Short Columns

Journal of Advanced Concrete Technology Vol. 3, No.
2, 267-281, June 2005 / Copyright 2005 Japan Concrete Institute
267
Scientific paper
Behavior of Confined High Strength Concrete Columns under Axial

Compression
Umesh K. Sharma1, Pradeep Bhargava2 and S.K.Kaushik3
Received 12 October 2004, accepted 1 February 2005
Abstract
An experimental study was carried out to investigate the behavior of high strength concrete short columns confined by
circular spirals and square ties under monotonically increasing concentric compression. The test variables included
volumetric ratio, spacing and yield strength of transverse reinforcement, longitudinal reinforcement ratio, lateral steel
configuration, shape of cross section and concrete compressive strength. The effects of these variables on the uniaxial
behavior of high strength concrete columns are presented and discussed The results indicate that more confinement is
required in columns of high strength concrete than in columns of low strength concrete to achieve the desired post-peak
deformability. The behavior of high strength concrete columns is characterized by the sudden spalling of concrete cover,
leading to a loss of axial capacity. A comparative study of existing confinement models of high strength concrete columns was also conducted to assess their capabilities of predicting the actual test behavior. To this end, the stress-strain
curves of the specimens tested in the present study were compared with the ones predicted by the various models. It is
shown that Legeron & Paultre (2003) model estimates the actual experimental curves more closely as compared to the
other models employed in the study.
1. Introduction
Inelastic deformability of reinforced concrete columns is
essential for overall stability of structures in order to
sustain strong earthquakes. Deformability of columns
can be achieved through proper confinement of the core
concrete. It is now well documented that the desired ductility can be attained in case of normal strength concrete
columns by providing well-detailed lateral confinement
reinforcement (Richart et al. 1928; Sheikh & Uzmeri
1980; Park et al. 1982; Mander et al. 1988). However,
the gradual development of concrete technology has
promoted the use of high strength concrete owing to its
wide range of advantages over normal strength concrete
and as a result, concrete strengths much higher than
60Mpa have gained acceptance in the construction industry. The high strength concrete is significantly more
brittle than conventional normal strength concrete. While,
existing code provisions for minimum amount of confinement reinforcement are based on experiences with
normal strength concrete, questions have been asked as
to whether similar amount of confinement is suitable for
high strength concrete columns (Razvi & Saatcioglu
1994; Pessiki & Pieroni 1997; Foster 2001).
The few studies carried out on confinement of high
Research Scholar, Department of Civil Engineering,

Indian Institute of Technology, Roorkee, India.
E-mail:umuksdce@iitr.ernet.in
2
Associate Professor , Department of Civil Engineering,
Indian Institute of Technology, Roorkee, India.
3
Professor, Department of Civil Engineering, Indian
Institute of Technology, Roorkee, India.
strength concrete columns in the recent past indicate that

the confinement steel requirements for such columns
have not yet been fully developed (Bjerkli et al. 1990;
Muguruma et al. 1993; Li et al. 1994; Cusson & Paultre
1994; Razvi & Saatciouglu 1996; Foster & Attard 2001).
This is also evidenced by the lack of design provisions in
the current relevant building codes of various countries.
ACI Committee 441 (1997) also emphasized that more
data on the confinement of high strength concrete columns are needed. Therefore, as a part of ongoing combined experimental and analytical study to fully discover
the strength and ductility of high strength concrete columns, this paper reports the results of concentrically
tested circular and square high strength concrete columns. An attempt has also been made to evaluate the
capabilities of existing stress-strain models for high
strength concrete by comparing the experimentally
found curves with the ones estimated by the various
models.
2. Experimental program
2.1 Test specimens
A total of 44 high strength concrete columns, 600 mm in
length, were tested under
concentric compression.
They included 18 tie confined 150 mm square specimens
and equal numbers of spiral confined 150 mm diameter
circular section columns. The remaining 8 columns were
prepared as 4 square and 4 circular plain (unconfined)
concrete specimens to establish the properties of unconfined concrete and thereby to get comparison of in place
strength of concrete with standard cylinder compressive
strength. To properly investigate the behavior of high
strength concrete columns, the specimens were cast and
268
U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005
faces of the specimens to prevent direct loading of the

