ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title no. 106-S37

Seismic Performance of Circular High-Strength

Concrete Columns
by Patrick Paultre, Rami Eid, Hugo Ita Robles, and Najib Bouaanani

This paper presents results from tests of six large-scale high- performance-based equations depend on ductility demand
strength concrete (HSC) circular columns under reverse cyclic- and account for the level of axial load; the concrete compressive
flexure and constant axial loads. The objectives of the tests strength; and the transverse reinforcement configuration,
presented in this paper were to examine the post-elastic behavior spacing, and yield strength. These equations are the basis of
and the ductility level reached by HSC circular columns designed the prescriptive design confinement reinforcement requirements
according to the 2004 CSA A23.3 requirements for transverse steel
in the 2004 CSA A23.3 Standard.16
reinforcement. The columns were subjected to constant axial loads
and a cyclic horizontal load-inducing reversed bending moment. It A large database of concentric axial compression tests on
is shown that columns designed according to the confinement confined HSC columns is available in the literature. Only a
reinforcement requirements of the Canadian standard behave in a small number of experimental test results, however, are
ductile manner regardless of transverse steel reinforcement yield available for HSC columns under combined cyclic flexure
strength or axial load level. and axial load,17 predominantly for square (or rectangular)
columns. Furthermore, the authors considered that there is
Keywords: confined concrete; ductility; high-strength concrete; tied column. a lack of test data on HSC circular columns subjected to
such loading, except for the tests performed by Saatcioglu
INTRODUCTION and Baingo.18
In recent years, the use of high-strength concrete (HSC) This paper presents tests on circular HSC columns
has been steadily increasing due to its many advantages over subjected to combined constant axial load and reversed
normal-strength concrete (NSC). HSC is used primarily in cyclic flexure simulating earthquake loading. The objectives
high-rise buildings to significantly decrease the dimensions of the tests presented herein are to examine the post-elastic
of the columns in lower stories, thereby reducing the volume behavior and the ductility level reached by HSC circular
of concrete. Studies have shown, however, that HSC is more columns designed according to the 2004 CSA A23.3
brittle than NSC1,2 in compressive behavior. These studies Standard’s16 transverse steel reinforcement requirement.
show that when the concrete strength increases, the amount
of confinement reinforcement has to be increased to reach a RESEARCH SIGNIFICANCE
constant level of ductility for columns subjected to the same There is an urgent need for test data on spirally reinforced
level of axial load. circular HSC columns subjected to combined constant axial
While the ACI Code3 does not limit the concrete compressive load and reversed cyclic flexure. This paper presents new
strength, the 1995 New Zealand Standard NZS 31014 and the experimental results obtained from six HSC circular
1994 CSA A23.3 Standard5 limited the maximum strength columns subjected to constant axial load and reversed cyclic
that can be used for seismic design to 70 MPa and 55 MPa flexure. The targeted concrete strength was 100 MPa (14.5 ksi).
(10.15 ksi and 7.98 ksi), respectively. These limits were due The amount of transverse steel reinforcement was determined
in part to a lack of test results when the latter standards were according to the confinement reinforcement requirements
published. Légeron and Paultre6 and Azizinamini et al.7 have of the 2004 CSA A23.3 Standard.16 The post-elastic
shown that the axial load level has a significant influence on behavior and the level of ductility of the test specimens are
the ductility of concrete columns subjected to cyclic flexure used to evaluate the new confinement steel requirement of
and constant axial loads. Moreover, studies have shown that the CSA Standard16 as well as the requirements of the ACI
HSC columns reinforced according to the current ACI Code Code3 and the New Zealand Standard.4
confinement requirements behave in a ductile manner if the
axial load is less than 0.20Ag fc′ ,6-11 where Ag is the gross DESIGN EQUATIONS FOR
area of column section and fc′ is the concrete compressive CONFINING SPIRAL REINFORCEMENT
strength. The New Zealand Standard4 is the only one of the Confinement of concrete columns is required by design
three standards examined that take the axial load level into codes to provide ductile behavior to concrete members
account in the confinement reinforcement requirements. subjected to seismic loading. Minimum amounts of confinement
Note that extensive test programs were conducted on square reinforcement are recommended in design codes of practice
columns and very limited test data exist for circular columns to ensure a certain level of ductility. According to ACI 318-083
confined with spirals. and the 1994 CSA A23.3 Standard,5 the volumetric ratio of
Following a comprehensive research program at the University
of Sherbrooke,1,6,12-14 and using test results from other
researchers in the U.S., Canada, and New Zealand, new design ACI Structural Journal, V. 106, No. 4, July-August 2009.
MS No. S-2006-499.R1 received April 30, 2008, and reviewed under Institute
equations for the volumetric transverse steel reinforcement publication policies. Copyright © 2009, American Concrete Institute. All rights reserved,
ratio were proposed for columns in ductile and limited including the making of copies unless permission is obtained from the copyright proprietors.
ductility moment-resisting frame structures.15 These new Pertinent discussion including author’s closure, if any, will be published in the May-June
2010 ACI Structural Journal if the discussion is received by January 1, 2010.

