ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title no. 94-S03

Response Modification Factors for Seismic Design of

Circular Reinforced Concrete Bridge Columns

by Ahmad Itani, Peter Gaspersic, and M. “Saiid” Saiidi

The response modification factors for circular reinforced concrete over” analysis) is normally used to calculate the ultimate
single column bridge bents were calculated for a wide range of displacement capacity of reinforced concrete bents based on
periods based on three target values: the capacity to demand the plastic hinge rotation capacity of columns.6,7 Thus, the
displacement, displacement ductility ratio, and longitudinal rein- ratio of the ultimate displacement capacity to the elastic
forcement ratio. The validity of using the ratio of capacity to
displacement demand of bridge columns has become a check
demand displacements as a check in the column design process
and the effect of confinement steel ratio on the response modifica- in the column design process.
tion factors were also investigated. The study represented in this paper investigated the
validity of this check in a wide range of period spectra and
Keywords: bridges (structures); columns (supports); ductility; dynamic
determined its influence on the reinforced concrete column
structural analysis; earthquake resistant structures. design. The response modification factors were calculated
based on target values for the ratio of the ultimate displace-
INTRODUCTION ment capacity to displacement demand (ΔC/ΔD), the
The current practice in seismic design of concrete bridges displacement ductility ratio (μ), and the longitudinal rein-
is to limit the inelastic response to the columns, which are forcement ratio (ρs).
normally designed and detailed to withstand severe cyclic
deformations during strong earthquakes. It is customary to CAPACITY ANALYSIS OF REINFORCED
calculate the design strength of bridge columns by obtaining CONCRETE SINGLE COLUMN BRIDGE BENTS
The seismic design concept for bridge columns aims to
the moments from an elastic dynamic analysis based on a
design the column to a force level less than the elastic
specified response spectrum. To account for inelastic defor-
response and to adequately detail the column for ductility so
mation of the column, the elastic moments are subsequently
it can undergo large inelastic deformation without signifi-
divided by a response modification factor R that is specified
cant strength degradation. Therefore, in addition to the
in design codes.1,2 The force level of the response spectrum
strength, there is an implied level of deformation capacity to
and the value of R control the size and reinforcement of the
insure an adequate seismic performance. Hence, a check of
column. Hence, the R value can play a significant role in
the ultimate deformation capacity seems necessary in bridge
characterizing the bridge dynamic properties.
column design. This check may be carried out by an inelastic
AASHTO bridge specifications1,2 specify constant R values lateral load analysis of the column, also known as the “push-
for the entire range of the fundamental period when designing over” analysis.
bridge columns. The rationale used in the development of the
specified R values was based on consensus, engineering judg-
PUSHOVER ANALYSIS
ment, and the performance of highway bridges in previous The ultimate displacement capacity for a single column
earthquakes. Recent studies3,4 have questioned the use of bent in the transverse direction can be obtained by
constant R values in seismic codes since these values should performing a pushover analysis on a cantilever column fixed
be dependent on the structure’s strength and the fundamental at the base and free at the top where an incremental hori-
period of vibration. This has raised some questions and zontal load is applied. It is assumed that under the influence
concerns about the reliability of the specified constant R of the lateral load the column yields in flexure at the base and
values in the AASHTO design codes. thus the flexural deformation can be obtained by adding the
elastic curvature over column height to the inelastic curva-
RESEARCH SIGNIFICANCE ture concentrated in a plastic hinge at the base. Therefore, for
The recent introduction of the capacity design method in
concrete bridges5 clarified the need for additional checks
beyond obtaining moments from elastic dynamic analysis ACI Structural Journal, V. 94, No. 1, January-February 1997.
Received August 29, 1995, and reviewed under Institute publication policies. Copy-
and dividing them by R value. These additional checks right © 1997, American Concrete Institute. All rights reserved, including the making of copies
address the nonlinear behavior of reinforced concrete unless permission is obtained from the copyright proprietors. Pertinent discussion including
author’s closure, if any, will be published in the November-December 1997 ACI Struc-
sections. Nonlinear analysis under horizontal loads (“push- tural Journal if the discussion is received by July 1, 1997.