bars. In all the specimens, the ratio of gross area of the
section, Ac, to the core area, Acc, measured to the outside
of the lateral reinforcement was 1.33. Longitudinal reinforcement was provided in all the confined column
specimens. The lateral ties of all square specimens had
1350 hooks and a development length as per the ACI 318
code provisions. Fig.1 and Table1 show the various
properties of longitudinal and transverse reinforcement.
The notation fc shows standard cylinder compressive
strength on the day of testing of columns, l and fy are
respectively longitudinal steel ratio and yield strength of
longitudinal steel, s, s and fyh are respectively spacing,
volumetric ratio and yield strength of lateral steel. Two
different specified concrete strengths, two lateral steel
yield strengths and two longitudinal steel ratios and resulting lateral steel configurations were employed. The
spacing of spirals and ties were varied from approximately one fourth to half of the total lateral dimension of
column in order to give varying volumetric ratios of lateral confining steel.
Failure of the specimens was forced in the test region,
which was equal to 300 mm in the middle of the specimen height by reducing the spacing of the lateral reinforcement out side the test region to half of the specified
spacing in the test region. However, in the case of the
specimens where specified spacing of lateral steel was
30 mm, the spacing in the end zones was reduced to only
three fourth of that in test region to facilitate proper placing of concrete in the cover of columns in these end regions. All the specimens were also externally confined
by 10mm thick mild steel collars in the end regions of
150 mm to further prevent premature failure.
tested in duplicate for each case in order to get the average of two results thus making 22 independent column
designs. All the details regarding the various column
specimens are illustrated in Table 1 and Fig. 1. Most of
the results reported here in this study are the averages of
two tests with a few exceptions of SG, SC, CH and SF
columns where duplicate results could not be taken due
to the following reasons. One of the SG specimens failed
rather more suddenly compared to the other and all the
measuring instruments like LVDTs and strain gauges got
dislodged and further readings after peak could not be
taken. In case of one of the CH specimens the top end
concrete cover of the column got crushed and the loads
came directly on the end steel collars which got damaged and therefore the test was stopped. Whereas one of
the results of each of SC and SF specimens were rejected
due to the reason that in spite of exercising extra care to
avoid any eccentric loading of the specimens, these columns seemed to suffer from this problem as evidenced
by their considerable tilting and entire spalling on one
side only.
The specimens were cast in total eleven sets. Each set
consisted of two duplicate circular and two duplicate
square columns of one design. For example two specimens of CA and SA each were cast in one set. The columns were designed to investigate the parameters of
confinement, including volumetric ratio and spacing of
transverse reinforcement, yield strength of transverse
reinforcement, longitudinal reinforcement ratio, and lateral steel configuration in addition to shape of cross section and concrete compressive strength. Concrete cover
of 10 mm was provided in all the confined specimens. A
concrete cover of 15 mm was also provided between the
ends of the longitudinal bars and the top and bottom sur-
150 mm
End Region
250 mm
LVDT
Gauge Length
600 mm
Square Column
150mm
6 Nos
8mm dia.
6 Nos
12mm dia.
150mm
150 mm
4 Nos
12 mm
dia.
150 mm
150mm
Cover 10mm
Cover 10 mm
600 mm
150 mm
End Region
Cover 10mm
Cover 10mm
8 Nos
12 mm dia.
150 mm
150 mm
150 mm
Strain Gauges - Longitudinal Bar

Lateral Steel
Circular Column
Fig. 1 Details of column specimens, reinforcement arrangement and location of strain gauges.
269
Table 1 Properties of column specimens.