ACI Structural Journal/July-August 2009 395

 fy(0.85fc′ ), fy is the lower characteristic yield strength of
Patrick Paultre, FACI, holds a Canada Research Chair in Seismic Engineering at the
University of Sherbrooke, Sherbrooke, QC, Canada. He is a member of Joint ACI- nonprestressed longitudinal reinforcement, d′′ is the diameter
ASCE Committees 352, Joints and Connections in Monolithic Concrete Structures, of concrete core of circular column measured from out-to-
and 441, Reinforced Concrete Columns.
out of spirals or circular hoops, and db is the nominal diameter
Rami Eid is a Structural Engineer at Halcrow Yolles, Toronto, ON, Canada. He of longitudinal bar. In Eq. (3), Ag /Ac shall not be taken less
received his PhD in civil/structural engineering from the Technion-Israel Institute of than 1.2, ρgm shall not be taken greater than 0.4, and fyh shall
Technology, Haifa, Israel. His research interests include analytical and experimental
behavior of reinforced concrete elements subjected to static and seismic loads.
not be taken greater than 800 MPa (116.03 ksi). Note that
Eq. (4) is an antibuckling requirement.
Hugo Ita Robles is a Structural Engineer at Mesar, Montréal, QC, Canada. He In the case of circular concrete columns that are part of a
received his MSc in civil engineering from the University of Sherbrooke.
ductile moment-resisting frame structure, the 2004 CSA
Najib Bouaanani is an Assistant Professor in the Department of Civil, Geology, and A23.3 Standard16 requires that the volumetric transverse
Mining Engineering at École Polytechnique, Montréal, QC, Canada. He received his steel reinforcement ratio not be less than the greater of
PhD in civil engineering from the University of Sherbrooke.
Eq. (1) and

spiral reinforcement, ρs , for circular columns part of a special f c′

moment frame shall not be less than the greater of ρ s = 0.4k p -----
- (5)
f yh
ρs = 0.45(Ag /Ach – 1)fc′ /fyh (1)
where kp = P/P0. This equation has been calibrated for P0 =
and 0.85(Ag – Ast)fc′ + Ast fy (fy is the yield strength of the
longitudinal reinforcement) and this definition will be used
ρs = 0.12fc′ /fyh (2) in this paper. It should be noted that the main goal of the tests
presented herein is to examine the post-elastic behavior and
the ductility level reached by HSC circular columns
where ρs is the ratio of volume of spiral reinforcement or hoops designed according to the 2004 CSA A23.3 Standard16
to the total volume of core measured out-to-out of the spirals or requirement for transverse steel reinforcement (Eq. (5)) and
hoops, Ag is the gross area of the column section, Ach is the area to provide much needed test data on this type of column.
of concrete core measured to the outside diameter of transverse
reinforcement, and fyh is the specified yield strength of transverse
reinforcement but not more than 700 MPa (100 ksi) in the ACI EXPERIMENTAL PROGRAM
Code3 and 500 MPa (70 ksi) in the CSA standard.16 Test specimens
Studies have shown,6,8,19 however, that the axial load level The experimental program consisted of six circular HSC
significantly affects the flexural behavior of HSC columns. columns measuring 300 x 2150 mm (11.81 x 84.65 in.)
Based on these studies, NZS 31014 and the 2004 CSA A23.3 connected to a massive I-shaped stub and cast vertically. The
Standard16 recommend equations that account for the axial load test setup is shown in Fig. 1. The transverse load was applied
level. For circular columns, the New Zealand standard4 requires at the tip of the specimen 2.0 m (78.74 in.) from the base of
that the volumetric ratio ρs shall not be less than the greater of the column with a force/displacement-controlled hydraulic
actuator. The columns represent a 4.0 m (157.48 in.) high
column in a typical building with the point of contraflexure
( 1.3 – ρ g m ) A g f c′ P
ρ s = ---------------------------
- -------- ------ --------------- – 0.0084 (3) located at column midheight. Figure 2 shows the specimen
2.4 A ch f yh φf c′ A g reinforcement details.

and Test variables

As stated previously, the HSC columns were designed and
A st f y 1 tested to examine the design equation (Eq. (5)) adopted by
ρ s = --------------
- ------ ----- (4) the Canadian standard.16 The specimen test variables were:
100d″ f yh d b
1) the level of the axial load; 2) the volumetric transverse
steel reinforcement ratio; and 3) the transverse steel yield
where ρg = Ast /Ag is the ratio of nonprestressed longitudinal strength. The combination of the designed variables in each
column reinforcement to gross cross-sectional area, m = specimen was expected to provide a similar level of ductility
to all columns regardless of the axial load level. The targeted
concrete strength was 100 MPa (14.50 ksi). The axial load
level is defined as the index Pa/Ag fc′ , where P is the constant
axially applied compression load. Three targeted axial load
levels of 15, 25, and 40% and two transverse reinforcement
yield strengths of fyh = 440 MPa (64 ksi), corresponding to
Grade 400, and 560 MPa (81 ksi), corresponding to Grade
500, were chosen. These grades of steel recommended by the
Canadian standard were used to determine the volumetric
transverse steel ratio of each specimen according to Eq. (5).
According to this equation, increasing the axial load level or
decreasing the yield strength of the lateral steel results in a
higher demand of transverse reinforcement ratio to achieve a
Fig. 1—Experimental setup. minimum level of displacement ductility.

