal line drawn
at Vmax. Δy was also evaluated using a secant line passing through 0.6 Vmax. Little difference was
observed between results from the two intercepts. Results based on a 0.7 Vmax intercept were
used in parameter extractions.
The drifts at yield obtained as described above differ from estimates that would be obtained
using column stiffness relations provided in the ASCE 41 standard. ASCE 41 stiffness relations
were not used for extracting paraerters a and b to avoid skewing the extracted plastic rotations by
1.3
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
errors in stiffness estimates inherent in the ASCE 41 stiffness relations [10] (especially for
columns with low deformation capacities). The objective of the study was to extract the best
estimate of plastic rotations a and b such that the proposed relations defining them would
provide median estimates of the plastic rotations and the trends described in those relations
would not be artificially altered. The outcome for users of the standard that opt to use ASCE 41
stiffness relations will be a slightly skewed total deformation capacity for elements, but the
estimated behavior of the elements would not be skewed (i.e., the degree of inelastic
deformations prior to loss of strength estimated by the relations would be the median estimate).
Moreover, the proposed relations are intended to be used not only with the ASCE 41 elastic
stiffness relations but also with other methods for evaluating stiffness (such as fiber-section
models).
Due to the scarcity of column tests conducted to collapse, two sets of b values were produced.
The first set originates from column tests that were conducted to axial collapse. If a test was
conducted to axial collapse, the plastic rotation at axial failure was taken as b1 = (Δaxial-Δy)/L;
with Δaxial = drift at onset of axial collapse. Only 36 rectangular and 9 circular columns in the
database were pushed to axial failure. The webcast database was further bolstered by 12 recent
rectangular-column collapse tests [11, 12]. The second set of b values, b2, includes the first set
but also introduces b values for all other columns in the database based on the following:
- If a test reached a drift at which the lateral strength degraded to 25% of the peak but no
axial failure was reported, the plastic rotation at that drift was taken as b2. This limit on
deformations at axial failure was introduced to mitigate possible errors or omissions in
the reporting of axial failure in database tests. Columns that lost 75% of their lateral
strength were deemed unstable and close to axial collapse.
- If a columns was not tested to collapse or to a deformation causing a reduction in lateral
strength to below 25% of the peak lateral strength, the plastic rotation at the largest drift
a column was pushed was taken as b2.
This methodology provides a lower bound estimate on the parameter b but utilizes all the
available column tests. Additional details about the process by which the plastic rotations were
extracted from the database can be found in [7].
Values of Vy were extracted from experimental data through analytical means. In all
calculations, measured material properties were used. Vy is defined as the shear demand
corresponding to the development of moment strength in a column. Vy = My/La with My = column
moment capacity, and La = column shear span. My was calculated using fiber-section analyses
that utilized the parabolic stress-strain relation for concrete proposed by Hognestad [13] for
concrete in compression. The limiting strain in compression for the concrete was adopted as
0.003. Confinement effects on concrete material properties, as well as the tensile strength of
concrete were neglected. An elastic perfectly plastic material model was used for steel fibers and
capped at the measured yield stress of longitudinal bars.
ACI 318-11 specifies the use of the ultimate stress of longitudinal bars (approximated as 1.25
the specified yield stress) when evaluating the flexural strength of plastic hinges in flexural
members. The increment of 1.25 on yield stress is an arbitrary limit to ensure that an upper
bound estimate of the shear demand is used for proportioning the transverse reinforcement. In
the study that was performed, measured values of yield stress were used in calculating Vy for
following reasons:
1.4
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
1) In reinforced concrete columns, longitudinal bars will only reach ultimate stress in well
confined columns that are pushed to large inelastic deformations. For columns that
sustain shear failure prior to flexural yielding or at relatively low inelastic deformations,
evaluating Vy using the yield stress of bars is more appropriate.
2) Deriving the modeling parameter and acceptance criteria relations using lower Vy values
obtained using the yield stress of bars would provide lower modeling parameters and
acceptance criteria should users opt to evaluate Vy using 1.25 the yield stress of
longitudinal bars.
The beneficial effects of confinement on flexural strength were not included in calculations of
Vy for similar reasons.
Vo was evaluated based on the shear-strength equation of ASCE 41-06 for reinforced concrete
columns. The shear strength equation used was:
Φ ⁄
1 0.8 Eq. 1
where α = 6 in psi units and 0.5 in MPa units; Nu = axial compression force (= 0 for tension
force); M/Vd is the ratio of moment to shear times the effective depth and was bounded by the
values of 4 and 2; d is the effective depth; and Ag is the gross cross-sectional area of the column.
For circular columns d = 0.8D, where D is the diameter of the column. To account for the
inefficacy of transverse reinforcement in resisting shear when spaced beyond 75% of the
effective depth of a section, the steel contribution to shear strength (first term in Eq. 1) was
modified by a factor of Φ that was taken as 1.0 for s/d ≤ 0.75, zero for s/d ≥ 1.0, and linearly
interpolated between the two values of s/d.
1.5
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
Regression analyses indicated that spirally reinforced circular columns responded differently
to variations in influential parameters than rectangular columns reinforced with rectangular ties.
To account for such differences, separate regression equations were developed for spirally
reinforced circular columns than for all other columns. The most influential parameters for
rectangular and circular columns were: the axial load ratio, the transverse reinforcement ratio
(ρt), and the ratio Vy/Vo. All three parameters are currently used in the standard ASCE 41 to
estimate plastic rotation limits for concrete columns. The maximum shear stress applied to a
column section is also used in the standard as a parameter to estimate MP. That parameter was
not found to have a determining role in column responses analyzed in this work.
The proposed relations between plastic rotations at incipient loss of lateral strength and axial
failure are given below. The subscripts R and C in the equations indicate plastic rotations for
rectangular and circular columns, respectively. Only relations for b2 were derived using linear
regression because there are too few b1 values to provide meaningful regression-based estimates.
Plastic rotation capacities for columns other than spirally reinforced circular columns
P Vy
a R 0.042 0.043 0.63 t 0.023 0.0 (rad) Eq.2
Ag f c' Vo
P Vy
b2 R 0.051 0.051 '
1.3 t 0.023 a R (rad) Eq.3
Ag f c Vo
Plastic rotation capacities for columns for spirally reinforced circular columns
P Vy
aC 0.06 0.058 '
1.3 t 0.037 0.0 (rad) Eq.4
Ag f c Vo
P Vy
b2C 0.064 0.07 2.85 t 0.03 a C (rad) Eq.5
Ag f c' Vo
When evaluating the above parameters, ρt should not be taken greater than 0.0175 and Vy/Vo
should not be taken smaller than 0.2. The equation is not applicable for ρt ≤0.0005. An upper
bound on the transverse reinforcement ratio of 0.0175 is prescribed because few columns in the
database contained a ratio exceeding that limit. Equations for modeling parameters cannot be
used for columns with a transverse reinforcement ratio below 0.0005 because the equations are
not intended for unreinforced columns. A lower limit on Vy/Vo of 0.2 is prescribed because few
columns in the database had lower values of Vy/Vo.
Values of a and b2 for rectangular and spirally reinforced column are presented in Table 1 at
practical parameter boundaries. In Table 1, the notation is consistent with that used in Eq. 2 to 5
in which the subscript r is used in reference to rectangular columns while the subscript c is used
in reference to circular columns.. As can be observed in the table, for a given set of parameters,
spirally reinforced columns exhibited significantly larger deformation capacities than rectangular
columns.
1.6
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
0.9 0.9
Equation estimates
0.8 0.8
ASCE 41-06 estimates
Equation estimates
0.7 0.7
Cumulative Distribution
Cumulative Distribution
0.5 0.5
0.4 0.4
0.3 0.3
Number of Tests = 319 Number of Tests = 171
0.2 0.2
Standard Deviation Equation = 0.015 Standard Deviation Equation = 0.019
0.1 0.1
Standard Deviation ASCE41-06 = 0.015 Standard Deviation ASCE41-06 = 0.022
0 0
-0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Error in aR = Experiment - Estimate (rad) Error in aC = Experiment - Estimate (rad)
Figure 3: Cumulative distribution of the error between experimental values and estimates for aR and aC
1.7
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
The cumulative distribution of the error difference between experimental and regression
estimates of aR are plotted for various bins of rectangular columns in Fig. 4. Test data are divided
into bins to illustrate the fit of the proposed relation in various parameter quadrants. As can be
observed in the figure, the proposed relation produced estimates that were very close to the
median in all bins. Standard deviations on the error of the proposed equation vary from 0.005 to
0.018; which indicates variable accuracy of the proposed equation across bins. Figure 5 presents
similar data as Fig. 4 but for circular columns. Similar conclusions can be drawn from Fig. 5 for
spirally reinforced circular columns, although the estimates from the proposed equation deviate
conservatively from the median estimate for columns with low transverse reinforcement ratio
and intermediate values of Vy/Vo (ranging between 0.6 and 1.0). The standard deviations were
slightly higher for circular columns than rectangular ones and ranged from 0.005 to 0.029 across
bins.
Axial Load Ratio 0.2, Transv. Reinf. Ratio 0.006, Vy/VO 0.6 Axial Load Ratio>0.2, Transv. Reinf. Ratio 0.006, Vy/VO 0.6
1 1
0.5 0.5
Number of Tests = 21 Number of Tests = 21
0 Standard Deviation = 0.015 0 Standard Deviation = 0.018
-0.04 -0.02 0 0.02 0.04 0.06 -0.04 -0.02 0 0.02 0.04 0.06
Axial Load Ratio 0.2, Transv. Reinf. Ratio 0.006, 0.6<Vy/VO 1.0 Axial Load Ratio>0.2, Transv. Reinf. Ratio 0.006, 0.6<Vy/VO 1.0
1 1
0.5 0.5
Number of Tests = 39 Number of Tests = 34
0 Standard Deviation = 0.016 0 Standard Deviation = 0.01
-0.04 -0.02 0 0.02 0.04 0.06 -0.04 -0.02 0 0.02 0.04 0.06
Axial Load Ratio 0.2, Transv. Reinf. Ratio 0.006, Vy/VO>1.0 Axial Load Ratio>0.2, Transv. Reinf. Ratio 0.006, Vy/VO>1.0
1 1
0.5 0.5
Number of Tests = 26 Number of Tests = 18
Standard Deviation = 0.0084 Standard Deviation = 0.005
Cumulative Distribution
0 0
-0.04 -0.02 0 0.02 0.04 0.06 -0.04 -0.02 0 0.02 0.04 0.06
Axial Load Ratio 0.2, Transv. Reinf. Ratio>0.006, Vy/VO 0.6 Axial Load Ratio>0.2, Transv. Reinf. Ratio>0.006, Vy/VO 0.6
1 1
0.5 0.5
Number of Tests = 47 Number of Tests = 95
0 Standard Deviation = 0.016 0 Standard Deviation = 0.016
-0.04 -0.02 0 0.02 0.04 0.06 -0.04 -0.02 0 0.02 0.04 0.06
Axial Load Ratio 0.2, Transv. Reinf. Ratio>0.006, 0.6<Vy/VO 1.0 Axial Load Ratio>0.2, Transv. Reinf. Ratio>0.006, 0.6<Vy/VO 1.0
1 1
0.5 0.5
Number of Tests = 11 Number of Tests = 6
0 Standard Deviation = 0.018 0 Standard Deviation = 0.016
-0.04 -0.02 0 0.02 0.04 0.06 -0.04 -0.02 0 0.02 0.04 0.06
Axial Load Ratio 0.2, Transv. Reinf. Ratio>0.006, Vy/VO>1.0 Axial Load Ratio>0.2, Transv. Reinf. Ratio>0.006, Vy/VO>1.0
1 1
Proposed Equation
0.5 ASCE 41-06 Sup 1 Table 0.5
Number of Tests = 1 Number of Tests = 0
0 Standard Deviation = 0 0
-0.04 -0.02 0 0.02 0.04 0.06 -0.04 -0.02 0 0.02 0.04 0.06
Error in aR = Experiment - Estimate (rad) Error in aR = Experiment - Estimate (rad)
Figure 4: Cumulative distributions of the error between experimental values and estimates for aR; data
split into bins covering various ranges of parameters and standard deviations given for equation estimates
only.
1.8
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
Axial Load Ratio 0.2, Transv. Reinf. Ratio 0.006, Vy/VO 0.6 Axial Load Ratio>0.2, Transv. Reinf. Ratio 0.006, Vy/VO 0.6
1 1
0.5 0.5
Number of Tests = 40 Number of Tests = 8
0 Standard Deviation = 0.027 0 Standard Deviation = 0.018
-0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Axial Load Ratio 0.2, Transv. Reinf. Ratio 0.006, 0.6<Vy/VO 1.0 Axial Load Ratio>0.2, Transv. Reinf. Ratio 0.006, 0.6<Vy/VO 1.0
1 1
0.5 0.5
Number of Tests = 27 Number of Tests = 5
0 Standard Deviation = 0.015 0 Standard Deviation = 0.0046
-0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Axial Load Ratio 0.2, Transv. Reinf. Ratio 0.006, Vy/VO>1.0 Axial Load Ratio>0.2, Transv. Reinf. Ratio 0.006, Vy/VO>1.0
1 1
0.5 0.5
Number of Tests = 46 Number of Tests = 8
0 Standard Deviation = 0.0099 0 Standard Deviation = 0.0094
Cumulative Distribution
-0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Axial Load Ratio 0.2, Transv. Reinf. Ratio>0.006, Vy/VO 0.6 Axial Load Ratio>0.2, Transv. Reinf. Ratio>0.006, Vy/VO 0.6
1 1
0.5 0.5
Number of Tests = 13 Number of Tests = 20
0 Standard Deviation = 0.029 0 Standard Deviation = 0.018
-0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Axial Load Ratio 0.2, Transv. Reinf. Ratio>0.006, 0.6<Vy/VO 1.0 Axial Load Ratio>0.2, Transv. Reinf. Ratio>0.006, 0.6<Vy/VO 1.0
1 1
Proposed Equation
0.5 0.5 ASCE 41-06 Sup 1 Table
Number of Tests = 3 Number of Tests = 1
0 Standard Deviation = 0.024 0 Standard Deviation = 0
-0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Axial Load Ratio 0.2, Transv. Reinf. Ratio>0.006, Vy/VO>1.0 Axial Load Ratio>0.2, Transv. Reinf. Ratio>0.006, Vy/VO>1.0
1 1
0.5 0.5
Number of Tests = 0 Number of Tests = 0
0 0
-0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Error in aC = Experiment - Estimate (rad) Error in aC = Experiment - Estimate (rad)
Figure 5: Cumulative distributions of the error between experimental values and estimates for aC; data
split into bins covering various ranges of parameters and standard deviations given for equation estimates
only.
The effects of hook details of transverse reinforcement are explored in Fig. 6. In this figure,
the error between experimental values and equation-based estimates of a are plotted versus s/d.
Rectangular columns with ties having 135o hooks and 90o hooks are highlighted. Columns with
unknown hook details are also indicated. Spirally reinforced circular columns as well as those
with lapped ties are highlighted. As can be seen in the figure, there appears to be little overall
bias in equation estimates with respect to hook details. In fact, for rectangular columns with 90o
hooks the median of the error on the estimate of aR is 0.005 and the associated standard deviation
is 0.011, while for 135o hooks the median of the error on the estimate of aR is zero and the
associated standard deviation is 0.015. Similarly, for circular columns with lapped ties the mean
error of the estimate of aC is -0.002 and its associated standard deviation is 0.011, while for
spirally reinforced columns the mean error of the estimate of aC is 0.006 and its associated
standard deviation is 0.019. For large s/d ratios (above 0.75) the proposed equation for
rectangular columns provided conservative estimates even for columns with ties having 90o
hooks; 90% of all data points in that range have conservative estimates of aR. One should note
that columns in the database with lapped ties or ties having 90o hooks had relatively low
transverse reinforcement ratios, ρt (as is the case typically in older existing buildings). The
median ρt for rectangular columns with 90o ties is 0.0017, while that for circular columns with
lapped ties is 0.0008. At such low reinforcement ratios, the contribution of ties to the overall
1.9
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
deformation capacity of columns is relatively low, which may explain the observed insensitivity
of estimate errors to hook details of ties.
0.08 0.08
135o hooks spirals
Error in aR = Experiment - Estimate (rad)
0.04 0.04
0.02 0.02
0 0
-0.02 -0.02
-0.04 -0.04
0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.2 0.4 0.6 0.8 1 1.2 1.4
Rectangular column s/d Circular column s/d
Figure 6: Error between experimental values and equation-based estimates for a plotted versus s/d
with transverse reinforcement type highlighted
4.2.2 Estimates of b
The cumulative distribution of the error difference between experimental values and estimates
of b2 are plotted for rectangular and circular columns in Fig. 7. Errors are plotted for estimates
evaluated using the proposed regression equations as well as for those estimated using the ASCE
41 standard. The trends observed in Fig. 7 for values of b2 were similar to those observed for a
values. It is apparent from Fig. 7 that the proposed regression equations shifted estimates from
conservative ones based on the standard to median estimates; i.e., error = 0 at 0.5 probability of
exceedance. For b2R, both methods produced similar spread on estimate error as evidenced by the
similar slopes of the cumulative distribution curves and the standard deviations. For b2C
however, the proposed relation produced a large shift in estimates from those derived using the
standard. For brevity, figures similar to Fig. 4 and Fig. 5 are not presented for b2 values. Similar
trends as those for a values were observed in the various bins for b2 values.
1 1
0.9 0.9
0.8 0.8
0.7 0.7
Cumulative Distribution
Cumulative Distribution
0.6 0.6
0.5 0.5
Equation estimates Equation estimates
0.4 0.4
ASCE 41-06 estimates ASCE 41-06 estimates
0.3 0.3
Number of Tests = 319 Number of Tests = 171
0.2 0.2
Standard Deviation Equation = 0.019 Standard Deviation Equation = 0.018
0.1 0.1
Standard Deviation ASCE41-06 = 0.022 Standard Deviation ASCE41-06 = 0.02
0 0
-0.04 -0.02 0 0.02 0.04 0.06 0.08 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Error in b2R = Experiment - Estimate (rad) Error in b2C = Experiment - Estimate (rad)
Figure 7: Cumulative distribution of the error between experimental values and estimates for b2R and b2C
1.10
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
0.1 0.1
Estimate of b 2R (rad)
Estimate of b 2C (rad)
0.08 0.08
0.06 0.06
0.04 0.04
0.02 0.02
0 0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14
Experimental b2R (rad) Experimental b2C (rad)
Figure 8: Comparison between experimental and model estimates for plastic rotations at axial collapse b2
In Eq. 6, θ is the angle of the critical shear crack from horizontal. Elwood and Moehle
recommend adopting a value of θ of 65̊. If this value is used, the equation is reduced to:
∆ .
Eq. 7
.
When originally developed, the provisions in the ASCE 41 standard were intended to
provide a conservative estimate of the drift ratio at axial failure. The conservative nature of the
standard is reflected in the Fig. 9, in which the drift ratio at axial failure calculated following the
1.11
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
provisions in the standard and the measured drift ratio at axial failure for several collapse tests
[12, 15-18] are compared. The data set used for the comparison shown in Fig. 9 includes
rectangular columns that are both slender and short, columns that are shear-critical, and columns
that are flexure-shear critical (i.e., yield in flexure prior to sustaining shear failure).
Figure 9 corroborates that drift ratios at axial failure calculated according to the provisions
in ASCE 41 are very conservative for rectangular slender and short columns subjected to
uniaxial loading protocols with three cycles per target lateral drift. Contrary to that, estimates of
the drift ratio at axial failure for specimens subjected to biaxial loading with three cycles per
target drift in each direction or uniaxial loading with six cycles per target drift correspond
approximately to the mean, which is consistent with the observation that columns with these
types of loading protocols experience axial failure at lower drift ratios than columns subjected to
fewer loading cycles [12, 19].
Figure 9: Comparison between measured and calculated drift ratio at axial failure; calculations made
using the ASCE 41 provisions. Test data from [12, 15-18].
A similar comparison between experimental drift ratio at axial failure and that calculated
using the proposed simplified model for rectangular columns is presented in Fig. 10. A different
trend was observed in Fig. 10 than in Fig. 9, whereby calculated drift ratios at axial failure for
slender and short columns subjected to uniaxial loading protocols (3 cycles per drift level)
corresponded approximately to the mean, while estimates of drift ratio for specimens subjected
to biaxial loading and higher number of cycles were unconservative.
1.12
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
Figure 10: Comparison between measured and calculated drift ratio at axial failure; calculations
made using the proposed model. Test data from [12, 15-18].
On the basis of these observations, a reduction factor of 0.5 was applied to drift estimates
derived using the proposed model to calculate parameter b for columns subjected to biaxial
loading or columns subjected to loading protocols with large numbers of cycles per drift ratio.
The results of applying the reduction factor are shown in Fig. 11. As can be seen in the figure,
the reduction factor improved the fit of the model drastically. The statistical results for the data
set of rectangular columns evaluated are presented in Table 2.
Figure 11: Comparison between measured and calculated drift ratio at axial failure; calculations
made using the proposed model with reduction factor for loading protocol. Test data from [12, 15-18].
Table 2: Statistical comparison of measured and calculated drift ratio at axial failure
Model Average Measured/ Standard Coefficient of
Calculated drift ratio Deviation Variation
at axial failure
ASCE 41-06 Suppl. 1 1.97 0.93 0.47
Elwood-Moehle 0.97 0.44 0.46
Simplified Elwood – Moehle
1.13 0.38 0.33
with loading factor
1.13
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
Parameter b2 values for all rectangular and circular columns in the database are shown in
Fig. 12. Estimates of parameter b were obtained using the proposed simplified equation for drift
ratio at axial failure. Parameter b was calculated by subtracting the experimentally observed drift
ratio at yield Δy/L from the drift ratio at axial failure estimated by the model. Tests conducted to
axial collapse are highlighted in the figure.
0.14 0.14
Tests not to axial collapse Tests not to axial collapse
Tests to axial collapse Tests to axial collapse
0.12 0.12
Proposed behavioral model b (rad)
0.08 0.08
0.06 0.06
0.04 0.04
0.02 0.02
0 0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14
Experimental b2R (rad) Experimental b2C (rad)
Figure 12: Comparison between experimental and model estimates for plastic rotations at axial
collapse; model estimates not adjusted by a factor of 1.3 for circular columns
For rectangular columns, the mean error ratio for b1 = 1.07 (mean experimental b1 / model
estimate), and the corresponding coefficient of variation was 0.71. For circular columns the mean
error ratio for b1 was 1.64 and the corresponding coefficient of variation was 0.57. For a majority
of column tests not conducted to collapse (201 rectangular and 117 circular), the model produced
higher estimates of the b2 values (tests to the left of the diagonal lines in Fig. 12). However,
given that the experimental b2 values constitute a lower-bound of the actual plastic rotations at
axial failure, it is not possible to conclude whether the model produced conservative or
unconservative estimates for those tests. The model produced lower estimates of the b2 values for
many tests that were not conducted to collapse (82 rectangular and 45 circular tests to the right of
the diagonal lines in Fig. 12), which indicates conservative estimates of the plastic rotations at
axial failure for those tests. Such numbers suggest that the model provided a reasonable estimate
of the mean plastic rotations at axial failure for tests conducted to collapse and may not be
unconservative over the entire database. For circular columns however, the analysis indicates
that the model provided a conservative estimate of the parameter b, particularly considering that
the mean error ratio of b1 values was 1.64. In light of these results, and given the paucity of
experimental results to axial failure of circular column, a factor of 1.3 is proposed to increase
model estimates for spirally reinforced circular columns. This factor was chosen by comparing
the terms of statistically-derived equations for the parameter b2 of circular and rectangular
columns. While the 1.3 factor does generate a mean error ratio still significantly larger than 1.0
(=1.26) for b1 values, it is consistent with the approximate 30% difference between the
regression based b2R and b2C estimates. Increasing model estimates by more than 30% was not
deemed appropriate given the limited test data (only nine tests on circular columns were
conducted to collapse).
1.14
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
1.15
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
expected given the observed higher deformation capacities of circular columns that are not
accounted for in the standard.
0.05 0.07
0.045
0.06
0.04
0.035 0.05
ASCE 41-06 aR (rad)
0.015 0.02
0.01
0.01
0.005
0 0
0 0.01 0.02 0.03 0.04 0.05 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
Proposed equation aR (rad) Proposed equation aC (rad)
Figure 13: Difference in estimates of a between ASCE41 and the proposed equations
A comparison between b values estimated using of the proposed mechanics-based model
and those of the ASCE 41 standard are shown in Fig. 14. For both rectangular and spirally
reinforced circular columns, the proposed model estimates increased deformation capacities
significantly over standard estimates. This was to be expected because the standard targets
conservative deformation capacities while the model targets median estimates.
0.12 0.14
Tests not to axial collapse Tests not to axial collapse
Tests to axial collapse Tests to axial collapse
0.12
Proposed behavioral model b (rad)
Proposed behavioral model b (rad)
0.1
0.1
0.08
0.08
0.06
0.06
0.04
0.04
0.02
0.02
0 0
0 0.02 0.04 0.06 0.08 0.1 0.12 0 0.02 0.04 0.06 0.08 0.1 0.12
ASCE 41-06 b(rad) ASCE 41-06 b (rad)
Figure 14: Difference in estimates of b between ASCE 41 and the proposed model (including the
30% adjustment for circular columns)
8. Acceptance Criteria
Acceptance criteria (AC) should reflect the amount of damage that is acceptable for the
Immediate Occupancy (IO) performance objective (defined in [3]). For more severe performance
objectives related to safety and structural stability, considerations for selecting AC should not
only reflect acceptable damage levels but also the probability of exceeding a threshold
behavioral state such as a column initiating lateral-strength degradation and exhibiting a negative
lateral stiffness. . The ASCE 41 standard provides procedures for defining AC based on
experimental data. For all members, the standard defines AC for the Immediate Occupancy (IO)
performance objective as the critical deformation at which permanent visible damage occurs in a
1.16
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
member. The standard defines AC differently for primary and secondary members at the Life-
Safety (LS) and Collapse-Prevention (CP) performance objectives. For primary members, AC
are defined as 75% of the MP a for LS and equal to a but not greater than 75% of b for CP. For
secondary members, AC are defined as 75% of MP b for LS and equal to b for CP. It is useful to
recall here that current MP in the standard were selected conservatively and thus by limiting AC
to the MP a or b, the standard in effect aims to limit deformations to below values at which loss
of lateral strength or member failure occur.
It is proposed here that AC for IO be based on a percentage of the a value since columns
with better detailing and higher deformation capacities tend to exhibit visible damage at larger
deformations than less well detailed columns. A conservative value of 10% of the a parameter is
selected as the limiting deformation at which a reinforced column is deemed to need repair and
no longer satisfy the IO performance objective.
For more severe performance objectives (i.e., Life-Safety (LS) and Collapse-Prevention
(CP)), structural stability and safety are of paramount concern. For these performance objectives,
ensuring a fixed probability of exceeding the deformation at onset of lateral-strength degradation
(a) or the onset of member failure (b) is recommended. Thus for primary structural members, AC
should be defined such that the deformation at onset of lateral-strength degradation (MP a) has a
low probability of being exceeded. For secondary structural members, AC should be defined so
the deformation corresponding to member loss of strength (MP b) has a low probability of being
exceeded. Because the errors on the estimates of MP exhibit different dispersions for various
parameters, selecting a fixed fraction of those MP values for AC results in different probabilities
of exceeding MP deformations for members with deformations not exceeding the AC.
Consequently, it is recommended that AC be defined through specific percentiles rather than
percentages of MP. Percentiles for selecting AC were selected to remain consistent with the
intent of the current ASCE 41standard definitions for AC (i.e., percentiles were selected to
provide percentage values of MP equivalent to those currently defined in the standard). The
following summarizes the selected percentiles recommended for the various performance
objectives:
a. For LS of primary members, it is proposed that plastic rotations should not exceed
the 20th percentile of a. For a member critical to the stability of a structure,
satisfying the AC for LS would indicate a high level of confidence (80%) that the
member under consideration has not initiated lateral-strength degradation.
b. For CP of primary members, it is proposed that plastic rotations should not exceed
the 35th percentile of a. The ASCE 41 standard specifies that this AC be equal to
the MP a. Given that MP provided in the ASCE 41 standard are conservative
estimates of the behavioral milestones rather than median estimates, the AC is
given here as lower than the proposed median estimate of that parameter.
c. For LS of secondary members, it is proposed that plastic rotations should not
exceed the 10th percentile of b nor be less than AC for LS of primary members.
Due to the more critical nature of the behavioral milestone identified by b, a lower
percentile was selected for this AC than for primary members.
d. For CP of secondary members, plastic rotations should not exceed the 25th
percentile of b nor be less than AC for CP of primary members.
e. In no case should the AC for primary members be larger than those for secondary
members.
1.17
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
The percentages of MP a and b that satisfy the selected percentiles for reinforced concrete
columns are given in Table 3. As can be seen from the table, the resulting AC are on the order of
30 to 50% of a for LS of primary members, 65 to 75% of a for CP of primary members, 60% of
b for LS of secondary members, and 75% of b for CP of secondary members. The range of
percentages needed to achieve the same percentiles of MP illustrates the need to move to fixed
percentiles and variable percentages to remain consistent with the performance-based philosophy
of the ASCE 41 standard.
1.18
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
shown to reduce the drift at axial collapse by up to 50%. Because the loading protocols that
generated such reductions are unusually severe and it is difficult to correlate them to earthquake
duration, no further adjustments to the proposed relations were introduced. Engineers are advised
to reduce model estimates of drift ratio at axial failure should an unusually demanding
displacement history be expected to occur.
Acceptance criteria (AC) were proposed that identify acceptable deformation levels at various
performance objectives. Limiting deformations for Life-Safety and Collapse-Prevention
performance objectives were defined in terms of the a and b MP such that reaching the AC
deformations would correspond to a low probability of exceeding deformations at onset of lateral
or axial degradation.
Acknowledgements
The authors would like to thank Dr. Kenneth Elwood for his insightful advice and support during
the development of the proposed relations. The authors would also like to acknowledge the help
of Insung Kim, Garrett Hagen, and Degenkolb Engineers for providing information regarding the
compliance of columns in the database with ACI318-11 seismic provisions.
1.19
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
Notations
a = modeling parameter representing the plastic rotation at incipient lateral-strength degradation;
experientially extracted as a = (Δ0.8-Δy)/L
aC = modeling parameter a for circular columns
aR = modeling parameter a for rectangular columns
Ag = concrete gross section area
Av = area of transverse reinforcement in the direction of loading spaced at s
b = modeling parameter representing the plastic rotation at incipient axial degradation
b1 = modeling parameter b for column tests conducted to axial collapse = (Δaxial-Δy)/L
b2 = modeling parameter b for all column tests including those conducted to axial collapse
b2C = modeling parameter b2 for circular columns
b2R = modeling parameter b for rectangular columns
bw = section web width
c = modeling parameter representing the residual lateral strength of a column
d = section effective depth form edge of compression region to centroid of tension steel. d is
taken as 0.8D (D=section diameter) for circular columns
dc = depth of core (centerline to centerline of ties)
f’c = concrete compressive strength
fyt = transverse steel yield strength
L = column clear length
La = column shear span
M/Vd = ratio of applied moment to applied shear times the effective depth and is bounded by the
values of 4 and 2
My = column moment strength
Nu = applied axial load used for shear strength evaluation (= 0 for tension loads)
P = axial load (positive in compression)
P/ Ag f’c = axial load ratio
s = transverse reinforcement spacing
Vmax = maximum applied shear
Vo = nominal shear strength of columns as defined by ASCE 41-06 but with the following
modification: the shear strength provided by steel is fully accounted for up to a s/d ratio of
0.75, is taken equal to zero for a s/d ratio above 1.0, and linearly interpolated between full
strength and zero strength between s/d = 0.75 and s/d = 1.0
Vy = shear corresponding to the development of flexural yield moments = Mp / La
1.20
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
1.21
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
References
1.22
@Seismicisolation
@Seismicisolation
Nonlinear modeling parameters and acceptance criteria for concrete columns
19. Simpson, B. and A.B. Matamoros, Criteria for Evaluating the Effect of Displacement
History and Span-to-Depth Ratio on the Risk of Collapse of R/C Columns, in 15th World
Conference on Earthquake Engineering2012: Lisbon, Portugal.
20. American Concrete Institute (ACI) Committee 374, Acceptance Criteria for Moment
Frames Based on Structural Testing and Commentary (374.1-05), 2005, American
Concrete Institute: Farmington Hills, MI. p. 9.
1.23
@Seismicisolation
@Seismicisolation
W.M. Ghannoum and A.B. Matamoros
1.24
@Seismicisolation
@Seismicisolation
SP-297—2
A database comprised of 59 reinforced concrete columns subjected to strong ground shaking using earthquake
simulators (or shaking tables) is compiled. This paper will focus on insights provided by the database related to the
concrete column provisions in ASCE/SEI 41. In particular, the Shaking Table Test Column Database is used to
evaluate the accuracy of column effective stiffness models, column classification criteria, and the level of
conservatism provided by the plastic rotation capacities specified in ASCE/SEI 41. It is found that the Standard
generally overestimates the column effective stiffness, while providing a mean value estimate of the column shear
strength regardless of tie spacing. The modeling parameters specified in the standard provide conservative estimate of
the column drift capacities and are consistent with the targeted probability of failure. Refinements of the shear
strength model and the criteria for column classifications are suggested. This study also compares the measured
response of columns subjected to quasi-static cyclic loads and shaking table tests.
Keywords: Shaking table tests; ASCE/SEI 41; reinforced concrete columns; shear failure, axial-load failure
1
Design Engineer, Read Jones Christoffersen, Vancouver, Canada
2
Professor, Department of Civil Engineering, University of British Columbia, Canada
3
Professor, Department of Civil Engineering, National Taiwan University, Taiwan
2.1
@Seismicisolation
@Seismicisolation
Y. Li et al.
INTRODUCTION
ASCE/SEI 41 (2013) is commonly used throughout the United States and internationally for the seismic
assessment and retrofit of existing concrete buildings. In 2007, the concrete provisions of ASCE/SEI 41 were
updated to better reflect the observed performance of concrete components from quasi-static cyclic laboratory tests
(Elwood et al. 2007). These updates have been incorporated in the latest release of the seismic rehabilitation
standard, ASCE/SEI 41 (2013).
Examples of other backbone models or acceptance criteria for existing concrete columns include
Panagiotakos and Fardis (2001), Elwood and Moehle (2006) and Haselton et al. (2008). Panagiotakos and Fardis
(2001), used as the basis for acceptance criteria in Eurocode 8 Part 3 (EN 1998-3 2005), provided empirical
equations for estimating the chord rotation at yield and 20% loss of strength. Elwood and Moehle (2006) provided a
backbone model for columns experiencing shear failure after flexural yielding, including estimates of the column
effective stiffness and drift ratio at shear and axial load failure. Haselton et al. (2008) provided a backbone model
for flexure-controlled columns based on a database of 255 quasi-static cyclic column tests.
Due to the prevalence of application in engineering practice, this study will focus on the backbone model and
provisions for concrete columns in ASCE/SEI 41. In particular, a database comprised of 59 reinforced concrete
columns subjected to strong ground shaking using earthquake simulators (or shaking tables) has been compiled to
evaluate the accuracy of column effective stiffness models, the performance of column classification criteria, and the
level of conservatism provided by the plastic rotation capacities specified in ASCE/SEI 41. A comparison of the
measured response of columns subjected to quasi-static cyclic loads and shaking table tests is also included.
2.2
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
EXPERIMENTAL DATABASE
To assess the ASCE/SEI 41 backbone model, a Shaking Table Test Column Database is compiled (for
brevity, this database will be referred to as “Dynamic Database” herein). The Dynamic Database includes results of
shaking table tests that have been conducted on simple two-dimensional frames with relatively large-scale
reinforced concrete columns (length scale factor no less than 1:3.5), for which critical test data (column axial load,
column shear, horizontal drift ratio and column shortening/lengthening) are available. Based on this criteria, seven
shaking table test programs were identified, three of which were conducted at UC Berkeley (Elwood 2002,
Ghannoum 2007, and Shin 2007) and the remaining four were carried out at National Center for Research on
Earthquake Engineering (NCREE) in Taiwan (Su 2007, Kuo 2008, Wu et al. 2006, Yavari 2011). Most of the
selected shaking table tests were conducted up to collapse of the frame specimens. Typical test configurations are
shown in Figure 2. All specimens were single story frames with the exception of specimens tested by Yavari (2011)
and Ghannoum (2007) which were two and three stories, respectively. Details of the shaking table tests are
summarised in Li (2012).
Fifty-nine (59) first-story columns out of a total of 97 columns are included in the Dynamic Database (upper
story column shear forces and axial loads were not recorded directly and cannot be determined with sufficient
reliability to be included in the database). Figure 3 provides the range of column geometry, material properties and
reinforcement details for columns in the database; data for each column are provided in Appendix A.
Columns in the Dynamic Database were classified into three types based on the observed failure modes:
flexure critical, flexure-shear critical and shear critical columns. The classification criteria based on column
properties and recorded response during testing are summarized in Table 1.
Fifteen of the flexure-shear-critical columns were part of a larger multi-story frame and the top of the
columns were connected to flexible beams similar to the configuration shown in Figure 2(a). This is relevant to the
assessment below since the flexible beam results in additional deformations from beams and joints, contributing to
the measured interstory drift. For specimens with rigid beams (e.g. setups similar to Figure 2(b)), the additional
deformation from beams and joints can be neglected. All shear-critical columns were connected to rigid beams.
(a) (b)
Figure 2 Typical configuration of tests in Dynamic Database (a) columns connected with flexible beams
(source: Yavari 2011); (b) columns connected with rigid beam (source: Shin 2007)
2.3
@Seismicisolation
@Seismicisolation
Y. Li et al.
S FS F S FS F S FS F
40 40 40
35 35 35
Number of Columns
Number of Columns
30 30
Number of Columns
30
25 25 25
20 20 20
15 15 15
10 10 10
5 5 5
0 0 0
20~24 24~28 28~32 32~36 36~38 <300 <350 <400 <450 <500 <400 <500 <600 <700
Concrete Strength (MPa) Long. Rein. Yield Strength (MPa) Trans. Rein. Yield Strength (MPa)
S FS F S FS F S FS F
40 40 40
35 35 35
Number of Columns
Number of Columns
Number of Columns
30 30 30
25 25 25
20 20 20
15 15 15
10 10 10
5 5 5
0 0 0
<2.0 <2.5 <3.0 <3.5 <4.0 <4.5 <1.5% <2.0% <2.5% <3.0% <3.5% <0.20% <0.60% <1.00% <1.40%
Aspect Ratio a/d Long. Rein. Ratio Trans. Rein. Ratio
S FS F S FS F
40 40
35 35
Number of Columns
Number of Columns
30 30
25 25
20 20
15 15
10 10
5 5
0 0
<0.1 <0.2 <0.3 <0.4 <0.1 <0.2 <0.3 <0.4 <0.5
Initial Axial Load Ratio Pini/Agfc' Max. Axial Load Ratio PMax/Agfc'
Figure 4 shows the normalized envelopes of measured lateral response for columns experiencing shear or
flexure-shear failure modes in the Dynamic Database. Column lateral force is corrected for P-Delta effects to give
the effective lateral force instead of recorded shear force of each column (unless specified otherwise, the lateral
force used throughout this paper is the column shear force corrected for P-Delta effects). The x-axis is the ductility
demand, based on the measured yield displacement determined from the maximum force observed during the test
and an effective stiffness based on a straight line from the origin through the measured response of the column at 75%
of the maximum lateral force. Detailed descriptions of the tests and the development of the envelopes can be found
in Li (2012). Despite the fact that all columns in Figure 4 experienced shear failures (before or after flexural
yielding), it is noted that all columns achieved ductilities greater than unity and the majority of flexure-shear-critical
columns were able to sustain lateral loads beyond a ductility demand of 2.0.
2.4
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
Normalized Force (V/Vmax)
0.4 0.4 0.4
Normalized Force (V/Vmax)
0.0 0.0 0.0
Failure mode
Flexure-critical Flexure-shear critical Shear-critical
• Inadequate transverse • Non-ductile columns with
reinforcement for ductile inadequate transverse
• Adequate transverse response (ρt≤0.002; 90° hooks) reinforcement
Column
reinforcement • Considered to be (transverse reinforcement ratio
Properties
(ρt>0.002; 135° hooks) representative of columns found ρt≤0.002; 90° hooks)
in older reinforced concrete • Usually have low aspect ratio
structures. (i.e. "short column")
2.5
@Seismicisolation
@Seismicisolation
Y. Li et al.
2
Shear‐critical
Flexure‐Shear‐critical
Flexure‐Shear‐critical (Beam)
1.5
Flexure‐critical
EIMeas /EIASCE
Flexure‐critical (Beam)
Mean
1
0.5
0
0 0.1 0.2 0.3 0.4
Initial axial load ratio Pini /Agfc'
Figure 5 Evaluation of column effective stiffness estimated by ASCE/SEI 41
2
Shear‐critical
Flexure‐Shear‐critical
Flexure‐Shear‐critical (Beam)
1.5
Flexure‐critical
EIMeas /EIEE
Flexure‐critical (Beam)
Mean
1
0.5
0
0 0.1 0.2 0.3 0.4
Initial axial load ratio Pini /Agfc'
Figure 6 Evaluation of column effective stiffness estimated by Elwood and Eberhard (2009)
EFFECTIVE STIFFNESS
Most columns included in the Dynamic Database could be considered as fixed against rotation at both ends,
thus the relationship between measured column lateral stiffness, kMeas, and effective flexural rigidity, EIeff, can be
expressed by Equation 1:
where Vy is the effective yield strength of the column; θy is the yield rotation; L is the column clear height.
The effective flexural rigidity (EIeff) is often expressed as a fraction of the gross-section flexural rigidity EIg. In the
interest of using common terminology from engineering practice, the ratio (EIeff /EIg) will herein be referred to as
effective stiffness ratio.
In this research, the measured effective stiffness ratio, EIMeas/EIg, and yield drift for columns in the Dynamic
Database are based on a secant stiffness to 75% of the maximum lateral force and corresponding drift on the
measured lateral force-drift envelope. According to ASCE/SEI 41, the effective stiffness ratio (EIeff /EIg) for
columns can be estimated as 0.3 for axial load ratios (P/Agfc’ ) less than 0.1, and 0.7 for axial load ratios higher than
0.5. Linear interpolation of the effective stiffness ratio is used between these limits.
Results of measured to calculated effective stiffness ratio, EIMeas/EIASCE, are presented in Figure 5 and
grouped by column failure mode. For specimens subjected to multiple ground motions, only the first test is
considered here to avoid consideration of stiffness reductions associated with previous testing. Specimens connected
to a flexible beam are indicated by a different symbol in Figure 5 (i.e. “Flexure-critical (Beam)” and “Flexure-shear-
critical (Beam)”), since such cases are expected to result in a lower bound estimate of the column stiffness due to the
additional flexibility of the beam and joint. Statistics are provided in Table 2 for all specimens and just those
specimens with rigid beams to remove the additional variability introduced by the beam flexibility.
2.6
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
Based on the test results of 150 columns subjected to quasi-static cyclic loads, Elwood and Eberhard (2009)
proposed an effective stiffness model taking into account the deformation due to flexure, shear and bar slip, as
shown in Equation 2:
EI EE 0.45 2.5 P Ag f c
'
0.2 1.0
EI g db D Equation 2
1 110
D a
where P is the column axial load; Ag is the gross cross-section area of the column; f’c is the concrete compressive
strength; db is the nominal diameter of longitudinal bars; D is the diameter of circular column or column depth in
direction of loading and a is the shear span of the column.
Ratio of measured to calculated effective stiffness EIMeas/EIEE by Elwood and Eberhard (2002) for the Dynamic
Database is presented in Figure 6 and Table 2.
ASCE/SEI 41 tends to overestimate the stiffness of nearly all columns regardless of failure mode. Elwood
and Eberhard (2009) gives an improved estimation (closer to mean value of the test results) but still generally
overestimates the measured stiffness, except for shear critical columns. Since the Elwood and Eberhard model was
developed to provide a close estimate of the mean response from quasi-static cyclic tests, these results suggest that
columns subjected to real ground motions tend to exhibit a slightly lower effective stiffness compared to quasi-static
cyclic tests. This result may be attributed to the large number of below-yield cycles typically imposed on columns
during real ground motions; however, further testing would be required to generalize this conclusion.
Standard Coefficient of
Min Max Mean Median Deviation Variation(COV)
Flexure-critical columns
EIMeas/EIASCE 0.27 0.98 0.58 0.57 0.14 25%
EIMeas/EIASCE (Rigid beam only) 0.27 0.76 0.58 0.60 0.12 21%
EIMeas/EIEE 0.41 1.24 0.75 0.73 0.20 26%
EIMeas/EIEE (Rigid beam only) 0.41 1.07 0.75 0.74 0.16 21%
Flexure -shear-critical columns
EIMeas/EIASCE 0.41 1.21 0.70 0.67 0.17 24%
EIMeas/EIASCE (Rigid beam only) 0.50 1.21 0.71 0.65 0.17 24%
EIMeas/EIEE 0.52 1.33 0.89 0.87 0.18 20%
EIMeas/EIEE (Rigid beam only) 0.63 1.31 0.88 0.86 0.16 18%
Shear-critical columns
EIMeas/EIASCE (Rigid beam only) 0.65 1.27 0.79 0.75 0.18 23%
EIMeas/EIEE (Rigid beam only) 0.97 1.90 1.19 1.13 0.27 22%
COLUMN CLASSIFICATION
To establish the deformation capacity of a concrete column, ASCE/SEI 41 classifies columns into three
“conditions”, or failure modes, based on the ratio of the shear strength at low ductility demand, V0, to the plastic
shear demand, Vp. The plastic shear demand, Vp, is the shear force corresponding to flexural yielding of plastic
hinges. As shown in Figure 1, for columns with both ends fixed (as assumed for columns in the Dynamic Database),
Vp is calculated by dividing the maximum moment strength obtained from section analysis by shear span, a=L/2.
V p 2M p / L Equation 3
ASCE/SEI 41 adopts the shear strength model developed by Sezen and Moehle (2004) and recommends
calculating column shear strength, Vn, by summing the contribution from concrete, Vc, and transverse reinforcement,
2.7
@Seismicisolation
@Seismicisolation
Y. Li et al.
Vs, as shown in Equation 4. The term V0= Vn /k is the shear strength excluding the reduction factor related to
ductility demand.
Vn kV0 k Vc Vt
0.5 f ' Nu
Vc c
1 0.8 Ag
M / Vd 0.5 f '
A
c g
Av f y d
s / d 0.5 Equation 4
s
Av f y d
Vs 0.5 0.5 s / d 1.0
s
0 s / d 1.0
As indicated in Equation 4, ASCE/SEI 41 specifies that if the spacing of column transverse reinforcement is
larger than half of the effective depth of column section but less than effective depth (d/2< s ≤ d) in the direction of
applied shear force, the transverse reinforcement should be taken as no more than 50% effective when resisting
shear or torsion, and thus Vs is set to Vs = 0.5Avfyd / s. If the spacing is larger than the effective depth of the column
section (s > d), the transverse reinforcement shall be considered ineffective and Vs =0. This modification of Vs is
not included in the original model by Sezen and Moehle (2004).
Su (2007) observed that the shear strengths for shear-critical columns in the NCREE 2007shake table tests
were higher than for columns subjected to quasi-static test with similar configurations. This difference may be
attributed to the higher strain rate and fewer pre-failure cycles that shear-critical columns experienced during the
dynamic tests. This observation suggests that the Sezen and Moehle model may be conservative for the columns in
the Dynamic Database as this original model was developed based on a database of quasi-static cyclic tests.
Based on the calculated shear strength, V0, and plastic shear demand, Vp, ASCE/SEI 41 classifies columns
with 135-degree hooks on transverse reinforcement into three conditions: condition i, ii and iii. These conditions are
intended to correspond to the following failure modes:
Poor transverse reinforcement detailing reduces the column lateral deformation capacity and therefore,
columns with transverse reinforcement including 90-degree hooks or having lap spliced transverse reinforcement are
adjusted to a poorer condition (e.g., from condition i to condition ii) as shown in Table 3.
The accuracy of the ASCE/SEI 41 classification method is evaluated using the matrix shown in Table 4. Only
columns that were observed to experience column failure (i.e. 20% drop in lateral resistance) during the tests are
included. The data along the diagonal in the blue cells indicates the number of columns that are classified properly
by ASCE/SEI 41. The column data that falls in the green cells provides a “conservative” estimation of the column
failure type in terms of column drift capacity (for example “23” indicates that there were 23 columns that are
classified as shear-critical columns, yet were observed to have experienced flexure-shear failures during the tests).
2.8
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
While “green zone” classification may result in unnecessary retrofit costs, it is unsafe and hence unacceptable to
have column tests falling in the red cells in Table 4. Results in Table 4 show that ASCE/SEI 41 is over-conservative
since the 23 of the 27 flexure-shear-critical columns are categorized in Condition iii.
ASCE/SEI 41 Classification
Flexure-Shear 0 4 23 27
Shear 0 0 5 5
Total 0 6 28 34
Flexure-Shear 0 27 0 27
Shear 0 0 5 5
Total 0 29 5 34
To provide a better match with the observed failure modes from the Dynamic Database, two modifications to
the classification methodology in ASCE/SEI 41 are proposed. First, the shear strength, V0, is not reduced for widely
spaced ties (d > s > d/2). This is consistent with the development of the shear strength model in Sezen and Moehle
(2004) where columns with widely spaced ties are included in the database without modification. Note that the
Dynamic Database does not include any columns with s>d, hence the impact of the penalty for such large spacing
cannot be evaluated here. It is considered prudent to maintain this limit considering the potential for a diagonal
crack not crossing any hoops for such wide spacing.
Secondly, it is found that most of the flexure-shear-critical columns that are misclassified as Condition iii
have ratio Vp /V0 only slightly over unity. If the criterion for flexure-shear-critical columns is modified to
1.1≥ Vp /V0 >0.6 instead of 1.0≥ Vp /V0 >0.6, the classification improves substantially. This relaxation of the upper
limit on the shear ratio for Condition ii reflects the fact that columns with plastic shear demand very close to the
shear capacity are expected to exhibit some limited plastic deformation capacity even if post yield deformations are
dominated by shear deformations. The relaxation of the upper limit is also consistent with the observation noted
above that columns subjected to dynamic loads tend to exhibit slightly higher shear strengths (Su 2007).
As shown in Table 5, by using the unreduced column shear strength and the modified upper limit on Vp /V0
for condition ii, flexure-shear-critical columns could be properly identified, resulting in more accurate modeling
parameters and acceptance criteria, and subsequently, a more economical retrofit solution. The modified
classification criteria are used to classify columns and evaluate the recommended modeling parameters in the
following section.
By classifying columns into the aforementioned three conditions based on failure modes, ASCE/SEI 41
provides recommendations on modeling parameters and allowable plastic rotations for columns in each condition.
As shown in Figure 1, the modeling parameters a and b are provided to represent the plastic rotation at significant
loss of lateral resistance and loss of axial load resistance, respectively. Plastic Rotation Ratio (PRR), calculated as
the ratio of measured plastic rotation to the modeling parameter, is adopted here to assess the accuracy of the
ASCE/SEI 41 modeling parameters.
2.9
@Seismicisolation
@Seismicisolation
Y. Li et al.
Non-ductile columns with inadequate transverse reinforcement are vulnerable to damage associated with
shear failure when subjected to seismic loads. Consistent with the definition of modeling parameter a in ASCE/SEI
41, the column shear failure point is taken as the point that column lateral resistance degrades to 80% of its peak
lateral resistance. Most drift capacity models available also have been developed to provide similar estimations of
column drift at 80% of maximum effective force (Elwood and Moehle 2005; Kato and Ohnishi 2002, etc.).
The measured plastic rotation capacity was taken equal to the measured drift ratio corresponding to a 20%
reduction in the maximum measured shear resistancel) minus the calculated yield drift ratio (y) using the
recommended effective stiffness from ASCE/SEI 41 and the plastic shear demand, Vp, as defined previously. Hence,
the plastic rotation ratio at loss of significant lateral resistance is defined as:
Plastic Rotation Ratio (PRR) = (l - y)/a Equation 5
Figure 7 shows the PRR at loss of significant lateral resistance for columns in the Dynamic Database. Data in
solid markers provide the PRR for columns that experienced failure (20% loss in lateral resistance) during the test,
while the hollow markers represent the lower bound of PRR based on the maximum recorded drifts for columns that
did not fail in the shake table test. It is noted that all flexure-critical columns (Condition i) did not experience
significant loss of lateral strength during the shaking table tests. The red solid diamond marker (Condition ii)
represents the PRR for flexure-shear critical columns that were connected by a rigid beam in the reinforced concrete
frame specimens. The PRR of flexure-shear-critical columns that were connected by flexible beam are represented
by purple solid diamond marker, namely “Condition ii (Beam)”. In such cases the measured drift ratio is influenced
by the flexibility of the beams and joints, indicating that the resulting PRR is likely higher than the true value for the
column alone and should be evaluated with caution.