ACI Structural Journal/January-February 1997 23

 The longitudinal reinforcement steel behavior that is used
Ahmad Itani is an assistant professor of civil engineering at the University of Nevada,
Reno. He received his MS and PhD in civil engineering from the University of Michi- in the program is described by a monotonic uniaxial stress-
gan, Ann Arbor, in 1991. A registered professional civil engineer, he has designed a strain response. This response is characterized by an elastic
variety of concrete and steel bridges in California.
region, a yield plateau, and a strain-hardening region
Peter Gaspersic is a visiting graduate student at the University of Nevada, Reno. He is followed by a descending branch after peak stress up to bar
a PhD candidate in the Department of Civil Engineering, University of Ljubljana, fracture. The mathematical equations that are used in the
Slovenia. program to define the behavior of different regions in the
ACI member M. “Saiid” Saiidi is a professor of civil engineering at the University of
stress-strain curve of reinforcement steel are those devel-
Nevada, Reno. He is the chairman of ACI Committee 341, Earthquake Resistant Con- oped by Park et al.9 In the current study, Grade 60 steel is
crete Bridges, and a member of joint ACI-ASCE Committees 343, Concrete Bridge assumed for both longitudinal and transverse reinforcement.
Design, and 352, Joints and Connections in Monolithic Concrete Structures.
The reinforcement steel was assumed to have a modulus of
elasticity (E) of 200,000 MPa (29,000 ksi) and a yield stress
fy of 414 MPa (60 ksi), which is the specified nominal stress.
a single column bent responding to transverse lateral Strain-hardening behavior is assumed to start at a strain
loading, the ultimate displacement capacity (ΔC) will be the equal to five times the yield strain, and the ultimate strain is
sum of the elastic and inelastic displacements. assumed to be equal to 0.12.
The lateral elastic displacements for a single column bent The concrete model that is used in the program has two
may be calculated as different regimes depending on the location of the concrete
portion, i.e., the core and the cover. The core portion is
2 defined to be the circle contained by the centerline of the
φy h
δ y = ---------- hoop. The behavior of this concrete is affected by the pres-
3
ence of the transverse reinforcement, since a passive
confinement is provided by this reinforcement. The program
where φy is the yield curvature of the reinforced concrete section utilizes Mander’s model10 to describe the behavior of
and h is the column height. The inelastic displacement, on confined concrete. The cover concrete, on the other hand, is
the other hand, can be derived by assuming the plastic rotation unconfined. A strain of 0.005 is used by the program to
is concentrated at the center of the plastic hinge at the define the spalling concrete strain after which the concrete
column base. The plastic rotation may be obtained, as cover will develop longitudinal cracks and will separate
discussed in Reference 5, by from the core. This study assumed the compressive strength
of concrete f c′ is equal to 27.6 MPa (4000 psi).
θp = (φu – φy)lp
CONFINEMENT STEEL RATIO
and thus the inelastic displacement can be calculated as To maintain the integrity of the concrete core and increase
the rotational capacity of the column, the confinement steel
δp = θp(h – lp /2) equations were adopted from Caltrans Bridge Design Speci-
fications.11 In all the column sections that were investigated
where φu is the ultimate curvature and lp is the length of the in this study, the confinement steel ratio ρ was governed by
plastic hinge that may be estimated by7 the following equation