Specimens
Circular
Square
f 'c
Longitudinal Reinforcement
Number &
l
fy
Diameter, mm
%
MPa
6 No 8
1.70 412
Diameter
mm
Transverse Reinforcement
s
Arrangement
s
mm
%
A
50
3.3
f yh
ke s f yh / f 'c
MPa
412
0.186
CA
MPa
62.20
CB
62.80
6 No 8
1.70
412
75
2.2
412
0.107
CC
61.85
6 No 8
1.70
412
50
3.3
520
0.236
CD
63.35
6 No12
3.84
395
50
3.3
412
0.189
CPL
61.90
--
--
--
--
--
--
--
--
--
CE
82.50
6 No 8
1.70
412
30
5.5
412
0.256
CF
81.75
6 No 8
1.70
412
50
3.3
412
0.141
CG
83.15
6 No 8
1.70
412
75
2.2
412
0.081
CH
81.80
6 No 8
1.70
412
50
3.3
520
0.178
CI
82.55
6 No12
3.84
395
50
3.3
412
0.145
CPH
82.25
--
--
--
--
--
--
--
--
--
SA
62.20
4 No12
2.01
395
50
3.3
412
0.0984
SB
62.80
4 No12
2.01
395
75
2.2
412
0.0505
SC
61.85
4 No12
2.01
395
50
3.3
520
0.1248
SD
63.35
8 No12
4.02
395
50
5.62
412
0.2303
SPL
61.90
--
--
--
--
--
--
--
--
--
SE
82.50
4 No12
2.01
395
30
5.5
412
0.1483
SF
81.75
4 No12
2.01
395
50
3.3
412
0.0748
SG
83.15
4 No12
2.01
395
75
2.2
412
0.0381
SH
81.80
4 No12
2.01
395
50
3.3
520
0.0944
SI
82.55
8 No12
4.02
395
50
5.62
412
0.1767
SPH
82.25
--
--
--
--
--
--
--
--
--
The fabricated steel cages were fixed in the steel

moulds and the casting was carried out in the laboratory.
After 24 hours, the specimens were removed from the
moulds and submerged in a water tank for curing. The
water-curing period lasted for 28 days, after which the
specimens were left in the laboratory at ambient temperature until the time of testing. The testing was commenced after 90 days of ageing.
2.2 Material properties
Two high strength concrete mixes were employed in the
study. The materials consisted of Ordinary Portland Cement, natural river sand, crushed stone aggregate of
maximum size 10mm, tap water for mixing and curing,
silica fume and superplasticizer admixture to maintain
adequate workability of mix. To obtain the desired
strength, several preliminary mix designs were practiced
and then optimal mix designs were determined after testing 28 days strength. The final mixes had 28 days average cube (150x150x150mm) compressive strengths of
68.4 & 87.5 MPa and average cylinder (100x200mm)
strengths of 58.03 and 76.80 MPa. Table 2 shows the
mix proportions and 28 days compressive strengths for
both the mixes. The measured tangent elastic moduli
(Ec) were 31645 MPa and 34465 MPa for lower and
higher strength concrete mixes respectively. The slump

value ranging from 100-120mm was maintained to ensure that the concrete could be placed through the dense
reinforcement cages.
Three standard concrete cylinders (100 x 200 mm)
were cast along with each set of four columns. They
were tested at various ages to determine the strength of
concrete fc on the day of testing of columns. The properties of unconfined concrete were obtained by testing
plain unreinforced column specimens (CPL, CPH, SPL
& SPH). It was observed that concrete strength of column specimens were generally lower than the concentric
strength measured on standard cylinders. The average
unreinforced specimen concrete strength was measured
as 88% and 90% of the average concrete cylinder
strength for the lower high strength (CPL & SPL) and
upper high strength (CPH and SPH) concretes respectively. The commonly used ratio of 0.85 was then used
for evaluating the concrete section capacity of the
specimens tested in the present study.
The longitudinal reinforcement consisted of 8 mm and
12mm deformed bars for circular columns and 12 mm
deformed bars for square columns. Two different grades
(412 MPa and 520 MPa) of reinforcing steel with a diameter of 8 mm were used as lateral reinforcement.
270
taken to complete each test ranged from 30 minutes to 1

hour depending upon the degree of confinement in the
specimen. To ensure concentric loading, an initial load
of approximately 20% of the total ultimate load was applied and the readings of the four LVDTs were monitored. If the readings of the LVDTs were not approximately equal, this signals that the load is not concentric.
The column was then unloaded and the additional packing was given under the wooden plies unless the load
becomes almost concentric. In spite of the rigorous procedure followed for aligning the specimens, some eccentricities were unavoidable.
Stress-strain relationships, established by performing at