396 ACI Structural Journal/July-August 2009

 Table 1—Concrete characteristics
fc′ , Ec , fr ,
Specimen MPa (ksi) εc′ MPa (ksi) εc50u MPa (ksi)

C100S100N15 109 0.00307 42,300 0.00420 8.86

(15.81) (6135) (1.29)

C100SH100N15 100.5 0.00294 40,700 0.00438 8.19

(14.58) (5903) (1.19)
103 41,400 9.61
C100S70N25 (14.94) 0.00312 (6005) 0.00468 (1.39)

C100SH70N25 97 42,700 8.78

0.00316 0.00427
(14.07) (6193) (1.27)
C100S37N40 100 0.00318 41,600 0.00467 9.75
(14.50) (6034) (1.41)

C100SH37N40 103 0.00305 43,400 0.00403 8.90

(14.94) (6295) (1.29)

Table 2—Steel characteristics

Steel type fy , MPa (ksi) εsh εsu fsu , MPa (ksi)
Fig. 2—Reinforcing cage and instrumentation details.
No. 3 425 (61.64) 0.00843 0.112 679 (98.48)
No. 10M 440 (63.82) — 0.173 656 (95.14)
The specimens are identified by concrete strength (C100
No. 20M 560 (81.22) — 0.050 697 (101.09)
for 100 MPa [14.50 ksi]), spiral transverse reinforcement
grade (S for Grade 400 and SH for Grade 500), spiral pitch
in mm (100, 70, and 37 mm [3.94, 2.76, and 1.46 in.]), and in Table 2, where fy is the yield strength, εsh is the strain at
the axial load level (N15, N25, and N40 for 15, 25, and 40% the commencement of strain hardening, and εsu is the ultimate
of Ag fc′ , respectively). It should be noted that although a strain corresponding to the ultimate stress fsu. Unlike the
spacing of 100 mm (3.94 in.) does not satisfy the CSA 20M Grade 400 steel, which exhibited a well-defined yield
standard16 and ACI Code3 requirements (maximum of 45 mm plateau from the beginning of yielding to the commencement
[1.8 in.] center-to-center and 75 mm [3 in.] clear spacing, of strain hardening, the 10M Grade 400 and the No. 3 Grade
respectively), this spacing was chosen intentionally to meet 500 steel had no distinct yield plateaus.
the research goal of examining the ductility level of HSC
columns containing different ratios of transverse steel Reinforcing cages
reinforcement. Details of reinforcing cages are shown in Fig. 2. A volumetric
longitudinal steel reinforcement ratio of 2.55% was provided
Material properties in each specimen with six 20M Grade 400 deformed bars.
The spiral reinforcement was provided with two types of
Concrete—The specified 100 MPa (14.50 ksi) concrete
bars: 10M Grade 400 and No. 3 Grade 500 deformed bars.
was mixed in the concrete laboratory at the University of
Different steel grades were used to obtain similar confinement
Sherbrooke. The concrete formulation and specimen casting
levels for different volumetric transverse steel reinforcement
were similar to those used and reported by Cusson and
ratios. To prevent crushing of the concrete, spiral spacing
Paultre1 and by Légeron and Paultre,6 respectively. To determine
was reduced at the top of the columns where the axial load
the concrete characteristics, the reinforced concrete column
was applied. Elsewhere, the spirals were equally spaced. The
specimens and the concrete control cylinders and prisms
stub was designed to prevent excessive cracking and provide
were cured under similar conditions.
proper anchorage for the column’s longitudinal bars.
Table 1 summarizes the material properties. The concrete
compressive strength fc′ and the corresponding strain εc′ Instrumentation and testing procedure
were determined from standard compressive tests on at least Instrumentation—Electrical strain gauges and linear
three 150 x 300 mm (6 x 12 in.) cylinders. The concrete variable differential transformers (LVDTs) were used to
compressive strength ranged from 97 to 109 MPa (14.07 to measure strain and displacement. Sixteen electrical strain
15.81 ksi) and the corresponding strain ranged from 0.00294 gauges were placed on the longitudinal bars above and in the
to 0.00318. The postpeak strain at 50% maximum stress, stub (Fig. 2). Within the lower 200 mm (7.87 in.) of the
εc50u , was determined from the complete stress-strain curves column, another nine strain gauges were placed on the spiral.
obtained on 100 x 200 mm (4 x 8 in.) cylinders. The strain Curvatures were calculated using three sets of LVDTs that
ranged from 0.00403 to 0.00468. The secant modulus of were placed in the plastic hinge region. Three LVDTs with a
elasticity Ec ranged from 40,700 to 43,400 MPa (5903 to range of 150 mm (5.91 in.) were attached to steel collars to
6295 ksi). The concrete cracking strength fr , obtained from measure average concrete strain over a gauge length of 560 mm
modulus of rupture tests on at least three 100 x 100 x 400 mm (22.05 in.). The second series included four LVDTs with a
(3.94 x 3.94 x 15.75 in.) prisms for each specimen, ranged range of 25 mm (0.98 in.) and a gauge length of 300 mm
from 8.19 to 9.75 MPa (1.19 to 1.41 ksi). (11.81 in.) that were supported by steel rods passing through
Reinforcement—Three different types of reinforcing bars the core and extending from one side of the column to the
were used: 10M (100 mm2 [0.155 in.2]) and 20M (300 mm2 other. These rods were attached to the longitudinal bars
[0.465 in.2]) Grade 400 bars, and No. 3 (71 mm2 [0.110 in.2]) before concreting. The third series of three LVDTs with a
Grade 500 bars. Complete stress-strain curves were obtained range of 5 mm (0.20 in.) were attached to a steel collar and
from test coupons on each of the steel types used. The averages to the column base to measure average concrete strain over a
of at least three steel coupons for each steel type are shown gauge length of 30 mm (1.18 in.) (Fig. 2).