lp = 0.08h + 9db f′
ρ ≥ 0.12 ----c- 0.5 + 1.25 ⎛ -----------
P ⎞
fy ⎝ f c′ A g⎠
where db is the diameter of the longitudinal bar reinforce-
ment of Grade 60.5 Therefore, the total displacement
capacity is equal to where ρ is the ratio of the volume of circular hoop reinforce-
ment to the volume of the concrete core, Ag is the gross area
Δ c = δy + δp of concrete, fy is the yield strength of confining reinforce-
ment, P is the column axial load, and f c′ is the compressive
COMPUTER MODEL strength of concrete. The new AASHTO LRFD2 uses a
This study utilized Caltrans column ductility program similar equation for the design of confinement steel; however, it
“Col-Duct”8 to perform the pushover analyses. This uses 0.16 instead of 0.12 in the previous equation.
program has been calibrated to experimental results of The main function of the transverse reinforcement is to
bridge column tests that were conducted at the University of insure that the axial load carried by the concrete column after
California at San Diego.8 The program calculates the mono- spalling of the concrete cover is at least equal to the load
tonic moment-curvature responses of reinforced concrete carried before spalling and that buckling of the longitudinal
sections and then determines the magnitude of deflection at reinforcement is prevented. Therefore, the spacing of the
various curvatures. The program is based on an assumption confinement steel plays a vital role in the seismic perfor-
that plane sections remain plane, and therefore the longitu- mance of RC columns. The spacing requirements for
dinal steel and concrete strains vary linearly across the confinement that were used in this study were based only on
section with a gradient equal to the curvature φ. Also, it is the previous equation and the maximum spacing that is
assumed that the stress gradient does not significantly affect suggested by Caltrans Bridge Design Specifications. There-
the behavior of the constituent materials within the section, fore, this study is limited only to the flexural behavior of
and thus the stress-strain models for these materials can be bridge columns since the lateral reinforcement was not
based on uniaxial behavior. checked for shear requirements.

24 ACI Structural Journal/January-February 1997

 ANALYSIS OF COLUMNS To achieve these objectives, this study investigated a
With the use of the previous material models, the program circular single column bridge bent subjected to transverse
performs an iteration process to determine the depth of the loading. Three characteristics of bridge column properties
neutral axis and the corresponding curvature by assuring that were set as target values for which the response modification
the internal stresses are in equilibrium with the external axial factors (R) were calculated. The definition of “R” is the same
load and moment. The program operates by subdividing a as that used in the codes.1,2 However, the values used in this
cross section into numerous elements representing sections of paper are based on elastic and inelastic analyses and vary
unconfined concrete, confined concrete, and longitudinal steel. depending on the parameters considered in the study. The
The strains in these elements are calculated from the product of three target values that were chosen for this study are the
the section curvature and the distance of the fiber from the ratios of:
neutral axis. Thus, the stress in these fibers is calculated based 1. Capacity to demand displacement (ΔC /ΔD)
on the strain and the corresponding material model. 2. Displacement ductility, μ = 2, 4, and 8
The ultimate curvature is defined as the curvature when 3. Longitudinal reinforcement, ρs = 1, 2, 3 percent
either the strain of the extreme fiber of confined concrete in
In general, to insure minimum satisfactory seismic perfor-
compression reaches ultimate strain value that results in a
mance of structures, the following basic design equation
failure of confinement steel, or the longitudinal tension steel
must be satisfied:
strain reaches 0.75 of the ultimate strain. Whichever of these
two conditions occurs first defines the value of ultimate
curvature. The program idealizes the moment-curvature supply ≥ demand
response of the reinforced concrete section based on an
energy concept where the area under the actual moment- In a displacement-based design environment, the previous
curvature response is equated to that of an idealized elastic- equation leads to
perfectly-plastic curve. Therefore, the idealized yield curvature
is defined as the break point in the idealized moment-curvature displacement capacity ≥ displacement demand
response. Based on the idealized yield curvature and the ulti-
mate curvature, the program calculates the idealized yield or in a simple term
displacement δy, the inelastic displacement δp, and thus the
ultimate displacement ΔC of the column as discussed earlier.
The displacement ductility μ that is used in this study is ΔC /ΔD ≥ 1
defined as
where ΔC is the ultimate displacement capacity of a structure
δy + δp that is based on the interaction between the nonlinear
μ = ---------------
- behavior of reinforced concrete section and the inelastic
δy structural analysis. On the other hand, ΔD is the displacement
demand that is the inelastic response due to seismic forces.
and the effective flexural rigidity EIeff as the slope of the Many researchers12-14 have investigated the concept of using
elastic part of the idealized moment-curvature diagram. the information of displacement capacity and displacement
The displacement demand (ΔD), on the other hand, is demand to design and thus control the behavior of RC struc-
generally based on an elastic dynamic analysis. For regular tures during severe ground motions. This study has chosen
bridges that do not have significant changes in the stiffness ΔC/ΔD as a target in determining the response modification
among their columns and have similar span lengths, the factors R since previous investigators12-15 have shown that
displacement demand may be obtained from an equivalent the concept of relating demand and capacity ratios has
static analysis. The period of the structure is calculated based insured that the RC structures would possess an adequate
on the mass of the superstructure tributary area and the seismic performance.
column cracked stiffness. Based on the period and an elastic The evaluation of the displacement demand and the
response spectrum, a static lateral force is applied at the top prediction of the displacement capacity is not a straightfor-
of the column to calculate the displacement demands. The ward process and most of the time it involves an interaction
ratio of the displacement capacity (ΔC) to the displacement between both of the demand and capacity displacements.
demand (ΔD) is checked against a specific value to determine Therefore, this study chose a ratio for ΔC /ΔD greater than
if the seismic performance of the column is satisfactory. The one to account for the variability in prediction of the
acceptable performance level is selected by the analyst and it displacement capacity and in the evaluation of the displace-
is usually set to prevent collapse. ment demand. A value of 1.33 for the ratio of ΔC /ΔD was
chosen to be a target value in this study, for which the R
OBJECTIVES AND SCOPE values were obtained for a wide range of periods.
The objectives of this research were as follows: The second target value that was chosen for this study was
• To obtain response modification factors for various peri- the displacement ductility ratio. Three values of constant
ods of single column bridge bents based on target values displacement ductility ratios (2, 4, 8) were selected as a basis
of capacity to demand displacement ratio, longitudinal for which the corresponding R values are determined. The R
reinforcement ratio, and displacement ductility ratio. values determined from a constant displacement ductility are
• To investigate the validity of using the ratio of capacity significant because they generally produce columns that
to demand displacement as a check in the column have the same performance level in a wide range period.
design process. The ratio of the longitudinal reinforcement was chosen as
• To determine the effect of lateral confinement on the the third target value in this study. The R values were calculated
response modification factors. for three ratios of longitudinal reinforcement equal to 1, 2,