least three coupon tests for each reinforcement, are illustrated in Fig. 2. The yield strength of the lateral reinforcement is defined as an offset strain of 0.2%.
2.3 Instrumentation and testing procedure
Longitudinal and lateral steel strains were measured by
electrical resistance strain gauges glued to the steel bars.
The strain gauges were pasted to the two opposite longitudinal steel bars at their middle lengths. Similarly two
gauges were glued at the two locations on the lateral
steel at approximately middle length of the specimen as
shown in the Fig. 1. The axial displacement of the
specimens was recorded using four linear variable differential transducers (LVDTs). Two LVDTs were attached on the opposite faces to a top and bottom steel
clamps to the specimens to give a gauge length of 250
mm. Whereas, two LVDTs were mounted on the other
two faces of the specimen and were held by the attachments provided to the end steel collars. If a sudden failure occurs or due to spalling, the steel clamps get dislodged, the LVDTs mounted on the end collars could be
used instead of the gauge length over the central 250 mm.
LVDTs were of 50 mm stroke. Loads were recorded
through a 3000 kN load cell. The recorded strain data
from the four LVDTs, four strain gauges and corresponding load data from load cell were fed to a data acquisition system and stored on a computer. Overall view of
the test setup is shown in Fig. 3.
The top and bottom ends of the specimens were
slightly ground by a grinding machine to remove any
surface unevenness. But, at the same time it was ensured
that the excessive grinding is not carried and the required cover is maintained to the longitudinal bars at the
ends. In addition, 6mm wooden ply pieces were put at
the top and bottom ends of the specimens to ensure parallelism of specimen end surfaces and uniform distribution of the load on the specimen. The test specimens
were loaded using a 5000 kN capacity hydraulic universal testing machine with load controlled capabilities. The
monotonic concentric compression was applied at very
slow rate to capture the post-peak part of the measured
load deformation curves by manually controlling the oil
pressure. The load was applied from zero to failure,
which was determined primarily by either rupture of the
lateral reinforcement or excessive crushing of the core,
together with buckling of the longitudinal bars. The time
800
8 m m ( f y h = 5 2 0 M P a , y h = 0 .0 0 2 6 )
700
Stress (MPa)
600
500
8 m m ( f y h = 4 12 M P a , y h = 0 .0 0 2 4 )
400
300
12 m m ( f y = 3 9 5 M P a , y = 0 .0 0 2 3 )
200
100
0
0
0.01
0.02
0.03
0.04
0.05
Strain (m m /m m )
Fig 2 Stress-strain curves for reinforcing bars.
Fig. 3 Experimental setup.
Table 2 Mix proportions.
Mix
Cement
Kg/m3
Water
Kg/m3
Coarse
Aggregate
Kg/m3
Sand
Kg/m3
Silica
Fume
Kg/m3
Superplasticizer
Kg/m3
Mix 1
545
169
1105
700
--
5.45
28 Days
Cylinder
Compressive
Strength*
f c ' , MPa
58.03
Mix 2
600
168
1055
625
50
7.5
76.80
* Average of 5 Specimens.
28 Days
Cube Compressive
Strength*
f ck , MPa
68.40
87.50
0.06
271
3. Observed behavior
velopment of strains associated with the failure of unconfined concrete specimens.

The cover spalling often resulted in a sudden drop in
load, which was also more pronounced in higher concrete strength columns. Load resistance of columns
again increased to a second peak normally for wellconfined specimens only. This behavior indicates that
the passive confinement becomes active only after the
cover has spalled and post-spalling behavior depended
solely upon the confinement level of specimens. Figures
5 and 6 show the total column axial load normalized to
unconfined column capacity versus longitudinal strain
curves for circular and square specimens respectively.
All columns initially behaved in a similar manner and

exhibited relatively linear load deformation behavior in
the ascending part, which is typical of high strength concretes. The plain concrete specimens CPL, CPH, SPL &
SPH, had a sudden explosive type of failure at the
maximum axial load. The complete load deflection
curves for these specimens could not be obtained with
the present testing facility, and no readings could be
taken after their brittle failure. The strains corresponding
to the peak loads were 0.00232, 0.00256, 0.00237 and
0.0026 for CPL, CPH, SPL and SPH specimens respectively.
The general behavior of confined specimens was comparatively ductile and complex unlike plain unconfined
columns. These columns were characterized sequentially
by the development of surface cracks, cover spalling,
yielding of longitudinal steel, yielding of lateral steel,
fracture of spiral or ties, buckling of longitudinal bars
and crushing of core concrete. The vertical longitudinal
cracks were noticed invariably first of all for almost all
the columns just before the cover spalling. These cracks
thereafter eventually led to the spalling of cover concrete,
which was marked by the separation of large pieces of
cover from core concrete. Figuer 4 illustrates the appearance of a typical specimen at various stages of loading. The spalling was comparatively more sudden and
explosive for columns with higher strength concrete mix.
The beginning of cover spalling was carefully observed
and marked during testing. The strain data (Table 3)
indicated that spalling of cover occurred prior to the de-
(a)
(a)
(b(b))
(c)
(c)
Fig. 4 Appearance of specimen CC at various

stages:(a) initial vertical cracking, (b) cover spalling,
and (c) ultimate failure.
Table 3 Test results.