To compare the PPR for columns subjected to dynamic loads and static cyclic loads, 50 non-ductile columns
subjected to quasi-static cyclic lateral loads are selected from the PEER Structural Performance Database (Berry et
al. 2004) to form a Static Dataset A (detailed column properties could be found in Appendix B). Figure 8 plots the
PRR at loss of significant lateral resistance for columns in Static Dataset A.
Condition i(lower) Condition ii Condition ii (Beam) Condition ii (lower)
5
PPR ( Plastic Rotation Ratio)
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
Initial Axial Load Ratio (Pini /Agfc')
Figure 7 Plastic rotation ratio at significant loss of lateral resistance for Dynamic Database
Condition ii
5
PPR (Plastic Rotation Ratio)
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
Axial load ratio (P/Agfc ')
Figure 8 Plastic rotation ratio at significant loss of lateral resistant force for Static Dataset A
2.10
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
Significant scatter exists among the data for PRR in Figure 7 and Figure 8, emphasizing the need to
determine the level of safety provided by ASCE/SEI 41 provisions. Results of plastic rotation ratio can be
represented by a lognormal cumulative distribution, commonly referred to as a fragility curve. The fragility curve in
Figure 9 focuses on the probability of failure for flexure-shear critical columns (i.e. Condition ii). In Figure 9(a), the
fragility curve is constructed based on the test data shown in red diamonds in Figure 7 (Condition ii columns
connected to rigid beams and observed to experience flexure-shear failures). The fragility curves for flexure-shear
critical columns in the Dynamic Database and Static Dataset A are compared in Figure 9 (b).
1 1
0.9 0.9
0.7 0.7
Probability of Failure (Pf)
Probability of Failure
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
From Test Data Dynamic Database
0.1 0.1 Static Dataset A
Lognormal CDF
0 0
0 1 2 3 4 5 0 1 2 3 4 5
Plastic rotation ratio Plastic Rotation Ratio
(a) (b)
Figure 9 Lognomal fragility curve for plastic rotation at significant loss of lateral resistance in
Dynamic Database and Static Dataset A
According to the target safety levels selected in the development of the ASCE/SEI 41 tables (as reported in
Elwood et al 2007), at a plastic rotation demand equal to parameter a the probability of failure, Pf, should be less
than 15% for columns vulnerable to shear failure, (i.e. columns in condition ii and iii); while for flexure-critical
columns in condition i, Pf up to 35% is acceptable since flexure failure is less sudden and brittle. (The “failure” here
refers to the point at which the column has lost more than 20% of its lateral resistance.) Based on Figure 9 (a), there
is a probability of failure of 3% for “Condition ii” columns in the Dynamic Database when subjected to plastic
rotation demand equal to parameter a specified in ASCE/SEI 41 (PRR=1), and thus, is lower than the ASCE/SEI 41
targets. As shown in Figure 9 (b), the probability of failure at PRR =1 for columns in the Static Dataset A is more
than four times that of the Dynamic Database, but still consistent with the ASCE/SEI 41 targets.
Most quasi-static cyclic tests are displacement-controlled (column specimens are typically subjected to two
or three reversed lateral force-deformation cycles at each horizontal drift ratio level until failure) and commonly
include more cycles at high displacement demands than expected from real ground motions used in the shaking table
tests. During multiple cycles in quasi-static cyclic tests, diagonal shear cracks open and close repeatedly, increasing
the column shear damage and thus resulting in faster degradation of column lateral resistance. This provides a
possible explanation for why the fragility curve for columns subjected to shaking table tests shifts to lower
probabilities of failure compared to that from quasi-static cyclic tests.
For columns expected to experience shear failure prior to flexural yielding (Condition iii), plastic deformations
cannot be relied upon prior to shear failure according to ASCE/SEI 41; and hence, parameter a is set to zero and the
drift ratio at lateral load failure is taken as V0/keff. Hence, the Total Rotation Ratio (TRR, Equation 6) is used here in
place of the Plastic Rotation Ratio to assess the level of conservatism provided by ASCE/SEI 41 for Condition iii
columns.
Total Rotation Ratio = l / (V0/keff Equation 6
2.11
@Seismicisolation
@Seismicisolation
Y. Li et al.
It is observed that all Condition iii columns exhibited TRR larger than unity, indicating that the ASCE/SEI 41
limits for shear-critical columns are conservative. Since only five shear-critical columns are included in the database,
the data are not sufficient to construct a fragility curve or recommend revisions to ASCE/SEI 41.
Condition iii
5
TPR ( Total Rotation Ratio)
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
Initial Axial Load Ratio (Pini /Agfc')
Figure 10 Total Rotation Ratio at significant loss of lateral resistance for Condition iii columns in Dynamic
Database
Modeling parameter b in ASCE/SEI 41 represents the plastic rotation at the point that a column experiences
axial-load failure and loses axial-load support. While it is intuitive to define the axial-load failure based on the axial
load response, load redistribution between columns in frame systems and column elongation/shortening due to
imposed lateral deformations can significantly influence the axial load data, rendering it difficult to capture the true
point of column axial-load failure from frame tests such as those in the Dynamic Database. For consistency among
all tests, this research adopts the method used in Yavari (2011) to define the column axial-load failure point as “the
point at which the peak vertical elongation of the column is recorded immediately prior to a sudden shortening of the
column”. Yavari (2011) found that this definition resulted in a failure point corresponding to a sudden drop in axial
load and visual failure of the column in videos. An example of the application of this definition is shown in Figure
11.
2.12
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
The measured plastic rotation capacity at axial-load failure was taken equal to the measured drift ratio at loss
of axial-load support a) minus the calculated yield drift ratio using the recommended effective stiffness from
ASCE/SEI 41 and the plastic shear demand, Vp, as defined previously. Hence, the plastic rotation ratio at loss of
axial load support is defined as:
Plastic Rotation Ratio (PRR) = (a - y)/b Equation 7
To compare the plastic rotation at axial-load failure for columns subjected to quasi-static cyclic tests and
shaking table tests, Static Dataset B consisting of 21 Condition ii columns is introduced and summarized in
Appendix C. These columns are subjected to unidirectional displacement cycles and constant axial load and
experience axial-load failure after shear failure. Static Dataset A could not be used for this assessment since only a
small subset of these tests experienced axial load failure. Figure 12 and Figure 13 plot the PRR (Plastic Rotation
Ratio) at loss of axial-load support for columns in the Dynamic Database and Static Dataset B, respectively.
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
Initial Axial Load Ratio (Pini /Agfc ')
Figure 12 Plastic rotation ratio at loss of axial-load support for Dynamic Database
Condition ii
5
PPR (Plastic Rotation Ratio)
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
Axial load ratio (P/Agfc ')
Figure 13 Plastic rotation ratio at loss of axial-load support for Static Dataset B
Similar to the previous section for modeling parameter a, results of plastic rotation ratio for Condition ii
columns in Dynamic Database are represented by a fragility curve, as shown in Figure 14 (a). The fragility curve is
constructed based on the test data of flexure-shear-critical columns that experienced axial-load failure during the
tests (data in red diamond marker in Figure 12).
2.13
@Seismicisolation
@Seismicisolation
Y. Li et al.
1 1
0.9 0.9
0.8 0.8
0.7 0.7
Probability of Failure
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
Test Data Static Dataset B
0.1 0.1
Fragility curve Dynamic Database
0 0
0 1 2 3 4 5 0 1 2 3 4 5
Plastic rotation ratio Plastic Rotation Ratio
Figure 14 Lognomal fragility curve for plastic rotation at loss of axial load support in Dynamic
Database and Static Dataset B
According to the target safety levels selected in the development of the ASCE/SEI 41 tables, when columns
experience plastic rotation demand equal to the specified parameter b, the probability of failure, Pf, should be less
than 15% (Elwood et al. 2007). (“Failure” here refers to the point that the column loses axial-load support.) Figure
14 (a) shows a 9% probability of failure for PRR=1.0, indicating that the probability of failure at loss of axial-load
support for flexure-shear-critical columns in the Dynamic Database is consistent with ASCE/SEI 41 targets. The
probability of failure for PRR = 1.0 for Static Dataset B is slightly higher, at 11%, as shown in Figure 14 (b). The
results indicate that the level of conservatism against axial load failure in ASCE/SEI 41 is consistent regardless of
the dataset used in the assessment, suggesting that axial load failure is not sensitive to the form of load application
(dynamic or static) or the number of cycles.
Considering the conclusions for the drift at lateral load failure (drifts from dynamic tests greater than drifts
from static tests) and axial load failure (drifts from dynamic tests similar to drifts from static tests) together, it is
noted that the gap between shear failure and axial load failure for columns subjected to shaking table tests is smaller
than that for columns in static cyclic tests. Indeed 56% of the flexure-shear-critical columns in the Dynamic
Database experience axial-load failure at a drift ratio within +/-0.1% of the drift at 20% drop in the lateral resistance.
This limited “buffer” between lateral and axial load failure for the Dynamic Database columns may be a result of
dynamic amplification of drift demands with the development of a negative stiffness (softening) and a rapid
acceleration of response after lateral load failure. Caution should be exercised in interpreting this observation,
however, since the specimens in the database included only a limited number of columns and the dynamic response
characteristics were heavily influenced by strength degradation in any one column. In a larger building structure,
particularly one with walls, failure of one column will not typically result in the same degree of softening and
associated dynamic amplification of drift demands.
CONCLUSIONS
Acceptance criteria in codes and standards (e.g. ASCE 41-13) have generally been developed and validated
based on comparisons with test data from quasi-static tests. To investigate the influence of real earthquake loading
on the performance of concrete columns a database consisting of 59 large-scale reinforced concrete columns tested
on a shaking table is compiled. This “Dynamic Database” is used to evaluate the accuracy and conservatism of the
concrete column provisions in ASCE/SEI 41-13.
ASCE/SEI 41 overestimates the effective stiffness ratio for columns in Dynamic Database. The mean
value of EIMeas/EIASCE for flexural-shear-critical columns connected with rigid beam in Dynamic
2.14
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
Database is 0.65, with coefficient of variation of 24%. This result is consistent with the observation
for columns subjected to quasi-static cyclic test.
Based on the current criteria of ASCE/SEI 41, most of the flexure-shear-critical columns in the
Dynamic Database are categorized into Condition iii (pure shear failure). This is too conservative,
since Condition iii columns are assumed to have zero plastic rotation capacity prior to shear failure.
The following modifications to the ASCE/SEI 41 column classification criteria are recommended: (1)
remove the penalty on column shear strength (V0) for columns transverse reinforcement spaced at
greater than d/2, while maintaining the penalty (Vs = 0) for s>d; (2) change the criteria for Condition ii
from 0.6<Vp/V0≤1.0 to 0.6<Vp/V0≤1.1.
Flexure-shear critical columns tend to develop larger plastic drifts at loss of significant lateral
resistance in shaking table tests, compared with columns subjected to quasi-static cyclic loads. Based
on Condition ii columns in the Dynamic Database, modeling parameter a in ASCE/SEI 41 provided a
3% probability of failure at significant loss of lateral resistance, which is significantly lower than the
probability of failure assessed based on columns subjected to quasi-static cyclic loads.
Flexure-shear critical columns tend to develop similar plastic drifts at loss of axial-load support in
shaking table tests, compared with columns subjected to quasi-static cyclic loads. Based on Condition
ii columns in the Dynamic Database, modeling parameter b in ASCE/SEI 41 provided a similar
probability of failure (9%) at loss of axial-load support compared to that assessed using columns
subjected to quasi-static cyclic loads.
Dynamic tests tend to result in a smaller difference between the drifts at lateral-load and axial-load
failure compared with static cyclic tests. For systems where strength deterioration of columns may
dominate the dynamic response of the structure, this observation suggests that real earthquakes may
result in more rapid degradation of lateral resistance for flexure-shear columns after shear failure than
had been previously assumed based on quasi-static cyclic tests.
REFERENCES
ASCE 2008. Seismic Rehabilitation of Existing Buildings. ASCE/SEI 41, Supplement 1. Reston,VA: American
Society of Civil Engineers.
ASCE 2013. Seismic Evaluation and Retrofit of Existing Buildings. ASCE/SEI 41. Reston,VA: American Society of
Civil Engineers.
Berry, M. Parrish, M. and Eberhard, M. 2004. PEER Structural Performance Database User’s Manual, Pacific
Earthquake Engineering Research Center, University of California, Berkeley. [available at:
www.ce.washington.edu/~peera1].
EN 1998-3. 2005. Eurocode 8: Design of structures for earthquake resistance – Part 3: Assessment and retrofitting
of buildings, European Committee for Standardization, Brussels.
Elwood, K. 2002. Shaking Table Tests and Analytical Studies on the Gravity Load Collapse of Reinforced Concrete
Frames. PhD Dissertaiton, University of California, Berkeley.
Elwood, K., Matamoros, A., Wallace, J., Lehman, D., Heintz, J., Mitchell, A., Moore, M., Valley, M. and Lowes, L.
2007. Update to ASCE/SEI 41 Concrete Provisions. Earthquake Spectra, 23(3), 493-523.
Elwood, K. J. and Eberhard, M. O. 2009. Effective Stiffness of Reinforced Concrete Columns. ACI Structural
Journal, 106 (4), 476-484.
2.15
@Seismicisolation
@Seismicisolation
Y. Li et al.
Elwood, K. J. and Moehle, J. P. 2005.Drift Capacity of Reinforced Concrete Columns with Light Transverse
Reinforcement. Earthquake Spectra, 21(1), 71-89.
Elwood, K.J., and Moehle, J.P., 2006. “Idealized backbone model for existing reinforced concrete columns and
comparisons with FEMA 356 criteria”, Structural Design of Tall and Special Buildings, 15(5), 553-569.
Ghannoum, W. M. 2007. Experimental and Analytical Dynamic Collapse Study of a Reinforced Concrete Frame
with Light Transverse Reinforcement. Ph.D. Dissertation, University of California, Berkeley.
Haselton, C. B., Liel, A. B., Taylor Lange, S., and Deierlein, G. G. (2008). Beam-Column Element Model
Calibrated for Predicting Flexural Response Leading to Global Collapse of RC Frame Buildings. Pacific
Earthquake Engineering Research Center, Berkeley, CA.
Kato, D., and Ohnishi, K. 2002. Axial Load Carrying Capacity of R/C Columns under Lateral Load Reversals. Third
US-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced concrete
Building Structures, Seattle, WA, PEER Report 2002/02.
Kuo, W. W. 2008. Study on the collapse behavior of nonductile reinforced concrete frames subjected to earthquake
loading. Ph.D. Dissertation, Department of Construction Engineering, National Taiwan University of
Science and Technology, Taipei, Taiwan, 542pp.
Li, Y. 2012. The shaking table tests column database and evaluation of drift capacity models for non-ductile
columns. Master Thesis, University of British Columbia.
Lynn, A.C. 2001. Seismic evaluation of existing reinforced concrete building columns. Ph.D. Dissertation,
Department of civil and Environmental Engineering, University of California, Berkeley.
Panagiotakos T.B. and Fardis, M.N., 2001 "Deformations of RC Members at Yielding and Ultimate", ACI Structural
Journal, 98(2), 135-148.
Sezen, H. 2002. Seismic response and modeling of reinforced concrete building columns, Ph.D. Dissertation,
Department of Civil and Environmental Engineering, University of California, Berkeley.
Sezen, H. and Moehle, J. 2004. Shear strength model for lightly reinforced concrete columns. ASCE Journal of
Structural Engineering, 130(11), 1692-1703.
Shin, Y. B. 2007. Dynamic Response of Ductile and Non-Ductile Reinforced Concrete Columns. Ph.D. Dissertation,
University of California, Berkeley.
Su, R.S. 2007. Shake table tests on reinforced concrete short columns failed in shear. Master thesis, Department of
Construction Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, 195pp.
Wu C.L., Yang Y.S. and Loh C.H. 2006. Dynamic Gravity Load Collapse of Non-ductile RC Frames I :
Experimental Approach. 8th US National Conference on Earthquake Engineering - the 100th Anniversary
Earthquake Conference Commemorating the 1906 San Francisco Earthquake, San Francisco, Moscone
Center, April 18-22.
Yavari, S., Elwood, K. J. and Wu, C. L. 2009. Collapse of a non-ductile concrete frame: Evaluation of analytical
models. Earthquake Engineering and Structural Dynamics, 38(2), 225-241.
Yavari, S. 2011. Shaking table tests on the response of reinforced concrete frames with non-seismic detailing. Ph.D.
Dissertation, University of British Columbia.
2.16
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
2.17
@Seismicisolation
@Seismicisolation
Y. Li et al.
2.18
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
Note: a = the shear span (equals to L/2 for fixed-fixed columns); b = width of column section; c = clear cover (measured from the outer face of the concrete to
outer face of longitudinal reinforcement); D = height of column section; dl = the diameter of longitudinal reinforcement; dt = the diameter of transverse
reinforcement; fc’= concrete compressive stress at test day; fyl = yield stress of longitudinal reinforcement; fyt = yield stress of transverse reinforcement; L= clear
height of the column; s = spacing of transverse reinforcement; ρl = longitudinal reinforcement ratio; ρt = volume transverse reinforcement ratio; Tie Type = the
type of the transverse reinforcement, “r135” means rectangular-shaped hoops with 135-degree hooks and “d90” means diamond-shaped transverse reinforcement
with 90-degree hooks; s = spacing of transverse reinforcement; ∆80% = column displacement at 80% of peak lateral resistance.
2.19
@Seismicisolation
@Seismicisolation
Y. Li et al.
Aspect
b D L f’c fyl s fyt Δ80%/L
Specimens Ratio ρl ρt P/Agf’c
(×10-2)
(mm) (mm) (mm) (MPa) (MPa) (mm) (MPa)
(a/D)
Bechtoula, Kono, Arai and Watanabe, 2002,
D1N30
250 250 625 37.6 2.50 0.0243 461 40 0.0050 485 0.30 3.96
Esaki, 1996 H-2-1/3 200 200 400 23.0 2.00 0.0253 363 40 0.0065 364 0.33 1.99
Esaki, 1996 H-2-1/5 200 200 400 23.0 2.00 0.0253 363 50 0.0052 364 0.20 2.52
Esaki, 1996 HT-2-1/3 200 200 400 20.2 2.00 0.0253 363 60 0.0065 364 0.33 2.49
Esaki, 1996 HT-2-1/5 200 200 400 20.2 2.00 0.0253 363 75 0.0052 364 0.20 2.94
Galeota et al. 1996, AB2 250 250 1140 80.0 4.56 0.0603 579 150 0.0054 579 0.30 4.02
Galeota et al. 1996, AB3 250 250 1140 80.0 4.56 0.0603 579 150 0.0054 579 0.30 3.74
Galeota et al. 1996, AB4 250 250 1140 80.0 4.56 0.0603 579 150 0.0054 579 0.20 4.06
Lynn et al. 1998, 3CMD12 457.2 457.2 1473 27.6 3.22 0.0303 331 305 0.0017 400 0.26 1.74
Nosho et al. 1996, No. 1 279.4 279.4 2134 40.6 7.64 0.0101 407 229 0.0010 351 0.34 1.61
Ohue et al. 1985, 2D16RS 200 200 400 32.0 2.00 0.0201 369 50 0.0048 316 0.14 3.83
Ohue et al. 1985, 4D13RS 200 200 400 29.9 2.00 0.0265 370 50 0.0048 316 0.15 1.78
Pujol 2002, No. 10-1-2.25N 152.4 304.8 686 36.5 2.25 0.0245 453 57 0.0073 411 0.08 3.21
Pujol 2002, No. 10-1-2.25S 152.4 304.8 686 36.5 2.25 0.0245 453 57 0.0073 411 0.08 3.14
Pujol 2002, No. 10-2-2.25N 152.4 304.8 686 34.9 2.25 0.0245 453 57 0.0073 411 0.08 3.21
Pujol 2002, No. 10-2-2.25S 152.4 304.8 686 34.9 2.25 0.0245 453 57 0.0073 411 0.08 3.17
Pujol 2002, No. 10-2-3N 152.4 304.8 686 33.7 2.25 0.0245 453 76 0.0055 411 0.09 3.19
Pujol 2002, No. 10-2-3S 152.4 304.8 686 33.7 2.25 0.0245 453 76 0.0055 411 0.09 3.05
Pujol 2002, No. 10-3-2.25N 152.4 304.8 686 27.4 2.25 0.0245 453 57 0.0073 411 0.10 3.05
Pujol 2002, No. 10-3-2.25S 152.4 304.8 686 27.4 2.25 0.0245 453 57 0.0073 411 0.10 3.22
Pujol 2002, No. 10-3-3N 152.4 304.8 686 29.9 2.25 0.0245 453 76 0.0055 411 0.10 3.13
Pujol 2002, No. 10-3-3S 152.4 304.8 686 29.9 2.25 0.0245 453 76 0.0055 411 0.10 3.15
Pujol 2002, No. 20-3-3N 152.4 304.8 686 36.4 2.25 0.0245 453 76 0.0055 411 0.16 3.33
Pujol 2002, No. 20-3-3S 152.4 304.8 686 36.4 2.25 0.0245 453 76 0.0055 411 0.16 3.36
Ramirez and Jirsa, 1980, 00-U 305 305 458 34.5 1.50 0.0245 374 65 0.0032 455 0.00 3.71
Saatcioglu and Ozcebe 1989, U1 350 350 1000 43.6 2.86 0.0321 430 150 0.0030 470 0.00 4.87
Saatcioglu and Ozcebe 1989, U2 350 350 1000 30.2 2.86 0.0321 453 150 0.0030 470 0.16 4.20
Saatcioglu and Ozcebe 1989, U3 350 350 1000 34.8 2.86 0.0321 430 75 0.0060 470 0.14 5.11
Sezen and Moehle No. 1 457.2 457.2 1473 21.1 3.22 0.0247 434 305 0.0017 476 0.15 2.56
Sezen and Moehle No. 2 457.2 457.2 1473 21.1 3.22 0.0247 434 305 0.0017 476 0.60 1.33
2.20
@Seismicisolation
@Seismicisolation
Assessment of ASCE/SEI 41 Concrete Column Provisions using Shaking Table Tests
Aspect
b D L f’c fyl s fyt Δ80%/L
Specimens Ratio ρl ρt P/Agf’c
(×10-2)
(mm) (mm) (mm) (MPa) (MPa) (mm) (MPa)
(a/D)
Sezen and Moehle No. 4 457.2 457.2 1473 21.8 3.22 0.0247 434 305 0.0017 476 0.15 3.49
Soesianawati et al. 1986, No. 4 400 400 1600 40.0 4.00 0.0151 446 94 0.0030 255 0.30 2.74
Takemura and Kawashima, 1997, Test 1 (JSCE-4) 400 400 1245 35.9 3.11 0.0158 363 70 0.0020 368 0.03 3.51
Takemura and Kawashima, 1997, Test 2 (JSCE-5) 400 400 1245 35.7 3.11 0.0158 363 70 0.0020 368 0.03 3.90
Takemura and Kawashima, 1997, Test 3 (JSCE-6) 400 400 1245 34.3 3.11 0.0158 363 70 0.0020 368 0.03 5.96
Takemura and Kawashima, 1997, Test 4 (JSCE-7) 400 400 1245 33.2 3.11 0.0158 363 70 0.0020 368 0.03 8.15
Takemura and Kawashima, 1997, Test 5 (JSCE-8) 400 400 1245 36.8 3.11 0.0158 363 70 0.0020 368 0.03 6.79
Umehara and Jirsa 1982, 2CUS 230 410 455 42.0 1.11 0.0301 441 89 0.0055 414 0.27 1.04
Wehbe et al. 1998, A1 380 610 2335 27.2 3.83 0.0222 448 110 0.0027 428 0.10 5.23
Wehbe et al. 1998, A2 380 610 2335 27.2 3.83 0.0222 448 110 0.0027 428 0.24 4.38
Wight and Sozen 1973, No. 00.033(East) 152.4 304.8 876 32.0 2.88 0.0245 496 127 0.0032 345 0.00 3.00
Wight and Sozen 1973, No. 00.033(West) 152.4 304.8 876 32.0 2.88 0.0245 496 127 0.0032 345 0.00 5.50
Wight and Sozen 1973, No. 25.033(East) 152.4 304.8 876 33.6 2.88 0.0245 496 127 0.0032 345 0.07 3.59
Wight and Sozen 1973, No. 25.033(West) 152.4 304.8 876 33.6 2.88 0.0245 496 127 0.0032 345 0.07 4.80
Wight and Sozen 1973, No. 40.048(East) 152.4 304.8 876 26.1 2.88 0.0245 496 89 0.0046 345 0.15 4.88
Wight and Sozen 1973, No. 40.048(West) 152.4 304.8 876 26.1 2.88 0.0245 496 89 0.0046 345 0.15 5.49
Wight and Sozen 1973, No. 40.067(East) 152.4 304.8 876 33.4 2.88 0.0245 496 64 0.0064 345 0.11 6.85
Wight and Sozen 1973, No. 40.067(West) 152.4 304.8 876 33.4 2.88 0.0245 496 64 0.0064 345 0.11 6.88
Xiao and Martirossyan 1998, HC4-8L16-T6-0.1P 254 254 508 86.0 2.00 0.0246 510 51 0.0075 449 0.10 6.40
Xiao and Martirossyan 1998, HC4-8L16-T6-0.2P 254 254 508 86.0 2.00 0.0246 510 51 0.0075 449 0.19 4.26
Note: a = the shear span (equals to L/2 for fixed-fixed columns); b = width of column section; D = height of column section; fc’= concrete compressive stress at
test day; fyl = yield stress of longitudinal reinforcement; fyt = yield stress of transverse reinforcement; L= clear height of the column; s = spacing of transverse
reinforcement; ρl = longitudinal reinforcement ratio; ρt = volume transverse reinforcement ratio; P = column axial load; Ag = area of column cross section; ∆80% =
column displacement at 80% of peak lateral resistance.
2.21
@Seismicisolation
@Seismicisolation
Y. Li et al.
2.22
@Seismicisolation
@Seismicisolation
SP-297—3
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
Frames
Jong‐Su Jeon, Laura N. Lowes, Reginald DesRoches
Synopsis: The results of laboratory testing and earthquake reconnaissance studies of reinforced
concrete frames indicate that beam‐column joint deformation can determine total frame deformation
and that for older buildings joint failure can result in frames losing lateral and gravity load carrying
capacity. Given the impact of joints on frame response, numerical models used to evaluate the
earthquake performance of reinforced concrete frames must include nonlinear joint models. This paper
reviews previously proposed models for simulating joint response with the objective of identifying
models that provide i) accurate simulation of response to earthquake loading, ii) simple implementation
in nonlinear analysis software, iii) numerical robustness, iv) computational efficiency, and v) objective
calibration procedures. Ultimately, no set of models was identified that met all of these requirements
for the range of geometric and design parameters found in reinforced concrete buildings in the United
States. With the objective of extending current modeling capabilities for interior joints, an experimental
data set was assembled. The data set was used to evaluate existing envelope response models and used
to calibrate cyclic response parameters for use with the preferred existing model. A new response
model for interior beam‐column joints is presented that meets the above requirements for the range of
geometric and design parameters found in reinforced concrete buildings in the United States.
Key words: Reinforced concrete, beam‐column joint, earthquake, seismic behavior, model,
simulation, bond, shear strength
3.1
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
Laura N. Lowes, FACI, is an Associate Professor of Civil and Environmental Engineering at the University
of Washington, Seattle. Her research and teaching focusses on nonlinear analysis of civil structures
subjected to severe loading including earthquake loading. She is a member of ACI Committees 369,
Seismic Repair and Rehabilitation, and 447, Finite Element Analysis of Reinforced Concrete Structures.
Reginald DesRoches is the Karen and John Huff School Chair, and Professor of Civil and Environmental
Engineering at the Georgia Institute of Technology. His research addresses earthquake engineering and
risk assessment of civil infrastructure. He is a member of the executive committee of the National
Academy of Sciences Disasters Roundtable, the Board of the Earthquake Engineering Research Institute,
and the advisory board for the Natural Disasters, Coastal Infrastructure and Emergency Management
Research Center.
Introduction
Numerical simulation to evaluate the earthquake performance of reinforced concrete frames requires
simulation of the nonlinear response of beam‐column joints. Laboratory testing of building frame
subassemblages shows that beam‐column joints with design details typical of pre‐19671 construction
exhibit severe stiffness and strength loss (Walker 2001; Leon 1990; Meinheit and Jirsa 1981; Pantelides
et al. 2002). Significant strength and stiffness loss have been observed also for joints designed to achieve
superior performance, such as those compliant with ACI 318 Code requirements (e.g., Park and Ruitong
1988; Durrani and Wight 1985). Finally, earthquake damage to reinforced concrete frames observed
during post‐earthquake reconnaissance efforts (e.g., Hall 1996; Holmes and Somers 1996) suggests also
that joint failure may result in structural collapse. Given the potential for beam‐column joints to
determine frame performance, simulation of nonlinear joint response is required for accurate
evaluation of the earthquake performance of older and modern reinforced concrete frame buildings.
Experimental investigation of the earthquake response of reinforced concrete beam‐column joints has
been the focus of numerous research studies during the last 40 years. Experimental testing has resulted
in data characterizing the response of frame subassemblages, which comprise column and beam
segments as well as joints and, in some cases, characterizing the local response of the beam‐column
joint. Experimental tests have addressed a range of joint configurations and designs, including interior
and exterior joints from two‐dimensional and three‐dimensional frames with detailing representing pre‐
1967 and modern construction. However, the objective of many of these laboratory tests was to
establish that modern building code requirements result in acceptable earthquake performance; thus,
many of these tests employed joint designs representative of modern construction, including transverse
reinforcement, moderate joint shear stress demands, and moderate bond stress demands for beam
1
In 1967 UBC provisions changed to include requirements that i) joint shear strength defined using the column shear strength
equation exceed joint shear demand, ii) transverse reinforcement be provided in the joint to achieve required shear strength,
iii) the sum of column flexural strengths exceed the sum of beam flexural strengths, and iv) splices in beam and column
longitudinal reinforcement be located away from the joint. Similar revisions were made to the ACI Code in 1971. These Code
changes are reflected in drawings for buildings designed for construction on the West Coast from 1923 – 1979 (Mosier 2000).
3.2
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
longitudinal reinforcement. Fewer tests have investigated the earthquake response of older joints with
no transverse reinforcement, high shear stress demands and high bond stress for beam longitudinal
reinforcement anchored in the joint. Additionally, most experimental investigations have employed
interior beam‐column joint subassemblages from two‐dimensional frames (the test specimen is
cruciform‐shaped and comprises a column that is continuous through the joint, a beam that is
continuous through the joint, and the joint region with no out‐of‐plane beam segments). Fewer
investigations have employed interior joints from three‐dimensional frames (the test specimen
comprises a column that is continuous through the joint, a beam that is continuous through the joint, a
second orthogonal beam that is continuous through the joint or terminates in the joint, and the joint
region). A very few tests have considered the behavior of exterior corner joints in three‐dimensional
frames (the test specimen comprises a column that is continuous through the joint, two orthogonal
beam segments that terminate in the joint, and the joint region). Despite the fact that a plethora of
experimental data do not exist for all joint geometries and designs, sufficient data exist to enable
calibration of numerical models for use in evaluating the earthquake performance of concrete frames.
Given the impact on frame response of joint flexibility and strength loss, many previous studies have
addressed the development of beam‐column joint models for use in assessing the earthquake behavior
and performance of concrete frames. Previously proposed models have ranged in sophistication from i)
“super‐elements” comprising multi‐dimensional continuum elements in which model response is
defined purely by fundamental material properties and joint geometry (Pantazopoulou and Bonacci
1994; Mitra 2007; Baglin and Scott 2000), to ii) “super‐elements” comprising one‐dimensional springs
and employing simplifying assumptions about joint response mechanisms (Lowes and Altoontash 2003;
Altoontash 2004; Mitra and Lowes 2007; Tajiri et al. 2006), to iii) rotational hinge models calibrated to fit
experimental data (Otani 1974; Anderson and Townsend 1977). In recent years, validation and
calibration efforts have employed larger and larger data sets; for example for interior joints, Mitra and
Lowes (2007), Shin and LaFave (2004), and Jeon (2013) employed data from 57, 26 and 124 tests,
respectively. Unfortunately, however, previous research has not resulted in a model or set of models
that fully meets the needs of the earthquake engineering community with respect to modeling accuracy
and precision, computational efficiency and robustness, and ease of use for the full range of beam‐
column joint configurations found in the building inventory and for the case of dynamic cyclic as well as
quasi‐static monotonic loading.
1. Accurate simulation of the fundamental characteristics of joint response to earthquake loading.
Multiple models may be required to provide accurate response simulation for a range of joint
designs and configurations. Simulation of response through significant loss of joint shear
strength was considered to be an important attribute; simulation of the loss of joint axial load
carrying capacity was not considered to be an important attribute. In the laboratory, beam‐
column joint subassemblages have typically maintained moderate axial loads following
significant loss of joint shear strength. A recent study by Hassan (2011) found that for joints with
beam‐to‐column depth ratios less than 2.5, joint axial load failure was unlikely to precede
column axial load failure.
2. Simple implementation of the model in OpenSees using existing element formulations and
material models. The OpenSees software platform was chosen for use because it is free, widely
3.3
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
used within the earthquake engineering research community, includes element and material
models that can readily be employed to simulate joint response, and provides efficient and
robust solution algorithms to support analysis of building frames subjected to quasi‐static and
dynamic loads. In particular, OpenSees includes joint elements and 1D spring elements that can
be used to simulate joint kinematics and material models that are capable of simulating the
stiffness and strength loss and pinched hysteretic response exhibited by joints subjected to
cyclic loading.
3. Numerical robustness and stability, so that analyses are not hampered by “failures to converge”,
and computational efficiency.
4. Objective calibration procedures, so that the model‐building process is as efficient as possible
and modeling parameters do not need to be adjusted to define the model for a specific beam‐
column joint with a specific set of design characteristics.
To accomplish the above objective, previously proposed models were reviewed and evaluated with
respect to the above criteria. As needed, model development activities were accomplished to fill gaps in
the capabilities of existing models.
Using a four‐node, finite‐area joint element, joint behavior is typically simulated through the action of
multiple 1D spring elements that compose the joint element (Altoontash 2004; Youssef and Ghobarah
2001; Shin and LaFave 2004; Lowes and Altoontash 2003; Mitra and Lowes 2007). However, the joint
element may comprise only a single rotational spring (Celik 2007; Park 2010; Jeon 2013) or a mesh of
continuum elements (Baglin and Scott 2000; Hegger et al. 2004; Deaton 2013; Mitra 2007). Joint super‐
elements that comprise more than one spring or continuum element typically require an intra‐element
solution2. Typically the user has minimal control of the intra‐element solution algorithm; as such, the use
of super‐elements introduces the potential for numerical problems and reduced numerical robustness
and efficiency. For 2D analysis, element nodes are typically located at the joint‐beam and joint‐column
interface; thus, no additional modeling effort is required to ensure that inelastic flexural action in beams
and columns occurs at the beam‐joint / column‐joint interface and that beam and column elements do
not contribute flexural deformation within the joint region. Appropriate element formulations are
required to enable simulation of geometric nonlinearity.
2
The OpenSees joint elements are essentially small structural subassemblages within the global structural model. These
subassemblages comprise multiple nodes and multiple nonlinear springs that are not (explicitly) part of the global structural
model. For each iteration to establish equilibrium of the global structural model, it is necessary to solve for equilibrium of the
joint subassemblage model. While the user has control over the global solution algorithm, with the ability to change the
tolerance or step size to improve the rate of convergence, the user has no control over the solution algorithm used to solve the
joint subassemblage model. Thus, the user has limited ability to continue the analysis when a “failure to converge” occurs
within the joint element.
3.4
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
Using a two‐node, zero‐length rotational joint spring (Figure 1 and 2), joint response is simulated
through the action of a single rotational spring located at the centerline intersection of the beam and
column. Since the beam and column elements extend to the center of the joint, it is necessary to add
rigid links or rigid‐end‐offsets to the ends of beams and columns, within the joint volume, to ensure that
inelastic flexural action in beams and columns occurs at the beam‐joint / column‐joint interface and that
beam and column elements do not contribute flexural deformation within the joint region. The
rotational spring element is appropriate for use with linear and nonlinear geometric transformations.
The rotational joint spring requires a 1D joint moment versus joint rotation response model. Joint
response models are discussed below; however, most of these models define joint shear stress versus
strain. Since joints are not subjected to pure shear loading at the geometric perimeter of the joint, Celik
(2007) provides equations relating shear stress to joint moment. Joint shear deformation may be
assumed equal to joint rotation.
Two‐node, zero‐length slip elements (Figure 2) may be employed in combination with joint elements to
simulate slip of longitudinal reinforcement in the joint region. Typically the joint element is used to
simulate shear deformation of the finite‐area joint region and the slip element is used to simulate
stiffness and strength loss due to bond failure for beam or column reinforcement anchored in the joint.
At one location, multiple independent zero‐length elements or a single section element may be used to
simulate load‐deformation response for transfer of moment, axial load and shear.
Two nodes
connected by Steel
Concrete
zero-length stress-slip
stress-slip
Nonlinear fiber- rotational spring model
model
Rotational spring type beam-column Slip section
elements
Figure 1: OpenSees Model of a 2D Interior Joint Figure 2: OpenSees Model of a 2D Exterior Joint
Subassemblage Subassemblage
Figure 2 shows a zero‐length barslip fiber‐section model located at the beam‐joint interface of an
exterior joint. These models have been used successfully to simulate rotation due to slip of
reinforcement anchored in footings (Berry 2006; Oyen 2006) and are consistent with the barslip springs
introduced into joint models by Youssef and Ghobarah (2001) and Mitra and Lowes (2007). The barslip
section model may be used to simulate deformation due to barslip for interior joints with continuous
reinforcement, but is likely most relevant for joints in which beam longitudinal reinforcement
terminates within the joint with a short straight anchorage. Berry (2006) provides general guidance on
defining the barslip model; Mitra and Lowes (2007) provides recommendations for bond strength within
joints and cyclic response parameters.
3.5
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
50 40 Center Joint
Center Joint Cracking Cracking
Column Shear (kips)
20
0
0
-50 -20
-40
-100 -60
-80
-150
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 -100
Drift (%) -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Drift (%)
Figure 3: Column shear versus drift for a joint‐ Figure 4: Column shear versus drift for a beam‐
controlled specimen (PEER 4150 from Walker controlled specimen (PEER 0995 from Alire (2002)
(2001) joint = 0%, = 47 , = 46 ). joint = 0%, = 8 , = 32 ).
Numerous studies have investigated the parameters that affect joint response to earthquake loading
(Kim and LaFave 2009; Mitra 2007). Kim and LaFave (2009), using a data set comprising 212 interior 2D
and 3D joint test specimens (3D specimens were subjected to unidirectional lateral loading), identifies
the following as the primary parameters controlling joint strength: presence of out‐of‐plane beams,
beam eccentricity, normalized joint transverse reinforcement ratio, normalized beam reinforcement
ratio, and concrete compressive strength. Birely et al. (2012), considering a data set comprising 45
interior 2D joint test specimens, proposes that joint versus beam controlled response is determined by
joint shear stress demand per ACI 352R‐02, , and joint bond stress demand assuming beam bars yield
on both sides of the joint, (Figure 5). In Figure 5, joint demands are defined as follows:
3.6
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
Eq. 1
and
Eq. 2
It should be noted that the above discussion of interior joint
Figure 5: Relationship of ductility
behavior does not consider the case of interior joints for
classification to design shear stress
which beam longitudinal reinforcement terminates within
and bond stress.
the joint. Prior to 19671, anchorage for beam longitudinal
reinforcement was often determined on the basis of gravity loads with the result that beam bottom
reinforcement often terminated within the joint with short (6 in. typ.) embedment lengths.
Experimental testing (e.g., Pessiki et al. 1990) shows that these joints typically exhibit strength loss prior
to beam yielding due to pull‐out of the reinforcement. Exterior joints have similar design details and
simulation of this type of response is discussed below.
3.7
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
Table 1: Response Models for Interior Beam‐Column Joints in 2D Frames
Elem.
Reference C/M1 Defining Characteristics Strengths Limitations
Type2
Mitra and C FA ‐ Model includes vs. for joint ‐ Validated using a large data set ‐ Complicated to define model
Lowes core derived from concrete strut (57 specimens). parameters for specific set of joint
(2007) model as well as bar‐slip springs ‐ Developed for OpenSees. characteristics.
derived from bond stress ‐ Requires intra‐element solution.
capacity.
Andreson et C RS vs. model only. ‐ Developed using data from test ‐ Model definition is relatively
al. (2008) ‐ Developed using experimental specimens with no trans. complicated.
vs. data. reinforcement. ‐ May not possible to implement
‐ Joint strength is 0.95ACI. ‐ Demonstrated to provide model in OpenSees without
‐ Joint strength degrades with accurate response of specimen approximations.
cyclic loading once maximum with limited transverse ‐ Because beam strength
strength is reached; rate of reinforcement. determines joint strength,
degradation depends on ACI. accurate simulation of post‐peak
response and impact of ACI on this
response is critical.
‐ Model was validated using joint
vs. data, not subassemblage
response data.
Kim and M RS vs. model. ‐ Developed using large data set ‐ Implementation in OpenSees
LaFave ‐ Developed using experimental (212 specimens). ‐ Defines requires definition of cyclic
(2009) vs. data. response for joints with a wide response quantities.
‐ Quadrilinear envelope, trilinear range of design parameters. ‐ Model calibration, for individual
to max. strength and then joint design, requires calculation
descending branch to 90% max. of a number of design parameters.
strength. ‐ Model only predicts response
‐ Envelope points are functions through 10% strength loss.
of multiple design parameters
(geometry, joint reinforcement
ratio, concrete strength, etc.).
Birely et al. M RS Joint moment versus joint ‐ Developed using large data set ‐ Requires definition of post‐peak
(2012) deformation model (i.e. includes (45 specimens). response.
all non‐frame member ‐ Simple. ‐ Requires definition of cyclic
deformation). ‐ Joint model includes all non‐ response quantities.
‐ Bilinear to max. frame member deformation, so no ‐ Model calibration, for individual
strength/failure; does not barslip model required; model was joint design, requires calculation
include post‐peak response. developed using subassemblage of beam yield moments.
test data and OpenSees lumped‐
plasticity elements used to model
beams and columns.
‐ Model validation demonstrates
that model can predict joint failure
versus beam‐yielding; specimens
exhibiting limited ductility are
predicted to exhibit joint failure.
Jeon (2013) C RS ‐ vs. model. ‐ Developed using large data set ‐ Joint shear strain at maximum
‐ Developed using experimental (375 specimens). strength and residual strength are
vs. data. ‐ Validated using large data set not defined.
‐ Mult‐linear envelope, trilinear (124 specimens).
to max. strength, descending ‐ Defines response for joints with a
branch to 20% max. strength, wide range of design parameters.
residual strength of 20% max.
strength.
‐ Envelope points are functions
of multiple design parameters
(geometry, joint reinforcement
ratio, concrete strength, etc.).
Notes: 1. Model defines cyclic response history and is appropriate for cyclic analysis, C, or model defines envelope to cyclic response history
and is appropriate for analysis under monotonic loading. 2. Model is appropriate for use with a 4‐node, finite‐area element, FA, or zero‐length
rotational spring element, RS.
3.8
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
On the basis of the information provided in Table 1, the models by Birely et al. (2012) and Kim and
LaFave (2009) were identified as warranting further evaluation and, potentially, development. Both
models provide a relatively simple approach for defining a shear strength versus deformation envelope
that is appropriate for use for a range of design configurations and that can provide simulation of ductile
as well as non‐ductile response mechanisms. However, both models require further evaluation. The Kim
and LaFave model was developed using experimental joint shear stress versus strain data and was not
validated using subassemblage data; thus, there is the potential for the model to neglect bar slip or
other deformation not captured by laboratory instrumentation of the joint‐core region. The Birely
model was developed and calibrated using beam‐column joint subassemblage data and a numerical
model such as shown in Figure 1; thus, simulated joint deformation includes all frame subassemblage
deformation not attributed to beams or columns. However, the Birely model defines a hardening type
response to the point of maximum strength with definition of post‐peak response left to the user;
because initial strength loss may not be extremely rapid (Figure 3 and 4), there is the potential for the
model to neglect and/or provide highly inaccurate results for the post‐peak response history. To enable
use of the models for analysis of frames subjected to dynamic earthquake excitation, both models
require definition of the post‐peak response envelope to the point of residual joint strength and of cyclic
response quantities.
Figure 6: Column shear versus drift for an Figure 7: Column shear versus drift for a shear‐
anchorage‐controlled exterior joint specimen (Unit controlled exterior joint specimen (Unit 6 from
2 from Pantelides et al. (2002), joint = 0%, max Pantelides et al. (2002) joint = 0%, max = 11.3
=7 (up) = 10.6 (down), beam , beam bar anchorage is 180‐degree hooks,
bottom bar anchorage length = 6 in., column axial column axial load = 0.25Agfc. Joint design is not
load = 0.25Agfc. Joint design is not compliant with compliant with ACI Code requirements for SMF).
ACI Code requirements for SMF).
3.9
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
Table 2: Response Models for Exterior Beam‐Column Joints
Elem.
Reference C/M1 Defining Characteristics Advantages Limitations
Type2
Sharma et al. M RS ‐ Model includes vs. for joint ‐ Validated using ‐ Model cannot be directly applied to joints with
(2011) core derived from Priestley data from 12 joint straight anchorage lengths other than 6 in.
(1997) principal tensile stress subassemblage tests; ‐ Proposed implementation / spring model is
model. specimens have different from that shown in Figure 2.
‐ Max tensile stress is empirically range of design ‐ Validation was not in OpenSees, used zero‐
calibrated; one strength for well‐ characteristics and length plastic hinges for beams and columns and
anchored bars and one strength exhibit anchorage used a spring configuration different from that
for bars with 6 in. straight and joint failure. in Figure 2.
anchorage length. ‐ Simple. ‐ Requires definition of cyclic response
‐ Does not require quantities; achieving accurate simulation may be
barslip section. challenging given different strengths in +/‐
directions.
Kim and M RS & ‐ Defines response for joints with ‐ Developed using ‐ Implementation requires definition and use of
LaFave (2009) BSS well‐anchored beam longitudinal large data set (111 barslip section model to simulate strength loss
reinforcement (90‐ or 180‐degree specimens). due to anchorage failure.
hooks). ‐ Includes ‐ Requires definition of cyclic response
‐ vs. model. descending branch quantities.
‐ Developed using experimental of response curve, so ‐ Model calibration, for individual joint design,
vs. data. no additional requires calculation of a number of design
‐ Quadrilinear envelope, trilinear modeling is required parameters.
to max. strength, descending to simulate strength ‐ Requires definition of post‐peak response;
branch to 90% max. strength. loss. model only predicts response through 10%
‐ Envelope points are functions of strength loss.
multiple design parameters
(geometry, joint reinforcement
ratio, concrete strength, etc.).
Mitra and M,C RS & ‐ vs. models for interior beam‐ ‐ Simple. ‐ Requires validation.
Lowes (2007), BSS column joints. ‐ Allows for use of ‐ Requires definition and use of barslip section
Anderson et ‐ Assume that these can be same model for model to simulate strength loss due to
al. (2008), applied to exterior joints. interior and exterior anchorage failure.
Birely et al. ‐ Employ barslip section model to joints. ‐ Requires definition of cyclic response
(2012) simulate pullout of beam bars quantities.
with short, straight anchorage.
Hassan C RS & ‐ vs. model to simulate joint ‐ Simple. ‐ Not validated for exterior joints in which beam
(2011) BSS core response combined with M‐ ‐ Developed for use bars have inadequate straight anchorage.
model to simulate barslip. in OpenSees. ‐ Validation effort for 2D joints is limited.
developed for exterior joints in ‐ Validated using
3D frames and applied to 2D subassemblage
joints. model and test data.
Park (2010) C RS & ‐ M vs. model to simulate joint ‐ Simple. ‐ Does not include simulation of barslip and
BSS response. ‐ Developed for use cannot simulate failure due to pullout of beam
developed for exterior joints in in OpenSees. bars with short anchorage lengths.
3D frames and applied to 2D ‐ Validated using ‐ Validation effort for 2D joints is limited.
joints. subassemblage
model and test data.
Jeon (2013) C RS ‐ vs. model. ‐ Developed using ‐ Joint shear strain at maximum strength and
‐ Developed using experimental large data set (340 residual strength are not defined.
vs. data. specimens).
‐ Mult‐linear envelope, trilinear ‐ Validated using
to max. strength, descending large data set (72
branch to 20% max. strength, specimens).
residual strength of 20% max. ‐ Defines response
strength. for joints with a wide
‐ Envelope points are functions of range of design
multiple design parameters parameters.
(geometry, joint reinforcement
ratio, concrete strength, etc.).
Notes: 1. Model defines cyclic response history and is appropriate for cyclic analysis, C, or model defines envelope to cyclic response history
and is appropriate for analysis under monotonic loading, M. 2. Model is appropriate for use with a zero‐length rotational spring element, RS, or
requires use of a rotational spring element in combination with one or more bar‐slip springs, RS & BSS.
3.10
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
Knee‐Joints in 2D Frames
A knee joint in a 2D frame represents the intersection of an exterior column and the beam at the top of
a frame. The knee joint on the left side of a frame is modeled in OpenSees by removing the top column
and its associated nodes from the model shown in Figure 2.
Given that the top story of a frame is typically not a critical region of the frame, with respect to loss of
lateral load carrying capacity or frame collapse, accurate simulation of response for these components
should be considered less important than for interior or exterior joints. Relatively little research has
addressed the earthquake response of knee‐joints in older or modern buildings frames. For knee‐joints,
joint shear demands are likely similar in magnitude to those in exterior joints; for older joints,
inadequate anchorage detailing for beam and/or column bars terminating in the knee joint likely
controls response. Kim and LaFave (2009) provides a vs. model for the envelope to the cyclic
response history for knee joints (18 specimens) with transverse reinforcement and adequate anchorage
detailing for beam and column reinforcement; recommendations are provided for extending this model
to the case of no transverse reinforcement with adequate anchorage detailing, but no data exist for
validation. Barslip section models at the beam‐joint and column‐joint interfaces are recommended for
simulating the response of older knee‐joints with inadequate straight bar anchorages for beam and/or
column longitudinal reinforcement anchored in the joint. Development activities must address
validation of models using subassemblage data, validation of bar‐slip models and definition of cyclic
response parameters.
Joints in 3D Frames
In 3D frames, the interior, exterior and knee joints described above are extended to include continuous
or terminating beams in the out‐of‐plane direction and are subjected to loading in the out‐of‐plane
direction. The addition of beams framing into the joint is typically considered to improve joint core
response by providing additional confinement of the core (Kitayama et al. 1991; Oka and Shiorhara
1992); however, 3D loading may increase joint damage and reduced joint deformation capacity (Leon
and Jirsa 1986). Relatively few experimental studies have employed joint subassemblages from 3D
frames subjected to bidirectional lateral. Model development activities have employed data from 3D
joint tests to calibrate 2D joint response models (i.e. models for use in 2D frames subjected to
unidirectional lateral loading); these models are included in Tables 1 and 2. Few models for use in
simulating the response of joints in 3D frames subjected to bidirectional lateral loading for which
response in orthogonal lateral directions is coupled do not exist.
To enable accurate simulation of the response of beam‐column joints in 3D frames, extensive
experimental testing and model development are required. Existing models are reviewed below.
3.11
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
only a few data exist for validation of the model. For the case of older interior beam‐column joints with
three beam segments framing into the joint (i.e. an interior joint in an exterior frame), inadequate
anchorage of beam reinforcement may control response. In this case, the model by Kim and LaFave
(2009) can be supplemented by a barslip section model at the joint‐beam interface.
The response of the joint subassemblages was simulated using OpenSees and a model configuration
such as shown in Figure 1. Jeon (2013) provides details of the model. A zero‐length rotational spring
element and the Pinching4 material model were used to model the joint. Lateral and vertical
displacements were constrained to be equal for duplicate nodes located at the center of the joint.
Elastic beam‐column elements with relatively large stiffnessess were introduced within the joint area.