ACI Structural Journal/January-February 1997 25

 freedom oscillators. The Caltrans ARS (A = Acceleration; R
= Rock; and S = Soil) Spectra Curve D with 0.7 g peak rock
acceleration11 for 5 percent damping was used as the ground
motion spectrum. Curve D represents relatively strong earth-
quakes on soft soil. Fig. 1 shows this curve, which is
normally used for sites where the depth of alluvium exceeds
45.7 m (150 ft). The spectrum was obtained from the product
of A, R, and S, where A is the maximum credible peak rock
acceleration, R is the acceleration spectra in rock, and S is
the soil amplification factor. Caltrans usually specifies an
equivalent seismic force for regular bridges, i.e., bridges
with nearly equal spans and columns of equal stiffness, as the
product of the ARS and the tributary weight.
The column diameter d was varied between 1.22 and 2.44
m (4 and 8 ft) in steps of 0.3 m (1 ft) to encompass practical
bridge column sizes. The axial force ratio, defined as the
ratio of axial load to the gross section, was equal to 0.1 f c′ in
all cases of this study. This ratio was chosen because it repre-
sents an average value of axial force on single column bents
of concrete bridges. An initial column height is specified at
this stage so an equivalent natural period of vibration of the
cantilever column idealized as a single degree for freedom
oscillator may be determined as