ACI Structural Journal/July-August 2009 397

Table 3—Details of test specimens
fc′, MPa dh, mm s, mm fyh, MPa
Specimen (ksi) (in.) (in.) ρs, % (ksi) P/Ag fc′ kp
109
C100S100N15 (15.81) 11.3 100 0.014 440
(0.44) (3.94) 3 (63.82) 0.15 0.16
100.5
C100SH100N15 (14.58) 9.5 100 0.010 560 0.15 0.16
(0.37) (3.94) 0 (81.22)
103
C100S70N25 (14.94) 11.3 70 0.020 440 0.25 0.27
(0.44) (2.76) 4 (63.82)
97
C100SH70N25 (14.07) 9.5 70 0.014 560 0.25 0.26
(0.37) (2.76) 3 (81.22)
100 11.3 37 0.038 440
C100S37N40 (14.50) (0.44) (1.46) 5 (63.82) 0.40 0.43

C100SH37N40 103 9.5 37 0.027 560 0.40 0.43

(14.94) (0.37) (1.46) 1 (81.22) Fig. 3—Loading history.

Table 4—Summary of test results

Specimen M′max , kN·m (kip·in.) μΔuI μφuI μφult EN IW DEW δu, % Ic Ie′ Ik
C100S100N15 168.4 (1490.54) 5.3 14.0 17.1 37.8 31.3 128.5 6.7 0.058 0.026 0.36
C100SH100N15 172.3 (1525.06) 5.0 12.0 16.1 33.7 28.2 99.9 6.8 0.056 0.019 0.31
C100S70N25 203.0 (1796.79) 4.8 11.4 20.8 39.7 26.8 113.8 6.3 0.087 0.042 0.33
C100SH70N25 190.7 (1687.92) 5.8 15.4 32.8 37.5 30.2 155.0 6.1 0.083 0.030 0.28
C100S37N40 227.8 (2016.30) 7.4 26.0 26.4 48.8 37.1 209.0 8.0 0.170 0.090 0.40
C100SH37N40 235.3 (2082.68) 6.7 21.3 21.3 42.7 30.0 147.7 8.2 0.147 0.077 0.31

The column specimens were tested in a frame that was events occurred: 1) the column was not able to sustain axial
specifically designed for this research (which includes the load, characterized by a loss of axial load during a quarter of
tests reported in this paper and those performed by Légeron a cycle; 2) flexural resistance dropped more than 50% of the
and Paultre6 and by Paultre et al.13). The axial constant maximum experienced capacity; and 3) a longitudinal bar
compression load was applied to the columns through four to ruptured, inducing a large drop in flexural capacity.
six high-strength 36 mm (1.42 in.) diameter bars tensioned
by four 1500 kN (337.23 kip) and two 1000 kN (224.82 kip) TEST RESULTS
hydraulic jacks. Strain gauges were placed on each bar to General behavior
determine the applied axial load. This load was adjusted The lateral load versus tip displacement of all the
during the test when the column was in its initial position. specimens is shown in Fig. 4. The lateral load is reported as
The horizontal load was applied by a 500 kN (112.41 kip) measured and is not corrected for the P-Δ effect, which is
MTS actuator with displacement and force control capabilities, indicated by the dashed line (refer to Fig. 4). It should be
supported by four braced steel columns. The horizontal load noted that a strength gain or loss is obtained when the
was measured by the load cell on the actuator, which was response curve in absolute terms lies above or below the P-Δ
also equipped with an LVDT to control the actuator while in dashed line, respectively. Occurrences of special events such
displacement control mode. The horizontal tip displacement as spalling of concrete cover, yielding of longitudinal bars,
was measured by an LVDT with a range of 300 mm (11.81 in.). yielding of spiral, and loss of axial load capacity are
This LVDT was fixed to the laboratory strong floor to obtain indicated in Fig. 4.
the column tip displacement independent of the experimental Tables 3 and 4 summarize the main variables and the
experimental results obtained for each specimen, respectively.
frame movement. It should be noted that due to the high
Table 3 presents the specimens’ variables including
stiffness of the frame, the difference between the actuator
measured concrete compressive strength fc′ , spiral spacing s,
LVDT and the independent LVDT was negligible (a
spiral diameter dh, volumetric spiral ratio ρs, spiral yield
maximum of 1% difference).
strength fyh, ratio P/Ag fc′ , and kp. Table 4 summarizes the
Test procedure—The axial load was applied at the target main indexes quantifying the hysteretic behavior of each column.
value, and the cyclic horizontal loading commenced. The These indexes will be explained in the following section.
first loading cycle was applied under load control to 75% of All specimens experienced failure by concrete crushing.
the predicted analytical yield load. The second cycle under Buckling of longitudinal bars was clearly noticed in the
displacement control reached the yield load, with yield columns with 100 mm (4 in.) spiral spacing, whereas the
displacement being defined as the point at which the longitudinal bars buckled slightly or not at all for the specimens
longitudinal bars first yielded in tension. After reaching the with 70 or 37 mm (2.75 or 1.46 in.) spacing, respectively.
yield displacement, each cycle was under displacement This behavior indicates that the maximum spacing requirement
control with a maximum displacement equal to 1.5, 2, 3, ..., of the CSA standard16 (maximum spacing for the specimens is
times the measured yield displacement up to failure (Fig. 3). controlled by Dc /6 = 45 mm [1.77 in.], where Dc is the core
Except for the first cycle, whose sole purpose was to crack diameter) is too restrictive, especially for columns of small
the member to simulate real conditions and obtain elastic dimension (Dc < 420 mm [16.5 in.]). The spacing requirements
characteristics, all the subsequent cycles were repeated of the ACI Code3 seem to be more appropriate for the
twice. The test ended when at least one of the following three circular specimens tested in this study with a maximum