The beamWithHinges element was used to model beams and columns. This element is a force‐based
element for which a linear moment distribution and constant shear and axial load distribution are
3.12
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
Table 3: Data set used in this study
Test program Specimen (√psi) Error estimate (Eτ)
(√psi) Jeon Kim and LaFave Birely Jeon Kim and LaFave Birely
1 Alire (2002) PEER0995 11.1 13.0 12.8 12.4 1.18 1.16 1.12
2 PEER1595 16.5 15.7 14.6 17.2 0.95 0.88 1.04
3 PEER4150 25.4 19.7 17.9 22.0 0.78 0.71 0.86
4 Goto and Joh (1996) J‐OH 18.3 18.5 17.3 22.0 1.01 0.94 1.20
5 Ohwada (1984) LJO‐6 20.1 14.6 15.0 22.0 0.73 0.74 1.09
6 LJXY‐6 19.1 16.8 17.7 22.0 0.88 0.92 1.15
7 LJXY‐8 18.2 16.8 17.7 22.0 0.93 0.97 1.21
8 JO‐5 17.6 14.9 14.7 21.5 0.85 0.83 1.22
9 JXY‐3 21.8 16.4 17.3 21.5 0.75 0.80 0.99
10 Walker (2001) PEER14 10.9 11.0 11.8 11.6 1.02 1.09 1.07
11 CD1514 11.6 11.1 11.8 12.0 0.96 1.03 1.04
12 CD3014 11.2 10.8 11.6 10.2 0.97 1.04 0.91
13 PADH14 10.8 10.8 11.6 10.2 1.00 1.07 0.94
14 PEER22 14.6 15.2 14.3 15.5 1.04 0.99 1.07
15 CD3022 15.5 14.9 14.2 15.7 0.96 0.92 1.01
16 PADH22 15.7 15.1 14.4 15.8 0.96 0.92 1.00
17 Endoh et al. (1991) LA1 20.0 22.5 22.2 20.1 1.12 1.11 1.00
18 A1 20.1 22.6 22.7 20.5 1.12 1.13 1.02
19 Fujii and Morita A1 18.7 22.4 23.2 22.0 1.20 1.24 1.18
20 (1991) A3 18.7 22.4 23.2 22.0 1.20 1.24 1.18
21 A4 19.0 24.8 27.8 22.3 1.31 1.46 1.17
22 Goto and Joh (1996) J‐HH 20.3 21.1 23.7 30.1 1.04 1.17 1.49
23 Goto and Joh (2003) LM‐60 16.1 14.5 15.6 16.7 0.90 0.97 1.04
24 LM‐125 12.7 12.5 12.5 13.1 0.98 0.99 1.03
25 HM‐60 20.1 15.9 17.2 21.6 0.79 0.85 1.07
26 HM‐125 14.4 13.9 13.6 16.3 0.97 0.95 1.14
27 HH‐125 14.5 14.3 14.4 15.8 0.98 0.99 1.09
28 Kurose et al. (1991) J1 21.3 17.2 19.8 16.9 0.81 0.93 0.80
29 Matsumoto et al. J‐5 22.0 21.1 23.5 21.9 0.96 1.07 1.00
30 (2010) J‐10 20.4 20.3 21.8 22.0 0.99 1.07 1.08
31 Noguchi and OKJ‐1 20.4 22.8 24.9 21.2 1.12 1.22 1.04
32 Kashiwazaki (1992) OKJ‐4 21.6 21.0 23.5 21.2 0.98 1.09 0.98
33 OKJ‐5 21.2 23.8 27.1 21.9 1.12 1.28 1.03
34 OKJ‐6 21.5 22.7 25.7 21.9 1.05 1.20 1.02
35 Oka and Shiohara J‐5 22.9 22.8 23.6 21.9 0.99 1.03 0.96
36 (1992) J‐6 20.4 18.7 17.9 18.6 0.92 0.88 0.91
37 Ozaki et al. NO1 18.1 17.4 19.7 18.2 0.96 1.09 1.00
38 (2010) NO2 16.9 16.7 17.1 17.7 0.99 1.01 1.05
39 Raffaelle and Wight SP1 13.5 14.6 15.4 13.8 1.08 1.14 1.03
40 (1992) SP2 10.2 11.3 14.3 14.7 1.11 1.40 1.44
41 SP3 9.6 10.8 13.1 11.0 1.13 1.36 1.14
42 SP4 11.5 12.1 13.0 12.1 1.06 1.13 1.06
43 Shin and LaFave SL1 11.6 13.4 13.3 12.0 1.15 1.15 1.03
44 (2004) SL2 12.4 13.9 13.3 12.4 1.12 1.08 1.00
45 SL4 16.5 16.0 16.6 13.0 0.97 1.00 0.78
46 Takamori et al. NO1 14.4 14.4 14.4 14.1 1.00 1.00 0.98
47 (2006) NO2 15.1 15.1 15.3 15.0 0.99 1.01 0.99
48 Teng and Zhou S1 18.0 16.7 18.5 16.7 0.93 1.03 0.93
49 (2003) S2 17.7 16.5 17.3 16.6 0.93 0.98 0.94
50 S5 14.5 14.5 15.2 14.7 1.00 1.05 1.02
51 Teraoka (1997) NO01 22.8 20.0 23.1 22.2 0.88 1.01 0.97
52 NO04 25.5 25.3 26.1 22.2 0.99 1.03 0.87
53 NO07 24.2 24.7 27.7 22.2 1.02 1.14 0.92
54 NO08 24.2 24.6 29.1 22.2 1.02 1.20 0.92
55 NO09 21.2 22.5 23.5 22.2 1.06 1.11 1.05
56 NO10 20.9 24.8 28.3 22.2 1.19 1.36 1.07
57 NO35 17.0 18.6 19.4 15.5 1.10 1.14 0.91
58 NO36 17.3 20.7 23.6 17.8 1.20 1.37 1.03
Mean 1.01 1.06 1.04
COV 0.12 0.15 0.12
3.13
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
assumed along the length of the element. Inelastic flexural response is assumed to occur only in “plastic‐
hinge regions” at the ends of the element; for the system shown in Figure 1, hinges were introduced
only where the beam (column) framed into the joint. A fiber‐type discretization of the beam (column)
cross section and typical material response models were used to simulate inelastic response within the
hinge regions; a plastic hinge length equal to half the member depth was used. Outside the hinge
region, beam and column elastic flexural stiffness was defined per ASCE 41‐06 (Elwood et al. 2007).
Lateral load was applied via displacement control to the top of the column. A constant gravity load was
applied at the top of the column as required to simulate laboratory testing.
Figure 8 shows the response models by Jeon (2013), Kim and LaFave (2009) and Birely et al. (2012). The
Jeon model includes parameters defining cyclic response (not shown); the Kim and LaFave and Birely
models define only the envelope of the response and are not directly applicable for quasi‐static cyclic or
dynamic analysis. The Jeon model (Figure 8a) represents a simplification of the model by Anderson et al.
(2008); experimental data characterizing the response of interior joints from 2D and 3D frames were
used to develop an expression defining maximum joint shear strength, max, as a function of joint design
parameters and post‐peak stiffnesses, kdeg, as a function of joint transverse reinforcement configuration
(kdeg = ‐2max and kdeg = ‐1.2max for joints that are ASCE 41‐06 non‐compliant and compliant, respectively).
An expression was not developed for joint shear strain at maximum strength; instead measured
quantities were used in evaluating the model. Thus, the Jeon model is not appropriate for predictive
analysis. Figure 8b shows the Kim and LaFave model, which was developed using statistical observation
methods and an extensive experimental data base comprising interior joint specimens from 2D and 3D
frames. The database included relatively few joints without joint transverse reinforcement, and the
model may not provide accurate response for this class of joints. The response envelope for the Kim and
LaFave model is defined by maximum joint shear stress, max, and the joint shear strain at maximum
strength, γmax; expressions are provided defining these quantities as a function of joint material,
geometric and reinforcement parameters. The Birely model shown in Figure 8c. It was developed using
data from tests of interior joints from 2D frames; it does not simulate the impact of confinement
provided by out‐of‐plane beam segments framing into the joint as the Jeon and Kim and LaFave models
do. The Birely model defines a joint shear stress‐strain response envelope that is bilinear to maximum
strength; brittle post‐peak response is conservatively assumed and response is not defined beyond
maximum strength. Initial stiffnesses are a function of the concrete shear modulus, G; the transition
from initial to second stiffness occurs once beam longitudinal reinforcement yields and analysis of the
beams framing into the joint is required to determine this point. For all three models, equilibrium of the
joint is considered to convert joint shear stress versus strain response to joint moment versus rotation
response for use with the OpenSees model shown in Figure 1.
3.14
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
Models were evaluated on the basis of the accuracy with which subassemblage and joint response
quantities were simulated. Table 3 lists, for each specimen, the ratio of simulated to measured joint
shear strength, , as defined below. Data in Table 4 and Table 5, respectively, quantify the accuracy
with which subassemblage and joint response are simulated; the mean and coefficient of variation (COV)
of various error measures defined below are provided for all specimens in the data set. Note that
statistics provided in Tables 4 and 5 are similar when specimens are grouped as compliant and non‐
compliant per ASCE 41‐06 (Elwood et al. 2007). Figures 9 and 10Error! Reference source not found.
show, respectively, measured and simulated subassemblage and joint response envelopes for specimens
exhibiting joint‐ and beam‐controlled response. The following response quantities were considered in
model evaluation and are listed in Tables 3, 4 and 5:
1. Ratio of simulated to measured maximum joint shear stress, as controlled by either joint shear
,
failure or beam flexural strength. For a test specimen, this ratio is defined as where
,
τ , and τ , are the simulated and measured maximum joint shear stresses, respectively.
2. Ratio of simulated to measured drift at computed yield strength. This is used to evaluate the
accuracy with which initial stiffness is computed. For a test specimen, the ratio is defined as
,
where Δ , is the simulated drift at the computed initial yield of beam (or column)
,
longitudinal steel or, for specimens that did not reach the computed yield strength of the beams
(columns), to 90% of maximum strength and Δ , is the measured drift at the same load level.
3. Ratio of simulated to measured maximum column shear, as controlled by either joint shear
,
failure or beam flexural strength. For a test specimen, this ratio is defined as where
,
V , is the simulated maximum shear load carried by the column and V , is the measured
maximum load. Note that and that both are provided for completeness.
4. Ratio of simulated to measured drift at maximum subassemblage strength. For a test specimen,
,
this ratio is defined as , where Δ , is the simulated drift at V , and Δ , is
,
the measured drift at V , . Note that ∆, is not provided for the Jeon model as the model
does not predict joint shear strain at maximum shear strength; thus, measured shear strain at
maximum shear strength was used to define the model and , is approximately 1.0 for this
model.
5. Ratio of simulated to measured drift at 10% loss of subassemblage strength. For a test
,
specimen, this ratio is defined as , where Δ , is the simulated and Δ , is the
,
measured drift at 10% loss of subassemblage strength.
6. Normalized error in the simulated response envelope. For a given specimens, this is computed
as ∑ where n is the number of drift demands for which the
,
difference between simulated, , and measured, , column shear is computed and , is
the measured maximum column shear. For the Birely model and subassemblages predicted to
exhibit joint‐controlled response, this quantity was computed to the point of maximum
strength; for all for models and subassemglages, this quantity was computed to the point of 10%
strength loss.
7. Ratio of simulated to measured initial joint shear stiffness. For a test specimen, this ratio is
,
defined as where γ , and γ , are, respectively, the simulated and measured shear
,
strain at 90% of computed joint shear strength.
3.15
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
8. Ratio of simulated to measured joint shear strain at maximum joint shear stress. For a test
,
specimen, this ratio is defined as , where γ , is the simulated joint shear strain
,
at simulated maximum shear stress, τ , , and γ , is the measured joint shear strain at the
measured maximum shear stress, τ , . Note that γ , is not predicted by the Jeon model
and , is not provided for this model.
9. Ratio of simulated to measured joint shear strain at 10% loss of joint shear strength. For a test
,
specimen, this ratio is defined as , where γ , is the simulated and γ , is the
,
measured joint shear strain at 10% loss of maximum joint shear strength.
10. Normalized error in the simulated joint response envelope. For a given specimen, this is
computed as , ∑ where n is the number of joint shear strains for
,
which the difference between simulated, , and measured, , joint shear stress is computed
and , is the measured maximum joint shear stress. For the Birely model and
subassemblages predicted to exhibit joint‐controlled response, this quantity was computed to
the point of maximum strength; for all for models and subassemglages, this quantity was
computed to the point of 10% strength loss.
150 70
125 60
Column shear (kips)
50
100
40
75
30
50 Experiment 20 Experiment
Jeon (2013) Jeon (2013)
25 Kim and LaFave (2009) 10 Kim and LaFave (2009)
Birely et al. (2012) Birely et al. (2012)
0 0
0 1 2 3 4 5 0 1 2 3 4 5
Drift (%) Drift (%)
a) PEER4150 (Alire 2002), joint = 0%, max =25.4 b) PEER14 (Walker 2001), joint = 0%, max=10.9
and λ=45.6 . Strength is controlled by and λ=28.1 . Strength is controlled by
joint failure. beam yielding.
60 400
50
Column shear (kips)
300
40
30 200
20 Experiment Experiment
Jeon (2013) 100 Jeon (2013)
10 Kim and LaFave (2009) Kim and LaFave (2009)
Birely et al. (2012) Birely et al. (2012)
0 0
0 1 2 3 4 0 1 2 3 4 5
Drift (%)
Drift (%)
c) NO01 (Teraoka 1997), joint = 1.09%, max d) NO02 (Takamori et al. 2006), joint = 0.60%, max
=22.8 and λ = 33.0 . Strength is =14.9 and λ = 19.7 . Strength is
controlled by joint failure. controlled by beam yielding.
3.16
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
Figure 9: Observed and simulated response envelopes for interior joint subassemblages.
2 1
0.8
Joint shear stress (ksi)
0.6
1
0.4
Experiment Experiment
0.5 Jeon (2013) Jeon (2013)
0.2
Kim and LaFave (2009) Kim and LaFave (2009)
Birely et al. (2012) Birely et al. (2012)
0 0
0 0.005 0.01 0.015 0.02 0.025 0.03 0 0.005 0.01 0.015 0.02 0.025 0.03
Joint shear strain (rad)
Joint shear strain (rad)
b) PEER14 (Walker 2001), joint = 0%, max =10.9
a) PEER4150 (Alire 2002), joint = 0%, max =25.4 and λ=28.1 . Strength is controlled by
and λ=45.6 . Strength is controlled by beam yielding.
joint failure. 1.4
2
1.2
Joint shear stress (ksi)
1
Joint shear stress (ksi)
1.5
0.8
1 0.6
0.4 Experiment
Experiment Jeon (2013)
0.5 Jeon (2013) 0.2 Kim and LaFave (2009)
Kim and LaFave (2009) Birely et al. (2012)
Birely et al. (2012) 0
0 0.005 0.01 0.015 0.02 0.025 0.03
0 Joint shear strain (rad)
0 0.005 0.01 0.015 0.02 0.025 0.03
d) NO02 (Takamori et al. 2006), joint = 0.60%, max
Joint shear strain (rad)
c) NO01 (Teraoka 1997), joint = 1.09%, max =14.9 and λ = 19.7 . Strength is
=22.8 and λ = 33.0 . Strength is controlled by beam yielding.
controlled by joint failure.
Figure 10: Observed and simulated joint shear stress versus strain response envelopes for interior joints.
Table 4. Comparison of subassemblage load versus drift error values for the three models
3.17
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
Mean COV Mean COV Mean COV Mean COV Mean COV
Birely 1.12 0.41 1.04 0.12 0.48 0.48 0.29 0.50 0.077 0.46
Kim and LaFave 0.96 0.45 1.06 0.15 1.28 0.51 1.53 0.50 0.011 0.77
Jeon 1.12 0.35 1.01 0.12 NA NA 1.43 0.37 0.009 0.79
Simulated and observed response envelopes for the entire data set, such as those shown in Figures 9
and 10, and the data in Table 4 and Table 5 support the following observations and conclusions:
1. The Kim and LaFave and Jeon models, on average, provide acceptable accuracy for simulation of
subassemblage and joint response; though both models provide poor prediction of joint shear
strain at failure. Both the Kim and LaFave models provide relatively imprecise simulation of
response, with relatively large COVs for most response quantities (initial joint stiffness, drift
capacity, etc.).
2. The Birely model, on average, provides acceptable accuracy for simulation of subassemblage
and joint response up to maximum strength. The model provides acceptable accuracy of
simulation of response beyond maximum strength for beam‐controlled specimens. For joint‐
controlled specimens, the model makes the conservative assumption of a relatively brittle joint
failure mode and does not define response beyond the point of maximum joint strength. As
shown in Figures 9 and 10, many specimens exhibiting joint‐controlled failure do not exhibit
classical brittle response. Thus, on average, the model provides poor prediction of response
beyond maximum strength.
3. For all models, corresponding subassemblage and joint response quantities (e.g. initial
subassemblage stiffness and initial joint shear stiffness) are simulated with approximately the
same accuracy and precision; the COV for simulation of a particular joint response quantity is
typically equal to or slightly larger than the COV for the corresponding subassemblage response
quantity.
4. The initial stiffness of the subassemblage and joint region are accurately simulated by the Kim
and LaFave model (average error of 12% and ‐4% for the subassemblage and joint, respectively)
and are simulated with acceptable accuracy by the Birely and Jeon models (average errors of 12%
for the subassemblage and approximately 25% for the joint). For all of the models,
subassemblage stiffness is simulated with greater precision (COVs of 0.20, 0.23 and 0.27) than
joint stiffness (COVs of 0.35, 0.41 and 0.45). On the basis of these data, the Kim and LaFave
model is considered to be the best modeling approach to estimate the initial stiffness.
5. Maximum column (joint) shear strength is simulated accurately and precisely for all three
models. For simulation of column strength the Birely, Kim and LaFave, and Jeon models,
respectively, provide average errors of 4%, 7%, and 1% with respective COVs of 0.13, 0.14, and
0.11. Statistics are essentially the same for joint shear strength.
6. Subassemblage drift and joint shear strain at maximum strength are simulated poorly by the
Birely model (average errors of ‐41% and ‐52% for these two quantities) and relatively poorly by
the Kim and LaFave model (average errors of 29% and 28% for these two quantities). These
relatively large errors are attributed to the relatively low stiffness of the system in the vicinity of
maximum strength, which results in significant variation in both measured and simulated
deformation at maximum strength. The Jeon model does not predict this quantity, and
measured quantities were used to complete definition of the response envelope; thus no error
value is provided for this model.
7. Drift capacity of the subassemblage is simulated with acceptable accuracy by both the Joen and
the Kim and LaFave models (average errors of 7% and 18% for the two models); however joint
3.18
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
shear strain at subassemblage drift capacity is poorly predicted by the models (average errors of
43% and 53%). In all cases COVs are large (greater than 0.28). The Birely model does not
simulate response beyond the point of maximum joint strength; thus, drift capacity and joint
shear strain at drift capacity are, on average, relatively poorly.
Given the above evaluation of the Birely, Kim and LaFave, and Jeon models, it was decided that further
model development activities would employ the Kim and LaFave model. The Kim and LaFave model
supports predictive modeling, while the Jeon model, which employs measured joint shear strain values,
does not. The Kim and LaFave model also provides definition of the response envelope beyond the point
of maximum strength, while the Birely model does not.
Development of Cyclic Response Parameters for the Kim and LaFave Model for Interior Joints
Model development activities employed the Kim and LaFave model and focused on defining parameters
to support cyclic analysis using the OpenSees Pinching4 material model. Figure 11 shows the response
envelope and response path under cyclic loading as well as key parameters used in defining the
Pinching4 material model. For each test specimen, the Kim and LaFave model was used to define the
envelope points (ePd*, ePf*, eNd*, dNf*) in Figure 11, and cyclic response parameters (rDispP, rDispN,
rForceP and rForceN) in Figure 11 were determined that provided a least‐squares, best‐fit to the cyclic
column shear versus story drift data. Average values of the cyclic response parameters were computed
for the entire data set (124 specimens) used in Jeon (2013), for specimen group according to whether
joint transverse reinforcement was conforming or non‐conforming per ASCE 41‐06, for specimens
grouped according to whether or not joint transverse reinforcement was provided, and for specimens
grouped according to joint‐ or beam‐controlled response as predicted using Birely et al. (2012). In all
cases, the average value of the cyclic response parameters ranged from 0.19 to 0.27, with the averages
for the entire data set being rDispP = rDispN = rForceN = 0.24 and rForceP = 0.23. On the basis of these
results, rDispP = rDispN = rForceN = rForceP = 0.25 was chosen as the preferred values for use. Figures
12 and 13 show simulated and observed cyclic response histories for column shear versus specimen drift
(Figure 12) and joint shear versus joint strain (Figure 13). In these figures, simulated response was
generated using the previously described OpenSees modeling approach with the Pinching4 material
model used to define the cyclic response of the joint rotational spring, the Kim and LaFave model used
to define the envelope to the joint moment‐rotation history, and the Pinching4 cyclic response
parameters defined equal to 0.25. Pinching4 model parameters related to cyclic stiffness and strength
deterioration were not considered in this study. Comparison of simulated and observed cyclic response
histories, such as those shown in Figures 12 and 13, shows that the proposed Pinching4 cyclic response
parameters provide relatively accurate simulation of the observed “pinched” behavior of the specimens.
However, the data in Figures 12 and 13 show the potential for poor simulation of joint strain demands
and thus joint response as well as the potential for relatively accurate simulation of subassemblage
response to be achieved even when joint response is poorly simulated.
3.19
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
(dmax,f(dmax))
(dmin,f(dmin))
Figure 11: OpenSees Pinching4 uniaxial material model (Lowes and Altoontash 2003).
150 80
60
100
Column shear (kips)
−50 −20
−40
−100 Experiment
−60 Experiment
Simulation Simulation
−150 −80
−6 −4 −2 0 2 4 6 −6 −4 −2 0 2 4 6
Drift (%)
Drift (%)
a) PEER4150 (Alire 2002), joint = 0%, max =25.4 b) PEER14 (Walker 2001), joint = 0%, max =10.9
and λ = 45.6 . Strength is controlled by and λ = 28.1 . Strength is controlled by
joint failure. beam yielding.
60 400
40
Column shear (kips)
200
20
0 0
−20
−200
−40 Experiment Experiment
Simulation Simulation
−60 −400
−8 −6 −4 −2 0 2 4 6 8 −6 −4 −2 0 2 4 6
Drift (%) Drift (%)
c) NO01 (Teraoka 1997), joint = 1.09%, max d) NO02 (Takamori et al. 2006), joint = 0.60%, max
=22.8 and λ = 33.0 . Strength is =14.9 and λ = 19.7 . Strength is
controlled by joint failure. controlled by beam yielding.
Figure 12: Simulated and observed column shear versus story drift response histories for interior joint
subassemblage test specimens.
3.20
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
With the objective of advancing current modeling capabilities, three recently proposed models for
interior beam‐column joints (Jeon 2013; Kim and LaFave 2009; Birely et al. 2012) were evaluated using
an experimental data set comprising 58 subassemblage test specimens. Test specimens had a wide
range of material properties, geometric configurations and design parameters and were representative
of modern and older construction in the United States. The results of the evaluation include 1) all
models provide accurate and precise simulation of strength, 2) all models provide acceptably accurate
simulation of initial stiffness, but the Kim and LaFave provides relatively accurate simulation of initial
stiffness, and 3) the Jeon model provides accurate and the Kim and LaFave model provides acceptably
accurate simulation of drift capacity, but all models provide poor simulation of joint shear strain demand
at drift capacity (90% strength loss).
2 2
Joint shear stress (ksi)
1 1
0 0
−1 −1
Experiment Experiment
Simulation Simulation
−2 −2
−0.06 −0.04 −0.02 0 0.02 0.04 0.06 −0.09 −0.06 −0.03 0 0.03 0.06 0.09
Joint shear strain (rad)
Joint shear strain (rad)
a) PEER4150 (Alire 2002), joint = 0%, max =25.4 c) NO01 (Teraoka 1997), joint = 1.09%, max
and λ = 45.6 . Strength is controlled by =22.8 and λ = 33.0 . Strength is
joint failure. controlled by joint failure.
1 12
8
Joint shear stress (ksi)
Joint shear stress (ksi)
0.5
4
0 0
−4
−0.5
Experiment −8 Experiment
Simulation Simulation
−1 −12
−0.04 −0.03 −0.02 −0.01 0 0.01 0.02 0.03 0.04 −0.06 −0.04 −0.02 0 0.02 0.04 0.06
Joint shear strain (rad) Joint shear strain (rad)
b) PEER14 (Walker 2001), joint = 0%, max =10.9 d) NO02 (Takamori et al. 2006), joint = 0.60%, max
and λ = 28.1 . Strength is controlled by =14.9 and λ = 19.7 . Strength is
beam yielding. controlled by beam yielding.
3.21
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
Figure 13: Simulated and observed joint shear stress versus shear strain response histories for interior
joint subassemblage test specimens.
The Kim and LaFave model was chosen for use in development of a cyclic response model for interior
joints; this decision was based on the results of the model evaluation as well limitations of the Jeon and
Birely models (the Jeon model does not provide prediction of joint shear strain at maximum strength
and the Birely model does not predict joint response beyond maximum strength). Cyclic response
parameters for the OpenSees Pinching4 material model were determined using a least‐squares, best‐fit
method and cyclic column shear versus story drift data for the 58 specimens in the assembled data set.
Results show that the proposed cyclic response parameters provide accurate simulation of the
hysteretic response exhibited by beam‐column joint subassemblages tested in the laboratory. However,
results show also the potential for relatively poor simulation of joint shear strain demands and that
relatively accurate simulation of subassemblage response can be achieved with relatively poor
simulation of joint shear response.
The results of this study suggest that additional research is need to develop a suite of numerical models
that enable accurate simulation of the nonlinear response of beam‐column joints with a range of
geometric and design configurations found in the US building inventory.
Acknowledgements
This material is based in part upon work supported by the National Science Foundation under Grant
Number 1000700. Any opinions, findings, and conclusions or recommendations expressed in this
material are those of the authors and do not necessarily reflect the views of the National Science
Foundation.
References
Alire, D.A. (2002). “Seismic evaluation of existing unconfined reinforced concrete beam‐column joints,”
MS Thesis. University of Washington, Seattle.
Altoontash, A. (2004). “Simulation and damage models for performance assessment of reinforced
concrete beam‐column joints,” PhD Dissertation. Stanford University, CA.
Anderson, M., Lehman, D.E., Stanton, J. (2008). “A cyclic shear stress–strain model for joints without
transverse reinforcement.” Engineering Structures, 30: 941‐954.
Anderson, J.C. and Townsend, W.H. (1977). “Models for reinforced concrete frames with degrading
stiffness,” Journal of the Structural Division, ASCE, 103 (ST12): 1433‐1449.
Baglin, P.S. and Scott, R.H. (2000). “Finite Element Modeling of Reinforced Concrete Beam‐Column
Connections,” ACI Structural Journal, 97 (6): 886‐894.
Berry, M. (2006). “Performance Modeling Strategies for Modern Reinforced Concrete Bridge Columns,”
PhD Dissertation. University of Washington, Seattle.
Birely, A., Lowes, L.N., Lehman, D.E. (2012). “A model for the practical nonlinear analysis of reinforced
concrete frames including joint flexibility,” Engineering Structures 34: 455–465.
Celik, O.C. (2007). “Probabilistic Assessment of Non‐Ductile Reinforced Concrete Frames Susceptible to
Mid‐America Ground Motions,” PhD Dissertation. Georgia Institute of Technology.
3.22
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
Deaton, J.B. (2013). “Nonlinear finite element analysis of reinforced concrete exterior beam‐column
joints with nonseismic detailing,” PhD Dissertation. Georgia Institute of Technology.
Durrani, A.J. and Wight, J.K. (1985). “Behavior of interior beam‐to‐column connections under
earthquake‐type loading,” ACI Journal 82 (3): 343‐349.
Elwood, K.J., Matamoros, A.B., Wallace, J.W., Lehman, D., Heintz, J.A., Mitchell, A.D., Moore, M.A.,
Valley, M.T., Lowes, L.N., Comartin, C.D., and Moehle, J.P. (2007). “Update to ASCE/SEI 41 concrete
provisions,” Earthquake Spectra, 23 (3): 493‐523.
Hall, J., ed. (1996). Northridge Earthquake of January 17, 1994 Reconnaissance Report Vol 1 Oakland:
Earthquake Engineering Research Institute.
Hassan, W. (2011). “Analytical and Experimental Assessment of Seismic Vulnerability of Beam‐Column
Joints without Transverse Reinforcement in Concrete Buildings,” PhD Dissertation. University of
California, Berkeley.
Hegger, J., Sherif, A., and Roeser, W. (2004). “Nonlinear Finite Element Analysis of Reinforced Concrete
Beam‐Column Connections,” ACI Structural Journal, 101 (5), 604‐614.
Holmes, W. and Somers, P., eds. (1996). Northridge Earthquake of January 17, 1994 Reconnaissance
Report Vol 2. Oakland: Earthquake Engineering Research Institute.
Jeon, J.‐S. (2013). “Aftershock Vulnerability Assessment of Damaged Reinforced Concrete Buildings in
California,” PhD dissertation. Georgia Institute of Technology.
Kim, J. and LaFave, J. (2009). “Joint Shear Behavior of Reinforced Concrete Beam‐Column Connections
Subjected to Seismic Lateral Loading,” NSEL Report NSEL‐020, University of Illinois at Urbana‐
Champaign.
Kitayama, K., Otani, S., and Aoyama, H. (1991). “Development of design criteria for RC interior beam‐
column joints,” Design of Beam‐Column Joints for Seismic Resistance (SP‐123), American Concrete
Institute, Detroit, MI, 97‐113.
Leon, R.T. (1990). “Shear strength and hysteretic behavior of interior beam‐column joints,” ACI
Structural Journal, 87(1): 3‐11.
Leon, R. and Jirsa, J.O. (1986). “Bidirectional loading of R.C. Beam‐Column Joints,” Earthquake Spectra, 2
(3): 537‐564.
Lowes, L.N. and Altoontash, A. (2003). “Modeling Reinforced Concrete Beam‐Column Joints Subjected to
Cyclic Loading,” Journal of Structural Engineering, ASCE, 129 (12): 1686‐1697.
Meinheit, D.F. and Jirsa, J.O. (1981). “Shear strength of RC beam‐column connections,” Journal of the
Structural Division, ASCE, 107 (ST11): 2227‐2244.
Mitra, N. (2007). “An analytical study of reinforced concrete beam‐column joint behavior under seismic
loading,” PhD Dissertation. University of Washington, Seattle.
Mitra, N. and Lowes, L.N. (2007). “Evaluation, Calibration, and Verification of a Reinforced Concrete
Beam–Column Joint Model,” Journal of Structural Engineering, ASCE, 133 (1): 105‐120.
3.23
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
Moiser, W.G. (2000). “Seismic assessment of reinforced concrete beam‐column joints,” MS Thesis.
University of Washington, Seattle.
Oka, K. and Shiohara, H. (1992). “Tests of high‐strength concrete interior beam‐column‐joint
subassemblages,” 10th World Conference on Earthquake Engineering, Madrid, Spain, 3211‐3217.
Otani, S. (1974). “Inelastic analysis of reinforced‐concrete frame structures,” Journal of the Structural
Division, ASCE, 100 (ST7): 1422‐1449.
Oyen, P. (2006). “Evaluation of Analytical Tools for Determining the Seismic Response of Reinforced
Concrete Shear Walls,” PhD Dissertation. University of Washington, Seattle.
Pantazopoulou, S.J. and Bonacci, J.F. (1994). “On earthquake resistant reinforced concrete frame
connections,” Canadian Journal of Civil Engineering, 21 (2): 307–328.
Pantelides, C.P., Hansen, J., Nadauld, J., Reaveley, L.D. (2002). “Assessment of reinforced concrete
building exterior joints with substandard details,” PEER Report 2002/18, Pacific Earthquake Engineering
Center, University of California, Berkeley, CA.
Park, S. (2010). “Experimental and Analytical Studies on Old Reinforced Concrete Buildings with
Seismically Vulnerable Beam‐Column Joints,” PhD Dissertation. University of California, Berkeley.
Park, R. and Ruitong, D. (1988). “A Comparison of the behavior of reinforced concrete beam‐column
joints designed for ductility and limited ductility,” Bulletin of the New Zealand National Society of
Earthquake Engineering, 21 (4): 255‐278.
Pessiki, S.P., Conley, C.H., Gergely, P., and White, R.N. (1990). “Seismic behavior of lightly‐reinforced
concrete column and beam‐column joint details,” Technical Report NCEER‐90‐0014, National Center for
Earthquake Engineering Research, State University of New York, Buffalo.
Sharma, A., Eligehausen, R., and Reddy, G.R. (2011). “A new model to simulate joint shear behavior of
poorly detailed beam‐column connections in RC structures under seismic loads, Part I: Exterior joints,”
Engineering Structures, 33 (3): 1034‐1051.
Shin, M. and LaFave, J.M. (2004). “Modeling of Cyclic Joint Shear Deformation Contributions in RC Beam‐
Column connections to overall frame behavior,” Structural Engineering and Mechanics, 18 (5): 645‐669.
Tajiri, S., Shiohara, H., and Kusuhara, F. (2006). “A new macroelement of reinforced concrete beam‐
column joint for elasto‐plastic plane frame analysis,” Proceedings of the 8th U.S. National Conference on
Earthquake Engineering, San Francisco, California.
Takamori, N., Hayashi, K., Sasaki, S., and Teraoka, M. (2006). “Experimental study on full‐scale R/C
interior beam‐column joints (in Japanese),” Proceedings of the Japan Concrete Institute, 28 (2): 283‐288.
Teraoka, M. (1997). “A study on seismic design of R/C beam‐column joint in high rise frame structure (in
Japanese),” Research Report, Extra Issue No. 5, Fujita Institute of Technology, Tokyo, Japan.
Walker, S. (2001). "Seismic performance of existing reinforced concrete beam‐column joints," MS Thesis.
University of Washington.
Youssef, M. and Ghobarah, A. (2001). “Modeling of RC Beam—Column Joints and Structural Walls,”
Journal of Earthquake Engineering, 5 (1):93‐111.
3.24
@Seismicisolation
@Seismicisolation
Numerical Models for Beam‐Column Joints in Reinforced Concrete Building
References Used in the Database
Alire, D. 2002. "Seismic evaluation of existing unconfined reinforced concrete beam‐column joint," MS
Thesis. University of Washington.
Endoh, Y., Kamura, T., Otani, S., and Aoyama, H. (1991). “Behavior of RC beam‐column connections
using light‐weight concrete,” Transactions of the Japan Concrete Institute, 13: 319‐326.
Fujii, S. and Morita, S. (1991). “Comparison between interior and exterior RC beam‐column joint
behavior,” Design of Beam‐Column Joints for Seismic Resistance (SP123), American Concrete Institute,
Detroit, MI, 145‐165.
Goto, Y. and Joh, O. (1996). “An experimental study on shear failure mechanism of RC interior beam‐
column joints,” 11th World Conference on Earthquake Engineering, Paper No. 1194, Acapulco, Mexico.
Goto, Y. and Joh, O. (2003). “Experimental study on shear resistance of RC interior eccentric beam‐
column joints,” 9th East Asia‐Pacific Conference on Structural Engineering and Construction, Bali,
Indonesia, RSC‐170.
Kurose, Y., Guimaraes, G.N., Zuhua, L., Kreger, M.E., and Jirsa, J.O. (1991). “Evaluation of slab‐beam‐
column connections subjected to bi‐directional loading,” Design of Beam‐Column Joints for Seismic
Resistance (SP123), American Concrete Institute, Detroit, MI: 39‐67.
Matsumoto, T., Nishihara, H., Nakao, M., and Castro, J.J. (2010). “Performance on shear strength of
reinforced concrete eccentric beam‐column joints subjected to seismic loading,” Proceedings of the 9th
U.S. National and 10th Canadian Conference on Earthquake Engineering, Paper No. 1457, Toronto,
Canada.
Noguchi, H. and Kashiwazaki, T. (1992). “Experimental studies on shear performances of RC interior
column‐beam joints with high‐strength materials,” 10th World Conference on Earthquake Engineering,
Madrid, Spain, 3163‐3168.
Ohwada, Y. (1984). “Studies on reinforced concrete beam‐column joints under biaxial seismic load (2) (in
Japanese),” Summaries of Technical Papers of Annual Meeting Architectural Institute of Japan, 1869‐
1870.
Oka, K. and Shiohara, H. (1992). “Tests of high‐strength concrete interior beam‐column‐joint
subassemblages,” 10th World Conference on Earthquake Engineering, Madrid, Spain, 3211‐3217.
Ozaki, E., Nishio, A., Tajima, K., and Shirai, N. (2010). “Damage evaluation of RC beam column joints
based on the image measurement (in Japanese),” Proceedings of the Japan Concrete Institute, 32 (2):
301‐306.
Raffaelle, G.S. and Wight, J.K. (1995). “Reinforced concrete eccentric beam‐column connections
subjected to earthquake‐type loading,” ACI Structural Journal, 92 (1): 45‐55.
Shin, M. and LaFave, J.M. (2004) “Modeling of cyclic joint shear deformation contributions in RC beam‐
column connections to overall frame behavior,” Structural Engineering and Mechanics, 18 (5): 645‐669.
3.25
@Seismicisolation
@Seismicisolation
Jong-Su Jeon et al.
Takamori, N., Hayashi, K., Sasaki, S., and Teraoka, M. (2006). “Experimental study on full‐scale R/C
interior beam‐column joints (in Japanese),” Proceedings of the Japan Concrete Institute, 28 (2): 283‐288.
Teng, S. and Zhou, H. (2003). “Eccentric reinforced concrete beam‐column joints subjected to cyclic
loading,” ACI Structural Journal, 100 (2): 139‐148.
Teraoka, M. (1997). “A study on seismic design of R/C beam‐column joint in high rise frame structure (in
Japanese),” Research Report, Extra Issue No. 5, Fujita Institute of Technology, Tokyo, Japan.
Walker, S.G. (2001). “Seismic performance of existing reinforced concrete beam‐column joints,” MS
Thesis. University of Washington, Seattle.
3.26
@Seismicisolation
@Seismicisolation
SP-297—4
Evaluation of ASCE 41 Modeling Parameters for Slender Reinforced Concrete Structural Walls
Synopsis: ASCE/SEI 41-06 provides guidelines for evaluating the seismic adequacy of existing buildings. For
nonlinear dynamic analysis of a building, ASCE 41 provides modeling parameters to define the backbone curve for
the response of structural components. Seismic adequacy is then determined by comparing simulated response to
predetermined acceptance criteria. In the reinforced concrete (RC) community, there is interest in evaluating the
modeling parameters and acceptance criteria for RC components, and if deemed necessary, developing updated
values that reflect the current state of understanding of the seismic performance of RC components. For some
structural components (e.g. columns), large databases of experimental data can be used to develop empirical
acceptance criteria that reflect the behavior of the component. In the case of slender structural walls, relatively
limited tests have been conducted such that sufficient variation in critical design and loading characteristics
including shape, aspect ratio (elevation and cross-sectional), confinement, and axial load are not represented by
experimental data to justify use of an experimental database to develop acceptance criteria. Evaluation of this
limited set of experimental data indicates current ASCE 41 modeling parameters and acceptance criteria for flexure-
controlled walls is inappropriate, generally resulting in overprediction of wall deformation capacity at high axial
load ratios and underprediction at low axial load ratios and low shear demands. Although suitable for evaluation of
criteria, the data set is not sufficiently varied such that revised provisions can be developed. To overcome the lack of
sufficient experimental data, a parameter study was conducted to provide data to support development of updated
acceptance criteria. The parameter study was conducted using a modeling approach validated to provide accurate
simulation of flexural failures in slender reinforced concrete walls. Simulation results were used to develop
preliminary recommendations for revised modeling parameters for slender RC walls. An evaluation of these
simulation results and preliminary recommendations for revised flexure-controlled RC wall modeling parameters are
presented in this paper.
Keywords: ASCE 41; reinforced concrete; walls; flexure-control; acceptance criteria; plastic hinge rotation
4.1
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
Anna C. Birely is an assistant professor in the Zachry Department of Civil Engineering at Texas A&M University.
She received her BS from the University of Colorado and her MS and PhD in civil engineering from the University
of Washington. She is a member of ACI Committees 369, Seismic Repair and Rehabilitation; 374, Performance-
Based Seismic Design of Concrete Buildings; and 133, Disaster Reconnaissance. Her research interests include the
performance of reinforced concrete structures subjected to hazardous loads, including earthquakes.
Laura N. Lowes is an associate professor in the Department of Civil and Environmental Engineering at the
University of Washington. She received her BS from the University of Washington and her MS and PhD from the
University of California, Berkeley. She is a member of ACI Committee 369, Seismic Repair and Rehabilitation, and
Joint ACI-ASCE Committees 445, Shear and Torsions and 447, Finite Element Analysis of Reinforced Concrete
Structures. Her research interests include numerical modeling of concrete structures and earthquake engineering.
Dawn E. Lehman is an associate professor in the Department of Civil and Environmental Engineering at the
University of Washington. She received her BS from Tufts University and her MEng and PhD from the University
of California, Berkeley. She is a member of ACI Committees 341, Earthquake Resistant Concrete Bridges; 374,
Performance-Based Seismic Design of Concrete Buildings; and Joint ACI-ASCE Committee 352, Joints and
Connections in Monolithic Concrete Structures. Her research interests include seismic design of reinforced concrete
and composite structures.
1. INTRODUCTION
ASCE/SEI 41-06 Seismic Rehabilitation of Existing Buildings (ASCE-41) provides provisions to guide in the
seismic evaluation of existing buildings. The first generation of ASCE 41 was based extensively on the precursor
document FEMA 356 Prestandard and Commentary for the Seismic Rehabilitation of Buildings (FEMA-356, 2000).
A supplement was developed in 2007 (ASCE-41, 2007) to improve the accuracy of modeling parameters and
acceptance criteria for structural components. Ideally, such values are based on extensive experimental evidence
such that empirical models can be established for the range of component characteristics that may be found in
existing buildings. In anticipation of future editions of ASCE 41, modeling parameters and acceptance criteria for
some reinforced concrete components have been reevaluated. Evaluation of modeling parameters and acceptance
criteria for slender reinforced concrete walls are addressed in this paper. ASCE 41 classifies walls based on the
expected behavior being either shear-controlled and flexure-controlled. Provisions for shear-dominated walls were
updated in the most recent supplement; however, provisions for flexure controlled walls have remained largely
unchanged since the original FEMA 356 document, for which experimental justification of the modeling parameters
and acceptance criteria is not available. An evaluation of the provisions for slender (flexure-controlled) reinforced
concrete structural walls and recommendations for modifications to future versions of ASCE 41 are presented here.
The current ASCE 41 provisions classifies structural walls as being either flexure-controlled or shear-
controlled. While no explicit distinction is made between the two, commentary in Section C.7.1 indicates that
flexure-controlled walls should be taken as those with an aspect ratio (AR = height/length) of 3.0 or greater, shear-
controlled walls should be taken as those with an aspect ratio of 1.5 or less, and walls with intermediate response
should be considered to have response controlled by both flexure and shear. This assumption for wall behavior
based on the aspect ratio is commonly used to define walls as squat (shear-controlled) or slender, with an aspect
ratio of 1.5 or 2.0 typically used to indicate the transition between the two wall types. This aspect ratio has, for
experimental subassemblage tests, proved to be an accurate indicator of whether the failure mode of the wall is shear
(AR < 2.0) or flexure (AR > 2.0) dominated. Most such specimens are loaded with a single lateral force at the top of
the wall, and thus, the aspect ratio equal to the shear span ratio ((effective height of lateral load)/length), which
essentially describes the relationship between shear and flexure base reactions and is a more appropriate descriptive
quality to consider in a multi-story building. Thus, flexure-controlled walls in this study are considered to have a
shear span ratio, M/Vlw, greater than 1.5. In the case of investigating multistory walls, the shear span ratio is
determined by assuming a uniform distribution of lateral forces at each floor.
Section 2 provides an overview of the existing provisions for flexure-controlled walls. Section 3 provides
an evaluation of these provisions based on a database of experimental data. Due to a fairly limited number of tests
available from which to develop comprehensive empirically based recommendations for future versions of the
ASCE 41 provisions, a parameter study was conducted to provide simulation data to support the development new
recommendations. Section 4 provides an overview of the parameter study, which was conducted using numerical
modeling techniques validated using much of the experimental data used to evaluate existing provisions. Section 5
provides an analysis of trends in the experimental and simulation data to identify the parameters that should be used
4.2
@Seismicisolation
@Seismicisolation
Evaluation of ASCE 41 Modeling Parameters for
Slender Reinforced Concrete Structural Walls
identifying modeling parameters and acceptance criteria. Finally, Section 6 provides recommendations for updated
values.
ASCE 41 provides provisions for both linear and nonlinear analysis of existing structures. Linear procedures
determine the adequacy of a structural component based on the ratio of the simulated demand to the strength, with
consideration for over strength in linear models accounted for by `m-factors’, where m-factor values are provided in
tabulated form based on key characteristics of a component. Linear procedures are not evaluated here.
For nonlinear analysis of a structure, the response is represented using a load-deformation response such as
the one shown in Figure 1. For structural walls controlled by the flexural response, the load-deformation response is
the moment-rotation response, where the rotation is the rotation in the plastic hinge region, Lp, at the base of the wall
and Lp is defined as equal to one-half the wall length, but not greater than the floor height. Point ‘B’ in Figure 1 is
the yield point of the wall. Points ‘C’, ‘D’, and ‘E’ are determined by the values ‘a’, ‘b’, and ‘c’, defined by
tabulated values in ASCE 41 (see Figure 2). For flexure-controlled walls, these parameters are specified as a
function of the axial load ratio, the shear demand, and confinement of the boundary regions, with linear interpolation
permitted between the tabulated values.
In addition to specifying modeling parameters for use in nonlinear models, ASCE 41 provides acceptance
criteria. For structural walls controlled by the flexural response, the acceptance criteria specify the maximum plastic
hinge rotations that shall not be exceeded. The maximum plastic hinge rotations are specified as a function of the
axial load ratio, the shear demand, and confinement in the boundary regions (see Figure 2). Six sets of acceptance
criteria are provided; selection of the appropriate set is driven by the desired performance level (Immediate
Occupancy (IO), Life Safety (LS), or Collapse Prevention (CP)) and if the component studied is a primary
component or a secondary component.
Figure 1 - Generic load-deformation response for flexural members (from ASCE/SEI 41 supplement 1).
The current ASCE/SEI 41-06 provisions were evaluated by comparing modeling parameters and acceptance criteria
to the equivalent response from experimental tests. The modeling parameters that are provided are the plastic hinge
rotations ‘a’ and ‘b’, and the residual strength ratio ‘c’ (see Figure 1). The plastic hinge rotation ‘a’ is the change in
rotation from yield of the component to loss of lateral load carrying capacity. The plastic hinge rotation ‘b’ is the
change in rotation from the loss of lateral load carrying capacity to the rotation at which the wall lateral capacity is
reduced. For flexure-controlled walls, the critical value to be evaluated is ‘a’; the additional plastic hinge rotation
following loss of lateral load carrying capacity and the residual strength ratio are secondary response characteristics
that do not have dramatic effects on determining whether or not a structural component is seismically adequate.
Additionally, available experimental data typically does not include the residual strength and/or the decay from
initial failure to the residual strength. Thus, the discussion presented here will be limited to a discussion of the
plastic hinge rotation value ‘a’.
ASCE 41 modeling parameters and acceptance criteria for flexure-controlled walls were evaluated using a
database of experimental wall tests consisting of 66 walls with shear span ratios equal to or greater than 2.0. Details
of the database are provided by Birely (2012). Not all experimental test programs reported measurements that could
4.3
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
Figure 2 - Acceptance criteria for reinforced concrete walls controlled by flexure (from ASCE/SEI 41
supplement 1). Shear stress demands, V/twlw√f’c, are for concrete strengths in psi; for MPa, multiply by
1/12.043.
be used to calculate rotation and, for those that did, the measurements were not consistent between test programs.
Thus, to provide a consistent measure for the rotation in the walls, an equivalent plastic hinge rotation was
calculated from the reported load-displacement histories and geometric properties of the walls. The equivalent
plastic hinge rotation was defined as the rotation in a lumped-plasticity model that would result in the experimental
load-displacement history. The lumped-plasticity model comprised a rotational hinge at the base of the wall with a
hinge length of one-half the length of the wall. The region outside the hinge was modeled as elastic with an effective
flexural stiffness of 0.5EcIg and an effective shear stiffness of 0.4EcIg.
Experimental ‘a’ values were calculated by subtracting the equivalent plastic hinge rotation at yield from the
equivalent plastic hinge rotation at the drift capacity. To ensure consistency between the experimental specimens,
for which varying levels of details are available in the literature, the yield points were defined as the moment at
which the extreme reinforcing bar in tension yielded in a moment-curvature analysis of the cross-section (note that
this is different than the calculation of the yield point specified in ASCE 41). The drift capacity was defined as either
the maximum experimental drift or the drift at which a 10% loss of lateral load carrying capacity occurred.
The literature for walls supports the general understanding by the engineering community that the performance
of structural walls, particularly when measured by the drift capacity of the wall, is greatly impacted by the axial load
ratio, the shear demand, and the confinement of boundary regions. Other factors have also been shown to impact the
performance, but it is these three that are the biggest factors, and it is these on which the current ASCE 41
provisions are based.
4.4
@Seismicisolation
@Seismicisolation
Evaluation of ASCE 41 Modeling Parameters for
Slender Reinforced Concrete Structural Walls
The experimental values are compared to the ASCE 41 values in Figures 3 and 4. Figure 3 compares the values
for the modeling parameter ‘a’. The values are shown versus the experimental values for axial load ratio (left
column) and maximum shear stress demand (right column). The first row shows all test specimens (rectangular,
barbell, C-shape, H-shape, and T-shape walls). The second and third rows show the rectangular and flanged walls
(C-, H-, and T-shape), respectively. For tests with bi-directional loading, reported data is for strong-axis bending.
Figure 4 provides the same set of comparisons but for the acceptance criteria (maximum plastic hinge rotation). The
following general observations can be made:
1. Overall, the modeling parameter ‘a’ is conservative; however, for the limited data points at higher axial
load ratio in rectangular walls, the ASCE 41 values are unconservative.
2. There is a lack of experimental data in the range of 3-4√f’c psi (0.25-0.33√f’c MPa), a reasonable shear
demand in many walls, particularly rectangular walls.
3. For rectangular walls, there is a general decrease in the experimental values as the axial load ratio and the
maximum shear stress demand increase, confirming observations in the literature that these factors impact
the performance of structural walls and justifying the use of these values as selection criteria for the ASCE
41 modeling and acceptance criteria.
4. For flanged walls, there are no distinct relationships between the experimental values for ‘a’ and the axial
load ratio or shear stress demand.
5. Relative to the magnitude of the experimental values, the distinction between acceptable plastic hinge
rotations for Life Safety (LS) and Collapse Prevention (CP) are minor. In applying these limits, it seems
unlikely to achieve LS without achieving CP, an observation that is consistent with analysis of buildings
damaged in the 2010 Maule earthquake using the ASCE 41 acceptance criteria (Birely et al. 2013).
Although the experimental data indicates trends, there is limited data, particularly for walls with axial load
ratios in excess of 10% and flanged walls. The top-left plot in Figure 3 clearly shows that many experimental tests
have no axial load, an unrealistic condition in buildings. Figure 5 shows, for rectangular and flanged walls, the
relationship between ‘a’ and the shear stress demand for walls with axial loads, illustrating the limited data for use in
developing recommendations for revised ASCE 41 modeling parameters and acceptance criteria. Additionally,
despite visible trends in Figures 3-5, other parameters may also be crucial to the performance of walls. Birely (2012)
investigated the impact of a wide range of wall characteristics on the drift capacity and damage progression for the
same set of data presented here and found that other characteristics may impact the performance of walls as well, yet
there is insufficient data to state this with certainty.
4. PARAMETER STUDY
The experimental data for slender RC walls is sufficiently limited that the empirical calibration of
recommendations for revised acceptance criteria is not practical for slender walls as it may be for other components
for which hundreds of experimental data points are available. To provide a more diverse set of data for use in
developing proposed revisions to ASCE 41, a parameter study was conducted to generate additional data. The
parameter study was conducted using OpenSees following modeling recommendations by Pugh (2012). The
following sections provide an overview of the models used and the characteristics of the parameter study.
The parameter study was conducted using OpenSees nonlinear force-based beam-column elements. One element
was used per floor, with five integration points per element. Each integration point consisted of an aggregated
section made up of independent responses of the wall to i) shear demand, and ii) flexural and axial demands. Out-of-
plane displacements were restrained, thus, the torsional stiffness of the walls was not considered.
Shear was model using an elastic shear stiffness equal to the gross shear stiffness of the wall in the
direction of bending. It should be noted that the other more sophisticated models for the shear response, including
those that couple the shear and flexure response of the walls, would likely result in more accurate models of the
cyclic response of the walls; however, as the focus of the developed acceptance criteria is on the ability to capture
the flexural response of the wall, the approach used for modeling the shear is reasonable, is the simplest method
available for including shear, and is consistent with the shear stiffness in the current version of ASCE 41.