Specimens
Pmax,
kN
CA 1109
CB 1059
CC 1148
CD 1241
CPL 963
Circular CE 1381
CF
1294
CG 1352
CH 1321
CI
1379
CPH 1312
SA 1334
SB
1364
SC
1308
SD 1626
Square SPL 1239
SE
1641
SF
1604
SG 1730
SH 1621
SI
1819
SPH 1684
Axial Loads
Pmax/Po Pc,
Pc/Poc
kN
1.063 986 1.074
1.006 947 1.021
1.106 1024 1.115
1.049 973 1.063
1.035 --1.03
1263 1.04
0.972 1175 0.973
0.980 1206 0.982
0.992 1202 0.995
0.956 1111 0.932
1.062 --0.992 1155 0.990
1.005 1185 1.006
0.977 1133 0.977
1.069 1268 1.090
1.046
--0.951 1463 0.946
0.937 1433 0.935
0.995 1551 0.995
0.946 1446 0.943
0.971 1461 0.964
1.070
---
-- Not applicable, * Could not be measured.
Pcc,
kN
978
796
1011
967
-1262
1069
964
1117
1032
-985
917
1042
1268
-1399
1220
1122
1304
1403
--
Pcc/Pocc
1.624
1.309
1.687
1.630
-1.581
1.351
1.199
1.410
1.335
-1.290
1.189
1.372
1.683
-1.382
1.216
1.10
1.30
1.43
--
c
0.00248
0.00236
0.00241
0.00235
0.00232
0.00228
0.00231
0.00250
0.00240
0.00235
0.00256
0.00232
0.00235
0.00224
0.00263
0.00237
0.00228
0.0022
0.0025
0.00232
0.00252
0.00260
Axial Strains
cc/co c50c
c/co
cc
1.068
1.017
1.038
1.013
1.00
0.890
0.902
0.976
0.937
0.917
1.00
0.978
0.991
0.945
1.109
1.00
0.876
0.846
0.961
0.892
0.969
1.00
0.00596
0.00357
0.00813
0.00638
-0.0074
0.0046
0.0034
0.0050
0.00564
-0.00386
0.00312
0.00427
0.00713
-0.0051
0.00344
0.00285
0.00361
0.0063
--
2.568
1.538
3.504
2.75
-2.891
1.797
1.328
1.953
2.203
-1.628
1.316
1.801
3.008
-1.961
1.323
1.096
1.388
2.423
--
0.0278
0.0110
0.0323
0.0291
-0.0298
0.0148
0.0080
0.0155
0.018
-0.011
0.0055
0.014
0.032
-0.014
0.0064
0.0042
0.0098
0.021
--
c50c/co
11.98
4.741
13.92
12.54
-11.65
5.781
3.12
6.051
7.031
-4.641
2.320
5.907
13.50
-5.384
2.461
1.615
3.769
8.076
--
I10
fhcc
MPa
7.8
7.2
8.58
8.3
-8.8
7.6
5.56
7.64
8.4
-6.63
4.32
6.72
8.4
-6.44
4.54
3.38
5.05
7.86
--
412
*
520
412
-412
*
274
458
394
-394
*
420
412
-395
336
*
*
412
--
272
Relative Column Load (P/P o)
The second peak was observed for many circular columns CA, CC, CD, CE, CH and CI, and a few square
columns SA, SD, SE and SI which have better efficiently
1.2
1
CB
0.8
CD
CC
CA
0.6
0.4
0.2
0
0
0.03
0.06
0.09
0.12
0.15
Colum n Axial Strain (m m /mm )

(a)
1.2
1
CF
0.8
CH
CG
CI
CE
0.6
0.4
0.2
0
0
0.03
0.06
0.09
0.12
0.15
Colum n Axial Strain (mm /m m )