398 ACI Structural Journal/July-August 2009

 Fig. 4—Uncorrected lateral load versus tip displacement.

controlled by a clear spacing of 75 mm (3 in.). Figure 5 splitting plane between the cover and core was found to be
shows the appearance of the plastic hinge zone after testing quite smooth, due to cracks passing through the aggregate.
for all specimens. Figures 5 and 6 show that the higher the
axial load level, the larger the damaged region of the column. Ductility and energy dissipation
For example, the length of the damaged region of Column Ductility parameters and energy dissipation capacity are
C100S37N40 with 40% axial load level is approximately 1.6 usually used to evaluate the seismic response of reinforced
times the damaged length of Column C100S100N15 with concrete members.6,13 While the ductility of long-period
15% axial load level. It can also be seen in Fig. 5 and 6 that structures is directly related to the strength reduction factor
the damaged zone for all columns started at a distance of 25 used in most codes21 to calculate the seismic base shear, the
to 30 mm (0.98 to 1.18 in.) above the base stub. This behavior, energy dissipation capacity can be used as a response indicator
observed in similar specimens by other researchers,6,8,9,13,19 in the design of short-period structures and structures
is attributed mainly to the confinement provided by the base subjected to a long-duration earthquake.6,13
stub to the column sections just above it.19,20 Therefore, the To overcome the nonlinear response of the reinforced
resisting moment M′max shown in Table 4 is calculated at concrete column, the ductility parameters are defined from
30 mm (1.18 in.) above the column-stub interface. an idealized diagram.19,22 Therefore, the load-displacement
The overall behavior of all the specimens was similar and the moment-curvature curves are idealized as bilinear
(Fig. 4). The cover of Columns C100S100N15 and diagrams, constituted of an elastic branch and an inclined
C100SH100N15 spalled off after yielding of longitudinal inelastic branch (Fig. 7). In the load-displacement curve, the
reinforcement, whereas for the other specimens, it spalled off elastic branch is secant to the real curve at 75% of the
before yielding. No warning sign, such as vertical cracking, maximum horizontal load, and reaches the maximum load to
was noticed prior to spalling in any of the specimens. The define the idealized yield displacement ΔyI. Column failure

ACI Structural Journal/July-August 2009 399

 Fig. 6—Sketches of most damaged regions of specimens.

Fig. 5—Damaged regions and plastic hinge zones for all

specimens.

is conventionally defined at the postpeak displacement Δ2,

where the remaining capacity of the column has dropped to
80% of the peak load. The idealized post-elastic branch is
defined between the two points (ΔyI, Hmax) and (Δ2, H2),
where H2 is determined from equating the area under the
actual and ideal curves. The moment-curvature curve is Fig. 7—Ideal curve definitions.
idealized with a similar procedure (Fig. 7(b)).
The idealized ultimate displacement ductility is defined as where z = 2000 mm (78.74 in.) is the column height. It
should be noted that the interstory drift ratio is reported on
Δ the top horizontal axes in Fig. 4.
μ ΔuI = ------2- (6) Energy dissipation is defined for a cycle i by the hatched
Δ yI
area in Fig. 8 or mathematically by
and the idealized ultimate curvature ductility is defined as
B

φ
μ φuI = ------2 (7)
Ei =
∫ F du (9)
φ yI A

The total energy dissipated during the test until failure is

The main drawback of using the ductility parameters is the
lack of consensus in the research community on a definition n
of yielding of a reinforced concrete member. The maximum
interstory drift ratio δu is a simpler parameter to use and is
E hyst = ∑ Ei (10)
i=1
defined based on the measured displacement at failure
where n is the number of cycles to failure. For comparison
Δ purposes, it is convenient to normalize the dissipated energy
δ u = -----2- (8)
z as follows

400 ACI Structural Journal/July-August 2009

 n
1
E N = ---------------------
H max′ Δ yI ∑ Ei (11)
i=1

where H′max is the maximum equivalent lateral load, which

includes the applied horizontal load and the equivalent
horizontal load due to the P-Δ effect.
Energy dissipation and inelastic deformation capabilities
may also be assessed by work and damage indexes. The
work index IW is defined as23