The flexural and axial response of the walls was modeled using fiber sections followed the
recommendations of Pugh (2012), which were validated for simulation of planar walls using an extended data set;
4.5
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
results show accurate simulation of stiffness to yield, strength, and drift capacity. Initial evaluation of the model for
non-planar walls using limited experimental data suggests that the model is appropriate for C-shaped walls but not
for T-shaped walls. For T-shaped walls, the assumptions of a linear strain field on the wall cross-section, which is
inherent to fiber-based beam-column elements, can result in significant error in simulated response. The basis for
Pugh’s recommendations is that distributed plasticity elements have the potential for localization of damage at the
critical section and that the global response is sensitive to the number of integration points used. To achieve an
objective response regardless of mesh size, the proposed recommendations use post-peak regularization of the
stress-strain response of the material models.
Figure 3 - Comparison of experimental values for the parameter ‘a’ and the ASCE 41 values for ‘a’;
experimental specimens used are detailed in Birely (2012). Shear stress demands shown are for concrete
strengths in psi; for MPa, multiply by 1/12.043.
4.6
@Seismicisolation
@Seismicisolation
Evaluation of ASCE 41 Modeling Parameters for
Slender Reinforced Concrete Structural Walls
All Walls
Rectangular Walls
Flanged Walls (H-, T-, C-shape) Axial Load Ratio Shear Stress Demand
Figure 4 - Comparison of experimental rotation at failure and the ASCE 41 values for acceptable plastic
hinge rotation; experimental specimens used are detailed in Birely (2012). Shear stress demands shown are
for concrete strengths in psi; for MPa, multiply by 1/12.043.
4.7
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
Figure 5 - Relationship between shear and experimental values for the parameter ‘a’ and the ASCE 41 values
for ‘a’ for experimental specimens with axial load. Shear stress demands shown are for concrete strengths in
psi; for MPa, multiply by 1/12.043.
The geometry of the parameter study walls was selected to provide a range of demands while falling within a range
representative of building characteristics for low- to mid-rise buildings. Two wall heights, 4 and 8 floors (10 ft. floor
heights), were considered. Wall thicknesses of 8 in. (200 mm) and 12 in. (300 mm) were used in combination with
cross-sectional aspect ratios of 10, 15, and 20. This resulted in aspect ratios (wall height/wall length) ranging from
2-12. If a uniform distribution of lateral loads is assumed, the resulting shear span ratios could range from 1.25-6.75.
To ensure that only flexure dominated walls were considered in the study, only shear span ratios equal to or greater
than 1.5 were used in developing acceptance criteria. The cross-sectional aspect ratios used are representative of the
range found in most building. The thicknesses used are at the lower range of those found in mid-rise buildings,
however, it is worth noting that thicker walls would typically be found in taller walls and that the corresponding
range of shear span ratio and cross-sectional aspect ratio is accounted for in the limited parameter study presented
here.
Planar and symmetric flanged walls were considered. Symmetric walls were defined to have two identical
flanges at both ends of the walls. The lengths of the flanges were defined as 10%, 25%, or 50% of the length of the
wall. For the range of wall heights and lengths used in the parameter studies, no flanges exceeded the effective
flange width specified by ACI 318.
The distribution of longitudinal reinforcing bars was considered to follow two general layouts. The first
longitudinal reinforcement is a uniform distribution of the bars throughout the wall. For these walls, reinforcement
ratios of 0.25%, 0.5%, 1.0%, and 1.5% were used. Walls with 0.25% reinforcement meet the minimum required by
the present ACI 318 code (ACI-318 2011). For older walls that may be evaluated with ASCE 41, the gross
reinforcement ratio may be less than 0.25%; however, such walls are essentially unreinforced and therefore not
consider here. The second longitudinal reinforcement layout is that commonly used in modern walls, with most
reinforcement concentrated in the end region (boundary element). For the planar and symmetric flanged walls, a
reinforcement ratio of 0.25% was used in web (the region between the boundary elements) and the reinforcement in
the boundary elements was set such that the gross reinforcement ratio of the wall was 0.5%, 1.0%, or 1.5%. In the
walls with boundary elements, 0.25% was not considered as adding boundary elements would not impact the
distribution of the steel. For unconfined walls, this would result in walls similar to the uniform distribution of steel
and confined boundary elements are unlikely to be found in such lightly reinforced walls. In the planar walls, the
boundary element length was considered to be 10% or 20% of the total wall length. For the symmetric flanged walls,
each flange was considered to be the boundary element.
The material properties were kept consistent for all walls considered in the parameter study. Lower-bound
values were adopted from ASCE 41 Tables 6-2 and 6-3 for the most recent time periods; for concrete, compressive
strength of 4,000 psi (28 MPa) and for reinforcing steel, a yield strength of 60 ksi (420 MPa). The lower-bound
properties were multiplied by the factors in ASCE 41 Table 6-4 (1.5 for concrete and 1.25 for reinforcing steel) to
4.8
@Seismicisolation
@Seismicisolation
Evaluation of ASCE 41 Modeling Parameters for
Slender Reinforced Concrete Structural Walls
convert to expected material properties. Concrete was modeled using the OpenSees Concrete02 uniaxial material,
with a residual concrete strength of 20% of the peak concrete strength. Steel was modeled using the OpenSees
Steel02 uniaxial material with strain limits included to indicate failure of the specimens. The compressive strain
limit was set equal to the concrete compressive strain when the residual strength is first reached, essentially
indicating that concrete crushing and buckling will occur simultaneously; such an approach was deemed reasonable
by Pugh (2012) for modeling the response of experimental test specimens. The tension strain limit was set to be
30%, ensuring a very ductile steel response; no considerations were made for fracture due to low-cycle fatigue.
Three levels of confinement were considered for the boundary elements: unconfined, confined, and well-
confined. The distinction between well-confined and confined was intended to capture the difference between
boundary elements that, essentially, meet the ACI 318 provisions for special boundary elements (well-confined) and
boundary elements for which some level of confinement is provided to delay bar bucking but which cannot be
considered as special boundary elements. The distinction between the two levels of confinement is intended to
provide data to support maintaining the current ASCE 41 practice of not only defining values for confined and
unconfined boundary elements, but also reduced confined values for boundary elements that are confined but do not
fully meet the ACI 318 provisions for special boundary elements.
4.3 Loading
For each wall design analyzed in the parameter studied, six separate analyses were conducted, each with a different
axial load ratio. The lowest axial load ratio used was 2% of the gross axial capacity of the wall, which was
considered to be a lower bound on the axial load ratio that would account for the self-weight of the wall and a very
small tributary area of load. Other axial load ratios used ranged from 5-25% of the gross-capacity axial capacity of
the wall, in increments of 5%. An axial load ratio of 25% of the gross-capacity would be considered to be a very
large axial load ratio in a structural wall; however, coupled walls, particularly planar walls, have been estimated to
sustain axial compressive demands as large as 40% of the gross-capacity (Lehman et al. 2013) due to the axial
demands introduced by coupling beam shear.
For all walls, lateral loads were distributed uniformly at all floors. Taking into account the cross-sectional
geometry of the walls studied and the use of four- and eight-story walls, the resulting shear span ratios ranged from
1.25-6.75; however, only walls with shear span ratios greater than 1.5 were used.
The same reverse-cyclic drift history was used for each wall to ensure consistency of the compared data;
this meant using different displacement histories for the four- and eight-story walls. Two full cycles were completed
for each peak drift. The drifts used were intended to capture the cyclic response of the walls at both pre- and post-
yield levels. Drifts of 0.01%, 0.02%, 0.05%, 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, 0.75%, 1.0%, 1.5%, 2.0%, 2.5%, 3.0%,
3.5%, 4.0%, 4.5%, and 5.0% were used. If failure had not been simulated at 5% drift, additional cycles were
completed to peak drifts increasing by increments of 1.0%.
4.4 Post-processing
Results of the parameter study were evaluated and any walls with maximum shear stress demand in excess
of 8√f’c psi (0.66√f’c MPa), the limiting shear stress in ACI 318, were removed. Yield of the walls was defined as
when the extreme tension reinforcing bar reached the yield strain. Failure of the walls was defined to occur when a
10% loss of shear occurred.
Although the use of the distributed plasticity beam-column elements would permit the calculation of the
wall rotation at yield and failure, it was desired to provide values that were consistent with the lumped-plasticity
approach in ASCE 41 and that were compatible with the experimental rotations considered. Thus, the equivalent
plastic hinge rotation approach was used, in which the rotation reported is the plastic hinge rotation back-calculated
from the simulated lateral drift of the wall.
5. RESULTS
The simulation results were evaluated to determine the appropriate parameters to use in providing recommendations
for revised modeling parameters and acceptance criteria for ASCE 41. While all modeling parameters (‘a’, ‘b’, and
‘c’) and acceptance criteria (for three performance levels and two component importance categories) matter, the
most critical is the parameter for ‘a’. The value ‘a’, a key aspect of the backbone curve shown in Figure 1, defines
the plastic hinge rotation from initial yield to loss of significant loss of load carrying capacity. The other modeling
parameters define the response after loss of load carrying capacity and the acceptance criteria provide acceptable
4.9
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
total plastic hinge rotations for specified performance levels. These values can be defined in relationship to the
parameter ‘a’. In ASCE/SEI 41-06, the acceptance criteria for the collapse prevention performance level of primary
components are defined as equal to ‘a’ and all other acceptance criteria are a percentage of this value. Thus, the
evaluation of the simulation data presented here focuses solely on the ‘a’ parameter, with the objective of achieving
the following:
1. Evaluate the expectation that shear and axial load ratio are the primary variables that impact the
performance of walls and thus should be the characteristics upon which the modeling parameter ‘a’ is
defined.
2. Establish appropriate limits, if any, for shear and axial demands if tabulated values of modeling parameters
and acceptance criteria are desired.
3. Evaluate the impact of confinement on the value of ‘a’, including varied levels of confinement (i.e. well
confined versus poorly confined), and establishing how, if at all, confinement should be accounted for in
determining modeling and acceptance criteria.
4. Determine if other wall characteristics should impact the definition of modeling and acceptance criteria.
Values for ‘a’ from the parameter study walls are shown against the shear stress demand in Figure 6. The
simulation data are indicated by open black circles. For comparison, values for experimental data with axial load are
shown by solid blue diamonds. It is evident that the simulated results are reasonable in comparison to the
experimental data. The ‘a’ parameter values for each simulated wall as specified by ASCE 41 are indicated by red
lines. A detailed examination of the simulated data is presented in the sections that follow.
Figure 6 - Relationship between shear stress demand and simulated values for the parameter ‘a’ and the
ASCE 41 values for ‘a’. Shear stress demands shown are for concrete strengths in psi; for MPa, multiply by
1/12.043.
A cursory evaluation of the wall design and loading characteristics impacting the value of ‘a’ indicates that,
as anticipated based on experimental data, shear stress demand and axial load ratio have a significant relationship
with the value of ‘a’. These two characteristics are highlighted in Figure 7, where ‘a’ is shown as a function of the
shear demand, with separate plots for each axial load ratio considered. Planar and flanged walls are shown
separately (Figure 7a and 7b, respectively). Simulated data is indicated by black circles. ASCE/SEI 41-06 values are
indicated by red lines; the three distinct lines shown are a result of the three confinement levels considered.
Key observations from evaluation of the data presented in Figure 7 include:
1. The impact of the axial load ratio is significantly greater in the simulated values for ‘a’ than is provided for
in the ASCE 41 modeling parameters. At low axial load ratios (2-5% of the gross axial capacity) the ASCE
4.10
@Seismicisolation
@Seismicisolation
Evaluation of ASCE 41 Modeling Parameters for
Slender Reinforced Concrete Structural Walls
41 values fall within the broad range of ‘a’ values. At axial load ratios of 10% or greater, the ASCE 41
values are unconservative, with the values for confined and well confined walls falling above the maximum
simulated values.
4.11
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
2. Current ASCE 41 values have lower limits at 10% axial load ratio and upper limits at 25% axial load ratio,
with interpolation permitted in between the two limits. Figure 7 indicates that such limits are inappropriate.
With increasing axial load ratio from 2% to 5% to 10%, the range of values for ‘a’ decreases significantly.
At axial load ratios greater than 10%, the range of values for ‘a’ continues to decrease, but the drop is less
dramatic than at smaller axial loads. Consequently, revised limits would be warranted.
3. At high axial load ratios, there is limited variation in the values for ‘a’ as a function of the shear stress
demand. At lower axial load ratios, there is a clear decrease in the value for ‘a’ as the shear stress demands
increases, however, this trend levels out at shear stress demands of approximately 3-5√f’c psi (0.25-0.42√f’c
MPa), indicating that the current lower and upper bound limits for shear demands may not be the
appropriate limits to use.
4. The planar and flanged walls are shown separately to illustrate the observation that, for a specified shear
demand, particularly at lower axial load ratios, the value ‘a’ can be expected to be much larger in flanged
walls. Thus, it may be necessary to provide separate modeling parameters and acceptance criteria for planar
and flanged walls, or to define these values based on characteristics that will account for these differences.
5.2 Confinement
Current ASCE 41 modeling parameters and acceptance criteria for flexure-controlled walls provide
different sets of values for walls with unconfined and confined boundary elements, and permits a reduction in
confined values for poorly confined walls. It is well accepted by structural engineers that confinement may increase
the lateral capacity of walls. Evaluation of the impact of confinement based on the parameter study was done to
establish the most appropriate method for accounting for the impact of confinement. The parameter study was
structured to allow for a direct method of considering the confinement. For each wall geometry and reinforcement
ratio for walls with boundary element, three levels of confinement were considered, with the all other wall
characteristic remaining unchanged. The levels of confinement were achieved by adjusting the crushing energy of
the concrete and thus increasing the strain at which crushing occurred. For the intermediate level of confinement
(referred to as confined), the crushing strain was adjusted but the confined concrete was not considered to have an
increase in compressive strength. For the highest level of confinement (referred to as well confined), the crushing
strain was adjusted using a different crushing energy and the confined concrete was considered to have a 40%
increase in compressive strength. Although all other characteristics of the walls stayed the same, the increased
compressive strength results in higher nominal moment strength and corresponding higher based shear demand
relative to unconfined concrete. Thus, evaluation of the impact of confinement on ‘a’ should account for the increase
in shear strength; Figure 8 shows the ratio of the change in ‘a’ due to the addition of confinement relative to the
change in the shear stress demand for planar walls with a 10% axial load ratio. For confined walls, there is
essentially no increase in the shear demand and an increase in ‘a’ of up to 10% from that of the unconfined walls.
For well confined walls, the value for ‘a’ increases as much as 80% of the confined value and there is a distinct
positive slope for the increase in ‘a’ relative to the increase in the shear demand. These observations from Figure 8
clearly indicate that the level of confinement is an appropriate parameter to consider in defining modeling
parameters for flexure-controlled walls. It is not clear, however, that the method used in the current ASCE 41
provisions is appropriate. Rather, it may be more appropriate to define values for walls with unconfined and well
confined boundary regions, with an increase in the unconfined values permitted for poorly confined boundary
regions. Additional analysis and comparison to experimental data is necessary to provide improved
recommendations.
For planar walls, Figure 7a indicates that consideration for shear and axial demands is capable of capturing
distinct trends; other parameters may affect the modeling parameters, but the impact is expected to be relatively
minimal. For symmetric flanged walls, the trends are less distinct and it is necessary to further explore relationships
to wall characteristics to determine impact on the modeling parameter ‘a’.
A key characteristic of the symmetric flanged walls is the length of the flange (bf) relative to the length of
the wall (lw), referred to here as the flange aspect ratio (FAR). Three FAR values (10%, 20%, and 50%) were
considered. For larger FAR values, the moment capacity is increased, and since the load distribution is held
constant, there is an increased shear demand at the base of the wall; this impact of the FAR on the shear demand
should be accounted for in evaluating the impact of shear demand on the modeling parameter ‘a’ for the symmetric
4.12
@Seismicisolation
@Seismicisolation
Evaluation of ASCE 41 Modeling Parameters for
Slender Reinforced Concrete Structural Walls
flanged walls. Figure 9 shows, for symmetric flanged walls with an axial load demand of 10% of the gross capacity,
the relationship between ‘a’ and the shear demand for each of the three FAR values considered.
Figure 9a, b, and c individually indicate less variation in ‘a’ relative to shear than for the same axial load in
Figure 7b, which makes no distinction for the FAR. From Figure 9, it is evident that the FAR has an impact on the
value of ‘a’. For relatively short flanges (FAR = 0.10, Figure 9a), the relationship between ‘a’ and the shear demand
is very similar to that for planar walls. For wider flanges (FAR = 0.25 and FAR = 0.50), the median values for ‘a’
are larger. Figure 9b and 9c show more scatter in the value for ‘a’, but the lower-bound and median values are larger
for FAR = 0.50 than for FAR = 0.25. Collectively, this suggests the need to provide modeling parameter
recommendations for symmetric flanged walls that are dependent on the flange aspect ratio (FAR).
100
90
80
70
a / aunconfined
60
50
40
30
20
10
0
-0.5 0 0.5 1 1.5 2 2.5 3 3.5
v / vconfined
Figure 8 - Ratio of the increase in the modeling parameter ‘a’ relative to the change in the shear demand
when the level of confinement is increased from unconfined to confined (gray circles) and unconfined to well
confined (black circles). Only planar walls with axial load of 10% are shown.
6. PRELIMINARY RECOMMENDATIONS
Evaluation of the simulated values of the modeling parameter ‘a’ confirms that it is appropriate to use the axial load,
shear demand, and level of confinement to define modeling parameters and acceptance criteria for flexure-controlled
walls. This is consistent with the current provisions of ASCE 41; however, the current values do not predict
simulation and experimental values well, thus, it is necessary to provide recommendations for revised provisions.
Additionally, the presence of flanges and the length of the flanges relative to the wall length should be accounted
for. Using the simulated data, linear regression was used to provide preliminary recommendations for updated
modeling parameters. Two sets of recommendations were developed: i) for planar walls, with ‘a’ defined as a
function of axial load ratio and the shear stress demand, and ii) for flanged walls, with ‘a’ defined as a function of
4.13
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
axial load ratio, shear stress demand, and flange aspect ratio (FAR). For each set of recommendations,
recommendations are provided for the three levels of confinement considered: unconfined, confined, and well
confined.
Evaluation of the simulated results indicates that it may be appropriate to a) specify upper and lower bound
values with interpolation between the limits or b) define a bilinear model that is steep at low shear and axial load
demands but nearly flat at high demands. However, it was desired to provide proposed revisions that are simple to
implement and multi-parameter linear regression was used to develop proposed recommendations. Thus, the
recommendations for the planar and flanged walls are defined by Equations (1) and (2), respectively:
The regression coefficients for each geometry (planar and flanged) and confinement level (unconfined, confined,
and well confined) are provided in Table 1, along with the R2 values for the regression plane. Although it was
desired to have a consistent linear expression for all conditions, at high axial loads and high shear demands,
application of the linear regression equations to the simulation data resulted in negative values for ‘a’ at high shear
stress demands and high axial demands. To provide realistic conditions, an lower bound limit was specified at an
axial load ratio of 20% and shear stress demand of 4√f’c psi (0.33√f’c MPa). The value predicted at this point is set
as the lower bound for all walls with higher axial load or for the same axial load and greater shear stress demand.
This point was found by inspection and found to produce a reasonable model for both planar walls and symmetric
flanged walls. Revision may be appropriate upon revising recommendations to include asymmetric flanged walls
and greater diversity of confinement values. Figure 10 compares the proposed modifications of Equations (1) and
(2) to the current values for ‘a’ in ASCE 41.
Table 1 - Regression coefficients and R2 values for proposed regression equation for revised modeling
parameter ‘a’.
Shape Confinement Level α αλ αv1 αFAR R2
To evaluate the proposed regression expressions for the modeling parameter ‘a’, an error measure was
specified as the ratio of the ‘a’ parameter from the simulation results, asim, to the ‘a’ parameter predicted by the
regression expressions, apredict. Using this error measure, a value of 1.0 indicates exact prediction of the simulated
parameter; a value less than 1.0 indicates an over-prediction of the simulated parameter (unconservative); and a
value greater than 1.0 indicates that an under-prediction of the simulated parameter (conservative). Figure 11a and
11b summarize this error value for the planar and flanged walls, respectively, by showing the cumulative
distribution of the error as a solid black line. For comparison, the same error measure using the ‘a’ value predicted
by ASCE/SEI 41-06 are shown in as red lines.
For planar walls at low axial loads (2-5%), the proposed regression equation for the modeling parameter ‘a’
presents approximately the same cumulative error distribution as does the current ASCE 41 modeling. For these
walls, approximately 50% of the ‘a’ values are over-predicted and 50% are under-predicted. Given that the error is
largely unchanged for low axial load in planar walls, the proposed equation for the modeling parameter ‘a’ offers a
benefit to the current as it is continuous with the values for higher axial loads. For planar walls with axial loads from
10-20%, the median error for the proposed regression equation again is such that 50% of the ‘a’ values are over-
predicted and 50% are under-predicted. At these higher demands, this is a significant improvement over the current
ASCE 41 model parameter, for which the modeling parameter is generally much larger than the ‘a’ value from the
simulation results. At an axial load of 25%, the median for proposed regression equations is significantly less than
1.0; although this is unconservative, it does offer a slight improvement over the ASCE 41 prediction.
4.14
@Seismicisolation
@Seismicisolation
Evaluation of ASCE 41 Modeling Parameters for
Slender Reinforced Concrete Structural Walls
For flanged walls, the proposed regression equations, which take into account the flange aspect ratio,
provide an overall improvement in the prediction of the ‘a’ value. This is evident in Figure 11b, which shows the
propsed equations shifting the median error measure closer to 1.0 relative to the median error measure for the ASCE
41 values, which overall provided unconservative values for ‘a’. As is the case for the planar walls, the prediction of
‘a’ at high axial demands is significantly unconservative.
4.15
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
Figure 11 presents the evaluation of the recommended modeling parameter ‘a’ for the simulation results
used in establishing the recommendations. The recommended values were then used to model the experimental data
set described in the evaluation of the current ASCE 41 provisions. Figure 12 shows a comparison of the current and
recommended values for rectangular and flanged walls. For rectangular walls, the recommended values are slightly
more conservative than the current values. The largest impact is for flanged walls, for which the prediction of the
experimental values is significantly improved over that of the current ASCE 41 modeling parameters which
significantly underestimated the value for ‘a’.
4.16
@Seismicisolation
@Seismicisolation
Evaluation of ASCE 41 Modeling Parameters for
Slender Reinforced Concrete Structural Walls
Figure 12 – Comparison of existing ASCE 41 modeling parameter ‘a’ values and proposed values to values
from experimental tests. Shear stress demands shown are for concrete strengths in psi; for MPa, multiply by
1/12.043.
ASCE/SEI 4106 Seismic Rehabilitation of Existing Buildings is a guideline used by engineers to determine the
seismic adequacy of buildings based on the response of individual building components. The modeling parameters
for flexure-controlled reinforced concrete walls were evaluated using a database of experimental tests and were
found to be conservative, particularly for flanged walls; however, for the limited data points at higher axial loads
and for flanged walls, the values were found to be unconservative. In the study presented here, recommendations for
improved modeling parameters for flexure-controlled walls were developed. To provide a diverse range of data from
which to develop the recommendations, a parameter study was conducted using models validated using
experimental data. Both rectangular and symmetrically-flanged walls were considered, with a variety of geometry,
reinforcement ratios (0.25-1.5%), confinement levels, and axial load ratios (0.02-0.25Agf’c). Simulation results were
used to develop regression equations for the recommended modeling parameters.
Separate regression equations for the recommended modeling parameters were developed for rectangular
walls and for symmetric flanged walls. The regression equations and associated coefficients are provided in
Equations 1 and 2 and Table 1. For both geometries, three sets of equations are provided based on the level of
confinement in the boundary regions (unconfined, confined, and well confined). All equations are a function of the
axial load ratio and the shear stress demand in the walls. The confinement, axial, and shear dependence are
consistent with the current ASCE 41 modeling parameters, thereby providing a direct comparison and allowing for
familiarity in implementing the recommended modeling parameters. For symmetric flanged walls, the recommended
modeling parameters introduce the consideration of the flange aspect ratio (ratio of the flange length to the wall
length). Another deviation the recommended modeling parameters make from the current ASCE 41 modeling
parameters is the elimination of tabulated values in favor of linear regression equations, with the exception of lower
bound limits provided on walls with high axial load and high shear demands.
The proposed regression equations were developed from a preliminary parameter study for which further
analysis is intended. There remains a need for additional simulations and evaluations prior to providing finalized
recommendations. Needs for further investigation include:
1. Observations for the simulated data have focused on the modeling parameter ‘a’ with the assumption that
relating all other modeling parameters & acceptance criteria to this would be appropriate. Further
evaluation of the data is necessary to determine if this is true and to establish relationships between ‘a’ and
all other modeling parameters and acceptance criteria.
2. There is a need to expand evaluation of the parameter study to consider asymmetric flanged walls. For
asymmetric flanged walls, it is necessary to not only consider the ratio of the flange length to the web
length, but the length of the boundary region at the tip of the flange and the unbalanced reinforcement of
the two boundary regions (web and flange). The current modeling parameters of ASCE 41 are based on
4.17
@Seismicisolation
@Seismicisolation
A. C. Birely et al.
axial load ratios that include consideration for unbalanced reinforcement in the tension and compression
regions, which along with the recommended inclusion of the flange aspect ratio, would likely provide a
simple method for addressing the difference in response between rectangular, symmetric flanged, and
asymmetric flanged walls. For asymmetric flanged walls, the modeling parameters would necessarily be
different for bending with the flange in compression verses the flange in tension, therefore resulting in
modeling parameters that will would likely be bound by the recommendations for the rectangular and
symmetric flanged walls.
3. Expand the parameter study to further evaluate not just the impact of confinement, but also the degree of
confinement. The data presented here was for unconfined, confined, and well-confined boundary regions.
In the case of well-confined boundary regions, the modification to the confined concrete properties was an
upper bound for reasonable values. Additional refinement could provide useful data for establishing
recommendations for the impact of confinement on the modeling and acceptance criteria.
4. Expand the parameter study to include more walls with shear stress demands greater than 6√f’c psi (0.5√f’c
MPa). The data for such data points is sparse relative to the data for low shear demands.
5. For flanged walls, investigate the impact of varied reinforcement configurations in the flange (discrete
confined elements in the toes of the flanges as opposed to the uniform distribution of reinforcement and a
fully confined flange (for well-confined walls, this may lead to simulated data in excess of reasonable
characteristics for buildings))
6. Results presented here are based on the assumption that the failure mode is flexure-compression, where
failure is defined in the models by crushing of the concrete. Flexure-tension failure should also be
considered, where failure of the walls is defined by fracture of the reinforcing bars in tension. It is expected
that such consideration would result in a decrease in the simulated ‘a’ parameter for walls with low axial
demands and low shear stress demands. When using the recommended modeling parameters, engineers
should make consideration for whether or not strain demands are large enough to cause fracture in
reinforcing bars, taking into consideration of the ductility of the steel in the wall analyzed.
7. The experimental data used in validating the modeling technique used and evaluating the proposed
recommendations has upper bound axial load ratios of approximately 10% for flanged walls and 15% for
planar walls. Where walls have axial loads in excess of this, such as in coupled wall systems, the proposed
modeling parameters should be used with caution due to the lack of validating data. Further improvement
of the prediction at high axial loads may be practical when a larger experimental data set is available.
8. REFERENCES
ASCE-41, 2006. Seismic Rehabilitation of Existing Buildings (ASCE/SEI Standard 41-06), Structural Engineering
Institute, American Society of Civil Engineers, Reston, Virginia.
ASCE-41, 2007. Supplement No. 1, Seismic Rehabilitation of Existing Buildings (ASCE/SEI Standard 41-06),
Structural Engineering Institute, American Society of Civil Engineers, Reston, Virginia.
Birely, A.C., 2012, “Seismic Performance of Slender Reinforced Concrete Structural Walls”, PhD Dissertation,
University of Washington, Seattle, WA.
Birely, A.C., Lowes, L.N., Lehman, D.E., Aviram, A., Kelly, D.J., 2013. ASCE 31/41 Evaluation of Damaged
Chilean Walled Buildings. Proceedings of ASCE Structures Congress 2013, Pittsburg, PA.
FEMA-356, 2000, Prestandard and Commentary for the Seismic Rehabilitation of Buildings, American Society of
Civil Engineers (ASCE) for the Federal Emergency Management Association (FEMA), Washington, D.C.
Pugh, J.S., 2012 “Numerical Simulation of Walls and Capacity Design Recommendations for Walled Buildings”,
PhD Dissertation, University of Washington, Seattle, WA.
Lehman, D.E., Turgeon, J.A., Birely, A.C., Hart, C.R., Marley, K.P., Kuchma, D.A., Lowes, L.N., 2013. Seismic
Behavior of a Modern Concrete Coupled Wall. ASCE Journal of Structural Engineering, 139(8)
4.18
@Seismicisolation
@Seismicisolation
SP-297—5
Synopsis: Over the past few decades, flat plate concrete building systems have been widely adopted in the United
States and other countries because it enables not only to save construction time and cost but also to make better use
of interior spaces. However, it has been observed that such buildings whose columns are cast into the concrete flat
plate are highly vulnerable to collapse. Such integrated slab-column frames have suffered severe damage or
completely collapsed during the past earthquake events. The main failure mode of these structures is punching shear.
Although a general guideline for seismic design and a seismic rehabilitation design guideline for both existing and
new reinforced concrete frame structures are specified in ACI 318 and ACI 369R, respectively, many uncertainties
have still resided in the modeling parameters to accurately predict seismic behavior of the flat plate concrete frame
structures. Therefore, a new chart of allowable plastic rotation values for correlating values of gravity shear ratio is
presented in this study with the objective of updating the modeling parameters of ACI 369R. The major issues which
lead to errors in evaluating flat plate concrete structural behavior under seismic loads are thoroughly investigated.
Also, dominant parameters from previous and recent experiments on integrated slab-column connections subjected
to seismic loading are utilized to assess the estimation of seismic performance of the flat plate system based on ACI
369R. Consequently, although it is confirmed that there is a trend in correlation between the allowable plastic
rotation and gravity shear ratio while almost no correlation is observed with reinforcement ratio, more experimental
data are necessary to enhance this correlation study. It is also noticed that the current ACI 369 recommendations for
allowable plastic rotation values for slab-column connection under seismic and gravity loading are unconservative.
Keywords: reinforced concrete; slab-column connections; modeling; plastic rotation; seismic; evaluation.
5.1
@Seismicisolation
@Seismicisolation
A. C. Hufnagel et al.
Amy C. Hufnagel is a Structural Engineer in the Walter P. Moore, Houston. She received her BS and MS from the
University of Oklahoma, Norman. Her research interests include the design and analysis of reinforced and
prestressed concrete structures.
YeongAe Heo, PhD, is a Post-doctoral Researcher in the Department of Architecture and Architectural Engineering,
Seoul National University, Korea. She received her BS and MS from Dong-A University, Korea, and her PhD from
the University of California at Davis. Her research interests include the seismic evaluation and nonlinear dynamic
modeling of reinforced concrete structures.
Thomas H.-K. Kang, PhD, PE, FACI, FPTI, is an Associate Professor in the Department of Architecture and
Architectural Engineering, Seoul National University, Korea. He received his BS from Seoul National University,
Korea, his MS from Michigan State University, and his PhD from the University of California at Los Angeles. His
research interests include the design and behavior of reinforced and prestressed concrete structures.
INTRODUCTION
Flat plate concrete building systems consist of monolithically cast slab and column frames, and are widely used
with and without shear walls in the United States and other countries for several advantages, including economy and
building efficiency. This setup reduces story heights, which allows for lower costs for building material, cladding,
electrical and mechanical ductwork, and annual heating and air-conditioning. This lowers both dead loads and lateral
loads on the structure. Despite these many advantages, there are also many issues such as complexity in modeling
and punching shear failure with flat plate building systems.
Because of the many benefits they provide, there is high motivation for the implementation of flat plate
building systems in every type of environment and loading situation. This includes areas of moderate to high
seismicity, despite the additional risks of increased lateral loads during earthquake events. Over the years, this has
resulted in a high percentage of flat plate buildings that have suffered large amounts of damage or even completely
collapsed during past seismic events, like the 1971 San Fernando earthquake, which initially called attention to the
weaknesses in flat plate building design. During the 1985 Mexico City earthquake, many non-ductile flat plate slabs
suffered severe damage caused by punching shear failure (Figure 1). During this event, 91 buildings completely
collapsed, while 44 also suffered severe damage (Robertson and Johnson, 2006). This triggered increased scrutiny
into the behavior and the code design methods for these structures in both the United States and Mexico. One source
of weakness in these buildings is the slab-column connections and the possibility of punching shear failure, which
occurs when the slab area around the column fails and collapses. This failure is caused by excessive shear stress
applied to the connection, as well as poor design for lateral loads, including non-ductile connections. It is important
that these connections withstand certain levels of lateral drift, while maintaining their gravity load carrying capacity
during an earthquake. There is a need to better understand the behavior of slab-column connections in flat plate
structures in order to improve design methods, as well as to suggest methods for investigation and retrofit of older
buildings in seismic regions.
Figure 1 — Progressive collapse of flat plate buildings in Virginia and Mexico City.
5.2
@Seismicisolation
@Seismicisolation
Modeling Parameters for Reinforced Concrete Slab-Column Connections
The purposes of this research are: 1) to investigate several issues with flat plate concrete structures in order to
better understand their behavior under gravity and lateral loads; 2) to compile past research and experiments
completed on slab-column connections in which specimens were subjected to seismic loading, and record relevant
information in a database; and 3) to examine the existing ACI 369 modeling parameters and provide a new chart of
modeling parameters of allowable plastic rotation values for correlating values of gravity shear ratio. The research
consists of a review of previous large-scale experiments, as well as a series of analyses on the data contained therein,
and is concluded by presenting experimentally suggested modeling parameters of allowable plastic rotation for slab-
column connections. This study is applicable to both pure flat plate systems and flat plate systems with walls and/or
moment frames, though slab-column connections in buildings with walls and/or frames may not be modeled. In such
cases, the acceptance criteria of plastic rotation become very important, which will be investigated in depth in the
next study of Slab-Column Connections Task Group of ACI Committee 369.
Complexity in Modeling
Despite the many advantages for flat plate design and construction, the nonlinear behavior of these structures is
difficult to predict (Farhey et al., 1993). The complexity of the nonlinear behavior of slab-column connections is the
result of several contributing factors. The method of load transfer between the slab and column is complicated and
difficult to calculate, because of the combination of torsional and flexural moments and the three-dimensional stress
distribution as well as shear forces concentrated on the connection. Several analytical modeling methods have been
developed for slab-column frames (e.g., Kang et al., 2009; see Fig. C4.2 of ACI 369R-11). Modeling is especially
important in gaining a higher understanding of the complicated behavior of structural components, like slab-column
connections.
A major issue in flat plate concrete structures is the possibility of punching shear failure at the slab-column
connections, especially when subjected to cyclic lateral loading combined with gravity loading. This is a result of
several contributing factors, and is specifically problematic for those built in the 1950s and 1960s. These older
buildings were designed with non-ductile connections to resist gravity loads only, without shear walls or frames to
handle the applied lateral loads. The slab-column connection is responsible for transferring the unbalanced moment
and the shear forces. If shear stress due to these applied forces and moments surpass the shear capacity of the slab at
the face of the column, punching shear failure is a risk.
Another factor contributing to punching shear failure in the design of these older buildings is the lack of
continuous bottom reinforcement through the interior columns, which prevents the rebar from developing ultimate
strength at the location at the face of the column by acting as “a membrane to suspend the slab following failure of
the concrete” (Moehle et al., 1988). This lack of continuity in the connection makes the structure more vulnerable
and can lead to sudden punching shear failure or possibly a progressive collapse of the entire building. This collapse
may be the result of excessive lateral loading or the redistribution of gravity loads following lateral deformations.
When punching shear failure initially takes place, the slab drops and collapses onto the floor below. This, in turn,
overloads the floor below, eventually leading to a progressive collapse of the entire building. Punching shear failure
can happen for several reasons, including poor construction practices or premature removal of shoring, like the
collapse of the Skyline Plaza in Bailey’s Crossroads, Virginia (Figure 1); however, the main focus of this research is
the punching shear failure of flat plate systems subjected to lateral loading during seismic events. Even in flat plate
buildings that are designed with shear walls, the slab-column connections still need to have the capability of
maintaining gravity load carrying capacity under lateral load deformation. In order to prevent punching failure, the
connections must be capable of exhibiting ductility in the inelastic range, while still carrying the design loads (Pan
and Moehle, 1989). Non-ductile detailing in these connections does not provide enough anchorage between the
reinforcement and the concrete at the interior columns, which can cause a sudden brittle pull out and failure (Dovich
and Wight, 1996).
As such, several contributing factors aside from overly intensive ground motion may have led to the failure of
so many existing flat plate structures. Throughout the years, it is common for the function of certain buildings to
change. When these changes call for building modifications, this can sometimes cause an increase in gravity loading
that was not considered in the original design (Tian et al., 2008). Any building modifications that cause an increase
5.3
@Seismicisolation
@Seismicisolation
A. C. Hufnagel et al.
in loading must be accounted for by checking the capacity of the existing structure prior to use. Another source of
weakness is that the connections are assumed to take the majority of the shear resistance from the concrete, but
damage to the slab during a seismic event may reduce connection strength and shear capacity, thereby significantly
overestimating the shear strength of the connection.
ACI 369R-11
ACI 369R-11 is the American Concrete Institute’s Guide for Seismic Rehabilitation of Existing Concrete
Frame Buildings and Commentary. The guide was based on ASCE/SEI 41-06 and is intended to be used in
combination with the code in analyzing the behavior of existing and new concrete elements in buildings subjected to
seismic loading. The purpose of ACI 369R-11 is to provide recommendations for acceptance criteria and modeling
parameters that are necessary for the linear and nonlinear analysis of concrete components in the seismic
rehabilitation design process. This research focuses on Chapter 4, Section 4.4 of the guide, which discusses slab-
column moment frames and the analytical processes used in assessment. Section 4.1.3 of ACI 369R-11 defines these
frames as monolithically cast concrete structures with slabs, columns, and connections as the main structural
framing components, and with non-prestressed or prestressed reinforcement to serve as the primary reinforcement in
the slab. General considerations outlined in Section 4.4.1 of ACI 369R-11 state that stiffness, strength, and
deformation capacity of the building components should all be represented in analytical models for these frames.
This section also warns against the possibility of punching shear failure as a result of a combination of flexural,
shear, and torsional transfer at the connections.
Section 4.4.3 of ACI 369R-11 outlines the method for determining the flexural strength of a slab to resist
moment resulting from lateral deformations using Eq. (1).
A major portion of the research consisted of a review of past studies on interior reinforced concrete slab-
column connections subjected to combined lateral and gravity loading. This included individual interior connections,
as well as two-bay frames with interior connections. Other qualifications for relevant specimens include absence of
shear reinforcement in the slab in the form of either rebar reinforcement, like stirrups, shearbands or headed shear
studs, or a shear capital or drop panel.
In these past experiments, reinforced concrete slab-column connections composed of a square or rectangular
concrete slab with both an upper and lower column, most commonly measuring half a floor height each, were placed
in a tested rig and lateral loading was applied. In most cases, the connections were loaded cyclically and uniaxially,
in which a load was applied to one face at the top of the column up to a certain drift ratio, and then the process was
reversed in the opposite direction, simulating loads from a seismic event. This uni-directional cycle is typically
repeated at increasing levels of drift until failure of the connection. For specimens tested biaxially, force is applied
5.4
@Seismicisolation
@Seismicisolation
Modeling Parameters for Reinforced Concrete Slab-Column Connections
to the column in two principal directions. The order and configuration of the load applications are discussed with
each relevant specimen (Hufnagel, 2011).
The critical section perimeter of the interior connection (bo) was determined based on the information in the
text or figure of the literature. The critical section perimeter affects the shear capacity of the concrete (Vo) at the
connection, which is defined in ACI 369R-11 (and ACI 318-11) and typically determined using Eq. (2).
Figure 2 — Selected lateral load vs. drift curves for recently tested slab-column connections (ND4LL – Robertson
and Johnson, 2006; CO – Kang and Wallace, 2008).
5.5
@Seismicisolation
@Seismicisolation
A. C. Hufnagel et al.
Each specimen is investigated so as to determine whether or not punching shear failure could be predicted. The
connections are analyzed using the eccentric shear stress model for a slab width of c2 + 5h, the center of the column
strip, which should have the highest reinforcement ratio. The nominal moment of the slab over this width is
currently the representative recommendation from ACI 369R as to the yield moment of the slab. This width should
have the strength to transfer the unbalanced moment by flexure. If the applied shear stress exceeds the nominal shear
stress capacity of the concrete, a brittle failure may take place; otherwise, a flexural failure or flexure yielding
followed by drift-induced punching failure can be expected (Kang and Wallace, 2006). The eccentric shear stress
model was used in order to compare the nominal shear capacity of each connection to the applied shear stresses
[Eqs. (3) and (4)].
vc vu (3)
Vg M c
vu v u AB (4)
bo d Jc
where vc is the nominal shear stress capacity provided by the concrete as per ACI 318-11, vu is the total applied
shear stress, Vg is the applied gravity shear, v is the factor to determine the portion of unbalanced moment
transferred by shear, Mu is the unbalanced moment (M+n_c+5h + M–n_c+5h was used for this research), cAB is the
distance between the centroid of the column critical section to the edge of the critical section, and Jc is the property
of the assumed critical section analogous to the polar moment of inertia.
From the information collected during the review of previous experiments, the data are analyzed to estimate the
approximate value of allowable plastic rotation for each specimen. This includes the determination of the parameters
a and b, which, as defined in ACI 369R, are used to measure deformation capacity in component load-deformation
curves. These parameters are defined in Figure 3. Current specifications for these modeling parameters can be found
in Table 4.9 in ACI 369R-11 (Table 1 of this paper).
Figure 3 — Generalized force-deformation relations for concrete elements or components (ACI 369R-11).
In this paper, the modeling parameters a and b are calculated using Eqs. (5) and (6), respectively; however, the
method to determine the yield point should be discussed further in the ACI Committee 369.
where 0.95Peak and 0.2Peak are the drifts at a 5% drop from the peak (green dots in Figure 2) and at a 80% drop from
the peak (purple dots in Figure 2), respectively, and M is the drift at which the lateral load corresponds to
n _ c 5h
the sum of M+n_c+5h and M–n_c+5h during the ascending branch of the backbone curve (blue dots in Figure 2). Here,
M+n_c+5h and M–n_c+5h are design flexural strengths of the c2 + 5h width for positive and negative bending,
respectively.
5.6
@Seismicisolation
@Seismicisolation
Modeling Parameters for Reinforced Concrete Slab-Column Connections
In this paper, these parameters are compared to the reinforcement ratios (ρc+5h) for the width of c2 + 5h, as well
as to the gravity shear ratios (Vg/Vo) in order to develop a chart for recommendations to ACI 369R-11.
Table 1 — Modeling parameters and numerical acceptance criteria for nonlinear procedures – Two-way slabs and
slab-column connections (ACI 369R-11)
*
Values between those listed in the table should be determined by linear interpolation.
†
Primary and secondary component demands should be within secondary component acceptance criteria where the full
backbone curve is explicitly modeled, including strength degradation and residual strength, in accordance with Section 3.4.3.2
of ASCE/SEI 41-06.
‡
Where more than Condition i occur for a given component, use the minimum value from Table 4.9 of ACI 369R-11.
#
Action should be treated as force-controlled
A summary of the data recorded in the database during the literature review can be found in Table 2 for the
specimens with and without continuous bottom reinforcement. The results for predicted punching shear failure can
be found in Table 3. This table includes the values for nominal shear stress capacity (vc) and total applied shear
stress (vu_c+5h) calculated as per Eq. (4) and using unbalanced moment values of (M+n_c+5h + M–n_c+5h) at an interior
connection and M+n_c+5h or M–n_c+5h for an exterior connection depending on the direction of slab bending. The width
of c2 + 5h was evaluated, because this width typically falls inside the column strip or close to the width of the
column strip, and therefore has the highest amount of slab reinforcement. The stress applied at the column is
transferred away from the joint as the area weakens, so the region closest to the column is the most crucial. These
results do not take into consideration the presence or absence of continuous bottom reinforcement through the
column. While many of these specimens seemed to exhibit enough strength to avoid stress-induced punching shear
failure, the applied cyclic loading at higher drift levels resulted in drift-induced punching failure anyway, despite the
presence of continuity in reinforcement (Kang and Wallace, 2006). While connections are not expected to withstand
drift levels beyond a certain level, the maximum allowable inelastic drift needs to be quantified.
5.7
@Seismicisolation
@Seismicisolation
A. C. Hufnagel et al.
Table 2 — Summary of database for specimens with and without continuous bottom reinforcement
5.8
@Seismicisolation
@Seismicisolation
Modeling Parameters for Reinforced Concrete Slab-Column Connections
Based on the data and the results in Table 3, the nominal shear stress capacity provided by the concrete (vc)
appears to be quite conservative in terms of the stress-induced punching failure (that is, brittle punching shear failure
with limited slab bar yielding and ductility); therefore, the nominal stress capacity of 4√f’c (psi) or 0.33√f’c (MPa) is
recommended to be increased to some degree [e.g., 5√f’c (psi) or 0.42√f’c (MPa)] for more realistic modeling of
stress-induced punching and reasonable seismic evaluation of reinforced concrete interior slab-column connections
with typical slab dimension and reinforcement detailing.
Tables 3 and 4 show the calculations for modeling parameters a and b, respectively. The tables provide
information for specimens with and without continuous bottom reinforcement. The positive/negative sign represents
just the direction of lateral drift. Table 4 includes data for the modeling parameters with respect to gravity shear ratio
(Vg/Vo) only, like in the existing Table 4.9 in ACI 369R-11 (or Table 1 of this paper). Note that for interior slab-
column connection subassemblies with a pin condition at the column below, story drift due to column elastic
bending is typically negligible and no plastic deformation exists in the column (Pan and Moehle, 1992). Thus, the
plastic drift is assumed to be equal to the plastic rotation of the slab on either side of an interior slab-column
connection. The information shows the relationship between the measured modeling parameter and gravity shear
ratio. The values shown in Table 4 are mean values for the specimens with and without continuous bottom
reinforcement under variable gravity levels. However, because the data used in this study did not apply to each
gravity shear ratio, the data is lacking and should be expanded in future research studies. The relevant data including
the mean values are analyzed in this study to choose the best recommendation for each range.
Table 3 — Failure mode and modeling parameters (Gray shades: b is based on the maximum drift during testing)
C v/ f c a (%) b (%) Stress-Induced
Authors Specimen Vg/Vo Punching Expected? /
v=vc v=vu + - + - Occurred?
AP1 4 3.6 0.37 0.8 0.8 0.8 1.9 No / No
Pan and Moehle (1992) AP3 4 2.7 0.18 3 2.9 4.6 3.0 No / No
AP4 4 2.7 0.19 2.7 1.4 4.0 0.9 No / No
Robertson and Durrani 2C 4 5.6 0.20 3.4 3.4 4.4 4.1 Yes / No
(1991) 7L 4 6.7 0.40 0.5 0.7 1.8 2.0 Yes / Yes
3 4 3.0 0.20 3.0 2.0 4.4 2.0 No / No
Farhey et al. (1993)
4 4 4.9 0.20 2.2 1.8 4.9 1.8 Yes / No
Robertson et al. (2002) 1C 4 2.4 0.17 2.4 1.9 2.9 2.8 No / No
Zee and Moehle (1984) Interior 4 4.1 0.14 2.9 2.6 2.9 2.6 Yes / No
Wey and Durrani SC0 4
0.05 2.3 2.6 4.2 4.2 No / No
(1992) 2.5
Kang and Wallace CO 4
0.30 2.3 1.2 3.9 3.4 No / No
(2008) 3.1
ND1C 4 2.7 0.20 3.5 2.5 7.1 4.4 No / No
ND4LL 4 2.8 0.30 1.9 1.8 3.5 3.4 No / No
Robertson and Johnson
ND5XL 4 3.7 0.50 0.6 0.6 0.6 0.6 No / No
(2006)
ND6HR 4 4.2 0.30 1.1 1.4 3.2 3.3 Yes / No
ND7LR 4 2.5 0.30 2.0 2.2 3.7 3.9 No / No
L0.5 4 1.9 0.23 1.3 1.5 1.7 1.7 No / No
Tian et al. (2008) LG0.5 4 1.9 0.23 1.0 1.0 1.0 1.0 No / No
LG1.0 4 2.5 0.23 1.0 0.9 1.1 1.0 No / No
Stark et al. (2005) C-63 4 2.5 0.10 0.6 1.0 1.6 2.0 No / No
Durrani et al. (1995) DNY_3 4 4.0 0.22 2.6 2.4 3.7 3.2 Yes / No
Luo et al. (1994) II 4 1.6 0 3.8 3.9 3.8 3.9 No / No
5.9
@Seismicisolation
@Seismicisolation
A. C. Hufnagel et al.
Modeling Parameters
Conditions
Plastic Rotation, Radians
Continuity a b
x = Vg/Vo
Reinforcement ACI 369R-11 Mean Proposed ACI 369R-11 Mean Proposed
0 ≤ x < 0.2 0.035 0.025 0.025 0.050 0.032 0.035
0.2 ≤ x < 0.4 0.030 0.021 0.020 0.040 0.032 0.03
Yes
0.4 ≤ x < 0.6 0.020 0.006 0.015 0.030 0.019 0.025
x ≥ 0.6 0 - - 0 - -
0 ≤ x < 0.2 0.025 0.023 0.023 0.025 0.028 0.025
0.2 ≤ x < 0.4 0.020 0.018 0.018 0.020 0.029 0.020
No
0.4 ≤ x < 0.6 0.010 0.006 0.008 0.010 0.006 0.010
x ≥ 0.6 0 - - 0 - -
The review of previous experiments on reinforced concrete slab-column connections subjected to lateral
loading resulted in a database of test details including dimensions, detailing, and calculated moment strength. This
database will be useful as this research continues to expand to include more specimens that are relevant to the study.
It also enabled the comparison of several different factors in order to determine which factors may contribute to the
punching shear strength and allowable plastic drift capacity for certain connections.
Many comparisons were made between the different variables of interest in this study in order to determine
trends in the data (Figures 4 to 7). Both positive and negative story drifts are plotted in Figures 4 to 7. The
correlation between gravity shear ratio and allowable plastic drift levels was already known, so it was analyzed in
order to provide recommendations for ACI 369R, which are discussed below. Figure 4 shows the relationship
between the gravity shear ratio and the measured modeling parameter a. As the applied gravity shear ratio increases,
the modeling parameter a tends to decrease for both the specimens with and without continuous bottom
reinforcement. Similar comparisons were made for the modeling parameter b. A trend of a positive correlation
between gravity shear ratio and the measured modeling parameter b was noticed (Figure 5). The scatter was
particularly large for the parameter b. Note that the “maximum” drift value was used to determine b for more than
half of the specimens, because the testing was stopped without significant loss of strength in order to avoid the
complete collapse of the specimen.
The comparisons also include an analysis between the reinforcement ratios and the modeling parameters, which
showed little trend (Figures 6 and 7). Figures 6 and 7 show the modeling parameters with respect to reinforcement
ratio (ρc+5h), in order to determine the relationship between allowable plastic drift and the level of reinforcing steel at
the face of the interior column. The total reinforcement ratio, top and bottom included, for the slab width of c2 + 5h
was used, and the reinforcement ratio is shown as a percentage (%). These figures also have some missing
information for lower reinforcement levels, but this is mainly due to minimum code requirements for reinforcement
amounts. The critical section area was also compared to the modeling parameters with the thought that larger joint
proportions would allow increased levels of plastic deformation; however, no correlation was noted in the data. This
may be due to many different variables that are taken into account in testing a slab-column specimen, so the direct
comparison of two joints differing only in critical section area was not possible at this time.
5.10
@Seismicisolation
@Seismicisolation
Modeling Parameters for Reinforced Concrete Slab-Column Connections
(a) with continuous bottom reinforcement (b) without continuous bottom reinforcement
Figure 4. Gravity shear ratio vs. measured modeling parameter a.
(a) with continuous bottom reinforcement (b) without continuous bottom reinforcement
Figure 5. Gravity shear ratio vs. measured modeling parameter b.
(a) with continuous bottom reinforcement (b) without continuous bottom reinforcement
Figure 6. Reinforcement ratio vs. measured modeling parameter a.