M
T = 2Π -----
K
Fig. 1—Caltrans ARS Spectra Curve D.
where M is the tributary mass and K is the lateral stiffness of
and 3 percent. These ratios were chosen because they represent the cantilever column that is equal to
common bridge column reinforcement ratios.
K = 3EIeff /h3
PROCEDURE OF DETERMINING R VALUES
TO SPECIFIED TARGET RATIO
To obtain the response modification factors for the three However, at this point the EIeff is not known yet; therefore, the
target values that were identified earlier, several aspects of initial lateral stiffness of the section is calculated by assuming
column properties were considered in the parametric anal- that the section cracked moment of inertia is approximately
ysis, such as height, diameter, axial force, and longitudinal equal to one-half of the section gross moment of inertia. Thus,
reinforcement ratio. To facilitate this parametric study, a the lateral seismic force H and its corresponding lateral
preprocessor and postprocessor were developed for the displacement ΔD can be calculated as discussed earlier.
“Col-Duct” program to automate the calculations. The The design moment at the base of the column can be
following procedure was used in this study: approximated by
1. Choose a design response spectrum.
2. Choose a column diameter. This defines the column ME = H × h + P × ΔD
axial load based on a selected axial load ratio.
3. Choose a column height. This height will be varied at the where ME is the elastic moment demand on the column base,
end of the operation to obtain a wide range of natural periods. H is the applied lateral force, P is the axial load on the
4. Choose a target value for ΔC /ΔD, μ, or ρs. This permits column, and ΔD is the demand lateral displacement due to the
the lateral seismic force to be calculated by an iterative applied lateral force. P-Δ accounts for the second order
procedure requiring sequential estimation of the member effect that is caused by the lateral displacement of the
stiffness, natural period, and therefore the lateral force. column under seismic load and the gravity load. This effect
5. Design (for ΔC /ΔD and μ target values) or choose (for ρs causes the moment at the column base to increase. At high
target value) the column longitudinal reinforcement. displacement ductility, the influence of the P-Δ effect
6. Design transverse reinforcement for ductility provisions. becomes significant due to the high value of the lateral
7. Calculate the effective flexural rigidity, ultimate displacement. Previous study16 has shown that the deflected
displacement capacity, plastic moment, period, lateral shape and hence the bending moment diagram are almost
seismic force, and corresponding displacement demand linear at high ductility levels, with the majority of the rota-
based on the column reinforcement. tion being concentrated at the plastic hinge forming at the
8. Calculate the design base moment (“moment demand”) column base. This linear approximation was used in this
at the base of the column. study to account for the second-order effect.
9. Calculate the response modification factor The longitudinal reinforcement is normally assumed in the
10. Repeat this operation by varying the column height. first cycle of iteration; thus the lateral stiffness of the column
The conventional method for describing the ground motion is is modified according to the calculated EIeff, which is based
a smoothed elastic response spectrum for single degree of on the actual longitudinal reinforcement. A revised period is

26 ACI Structural Journal/January-February 1997

then calculated based on the effective flexural rigidity, and a
new lateral seismic force is applied at the top of the column.
Then the actual target value is calculated and the entire
seismic design procedure is repeated until the desired target
value is achieved. At the end of the iteration, the response
modification factor is calculated as the ratio of

R = ME /MP

where MP is the plastic moment capacity of the section as

determined by the “Col-Duct” program and ME is the
demand moment at the base of the column.
Fig. 2—Average values of parameters for target value of
ΔC /ΔD = 1.33.
RESULTS OF ANALYSES
Figures 2 to 8 present the main results of the study. These
figures show the average values for the five different column
diameters that were used in this investigation.

RESPONSE MODIFICATION FACTOR SPECTRA

FOR CONSTANT ΔC /ΔD RATIO
Figure 2 shows the average values of various parameters:
ρs, h/d, μ, and R values obtained for the target value of ΔC/
ΔD = 1.33. However, as Fig. 2 shows, the target value of ΔC/
ΔD = 1.33 could not be achieved for periods less than 0.8 sec.
This is due to the fact that short period columns have a rela-
tively small displacement demand. In the range of periods
larger than 0.8 sec, the target value of ΔC/ΔD was achieved
by increasing the amount of required reinforcement (up to 4
percent). During the iteration process the reinforcement ratio
was allowed to vary between 0.01 and 6 percent to obtain the
desired target value. However, a common range of longitu-
dinal reinforcement ratio for bridge columns lies between 1
and 4 percent, thus limiting the acceptable range of the R
spectrum to periods longer than 1.25 sec. The values of R in
this range decrease from an approximate value of 5 to 3. The Fig. 3—Average R values for target ρs = 1, 2, 3 percent.
R value specified in the code1,2 for single column bents is 3.
Displacement ductility in this range of periods follows a very
similar trend as R values, with ductility values being approx-
imately 25 percent higher than the R values.
The following observations were made during the iterations:
• The R values increase in short period range and then
decrease in the long period range.
• Increasing the amount of the longitudinal reinforce-
ment would result in an increase of ΔC/ΔD ratio and
thus a decrease in the R values.
• The R values were relatively sensitive to small changes
in the ΔC/ΔD ratio. For example, increasing the ratio of
ΔC/ΔD by 10 percent would decrease R by 20 percent.
• Varying the column diameter did not have a significant
influence on R. However, it had a noticeable impact on
the reinforcement ratio required to obtain the target ΔC/ΔD,
especially in the long period range. For example, to
achieve ΔC/ΔD of 1.33, the reinforcement ratio for the
1.22-m-(4-ft)-diameter column had to be increased to 6
percent, while the reinforcement ratio for the 2.44-m- Fig. 4—Average displacement ductility μ for target ρs = 1,
(8-ft)-diameter column remained at only 2 percent. 2, 3 percent.
It is interesting to note that the displacement ductility asso-
ciated with the constant ΔC/ΔD ratio varied significantly as however, columns designed for constant ΔC/ΔD do not have
shown in Fig. 2. The displacement ductility dropped from a the same performance level.
value close to 8 for periods of 1 to less than 4 sec for periods
higher than 2 sec. This proves that the damage in concrete EFFECT OF LONGITUDINAL
columns designed to a constant ΔC/ΔD ratio varies depending REINFORCEMENT RATIOS
on the period of vibration. This implies that the ΔC/ΔD seems Figures 3 through 5 show, respectively, the average R
to be a viable check in the column design procedure; spectra, average displacement ductility, and the corre-