n
Hi Δi
IW = ∑ ---------------------
H max′ Δ yI
(12)
i=1

A damage index DEW that combines the cyclic dissipated

energy and the elastic energy is defined as24
Fig. 8—Energy dissipation.
n
K Δ 2
∑ Ei ⎛⎝ -------
i⎞⎛ i⎞
1 yield strengths. It was found that the effective confinement
D EW = --------------------- ------- (13)
H max′ Δ yI KyI ⎠ ⎝ ΔyI ⎠ index Ie′ is more appropriate to evaluate confinement efficiency
i=1
for concrete columns with similar axial load levels. This
where Ki and Δi are defined in Fig. 8. The defined indexes index was introduced by Légeron and Paultre14 as
μΔuI, μφuI, EN, IW, and DEW are computed for all the columns
and are given in Table 4. f le′
I e′ = ------
- (14)
f c′
ANALYSIS OF RESULTS
The purpose of this research program was to examine the where fle′ is the effective confinement pressure or the passive
post-elastic behavior and level of ductility of HSC columns pressure applied by the ties to the concrete core when
containing transverse steel reinforcement amounts determined confined concrete reaches its maximum axial stress. This
according to the requirement given in the 2004 CSA A23.3-04 pressure is derived from force equilibrium on the column
Standard16 (Eq. (5)). This requirement was proposed after a half cross section14
comprehensive research program1,6,13-15 to ensure a certain
level of ductility in reinforced concrete columns under axial
and cyclic lateral loadings. Figure 9 and Table 4 show that A sh
f le′ = K e -------f ′ (15)
although the maximum moment capacity is affected by the cs h
level of the axial load, the post-elastic behavior of the six
specimens is quite similar with μΔuI > 4.8. The comparable where Ke is the geometric effectiveness coefficient of the
values of the energy dissipation indexes indicate quite clearly transverse steel reinforcement defined by Sheikh and
the similarity in the inelastic behavior of the specimens. Hence, Uzumeri25 and Mander et al.,26 Ash is the total cross-sectional
concrete columns that contain different amounts of transverse area of the ties in one direction, c is the cross-sectional
steel reinforcement, different transverse steel yield strengths, dimension of the column core (center-to-center of ties), and
and different axial load levels have similar postpeak fh′ is the stress in the transverse steel at peak confined
behavior and energy dissipation (compare, for example, the concrete stress. Nonetheless, the test results presented in this
results of Columns C100SH37N40 and C100S100N15 in study show (Table 4) that columns with different effective
Table 4). Although there is a high probability that the confinement index values have quite similar ductility and
inelastic behavior of the columns designed according to energy dissipation parameters (compare, for example,
Eq. (5) will be quite similar, it should be emphasized that this Columns C100S70N25 and C100SH37N40 in Table 4). This
design equation was developed to ensure a sufficient actual behavior is attributed to the difference in the applied axial load
ultimate sectional ductility level15 for the reinforced levels of the reported specimens and can be demonstrated by
concrete columns and not to ensure a similar post-elastic introducing a simplified confinement index, which takes into
behavior. Note that the results of the sectional (curvature) account the influence of the axial load level on the confinement
ductility μφult corresponding to the ultimate curvature (φu effectiveness as follows
instead of φ2 in Eq. (7); refer to Fig. 7) are also reported in
Table 4. These results support those from other indexes in ρ s f h′
Table 4 and the behavior shown in Fig. 4, that is, the six tested I k = ----------
- (16)
specimens achieved an adequate ductility level. f c′ k p
Paultre et al.13 have shown that the confinement index
Ic = ρs fyh/fc′ (which has been recommended by Joint ACI- According to Eq. (5) and (16), a circular reinforced
ASCE Committee 44110 to evaluate confinement efficiency) concrete column can achieve sufficient sectional ductility if
cannot capture the effect of high-yield-strength confinement the index Ik is at least 0.4. The variable fh′ can be determined
steel and is not appropriate when comparing columns under according to Légeron and Paultre;14 but for simplification, fh′
different axial load levels or reinforced with steel of different can be taken as fyh up to 500 MPa (72.5 kips) as recommended

ACI Structural Journal/July-August 2009 401

 Fig. 9—Experimental and predicted moment-curvature response for all specimens.

by the CSA standard.16 The similar general behavior and the loading. Figure 9 compares the experimental moment-
close values of the ductility and energy dissipation parameters curvature of the six columns with analytical curves
of the six tested columns can be explained by the narrow computed using a computer program28 with a layer-by-layer
range of their confinement index Ik, which varies between analysis and by applying the Légeron and Paultre14 concrete
0.28 to 0.4, as shown in Table 4. Note that Ik is not always model. The experimental curvatures were obtained from
equal to 0.4 for all the tested specimens because the ρs used LVDT readings on a gauge length of 300 mm (11.81 in.), as
is less than the ρs required according to Eq. (5), except for shown in Fig. 2. The predictions and experimental moment-
Column C100S37N40. curvature responses are in very good agreement, as can be
seen from Fig. 9.
Prediction of test results
The study presented in this paper is part of a comprehensive
Comparison with code requirements
program underway at the University of Sherbrooke. Under
this research program, Légeron and Paultre14 proposed a In seismic zones, structural members designed to behave
stress-strain concrete model suitable for representing the in a ductile manner should be well-confined. The ACI Code3
axial behavior of circular and square concrete columns made and the New Zealand4 and the Canadian16 standards’
of NSC and HSC (20 to 140 MPa [2.9 to 20.31 ksi]) confined requirements for the minimum volumetric transverse steel
by normal- or high-strength (300 to 1400 MPa [43.51 to reinforcement ratio in circular columns were given
203.05 ksi]) confinement steel. This model, which was previously (refer to Eq. (1) to (5)). Table 5 compares these
found to predict experimental results with good accuracy,27 requirements for the six columns presented in this study. It
was used to develop the new confinement reinforcement should be noted that, according to the confinement reinforcement
requirement of the Canadian standard16 (Eq. (5)). Therefore, requirement of the Canadian standard16 (Eq. (5)), fyh in
it was important to evaluate the model’s predictions with Eq. (5) should not be taken greater than 500 MPa (72.52 ksi)
experimental results from rectangular6,13 and circular and the axial load level is accounted for by the parameter kp.
reinforced concrete columns under axial and cyclic lateral The designed volumetric ratios of the tested specimens,