5.11
@Seismicisolation
@Seismicisolation
A. C. Hufnagel et al.
(a) with continuous bottom reinforcement (b) without continuous bottom reinforcement
Figure 7. Reinforcement ratio vs. measured modeling parameter b.
(a) with continuous bottom reinforcement (b) without continuous bottom reinforcement
Figure 8. Gravity shear ratio vs. proposed modeling parameter a.
(a) with continuous bottom reinforcement (b) without continuous bottom reinforcement
Figure 9. Gravity shear ratio vs. proposed modeling parameter b.
5.12
@Seismicisolation
@Seismicisolation
Modeling Parameters for Reinforced Concrete Slab-Column Connections
The majority of the specimens also had continuous bottom reinforcement, which should be used to prevent
collapse and improve punching shear capacity (by providing dowel action and connection integrity and by
increasing flexural transfer capacity of unbalanced moment). The eccentric shear stress model does not take into
account the lack or presence of continuity in bottom reinforcement through the column. This is important, because
an insufficient development length may lead to brittle and unexpected failure by pullout of the rebar, which in turn
could lead to a progressive collapse of an entire structure. This emphasizes the importance of redundancy in
structural design. It is also important to know the varying strength of the slab with increasing flexural transfer
widths, in order to better understand the transfer mechanism and failure behavior of a structure. It is important that a
slab region around a column has the ability to transfer the load to the column, rather than punch suddenly and
catastrophically under load application. ACI 369R recommends the use of a slab width of c2 + 5h for flexural
transfer width, unlike ACI 318 where a width of c2 + 3h is specified. This study follows the current recommendation
of ACI 369R, because the database suggests that the c2 + 5h width is more reasonable.
Figures 8 and 9 compare the values preliminarily proposed from this research to the current values for the
modeling parameters a and b that are recommended in ACI 369R. The values in Figures 8 and 9 and Table 4 are
mainly based on the mean values for both specimens with and without continuity in reinforcement, with some
engineering judgment applied to address cases where limited data and significant scatter exist (see Figures 4 and 5).
The standard deviations of the errors between the experimental values and estimated modeling parameters are also
provided (Table 5). The proposed values follow the same basic trends as the current numbers, in that with increasing
gravity shear ratio, the modeling parameters decrease. However, due to unconservative estimates of the plastic
rotation capacity, smaller values are proposed for both the modeling parameters a and b. The median values are
quite close to the mean values (Table 6), so the use of the median would not affect the proposed values for
reinforced concrete slab-column connections. For specimens with a gravity shear ratio of 0.6 or greater, ACI 369R
typically recommends zero allowable plastic rotation angles for all categories. The research results did not include
any specimens that fell in this category, but with such a high amount of gravity loading, it appears that zero would
be an appropriate and conservative recommendation. In general, the updated numbers are smaller than those
currently recommended, meaning that the ACI 369 recommendations may have areas that are unconservative, and
therefore may not be safe for reinforced concrete slab-column connections.
In terms of modeling parameter c, residual strength ratio (see Table 1), at this point no updates are made to the
current values. Using the compiled database, acceptance criteria can also be assessed. ASCE/SEI 41-06 provides
procedures to define acceptance criteria for the Immediate Occupancy (IO), Life Safety (LS) and Collapse
Prevention (CP). For primary members, acceptance criteria for LS and CP are defined as 0.75(a) and the larger of
1.0(a) and 0.75(b), respectively. For secondary members, acceptance criteria for LS and CP are defined as 0.75(b)
and 1.0(b), respectively. However, to have the same (or similar) probability of exceeding a modeling parameter a or
b for various members/connections that meet acceptance criteria, ACI Committee 369 is proposing a new procedure
utilizing specific percentiles. While specific numbers are still being discussed, the same or similar procedure will be
applicable to the assessment of slab-column connections’ acceptance criteria. In Table 6, a few selected percentiles
are presented to provide a general idea.
Table 5 — Standard deviations of errors between experimental values and estimated modeling parameters (proposed
and ACI 369R-11) for a and b in %
Continuity
Vg/Vo STDa_proposed STDa_ACI369 STDb-proposed STDb_ACI369
Reinforcement
0 ≤ x < 0.2 0.47 1.13 1.07 2.07
0.2 ≤ x < 0.4 0.93 1.30 1.35 1.58
Yes
0.4 ≤ x < 0.6 1.17 1.40 0.61 1.10
x ≥ 0.6 - - - -
0 ≤ x < 0.2 1.53 1.54 1.09 1.09
0.2 ≤ x < 0.4 0.72 0.76 1.84 1.84
No
0.4 ≤ x < 0.6 0.20 0.40 0.40 0.40
x ≥ 0.6 - - - -
5.13
@Seismicisolation
@Seismicisolation
A. C. Hufnagel et al.
CONCLUSION
The primary purpose of the research was to review the experimental data of previously tested reinforced
concrete slab-column connections and to recommend potential updates to ACI 369R concerning modeling
parameters for allowable plastic rotations based on the investigation. The thorough review of previous and recent
experiments resulted in a database of useful information, which was used in further calculations and analysis of the
punching shear strength of reinforced concrete slab-column connections in flat plate systems. This study showed
that the nominal punching shear strength of connections subjected to combined seismic and gravity loading provides
a conservative estimate and that conversely the current modeling parameters for plastic deformation outlined in ACI
369R are unconservative. This would help provide more reasonable and safer guidelines for seismic evaluation and
rehabilitation, and may work to limit the damage caused to rehabilitated flat plate systems by strong seismic events.
This research was also successful in confirming that flat plate concrete structures without continuous
reinforcement are more susceptible to punching shear failure, and therefore have lower recommended allowable
plastic deformation values. The modeling parameters correlate with different values of gravity shear ratio and
reinforcement ratio; however, the trends suggest that the main variable of interest is the gravity shear ratio. Little
trend was noticed in the data between the reinforcement ratio and the modeling parameters. Given that the current
ACI 369 recommendations for plastic rotation values for slab-column connections were found to be unconservative,
the authors have made preliminary proposals on the values of a and b based on limited experimental data.
This study was especially important in pointing out the lack of information available for this area of research.
While many studies have taken place, the large amount of different variables among the specimens allows for less
comparable data between the specimens, especially for those without continuity in bottom reinforcement. Although
buildings are not designed this way anymore, it is important to determine if those that currently exist with non-
ductile connections have the strength to survive a seismic event or if retrofit methods are necessary. Further analysis
and specimen testing would be useful in expanding this study to provide more accurate recommendations to ACI
369R.
ACKNOWLEDGEMENTS
Financial support from Seoul National University is greatly appreciated. The authors acknowledge members of
ACI Committee 369 who provided constructive comments during the committee meetings, particularly John
Wallace, Ying Tian, Ken Elwood, Wassim Ghannoum, Roberto Stark, Murat Melek, and Mary Beth Hueste.
REFERENCES
ACI Committee 318 (2011), Building Code Requirements for Structural Concrete (ACI 318-11) and Commentary,
American Concrete Institute, Farmington Hills, MI.
5.14
@Seismicisolation
@Seismicisolation
Modeling Parameters for Reinforced Concrete Slab-Column Connections
ACI Committee 369 (2011), Guide for Seismic Rehabilitation of Existing Concrete Frame Buildings and
Commentary (ACI 369R-11), American Concrete Institute, Farmington Hills, MI.
American Society of Civil Engineers/Structural Engineering Institute (ASCE/SEI) Committee 41, Seismic
Rehabilitation of Existing Structures, in ASCE Standard 2007, Reston, VA.
American Society of Civil Engineers/Structural Engineering Institute (ASCE/SEI) Committee 41, Supplement 1 to
ASCE 41, in ASCE Standard 2007, Reston, VA.
Dovich, L. and Wight, J. K. (1996), “Lateral Response of Older Flat Slab Frames and the Economic Effect on
Retrofit,” Earthquake Spectra, Vol. 12, No. 4, pp. 667-691.
Durrani, A. J., Du, Y., and Luo, Y. H. (1995), “Seismic Resistance of Nonductile Slab-Column Connections in
Existing Flat-Slab Buildings,” ACI Structural Journal, Vol. 92, No. 4, pp. 479-487.
Farhey, D. N., Adin, M. A., and Yankelevsky, D. Z. (1993), “RC Flat Slab-Column Subassemblages Under Lateral
Loading,” ASCE Journal of Structural Engineering, Vol. 119, No. 6, pp. 1903-1916.
Hufnagel, A. C. (2011), “Analytical Studies of Reinforced and Post-Tensioned Concrete Flat-Plate Systems,” MS
thesis, School of Civil Engineering and Environmental Science, University of Oklahoma, Norman, OK.
Kang, T. H.-K. and Wallace, J. W. (2006), “Punching of Reinforced and Post-Tensioned Concrete Slab-Column
Connections,” ACI Structural Journal, V. 103, No. 4, pp. 531-540.
Kang, T. H.-K. and Wallace, J. W. (2008), “Seismic Performance of Reinforced Concrete Slab-Column Connections
with Thin Plate Stirrups,” ACI Structural Journal, Vol. 105, No. 5, pp. 617-625.
Kang, T. H.-K., Wallace, J. W., and Elwood, K. J. (2009), “Nonlinear Modeling of Flat-Plate Systems,” ASCE
Journal of Structural Engineering, V. 135, No. 2, pp. 147-158.
Luo, Y. H., Durrani, A. J., and Conte, J. P. (1994), “Equivalent Frame Analysis of Flat Plate Buildings for Seismic
Loading,” ASCE Journal of Structural Engineering, Vol. 120, No. 7, pp. 2137-2155.
Moehle, J. P., Kreger, M. E., and Leon, R. (1988), “Background to Recommendations for Design of Reinforced
Concrete Slab-Column Connections,” ACI Structural Journal, Vol. 85, No. 6, pp. 636-644.
Pan, A. and Moehle, J. P. (1989), “Lateral Displacement Ductility of Reinforced Concrete Flat Plates,” ACI
Structural Journal, Vol. 86, No. 3, pp. 250-258.
Pan, A. and Moehle, J. P. (1992), “An Experimental Study of Slab-Column Connections,” ACI Structural Journal,
Vol. 89, No. 6, pp. 626-638.
Robertson, I. N. and Durrani, A. J. (1991), “Gravity Load Effect on Seismic Behavior of Exterior Slab-Column
Connections,” ACI Structural Journal, Vol. 88, No. 3, pp. 255-267.
Robertson, I. and Johnson, G. (2006), “Cyclic Lateral Loading of Nonductile Slab-Column Connections,” ACI
Structural Journal, Vol. 103, No. 3, pp. 356-364.
Robertson, I. N., Kawai, T., Lee, J. and Enomoto, B. (2002), “Cyclic Testing of Slab-Column Connections with
Shear Reinforcement,” ACI Structural Journal, V. 99, No. 5, pp. 605-613.
Stark, A., Binici, B., and Bayrak, O. (2005), “Seismic Upgrade of Reinforced Concrete Slab-Column Connections
Using Carbon Fiber-Reinforced Polymers,” ACI Structural Journal, Vol. 102, No. 2, pp. 324-333.
5.15
@Seismicisolation
@Seismicisolation
A. C. Hufnagel et al.
Tian, Y., Jirsa, J. O., Bayrak, O., Widianto, and Argudo, J. F. (2008), “Behavior of Slab-Column Connections of
Existing Flat-Plate Structures,” ACI Structural Journal, Vol. 105, No. 5, pp. 561-569.
Wey, E. H. and Durrani, A. J. (1992), “Seismic Response of Interior Slab-Column Connections with Shear
Capitals,” ACI Structural Journal, Vol. 89, No. 6, pp. 682-691.
Zee, H. L. and Moehle, J. P. (1984), “Behavior of Interior and Exterior Flat Plate Connections Subjected to Inelastic
Load Reversals,” Report No. UCB/EERC-84/07, Earthquake Engineering Research Center, University of
California, Berkeley, CA.
5.16
@Seismicisolation
@Seismicisolation
SP-297—6
Abstract
Rehabilitation of structures requires knowledge on the anticipated nonlinear
behavior of building components. Column retrofitting using jackets made of
different materials, designed to counter deficiencies in the original design, can be
used to improve the overall performance of an existing building. This paper
presents a procedure to obtain the backbone nonlinear response of jacketed
columns based on existing experimental results for this class of elements. Column
jacketing has been used in the past to increase ductility and strength of reinforced
concrete columns. Jacket designs can be varied to selectively improve the lateral-
load response of columns depending on the original design deficiency. A uniform
methodology or specific recommendations on how to estimate the response of
jacketed columns with different jacket materials types do not exist. Existing
experimental data are used to determine the backbone nonlinear response of
jacketed columns. These data are then used to develop recommendations to
determine non-linear force and deformation parameters of jacketed columns.
Although the most common types of jackets used are made from: concrete, steel,
and FRP, only the last two are included in this paper. The backbone curves thus
generated can then be used in performance assessment of reinforced concrete
buildings where columns have been jacketed.
Introduction
Performance-based design and seismic rehabilitation of structures requires a thorough
understanding of the expected response of a structure under various levels of seismic demand.
The elastic and inelastic cyclic behavior of structural components needs to be well understood in
order to be able to get better estimates of overall structure response under different ground
shaking scenarios. Current practice is to approximate the nonlinear cyclic response of structural
components by using force-deformation curves that bound the hysteretic response of these
components (backbone curves). To model the nonlinear response of components, average
backbone curves are simplified by determining only those points that impact the response of the
component importantly. These simplified backbone curves are constructed using values that
6.1
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
represent the mean response of components tested in the laboratory. Modeling parameters in
ASCE/SEI Standard 41-06 (2006), for example, allow users to determine force and deformation
values at yield, at peak strength, and at residual strength. Multi-linear backbone curves are
constructed by joining these three key points. Backbone curves for positive and negative
directions of loading are assumed identical for components with symmetrical cross-section and
reinforcement.
Modeling parameters included in ASCE/SEI 41-06 and ACI 369R-11 were developed for
existing structural components. In a seismic rehabilitation project, the ability to estimate the
response of the retrofitted structure is important to assess the success of a proposed intervention.
Nonlinear modeling of rehabilitated components is therefore needed. A common feature of
existing frame structures that predate modern seismic procedures is the presence of columns with
details that make them vulnerable to collapse during earthquakes. Columns in these frames may
contain short lap splices of longitudinal reinforcement, inadequate shear strength, and/or
insufficient confinement to resist the anticipated seismic demands. Column jacketing is a
technique that has successfully been used in the past, as shown by experiments, to correct these
design and detailing deficiencies. There is little guidance, however, on procedures to determine
nonlinear modeling parameters for jacketed columns so the seismic response of retrofitted frames
cannot be determined consistently. This paper provides recommendations for construction of
backbone curves by determining modeling parameters of columns retrofitted using fiber-
reinforced polymer (FRP) jackets or steel jackets. Backbone curves of jacketed columns are first
constructed using existing experimental data to determine force-deformation parameters at yield
and peak strength. These data are then used for comparison with values determined from models
proposed to determine relevant force-deformation parameters of jacketed columns.
All columns in the database were tested under cyclic static loading applied incrementally
following a prescribed protocol. Columns were tested in original (un-retrofitted) condition and
retrofitted using jackets made with fiber-reinforced polymer materials or steel. The concrete
compressive strength of columns in the database ranged between 1300 and 8000 psi (9 and 55
MPa). Columns were tested under either single curvature bending (cantilever) or double
curvature bending (fixed-fixed). Although some tests were intended to simulate bridge column
behavior, data are still included in the database for these specimens.
The database contains 146 columns of which 52 are circular columns and 94 are
rectangular columns. Figure 1 summarizes the classification of columns in the database
6.2
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
according to jacket type and type of deficiency encountered in the original design. Figure 2
illustrates the distribution of key parameters of the columns in the database. The parameters
presented in this figure are spacing of transverse reinforcement (s) normalized by effective depth
to the tension force resultant, d, assumed equal to 0.8 dc; axial force ratio, defined as the axial
load divided by the nominal compressive strength of the concrete (f’c) and the gross cross-
sectional area (Ag); type of jacket used to retrofit each column (steel or FRP); the ratio between
shear at plastic hinging (Vo) and nominal shear strength (Vn). Nominal shear strength was
calculated using the shear strength equation 4-1 in ACI 369R-11 using a ductility of 8.
6.3
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
A short lap splice condition refers to splicing longitudinal reinforcement within the
plastic hinge region of the column over a length of 20 to 36 longitudinal bar diameters. Columns
with inadequate shear strength are those that fail in shear prior to developing the plastic hinge
capacity of the column at high displacement demands. Inadequate confinement of longitudinal
reinforcement is caused by a low volume of transverse reinforcement because of either small
diameter hoops and/or hoops at a large spacing. Poor seismic detailing of hoops with 90° hooks
at their ends also leads to inadequate confinement after cover concrete spalls at moderate
displacement demands.
The column database includes information about the material properties and geometry of
the jacket. The diameter of circular columns are between 9.5 and 30 in. (241 and 762 mm), while
the width of rectangular columns is between 6 and 36 in. (152 and 914 mm) and the depth is
between 7.5 and 36 in. (191 and 914 mm). Key jacket properties include jacket material (steel or
FRP), jacket thickness, length of jacket, and jacket mechanical properties. The properties of the
jacket vary widely depending on the retrofitting objective and the original column deficiency.
Appendix A includes tables that list force-deformation parameters of individual columns in the
database. Force-deformation values are provided at three levels: yield force, peak force, and
maximum force and deformation. Force is given as applied lateral load, V, at these different
levels. Peak values are those measured at the highest force applied to the specimens. Maximum
values are those measured at the maximum deformation imposed to the specimens. The
deformation parameter used in the tables and throughout this paper was drift, given as lateral
displacement divided by specimen height.
Backbone curves were constructed by digitizing data from figures of hysteresis plots
included in the original test references. The data were digitized by obtaining the force and
displacement (or drift) at each displacement level applied to the columns. Force and
6.4
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
displacement measured during the first cycle at each displacement level were used to construct
these curves. If authors directly reported backbone curves, then these plots were digitized to
obtain the force-deformation relationships instead of extracting them from hysteresis curves.
Generation of a backbone curve from the hysteresis response of one column in the database is
illustrated in Figure 3. The key points needed to construct simplified backbone curves were
obtained from these curves as illustrated in the figure; they are listed by column specimen in
Appendix A.
The coordinates at peak force (VPeak, ΔPeak) were determined by using the average of the
peaks measured in the positive and negative directions of loading. The residual force (Vres) and
maximum displacement (max) are not frequently reported in the literature. Therefore Vres was
defined as a fraction (20%) of the peak force (Vpeak) and max was defined conservatively as the
maximum displacement reported for each tested column. The definitions of nonlinear modeling
parameters (a, b, and c) in ASCE/SEI 41-06 were used, wherever possible, to construct the
simplified backbone curves of jacketed columns. Parameter a is defined in ASCE/SEI 41-06 as
the difference between the deformation at lateral strength degradation of 20% and the
deformation at yield. In some tests the columns were not displaced to cause a 20% strength
degradation. In these cases parameter a was defined using the maximum deformation imposed
during the test. Parameter b is defined in ASCE/SEI 41-06 as the deformation corresponding to
loss of axial capacity, accompanied by a significant drop in lateral-load strength. Most of the
jacketed columns were not tested to this level of strength degradation, so parameter b was
6.5
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
determined as the deformation corresponding to a degradation of 25% from Vpeak. In cases where
columns did not exhibit a strength degradation level of 20% or more, parameter b could not be
determined and is not reported in Appendix A.
Effect of Column Jacketing on Yield Force, Peak Strength, and Flexural Stiffness
The key parameters that define simplified backbone curves of jacketed columns need to
be calculated reliably for use in nonlinear analyses. Techniques used to determine these
parameters for jacketed columns must be developed to allow evaluation of the response of
retrofitted frames. Models that can be used to determine the yield force and peak strength of
jacketed columns are presented in this section. Because of lack of accurate deformation models,
drift at these key points may be based on results collected from the column database, so no
model was developed.
Depending on the jacket material used and the way it is connected to the existing
concrete column, jackets may also contribute to the flexural and shear strength of the column.
The two types of jackets discussed in this paper, steel and FRP, contribute in different degrees to
confinement, shear, and flexural strength. Steel jackets may contribute to the shear strength and
flexural strength of the columns since steel mechanical properties can be considered as isotropic.
FRP jackets, on the other hand, primarily contribute to strength and stiffness in the direction of
6.6
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
fiber orientation. Fibers in FRP jackets used for confinement and shear strength increase are
applied transversely to the column axis so their contribution to flexural strength is negligible.
Only the effects of jacket confinement are considered here; contribution of jackets to shear
strength is not discussed in detail.
Steel jackets contribute to bending stiffness of the columns due to the increase in cross-
sectional dimensions (grout between original column and steel jacket) and the contribution of the
external steel jacket to bending stiffness. The contribution of FRP jackets to bending stiffness is
negligible so it was therefore not considered. Stiffness reductions to account for concrete
cracking (ACI 369R-11, Table 3.1) were used to modify the gross cross-sectional stiffness of
jacketed columns.
Short lap splices or inadequate confinement of existing older columns may cause
localized failures primarily within the plastic hinge region at low ductility demands. The global
response of a column with these deficiencies is dominated by the behavior within the plastic
hinge region. In these cases it is appropriate to use a sectional model within the plastic hinge
zone to capture the global column behavior. The goal of localized column jacketing is to
eliminate a brittle splice failure from occurring within the plastic hinge region so that the global
response of the column becomes ductile. The jackets, however, may not necessarily provide the
level of clamping force necessary to develop yielding of longitudinal bars in the spliced region.
The stress developed in the spliced reinforcement was therefore reduced using equation 3-2 of
the ACI 369R-11, which considers the effect of bar slip in stress reduction. This modification
resulted in better agreement with the measured backbone behavior.
Column yield and peak force were estimated by including the effects of concrete
confinement using existing models with some modification to account for the properties of the
jackets. The cross–sectional behavior of a jacketed column was calculated by dividing the
section into fibers. The contribution of each fiber to flexural capacity at various deformations
(curvatures) was calculated using the uniaxial stress-strain behavior of the material composing
6.7
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
the fiber. For a jacketed column fibers representing reinforcing steel, confined concrete, and
jacket were used. Unconfined concrete fibers were not included since jackets are applied
externally so all the concrete in the cross section is confined.
(1)
and for circular FRP jackets, the lateral confining stress was calculated as:
(2)
where,
To allow use of the popular confinement model developed by Mander et al. (1988),
lateral confining stressed developed by steel or FRP jackets was converted to an equivalent
quantity of transverse reinforcement that would result in the same lateral confining stress. The
jacket confining stresses were equated to the lateral confining stress equation in the Mander et al.
(1988) model to determine the spacing required of No. 3 bar hoops that would result in the same
confining stress using:
∙ ∙ (3)
where ke equals
′
(4)
ke represents the confinement efficiency factor that accounts for the reduced confining
effect of discretely spaced transverse hoops at spacing s. The spacing s' refers to the clear
spacing between transverse hoops. The ratio between the area of longitudinal reinforcement (As)
to concrete area (Ac), denoted as cc, may be taken equal to zero in the case of jacketed columns.
6.8
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
Equations (3) and (4) are solved for s for an assumed area of transverse reinforcement Atr, in this
case taken equal to 0.11 in2 for No. 3 transverse reinforcing bars.
2 (5)
2 (6)
The lateral confining stress of the equivalent elliptical jacket is calculated using the
equations proposed by Priestley et al. (1994), for confining stresses in the strong direction:
(7),
(8)
6.9
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
For rectangular FRP jackets, the column cross section was transformed into an equivalent
circular cross section instead of an ellipse given the higher flexibility of this material. The
diameter of the equivalent section was calculated equal to the diagonal dimension of the column
cross section, or:
(9)
Following the same philosophy as for circular columns, the confining stress generated by
the steel or FRP jackets was equated to the confining stress provided by an equivalent amount of
transverse hoops determined using the model by Mander et al. (1988). The difference in cross-
sectional dimensions, and the difference in longitudinal and transverse reinforcement
arrangement in the two directions of rectangular columns result in different confining stresses in
each direction. The effective confining stress provided by transverse reinforcement in the x and
y directions of a rectangular column is given by:
(10)
(11)
where x and y represent the transverse reinforcement content in the x (Atr-x/s·dc) and y
(Atr-y/s·bc) directions, respectively; fyh is the yield stress of transverse hoops; and ke is an
effectiveness factor to account for unconfined concrete zones in the vertical direction between
hoops and in the horizontal direction between longitudinal reinforcing bar positions. This
effectiveness factor may be estimated using:
∑ ′ ′
1 1 1 (12)
where bc and dc are the width and depth of the rectangular column. The term Σwi2 is used
to evaluate the reduction in confinement efficiency between longitudinal bars of the confined
core. Rectangular jackets were assumed restrained only at the column corners so the length of
each side of the column cross-section was used to estimate the ∑wi2 variable, resulting in a value
equal to 2 2 . Substituting this value into equation
′ ′
1 1 1 (13)
Using the area of No. 3 hoops for Atr-x and Atr-y the spacing required to give an equivalent
confining stress as generated by steel or FRP jackets was determined from equations (10), (11),
and (13).
Once the confining stress acting on the concrete was estimated, the confined concrete
material properties were determined using the Mander et al. (1988) model. Confined concrete
properties were used throughout the cross section to compute the moment-curvature response. It
6.10
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
was assumed that the response of each column was governed by the sectional behavior within the
plastic hinge region. Using this assumption, the force-displacement response was calculated for
each column in the database where the plastic deformation concentrated in the column plastic
hinge region.
Comparison between Calculated and Measured Column Yield and Peak Capacities
The calculated forces at yield and at peak of the columns in the database were compared
with measured values. An assessment of the proposed force models was conducted to verify the
assumptions that were made and propose modifications if necessary. The proposed models only
incorporate the effects of confinement generated by the column jackets. Future efforts will
concentrate on determining the contribution of jackets to column shear strength.
A comparison between calculated and experimental yield force of the columns in the
database is shown in Figure 6. A similar comparison is given in Figure 7 for the peak force in
the columns. The data are divided into circular and rectangular columns because the confining
effects are markedly different between these two cross-sectional shapes. The data shown
indicate that the proposed models match the experimental data much better in the case of circular
columns than rectangular columns, but significant scatter is observed. Bending flexibility of the
jackets in rectangular columns, which is not captured adequately by the proposed models, may
be partially responsible for a better fit of model predictions to experimental values. The
proposed model overestimates the effect of confinement leading to higher calculated forces at
yield and peak. In some of the columns tested with partial jackets, failure moved to a section
above the jacket causing the differences observed between calculated and measured force values.
The columns in the database include details leading to three different types of
vulnerabilities and various jacket configurations. The lack of uniformity in the original column
designs and geometry and configuration of jackets used to correct the column vulnerabilities
introduces variability into the database that is difficult to capture with a single model. However,
a modification to the model by a reduction in the jacket confining efficiency seems warranted.
Summaries of the mean and standard deviation of the ratio of model predictions to
experimental values of force at yield and peak are given in Table 1 and Table 2 for circular and
rectangular columns, respectively. The values in these tables show that the discrepancies
between measured and calculated values vary significantly by jacket type and type of original
column vulnerability. Much better correlation in calculated to measured values is obtained if the
entire column database is considered regardless of vulnerability in the original design (Table 3).
6.11
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
Columns with Steel Jackets Columns with FRP Jackets
Circular Columns
Lap Splice Shear Confinement Lap Splice Shear Confinement
Mean STD Mean STD Mean STD Mean STD Mean STD Mean STD
Vy calc/Vy test ‐ ‐ 0.95 0.02 1.17 0.11 1.43 0.33 1.24 0.06 ‐ ‐
Vpeak calc/Vpeak test ‐ ‐ 0.98 0.04 1.13 0.17 1.23 0.24 0.99 0.03 ‐ ‐
6.12
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
Columns with Steel Jackets Columns with FRP Jackets
Vy calc/Vy test Vpeak calc/Vpeak testt Vy calc/Vy test Vpeak calc/Vpeak test
Mean 1.09 1.08 1.38 1.16
All Circular Specimens
STD 0.14 0.15 0.30 0.23
Mean 1.25 1.12 1.56 1.14
All Rectangular Specimens
STD 0.35 0.18 0.61 0.20
The database includes steel and FRP jacketed columns of circular or rectangular cross-
section. A large variety of jacket configurations and design parameters were found in the
literature because of the different design vulnerabilities that original columns had. Existing
models intended to predict the load capacity of columns were modified to include the effects of
the various jacket designs that may be encountered in practice. In this paper, focus is given to
the effect of confinement on column strength and ductility. Shear strength enhancement was not
included in the current discussion. Following are the main observations that can be drawn from
the work conducted to date:
6.13
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
3. The ratio of peak strength to yield force of jacketed circular columns is on average higher
than for rectangular columns. This is caused by the higher confining efficiency of
circular jackets.
4. Drifts at yield and peak varied depending on jacket type and original column deficiency.
Columns with short lap splices within the plastic hinge region had higher deformations at
the same loads as those with continuous bars because of slippage of spliced bars.
5. The maximum deformation of columns varied significantly, but no firm conclusion can
be reached regarding this parameter. Most columns found in the literature were not
tested to displacements corresponding to significant loss of lateral force capacity. Many
tests were stopped prior to or when strength had degraded to 80% of peak because of
equipment limitations.
6. The proposed models used to predict strength of jacketed columns were able to estimate
the yield force and strength of columns in the database with different degrees of
accuracy. Models for jacketed circular columns are more accurate than those used for
jacketed rectangular columns. Strength was better predicted than yield with the models
proposed in this paper. More work is needed to develop physical models that produce
more accurate results.
The number of columns found in the literature for each design category and jacket type was
in many cases too small to allow making firm conclusions. However, the observations given
above may serve as a guide on the anticipated performance of jacketed columns and allow
development of consistent modeling parameters for different types of jacketed columns. It is
clear that more tests are necessary to develop comprehensive design recommendations on this
important class of rehabilitated component.
6.14
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
References
Aboutaha, R. S., Engelhardt, M. D., Jirsa, J. O., and Kreger, M. E. (1999). "Rehabilitation of
Shear Critical Concrete Columns by Use of Rectangular Steel Jackets". ACI Structural
Journal. V. 96, No. 1, 68-78.
Aboutaha, R. S., Engelhardt, M. D., Jirsa, J. O., and Kreger, M. E. (November 1996). “Retrofit
of Concrete Columns with Inadequate Lap Splices by the Use of Rectangular Steel Jackets”.
Earthquake Spectra, V. 12, No. 4 , 693-714.
Aboutaha, R. S., Engelhardt, M. D., Jirsa, J. O., & Kreger, M. E. (1999). “Experimental
Investigation of Seismic Repair of Lap Splice Failures in Damaged Concrete Columns”. ACI
Structural Journal, V. 96, No. 2 , 297-307.
ACI Committee 369. (2011). “Guide for Seismic Rehabilitation of Existing Concrete Frame
Buildings”, ACI 369R-11, American Concrete Institute, Farmington Hills, MI.
ACI Committee 440. (2008). “Guide for the Design and Construction of Externally Bonded FRP
Systems for Strengthening Concrete Structures”, ACI 440.2R-08, American Concrete
Institute, Farmington Hills, MI.
Breña, S. F., and Schlick, B. M. (2007). "Hysteretic Behavior of Bridge Columns with FRP-
Jacketed Lap Splices Designed for Moderate Ductility Enhancement".Journal of Composites
for Construction. V. 11, No. 6, 565-574.
Chai, Y. H., Nigel Priestley, M. J. N., Seible, F., California., and University of California, San
Diego. (September – October 1991). “Seismic Retrofit of Circular Bridge Columns for
Enhanced Flexural Performance”. ACI Structural Journal, V. 88, No. 5, 572-584.
Esmaeily, A., and Xiao, Y. (2002). “Seismic Behavior of Bridge Columns Subjected to Various
Loading Patterns”. California: Pacific Earthquake Engineering Research Center (PEER)
2002/15.
Esmaeily, A., and Xiao, Y. (2005). “Behavior of Reinforced Concrete Columns Under Variable
Axial Loads: Analysis”. ACI Structural Journal ,V. 102, No. 5 , 736-744.
6.15
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
Galal, K., Arafa, A., and Ghobarah, A. (2005). "Retrofit of RC Square Short Columns".
Engineering Structures. V. 27, no. 5, 801-813.
Harajli, M. H., and Rteil, A. A. (2004). "Effect of Confinement Using Fiber-Reinforced Polymer
or Fiber-Reinforced Concrete on Seismic Performance of Gravity Load-Designed
Columns". ACI Structural Journal. V. 101, No. 1, 47-56.
Haroun, M. A., and Elsanadedy, H. M. (2005). "Fiber-Reinforced Plastic Jackets for Ductility
Enhancement of Reinforced Concrete Bridge Columns with Poor Lap-Splice
Detailing". Journal of Bridge Engineering. V. 10, No. 6, 749-757.
Haroun, M. A., and Elsanadedy, H. M. (2005). "Behavior of Cyclically Loaded Squat Reinforced
Concrete Bridge Columns Upgraded with Advanced Composite-Material Jackets". Journal of
Bridge Engineering. V. 10, No. 6, 741-748.
Iacobucci, R. D., Sheikh, S. A., and Bayrak, O. (2003). "Retrofit of Square Concrete Columns
with Carbon Fiber-Reinforced Polymer for Seismic Resistance". ACI Structural Journal. V.
100, No. 6, 785-794.
Li, Y.-F., Hwang, J.-S., Chen, S.-H., and Hsieh, Y.-M. (2005). “A Study of Reinforced Concrete
Bridge Columns Retrofitted by Steel Jackets”. Journal of the Chinese Institute of Engineers,
V. 28, No. 2, 319-328.
Mander, J. B., Nigel Priestley, M. J., and Park, R. et al. (1988). “Theoretical Stress-Strain Model
for Confined Concrete”. Journal of Structural Engineering, V. 114, No. 8 , 1804-1826.
Memon, M. S., and Sheikh, S. A. (2005). "Seismic Resistance of Square Concrete Columns
Retrofitted with Glass Fiber-Reinforced Polymer". ACI Structural Journal, V. 102, no. 5,
774-783.
Nigel Priestley, M. J., Seible, F., Xiao, Y., and Verma R. (July–August 1994). “Steel Jacket
Retrofitting of Reinforced Concrete Bridge Columns for Enhanced Shear Strength Part 1:
Theoretical Considerations and Test Design”. ACI Structural Journal , V. 91, No.4 , 394-405.
Nigel Priestley, M. J., Seible, F., Xiao, Y., and Verma, R. (September–October 1994). “Steel
Jacket Retrofitting of Reinforced Concrete Bridge Columns for Enhanced Shear Strength Part
6.16
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
2: Test Results and Comparison with Theory”. ACI Structural Journal, V. 91, No. 5 , 537-
551.
Seible, F., Priestley, M. J. N., Hegemier, G. A., and Innamorato, D. (1997). "Seismic Retrofit of
RC Columns with Continuous Carbon Fiber Jackets". Journal of Composites for
Construction. V. 1, No. 2, 52-62.
Wu, Y.-F., Wang, L., and Liu, T. (2008). "Experimental Investigation on Seismic Retrofitting of
Square RC Columns by Carbon FRP Sheet Confinement Combined with Transverse Short
Glass FRP Bars in Bored Holes". Journal of Composites for Construction. V. 12, No. 1, 53-
60.
Xiao, Y., and Wu., H. (2003). "Retrofit of Reinforced Concrete Columns Using Partially
Stiffened Steel Jackets". Journal of Structural Engineering. V. 129, No. 6, 725-732.
Xiao, Y., and Ma, R. (1997). "Seismic Retrofit of RC Circular Columns Using Prefabricated
Composite Jacketing". Journal of Structural Engineering. V. 123, No. 10, 1357-1364.
Youm, K.-S., Lee, H.-E., and Choi, S. (2006). "Seismic Performance of Repaired RC Columns".
Magazine of Concrete Research. V. 58, No. 5, 267-276.
6.17
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
FC9 Aboutaha 38.4 0.5 1.4 2.7 0.9 4.9 3.5 4.3
FC11 Aboutaha 45.4 0.8 1.4 2.4 0.5 5.5 2.6 4.7
FC12 Aboutaha 43.1 0.6 1.4 2.7 0.9 5.5 3.4 4.9
FC17 Aboutaha 45.6 0.2 1.4 2.4 1.3 5.4 5.2 5.2
Mean 43.4 0.7 1.4 2.8 1.0 4.9 3.3 4.5
Steel Jacketed Rectangular Columns
6.18
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
Specimen Author Vy (kip) Δy/H (%) Vpeak/Vy Δp/H (%) Vmax/Vy Δmax/H (%) a b
C2R Priestly 115.5 0.3 1.4 4.4 † † 4.1
C4R Priestly 150.5 0.3 1.4 4.1 † † 3.8
Shear
C6R Priestly 161.0 0.4 1.4 5.5 † † 5.1
C8R Priestly 193.2 0.3 1.4 5.2 1.0 ‡ 4.9
Steel Jacketed Circular Columns
6.19
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
Specimen Author Vy (kip) Δy/H (%) Vpeak/Vy Δp/H (%) Vmax/Vy Δmax/H (%) a b
RF-R1 Haroun 59.7 1.3 1.3 2.4 0.9 2.8 1.2 1.5
RF-R2 Haroun 58.4 1.2 1.3 4.5 0.9 1.9 0.3 0.7
RF-R3 Haroun Lap Splice
54.4 0.9 1.4 3.6 1.2 3.6 1.8 2.7
RF-R4 Haroun 59.1 1.2 1.3 3.5 1.0 2.7 1.3 1.5
Mean 57.9 1.2 1.3 3.5 1.0 2.7 1.1 1.6
STD 2.1 0.1 0.1 0.7 0.1 0.6 0.6 0.7
RS-R1 Haroun 107.1 0.2 1.3 1.7 0.9 4.7 3.4 4.5
RS-R2 Haroun 104.7 0.3 1.3 1.1 0.9 4.7 2.7 4.4
RS-R3 Haroun 105.2 0.2 1.3 1.5 1.1 4.4 4.1
RS-R4 Haroun 95.9 0.1 1.4 2.1 1.2 4.6 4.5
FRP Jacketed Rectangular Columns
6.20
@Seismicisolation
@Seismicisolation
Non-linear Modeling Parameters for Jacketed Columns
Used in Seismic Rehabilitation of RC Buildings
Δp/H
Specimen Author Vy (kip) Δy/H (%) Vpeak/Vy (%) Vmax/Vy Δmax/H (%) a b
C2FP2 Harajli 15.1 0.7 1.4 3.0 0.8 5.1 4.0 4.3
C2F1 Harajli 15.6 0.8 1.4 3.1 0.9 5.0 3.7 4.3
C2F2 Harajli 15.3 0.8 1.4 3.0 0.9 5.1 3.4 4.3
FRP Jacketed Rectangular Columns
KFRP-05 Brena 9.7 1.4 1.4 6.3 0.9 9.5 7.5 8.1
CFRP-15 Brena 9.1 1.3 1.8 6.0 1.5 9.2 7.9
Lap Splice
Columns
6.21
@Seismicisolation
@Seismicisolation
J. C. Alvarez and S. F. Breña
Specimen Author Vy (kip) Δy/H (%) Vpeak/Vy Δp/H (%) Vmax/Vy Δmax/H (%) a b
CF-R6 Haroun 30.3 0.4 1.4 4.8 1.1 7.0 6.2 6.6
Lap Splice Lap Splice Rₒ Seible 42.7 0.1 1.4 2.4 1.1 3.8 7.1
FRP Jacketed Circular Columns
6.22
@Seismicisolation
@Seismicisolation
SP-297—7
Analysis of Seismic Response of Masonry‐Infilled RC Frames through Collapse
P. Benson Shing and Andreas Stavridis
Synopsis:
The assessment of the seismic vulnerability and collapse potential of masonry‐infilled RC frame buildings
presents a significant challenge because of the complicated failure mechanisms they could exhibit and
the number of factors that could affect their behavior. In general, there are two types of analysis
methods that can be used to simulate the inelastic behavior of infilled frames. One is to use simplified
frame models in which infill walls are represented by equivalent diagonal struts, and the other is to use
refined finite element models that can capture the failure behavior of RC frames and infill walls in a
detailed manner. However, both types of models have limitations in simulating structural response
through collapse. While refined finite element models are not computationally efficient, simplified
models are less accurate because of their inability to represent some failure mechanisms that could
occur in an infilled frame. In this paper, possible failure mechanisms and causes of collapse of masonry‐
infilled RC structures are discussed, and both simplified and refined finite element analysis methods that
can be used to simulate the inelastic response of these structures and assess their vulnerability to
collapse are presented with numerical examples. Additional research and development work needed to
improve collapse simulations is discussed.
Keywords: masonry; infill walls; reinforced concrete frames; earthquakes; seismic; collapse; finite
element method; equivalent strut models.
7.1
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
P. Benson Shing is a Professor of Structural Engineering at the University of California, San Diego. He
received his BS, MS, and PhD from the University of California, Berkeley. He is a member of ACI and his
main research interests are in seismic design and analysis, and in the inelastic behavior and performance
of concrete and masonry structures, including large‐scale testing and analytical modeling.
Andreas Stavridis is an Assistant Professor of Civil Engineering at the University at Buffalo. He received
his BS degree from the National Technical University of Athens, Greece, and MS and PhD from the
University of California, San Diego. His main research interests include the experimental and analytical
evaluation of the seismic performance of structures, and the condition assessment and rehabilitation of
existing concrete and masonry buildings.
INTRODUCTION
Masonry‐infilled RC frames can be frequently found in seismically active regions around the world. They
represent a common form of construction in many countries. Many older buildings in the western
coastal regions of the United States have unreinforced masonry infill walls, and infill walls are frequently
used in newer buildings in the mid‐western and eastern parts of the country. Masonry infill walls are
normally neither reinforced nor tied to the RC frames. In spite of this, infill walls often enhance the
seismic resistance of non‐ductile RC frames, which would otherwise not be able to withstand a major
earthquake. However, the partial or total collapse of an infilled RC frame could be sudden as
demonstrated in past earthquakes. The assessment of their seismic vulnerability and collapse potential
presents a significant challenge because of the complicated failure mechanisms they can exhibit and the
number of factors that can affect their behavior. In general, there are two types of analysis methods
that can be used to simulate the inelastic behavior of infilled frames. One is to use simplified frame
models in which infill walls are represented by equivalent compression‐only diagonal struts, and the
other is to use refined finite element models that can capture the failure behavior of RC frames and infill
walls in a detailed manner. However, both types of models have limitations in simulating structural
response through collapse. While refined finite element models are not computationally efficient,
simplified models are less accurate because of their inability to represent some main failure mechanisms
that could occur in an infilled frame. In this paper, possible failure mechanisms and causes of collapse of
masonry‐infilled RC structures are first discussed, and both simplified and refined finite element analysis
methods that can be used to simulate the inelastic response of these structures and assess their
vulnerability to collapse are presented with numerical examples. Additional research and development
work needed to improve collapse simulations is discussed.
FAILURE MODES AND CAUSES OF COLLAPSE
Failure modes, load resistance, and ductility
When built with high quality masonry materials and workmanship, infill walls can significantly enhance
the earthquake resistance of RC frames. However, under severe seismic loads, their interaction with the
bounding frames can result in undesired failure mechanisms, which may lead to the collapse of the
structure. Hence, to evaluate the ability of this type of structures to survive an earthquake without
collapsing, the frame‐wall interaction and the resulting failure mechanism have to be accurately
simulated in the analytical model. Under earthquake loading, the interaction of an RC frame with
masonry infill walls can result in one of several possible failure mechanisms depending on the strength
and stiffness of the walls as compared to those of the frame. Figure 1 shows three failure mechanisms
that have been frequently observed in laboratories and in the field (Mehrabi et al. 1994, 1996).
7.2
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
Figure 1 – Failure mechanisms of infilled frames Figure 2 – Failure Mechanism (b) exhibited in a
(Mehrabi et al. 1996). quasi‐static cyclic loading test (Blackard et al.
2009).
Failure Mechanism (a) shown in Figure 1 may occur when infill walls develop profuse horizontal sliding
shear cracks because of weak mortar joints. This mechanism can result in flexural hinges at the top and
bottom of the columns. However, because of the relatively small horizontal forces exerted by the wall
against the columns in this case, the columns can avoid severe damage as long as the story drift is not
excessive. Mechanism (b) can occur in non‐ductile RC frames infilled with relatively strong masonry,
which tends to develop dominant horizontal and/or diagonal cracks as shown by a laboratory specimen
in Figure 2. In this case, shear failures can occur in the columns as a result of a large horizontal force
exerted by the cracked wall against the top or bottom of a column. Sometimes, shear failure can also
occur at the mid‐height of a column when the masonry bed‐joint at that level develops a through crack
leading to a short‐column effect. This can be the case when a weak horizontal sliding plane has been
introduced by a window opening. Mechanism (c) is associated with corner crushing in a wall as a result
of the racking force from the frame. This can be expected for walls constructed of weak masonry units,
such as hollow clay tile, which is vulnerable to compressive failure. This mechanism will result in flexural
hinges in the columns and may also introduce a short‐column effect.
In addition to the aforementioned failure modes, older multi‐story RC frames with strong infill walls
could develop lap‐splice failures at the base of the columns. In such a case, the infill walls and the
boundary columns act together to function like a monolithic shear wall with severe tensile and
compressive forces developing at the base of the columns.
Hence, the lateral load resistance contributed by the frame and the wall in an infilled frame is governed
by the frame‐wall interaction and the resulting failure modes. As a consequence, the resistance of an
infilled frame is not a simple sum of that of a bare frame and that of a masonry wall considered
independently.
The ability of an infilled frame to resist earthquake loads depends not only on its strength but also on its
ductility. As shown in Figure 3, a frame exhibiting Failure Mechanism (a) discussed above tends to be
more ductile than a frame developing Mechanism (b) because the latter has the shear failure of the
columns. Furthermore, as shown in Figure 4, increasing the axial load on an infilled frame governed by
Failure Mechanism (b) can increase its lateral load resistance but reduces its ductility. This can be
explained by the fact that the masonry infill is the main load resisting element of an infilled frame, and
7.3
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
the lateral resistance of a masonry wall is largely governed by the shear resistance of the bed joints,
which largely depends on the compressive stress in the joints. Nevertheless, masonry units will be more
susceptible to crushing failure when subjected to higher axial and shear stresses. Such crushing can
occur in the interior of a wall (which is different from Mechanism (c)), and will lead to a rapid strength
degradation. Such behavior can also be observed in frames exhibiting Mechanism (a).
Mechanism (a) Mechanism (b)
Figure 3 – Non‐ductile RC frames of the same design exhibiting Failure Mechanism (a) and Failure
Mechanism (b) due to different infill walls (Mehrabi et al. 1996).
Figure 4 – Influence of vertical load in non‐ductile RC frames exhibiting Failure Mechanism (b)
(Mehrabi et al. 1996).
Hence, to assess the shear strength of a masonry wall, it is important to know the amount of axial load
carried by the wall. The proportion of the gravity load initially carried by an infill wall depends on the
ratio of the elastic axial stiffness of the wall to that of the RC columns and the construction sequence,
i.e., whether the walls were constructed before the upper stories were built. If the infill walls were
constructed after the entire RC frame had been built, then these walls would experience almost no
gravity load right after construction. However, part of the sustained gravity load may be gradually
shifted to the walls as creep and shrinkage occurs in the RC columns and the clay units expand in time
due to moisture absorption. Therefore, one can expect that infill walls will eventually carry a portion of
the gravity loads unless gaps are introduced between the walls and the beams above. The vertical load
carried by a wall can be measured by in‐situ tests using apparatus like flat‐jacks. However, in the
7.4
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
absence of such data, it has to be estimated analytically. An analytical method to calculate the gravity
load distribution in an infilled frame considering the creep of the RC columns and masonry wall can be
found in Koutromanos (2011).
As compared to a bare frame, the ductility of an infilled frame is normally much lower. For a non‐ductile
frame with a strong infill wall, diagonal/sliding shear cracks usually occur in the infill at a story‐drift ratio
of 0.2 to 0.3%. This will be quickly followed by the shear failure of the RC columns, after which the in‐
plane lateral resistance starts to decrease rapidly with increasing drift. Normally, severe damage will be
inflicted in an infilled frame when the story‐drift ratio exceeds 1%.
In summary, the performance of an infilled frame and its ability to survive an earthquake depend on
many factors, including the strength and quality of the masonry infill, the reinforcing details of the
frame, and the resistance of the frame‐wall system as compared to the seismic load demand. In
addition, the strength and ductility of an infilled frame depend on the shapes, sizes, and locations of
openings in the walls (Stavridis 2009), and the locations of infill walls in the frame.
Causes of collapse
Even though an infilled frame may exhibit an undesired failure mode, such as Failure Mechanism (b)
shown in Figure 1, and the falling debris of damaged masonry can be a major life‐safety concern, both
field observations from past earthquakes and laboratory studies (e.g., see Stavridis et al. 2012) have
shown that masonry infill walls can significantly enhance the lateral resistance of a non‐ductile RC frame
and protect it from major damage or collapse in the event of a severe earthquake. The collapse of
infilled RC frames in past earthquakes was often associated with a weak story mechanism, which could
be attributed to the lack of infill walls in the bottom story of a building for certain practical reasons or to
the severe damage and subsequent loss of infill during strong shaking. Studies have shown that intact,
well‐built, multi‐wythe, infill walls can develop significant resistance to out‐of‐plane loads because of
the arching mechanism that can develop with the bounding frame (Dawe and Seah 1989, Angel et al.
1994, Mander et al. 1993, Bashandy et al. 1995, and Flanagan and Bennett 1999). However, the
effectiveness of the arching mechanism and the stability of a wall depend on its height‐to‐thickness
(slenderness) ratio, and such mechanism can be jeopardized when the masonry infill has been damaged
by the frame‐panel interaction due to in‐plane shaking. Single‐wythe infill walls are especially vulnerable
to out‐of‐plane collapse but their contribution to the structural resistance is small to begin with. Figure
5(a) shows the collapse of a single‐wythe clay tile wall during an earthquake while the building remained
standing. Figure 5(b) shows the collapse of an infilled‐frame building in the same earthquake, which can
be attributed to the lack of sufficient walls in the first story. One can see from the picture that significant
drift occurred in the first story of the building leading to the separation of the RC frame from the
exterior wall, which together with the damage inflicted on the columns might have contributed to the
loss of the vertical load carrying capacity of the structure. Walls having window openings are more
susceptible to collapse, which was observed in a shaking‐table test as shown in Figure 6.
Computational models
To simulate the earthquake response of a masonry infilled RC frame through collapse, it is important
that the model can directly or indirectly account for the aforementioned failure mechanisms including
the loss of stability of infill walls due to in‐plane and out‐of‐plane loads. In particular, it is important that
the model be able to account for the following possible local and global behavior of a structure.
7.5
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
1. The inelastic behavior of RC members, including flexural hinging, diagonal shear failure, axial
load failure of RC columns, lap‐slip failures, and the failure of beam‐to‐column joints.
2. The inelastic behavior of masonry walls, including the cracking and shear sliding of mortar joints
and the crushing failure of masonry units.
3. The loss of stability and out‐of‐plane collapse of infill walls, considering the influence of the
arching mechanism, wall damage induced by in‐plane loads, and wall openings.
4. The frame‐wall interaction and the resulting failure mechanism.
5. The P‐ effect on the RC frame.
6. The three‐dimensional response of a building to multi‐axial ground motions.
Figure 5 – Damage of buildings with masonry infill walls in 2008 Wenchuan Earthquake in China.
It is also important that the computational models should be general enough to account for the
construction quality. However, a model that has all the aforementioned simulation capabilities can be
computationally prohibitive, and simplifications will be normally needed. In the following sections, two
modeling methods are discussed. One is the use of detailed finite element models using continuum
elements and the other is the use of simplified models with frame elements.
7.6
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
Figure 6 – Collapse of masonry infill with a window opening in a shaking‐table test (Stavridis et al. 2012).
NONLINEAR FINITE ELEMENT MODELS
A detailed nonlinear finite element model can capture the global as well as local failure behavior of an
infilled frame. However, these models require a significant computational effort. There are two general
finite element modeling approaches to simulate the fracture behavior of quasi‐brittle materials, such as
concrete and masonry. They are the smeared and discrete crack approaches. The smeared‐crack
approach has been often used to model diffused tensile cracks as well as the compressive failure of
concrete in reinforced concrete structures. However, this approach suffers from several inherent
limitations, including stress locking (Rots 1991, Lotfi and Shing 1991) that limits its ability to simulate the
brittle behavior of an RC member failing in diagonal shear and the shear sliding behavior of masonry
joints. This limitation can be overcome by representing cracks in a discrete manner using zero‐thickness
interface elements. Hence, combining the smeared and discrete approaches is necessary to model the
failure behavior of infilled frames in a realistic way (Stavridis and Shing 2010, Koutromanos et al. 2011).