ACI Structural Journal/January-February 1997 27

Fig. 5—Average ΔC/ΔD ratio for target ρs = 1, 2, 3 percent. Fig. 7—Average displacement ductility for target μ = 2, 4, 8.

Fig. 6—Average R value for target μ = 2, 4, 8. Fig. 8—Average ρs for target μ = 2, 4, 8.

sponding ΔC /ΔD based on longitudinal reinforcement ratios constant R value would possess different displacement ductili-
of 1, 2, and 3 percent. Figure 3 shows that the values of R ties, and, therefore, would have distinct seismic responses.
increase with the period since the R values start from less Figure 5 illustrates the corresponding values of ΔC/ΔD
than one, i.e., elastic behavior, and reach maximum values of ratios. The curves start with values close to 15 in the short
5 and 3 for reinforcement ratios of 1 and 3 percent, respec- period range, which indicates high reserves in displacement
tively. Fig. 4 shows the corresponding displacement ductili- capacity in stiff structures, and then drops down to nearly 1.2
ties that decrease with the period. The maximum for long periods.
displacement ductility was equal to 11 and 8 and then drops The column diameter had no significant effect on the
down to 5 and 4 for reinforcement ratios of 1 and 3 percent, displacement ductility or on the ΔC/ΔD ratios. However, it
respectively. In general, these figures show an increase in affected the R values, especially in the long period range, by
both the R values and the displacement ductility as the rein- showing that columns with smaller diameters required
forcement ratio decreases. higher R values than columns with larger diameters. For
Figures 3 and 4 show that columns with less reinforcement example, for all ρs values, the 1.22-m (4-ft) column required
ratios have higher R values and do possess higher displacement R values that were approximately 35 percent higher than
ductility. This means that columns designed to high R values are those for the 2.44-m (8-ft) column.
more ductile than columns designed to low R values. For The influence of the selected bar sizes at specific rein-
example, using a constant R equal to 3, as specified in the forcement ratios was also checked because they affect the
code,1,2 would require ρs = 2 percent for a period equal to 1 sec length of the plastic hinge. It was concluded that the bar size
and ρs = 3 percent for a period equal to 2 sec. This shows that influence on R factors was negligible; however, it influenced
columns with different periods of vibrations designed for the ductility and ΔC/ΔD ratios in short period range.