402 ACI Structural Journal/July-August 2009

Table 5—Design code requirements and specimen ductility
ρ s, prov ρ s, prov ρ s, prov μΔu
-------------- -------------- ---------------------------
-
Specimen fc′ , MPa (ksi) fyh, MPa (ksi) ρs kp ρ s, ACI ρ s, NZS ρ s, CSA – Eq. (5) (μΔuI)
C100S100N15 109 (15.81) 440 (63.82) 0.0143 0.16 0.481 1.044 0.904 5 (5.3)
C100SH100N15 100.5 (14.58) 560 (81.22) 0.0100 0.16 0.465 1.373 0.781 5 (5.0)
C100S70N25 103 (14.94) 440 (63.82) 0.0204 0.27 0.728 0.784 0.817 4 (4.8)
C100SH70N25 97 (14.07) 560 (81.22) 0.0143 0.26 0.689 0.857 0.698 4 (5.8)
C100S37N40 100 (14.50) 440 (63.82) 0.0385 0.43 1.413 0.862 0.995 4 (7.4)
C100SH37N40 103 (14.94) 560 (81.22) 0.0271 0.43 1.227 0.780 0.769 4 (6.7)

however, were chosen at an early stage of development using ACKNOWLEDGMENTS

Eq. (5) with no limitation on fyh and with P/Ag fc′ instead of kp. The financial assistance provided by the Natural Sciences and Engineering
Research Council of Canada (NSERC) and by the Fonds pour la Formation de
The results in Table 1 show that HSC circular columns chercheurs et l’aide à la recherche of the Government of Quebec (FCAR) is
subjected to different levels of axial load can achieve the gratefully acknowledged.
same level of displacement ductility if they contain amounts
of transverse steel reinforcement that account for the axial NOTATION
load level and the transverse-steel yield strength. For Ach = cross-sectional area of concrete cores measured to outside
diameter of spiral
example, the results revealed that Column C100S100N15, Ag = gross section of concrete
which had been tested under low axial load level (P/Ag fc′ = Ash = total cross-sectional area of transverse reinforcement
15%) and contained 90.4%, 104.4%, and 48.1% of the Ast = total cross-sectional area of longitudinal reinforcement
confining transverse reinforcement proposed by the CSA Ec = secant modulus of elasticity of plain concrete
fc′ = maximum compressive strength of concrete measured on 150 x
standard,16 the New Zealand standard,4 and the ACI Code,3 300 mm (5.90 x 11.81 in.) cylinders
respectively, achieved an idealized displacement ductility of fr = modulus of rupture of concrete
5.3. On the other hand, Column C100S37N40, which had fsu = ultimate strength of transverse reinforcement steel
been under higher axial-load level (P/Ag fc′ = 40%) and fy = yield strength of longitudinal reinforcement steel
H = applied horizontal load
contained 99.5%, 86.2%, and 141.3% of the confining H′ = applied horizontal load H plus equivalent horizontal load due to
transverse reinforcement proposed by the CSA standard,16 the P-Δ effect
New Zealand standard,4 and the ACI Code,3 respectively, hc = cross-sectional dimension of column core measured perpendicular
achieved a comparable idealized displacement ductility of to direction of hoop bars
7.4. Thus, overall, the CSA16 and New Zealand4 standard P = axial load carried by concrete
P0 = nominal axial load of column: P0 = 0.85(Ag – Ast )fc′ + Ast fy
requirements present good correlation with the displacement s = center-to-center spiral spacing
ductility of HSC columns. From the test results, concrete Δ = tip displacement of column
columns designed with these requirements with different Δ2 = maximum tip displacement of column
transverse-steel yield strength and different axial-load levels ΔyI = ideal yield displacement of column
εc′ = axial strain in plain concrete corresponding to fc′
would have adequate sectional ductility. The test results εsh = commencement of strain hardening in steel bars
clearly indicate that the ACI Code3 requirements might be εsu = ultimate strain of reinforcement steel
too conservative for low level of axial load. For example, εy = strain in reinforcement steel at yield strength
Columns C100S100N15 and C100SH100N15 had only less than φ = curvature
φ2 = maximum curvature
half the required volumetric ratio of transverse reinforcement φyI = ideal yield curvature
by the ACI Code3 but achieved an idealized displacement μΔu = ultimate displacement ductility
ductility of 5.0 or more. The volumetric ratios of transverse μφu = volumetric ratio of longitudinal reinforcement in column cross
reinforcement for Columns C100S37N40 and C100SH37N40 section
ρg = volumetric ratio of transverse reinforcement in concrete core
were 40% and 22% more than what is required by the ACI measured center-to-center of ties
Code,3 respectively. These specimens supporting high axial
load presented idealized displacement ductility of REFERENCES
approximately 7.0. 1. Cusson, D., and Paultre, P., “High-Strength Concrete Columns
Confined by Rectangular Ties,” Journal of Structural Engineering, ASCE,
V. 120, No. 3, 1994, pp. 783-804.
CONCLUSIONS 2. Razvi, S. R., and Saatcioglu, M., “Strength and Deformability of
This paper presented a test program aimed at studying the Confined High-Strength Concrete Columns,” ACI Structural Journal, V. 91,
post-elastic behavior of six HSC circular columns under No. 6, Nov.-Dec. 1994, pp. 678-687.
3. ACI Committee 318, “Building Code Requirements for Structural
simulated earthquake loading designed according to the Concrete (ACI 318-08) and Commentary,” American Concrete Institute,
provisions of the 2004 CSA Standard.16 The results of this Farmington Hills, MI, 2008, 465 pp.
test series demonstrate that HSC circular columns can 4. NZS 3101, “Concrete Structures Standard, Part 1—The Design of
achieve adequate ductility if the transverse reinforcement Concrete Structures,” Standards Association of New Zealand, New
Zealand, 1995, 256 pp.
chosen accounts for the axial load level and the transverse- 5. CSA A23.3, “Design of Concrete Structures (CSA A23.3-94),”
steel yield strength. Hence, concrete columns that have Canadian Standard Association, Rexdale, ON, Canada, 1994, 199 pp.
different transverse-steel yield strengths and different axial- 6. Légeron, F., and Paultre, P., “Behavior of High-Strength Concrete
load levels will have adequate sectional ductility. An Columns Under Cyclic Flexure and Constant Axial Load,” ACI Structural
important parameter that is known to influence column Journal, V. 97, No. 4, July-Aug. 2000, pp. 591-601.
7. Azizinamini, A.; Baum Kuska, S. S.; Brungardt, P.; and Hatfield, E.,
ductility is the concrete strength. This parameter has not “Seismic Behavior of Square High-Strength Concrete Columns,” ACI
been studied in this research program. Structural Journal, V. 91, No. 3, May-June 1994, pp. 336-345.