An example of this is shown in Figure 7, in which zero‐thickness interface elements were used to model
flexural and diagonal shear cracks in RC columns and cracks in masonry joints, while smeared crack
elements were used to model diffused damage, including compressive failure in the concrete members
and brick units. In this model, each reinforcing bar in the columns was divided into multiple truss
elements (Stavridis and Shing 2010) so that each discrete crack was resisted by the right quantity of
reinforcement. This also improved the robustness of the numerical solution by having as many smeared‐
crack elements connected to a truss bar as possible.
Constitutive laws that can be used for the smeared‐crack and zero‐thickness interface elements are
described in Koutromanos (2011) and Koutromanos et al. (2011). A smeared‐crack model represents
cracks in a distributed fashion and can also simulate the compressive failure behavior of concrete and
masonry. The model used by Koutromanos et al. (2011) has an uncracked material represented by a J2‐
plasticity constitutive law. As shown in Figure 8(a), it has a von Mises failure surface with a tension cut‐
off. When the maximum principal stress reaches the tensile strength of the material, cracks initiate in a
direction normal to the direction of the maximum principal stress. For a cracked material, it adopts a
nonlinear orthotropic law with the axes of orthotropy normal and parallel to the crack, whose direction
is assumed fixed. At each point, two orthogonal cracks are allowed to occur. The uniaxial stress‐strain
relation for each direction of the orthotropic law is shown in Figure 8(b).
7.7
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
Figure 7 – Finite element modeling of a masonry infilled RC frame.
INITIAL YIELD SURFACE σ2
-f ’m -fo f ’t ft
Initial stiffness σ Exponential
unloading/reloading softening
σ1 ft
f ’ t ε2 ε1
LINEAR ε
ELASTIC -fo
Secant stiffness
unloading/reloading
ELASTOPLASTIC Exponential
softening
-f ’m f΄c
fc parabola
VON MISES
FAILURE SURFACE
Figure 8 – Smeared‐crack constitutive model.
The shear failure of RC columns and the sliding shear behavior of mortar joints are best captured by
zero‐thickness interface elements with a cohesive crack constitutive law. A cohesive crack elastic‐plastic
constitutive model has been proposed by Koutromanos and Shing (2012) to capture the mix‐mode
fracture as well as the cyclic crack opening‐closing and sliding shear behavior of concrete and masonry.
7.8
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
It is implemented in a 4‐node, zero‐thickness, isoparametric, line element. The model can simulate the
initiation and propagation of cracks under combined normal and shear stresses. It also accounts for
reversible shear dilatation induced by the roughness of the crack surface, which can have a significant
effect on the response of a confined crack. As shown in Figure 9(a), the yield surface for the model is
hyperbolic and its evolution is controlled with internal variables that are functions of the mode‐I and
mode‐II fracture energies as well as the frictional work. Figure 9(b) shows an example on the modeling
of the shear behavior of a mortar joint. As shown, the model accounts for the shear dilatation and joint
compaction due to damage.
Shear Displacement (in)
-0.4 -0.2 0.0 0.2 0.4
n 1.5 218
1.0 146
-10 -5 0 5 10
Shear Displacement (mm)
0.5 0.02
0.0 0.00
-0.5 -0.02
-1.0 -0.04
-1.5 -0.06
-2.0 experiment -0.08
-2.5 analysis -0.10
-3.0 -0.12
-10 -5 0 5 10
Shear Displacement (mm)
Figure 9 – Cohesive crack interface model.
It is well known that the interaction between the mortar joints and the masonry units due to the
different moduli of elasticity and Poisson effects of the two materials has a significant influence on the
strength and failure mechanism of a masonry assembly. However, when a zero‐thickness interface is
used to represent the behavior of a mortar joint, the 3‐D behavior of a mortar layer and its interaction
with the adjacent masonry units cannot be simulated. In this case, the influence of the brick‐mortar
interaction on the masonry strength has to be accounted for in an indirect way by assuming that the
compressive strength of the masonry units is equal to the compressive strength of the masonry prisms.
For brick masonry, prism strengths are normally lower than the brick strengths due to the tensile
splitting stress introduced by the softer mortar to the stiffer brick. In addition, the stiffness of the
interface elements representing mortar joints has also to be so determined that the stiffness of the
masonry assembly is properly represented. A procedure to calibrate these constitutive models is given
in Stavridis and Shing (2010), and Koutromanos et al. (2011).
7.9
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
This modeling approach has been proven to be able to accurately capture the failure mode and load‐
displacement response of an infilled frame up to a severe damage state, at which the lateral load
resistance decreases to as much as 50% of the peak resistance and the story‐drift ratio exceeds 1%.
Figure 10 shows the simulation of a quasi‐static test conducted by Blackard et al. (2009) on a masonry‐
infilled non‐ductile RC frame. A picture of the damaged specimen is shown in Figure 2. It can be seen
that the model was able to capture the crack pattern as well as the load‐displacement response of the
test specimen up to a state with severe load degradation.
800
400
force (kN)
0
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
-400 Experiment
Analysis
-800
drift ratio (%)
Figure 10 – Analysis of an infilled frame subjected to quasi‐static cyclic loading. (Koutromanos et al.
2011).
This modeling method has also been successfully applied to simulate the response of a three‐story,
masonry‐infilled, non‐ductile, RC frame tested on a shaking table by Stavridis et al. (2012). As shown in
Figure 11, the model is able to accurately capture the failure mechanism, response time histories, and
base shear‐vs.‐bottom story drift relations of the structure through the second last test on the shaking
table that had the Gilroy motion (from the 1989 Loma Prieta Earthquake) scaled to 120%. The specimen
collapsed out‐of‐plane in the subsequent test that had the El Centro record from the 1940 Imperial
Valley Earthquake scaled to 250%. This occurred after severe diagonal shear failures had developed in
the bottom‐story columns and the bottom‐story infill wall that had a window had partially collapsed (as
shown in Figure 6). However, neither the out‐of‐plane collapse of the frame nor the collapse of the wall
could be simulated by the model. This is because of the plane‐stress and small displacement
assumptions adopted in the model formulation.
7.10
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
0.4
0.3
0.2
1.2 2.0
1.5
0.8
1.0
drift ratio (%)
0.4 0.5
Vb/W
0.0 0.0
123.0 123.5 124.0 124.5 125.0 125.5 126.0 -1.2 -0.8 -0.4
-0.5 0.0 0.4 0.8 1.2
-0.4
-1.0
-0.8 Experiment Experiment
-1.5
Analysis Analysis
-1.2 -2.0
time (sec) drift ratio (%)
(c) Bottom‐story drift time history for Gilroy (d) Base shear vs. story drift for Gilroy motion
motion scaled to 120% scaled to 120%
Figure 11 – Analysis of a three‐story infilled frame tested on a shaking table (Koutromanos et al. 2011).
It should be noted that even though all the validation studies were based on laboratory specimens that
had good quality materials and workmanship, the material parameters, such as the mortar joint
properties, used in the models can be calibrated to account for the deterioration due to aging and for
the less‐than‐ideal construction quality that may be encountered in real structures. To model the failure
of lap splices in RC columns, a bond‐slip model can be used to connect reinforcing bars to smeared‐crack
elements. Such a model has been developed by Murcia‐Delso (2013) to simulate the bond‐deterioration
between the reinforcing steel and concrete, and can be used to simulate the failure of bar anchorages
and lap‐splices.
The constitutive models presented here can be extended to 3 dimensions to simulate the out‐of‐plane
collapse of an RC frame or that of a masonry infill wall. While this extension is conceptually
straightforward, it can significantly increase the computational effort. Furthermore, to simulate the
collapse of a masonry wall due to brick dislocation and the collapse of an RC column with dominant
shear cracks under an axial load, the cohesive crack interface model presented here needs to be
extended to account for the loss of contact due to shear sliding. This can be achieved with a contact law.
7.11
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
However, combining a cohesive crack model with a contact law will introduce additional computational
overheads.
SIMPLIFIED ANALYSIS METHODS
A simple and efficient approach to model infill walls in frame structures is to use the equivalent diagonal
strut concept as proposed by Holmes (1961), Stafford Smith (1967), and Mainstone and Weeks (1970).
This concept has been investigated and extended by others in numerous studies. However, the original
equivalent strut theory was based on the observed behavior of small‐scale steel frames infilled with
either brickwork or concrete. For masonry infilled RC frames, the equivalent strut representation
appears to be an over‐simplification of the actual behavior and fails to capture some main failure
mechanisms, such as Mechanisms (a) and (b) depicted in Figure 1. A study by Stavridis (2009) using
detailed nonlinear finite element models has demonstrated that the compressive stress field in a
masonry infill wall may not be accurately represented by a single diagonal strut. As shown in Figure 12,
the distribution and orientation of compressive forces developed in an infill wall can be quite different
from that introduced by a single diagonal strut. Hence, replacing a wall by a diagonal strut will not lead
to a realistic representation of the frame‐wall interaction. Furthermore, it will not simulate the possible
shear failure of a column that could be induced by the frame‐wall interaction. Models to simulate the
shear and axial load failures of columns using zero‐length springs have been proposed (e.g., Elwood
2004). The zero‐length spring concepts can be combined with a strut model in the way shown in Figure
13 to account for the axial and shear failures of a column as a result of the frame‐wall interaction.
However, three issues are associated with such a model. One is that it ignores the shear transfer that
develops between the beam and the wall, as shown in Figure 12, and it will, thereby, induce
unrealistically large shear demands on the columns causing their premature shear failure. Second, such
a model ignores the variation of the axial force along the height of each column due to the interface
stresses introduced by the infill. Finally, in such a model, the shear failure of a column may result in the
immediate collapse of the column under the axial load, which will lead to the collapse of the frame
without being able to account for the vertical load carrying capacity of the wall. Moreover, the
orientation of the compressive forces in a wall can change as damage evolves, while the angle of a
diagonal strut is fixed. Multiple‐strut models as proposed in some studies might partially overcome
these problems. Nevertheless, this will increase the complexity of a model and its benefits remain to be
demonstrated.
Vtot = 591 kN
Figure 12 – Strut model vs. forces on an infill from Figure 13 – Strut model with zero zero‐thickness
finite element analysis. springs.
7.12
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
Lateral Load
Qmax
Qy K2
1
Qres
K
1
y Q
max
Q Story Drift
res
(a) Load‐vs.‐displacement envelop for an infilled (b) Comparison of the simplified curve with the
fame experimental result for the structure shown in
Figure 10
Figure 14 ‐ Comparison of a simplified curve with experimental result.
In spite of the aforementioned issues, the use of strut models is probably the most efficient way for
Monte‐Carlo‐type simulations. For this purpose, one can treat diagonal struts as purely
phenomenological models. However, the struts should be calibrated in such a way that they not only
reflect the behavior of the infill walls but also the response of the frame members associated with
mechanisms that are not represented in the model, such as the shear failure of the RC columns. Two
steps are needed for such calibration. The first is to divide a multi‐bay, multi‐story, structure into
multiple single‐bay, single‐story frames, and the second is to derive the lateral load‐vs.‐displacement
curve for each of these frames. The second step can be accomplished by creating a detailed finite
element model for each bay of the structure. However, this can be impractical for structures with many
bays and stories. To circumvent this difficulty, Stavridis (2009) has used experimental data and finite
element analysis results to derive a set of simple rules to define ASCE 41‐type pushover curves for
infilled frames. His study focused on non‐ductile RC frames with strong infill walls consisting of multi‐
wythe solid clay brick units, which typically lead to Failure Mechanism (b) shown in Figure 1. The
idealized lateral‐vs.‐displacement curve adopted is shown in Figure 14 and it is calibrated with the
following steps.
1. The initial stiffness of an uncracked infilled frame is calculated with a shear beam model as proposed
by Fiorato et al. (1970).
1
K (1)
1 1
K fl K shl
in which K fl and K sh represent the flexural and shear stiffness of an uncracked cantilever wall. With this
approach, the structure is assumed to be a composite beam with the RC columns being the flanges and
the masonry wall being the web of the beam. Hence, for the flexural stiffness, K fl , the equivalent
properties of the composite beam should be used, but for the shear stiffness, only the contribution of
7.13
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
empirical equations have been proposed by Stavridis (2009) based on the parametric study.
1
Q 0.86 AR w for AR w 2.15 (4a)
max
3
Q max
0.15 for AR w 2.15 (4b)
7.14
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
corresponds to a brittle behavior with a high post‐peak negative stiffness.
The aforementioned procedure has been calibrated only for non‐ductile RC frames infilled with solid
masonry panels. All these frames exhibit Failure Mechanism (b) in Figure 1, with shear cracks developed
in the infill and the columns. If a different failure mechanism develops due to different infill and frame
properties, the pushover curve has to be recalibrated. It is also important to point out that this study has
considered in‐plane seismic loads only with the out‐of‐plane loads ignored.
The load‐vs.‐displacement relation of an infilled frame is also affected by the existence of window and
door openings in the infill walls. The openings reduce the strength and stiffness of an infilled frame.
Stavridis (2009) has examined the behavior of 30 infilled frames with various opening shapes, sizes, and
locations in the infill using nonlinear finite element models. In his study, the ratio of the opening area to
the gross surface area of the wall is between 7 and 19%. Based on this study, he has proposed the
following recommendations to derive a conservative backbone curve for infilled frames with openings.
1. The initial stiffness of a frame with an opening is to be estimated with the following formula.
K
1 RA (6)
K solid
in which K solid is the initial stiffness of the same frame with a solid infill calculated with Equation 1,
RA Aop AWtot , which is the ratio of the opening area, Aop , to the area of the solid infill wall, AWtot , and
α is a factor depending on the shape of the opening. Based on the results of the parametric study, it has
been suggested that 2 for infills with a window and 1.6 for infills with a door.
2. The yield force, Q y , corresponds to the load at which cracks initiate at the corners of the openings.
Based on the parametric study, the ratio, Qy / Qmax , is between 65 and 80%, similar to the case with
solid infill walls. Hence, this ratio can be assumed to be 2/3 for infilled frames with openings as well.
3. For RA between 7 and 19%, Qmax can be conservatively estimated as follows.
Qmax 0.8 Qmax
solid
(7)
Solid
in which Qmax is the peak strength of the same frame with a solid infill.
7.15
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
4. The drift at the peak load, Qmax , can be conservatively assumed to be equal to that for the frame with
a solid masonry panel, which can be estimated with Equation 4. Alternatively, the drift, Qmax , can be
determined with the post‐yield hardening stiffness, K 2 , as shown in Figure 14(a). This stiffness can be
estimated as a fraction of the initial stiffness.
K2 K (8)
with recommended to be 0.10 for infills with a window and 0.08 for infills with a door.
5. The residual strength Qres can be determined as a fraction of the peak strength.
Qres 0.5 Qmax (9)
6. The drift Qres , can be assumed to be equal to 1.40 Q , which is the same as that for frames with
max
solid infills.
Once the backbone curve for an infilled frame has been defined with the procedure described above or
with a finite element analysis, the curve can be used to determine the load‐displacement relation for an
equivalent diagonal strut. It should be emphasized that this curve accounts for the load degradation
exhibited by an infilled frame due to the shear and/or flexural failure of the RC columns as well as
damage in the infill. For a multi‐bay, multi‐story, structure, the calibration can follow these steps:
Step 1: Develop the backbone curve for the lateral load‐vs.‐drift relation for each bay and each story of
the structure using the aforementioned procedure or a finite element model.
Step 2: Construct a bare RC frame model for each bay and each story of the structure, and determine
the lateral load‐vs.‐displacement backbone curve for the frame considering the appropriate gravity load.
Step 3: Calibrate the strut model so that when it is added to the bare frame considered in Step 2, one
can obtain the behavior of the infilled frame estimated in Step 1.
Step 4: Develop a model for the entire structure consisting of the frame elements determined in Step 2
and the diagonal struts calibrated in Step 3.
The above method was applied to model the response of the three‐story infilled frame tested on a
shaking table by Stavridis et al. (2012), with the structural configuration shown in Figure 11. The frame
model with equivalent diagonal struts is shown in Figure 15(a). The analysis was carried out with the
software platform OpenSees. The axial behavior of the diagonal struts was modeled with a constitutive
law for concrete, which was calibrated with simplified pushover curves derived with the procedure
described above. The strut for each bay was individually calibrated. The load‐displacement curves for
the infilled frame models representing the two bottom‐story bays are shown in Figures 15(b) and 15(c).
One was for the solid infill while the other was for the infill with a window opening. In both cases, the
numerical model in OpenSees deviated slightly from the idealized (simplified) pushover curve because of
the specific constitutive law used. In Figure 15(d), the base shear‐vs.‐bottom story drift hysteresis curves
7.16
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
obtained from the dynamic analysis of the three‐story frame with the strut‐based model, for a sequence
of ground motion records up to the 120% Gilroy motion, are compared to the test results. The model
shows more rapid load degradation than the test results due to the conservatism introduced in the
simplified pushover curves. The quality of the simulation could probably be improved if a detailed finite
element model was used to calibrate the model in place of the simple rules.
This type of phenomenological model can be extended to simulate the response of an infilled frame to
out‐of‐plane loads by using two bi‐axial fiber‐section beam‐column elements to represent a diagonal
strut. A fiber‐section beam‐column model can simulate both the arching mechanism and the out‐of‐
plane failure of an infill wall. This approach has been considered by Kadysiewski and Mosalam (2008),
who have proposed an interaction curve to estimate the capacity of an infill wall subjected to
simultaneous in‐plane and out‐of‐plane loads. The proposed interaction curve is calibrated with finite
element models. This approach deserves to be further explored and evaluated with experimental data
(e.g., the data of Dawe and Seah 1989, Angel et al. 1994, Mander et al. 1993, Bashandy et al. 1995, and
Flanagan and Bennett 1999).
700
400
80
300
200
40
100
0 0
0 0.2 0.4 0.6 0.8 1 1.2
Drift ratio, %
(a) Diagonal strut model (b) Pushover analysis of a bay with solid infill
1800 400
700
1350 Shake-Table Tests 300
600 OpenSEES model
Simplified curve 900 Strut model 200
120
Base shear, kips
Base shear, kN
450 100
Lateral force, kN
400 0 0
80
300 -450 -100
-900 -200
200
40
-1350 -300
100
-1800 -400
0 0 -1.5 -1 -0.5 0 0.5 1
0 0.2 0.4 0.6 0.8 1 1.2 st
1 Story drift, %
Drift ratio, %
(c) Pushover analysis of an infilled bay
(d) Hysteretic curves from dynamic analysis
with a window
to up 120% Gilroy
Figure 15 – Frame analysis using equivalent diagonal struts.
7.17
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
COLLAPSE CRITERION
The detailed finite element modeling and simplified analysis methods presented above can be used to
simulate the response of an infilled frame up to a severe damage stage close to collapse. Nevertheless,
neither method can simulate the actual collapse of an infilled frame in a satisfactory manner without
further enhancements. For Monte‐Carlo‐type simulations, the use of detailed finite element models is
not practical and simplified strut‐based models should be used.
Regardless of the type of model being used, one critical issue in a collapse simulation of any structure is
the definition and identification of collapse. While collapse can be defined as the loss of the vertical load
carrying capacity of a structure, the distinction between the partial and total collapse needs to be made.
The simulation of either the partial or the total collapse of an infilled frame is a big challenge because of
the interaction and possible changes of the load transfer mechanisms among the columns, beams, and
walls. Once the RC columns in an infilled frame have suffered severe damage such as shear failure, the
gravity load can be shifted to the infill walls as long as these walls remain standing. A severely cracked
masonry wall could still carry significant gravity load even though it might have lost a large portion of its
in‐plane and out‐of‐plane lateral load resisting capacity. This can be perceived as an unstable
equilibrium state, which has been often observed in in‐plane quasi‐static tests (see Figure 4) as well as
nonlinear finite element analyses (Koutromanos et al. 2011). Hence, the collapse of an infilled frame has
to be preceded by the collapse of the infill walls. The collapse of infill walls could occur before or after
the RC frame has suffered severe damage. The walls could collapse prior to the development of severe
damage in the frame if the masonry is weak or large openings exist in the walls, and this will
immediately result in a weak‐story mechanism. For this situation, simple models like those proposed by
Elwood (2004) can be used to simulate the subsequent collapse of the RC columns.
In view of the aforementioned modeling challenges and the large uncertainties in the collapse
prediction for an infilled frame, it is prudent to treat the first attainment of an unstable equilibrium state
of the structure as the total collapse condition. This can be considered as a state in which the infilled
frame has lost a significant portion of its lateral in‐plane load carrying capacity due to the failure of the
walls at one or more stories. Until more experimental data are available, this state can be assumed to be
reached when the post‐peak in‐plane resistance of the infilled frame drops to 60% of the peak. The
story‐drift level at which this load degradation is reached can be taken as the collapse criterion for
Monte‐Carlo‐type simulations. This critical drift limit can be identified with a pushover analysis of a
planar structure using a detailed finite element model or a simplified model. Experimental results from
different studies have shown that under in‐plane loads only, this can occur at a story‐drift ratio between
1.0 and 1.5%. Quasi‐static tests and finite element analyses have shown that frames with weak infill
walls or walls with a large opening tend to have lower strengths but higher ductile. Hence, under static
loads, these frames may not reach 40% strength loss till a relatively large drift level has been reached.
Nevertheless, walls in these frames will be less stable under dynamic loads as mentioned in a previous
section. Therefore, to be prudent, the critical drift limit should not be greater than 1.5%. If a wall has
been damaged by the out‐of‐plane load, it can be expected that the in‐plane resistance will drop more
rapidly and a strength loss of 40% can occur at a smaller story‐drift level. Further studies are needed to
establish a drift limit for such cases.
CONCLUSIONS
Simplified models using equivalent compression‐only diagonal struts are most efficient for Monte‐Carlo‐
type collapse simulations for masonry‐infilled RC frames. However, because of the inability of such
models to represent some main failure mechanisms of an infilled frame, the strut model should be so
calibrated that the load degradation induced by the unaccounted mechanisms, such as the shear failure
7.18
@Seismicisolation
@Seismicisolation
Analysis of Seismic Response of Masonry-Infilled RC Frames through Collapse
of the RC columns, can be represented in the load‐displacement response of the entire frame. To this
end, the axial behavior of a strut can be modeled with a phenomenological law that reflects the damage
evolution of an entire infilled frame. Such a strut model can be calibrated with a detailed nonlinear finite
element analysis or the simple procedure described in the paper. The simple procedure presented here
was derived for structures that have non‐ductile RC frames and strong infill walls. Further work is
needed to extend the procedure to structures with different frame and wall properties.
In a simplified model, RC beams and columns can be modeled with beam‐column elements with flexural
hinging capabilities. To simulate the damage and possible collapse of an infill wall caused by out‐of‐
plane loads, a diagonal strut can be modeled with two bi‐axial fiber‐section beam‐column elements,
which account for the out‐of‐plane bending as well as the arching mechanism developed in an infill wall.
Nevertheless, further work is needed to calibrate these models with available experimental data.
For Monte‐Carlo‐type collapse simulations, a convenient criterion for collapse must be determined.
Until more data are available, it can be assumed that an infilled frame will be on the verge of collapse
when its in‐plane resistance drops to 60% of the peak resistance. This is based on test data currently
available. The story‐drift level at which this load degradation will be reached under monotonically
increasing loads can be taken as a collapse criterion. This drift limit can be identified with a pushover
analysis of a planar structure using a detailed finite element model or a simplified model, but the critical
drift ratio for collapse should not be greater than 1.5%.
For future numerical studies and the calibration of simplified models, finite element models that can
simulate collapse in a detailed fashion are desired. To simulate out‐of‐plane failures, 3‐D finite element
models are needed. Other simplified collapse simulation strategies such as element removal methods
should also be investigated. Different modeling methods should be compared and evaluated with
benchmark examples.
ACKNOWLEDGMENTS
The modeling work presented in this paper is based on a prior research project supported by the
National Science Foundation Grant No. 0530709 awarded under the George E. Brown, Jr. Network for
Earthquake Engineering Simulation Research (NEESR) program. However, opinions expressed in this
paper are those of the authors and do not necessarily represent those of the sponsor.
REFERENCES
Angel, R, Abrams, D, Shapiro, D, Uzarski, J, and Webster, M, 1994, “Behavior of Reinforced Concrete
Frames with Masonry Infills,” Report No. UILU‐ENG‐94‐2005, Dept. of Civil Engineering, University of
Illinois, Urbana‐Champaign, IL.
Bashandy, T, Rubiano, NR, and Klingner, RE, 1995, “Evaluation and Analytical Verification of Infilled
Frame Test Data,” Report No. 95‐1, Dept. of Civil Engineering, University of Texas, Austin, TX.
Blackard, B, Willam, K, and Mettupalayam, S, 2009, “Experimental Observations of Masonry Infilled RC
Frames with Openings,” ACI SP 265‐9, American Concrete Institute, pp. 199‐222.
Dawe, JL and Seah, CK, 1989, “Out‐of‐Plane Resistance of Concrete Masonry Infilled Panels,” Journal of
the Canadian Society for Civil Engineering, V. 16, pp. 854‐864.
Elwood, KJ, 2004, “Modelling Failures in Existing Reinforced Concrete Columns,” Canadian Journal of
Civil Engineering, V. 31, pp. 846‐859.
Fiorato, AE, Sozen, MA, and Gamble, WL, 1970, “An Investigation of the Interaction of Reinforced
Concrete Frames with Masonry Filler Walls,” Report No. UILU‐ENG 70‐100, Dept. of Civ. Eng., University
of Illinois at Urbana‐Champaign, Urbana, IL.
7.19
@Seismicisolation
@Seismicisolation
P. B. Shing and A. Stavridis
Flanagan, RD and Bennett, RM, 1999, “Bidirectional Behavior of Structural Clay Tile Infilled Frames,”
Journal of Structural Engineering, V. 125, No. 3, pp. 236‐244.
Holmes, M, 1961, “Steel Frames with Brickwork and Concrete Filling,” Proceedings of the Institute of
Civil Engineers, V. 19, No. 6501, pp. 473‐478.
Kadysiewski, S and Mosalam, KM, 2009, “Modeling of Unreinforced Masonry Infill Walls Considering
In‐Plane and Out‐of‐Plane Interaction,” PEER Report 2008/102, Pacific Earthquake Engineering Research
Center, University of California at Berkeley, Berkeley, CA.
Koutromanos, I, 2011, “Numerical Analysis of Masonry‐Infilled Reinforced Concrete Frames Subjected
to Seismic Loads and Experimental Evaluation of Retrofit Techniques,” Ph.D. Dissertation, University of
California at San Diego, La Jolla, CA.
Koutromanos, I, Stavridis, A, Shing, PB, and Willam, K, 2011, “Numerical Modeling of Masonry‐Infilled
RC Frames subjected to Seismic Loads,” Computers and Structures, V. 89, pp. 1026‐1037.
Koutromanos, I and Shing, PB, 2012, “A Cohesive Crack Model to Simulate Cyclic Response of Concrete
and Masonry Structures,” ACI Structural Journal, V. 109, No. 3, pp. 349‐358.
Lotfi, HR and Shing, PB, 1991,” An Appraisal of Smeared Crack Models for Masonry Shear Wall
Analysis,” Computers & Structures, V. 41, No. 3, pp. 413–425.
Mainstone, RJ and Weeks, GA, 1970, “The Influence of Bounding Frame on the Racking Stiffness and
Strength of Brick Walls,” Proceedings of the 2nd International Conference on Brick Masonry, Stoke‐on‐
Trent, pp. 165‐171.
Mander, JB, Nair, B, Wojtkowski, K and Ma, J, 1993, “An Experimental Study on the Seismic
Performance of Brick‐Infilled Steel Frames with and without Retrofit,” Report No. NCEER‐93‐0001,
National Center for Earthquake Engineering Research, State University of New York, Buffalo, NY.
Mehrabi, AB, Shing, PB, Schuller, MP, and Noland, JL, 1994, “Performance of Masonry‐Infilled R/C
Frames under In‐Plane Lateral Loads,” Report No. CU/SR‐94‐6, Dept. of Civil, Environmental, and
Architectural Engineering, University of Colorado, Boulder, CO.
Mehrabi, AB, Shing, PB, Schuller, MP, and Noland, JL, 1996, “Experimental Evaluation of Masonry‐
Infilled RC Frames,” Journal of Structural Engineering, V. 122, No. 3, pp. 228‐237.
Murcia‐Delso, J, 2013, “Bond‐slip Behavior and Development of Bridge Column Longitudinal
Reinforcing Bars in Enlarged Pile Shafts,” PhD Dissertation, University of California, San Diego, La Jolla,
CA.
Rots, JG, 1991, “Numerical Simulation of Cracking in Structural Masonry,” HERON, Netherlands School
for Advanced Studies in Construction, The Netherlands, V. 36, No. 2, pp. 49‐63.
Stafford Smith, B, 1967, “Methods for Predicting the Lateral Stiffness and Strength of Multi‐Story
Infilled Frames,” Building Science, V. 2, pp. 247‐257.
Stavridis, A, 2009, “Analytical and Experimental Study of Seismic Performance of Reinforced Concrete
Frames Infilled with Masonry Walls,” Ph.D. Dissertation, University of California at San Diego, La Jolla,
CA.
Stavridis, A and Shing, PB, 2010, “Finite Element Modeling of Nonlinear Behavior of Masonry‐Infilled
RC Frames,” Journal of Structural Engineering, V. 136, No. 3, pp. 285‐296.
Stavridis, A, Koutromanos, I, and Shing, PB, 2012, “Shake‐Table Tests of a Three‐Story Reinforced
Concrete Frame with Masonry Infill Walls,” Journal of Earthquake Engineering and Structural Dynamics,
V. 41, pp. 1089‐1108.
7.20
@Seismicisolation
@Seismicisolation
SP-297—8
Synopsis:
Probability of collapse is currently used to set targets for system performance and response measures of new
buildings. This study compares the probability of collapse for new, retrofitted and existing concrete buildings.
Retrofitting plays an important role in reducing seismic risk from older concrete buildings. In order to decide on the
most appropriate and economical retrofit strategy for an existing structure, it is necessary to assess the risk of
collapse of each rehabilitation measure. At present, it is frequently assumed that retrofitting a non-ductile concrete
building will enhance the seismic performance such that it can reach the same performance level as a ductile
building designed based on current seismic codes. However, based on the evaluation of the concrete frames
presented in this paper, typical retrofit schemes (such as: adding an additional lateral force restraint system;
increasing ductility of existing concrete columns; and weakening the existing beams) cannot achieve the same
performance as modern code-conforming structures. The study finds that retrofitting schemes where the columns or
beams are modified such that the frame satisfies the collapse prevention level of ASCE 41-13 have the least
beneficial effect regarding seismic collapse safety; and conversely, adding a shear wall will significantly improve
the seismic performance in terms of the probability of collapse.
Keywords:
Collapse assessment; Nonlinear analysis; Probability of collapse; Retrofitting; Reinforced concrete moment frames;
Seismic safety
8.1
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
ACI member Majid Baradaran Shoraka is a post-doctorate researcher at University of British Columbia,
Vancouver, BC, Canada. He received his PhD from the University of British Columbia, Vancouver, BC, Canada.
His research interests include developing performance-based evaluation methodology and code design procedures
for new and existing structures. He is an associate member of the ACI Committee 369.
ACI Fellow, Kenneth J. Elwood is a Professor of Civil Engineering at the University of British Columbia, Canada.
He received his PhD from the University of California, Berkeley. His research interests include the behavior and
performance-based design of reinforced concrete structures under seismic loading. He is Chair of ACI committee
133, Disaster Reconnaissance, past Chair of ACI committee 369, Seismic Repair and Rehabilitation, and member of
Subcommittee H, Seismic Design, of ACI Committee 318.
Tony Yang is an assistant Professor at the University of British Columbia, Vancouver, BC, Canada. He received his
PhD from the University of California, Berkeley, CA. His research focus is on improving the structural performance
through advanced analytical simulation and experimental testing. He is a corresponding member of AISC TC9
which responsible for publishing the seismic design guidelines for steel structures in United States, member of the
Tall Building Initiative project and one of the core developers for the next-generation performance-based earthquake
engineering assessment methodology currently adopted by the ATC-58/FEMA-P58 research project.
Abbie B. Liel is an assistant professor at the University of Colorado Boulder. She received her PhD from Stanford
University. Her research focuses on performance-based assessments of structural safety and collapse risk under
extreme loads, including seismic assessment of non-ductile concrete buildings and alternative retrofit strategies.
INTRODUCTION
With recent earthquakes worldwide, retrofitting plays an important role in mitigating the seismic risk for older RC
structures. Repair and upgrading of existing structures is becoming increasingly important and has significant
economic impacts. Seismic retrofitting of an existing building will either involve adding new components to the
existing structure, increasing the deformation capacity of existing components, or changing the collapse mode of the
structure. In order to decide on the most appropriate and economical retrofitting strategy, it is necessary to assess the
lateral load resistance, deformation capacity, and potential modes of collapse.
Current retrofit standards, including the latest version of ASCE 41 (ASCE, 2013), assist engineers with the seismic
assessment of existing buildings based on component acceptance criteria. In this standard, component demands are
compared with component acceptance criteria for different performance levels, namely immediate occupancy (IO),
life safety (LS), and collapse prevention (CP). The component with the worst performance level will define the state
of the entire structure. The performance of non-ductile reinforced concrete moment frames retrofitted to the LS and
CP performance levels are studied in this paper. Specifically, the collapse vulnerability of different retrofitting
schemes is studied using the system-level assessment procedure introduced in Baradaran Shoraka (2013) and
Baradaran Shoraka et al. (2013).
The collapse vulnerability of existing non-ductile and retrofitted archetype RC structures has previously been
studied by Liel and Deierlein (2013). In that study, 1960s era California type 4- and 8-story non-ductile RC space
and perimeter moment frame structures were chosen. Three retrofit techniques were considered: (1) jacketing of the
RC columns with reinforced concrete, (2) carbon fiber-wrapping of RC columns, and (3) construction of “super
column shear walls” around existing columns. For each type of retrofit, both ‘modest’ and ‘significant’ retrofits were
considered; however, it is not known how these retrofits relate to the criteria specified in ASCE 41-13 as they were
not designed to a specific standard. The collapse performance of these structures was compared to the collapse
performance of the un-retrofitted (original) structure. However, only side-sway collapse mechanisms were
simulated. Non-ductile collapse modes, such as column shear failure and subsequent axial failures, were considered
as non-simulated collapse modes. The study assumes that when the first occurrence of this non-simulated failure
mode occurs, it will lead to collapse of the structure, which produces a conservative assessment of collapse capacity,
ignores the system’s ability to redistribute the load after failure. Figure 1 shows the computed collapse fragilities for
the 4-story space frame structures presented in Liel’s study: including existing (un-retrofitted), retrofitted (using the
three retrofitting techniques) and modern (with 20 ft. and 30 ft. bay widths) buildings. As shown in Figure 1, the
8.2
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
performance of the retrofitted structures is superior to that of the unretrofitted frames, but does not achieve the low
collapse risk of the modern frames. In addition, Figure 1 shows large variability among the collapse risk of the
retrofitted frames, depending on the unique characteristics of the retrofit. Liel and Deierlein (2013) also conducted a
cost-benefit analysis of the retrofit schemes, the review of which is outside the scope of this paper.
The study presented herein builds on the Liel and Deierlein (2013) findings by explicitly modeling the gravity-load
collapse of the non-ductile frame and by directly considering the nonlinear acceptance criteria for both LS and CP in
ASCE 41-13 in selecting and designing the retrofit for the frames. The first advancement is critical, because it
ensures that the comparison of collapse performance between the unretrofitted and retrofitted frames is more
accurate. The use of ASCE 41-13 ensures that the retrofits considered in this study are consistent with those
designed in practice; ASCE 41-13 is widely used in practice for design and evaluation of seismic retrofit schemes.
The paper will first introduce the case study structure, apply different retrofitting strategies to upgrade the case study
building, and finally compare the seismic performance of each building by means of collapse fragilities and collapse
performance metrics. All the results presented in this paper are original for this manuscript.
Figure 1–Collapse fragility functions for 4-story space frames Liel and Deierlein (2013)
To compare the collapse performance of existing and retrofitted (using the ASCE 41-13 standard) RC moment
frames, a non-ductile perimeter concrete moment resisting frame building designed according to the 1967 UBC
(ICBO 1967) code was selected as the prototype structure. The 8-story perimeter frame is chosen from the Liel
database (Liel, 2008; Liel et al., 2011) and is summarized in Table 1. The prototype building was retrofitted using
three retrofitting techniques, each designed to satisfy two performance objectives (LS and CP) as specified in ASCE
41-13. In addition, the performance of a ductile perimeter moment frame building designed according to the
International Building Code (ICC 2003) was also included in this study.
The prototype building is assumed to be located in Los Angeles, California, site class D, with SS = 1.5g and S1 =
0.6g, where SS is the earthquake spectral response acceleration at short periods and S1 is the earthquake spectral
response acceleration at 1-second periods. The earthquake hazard level defined for the selected performance
objectives has a 2% probability of exceedance in 50 years.
8.3
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
Table 1 provides key characteristics of the prototype building. This structure was designed according to 1967 UBC
for Zone 3 (the highest seismic zoning criteria). Seismic design requirements such as the maximum and minimum
steel reinforcement ratios, maximum stirrup spacing, and requirements on hooks, bar spacing and anchorage were
included. It should be noted that there is no transverse reinforcement in the joints as is typical of buildings of this
era. A concrete strength of 4 ksi (28 MPa) was assumed for all members .
Numerical model– In numerical simulation, there is always the challenge of defining the collapse state during the
analysis and differentiating structural collapse from numerical non-convergence. In collapse analysis used in this
paper, the criterion is based on two types of global failure, gravity-load collapse and side-sway collapse. Side-sway
collapse develops due to P-delta instability at large lateral deformations, with an implicit assumption that the
structure is ductile enough to reach this state. However, because of the limited ductility of older reinforced concrete
structures, an alternate criteria is also defined, entitled here as gravity-load collapse. Gravity-load collapse may be
precipitated by axial-load failure of columns, punching shear failure of slab–column connections, failure of slab-
diaphragm connections, or axial-load failure of beam-column joints. Gravity load collapse is defined in this paper by
the point at which the axial load capacity at a single story drops below the gravity load demand (explained in detail
in Baradaran Shoraka (2013) and Baradaran Shoraka et al. (2013)). Both side-sway and gravity-load collapse have
been considered in the current study.
The numerical analyses were conducted using a two-dimensional frame modeled in OpenSees (PEER, 2009). Only
the perimeter moment resisting frames were included in the numerical model. The analytical model incorporated all
important features required to model collapse of the structure.
The beam and column elements were modeled using the lumped plastic hinge model developed by Ibarra et
al. (2005) to account for the strength and stiffness degradation under cyclic loads. Strain-softening
behaviour associated with concrete crushing, rebar buckling and fracture, and/or bond failure was
accounted for; including such degradation in the model has an important effect on the collapse response of
these structures. The backbone and cyclic response parameters used in the numerical model were developed
by Haselton et al. (2008).
The model includes large geometry transformations that take into account P-delta effects resulting from the
mass of the entire structure.
To ensure the numerical model is capable of modeling the joint failure, a two-dimensional joint model was
used. The numerical model developed by Lowes and Altoontash (2003) was used to define the shear
deformations of the joint and bond-slip behaviour.
Shear and axial failures in columns were modeled using the limit-state material developed by Elwood
(2004). These models define the shear and axial failures of concrete columns as a function of the
deformation capacity, considering the geometric, material and design parameters.
The first fundamental period of the structure is about 2.4 (s). The ASCE 41 (2013) estimate of the building period is
1.7 (s). The building period is high because this frame was designed to achieve a minimum column size and high
vertical reinforcement percentage.
Table 1 also provides key characteristics of the ductile building (Haselton et al., 2011). For comparison, this ductile
building was chosen to be similar to the non-ductile moment frame. This structure was designed according to the
International Building Code (ICC, 2003), ASCE 7 (ASCE, 2002), and ACI 318 (ACI, 2005), and meets all the
governing code requirements and detailing for special moment frames.
8.4
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
Table 1–Design documentation for 8-story non-ductile (Liel et al., 2011) and ductile (Haselton et al., 2011)
perimeter frame structure (1 in = 0.0254 m)
Column Beam
Column Beam
Design Column Hoop Beam Hoop Floor Bay
Size, Size,
Structure Base Shear Reinforcement Spacing, Reinforcement Spacing, Height Width,
hxb hxb
Coefficient Ratio, tot s Ratio, ’) s (ft) (ft)
(in x in) (in x in)
(in) (in)
8 story-
perimeter
frame 0.054g 30 x 36 0.033 15 26 x 36 0.008 (0.010) 17 15 25
(non-
ductile)
8 story-
perimeter
0.050g 26 x 34 0.018 3.5 26 x 30 0.007 (0.008) 5 15 25
frame
(ductile)
Retrofitted buildings
The non-ductile perimeter concrete moment resisting frame presented in the previous section was evaluated using
the procedures outlined in ASCE 41-13. A pushover analysis, using the first mode load distribution and the target
roof displacement calculated using Eqn. 1, was performed.
(1)
where δ is defined as the roof drift. S represents the elastic spectral displacement of an equivalent single degree
of freedom (SDOF) oscillatory at the earthquake hazard level of interest; chosen Maximum Considered Earthquake,
MCE (equivalent to a probability of occurrence of 2% in 50 years) shaking intensity for this study. Te is the
fundamental period of the existing structure. C0 is characterized as a dimensionless coefficient that relates the
spectral displacement of an equivalent SDOF to the roof displacement. C1 is a dimensionless coefficient that relates
expected maximum inelastic displacements to the displacements calculated using the linear elastic response and
finally C2 is a dimensionless coefficient that adjusts the roof drift ratio to account for the pinched hysteretic shape,
stiffness degradation and strength deterioration effect on the structure.
At such roof drift, the component deformation demands were checked using the acceptance criteria defined in ASCE
41-13. The prototype building was retrofitted to the Life Safety (LS) and Collapse Prevention (CP) performance
levels. For the Collapse Prevention performance level, the structural components can be severely damaged, but the
structure must be able to continue carrying gravity loads without collapse. Because the strength and stiffness
degradation was modeled in all the numerical components, the “Secondary Components” acceptance criteria, as
specified in ASCE 41-13, were used to assess the performance for all components.
There are various retrofitting schemes available. Using the recommendations presented in FEMA 547 (FEMA,
2006), three retrofitting schemes were identified and included in this study:
1) Strengthening the existing columns, beams and joints by steel jacketing, or adding new reinforced concrete,
steel, or fiber-reinforced polymer wrap overlays. This method enhances the strength and deformation
capacity of the existing non-ductile components.
8.5
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
2) Weakening a certain portion of the structure by removing or modifying a portion of the existing structure.
This technique, usually applied to weaken the beams, promotes formation of a strong-column weak-beam
mechanism.
3) Adding supplementary lateral force-resisting systems, such as adding ductile reinforced concrete shear walls,
to reduce deformation demands on the existing elements.
Retrofitted building – columns modified (strengthening technique) –The first approach to retrofit non-ductile
RC frames increases the strength and deformation capacity of the critical non-ductile columns. The critical non-
ductile columns were identified based on the acceptance criteria provided in ASCE 41-13. The critical columns were
jacketed with reinforced concrete in order to increase the column deformation capacity with increased transverse
reinforcement. It is assumed that the retrofitted columns will no longer sustain shear/axial failures due to the
additional transverse reinforcement provided in the jacket. The final retrofit design for the selected performance
objective is schematically illustrated in Figure 2a. The highlighted section in this figure shows the retrofitted
columns for each of the performance objectives. The highlighted numbers in the table also show the changes in the
column dimensions and transverse reinforcement for the structures upgraded to satisfy the CP and LS performance
levels.
Retrofitted building – beams modified (weakening technique)–The second approach is to retrofit the non-ductile
RC frames by weakening the existing beams such that the system can form a strong-column, weak-beam
mechanism. The concept of weakening for retrofit of RC frames has been previously proposed in retrofit guidelines
and by several researchers (FEMA-356, 2000; Pampanin, 2005b; Viti et al., 2006, Kam and Pampanin, 2009). This
technique was applied by cutting longitudinal reinforcement in the beams. The longitudinal reinforcement were cut
floor by floor and the demand to the system was recalculated after each iteration, until the component demands for
the critical columns fall within the selected performance level. This retrofitting measure will result in a decrease in
strength and stiffness for the beams. For this technique, it was not possible to weaken the beams to the extent
necessary to reach the LS level. The final design (for the CP level) is schematically illustrated in Figure 2b. The
highlighted section in this figure shows the weakened beams. The highlighted numbers in the table also show the
changes in the beam dimensions and longitudinal reinforcement for the structure upgraded to the CP performance
objectives.
Retrofitted building – wall added (adding technique)–The third approach to retrofit the non-ductile RC frame is
by adding shear walls to decrease the drift demand on the existing moment frame. The shear wall was designed with
the same cross sectional area and reinforcement ratio over the full height of the wall. The strength of the shear wall
was increased by means of expanding the depth and longitudinal steel reinforcement until the component demands
for the critical columns fall into the performance target. Note that in order to assess the performance of a retrofit
satisfying the criteria in ASCE 41-13, only enough wall stiffness was added to get rotations in critical elements
below selected performance target limits, hence final wall dimensions may not resemble a real wall. The final
retrofit design for the selected performance objective is illustrated in Figure 2c. The highlighted section in this figure
shows the wall added to this building. The numbers in the table also show the wall designs for the structures
upgraded to the CP and LS performance levels. The shear wall is modeled in OpenSees using the Nonlinear Beam-
Column element (de Souza, 2000). The shear wall was connected to the frame using elastic truss elements at each
floor (see Figure 2c).
8.6
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
Beams (CP)
8.7
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
Wall (LS)
tw
bf (in) tf (in) L (in) Lw (in) f w
(in)
15 10 75 55 8 0.020 0.0025
fw = ratio of longitudinal reinforcement in the flange and web; bf & tf = width and thickness of the flange;
1 in = 0.0254 m
(c) Retrofitted Building – Walls (adding technique; the shear wall and the truss elements are added)
Figure 2–Design documentation for 8-story retrofitted perimeter frame structure (members highlighted in
this picture are modified for the different retrofitting measures)
The performance assessments for the three types of retrofitted buildings conducted using the ASCE 41-13 document
is presented in the following sections.
Pushover results
ASCE 41-13 Nonlinear Static Procedure uses pushover analysis to assess the deformation demand in the
components. The target roof deformation for all buildings, calculated using Eqn. 1, is presented in Table 2. As seen
in this table, the target roof drift ratio varies from 1.0% to 1.9%. The pushover analysis will represent an idealized
force-deformation curve. The force–deformation was idealized using a bilinear curve to compute the yield and
ultimate drift. Table 3 summarizes the results of the static pushover analyses. Performance levels were assessed
using the acceptance criteria for concrete columns in Table 10-13 of ASCE 41-13.
Figure 3 presents the pushover curve for the non-ductile prototype building. This figure also highlights the sequence
of the critical component failures. In this building, yielding first started in the beams, then the joints and followed by
the columns. Next, the deformation in the non-ductile columns increased and resulted in shear failure in these
critical columns. Axial failure in the columns leads to the collapse of the structure under gravity load, this failure
pre-empts side-sway collapse. It should be noted that this building was not able to reach the target roof drift as
specified by Table 2.
8.8
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
Columns (Life Safety) 0.36% 569.7 1.06% 809.3 2.97 1.42 2.1
Beams (Collapse Prevention) 0.29% 252.1 1.44% 348.7 5.04 1.38 3.2
Walls (Collapse Prevention) 0.40% 633.3 1.36% 834.0 3.38 1.32 2.1
Wall (Life Safety) 0.37% 1163.1 1.02% 1233.4 2.78 1.06 1.3
Ductile 0.16% 360.2 0.76% 455.2 4.66 1.26 1.7
y/ht: yield drift; Vy: shear yield strength; d/ht: drift at maximum base shear; Vd: maximum base shear; T: displacement ductility; 1 kips = 4.448
kN
Failure
0.08 in Column
0.06
0.04
First Beam Yielding
Non-Ductile Frame
0.02
0
0.00% 0.20% 0.40% 0.60% 0.80% 1.00% 1.20% 1.40% 1.60% 1.80%
Roof Drift [%]
Figure 3–Force-deformation response of the non-ductile frame
Figure 4a, 4c and 4e present the pushover curves for the retrofitted building using the strengthening, weakening and
adding techniques, respectively. As explained before, the strengthening and adding techniques were selected to
achieve both the CP and LS levels, while using the weakening technique, the structure could only achieve the CP
level. The ASCE 41-13 target drifts in these figures refer to targets for the retrofitted buildings. Figure 4b, 4d and 4f
present the component demands for the non-ductile columns, the critical component, for the strengthening,
weakening and adding technique, respectively. The acceptance criteria specified by ASCE 41-13 for the non-ductile
columns are also presented in these figures.
8.9
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
900 70000
ASCE 41 LS
Retrofitted Acceptance Criteria Retrofitted
800 Retrofitted Column
ASCE 41
Frame 60000 Frame CP
LS Level CP Level Acceptance
700 Criteria
Base Shear [kips]
Moment [kips.in]
Frame Column
CP Level 40000
500 Retrofitted Frame LS
Level
400 30000
Non-Ductile Frame ASCE 41 Non-Ductile
300 Target Roof ASCE 41 CP Frame
Drift 20000 Acceptance
200 ASCE 41
Target Roof CP Level Criteria
UnRetrofitted
100 Drift 10000 Column
LS Level
0 0
0.00% 0.20% 0.40% 0.60% 0.80% 1.00% 1.20% 1.40% 1.60% 1.80% 0 0.02 0.04 0.06 0.08
Roof Drift [%] Plastic Rotation Angle [radian]
(a) Base shear vs. roof drift ratio of the non-ductile frame vs. retrofitted building (b) Force-deformation response of the first story column in the non-ductile
using the strengthening technique. frame vs. retrofitted building using the strengthening technique.
800 70000
ASCE 41 CP
700 ASCE 41 Acceptance
Target Roof 60000 Criteria
UnRetrofitted
Drift
600 Column
Base Shear [kips]
CP Level 50000
Moment [kips.in]
Retrofitted
500 Frame
Non-Ductile Frame
40000 CP Level
400
30000 Non-Ductile
300 Frame
Retrofitted 20000
200
Frame
CP Level
100 10000
0 0
0.00% 0.20% 0.40% 0.60% 0.80% 1.00% 1.20% 1.40% 1.60% 1.80% 2.00% 0 0.02 0.04 0.06 0.08
Roof Drift [%] Plastic Rotation Angle [radian]
(c) Base shear vs. roof drift ratio of the non-ductile frame vs. retrofitted building (d) Force-deformation response of the first story column in the non-ductile
using the weakening technique. frame vs. retrofitted building using the weakening technique.
Figure 4-Pushover results for the retrofitted buildings
8.10
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
1600 70000
ASCE 41 ASCE 41 LS ASCE 41 CP
Retrofitted Target Roof Acceptance Criteria Acceptance
1400 60000
Frame Drift UnRetrofitted Column Criteria
LS Level LS Level UnRetrofitted
1200 Column
Base Shear [kips]
Retrofitted
Frame 50000
Retrofitted
Moment [kips.in]
1000 CP Level Frame
40000 CP Level
800
30000
600 Non-Ductile
Frame
20000
400
Non-Ductile Frame ASCE 41
Target Roof Retrofitted
200 10000 Frame LS
Drift
CP Level Level
0 0
0.00% 0.20% 0.40% 0.60% 0.80% 1.00% 1.20% 1.40% 1.60% 0 0.02 0.04 0.06 0.08
Roof Drift [%] Plastic Rotation Angle [radian]
(e) Base shear vs. roof drift ratio of the non-ductile frame vs. retrofitted building (f) Force-deformation response of the first story column in the non-ductile
using the adding technique. frame vs. retrofitted building using the adding technique.
8.11
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
To account for the uncertainties of the structural response under different ranges of earthquake shaking intensities, a
suite of earthquake records was used in the nonlinear dynamic analysis. In this study, incremental dynamic analysis
(IDA) proposed by Vamvatsikos and Cornell (2002) was chosen as the method to quantify the maximum structural
response as the earthquake shaking intensity increases.
Typically a site-specific probability seismic hazard analysis (PSHA) is used for the selection and scaling of the
ground motion. However, due to the fact that nonlinear analyses were performed by IDA, ground motions were not
scaled to match the specific target spectrum at each hazard level. Instead, the ground motions were gradually scaled
until global collapse takes place in the structure.
In this study, the ground motions used for the nonlinear dynamic analyses were selected from earthquakes with
moment magnitude between 6.5 to 7.6 and fault rupture distances between 10 to 45 km. A total of 39 pairs of ground
motion records selected by Haselton and Deierlein (2007) were adopted in this study. This set of ground motions
represent an expanded version of the far-field ground motion set which was used in the FEMA P695 document
(FEMA, 2009). Table 4 summarizes the characteristic of the selected ground motions.