28 ACI Structural Journal/January-February 1997

 RESPONSE MODIFICATION FACTOR SPECTRA ductility. Therefore, columns with improved confinement
FOR CONSTANT DUCTILITY provided a higher strength at the same displacement ductility
Figures 6 through 8 show the average R spectra for than columns with minimum confinement steel. The longitu-
different displacement ductilities (μ = 2, 4, and 8), the dinal reinforcement and ΔC /ΔD ratios increased in the mid-
achieved average displacement ductilities, and the corresponding period range by factors of 3 and 1.6 as a result of increasing
average longitudinal reinforcement ratios, respectively. The the confining steel.
general shape of R spectra, as shown in Fig. 6, is similar to
inelastic spectra that were determined by others3,4 based on SUMMARY AND CONCLUSIONS
inelastic dynamic analysis of a single-degree-of-freedom The primary purpose of this investigation was to deter-
oscillator with a bilinear behavior. The target value μ = 2 mine the response modification factors for single column
could not be achieved for some periods because it required circular reinforced concrete bridge bents using three target
impractical values for some of the parameters. values: the capacity to demand displacement ratio, displace-
Based on the results that are shown in Fig. 6, the R values ment ductility ratio, and longitudinal reinforcement ratio.
for a period of 0.5 sec ranged between 1.75 and 3.25 for Based on the assumptions of this study, the following
displacement ductilities of 4 and 8, respectively. For higher conclusions may be derived:
periods, the R values reached 8 for a displacement ductility • The capacity to demand a displacement check in the
of 8 and nearly 4 for a displacement ductility of 4, while the column design procedure seems to be a viable check in
corresponding reinforcement ratios were equal to 0.4 and the column design procedure. However, it was found
2.25 percent, respectively. The results of R values obtained that the response modification factors determined for
from this analysis can be used for design purposes if the main this target value were very sensitive to the period and
intention is to control deformations. The designer could did not show a consistent trend. This contradicts the
choose a ductility level that the column has to achieve and constant R value specified in bridge seismic codes. For
pick the corresponding R value based on the period of vibra- example, for a specified capacity to demand displace-
tion. This means that the columns would be designed for ment ratio equal to 1.33, the R values varied between 3
variable R values depending on the bridge natural period of and 6. This would prove that designing to a specified
vibration. This would insure that the columns would have capacity to demand the displacement ratio would not
similar performance levels and uniform damage during produce reinforced concrete sections that have the same
severe earthquakes. performance level.
• For a given reinforcement ratio, the response modification
EFFECT OF CONFINING STEEL ON R SPECTRA factors exhibit important variation with changes in
A second set of calculations using twice the amount of the period, particularly in the period range greater than 1
minimum confining steel was performed for the same target sec. The maximum response modification factor for 1
values as in the first set of spectra (ΔC /ΔD = 1.33, reinforce- percent longitudinal reinforcement ratio was equal to 5
ment ratios = 1, 2, 3 percent, and displacement ductility = 2, while it was equal to 3 for 3 percent reinforcement.
4, 8). The R spectrum corresponding to a constant ΔC /ΔD • Design for a specified displacement ductility of 2 and 4
ratio of 1.33 could not be obtained since the results of ΔC /ΔD produces very close R values. Columns designed to a
ratios were always higher than the target value even with the constant displacement ductility ratio would generally
smallest amount of longitudinal reinforcement. For a period produce a consistent design since the performance level
of 2.75 sec, the average ΔC/ΔD of 1.75 was closest to the is expected to be the same.
target value. At that point, the reinforcement ratio was equal • Doubling the minimum confining steel had only a
to 1.3 percent, which resulted in an R value of 4.5 and a slight effect on the response modification factor while it
corresponding displacement ductility of 7. When comparing had a significant effect on the displacement ductility
these with the results of minimum confinement at periods ratio and ΔC /ΔD.
close to 2.75 sec, it can be concluded that higher confine-
ment steel increased the displacement capacity and reduced ACKNOWLEDGMENTS
the need for longitudinal reinforcement. A smaller amount of Financial support for this study was provided through NSF EPSCoR
Cooperative Agreement OSR-9353227.
longitudinal reinforcement is further reflected in higher R
values and higher ductility capacity than in the case of
minimum confining steel. NOTATION
A = maximum credible peak rock acceleration
The R spectra for constant reinforcement ratios were only Ag = gross area of concrete column
obtained for the reinforcement ratio of 2 percent. The R d = diameter of concrete column
values were approximately 5 percent smaller than those db = diameter of longitudinal reinforcement
E = modulus of elasticity
obtained with minimum confinement steel. Therefore, the EIeff = effective flexural rigidity
improved confinement slightly increased the strength of the f ′c = compressive strength of concrete
columns. The influence of doubling the confinement steel fy = yield strength of steel reinforcement
was much more pronounced on the displacement ductility h = column height
and ΔC/ΔD ratios since they were improved by approximately H = lateral seismic force
K = lateral stiffness
40 and 60 percent, respectively. lp = plastic hinge length
The R spectra for constant displacement ductility obtained M = tributary mass
for μ = 8 were significantly lower (40 percent) than in cases ME = elastic demand moment
with minimum confining steel for the same target value. The MP = plastic moment
P = axial load
improved confinement allowed for higher reinforcement R = response modification factor
ratios and hence a higher strength (i.e., smaller R) for the R = acceleration spectra in rock
columns while still providing the required displacement S = soil amplification factor