ACI Structural Journal/July-August 2009 403

 8. Li, B.; Park, R.; and Tanaka, H., “Strength and Ductility of Reinforced 19. Sheikh, S. A., and Khoury, S. S., “Confined Concrete Columns with
Concrete Members and Frames Constructed Using HSC,” Research Report Stubs,” ACI Structural Journal, V. 90, No. 4, July-Aug. 1993, pp. 414-431.
No. 94-5, Department of Civil Engineering, University of Canterbury, 20. Tanaka, H.; Park, R.; and McNamee, B., “Anchorage of Transverse
Christchurch, New Zealand, 1994, 373 pp. Reinforcement in Rectangular Reinforced Concrete Columns in Seismic
9. Bayrak, O., and Sheikh, S. A., “Confinement Reinforcement Design Design,” Bulletin of the New Zealand National Society for Earthquake
Consideration for Ductile HSC Columns,” Journal of Structural Engineering,
Engineering, V. 18, No. 2, 1985, pp. 165-190.
ASCE, V. 124, No. 9, 1998, pp. 999-1010.
10. Joint ACI-ASCE Committee 441, “High-Strength Concrete 21. Newmark, N. M., and Hall, W. J., “Earthquake Spectra and Design,”
Columns: State-of-the-Art,” ACI Structural Journal, V. 94, No. 3, May- Technical Report, Earthquake Engineering Research Institute, Berkeley,
June 1997, pp. 323-335. CA, 1980, 103 pp.
11. Sheikh, S. A.; Shah, D. V.; and Khoury, S. S., “Confinement of 22. Park, R., “Evaluation of Ductility of Structures and Structural
High-Strength Concrete Columns,” ACI Structural Journal, V. 91, No. 1, Assemblages from Laboratory Testing,” Bulletin of the New Zealand
Jan-Feb. 1994, pp. 100-111. National Society for Earthquake Engineering, V. 22, No. 3, 1989, pp. 155-166.
12. Cusson, D., and Paultre, P., “Stress-Strain Model for Confined High- 23. Gosain, K. N.; Brown, H. R.; and Jirsa, J. O., “Shear Requirements
Strength Concrete,” Journal of Structural Engineering, ASCE, V. 121, for Load Reversals on RC Members,” Journal of Structural Division,
No. 3, 1995, pp. 468-477. ASCE, V. 103, No. 7, 1977, pp. 1461-1476.
13. Paultre, P.; Légeron, F.; and Mongeau, D., “Influence of Concrete
Strength and Yield Strength of Ties on the Behavior of High-Strength 24. Ehsani, M. R., and Wright, J. K., “Confinement Steel Requirements
Concrete Columns,” ACI Structural Journal, V. 98, No. 4, July-Aug. 2001, for Connections in Ductile Frames,” Journal of Structural Division, ASCE,
pp. 490-501. V. 116, No. 3, 1990, pp. 751-767.
14. Légeron, F., and Paultre, P., “Uniaxial Confinement Model for 25. Sheikh, S. A., and Uzumeri, S. M., “Analytical Model for Concrete
Normal- and High-Strength Concrete Columns,” Journal of Structural Confinement in Tied Columns,” Journal of Structural Division, ASCE,
Engineering, ASCE, V. 129, No. 2, 2003, pp. 241-252. V. 108, No. 12, 1982, pp. 2703-2722.
15. Paultre, P., and Légeron, F., “Confinement Reinforcement Design for 26. Mander, J. B.; Priestley, M. J. N.; and Park, R., “Seismic Design of
Reinforced Concrete Columns,” Journal of Structural Engineering, ASCE, Bridge Piers,” Research Report No. 84-2, Department of Civil Engineering,
V. 134, No. 5, May 2008, pp. 738-749. University of Canterbury, Christchurch, New Zealand, 1984, 442 pp.
16. CSA A23.3, “Design of Concrete Structures (CSA A23.3-04),”
Canadian Standard Association, Mississauga, ON, Canada, 2004, 214 pp. 27. Sharma, U. K.; Bhargava, P.; and Kaushik, S. K., “Behavior of
17. University of Washington, UW-PEER Reinforced Concrete Column Confined High Strength Concrete Columns Under Axial Compression,”
Test Database, www.ce.washington.edu/peera1. Journal of Advanced Concrete Technology, V. 3, No. 2, 2005, pp. 267-281.
18. Saatcioglu, M., and Baingo, D., “Circular High-Strength Concrete 28. Paultre, P., “MNPhi: User Manual,” CRGP Report 2001-01,
Columns Under Simulated Seismic Loading,” Journal of Structural Department of Civil Engineering, University of Sherbrooke, Sherbrooke,
Engineering, ASCE, V. 125, No. 3, Mar. 1999, pp. 272-280. QC, Canada, 2001, 97 pp.

404 ACI Structural Journal/July-August 2009