Incremental dynamic analysis (IDA) was chosen as the method to quantify the structural response as the shaking
intensity increases. The outcome of the IDA was used to identify the earthquake shaking intensities when the
structure collapses. The spectral acceleration at the first mode period, Sa (T1), was recorded when the structure
collapsed and plotted using a log-normal distribution. Figure 5 illustrates the system collapse fragility curve for the
existing non-ductile building. This figure also demonstrates that the relationship between the Sa(T1) and probability
of collapse fits well with the log-normal distribution.
It should be noted that the ground motion records were selected and scaled without considering the distinctive
spectral shape of rare (extreme) ground motions, due to difficulties in selecting and scaling a different set of records
for a large set of buildings having a wide range of first mode periods. To account for the important impact of
spectral shape on collapse assessment, shown by Baker and Cornell (2006), the collapse predictions made using the
general set of ground motions and each of the building models were modified using a method proposed by Haselton
et al. (2011). The expected spectral shape of rare (large) California ground motions is accounted for through a
statistical parameter referred to as epsilon, which is a measure of the difference between the spectral acceleration of
a recorded ground motion and the median value predicted by a ground motion prediction equation. A target value of
=1.5 was used to approximately represent the expected spectral shape of severe ground motions that can lead to
collapse of code-conforming buildings in Los Angeles (Appendix B of FEMA P695, 2009).
As proposed in FEMA P695 (2009), the collapse fragilities were adjusted by a spectral shape factor, SSF, computed
using the following equation:
(2)
where depends on building inelastic deformation capacity; depends on the Seismic Design Category (SDC)
and is equal to 1.0 for SDC B/C, 1.5 for SDC D, and 1.2 for SDC E; and is the mean value of the Far-Field
record set listed in Table 4. In all cases, the buildings considered in the current study were designed for a single level
of high seismic ground motions representing Seismic Design Category (SDC) D buildings, therefore, and
are computed using the following equations:
.
. (3a)
. . (3b)
8.12
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
Table 4–Far Field Ground Motion Set (adapted from FEMA P695, (2009))
EQ Index Magnitude Year Event Fault Type Station Name Vs30 (m/s) Campell Distance (km)
1 6.7 1994 Northridge Blind thrust Beverly Hills - 14145 Mulhol 356 17.2
2 6.7 1994 Northridge Blind thrust Canyon Country - W Lost Cany 309 12.4
3 6.7 1994 Northridge Blind thrust LA - Saturn St 309 27
4 6.7 1994 Northridge Blind thrust Santa Monica City Hall 336 27
5 6.7 1994 Northridge Blind thrust Beverly Hills - 12520 Mulhol 546 18.4
6 7.1 1999 Duzce, Turkey Strike-slip Bolu 326 12.4
7 7.1 1999 Hector Mine Strike-slip Hector 685 12
8 6.5 1979 Imperial Valley Strike-slip Delta 275 22.5
9 6.5 1979 Imperial Valley Strike-slip El Centro Array 11 196 13.5
10 6.5 1979 Imperial Valley Strike-slip Calexico Fire Station 231 11.6
11 6.5 1979 Imperial Valley Strike-slip SAHOP Casa Flores 339 10.8
12 6.9 1995 Kobe, Japan Strike-slip Nishi-Akashi 609 25.2
13 6.9 1995 Kobe, Japan Strike-slip Shin-Osaka 256 28.5
14 6.9 1995 Kobe, Japan Strike-slip Kakogawa 312 3.2
15 6.9 1995 Kobe, Japan Strike-slip KJMA 312 95.8
16 7.5 1999 Kocaeli, Turkey Strike-slip Duzce 276 15.4
17 7.5 1999 Kocaeli, Turkey Strike-slip Arcelik 523 13.5
18 7.3 1992 Landers Strike-slip Yermo Fire Station 354 23.8
19 7.3 1992 Landers Strike-slip Coolwater 271 20
20 7.3 1992 Landers Strike-slip Joshua Tree 379 11.4
21 6.9 1989 Loma Prieta Strike-slip Capitola 289 35.5
22 6.9 1989 Loma Prieta Strike-slip Gilroy Array 3 350 12.8
23 6.9 1989 Loma Prieta Strike-slip Oakland - Outer Harbor Wharf 249 74.3
24 6.9 1989 Loma Prieta Strike-slip Hollister - South - Pine 371 27.9
25 6.9 1989 Loma Prieta Strike-slip Hollister City Hall 199 27.6
26 6.9 1989 Loma Prieta Strike-slip Hollister Diff. Array 216 24.8
27 7.4 1990 Manjil, Iran Strike-slip Abbar 724 13
28 6.5 1987 Superstition Hills Strike-slip El Centro Imp. Co. Cent 192 18.5
29 6.5 1987 Superstition Hills Strike-slip Poe Road (temp) 208 11.7
30 6.5 1987 Superstition Hills Strike-slip Westmorland Fire Sta 194 13.5
31 7 1992 Cape Mendocino Thrust Rio Dell Overpass - FF 312 14.3
32 7.6 1999 Chi-Chi, Taiwan Thrust CHY101 259 15.5
33 7.6 1999 Chi-Chi, Taiwan Thrust TCU045 705 26.8
34 7.6 1999 Chi-Chi, Taiwan Thrust TCU095 447 45.3
35 7.6 1999 Chi-Chi, Taiwan Thrust TCU070 401 24.4
36 7.6 1999 Chi-Chi, Taiwan Thrust WGK 259 15.4
37 7.6 1999 Chi-Chi, Taiwan Thrust CHY006 438 13.2
38 6.6 1971 San Fernando Thrust LA - Hollywood Stor FF 316 25.9
39 6.5 1976 Friuli, Italy Thrust (part blind) Tolmezzo 425 15.8
1 in = 0.0254 m
8.13
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
100%
90%
80%
70%
60%
P [Collapse]
50%
40%
30%
20%
10%
0%
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Sa (T1 = 2.36 s) [g]
Figure 5–Probability of collapse vs. Sa(T1) for non-ductile (unretrofitted) building, before spectral shape
adjustment
Figure 6 shows the fragility curves for all buildings and performance levels considered in this study adjusted for the
spectral shape. The figure clearly illustrates the significant variability in the probability of collapse of the retrofitted
structures, which fill the spectrum between the non-ductile and modern design buildings. It should be noted that the
retrofitted building using shear walls and upgraded to the LS performance level has a collapse performance identical
to the modern designed building (building similar to the 8-story non-ductile perimeter frame but designed using
modern building codes), while the weakening retrofit approach only provides a slight improvement on the
probability collapse over the existing building.
RC type b1 0 SSF
Existing 0.16 1.5 0 1.27
Columns (Collapse Prevention) 0.18 1.5 0 1.31
Retrofitt
The probability of collapse at the MCE (2% in 50 yr return period) intensity (for each building) is shown with
circles in Figure 6. The circles indicate a variety of collapse ratios at this intensity level for the different retrofitting
techniques and performance objectives. The first mode period of the existing, retrofitted, and modern designed
buildings are all considered as being in the intermediate period range and therefore the dominant period has an
inverse relationship with the spectral acceleration. This variety in collapse performance can be compared by
normalizing the collapse fragility curves with the spectral acceleration at the MCE intensity, as shown in Figure 7.
8.14
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
100%
90%
80%
70%
60%
P [Collapse]
50%
40%
Existing
Columns (Collapse Prevention)
30%
Columns (Life Safety)
Beams (Collapse Prevention)
20% Walls (Collapse Prevention)
Walls (Life Safety)
10% Ductile
P[Collapse | Sa 2% in 50yrs]
0%
0 0.5 1 1.5 2 2.5
Sa (T1) [g]
Figure 7 clearly indicates when the spectral acceleration is normalized to the MCE intensity, the collapse fragilities
for the different retrofitted buildings have a smaller variability compared to the original collapse fragility curve
(Figure 8). The acceptable probability of collapse at the MCE level for new buildings, defined as 10% by FEMA
P695, is also shown in Figure 7. The non-ductile existing building clearly does not meet this criterion and the
modern code-conforming structure clearly passes this criterion; while most of the retrofitted buildings are very close
to this criterion. The exception is the CP column retrofit which indicates a probability of collapse of 31% at the
MCE level. This poorer performance can be explained by the fact that the CP column retrofit only addresses the
vulnerability of the columns on the first two stories (see Figure 2). Since the 3rd story columns only barely satisfied
the CP acceptance criteria in the assessment of the CP column retrofit, it does not take a significant increase in
ground motion intensity to initiate shear and axial failure of the columns in this story. For 95% of the ground
motions considered, the analysis predicts collapse of the CP column retrofit due to shear and axial failure of the
columns on the 3rd story. In contrast, the CP beam retrofit changes the collapse mode of the structure from gravity-
load collapse due to column failure, to side-sway collapse of a strong-beam weak-column system. Such a system is
less sensitive to small increases in the ground motion intensity. The wall retrofit provides a reduction in drift
demands up the height of the building, and hence, protects all stories from failure.
8.15
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
100%
90%
80%
70%
60% Existing
P [Collapse]
10%
0%
0 1 2 3 4 5
Sa (T1)/Sa2%in50yrs
Different collapse performance metrics are reported in Table 6. The predicted median collapse capacity,
MedianSacol(T1), of the retrofitted structures is approximately 1.5 to 5 times larger than the existing non-ductile
building. The probability of collapse in 50 years shows a higher difference, this performance metric decreases from
52% for the non-ductile building down to a range of 7% - 11% for most of the retrofitted buildings, which is
comparable to the 4% probability of collapse for the modern design structure.
It should be noted that the probability of collapse in 50 years is the preferred collapse metric for comparison to the
probability of collapse at the MCE level. There are two main points for this preference. First of all, the probability of
collapse in 50 years, which is obtained by combining the collapse fragility curve with the hazard curve, covers a
larger range of hazard levels compared to the MCE level. Secondly, the probability of collapse at the MCE level is
influenced by the change in building period as well as change in the fragility curve. As a result, the probability of
collapse in 50 is used for the comparison of the different buildings.
An overall summary of the three sets of buildings is presented in Table 7. This table demonstrates an estimate of the
change in the material required to rehabilitate the existing non-ductile building. Modifying the beams would
8.16
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
probably cost the least to retrofit (with the assumption that the labor cost to perform all three techniques would
approximately be the same) but has less benefit in terms of the seismic collapse safety. On the other hand, adding
shear walls and retrofitting to the CP and LS performance levels would cost the most but has the lowest probability
of collapse.
Table 7–System performance of the three set of buildings
ASCE 41 (Reinforced
Material List
Concrete Columns) Collapse
Retrofit Performance
Building Type Margin (P[Col]
Technique Level Change (compared to
Acceptance Weight [kips] in 50 yrs)
Demand existing building)
Criteria
Concrete Steel Concrete Steel
Existing (Non-
- - 0.043 0.07 786 48 - - 52%
Ductile)
Enhance Collapse
0.06 0.05 803 49 2% 2% 16%
Columns Prevention
Enhance
Life Safety 0.035 0.03 836 50 6% 4% 8%
Columns
Weaken Collapse
Retrofitt 0.043 0.03 786 48 0% 0% 11%
Beams Prevention
Add Collapse
0.043 0.032 839 50 7% 6% 7%
ShearWall Prevention
Add
Life Safety 0.0161 0.012 868 56 10% 18% 7%
ShearWall
1 kips = 4.448 kN
CONCLUSIONS
Advancements in nonlinear dynamic analysis, seismic hazard analysis, and performance-based earthquake
engineering are enabling more scientific assessment of structural collapse risk. The main objective of this study is to
quantify and compare the seismic safety of non-ductile, retrofitted, and ductile RC frame buildings. At present, the
general assumption when retrofitting a non-ductile building according to ASCE 41-13 is that their seismic
performance is enhanced such that it can reach the performance corresponding to ductile buildings designed based
on current seismic codes. However, as shown in this study, different retrofitting measures do not exhibit the same
collapse performance metrics as modern ductile buildings.
Based on the evaluation of the eight-story RC frame in Los Angeles considered in this study, the following
observations are made:
The considered retrofit schemes (column strengthening, beam weakening, and adding shear wall) using
ASCE 41-13 acceptance criteria provided an intermediate level of collapse performance, falling between
that achieved by a non-ductile frame and that achieved by a modern code-conforming ductile RC moment
frame.
The study finds that retrofitting scheme of modifying the columns up to only the CP level has the least
beneficial effect regarding the seismic collapse safety and conversely adding a shear wall has the highest
improvement in collapse performance. The column CP retrofit is vulnerable to collapse due to failure of
unretrofitted columns above the retrofitted stories. Modification of collapse mode due to beam weakening
or reduction of drifts at all stories with an added shear wall resulted in better collapse performance metrics.
The probability of collapse in 50 years decreases from 52% (for the non-ductile building) down to a range
of 7% - 16% for the retrofitted buildings; higher than the 4% probability of collapse in 50 years for the
modern code-conforming structure.
8.17
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
The predicted median collapse capacities of the retrofitted structures are approximately 1.5 to 5 times larger
than the existing non-ductile building depending on the retrofit technique and the selected performance
level.
The above observations are valid for the building considered in this study. Future research is required to determine
if these observation hold for other structural systems, retrofit techniques and seismic hazard.
ACKNOWLEDGMENTS
This work is supported in part by the Natural Sciences and Engineering Research Council of Canada through the
Canadian Seismic Research Network. Any opinions, findings, and conclusion or recommendations expressed in this
work are those of the authors and do not reflect those of the organizations or individuals noted here.
REFERENCES
American Concrete Institute, 2005. Building Code Requirements for Structural Concrete (ACI 318). American
Concrete Institute, Farmington Hills, MI.
American Society of Civil Engineers, 2013. Seismic Evaluation and Retrofit of Existing Buildings, American
Society of Civil Engineers (ASCE/SEI 41-13), pre-publication edition for public comment and final review.
American Society of Civil Engineers. 2005. ASCE 7-05: Minimum Design Loads for Buildings and Other
Structures, Reston, VA.
Anagnos, T., Comerio, M.C., Goulet, C., Steele, J. and J. P. Stewart, 2010. Development of a concrete building
inventory: Los Angeles case study for the analysis of collapse risk. Proc. 9th US National & 10th Canadian Conf.
on Earthquake Eng., EERI and Canadian Assoc. for Earthquake Eng., July 25-29, 2010, Paper No. 48.
Baker J. W. and Cornell C. A., 2006. Spectral shape, epsilon and record selection, Earthquake Engineering &
Structural Dynamics 35, 1077-1095.
Baradaran Shoraka, M. 2013. Collapse assessment of non-ductile reinforced concrete buildings,PhD thesis,
Civil Engineering Department, University of British Columbia, BC, Canada.
Baradaran Shoraka, M., Yang, T. Y., and Elwood, K. J., 2013. Seismic loss estimation of non-ductile reinforced
concrete buildings, Earthquake Engineering and Structural Dynamics 42, 297–310.
Bazzurro, P., and Cornell, C. A., 1994. Seismic hazard analysis of nonlinear structures. I: Methodology, Journal
of Structural Engineering 120, 3320.
Bazzurro P. and Cornell C. A., 1999. Disaggregation of seismic hazard, Bulletin of the Seismological Society of
America 89, 501-520.
Elwood K. J., Matamoros A. B., Wallace J. W., Lehman D. E., Heintz J. A., Mitchell A. D., Moore M. A.,
Valley M. T., Lowes L. N., Comartin C. D. and Moehle J. P., 2007. Update to ASCE 41 concrete provisions,
Earthquake Spectra 23, 493-523.
Elwood K. J. and Moehle J. P., 2008. Dynamic collapse analysis for a reinforced concrete frame sustaining
shear and axial failures, Earthquake Engineering & Structural Dynamics 37, 991-1012.
Elwood K. J., 2004. Modelling failures in existing reinforced concrete columns, Canadian Journal of Civil
Engineering 31, 846-859.
Elwood, K.J. 2002. Shake table tests and analytical studies on the gravity load collapse of reinforced concrete
frames. Ph.D. Dissertation, Department of Civil and Environmental Engineering, University of California, Berkeley,
CA.
Federal Emergency Management Agency (FEMA P695A), 2009. Recommended Methodology for
Quantification of Building System Performance and Response Parameters, Washington, D.C.
8.18
@Seismicisolation
@Seismicisolation
Collapse Assessment of Non-Ductile, Retrofitted and
Ductile Reinforced Concrete Frames
Federal Emergency Management Agency (FEMA 547), 2006. Techniques for the Seismic Rehabilitation of
Existing Buildings, Washington, D.C., 571 pp.
Federal Emergency Management Agency (FEMA-356), 2000. Pre-Standard and Commentary for the Seismic
Rehabilitation of Buildings. Federal Emergency Management Agency, Washington, D.C.
Haselton, C. B., Liel, A. B., Taylor Lange, S., and Deierlein, G. G., 2008. Beam-Column Element Model
Calibrated for Predicting Flexural Response Leading to Global Collapse of RC Frame Buildings. Pacific Earthquake
Engineering Research Center, Berkeley, CA
Haselton, C. B., Liel, A. B., Deierlein, G. G., Dean, B. S., Chou, J. H., 2011. Seismic Collapse Safety of
Reinforced Concrete Buildings: I. Assessment of Ductile Moment Frames, Journal of Structural Engineering,
137(4), pp. 481-491.
Haselton, C. B., Baker, J. W., Liel, A. B. and Deierlein, G. G, 2011. Accounting for Ground Motion Spectral
Shape Characteristics in Structural Collapse Through an Adjustment for Epsilon, Journal of Structural Engineering,
137(3), pp. 332-344.
Ibarra L. F., Medina R. A. and Krawinkler H., 2005. Hysteretic models that incorporate strength and stiffness
deterioration, Earthquake Engineering & Structural Dynamics 34, 1489-1511.
ICC 2003. International Building Code, International Code Council, Washington, D.C.
Kam, W. Y., and Pampanin, S., 2009. Experimental and numerical validation of selective weakening retrofit for
existing non-ductile RC frames. Proc. Improving the Seismic Performance of Existing Buildings and Other
Structures.
Liel, A.B., 2008. Assessing the collapse risk of California’s existing reinforced concrete frame structures:
Metrics for seismic safety decisions. Ph.D. thesis, Stanford University. United States.
Liel, A. B. and Deierlein, G. G., 2013. Cost-Benefit Evaluation of Seismic Mitigation Alternatives for Older
Reinforced Concrete Frame Buildings, Earthquake Spectra.
Lowes L. N. and Altoontash A., 2003. Modeling reinforced-concrete beam-column joints subjected to cyclic
loading. Journal of Structural Engineering 129, 1686-1697.
Pacific Earthquake Engineering Research Center, 2009. Open system for earthquake engineering simulation
_OpenSees_ framework-Version 2.1.0.
Pampanin, S., 2005. Emerging Solutions for High Seismic Performance of Precast/Prestressed Concrete
Buildings. Journal of Advanced Concrete Technology (ACT), 3(2), 202-223.
Vamvatsikos D. and Cornell A., 2002. Incremental dynamic analysis, Earthquake Engineering and Structural
Dynamics 31, 491-514.
Viti, S., Cimellaro, G. P., and Reinhorn, A. M., 2006. Retrofit of a hospital through strength reduction and
enhanced damping. Smart Structures and Systems, 2(4), 339-355.
Yavari S., Elwood K. J., Lin S., Wu C., Hwang S. and Moehle J. P., 2009a. Experimental study on dynamic
behaviour of multi-story reinforced concrete frames with non-seismic detailing. Proc. Improving the Seismic
Performance of Existing Buildings and Other Structures, pp. 489-499. 489-499.
Yavari S., Elwood K. J. and Wu C., 2009b. Collapse of a nonductile concrete frame: Evaluation of analytical
models, Earthquake Engineering and Structural Dynamics 38, 225-241.
8.19
@Seismicisolation
@Seismicisolation
M. Baradaran Shoraka et al.
8.20
@Seismicisolation
@Seismicisolation
SP-297—9
Synopsis: Case studies on seismic assessment and rehabilitation of reinforced concrete buildings are discussed based
on the projects in which Degenkolb Engineers has been involved in the past 5 years. Design, analysis and challenges
are discussed to present applications of ASCE 31-03, Seismic Evaluation of Existing Buildings and ASCE 41-06,
Seismic Rehabilitation of Existing Buildings.
Keywords: reinforced concrete; seismic evaluation; seismic retrofit, repair, rehabilitation, non-ductile concrete structure
ACI member InSung Kim is a Structural Engineer at Degenkolb Engineers, San Francisco, CA. He received his BS
from Yonsei University, Seoul, South Korea; and his MS and PhD from the University of Texas at Austin, Austin,
TX. He is a member of ACI Committees 369, Seismic Repair and Rehabilitation; 374, Performance-Based Seismic
Design of Concrete Buildings; and 440, Fiber-Reinforced Polymer Reinforcement. His research interests include
seismic design and rehabilitation of reinforced concrete structures.
Garrett Hagen also works with Degenkolb Engineers, in their Los Angeles office. He graduated with his BS and
MS from California Polytechnic University, San Luis Obispo, where his graduate research involved performance-
based analysis of concrete structural walls. He is a member of ACI Committees 369 and 374. He actively
contributes to research regarding the modeling, behavior, and design of reinforced concrete structures.
INTRODUCTION
Degenkolb Engineers is a structural engineering firm specializing in seismic engineering and has five offices
along the West Coast of the USA. The firm has a variety of experience in evaluation and retrofit of building
structures in the high seismic regions of the USA. In this paper, case studies of their seismic evaluation and retrofit
projects of RC (reinforced concrete) buildings in the past 5 years are presented. The projects utilized ASCE 31-03,
Seismic Evaluation of Existing Buildings and ASCE 41-06, Seismic Rehabilitation of Existing Buildings, and its
supplement. Deficiencies of the RC buildings were identified, and retrofit schemes to mitigate the deficiencies were
developed, based on ASCE 31-03 and ASCE 41-06. In this paper, typical deficiencies and retrofit schemes are
discussed along with some of the challenges in the application of the documents. In addition, the areas requiring
further improvement in the concrete chapter of ASCE 41-13, Seismic Evaluation and Retrofit of Existing Buildings,
which is the combined document of the current ASCE 31 and 41 and ACI 369R-11 Guide for Seismic Rehabilitation
of Existing Concrete Frame Buildings and Commentary, are discussed.
9.1
@Seismicisolation
@Seismicisolation
I. Kim and G. Hagen
Building Descriptions
Seismic evaluation and retrofit of ten concrete buildings, located in California and built between the 1940’s and
1970’s, are reported in this paper. A summary of the building descriptions and their major deficiencies and retrofit
schemes is given in Table 1. The number of stories of the buildings ranges from 1 story to 10 stories, which can be
considered as low to mid-rise buildings. Most of the buildings had shear walls for their existing lateral force
resisting system (LFRS), and concrete frames were typically used to resist gravity force and/or as supplementary
LFRS. Various types of slabs were used as the diaphragms of the structures. One- or two-way RC slabs were used
throughout the years, and some of the buildings built after the 1960’s had post-tensioned slabs as their diaphragms.
Because the State of California requires mandatory seismic upgrade of hospital buildings, hospital buildings are
evaluated and retrofitted more often than other types of building occupancies. Six out of the ten buildings addressed
in this paper are hospitals.
Upon the request from the building owners, the information which could be used to identify the buildings is not
disclosed.
Major Deficiencies
Some of the classic deficiencies identified in ASCE 31-03 were observed in the building evaluations. Global
issues, such as torsional irregularities, vertical irregularities, potential pounding with adjacent structures, and weak
or soft stories, were observed in the various buildings. Specific issues, such as non-ductile concrete frames with
shear critical captive columns, large openings in shear walls, and unconfined boundary elements, were found as well.
Severe vertical irregularities were typically observed for the buildings with podiums (Building 1, 2, 4 and 8).
An example of a building with a podium is shown in Fig. 1. At the top of podium shown in Fig. 1 where the setback
occurred, the strength and stiffness of the building changed abruptly, requiring large forces to be transferred through
podium diaphragms and sometimes creating significant stress concentrations. For some of the buildings, changes in
LFRS occurred in the transition from above the podium level to below.
Captive columns were typically observed in the perimeter frames with spandrel beams (Building 2 and 8, Fig.
2). The spandrels framed into the columns eccentrically, and the eccentricity combined with high shear force in the
captive columns created the potential for damage concentration in these frames (Fig. 2). Eccentric beam column
connections were common not only in the spandrel frames but also in moment frames located in the building
perimeters.
9.2
@Seismicisolation
@Seismicisolation
Seismic Assessment and Retrofit of Existing RC Buildings:
Case Studies from Degenkolb Engineers
Figure 2 — “Captive” columns (elevation view) and eccentrically connected spandrel and column (plan view)
Diaphragm capacity and diaphragm connections to the LFRS and/or gravity force resisting system were
problematic in most of the buildings evaluated. Reinforcement in the slab of the existing buildings was typically
designed for gravity force only, so the amount of existing reinforcement which could be utilized for seismic load
transfer was typically not sufficient. The connection from diaphragm to the vertical lateral or gravity force resisting
system had deficiencies as well. For example, Fig. 3 represents a case in which the dowels connecting the slab to
shear wall were not sufficient and were not adequately anchored to develop yield. Large diaphragm openings
observed in Buildings 9 and 10 created a weak link for the lateral force transfer. PT slabs in Building 4 and 6 did not
contain conventional reinforcement so the diaphragm did not have enough capacity to transfer the seismic load, and
the connections to the columns were not adequate. The failure of the diaphragm and its connections may be one of
the main causes for collapse of a structure as observed in the CTV Building and Pyne Gould Building after the 2011
Christchurch Earthquake (C. Hyland and A, Smith 2012, Canterbury Earthquake Royal Commission, 2012).
Rehabilitation schemes for these issues generally consisted of installing additional collectors and/or adding new
frames or shear walls to reduce demand in the diaphragm.
High tensile or compressive force in the vertical member located close to the end of the LFRS, due to global
building overturning, was one of the common deficiencies observed in the mid-rise buildings. The continuous deep
(stiff) spandrel beams in the perimeter of the building created high uplift and compressive demands in the walls and
columns at the building corners. For building analysis, precautions were taken to accurately capture stiffness at
building transitions, as it is possible for axial forces to be overestimated when the building base is numerically
modeled with restraints to vertical transition (e.g. fixed or pinned supports).
Many of the deficiencies observed in the buildings are well summarized in Program Plan for the Development
of Collapse Assessment and Mitigation Strategies for Existing Reinforced Concrete Buildings (NEHRP, 2010).
9.3
@Seismicisolation
@Seismicisolation
I. Kim and G. Hagen
Retrofit Schemes
9.4
@Seismicisolation
@Seismicisolation
Seismic Assessment and Retrofit of Existing RC Buildings:
Case Studies from Degenkolb Engineers
9.5
@Seismicisolation
@Seismicisolation
I. Kim and G. Hagen
Analysis Techniques
After Tier 1 evaluation using the checklists in ASCE 31-03, 3D numerical models are typically created for
further evaluation or retrofit design. The initial analysis generally begins with linear static and dynamic analysis, but
nonlinear analysis may be adopted depending on the characteristics of the project. Depending on the analysis
techniques selected, different structural analysis programs may be used. 3D numerical models of a building
developed using different analysis programs are shown in Fig. 7. Typically, after an engineer reports that the
building does not comply with Tier 1 criteria to the owner, the engineer provides the evaluation results with
recommendations for detailed evaluation of the Tier 1 deficiencies, to determine whether the deficiencies in fact
require mitigation through retrofit.
Based on review of the existing drawings and completion of the checklists during the Tier 1 evaluation and
linear analysis, the engineer usually can identify the critical deficiencies which jeopardize building performance,
The engineer can then provide several options of retrofit schemes to the building owner. Some retrofit schemes may
require advanced nonlinear analysis to justify the design. In those cases, the added cost for advanced analysis has the
potential for making the most cost-effective retrofit scheme for the project possible. It is important that engineers
understand they should not use nonlinear analysis with the intent of justifying performance of existing non-ductile
components which are critical elements for resisting gravity or seismic load and may experience significant
deformation during earthquakes, even if they show acceptable performance in the nonlinear analysis. More complex
numerical models do not always result in more reliable or accurate results and nonlinear internal load distribution
calculated in a model can be very different from the load distribution in a real building. Engineers should be more
cautious about uncertainty of seismic loads and approximate natures of analysis results when their decisions are
9.6
@Seismicisolation
@Seismicisolation
Seismic Assessment and Retrofit of Existing RC Buildings:
Case Studies from Degenkolb Engineers
based on the insights obtained from nonlinear analysis. Of course, nonlinear analysis can be utilized to resolve
unrealistic concentrations of internal forces shown in linear analysis by allowing redistribution when the global
structural system of a building has adequate strength, deformation capacity and redundancy. However, the best use
of nonlinear analysis may be the verification of the retrofit design, employing the capacity-based design concept and
ductile details for load redistribution.
A good example for application of nonlinear analysis is the use of foundation springs. The springs allow the
reduction of unrealistic high axial forces due to global building overturning (Building 2, 5 and 8). The analysis
should be performed with upper and lower bound spring stiffness as described in ASCE 41 in order to capture
uncertainty of the foundation stiffness.
Nonlinear analysis is also effective for buildings in which deformation compatibility is of primary concern, in
which case localized deformations are critical.
Figure 7 — 3D numerical models developed using ETABS (left) and PERFORM 3D (right)
9.7
@Seismicisolation
@Seismicisolation
I. Kim and G. Hagen
1968
Linear static and
3 stories New RC walls and columns with
Bldg. 3 Vertical discontinuity dynamic analysis
Shear walls and moment frames new foundation
Diaphragm strength Nonlinear static
School New collectors
analysis
Northern California
Vertical discontinuity
Torsional irregularity
1968 New RC walls and boundary
Diaphragm strength Linear static and
3 stories elements
Shear wall strength dynamic analysis
Bldg. 4 Shear walls and moment frames Steel plate collector
Deformation capacity of post Nonlinear static
School Steel channel support at slab
tensioned slab with column (no analysis
Northern California column joints
continuous reinforcement
through column)
New shear wall
New structural slab-on-grade at the
first floor level to connect new wall
Weak story to basement wall
1949 (4 floors were added in Vertical discontinuities RC infill of wall openings
1965) Weak shear walls and coupling Strengthening of existing concrete
Linear static and
3 stories + basement original beam wall and spandrel
dynamic analysis
7 stories + basement after 1965 Torsional irregularity New below-grade concrete shear
Bldg. 5 Nonlinear static
Shear walls Non ductile concrete frames walls to strengthen load path at sub-
analysis
Hospital High drift basement foundation level
Northern California New structural connection between
basement floor slab and perimeter
concrete walls
Connecting two buildings
Chimney (stack) brace
1969
3 stories Non-ductile columns
New exterior BRB with foundation Linear static
Bldg. 6 Moment frames Two-way PT slab without
Concrete collectors analysis
Office conventional reinforcement
Northern California
9.8
@Seismicisolation
@Seismicisolation
Seismic Assessment and Retrofit of Existing RC Buildings:
Case Studies from Degenkolb Engineers
1948
Lack of diaphragm strength in New shear wall with footing Linear static and
4 stories
Transverse direction Slab-column connection dynamic analysis
Bldg. 8 Pier-spandrel shear walls
Weak story strengthening Nonlinear static
Hospital
Non-ductile concrete frames FRP column jacketing analysis
Northern California
1968
8 stories Vertical discontinuity New shear walls with axial and
Linear dynamic
Bldg. 9 Shear walls Inadequate load path near lateral force-resisting piles
analysis
Hospital diaphragm openings Steel plate collectors
Southern California
Vertical discontinuity
1966 Shear wall strength
FRP strengthening of walls
6 stories Diaphragm strength
FRP column jacketing Linear dynamic
Bldg 10 Shear walls Inadequate load path near
Addition of and strengthening of analysis
Hospital diaphragm openings
collector elements
Southern California Column strength under
discontinuous shear walls
9.9
@Seismicisolation
@Seismicisolation
I. Kim and G. Hagen
The ASCE 31-03 and 41-06 standards have been combined into a new addition of ASCE 41-13, Seismic
Evaluation and Retrofit of Existing Buildings, and most of the updates in the concrete chapter of ASCE 41-13 were
based on ACI 369R-11, Guide for Seismic Rehabilitation of Existing Concrete Frame Buildings and Commentary.
The following section discusses areas requiring further improvement in ASCE 41 and ACI 369R based on the
experience in seismic evaluation and retrofit of the ten buildings in this study.
When a RC column, shear wall, or diaphragm element is strengthened with fiber reinforced polymer (FRP),
the guides for selection of MPs and AC for the strengthened members are required (Fig. 5).
If a RC column or collector is strengthened with steel plates, engineers need not only MP’s and AC
information, but also guides for calculating stiffness of the strengthened member (Fig. 6).
If a shear wall has a concrete overlay on one side or an existing column is jacketed by additional concrete
and transverse reinforcement, guides for selecting the MP’s and AC are required.
Further guides for modeling and selection of AC and detailing requirements of strengthened members to
achieve specific performance (i.e. deformation capacity) are necessary for more reliable retrofit. Guides for
strengthened members with techniques commonly used in practice would be helpful for engineers to understand and
adopt these techniques.
9.10
@Seismicisolation
@Seismicisolation
Seismic Assessment and Retrofit of Existing RC Buildings:
Case Studies from Degenkolb Engineers
reinforcement in the beam-column joint, or deformation capacities of perimeter spandrel frames with captive
columns, may be helpful.
Nonlinear Response History Analysis
The current provisions for nonlinear analysis in the Concrete Chapter of ASCE 41 are mainly applicable to the
nonlinear static (push-over) analysis. Because nonlinear dynamic analysis is commonly used in practice and is
sometimes required by the jurisdiction, more guidelines for nonlinear dynamic analysis are needed. The hysteresis
information of a member (i.e., energy dissipation during cyclic behavior) and stiffness modification for the nonlinear
dynamic procedure might be critical information required by engineers to apply provisions for the nonlinear
dynamic analysis. Recommendations for magnitude of damping (both within and outside of the elastic range) are
needed to ensure engineers are not over-accounting for damping when explicitly modeling hysteretic behavior.
Additionally, in ASCE 41, there are currently not sufficient guidelines for cyclic degradation.
CONCLUDING REMARKS
Based on experiences of Degenkolb Engineers in seismic evaluation and retrofit of concrete buildings in the
past 5 years, major structural deficiencies, strengthening schemes, and analysis techniques were discussed in this
paper. The major structural deficiencies observed in the buildings corresponded to those recognized throughout past
earthquakes. To mitigate those deficiencies, new vertical and/or horizontal LFRS were added or the existing systems
were strengthened. New lateral systems or materials for the existing members were added to the buildings and some
of these consisted of materials other than concrete. Knowledge regarding the interaction between new and existing
elements was critical to achieve the intended performance of the building after rehabilitation. It would be helpful to
engineers if more guides regarding element interaction were provided in the future editions of ASCE 41 and ACI
369R.
Horizontal LFRS, diaphragm, and collectors were strengthened in all the buildings in this study and are as
important as the vertical LRFS for seismic force resistance and transfer. More guidelines for evaluation and retrofit
of the horizontal LFRS are desirable in ASCE 41 and ACI 369R to enhance engineers’ understanding of these
system and to achieve performance consistent with the performance of the vertical LFRS.
Advanced analysis like the nonlinear dynamic procedure is often used in the evaluation of existing and
retrofitted structures. It is desirable that more information regarding nonlinear dynamic analysis, such as cyclic
hysteresis and variation of stiffness of conventional concrete members or members strengthened with materials other
than concrete, should be provided in ASCE 41 and ACI 369R so engineers could have more consistent procedures
for determining load redistribution and energy dissipation in concrete members. It should be emphasized that
engineers should not use nonlinear analysis with the intent of justifying performance of existing non-ductile
components which are critical elements for resisting gravity or seismic load and may experience significant
deformation during earthquakes. Engineers should be more cautious about uncertainty of seismic loads and
approximate natures of analysis results when their decisions are based on the insights obtained from nonlinear
analysis. The best use of nonlinear analysis may be the verification of the retrofit design, employing the capacity-
based design concept and ductile details for load redistribution
ACKNOWLEDGEMENTS
The authors would like to extend their gratitude to their colleague engineers within Degenkolb Engineers who
provided the information regarding the projects for these case studies.
REFERENCES
ACI Committee 369 (2011), Guide for Seismic Rehabilitation of Existing Concrete Frame Buildings and
Commentary (ACI 369R-11), American Concrete Institute, Farmington Hills, Mich., USA
American Society of Civil Engineers (ASCE), (2003), Seismic Evaluation of Existing Buildings, ASCE 31-03,
Reston, VA., USA
9.11
@Seismicisolation
@Seismicisolation
I. Kim and G. Hagen
American Society of Civil Engineers (ASCE), (2007), Seismic Rehabilitation of Existing Buildings, ASCE 41-06,
Reston, VA., USA
American Society of Civil Engineers (ASCE), (2013, expected), Seismic Evaluation and Retrofit of Existing
Buildings, ASCE 41-13, Reston, VA.,
C. Hyland, A, Smith (2012), CTV Building Collapse Investigation, Department of Building and Housing, New
Zealand
Canterbury Earthquake Royal Commission (2012), Final Report Volume 2, The Performance of Christchurch CBD
Buildings, Christchurch, New Zealand
NEHRP Consultants Joint Venture, (2010), Program Plan for the Development of Collapse Assessment and
Mitigation Strategies for Existing Reinforced Concrete Buildings, US Department of Commerce, National
Institute of Standards and Technology, USA
9.12
@Seismicisolation
@Seismicisolation
SP-297—10
Synopsis: There are many vulnerable reinforced concrete (RC) buildings located in earthquake-
prone areas around the world. These buildings are characterized by the lack of seismic details
and corresponding non-ductile behavior and significant potential of partial and global collapse.
One of the current challenges of the earthquake engineering profession and research
communities is the identification of such buildings and determination of effective and
economical retrofit methods for response enhancement. Identification of these buildings is not a
trivial task due to the various sources of non-ductile behavior and the large number of involved
sources of uncertainty. Furthermore, accurate determination of collapse-prone buildings is
important from an economical perspective. Unfortunately, there are not enough economical
resources to retrofit all the non-ductile buildings that have the symptoms for collapse potential.
In order to use the available monetary resources in an effective manner, these buildings should
be accurately and reliably ranked to identify those that are most vulnerable to collapse. This
paper intends to provide a contribution to the accurate determination of the most collapse
vulnerable non-ductile RC buildings by discussing the methods from existing literature and
exploring the research needs related to (a) gravity load failure modeling and (b) consideration of
sources of uncertainty in an efficient manner.
10.1
@Seismicisolation
@Seismicisolation
K. M. Mosalam and S. Günay
Selim Günay is a research scientist at the NEES laboratory at the University of California,
Berkeley. He executes the experiments of various NEES projects and conducts research on
various aspects of earthquake engineering including progressive collapse, hybrid simulations,
and performance-based earthquake engineering of different structural systems.
INTRODUCTION
It is a well-known fact that there are many vulnerable reinforced concrete (RC) buildings located in earthquake-
prone areas around the world. These buildings are characterized by the lack of seismic details (such as lack of
confinement at the beam and column ends and the beam-column joints, strong beam-weak column proportions, and
presence of shear-critical columns) and the corresponding non-ductile behavior and significant potential of partial
and global collapse, posing threats to human life. One of the current challenges of the earthquake research and
profession is the identification of such buildings and the determination of effective and economical retrofit methods
for enhancing their seismic response. Identification of these buildings is not an easy task due to the various sources
of non-ductile behavior and the large uncertainties involved in material characteristics, amount of reinforcement,
geometry, etc. Furthermore, accurate determination of collapse-prone buildings is important from an economical
perspective. Unfortunately, there are insufficient economical resources to retrofit all the non-ductile buildings that
have the symptoms for collapse potential. In order to effectively use these limited monetary resources, these
buildings should be accurately and reliably ranked to identify those that are most vulnerable to collapse. This paper
intends to provide a contribution to the accurate determination of the most collapse vulnerable non-ductile buildings
by discussing relevant methods from existing literature and exploring the research needs related to (a) gravity load
failure modeling and (b) consideration of uncertainty in an efficient manner.
Current state of knowledge and practice in nonlinear static and dynamic analyses has the ability to determine
side-sway collapse that occurs due to lateral dynamic instability at excessive lateral displacements, when the lateral
strength of the structure degrades significantly. However, non-ductile RC buildings mostly collapse by losing
gravity load carrying capacity, much before reaching these excessive displacements, as demonstrated in Figure 1. At
point A of this figure, the lateral load resistance starts to degrade because of various events such as in-plane or out-
of-plane failure of infill walls or shear failure of columns. This is followed by the loss of the gravity load carrying
capacity of the lateral and gravity load resisting components, starting with points B and C, respectively. The first
part of this paper explores this relatively neglected, but significantly important, issue of gravity load carrying
capacity by discussing the methods of gravity load failure modeling in collapse simulations and pointing out further
research needs. It is to be noted that the term “gravity load failure” is used in the rest of the paper to refer to the loss
of gravity load carrying capacity.
There are two alternative options that can be considered for gravity load failure modeling of the elements of a
structure: (a) explicit modeling, (b) implicit modeling, i.e. non-simulated failure. Explicit modeling of gravity load
failure consists of two stages: (1) detection of gravity load failure and (2) post-failure modeling. In the next section
of the paper, the advantages and disadvantages of the implicit modeling are presented. Moreover, three different
approaches are discussed for the purpose of employing them in the second stage of explicit modeling. These
approaches are (i) element removal, (ii) assignment of low stiffness to a failed element, and (iii) representation of
the post-failure response of failed elements with degradation. A subsequent section of the paper explores the models
that can be used for failure detection of various structural members in explicit modeling. The final section related to
gravity failure modeling is comprised of the modeling of the structural elements primarily designed to resist the
gravity loads with insignificant contribution to the lateral load resistance. While the gravity load resisting system can
10.2
@Seismicisolation
@Seismicisolation
Towards an Accurate Determination of Collapse
Vulnerable Reinforced Concrete Buildings
be approximately considered for the case of side-sway collapse, its explicit modeling significantly complements the
explicit gravity load failure modeling of the primary lateral resisting system, as shown in Figure 1.
Lateral
Capacity
Vertical Ductile RC
Capacity Building
Non-ductile RC
Vertical Building
Capacity
Figure 1: Lateral and vertical responses of RC buildings (Modified from Holmes, 2000)
As mentioned above, non-ductile RC buildings generally involve significant amount of uncertainties. Therefore,
a substantial number of collapse simulations may need to be conducted for accurate identification of buildings which
are most vulnerable to collapse. The final part of the paper presents a deterministic sensitivity analysis method from
the literature, the so-called tornado diagram, as a method to identify the sources of uncertainties which are most
influential on the seismic response. Accordingly, this practical method can handle the effect of uncertainty more
efficiently in collapse simulations of non-ductile RC buildings by only incorporating these influential sources of
uncertainty as probabilistic, while those that are less influential can be incorporated as deterministic.
In order to model the gravity load failure of the lateral load resisting components of non-ductile RC buildings, there
are two alternative options that can be considered. First option is the explicit modeling of gravity load failure, while
the second one is the determination of gravity failure of the components implicitly, through post-processing without
modeling of gravity failure (non-simulated failure). In the second approach, which has been the commonly utilized
approach, engineering demand parameters (EDPs), such as drifts or accelerations, obtained as a result of the
analyses can be used to determine collapse by comparing these EDPs with the limits provided by available gravity
loss models (e.g. Elwood and Moehle, 2005).
Implicit modeling of gravity load failure (non-simulated failure) can only be feasible in some cases where the
first element failure, for example a column axial failure, is sufficient to define global collapse. Such a case may
occur when all the columns at a story have similar properties and failure of all columns is likely to take place almost
simultaneously. In this case, there is no need to explicitly consider the consequences of an axially-failed column and
first column axial failure can be sufficient to define global collapse. In all other cases, explicit modeling of gravity
load failure is essential for accurate determination of collapse. This distinction can be further supported by
considering one of the collapse indicators discussed in the NIST report for collapse assessment and mitigation
strategies for existing RC buildings (2010). The considered collapse indicator is the “maximum fraction of columns
at a story experiencing axial failures,” which requires the identification of the number of columns experiencing
gravity load failures. Such identification would potentially be inaccurate if executed by post-processing of results
without an explicit consideration of gravity load failure, because the gravity load failure of a column (or a beam-
column joint) is likely to affect the response of the other columns and the overall system.
10.3
@Seismicisolation
@Seismicisolation
K. M. Mosalam and S. Günay
Explicit modeling of gravity load failure consists of two stages. The first stage is the detection of gravity load
failure. Available models in literature for columns and beam-column joints that can be used for this purpose are
presented in the next section. The second stage is the post-failure modeling, possible options of which are: (1)
Element removal, (2) Assigning low stiffness to a collapsed element, and (3) Representing the post-failure response
with degradation. The first approach consists of the direct removal of the element from the structural model upon its
failure (e.g. Talaat and Mosalam, 2007, 2009). The second approach consists of reducing the stiffness of the
collapsed element using a small multiplier (e.g. Grierson et al., 2005) in order to eliminate its contribution to the
global structural stiffness matrix. In the third approach, the post-failure response is represented with a degraded
force-displacement relationship (e.g. Elwood and Moehle, 2005). Advantages and disadvantages of these three
approaches are summarized in Table 1 along with the implicit gravity load failure approach mentioned above.
Table 1: Advantages and disadvantages of different gravity load failure modeling methods
Component
Advantages Disadvantages
Failure Method
Explicit Modeling 1. Numerical problems associated with ill- 1. Requirement of additional book-keeping
Option 1: conditioned stiffness matrices are operations to update the nodal masses and to
eliminated. check nodal forces, constraints, restraints,
Element removal
2. Enforcing dynamic equilibrium dangling nodes, floating elements, etc.
enables:
a) Computation of the resulting 2. An additional computational burden
increase in nodal accelerations introduced by the redefinition of degrees of
b) Inclusion of the system’s complete freedom of a structural model and the
kinematic state at time of element corresponding connectivity upon removal of
collapse to determine if it can an element or several elements.
survive to a new equilibrium state.
3. Motion of the collapsed element can be 3. Convergence problems, not on the element or
tracked relative to the damaged system material levels, but on the numerical
to estimate the time and kinetics of a integration level, as a result of the sudden
subsequent collision with the intact updating of mass, stiffness and damping
structural part. matrices, as well as the local vibrations
4. Elimination of the numerical triggered as a consequence of the resulting
convergence problems related to the transient effect (refer to discussions on the
iterative formulation of some element methods to overcome these convergence
and material types by removing them problems).
(refer to item 1 in the disadvantages of
degraded post-failure response).
Explicit Modeling Additional tasks (Items 1 and 2 in 1. Numerical problems associated with ill-
Option 2: disadvantages of element removal) related conditioned stiffness matrices.
to the element removal process are 2. Not possible to explicitly consider the
Assigning low
avoided. consequences of component failure (Items 2
stiffness to a
and 3 in advantages of element removal).
failed element
Explicit Modeling Additional tasks (Items 1 and 2 in 1. Numerical convergence problems related to
Option 3: disadvantages of element removal) related the iterative formulation of some types of
to the element removal process are elements, e.g. force-based beam-column, and
Degraded post-
avoided. materials, e.g. Bouc-Wen, since the failure
failure response
state generally occurs at a negatively sloped
portion of the constitutive relationship.
2. Not possible to explicitly consider the
consequences of component failure (Items 2
and 3 in advantages of element removal).
Implicit Modeling Suitable for fast and simplified analyses Inaccurate results due to the lack of realistic
(Non-simulated and in some special cases, e.g. having representation of the post-failure response.
failure) similar columns in one story.
10.4
@Seismicisolation
@Seismicisolation
Towards an Accurate Determination of Collapse
Vulnerable Reinforced Concrete Buildings
The element removal approach of Talaat and Mosalam (2007) is based on dynamic equilibrium and the
resulting transient change in system kinematics. It constitutes the basis of a corresponding progressive collapse
algorithm. This algorithm is implemented in OpenSees for automatic removal of collapsed elements during an
ongoing simulation, Figure 2. The implementation is carried out as a new OpenSees module, designed to be called
by the main analysis module after each converged integration time step to check each element for possible violation
of its respective removal criteria, where the relevant models presented in the next section can be used for defining
the removal criteria. A violation of a pre-defined removal criterion triggers the activation of the algorithm on the
violating element before returning to the main analysis module. Activation of the element removal algorithm
includes updating nodal masses, checking if the removal of the collapsed element results in leaving behind dangling
nodes or floating elements, which must be removed as well as removing all associated element and nodal forces,
imposed displacements, and constraints. It is noted that the gravity loads at the node of a column, which is common
with the other elements, is not removed. Accordingly, the gravity loads on the structure are not reduced upon
removal of a column, allowing for the analysis model to capture the redistribution of the gravity loads to the other
intact columns.
Update
Start from main code Remove dangling nodes structural
End,
Remove floating elements model, back to
Check for dangling nodes, Delete element/node loads time step
Remove element and main
floating elements, and
Update nodal masses solution code
element loads and masses
parameters
Since the structural elements lose their ability to support gravity loads after gravity load failure, the removal of
a failed element is the most representative approach to model gravity load failure. Hence, the discussion in the
following two paragraphs are based on the comparison of the element removal approach (Option 1) with Options 2
and 3. It should be noted that Option 1 approach assumes that the gravity load support is lost instantaneously. As
mentioned previously, the first stage of gravity load failure in explicit modeling is the detection of this failure. Such
detection is based on equations derived from tests where the loss of the gravity load support of the test specimen was
defined by a single point and there is no data obtained from the tests beyond this point. Accordingly, the assumption
of instantaneous gravity load failure is dictated by the gravity failure detection models. As an alternative, the
detection equations can be constructed with a probability distribution, e.g. in the form of a set of equations for the
median and median plus/minus a dispersion. However, such equations require further experimental research.
The removal of a collapsed element requires several book-keeping operations to update the nodal masses and to
check nodal forces, constraints, restraints, dangling nodes, floating elements, etc. Also, there is an additional
computational burden introduced by the redefinition of degrees of freedom of a structural model and the
corresponding connectivity upon removal of one or more elements. Such additional tasks are avoided in Option 2,
which consists of assigning low stiffness to failed elements. However, there are three important advantages of
Option 1 compared to Option 2. First, it avoids numerical problems due to ill-conditioned stiffness matrices. Second,
enforcing the dynamic equilibrium enables: (1) the computation of the resulting increase in nodal accelerations and
(2) the inclusion of the system’s complete kinematic state at time of element collapse to determine if the structure
can successfully redistribute the forces from the removed element and survive to a new equilibrium state. Third, the
motion of the collapsed element can be tracked relative to the damaged system to estimate the time and kinetics of a
subsequent collision with the intact structural part.
Although representing the post-failure response with a degraded force-displacement relationship in Option 3 is
realistic for most of the failed components, it may introduce numerical problems. The failure state generally
corresponds to a negatively sloped portion of the constitutive relationship, where the iterative formulation of some
types of elements, e.g. force-based beam-column, and materials, e.g. Bouc-Wen, are likely to experience
convergence problems. The removal of such elements automatically eliminates the associated numerical problems.
Analyses conducted to estimate the responses obtained from shaking table tests of a non-ductile RC frame showed
that the analyses considering and not considering the element removal (ER) were both successful in predicting the
observed collapse of the non-ductile members (Mosalam et al., 2009). On the other hand, the response after collapse
was rather jagged and close to being unstable for the case without element removal whereas the analysis with
10.5
@Seismicisolation
@Seismicisolation
K. M. Mosalam and S. Günay
element removal provided a more reasonable response, Figure 3. It should be noted that the analyses without the
element removal used the degraded post-failure approach (Option 3).
Slip spring
C4 C3 C2 C1
Shear and
axial springs
50 50
40 Column C1 40 Column C2
30 30
20 20
10 10
0 0
-10 -10
-20 -20 Test
-30 -30 Analysis (ER: Option1)
-40 -40
Analysis ( w/o ER: Option 3)
-50 -50
-0.1 -0.05 0 0.05 0.1 -0.1 -0.05 0 0.05 0.1
Drift Ratio Drift Ratio
Figure 3: Response of a system with non-ductile columns experiencing axial failure from shaking table tests
and analyses with and without element removal (Mosalam et al., 2009)
Element removal may introduce convergence problems, not on the element or material levels, but on the
numerical integration level, as a result of the sudden updating of mass, stiffness and damping matrices, and the
triggered local vibrations as a consequence of the resulting transient effect due to the sudden changes in the
matrices. It is reminded that the external forces due to gravity loading and the ground motion excitation remain
unchanged when the matrices are suddenly updated. A possible solution to such convergence problems is adaptive
switching of solver type and convergence criteria and reduction of the integration time step (Talaat and Mosalam,
2007). It is to be noted that this strategy has been used for the analyses of the non-ductile RC frame mentioned in the
above paragraph. Another effective solution is the use of transient integrators which do not require iterations, e.g.
operator-splitting methods (Hughes et al., 1979). Analyses conducted on bridge systems (Mosalam et al., 2013)
showed that the operator splitting method results in exactly the same solution as the commonly used implicit
Newmark integration even for cases with highly nonlinear response.