ACI Structural Journal/January-February 1997 29

 T = natural period of vibration 7. “Seismic Assessment and Retrofit of Bridges,” Report No. SSRP -91/03,
δp = inelastic lateral displacement M. J. Nigel Priestly and Frieder Seible, eds., Department of AMES,
δy = elastic lateral displacement University of California, San Diego, July 1991.
ΔC = ultimate displacement capacity 8. Seyed, M., “Seismic Bridge Analysis Package,” Special Analysis Section,
ΔD = displacement demand Division of Structures, California Department of Transportation, Apr. 1993.
φu = ultimate curvature
9. Park, R., and Paulay, Reinforced Concrete Structures, John Wiley and
φy = yield curvature
Sons, Inc., New York, 1975.
μ = displacement ductility
ρ = volume of circular hoop reinforcement to volume of concrete core 10. Mander, J. B.; Priestley, M. J. N.; and Park, R., “Theoretical Stress-
ρs = longitudinal reinforcement ratio Strain Behavior of Confined Concrete,” ASCE Structural Journal, V. 114,
θp = plastic rotation No. 8, Aug. 1988, pp. 1804-1825.
11. “Bridge Design Specifications,” Caltrans, Sacramento, Calif., 1994.
REFERENCES 12. Veletsos, A. S., and Newmark, N. M., “Effect of Inelastic Behavior
1. “Standard Specifications for Highway Bridges,” 15th Edition, Amer- on the Response of Simple Systems to Earthquake Motions,” Proceedings,
ican Association of State Highway and Transportation Officials, Wash- Second World Conference on Earthquake Engineering, V. 2, Japan, 1960,
ington, D.C., 1992. pp. 895-912.
2. “AASHTO LRFD Bridge Design Specifications,” American Association 13. Sozen, M. A., “Review of Earthquake Response of RC Buildings
of State Highway and Transportation Officials, Washington, D.C., 1994. with a View to Drift Control,” State-of-the-Art in Earthquake Engineering,
3. Penzien, J., “Seismic Design Criteria For Transportation Structures,” 1981, O. Ergunay and M. Erdik, eds., Ankara, Oct. 1981, pp. 383-418.
Proceedings of the ASCE Structures Congress ‘93, V. 1, pp. 4-36. 14. Qi, X., and Moehle, J. P., “Displacement Design Approach for Rein-
4. Vidic, T.; Fajfar, P.; and Fischinger, M., “Consistent Inelastic Design forced Concrete Structures Subjected to Earthquakes,” Report No. UCB/
Spectra: Strength and Displacement,” Earthquake Engineering and Structural
EERC-91-02, Earthquake Engineering Research Center, University of
Dynamics, V. 23, 1994, pp. 507-521.
California at Berkeley, California, Jan. 1991, 186 p.
5. Priestley, M. J. N., and Park, R., “Strength and Ductility of Concrete
Bridge Columns under Seismic Loading,” ACI Structural Journal, V. 84, 15. “A Handbook for Seismic Evaluation of Existing Buildings,” ATC
No. 1, Jan.-Feb. 1987, pp. 61-67. 22, Applied Technology Council, Redwood City, Calif., 1989, 169 pp.
6. Saiidi, M., and Sozen, M. A., “Simple Nonlinear Seismic Analysis of 16. Verma, R., and Priestly, M. J. N., “Optimal Trends in Seismic Design
RC Structures,” Journal of the Structural Division, ASCE, V. 107, No. ST5, of Single Column Circular Reinforced Concrete Bridge Piers,” EERI,
May 1981, pp. 937-951. Earthquake Spectra, V. 10, No. 3, Aug. 1994, pp. 589-614.

30 ACI Structural Journal/January-February 1997