It is noted that the objective of this paper is not the recommendation of an explicit modeling option; rather it is
to present the available options of collapse determination to the interested reader. Furthermore, the most suitable
option may change from one building to another. Hence, the presented advantages and disadvantages are expected to
provide guidance to the readers in choosing the best starting point and possibly switching between different options.
As mentioned previously, the first stage of the explicit modeling of gravity load failure is the detection of gravity
load failure. Models that can be used for the detection of gravity load failure of columns, beam-column joints, slab-
column joints and infill walls are described in the following sub-sections. Collapse of buildings with shear-walls,
due to loss of gravity load carrying capacity, has rarely been observed in the last 50 years (Wallace et al., 2008).
Therefore, detection of gravity load failure of shear-walls is not covered in this paper.
10.6
@Seismicisolation
@Seismicisolation
Towards an Accurate Determination of Collapse
Vulnerable Reinforced Concrete Buildings
Columns
One of the models that can be used to detect the gravity load failure of columns is proposed by Elwood and Moehle
(2005), where the drift at axial failure of a shear-damaged column is represented as follows:
1 tan
2
4 (1)
L axial 100 tan P s Ast f yt d c tan
where (Δ/L)axial is the drift ratio at axial failure, P is the column axial force, Ast, fyt, and s are respectively area, yield
strength, and spacing of the transverse reinforcement, dc is the column core depth (center to center of tie) and θ is the
critical crack angle from the horizontal (assumed 65°).
Elwood and Moehle (2005) stated that this axial failure model is based on data from 12 columns, where all
columns were constructed from normal strength concrete, had the same height-to-width ratio, were designed to yield
the longitudinal reinforcement prior to shear failure, and were tested in uniaxial bending. Despite these limitations,
Fardipour et al. (2011) mentioned that the drifts estimated with a modified version of Equation (1) were in
reasonable agreement with the test results of four cantilever columns with axial load ratios of 20% to 40%, a
nominal transverse reinforcement ratio of 0. 07% and vertical reinforcement ratio of 0.5% to 1%. The modifications
of Fardipour et al. (2011) to Equation (1) consisted of: (1) change of the crack angle to 55+35P/Po for P/Po < 0.25
and 59 for P/Po > 0.25 where Po is the axial force capacity of the undamaged column and (2) addition of the yield
drift to the right-hand side of Equation (1) to calculate the drift corresponding to axial failure. However, it is to be
noted that the above change of the crack angle is inappropriate due to the way Equation (1) was developed.
This axial capacity model is implemented in OpenSees as a limit state material model and used as a spring
connected to a column end. Removal of the corresponding spring is also implemented in an earlier version of
OpenSees. It is to be noted that the authors can provide this version to an interested reader, until the code is
rearranged to comply with the standard version of OpenSees. When the drift during a simulation reaches the drift
corresponding to axial failure, the corresponding spring and the column, to which the spring is connected, are
removed using the element removal algorithm.
Beam-Column Joints
Beam column joints of old non-ductile RC buildings, e.g. designed in the sixties, are generally unreinforced without
any transverse steel bars. The beams connected to such joints rotate relative to the columns, i.e. right angle between
the beam and column is not maintained, due to joint shear failure and corresponding deformation. Joint panel
flexibility can be modeled by using a rotational spring located between the beam and column end nodes. It is noted
that rigid end offsets are used at the beam and column ends to consider the joint physical dimensions (Figure 4d).
The rotational spring is defined by a nonlinear constitutive relationship, which is characterized by a backbone curve
and a set of hysteresis rules (Park and Mosalam, 2013a). These characteristics are recently developed empirically
based on the measured joint responses and visual observations from tests of four corner beam-column-slab joint
specimens (Park and Mosalam, 2013b) and verified by comparison with other exterior and interior beam-column
joint tests. A strength model is developed to determine the peak force of the backbone curve, which also corresponds
to the joint shear strength (Park and Mosalam, 2009, 2012a). The practical strength model in Figure 4a, which
accounts for the effects of two main parameters, namely (1) the joint aspect ratio, defined as the ratio of beam to
column cross-sectional heights and (2) the beam reinforcement ratio, is verified through its accurate predictions of
various beam-column joint test results available in the literature (Park and Mosalam, 2012b, 2012c).
The rotational spring and the constitutive relationship mentioned above can be used to represent the axial failure
and the corresponding removal of a beam-column joint using the proposed extension by Hassan (2011). Hassan
proposed an axial capacity model for beam-column joints where the drift, i.e. beam tip displacement normalized by
the beam length, refer to the test setup in (Park and Mosalam, 2012), at axial collapse is represented as a function of
the axial force and the beam bottom reinforcement strength (Figure 5). Equation (2) is suitable to be used to remove
a beam-column joint as a part of the progressive collapse algorithm when the drift in this equation is replaced by the
joint rotation of the considered analytical model, Figure 4c. Note that the difference between the joint rotation and
the above-mentioned drift in the tests of Hassan (2011) is the sum of the flexural deformation of the beam and the
beam displacement due to column rotation (Figure 6). However, it should be noted that this expression is based on a
rather small database of joint axial failures. Therefore, more joint axial failure tests are needed to further verify this
10.7
@Seismicisolation
@Seismicisolation
K. M. Mosalam and S. Günay
relationship. Hassan (2011) also mentioned that the case of high axial load on a joint where the beam flexural
capacity is much smaller than the direct joint failure capacity is excluded from the application of this model.
Asb f yb
axial 0.057 (2)
P tan
where θaxial is the joint rotation at the joint axial failure, P is the axial force, Asb and fyb are the respective area and
yield strength of the beam bottom reinforcement, and θ is the crack angle.
Figure 4: Overview of unreinforced beam-column joint model (Park and Mosalam, 2012c)
The removal of a beam-column joint is not implemented in OpenSees yet. However, the idea is similar to the case of
column failure, i.e. when the rotation of the spring (representing the joint) during a simulation reaches the rotation
corresponding to axial failure defined by Equation (2), the joint (rotational spring and rigid end offsets) is removed
using the element removal algorithm. Because the generic removal algorithm is already implemented, including the
removal of beam-column joints using the above mentioned criteria is rather straightforward.
Slab-Column Joints
Gravity support loss of slab-column joints can be defined with the punching shear failure. In order to detect the
punching shear failure, available limit models in literature such as Hueste and Wight (1999) or Elwood et al. (2007)
can be used (Figure 7). In these models, drift or plastic rotation values corresponding to the punching shear failure
are determined as functions of the gravity shear ratio defined as the value of the vertical gravity shear divided by the
punching shear strength of the joint. Kang et al. (2009) used these limit models for the detection of punching shear
failure, while conducting analytical simulations to predict the results of shaking table tests. In that regard, they used
option 3, representing the post-failure response with degradation, for post-failure modeling.
10.8
@Seismicisolation
@Seismicisolation
Towards an Accurate Determination of Collapse
Vulnerable Reinforced Concrete Buildings
Figure 5: Axial capacity model in Equation (2) for beam-column joints (Hassan, 2011)
: Drift at axial collapse
L axial
θc j ,axial : Joint rotation at axial collapse
θj
c ,axial : Column rotation at axial collapse
bf
: Beam flexural drift at axial collapse
bf L axial
Figure 6: Contribution of different components to beam tip displacement in a beam-column joint test
10.9
@Seismicisolation
@Seismicisolation
K. M. Mosalam and S. Günay
When the seismic vulnerabilities present in the RC system are combined with the complexity of the interaction
between the infill walls and the surrounding frame and the brittleness of the unreinforced masonry (URM) materials,
the URM infill walls can increase the vulnerability and collapse potential of non-ductile RC buildings. Earthquakes
in the last two decades, e.g., 1999 Kocaeli, 2008 Wenchuan and 2009 L’Aquila earthquakes, led to several
observations related to URM infill walls (Mosalam and Günay, 2012), which are listed as follows:
1. URM infill walls contribute to the stiffness and strength of the frames as evidenced from the weak/soft
story damage of the open ground story buildings and the torsional response created by the non-uniform
distribution of infill walls around the building perimeter.
2. URM infill wall failure is a combination of in-plane (IP) and out-of-plane (OOP) effects as evidenced from
some of the URM infill wall failures taking place at the upper stories instead of the lower stories where the
story shear forces are the highest.
3. Failure of infill walls at a story lead to the formation of weak/soft stories during the earthquake, which may
result in the failure of a story as evidenced by intermediate story collapses.
4. Infill walls interact with the frame members as evidenced by shear cracks and failures of columns and
beam-column joints in infilled bays.
Accordingly, URM infill walls should be modeled to consider these observations. The first two observations can be
reflected by employing a practical model that considers IP-OOP interaction of infill walls, Figure 8, (Kadysiewski
and Mosalam, 2008). The fourth observation can be taken into account by modeling nonlinear shear springs at the
column ends to consider the effect of additional horizontal forces transferred from the infill walls to the columns.
The third observation can be considered by removal of failed infill walls where detection of failure is based on a
combination of IP and OOP displacements (OpenSees Wiki, 2011, Mosalam and Günay, 2013).
Modeling of infill walls considering the element removal due to IP/OOP interaction has been recently used for
the investigation of the earthquake response of buildings designed according to modern seismic codes without
considering the infill walls in the design process (Mosalam et al., 2013). It is noted that, different from the other
discussed elements, failure of infill walls is not directly considered as gravity load failure. It is rather considered to
negatively affect the lateral response. However, detection and post-failure modeling of infill walls are still important
modeling aspects for the objective of the accurate determination of collapse, because the consequences are likely to
affect the gravity load failure of other discussed elements.
IP displacement
Displacement history
Figure 8: Failure detection for URM infill walls considering IP-OOP interaction
Modeling of the gravity system as a single column that accounts for P-Δ effects (Figure 9) is a commonly utilized
approach (Liel et al., 2009, Lai and Mahin, 2013). Such modeling does not generally account for the strength and
stiffness of the gravity system, either because collapse is not of interest or the investigated collapse mechanism is
side-sway. For the collapse type investigated herein, namely the gravity load failure, explicit modeling of the gravity
system is essential. As shown in Figure 1, gravity support may not be completely lost after the lateral and axial
10.10
@Seismicisolation
@Seismicisolation
Towards an Accurate Determination of Collapse
Vulnerable Reinforced Concrete Buildings
failures of the primary system. Therefore, explicit modeling of the gravity system generally leads to a more accurate
and realistic determination of the global collapse.
Leaning
column
Non-ductile RC buildings generally involve significant number of uncertainty sources. Some of these sources
are concrete strength, longitudinal and transverse reinforcement yield strength, concrete and masonry modulus of
elasticity, masonry compressive and shear strengths, damping ratio, story mass, and ground motion record-to-record
variability. Considering all these uncertainty sources in the process of accurate determination of the most collapse-
vulnerable buildings is likely to result in an extensive number of collapse simulations. However, uncertainties of
some of these parameters may have insignificant effects on the variability of the structural response. Tornado
diagram analysis is a practical method used to identify and rank the effect of parameter uncertainties on the response
variability (Lee and Mosalam, 2006). Considering that time is of essence to rapidly determine the most collapse
vulnerable buildings, so that they can be retrofitted or demolished (if needed) before the next big earthquake,
Tornado diagram analysis comes forward as a suitable method to eliminate the burden of unnecessary simulations
by treating the parameters with insignificant effect as deterministic.
The tornado diagram, commonly used in decision analysis, has been used in sensitivity analysis in earthquake
engineering (Porter et al., 2002). The diagram consists of a set of horizontal bars, referred to as swings, one for each
random variable, i.e. considered parameter. The length of each swing represents the variation in the output, i.e. EDP,
due to the variation in the respective random variable. Thus, a variable with larger effect on the EDP has larger
swing than those with lesser effect. In a tornado diagram, swings are displayed in a descending order from top to
bottom. This wide-to-narrow arrangement of swings resembles a tornado. In order to determine the swing due to a
considered parameter, two extreme values, e.g. 10th and 90th percentiles, corresponding to pre-defined upper and
lower bounds of the assumed probability distribution for the parameter are selected. Considered EDP is determined
as a result of nonlinear response history analysis using the upper and lower bound values of the considered
parameter while the other input random variables are set to their best estimates such as the medians. This process
yields two bounding values of the EDP variation for each input parameter. The absolute difference of these two
values is the swing of the EDP corresponding to the selected input parameter. This process is repeated for all the
input parameters to compute the swings of the EDP, Figure 10. Finally, one builds the tornado diagram by arranging
the obtained swings in a descending order as mentioned above. The resulting Tornado diagram generally provides an
indicative picture for the selection of the necessary random variables to be used in the collapse simulations to
develop the fragility curves (Lee and Mosalam, 2005).
10.11
@Seismicisolation
@Seismicisolation
K. M. Mosalam and S. Günay
Figure 10: Construction of the Tornado diagram (Lee and Mosalam, 2006)
CONCLUDING REMARKS
Non-ductile RC buildings are one of the main seismic safety concerns worldwide. In order to use the limited
monetary resources in an effective manner, these buildings should accurately and reliably be ranked to identify those
that are most vulnerable to collapse. This paper attempted to provide a contribution to the accurate determination of
the most collapse vulnerable non-ductile buildings by discussing the methods from existing literature and exploring
the research needs related to gravity load failure modeling and consideration of uncertainty in an efficient manner.
These topics are presented to facilitate the accurate and efficient application of refined seismic assessment.
Concluding remarks of the above discussions are as follows:
Non-ductile RC buildings mostly collapse by losing gravity load carrying capacity, much before reaching
the lateral displacements that would be experienced in a side-sway collapse mechanism.
Implicit (non-simulated) modeling of gravity load failure is only adequate in the case where determining
the first element failure is sufficient to define global collapse. In all other cases, explicit modeling should
be utilized for accurate prediction of global collapse.
There are various advantages and disadvantages of the methods that can be used for post-failure modeling.
Furthermore, the most suitable option may change from one building to another. Advantages and
disadvantages presented in this paper are expected to provide guidance to the readers for choosing the best
starting point and switching between different options.
There are various methods in literature that can be used for failure detection of the RC elements of a
structure. However, there is still a need for enhancement of the failure detection of some of the elements,
for example the non-ductile RC beam-column joints.
There is an existing element removal algorithm implemented in OpenSees. This algorithm is used in the
standard version of OpenSees for the removal of URM infill walls. Removal of columns and beam-column
joints are planned to be included in the standard version of OpenSees in near future.
Gravity support may not be completely lost after the lateral and axial failures of the primary lateral load
resisting system. Therefore, explicit modeling of the gravity system generally leads to a more accurate and
realistic determination of the global collapse.
Non-ductile RC buildings generally involve significant amount of uncertainties. The presented Tornado
Diagram Analysis can be used as a method to handle uncertainty in an efficient and practical manner.
REFERENCES
Elwood K.J., Matamoros A., Wallace J.W., Lehman D.E., Heintz J.A., Mitchell A., Moore M.A., Valley M.T.,
Lowes L.N., Comartin C., Moehle J.P., 2007, “Update of ASCE/SEI 41 Concrete Provisions,” Earthquake
Spectra, 23(3):493-523.
Elwood K.J., Moehle J.P., 2005, “Axial Capacity Model for Shear-Damaged Columns,” ACI Structural Journal,
102: 578-587.
10.12
@Seismicisolation
@Seismicisolation
Towards an Accurate Determination of Collapse
Vulnerable Reinforced Concrete Buildings
Fardipour M., Lam N., Gad E., Wilson J., 2011, “Collapse Limit State Assessment of Lightly Reinforced Concrete
Columns,” Australian Earthquake Engineering Society 2011 Conference, 18-20 November, Barossa Valley,
South Australia.
Grierson D.E., Xu L., Liu Y., 2005, “Progressive-Failure Analysis of Buildings Subjected to Abnormal Loading,”
Computer-Aided Civil and Infrastructure Engineering 20(3):155-171.
Hassan W.M., 2011, “Analytical and Experimental Assessment of Seismic Vulnerability of Beam-Column Joints
without Transverse Reinforcement in Concrete Buildings,” PhD Dissertation, University of California,
Berkeley, CA.
Holmes W.T., 2000, “Risk Assessment and Retrofit of Existing Buildings”, 12th World Conference on Earthquake
Engineering, Auckland, New Zealand, Paper No: 2826
Hueste, M.B.D., Wight J.K., 1999, “Nonlinear Punching Shear Failure Model for Interior Slab-Column
Connections,” Journal of Structural Engineering, 125(9):997-1007.
Hughes T.J.R., Pister K.S., Taylor R.L., 1979, “Implicit-Explicit Finite Elements in Nonlinear Transient Analysis,”
Computer Methods in Applied Mechanics and Engineering 17/18:159-182.
Kadysiewski S., Mosalam K.M., 2008, “Modeling of Unreinforced Masonry Infill Walls Considering In-Plane and
Out-of-Plane Interaction,” PEER Report 2008/102, Berkeley, CA, 144 pp.
Kang TH.-K., Wallace J.W., Elwood K.J., 2009, “Nonlinear Modeling of Flat Plate Systems,” ASCE Journal of
Structural Engineering, 135(2): 147–158.
Lai J.W., Mahin S.A., 2013, “Experimental and Analytical Studies on the Seismic Behavior of Conventional and
Hybrid Braced Frames,” PEER Report 2013/20, Berkeley, CA, 633 pp.
Lee, T.-H. and K.M. Mosalam, 2005, "Seismic Demand Sensitivity of Reinforced Concrete Shear-Wall Building
Using FOSM Method," Earthquake Engineering and Structural Dynamics, 34(14): 1719-1736.
Lee T.-H., Mosalam K.M., 2006, “Probabilistic Seismic Evaluation of Reinforced Concrete Structural Components
and Systems,” PEER Report 2006/04, Berkeley, CA, 181 pp.
Liel A., Haselton C., Deierlein G., Baker J., 2009, “Incorporating Modeling Uncertainties in the Assessment of
Seismic Collapse Risk of Buildings,” Structural Safety, 31(2):197–211.
Mosalam K.M., Park S., Günay M.S., 2009, “Evaluation of an Element Removal Algorithm For Shear Critical
Reinforced Concrete Frames,” COMPDYN, Thematic Conf. on Computational Methods in Structural Dynamics
and Earthquake Engineering, June 22-24, Greece.
Mosalam, K.M. and Günay M.S., 2012, “Behavior and Modeling of Reinforced Concrete Frames with Unreinforced
Masonry Infill Walls”, in Structural Engineering and Geomechanics, [Ed. S.K. Kunnath], in Encyclopedia of
Life Support Systems (EOLSS), Developed under the Auspices of the UNESCO, Eolss Publishers, Oxford ,UK,
[http://www.eolss.net] [Retrieved February 19, 2013]
Mosalam K.M., Günay M.S., Sweat H.D., 2013, “Simulation of Reinforced Concrete Frames with Unreinforced
Masonry Infill Walls with Emphasis on Critical Modelling Aspects,” 12th Canadian Masonry Symposium, June
2-5, Vancouver.
Mosalam, K.M. and Günay S., 2013, “Progressive Collapse Analysis of RC Frames with URM Infill Walls
Considering In-Plane/Out-of-Plane Interaction,” Earthquake Spectra, In press.
Mosalam, K.M., Liang X., Günay S., and Schellenberg A., 2013, “Alternative Integrators and Parallel Computing
for Efficient Nonlinear Response History Analyses,” COMPDYN 2013, 4th ECCOMAS Thematic Conference
on Computational Methods in Structural Dynamics and Earthquake Engineering, M. Papadrakakis, V.
Papadopoulos, V. Plevris (eds.), June 12-14, Kos Island, Greece.
National Institute of Standards and Technology (NIST), 2010, “Program Plan for the Development of Collapse
Assessment and Mitigation Strategies for Existing Reinforced Concrete Buildings,” Report No. NIST GCR 10-
917-7, NEHRP.
OpenSees Wiki, 2011, http://opensees.berkeley.edu/wiki/index.php/Infill_Wall_Model_and_Element_Removal
OpenSees Wiki, 2012, http://opensees.berkeley.edu/wiki/index.php/Rigid_Diaphragm_Consequences
Park, S. and K.M. Mosalam, 2009, “Shear Strength Models of Exterior Beam-Column Joints without Transverse
Reinforcement,” PEER Report 2009/106, Berkeley, CA, 101 pp.
10.13
@Seismicisolation
@Seismicisolation
K. M. Mosalam and S. Günay
Park, S. and K.M. Mosalam, 2012a, “Parameters for Shear Strength Prediction of Exterior Beam-Column Joints
without Transverse Reinforcement,” Engineering Structures, 36(3):198-209.
Park, S. and K.M. Mosalam, 2012b, “Analytical Model for Predicting the Shear Strength of Unreinforced Exterior
Beam-Column Joints,” ACI Structural Journal, 2012, 102(2): 149-160.
Park S., Mosalam K.M., 2012c, “Experimental and Analytical Studies on Reinforced Concrete Buildings with
Seismically Vulnerable Beam-Column Joints”, PEER Report 2012/03, Berkeley, CA, 224 pp.
Park, S. and K.M. Mosalam, 2013a, “Simulation of Reinforced Concrete Frames with Non-Ductile Beam-Column
Joints,” Earthquake Spectra, 29(1): 233-257.
Park, S. and K.M. Mosalam, 2013b, “Experimental Investigation of Non-Ductile Reinforced Concrete Corner Beam-
Column Joints with Floor Slabs,” ASCE, Journal of Structural Engineering, 139(1): 1-14.
Porter, K.A., Beck, J.L., and Shaikhutdinov, R.V., 2002, “Sensitivity of Building Loss Estimates to Major Uncertain
Variables,” Earthquake Spectra, 18(4): 719-743.
Talaat M., Mosalam K.M., 2007, “Computational Modeling of Progressive Collapse in Reinforced Concrete Frame
Structures,” PEER Report 2007/10, Berkeley, CA, 310 pp.
Talaat, M. and K.M. Mosalam, 2009, “Modeling Progressive Collapse in Reinforced Concrete Buildings Using
Direct Element Removal,” Earthquake Engineering and Structural Dynamics, 38(5): 609-634.
Wallace, J.W., Elwood, K.J., and Massone, L.M., 2008, “An Axial Load Capacity Model for Shear Critical RC Wall
Piers,” Journal of Structural Engineering, 134(9): 1548-1557.
10.14
@Seismicisolation
@Seismicisolation
SP-297—11
IDENTIFYING BUILDINGS WITH HIGH SEISMIC RISK UNDER URBAN
RENEWAL LAW IN TURKEY
SYNOPSIS:
A new law known as the "Urban Renewal Law" for risk mitigation was passed in May 2012 with the
objective of reducing seismic risk associated with the existing building stock in Turkey. As stated in the
law, new provisions are set forth to assess and to identify seismically vulnerable residential buildings as
quickly as possible. The buildings that are classified as high risk are either demolished or strengthened.
New buildings are constructed through the financing options provided by the government. In this study,
first, the technical provisions of seismic risk assessment, based on linear elastic analysis, are briefly
described with special emphasis on the deformation limits. Because of the inability of the linear elastic
analysis to allow for redistribution, some flexibility is provided on how many vertical load bearing
elements are allowed to exceed their performance limits. Afterwards, three case study buildings are
analyzed by using the new provisions and ASCE/SEI 41-06 linear elastic procedure. Level and sources of
conservatism in the two approaches are critically evaluated.
11.1
@Seismicisolation
@Seismicisolation
B. Binici et al.
Biography:
Baris Binici is a Professor at Middle East Technical University, Department of Civil Engineering, Turkey. He
obtained his B.Sc. degree at Middle East Technical University and M.S.E. and Ph.D. degrees at The University
of Texas at Austin. His areas of research are reinforced concrete mechanics, seismic assessment and retrofit of
concrete structures.
Ahmet Yakut is a Professor at Middle East Technical University, Department of Civil Engineering, Turkey. He
completed his B.Sc. and M.Sc. degrees at Middle East Technical University and Ph.D. degree at The University
of Texas at Austin. His research interests are earthquake engineering, seismic vulnerability and risk assessment.
Sadun Taniser obtained his B.Sc. and M.Sc. degrees in civil engineering at Middle East Technical University.
He is currently a Ph.D candidate at Middle East Technical University. His research field is earthquake
engineering.
Guney Ozcebe is the Dean of Engineering and Architecture Faculty at TED University. He completed his B.Sc.
and M.Sc. degrees at Middle East Technical University and his Ph.D. degree at University of Toronto. He is a
member of ACI Committees 314 and 318-WA. His research interests include repair and upgrade of concrete
structures, high strength concrete, performance-based design, evaluation and assessment of structures.
11.2
@Seismicisolation
@Seismicisolation
Identifying Buildings with High Seismic Risk Under Urban Renewal Law in Turkey
INTRODUCTION
Seismic performance of the Turkish building stock during recent earthquakes revealed that majority of
existing buildings is extremely vulnerable to earthquakes (Kocaeli and Duzce 1999, Bingol 2003, Van 2011).
The poor performance prompted officials and government organizations to take some major actions in the last
decade including the assessment and strengthening of school buildings, hospitals, industrial facilities and some
of the government buildings. Although, the existing seismic code was updated in 2007 to include a part on the
assessment and rehabilitation of existing buildings, most of the vulnerable residential buildings received no
attention and they are at high seismic risk. A new campaign was started with a new urban renewal law passed on
May 16 2012 (MEU 2012) to mainly address the vulnerable residential building stock. According to the law,
local municipality authorities, Ministry of Environment and Urbanization (MEU) or any of the apartment owners
may request the seismic assessment of a building. If a building is found to be seismically vulnerable, occupants
are given 60 days before demolishing the building or strengthening it with a methodology approved by the local
municipality authorities. Within ten days of the technical report claiming that building is not expected to perform
satisfactorily in a future earthquake, seismically vulnerable buildings are banned from rent or sale. The
applicants are provided either 18 months of rent support as a grant or offered reduced interest rates for mortgage
by the government for the process of refinancing of the building as an encouragement.
According to the law a building is classified as high risk or critical if the building is expected to experience
collapse or very heavy damage under the earthquake level. The number of buildings in Turkey to be examined
under this law in the next 10 years is estimated to be in the order of several millions. The time and budget
required for the assessment of these buildings using existing code procedures is overwhelming. For these
reasons, The Ministry of Environment and Urbanization set up a committee to draft a relatively fast and
acceptable procedure for assessment of residential buildings (named as Specifications for Classification of High
Risk Buildings-SCHRB). The procedure is based on linear elastic analysis of the building model that may be
generated from the information collected for the ground floor. This article first summarizes the provisions with
emphasis on the member performance limit states. Then, applications of the new provisions and existing Turkish
code on a set of representative new and existing buildings are presented. The linear assessment procedure of
ASCE/SEI 41-06 was also applied on three of these buildings in order to investigate the differences in
assessment results among these procedures.
The new assessment provisions are intended to be practical, efficient and as accurate as possible. The speed
of the assessment procedure was established by allowing the building data (concrete strength, member
dimensions, detailing etc.) to be collected from the ground floor only unless vertical element discontinuity is
found. A number of methods are available for seismic assessment based on ground floor member dimensions
(Hassan and Sozen 1997, Gulkan and Sozen 1999, Yakut et al. 2006). These practical procedures, although
provide valuable information on the seismic vulnerability of a building stock, are not tailored to estimate the
expected seismic performance of a single building. On the other hand, nonlinear static procedures (FEMA 356,
ASCE/SEI 41-06, ACI 562-13) are not deemed essential by the committee to solely identify the buildings with
many critical deficiencies leading to unacceptable seismic performance. Hence, the committee employed the
linear elastic (equivalent lateral force or mode superposition) procedure with some major differences as
compared to that of the Turkish Earthquake Code (TEC 2007) and ASCE/SEI 41-06. The new provisions are
applicable to buildings having 8 stories or less with total building height less than 25 meters. It is noteworthy to
mention that the new provisions are not intended to be used for seismic retrofit as more detailed provisions are
given in TEC (2007). A brief summary of the seismic assessment procedure detailed in the new provisions is
given below:
1. A three dimensional model of the building is generated based on a detailed survey performed for the
critical floor (generally the ground floor) only. If preferred, a complete survey can be carried out.
Since the assessment is done for only columns and shear walls at the critical floor, material
properties and reinforcement detailing are determined for these members. At least five concrete core
samples shall be taken from the columns and walls to determine the concrete strength. Removing
the cover concrete for several columns and walls is required to determine the reinforcement details.
The complete building model may be obtained by the replication of the ground floor layout over the
building height for regular buildings. However, one must consider variations in height and
11.3
@Seismicisolation
@Seismicisolation
B. Binici et al.
irregularities in plan and elevation as defined in the TEC (2007) including the torsional irregularity,
plan irregularity, and possible discontinuity of vertical load bearing elements. If discontinuity of
vertical load bearing elements exists in a building the engineer must resort to obtaining the building
data from all floor levels.
2. Based on the results of linear elastic analysis (equivalent lateral load or response spectrum) under
the design response spectrum (given in TEC 2007) and using no response modification factor, the
bending moment demand capacity ratios (DCR) at member ends and interstory drift deformations
are determined. Moment capacities for columns are computed based on axial loads obtained from
G+nQ±E/6 load combinations (G: dead load, Q: live load, n: live load reduction factor, E:
earthquake load). In this way, over conservatism regarding expected earthquake induced axial loads
is avoided. As a further simplification in the new provisions, flexural stiffnesses are taken as 30
percent of EIg for beams and walls, and 50 percent of EIg for columns. Given the lack of accurate
modulus of elasticity information for each column in the buildings, dependence of flexural stiffness
on axial load ratio was ignored. Furthermore, sensitivity of results on case study buildings showed
almost no influence in the final decision upon using the proposed simplified rigidities.
3. For each column and wall at the critical floor, DCR and interstory drift deformations are compared
with the corresponding limit values. If the interstory drift ratios in any floor are higher than the ones
obtained in the ground floor then assessment of the members for the interstory drift ratio at that floor
is also carried out. If the column/wall does not satisfy either one of the limits, it is classified as
unacceptable. Depending on the number of unacceptable members, buildings are classified as
“critical” or “not critical.” A critical building represents a building that is expected to suffer heavy
damage or collapse under the design earthquake effect. Because of the inability of the linear elastic
analysis to allow for redistribution, some flexibility is provided on how many columns are allowed
to exceed their performance limits. When the average axial load ratio resulting from gravity loads in
the considered floor exceeds 0.65, none of the members are allowed to exceed their performance
limit to classify the building as "not critical, (NC).” When the average axial load ratio is less than or
equal to 0.1, columns/walls that carry up to 35 percent of the story shear are allowed to exceed their
performance limits in order to classify the building as NC. Linear interpolation is used to determine
the acceptable story shear ratio for intermediate average axial load ratios.
It is worth mentioning that the proposed provisions are applicable to low to mid rise residential buildings, in
which significant member size changes from floor to floor are not expected. For these buildings, possible minor
changes in the member dimensions do not lead to significant differences in the assessment results as evidenced
by the analysis of several typical residential buildings in Turkey during the process of calibration of the
provisions. Similar to ASCE/SEI 41-06 provisions based on linear elastic procedure, DCRs are used to compare
the demand with the performance limits for columns and walls. Additionally, interstory drift limit is used as
deformation based assessment parameter. When these limits are exceeded, the component is very likely to
experience significant strength degradation. Translating the expected damage into quantitative engineering
parameters is a quite challenging task, especially within the context of urban renewal as the outcome may mean
demolishing buildings under the current law in Turkey. The approach of the committee was to set the
performance limits between the life safety definition of ASCE/SEI 41-06 (i.e. 75% of the deformation level at
which 20% loss of lateral maximum strength occurs for a member) and axial collapse. Accurate estimation of
actual collapse of a column or wall member is very challenging, for which a number of recent studies were
conducted (for example Elwood 2003). The limit values proposed in the new provisions are given for various
groups of columns and walls similar to that of ASCE/SEI 41-06. In the new provisions, columns are classified
into three, walls are classified into two groups based on their expected failure modes. The performance limit
values were determined for each group based on the mean of experimental data and analytical simulations when
the expected failure modes are in the presence of low axial load and low shear force. For high axial load and/or
shear cases, limits were selected from the lower bound values. Some adjustments were made for certain
components in order to ensure that buildings that fully comply with the provisions of TEC (2007) are found as
NC.
The following methodology was employed while establishing DCR and interstory drift limits: First, ductility
and ultimate displacements of columns and walls were obtained from experimental results, empirical/mechanical
models or relevant guidelines (Solmaz 2010, Erguner 2009, Bae and Bayrak 2009, ASCE/SEI 41-06 Supplement
11.4
@Seismicisolation
@Seismicisolation
Identifying Buildings with High Seismic Risk Under Urban Renewal Law in Turkey
1 and Eurocode 8 were employed for this purpose). Ultimate deformations were taken as the displacement values
corresponding to 20 percent capacity drop beyond the peak in experimental results or collapse prevention
displacement limit when guidelines are considered. Afterwards displacement ductility was divided by a factor of
1.25 in order to obtain DCR limits, and ultimate inelastic displacement was divided by a factor of 1.4 in order to
obtain corresponding interstory drift ratio limits. In this way, inelastic deformation limits were converted to
elastic deformation limits to be used along with linear elastic analysis results as neither the TEC nor SCHRB
employ displacement coefficient while estimating the target elastic displacement demand.
The columns are classified in SCHRB (similar to ASCE/SEI 41-06) based on the shear demand capacity
ratio and transverse reinforcement detailing of columns. Tables 1 and 2 show the variables used in the
classification of columns in SCHRB and ASCE/SEI 41-06. Three different possible failure modes are considered
in the classification of columns. Class A and type i columns are expected to fail in flexure, class B and type ii
columns are expected to fail in shear-flexure, and class C and type iii columns are expected to fail in shear. The
calculation of shear strength Vr and Vn are similar in SCHRB and ASCE/SEI 41-06 if one uses the ACI 318-11
shear strength equation. Although the approach in general is similar, the definition of Ve (Vp in ASCE/SEI 41-06)
is slightly different. In SCHRB, Ve is calculated as the smaller of shear force computed based on plastic hinges
forming at column ends and based on plastic hinges forming at beam ends. On the other hand Vp is determined as
the shear demand at flexural yielding of plastic hinges in ASCE/SEI 41-06.
Performance limits for class A columns according to SCHRB are given in Table 3. Comparisons of SCHRB
and ASCE/SEI 41-06 DCR limits are presented in Figure 1. SCHRB limits were determined based on studies by
Bayrak and Bae (2009) and Solmaz (2010) and further comparisons of performance limits of typical column
sections using ASCE/SEI 41-06 and Eurocode 8 (2005). It can be observed that for class A columns there is a
general agreement between the recommended values of SCHRB and ASCE/SEI 41-06. Columns with low Ve/Vr
ratios and/or "insufficient" confining steel reinforcement experience failure due to a combined effect of flexure
and shear with limited ductility. A database of 41 columns compiled by Erguner (2009), which is extracted from
a wider database reported by Berry et al. (2004), was employed to study the columns classified as class B in
Table 1. The demand-capacity-ratios and drift deformation limits of the columns from the database are plotted
along with the proposed limits in Figure 2. Recommendations were made as a function of axial load ratio and
transverse steel reinforcement amounts. The shear stress ratio (Ve/bdfctm) was not considered as a separate
parameter due to the wide scatter of the data that shows no clear dependency as shown in Figure 2. For class B
11.5
@Seismicisolation
@Seismicisolation
B. Binici et al.
columns, recommended values of DCR for low and high axial load ratios are based on mean and lower bound
values, respectively. Interstory drift ratio limits are based on lower bound values. It can be observed that the
number of experiments involving high shear force together with high axial force is very limited.
In SCHRB and ASCE/SEI 41-06 approaches, m-factors are used to represent DCR (moment demand
calculated from G+nQ±E loading divided by moment capacity) and to assess performance of the member. IN
SCHRB, drift ratios of the critical floor are also checked in addition to the DCR limits. The major differences
between the performance limits of SCHRB and ASCE/SEI 41-06 can be summarized as follows:
NK (P in ASCE/SEI 41-06) parameter is determined using different load combinations in SCHRB and
ASCE/SEI 41-06. NK is calculated using (G + Q + E / 6) load combination in SCHRB while ASCE/SEI 41-06
leads to consider the axial load as a force-controlled parameter and use (Eq. 3-19) in the calculation of NK which
is (G + Q + E / C1C2J). Use of different NK values leads to consideration of different Vr values (Vn in ASCE/SEI
41-06) in column classification, and use of different section moment capacities in the evaluation of each column.
a) Class A b) Class B
11.6
@Seismicisolation
@Seismicisolation
Identifying Buildings with High Seismic Risk Under Urban Renewal Law in Turkey
Figure 1. Comparisons of SCHRB and ASCE/SEI 41-06 DCR and Interstory Drift Ratio Limits for Class A and
Type i Columns
Figure 2. Comparisons of DCR and Interstory Drift Ratio Limits with Experimental Results for Class B and
Type ii Columns
The reinforced concrete walls are classified into two groups in SCHRB, whereas they are classified into
three groups in ASCE/SEI 41-06. Kazaz et al.(2012) compiled a database for shear walls. The compiled
experimental work indicated that, unlike columns, the number of experiments on walls failing in flexure-shear
and shear mode is rather limited. Hence, it was not deemed necessary to differentiate between these two failure
modes for walls. The classification for walls is done in terms of (Ve/Vr) and (Hw/lw) ratios in SCHRB, whereas
only (Hw/lw) ratio is decisive in classification in ASCE/SEI 41-06. Tables 4 and 5 show the classification of RC
walls. According to SCHRB, type A walls are controlled by flexural behavior; type B walls are controlled by
either flexure-shear or shear behavior.
11.7
@Seismicisolation
@Seismicisolation
B. Binici et al.
Similar to the columns, m-factors and interstory drift ratios are used in the determination of acceptance
criteria of RC walls. If the lateral load carrying capacity of shear walls are above 50 percent of the total base
shear demand and the maximum interstory drift ratio demand is less than 0.75, such buildings can be evaluated
solely based on interstory drift ratio limits. The parameters used for each wall class can be summarized as
follows:
- Walls controlled by flexure: The same parameters control the m-factors in both guidelines. These
parameters are axial load, shear, and confinement in boundary of walls. However, the effect of axial load in m-
factors is slightly different in ASCE/SEI 41-06. Whereas the effect of axial load is represented by NK in SCHRB,
it is formulated by ((As–A’s)fy + P) in ASCE/SEI 41-06. In other words, the effect of longitudinal reinforcement
contribution to the ductility on axial load level of wall is taken into account in ASCE/SEI 41-06.
- Walls controlled by shear: m-factor is a function of normalized shear (Ve / bwdfctm) in SCHRB whereas it is
a function of axial load (((As–A’s)fy + P)/Acf’c) in ASCE/SEI 41-06.
The interstory drift limits were determined for the same parameters as DCR limits. Table 6 presents the
performance limits in SCHRB for columns and walls. The numerical simulation results of Kazaz et al. (2012)
were employed along with the compiled test results for class A walls as shown in Figure 3.a. The selected
performance limits for low axial load levels correspond to 50 percent probability of being exceeded based on the
experimental data points. For high axial loads, DCR limits were selected as lower bounds to the simulation
results as there were no experimental data. However, such an approach was not considered for the interstory drift
ratio limits owing to the superior performance of buildings with significant structural walls under earthquakes.
Instead, they were determined considering the assessment results of the case study buildings. For class B walls,
performance limits were selected for a 30 percent probability of being exceeded and their dependency on axial
force was neglected (Figure 3.b). This can be justified on the ground of having low axial force levels on walls for
practical purposes and lack of experimental data for such structural elements. Additionally, numerical results of
Kazaz et al. (2012) indicated that dependence of the performance limits on axial load ratio for type B walls is not
significant within the practical limits of axial load ratios.
TEC
(2007)
Nk /(fcm×Ag) Ve /(b×d× fc) Compliant DCRlimit (δ/h)limit Ve /(b×d× fcm) DCRlimit (δ/h)limit
Boundary
Region
Yes 6.0 0.030 ≤ 0.9 4.0 0.020
≤ 0.9
No 4.0 0.015 ≥ 1.3 2.0 0.010
< 0.1
Yes 3.5 0.015
≥ 1.3
No 2.0 0.0075
Yes 3.5 0.020
≤ 0.9
No 2.0 0.010
> 0.25
Yes 2.0 0.010
≥ 1.3
No 1.5 0.005
11.8
@Seismicisolation
@Seismicisolation
Identifying Buildings with High Seismic Risk Under Urban Renewal Law in Turkey
a) Class A Walls
b) Class B Walls
Analysis results from three case study buildings are presented in this section. Building 1 is a 6-story
residential RC moment resisting frame building designed in accordance with the provisions of TEC (2007). The
floor plan is designed to represent typical buildings in Turkey. The height of first story is 3.5 m and the height of
other stories is 3 m. The story plan and three dimensional model of the building are presented in Figure 4.
Building 2 is also a 6-story residential building having RC moment frame-wall system designed according to the
code. The building has a commonly used configuration of shear walls in both principal directions. The story
height is 3 m for all story levels. The story plan and three dimensional model of the building are presented in
Figure 5. The third building is building 10 of case study buildings that is a 5-story RC moment frame-wall
system. This building experienced 1999 Düzce earthquake and suffered moderate damage. The story height is
2.8 m for all story levels. The story plan and three dimensional model of the building are presented in Figure 6.
Material and site properties of the buildings are summarized in Table 7.
11.9
@Seismicisolation
@Seismicisolation
B. Binici et al.
11.10
@Seismicisolation
@Seismicisolation
Identifying Buildings with High Seismic Risk Under Urban Renewal Law in Turkey
Linear dynamic procedure (modal analysis) is used for all case studies by employing mode superposition
method with sufficient modes that contribute 90 percent of the participating mass of the building in each
orthogonal horizontal direction. The design response spectrum for the highest seismic zone in Turkey, which is
defined in TEC (2007), is used for Building 1 and Building 2. The assessment of Building 3 is carried out under
the response spectrum of 1999 Düzce earthquake, as this building was moderately damaged under Düzce
earthquake.
Figures 7 to 9 present m/mlimit results of the ground floor from the +X direction analysis of Buildings 1, 2,
and 3, respectively. Table 8 summarizes the number of acceptable and unacceptable column and shear walls of
buildings in terms of m/mlimit results. Analyses for other directions were also conducted, but they are not
presented here for brevity as the conclusions are similar. For Building 1, ASCE/SEI 41-06 produces more
conservative results as compared to SCHRB for all columns. A similar conclusion can be reached upon
examining the results of Building 2. All columns were found as acceptable according to both provisions for this
building. However ASCE/SEI 41-06 estimations were very conservative for the seismic performance of shear
walls. SCHRB provisions classified Building 3 as NC according to the assessment based on interstory drift ratio.
ASCE/SEI 41-06 results, on the hand, estimated insufficient seismic performance for about half of the columns
and all of the shear walls. This result is again on the over safe side from a point of seismic assessment.
Two main reasons of the ASCE/SEI 41-06 in providing overly conservative results for the seismic
assessment are attributed to the following factors:
1- The m-factors provided in SCHRB are more liberal than those given ASCE/SEI 41-06.
2- When calculating the component capacities the associated axial loads are obtained under a load
combination with reduced earthquake force with a reduction factor of 6 in SCHRB. However in the same
combination, ASCE/SEI 41-06 leads to use of larger earthquake force with a reduction factor of (C1C2J)
which is usually found between 2 and 2.5. This situation leads to use of smaller component capacities under
higher axial loads.
3- Owing to the superior performance of building with shear walls, SCHRB provisions allow drift based
assessment when the drift demands are sufficiently small (i.e. less than 0.75%). Such a favorable
consideration is not present in ASCE/SEI 41-06 and one must always employ the m-factors for linear elastic
assessment.
11.11
@Seismicisolation
@Seismicisolation
B. Binici et al.
2.00
ASCE/SEI 41‐06 S: Column
SCHRB
m/mlimit
1.00
0.00 S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
S13
S14
S15
S16
S17
S18
S19
S20
S21
S22
S23
S24
S25
S26
S27
S28
S29
S30
S31
S32
Member Names
Figure 7. Performance Comparison of Members according to ASCE/SEI 41-06 and SCHRB for Building 1
CONCLUSIONS
In this article, the technical provisions and the relevant background for the seismic assessment
under Urban Renewal Law in Turkey were presented. Special emphasis was given to member
ductility and deformation limits for use along with linear elastic assessment procedures. Afterwards
comparisons of assessment results on three case study buildings from the new provisions and
guidelines of ASCE/SEI 41-06 were discussed.
SCHRB provisions estimated that well designed buildings would exhibit better performance
levels than collapse prevention level. Similarly, for a building that experienced moderate damage
due to the presence of significant shear walls, the building was classified as NC meaning that it is
not expected to e collapse or suffer heavy damage. ASCE/SEI 41-06 estimated that for the well
designed buildings, many of the vertical load bearing elements would be in the collapse
performance state. A similar result was found from the examination of the third building. These
analyses results showed that ASCE/SEI 41-06 guidelines are on the over safe side as compared to
SCHRB provisions and observed building performance. When such overly safe building
performance estimations are made, retrofit costs tend to increase so high and discourage people
from employing seismic retrofit solutions. Apart from being conservative that is on the safe side,
another reason which may lead to overly conservative estimations may be the fact that these
guidelines were calibrated based on member deformation limits (from member tests) without
placing sufficient emphasis on structural system behavior and performance. The authors believe
that ASCE-SEI 41-06 approach can further be improved in the upcoming revision cycles from the
output of the research that reveals the uncovered link between the member and system
performance.
11.12
@Seismicisolation
@Seismicisolation
Identifying Buildings with High Seismic Risk Under Urban Renewal Law in Turkey
1.00
ASCE 41‐06
ASCE/SEI 41‐06 S: Column
m/mlimit
RBTY 2012
0.00
S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
S13
S14
S15
S16
S17
S18
S19
S20
S21
S22
S23
S24
S25
S26
S27
S28
Member Names
a) Columns
2.00
P: Shear Wall
m/mlimit
1.00
0.00
P1
P2
P3
P4
P5
P6
Member Names
b) Walls
Figure 8. Performance Comparison of Members according to ASCE/SEI 41-06 and SCHRB for Building 2
2.00
S: Column ASCE 41‐06
ASCE/SEI 41‐06
P: Shear Wall RBTY 2012
m/mlimit
1.00
0.00
S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
S13
S14
S15
S16
S17
S18
S19
S20
S21
S22
S23
S24
S25
S26
S27
S28
S29
S30
S31
S32
S33
S34
S35 Member Names
a) DCR
4.00
Interstory Drift Ratio Demand S: Column
3.50
3.00 Interstory Drift Ratio Demand acc. to SCHRB
P: Shear Wall
2.50
IDR (%)
2.00
1.50
1.00
0.50
0.00
S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
S13
S14
S15
S16
S17
S18
S19
S20
S21
S22
S23
S24
S25
S26
S27
S28
S29
S30
S31
S32
S33
S34
S35
P1
P2
P3
P4
P5
P6
Member Names
b) Interstory Drift Ratio
Figure 9. Performance Comparison of Members according to ASCE/SEI 41-06 and SCHRB for Building 3
11.13
@Seismicisolation
@Seismicisolation
B. Binici et al.
APPENDIX: NOTATION
Ash Area of transverse reinforcement
A g Area of cross section
bk Width of core concrete
b Width of cross section perpendicular to shear force
m, DCR Demand capacity ratio
mlimit, DCRlimit Demand capacity ratio limit
d depth of cross section
d b Diameter of the tension reinforcement
DRy Yield interstory drift ratio
DRu Maximum interstory drift ratio
EIg Flexural rigidity of gross cross section
EIeff Effective flexural rigidity
fcm, fc Concrete compressive strength in MPa
fywm Yield strength of transverse reinforcement in MPa
fy Yield strength of longitudinal reinforcement in MPa
fctm Tensile strength of concrete in MPa
h Column depth
Mu Moment capacity
Nk, N Axial load
N/N0 Axial load ratio
R Strength reduction factor
s Spacing of transverse reinforcement
Ve, V Flexural shear demand
Vr Shear strength
Effective stiffness factor
(/h)sınır Interstory drift ratio limit
w Wall length
µ, µ Displacement ductility
ρs = Ash/bks Transverse reinforcement ratio
REFERENCES
ACI Committee 562, 2013, Code Requirements for Evaluation, Repair, and Rehabilitation of Concrete Buildings
(ACI 562-13) and Commentary, American Concrete Institute, Farmington Hills, MI 48331, U.S.A.
ACI Committee 318, 2011. Building Code Requirements for Structural Concrete (ACI 318-11), American
Concrete Institute, Farmington Hills, MI 48331, U.S.A.
American Society of Civil Engineers, 2000. Prestandard and Commentary for the Seismic Rehabilitation of
Buildings, Report No. FEMA 356. Reston, Virginia.
American Society of Civil Engineers, 2007. Seismic Rehabilitation of Existing Buildings, Report No. ASCE/SEI
41-06, Reston, Virginia, 428.
American Society of Civil Engineers, 2007. Seismic Rehabilitation of Existing Buildings, Report No. ASCE/SEI
41-06-Supplement 1, Reston, Virginia, 428.
Bae S, Bayrak O, 2009. Drift Capacity of Reinforced Concrete Columns, ACI Structural Journal, Vol. 106(4),
405-415.
11.14
@Seismicisolation
@Seismicisolation
Identifying Buildings with High Seismic Risk Under Urban Renewal Law in Turkey
Berry, M. P., M. Parrish, and M. O. Eberhard. 2004. PEER structural performance database user’s manual.
Pacific Earthquake Engineering Research Center. University of California, Berkeley, California.
CEN. Eurocode 8, 2005., Design Provisions for Earthquake Resistance of Structures, European Prestandard
ENV- Part 3: Assessment and Retrofitting of Existing Buildings, Comité Européen de Normalisation,
Brussels.
Elwood, J.K., 2003. Shake table tests and analytical studies on the gravity collapse of reinforced concrete
frames. Ph.D. dissertation, University of California at Berkeley, Berkeley, California.
Ergüner K., 2009. Analytical examination of performance limits for shear critical reinforced concrete columns,
M.S. Thesis, Middle East Technical University, Department of Civil Engineering, Ankara, Turkey.
Gulkan, P., Sozen, A.M., 1999. Procedure for Determining Seismic Vulnerability of Building Structures, ACI
Structural Journal 96(3), 336-342.
Hassan, A.F., Sozen, M.A., 1997. Seismic Vulnerability Assessment of Low-rise Buildings in Regions with
Infrequent Earthquakes, ACI Structural Journal, 94(1), 31-39.
Kazaz, I., Gülkan, P. and Yakut, A. 2012. Deformation limits for structural walls with confined boundaries,
Earthquake Spectra, 28 (3):1019–1046.
Ministry of Public Works and Settlement, 2007. Specification for Buildings to be Built in Seismic Zones (TEC
2007), Government of Republic of Turkey.
Middle East Technical University-EERC., 2011. 23 October 2011 Mw=7.2 Van-Ercis Earthquake
Reconnaissance Report, METU-EERC Report No:2011-04, METU, Ankara.
Ministry of Environment and Urbanization (MEU), 2012. The Urban Renewal Law for Regions under Disaster
Risk, Law No: 6306, Official Gazette: Date, 28(309), Volume 52, Turkey.
Pujol, S., Sözen, M., Ramírez, J., 2000. Transverse Reinforcement for Columns of RC Frames to Resist
Earthquakes, Journal of Structural Engineering, 126(4), 461-466.
Sezen, H., 2002. Seismic behavior and modeling of the reinforced concrete building columns, Ph.D. dissertation,
University of California at Berkeley, Berkeley, California.
Solmaz T., 2010. Evaluation of performance based displacement limits for reinforced concrete columns under
flexure, M.S. Thesis, Middle East Technical University, Department of Civil Engineering, Ankara, Turkey.
Yakut, A., 2004, Preliminary seismic performance assessment procedure for existing RC buildings, Engineering
Structures, 26(10), 1447-1461.
Yakut A., Ozcebe, G., Yucemen M.S., 2006. Seismic Vulnerability Assessment using Regional Empirical Data,
Earthquake Engineering & Structural Dynamics, 35(10), 1187–1202.
Yakut, A., Binici, B., Ozcebe, G., Ay, B.O., Akansel, V.H., 2012. Calibration and Validation of the Provisions
for the Seismic Risk Evaluation of Existing RC Buildings in Turkey under Urban Renewal Law, Report
Submitted to Ministry of Environment and Urbanization.
11.15
@Seismicisolation
@Seismicisolation
B. Binici et al.
11.16
@Seismicisolation
@Seismicisolation