(b)
Fig. 5 (a) & (b) Column axial load versus axial strain
curves (circular columns).
1.2
1
SA
SD
SC
SB
0.8
0.6
0.4
0.2
0
0
0.02
0.04
0.06
0.08
0.1
Colum n Axial Strain (mm /m m)

(a)
1.2
1
SE
SG
SF
0.8
SI
SH
0.6
0.4
0.2
0
0
0.02
0.04
0.06
0.08
0.1
0.12
Colum n Axial Strain (mm /m m )

(b)
Fig.6 (a) and (b) Column axial load versus axial strain
curves (square columns).
confined core than the other specimens of the study as

implied by their respective volumetric ratios, spacing
and distribution of lateral steel. In the case of CE and SD
specimens the load values at second peak were in fact
more than those at respective first peaks. This may be
attributed due to the fact that CE specimen has a very
high volumetric ratio and closer spacing of spirals.
Square column SD had also high volumetric ratio and
the lateral tie arrangement of eight bars, which provided
better confinement than an arrangement of four bars.
However, the SG, CG and SB specimens failed immediately after the first maximum load was reached, indicating that the spacing of the lateral steel was too wide to
provide lateral confinement. Overall, it was observed
that for the similar confinement circular columns had a
better post peak behavior than the square tie confined
columns and the lower high strength mix columns failed
slowly in a better ductile fashion than the similarly confined higher concrete strength mix columns.
At the stage of cover spalling, in all the specimens
longitudinal steel yielded but, the level of stress in lateral
confining steel was considerably less than their respective yield strengths. Similar observations were made in
some earlier studies also (Cusson & Paultre 1994; Foster
1999). Lateral steel yielded either at the second peak or
even after that in the descending part of loaddeformation behavior. In fact yielding of lateral steel at
second peak was noticed only for well-confined columns
and mostly for lower high strength mix specimens. They
were CA, CC, CD, and CE circular columns and SD and
SI square column. Fracture of transverse reinforcement
was observed only for the specimens with closer spacing
of spirals or ties i.e. 30 and 50 mm. The stage of spiral or
tie fracture was invariably at the moment when the load
dropped to approximately 40-50% of the maximum
value in the descending segment. It was observed that
the fracture of lateral steel caused a sudden drop in load.
In the case of columns with 75 mm spacing of transverse
steel, buckling of longitudinal steel was observed prior
to the fracture of lateral steel, which ultimately lead to
the fracture of ties and crushing of core.
4. Test results
The columns were analyzed to obtain the stress-strain
curves of confined concrete, as suggested by Sheikh &
Uzumeri (1980) and Cusson & Paultre (1994). The concrete contribution Pc at a certain deformation was determined by subtracting the contribution of longitudinal
steel from the applied load P. The load carried by the
longitudinal steel was determined from the stress-strain
curves obtained from the tension test. In computing the
load carried by the longitudinal steel, the strain hardening regions of the curves were assumed to be a straight
line that resulted in a maximum error of less than 5% in
the computed concrete force. The concrete contribution
curves were nondimensionalized with respect to gross
concrete area force Poc and core concrete area force Pocc
273
(Fig. 7), where
(1)
Pocc = 0.85 f ' c Acc
(2)
While the gross concrete area Ac represented column

behavior before the cover started spalling, only the core
area Acc resisted the applied load after concrete cover
was completely spalled. A smooth transition was assumed to take place from the lower curve to the upper
curve. In Fig. 7, dark line shows the behavior of confined concrete. The load sustained by the confined core
concrete Pcc was calculated by subtracting the load carried by the cover concrete from Pc. The maximum loads
Pmax, Pc and Pcc were normalized by Po, Poc and Pocc respectively and these values were compared with each
other as shown in Table 3, where
Po = 0.85 f ' c ( Ag Ast ) + f y Ast
(3)
The ratio Pmax/ Po ranges from 0.95 to 1.10 for circular

columns and 0.93 to 1.07 for square columns. The lower
ratios are observed for specimens with higher strength
concrete and closer spacing of lateral steel indicating
loss in column capacity due to spalling. Similarly Pc/
Poc varies from 0.93 to 1.11 for circular columns and
0.93 to 1.09 for square columns. The strain ratio c/co,
ranges from 0.89 to 1.07 for circular columns and 0.84
to 1.10 for square columns indicating premature failure
of cover for higher strength concrete columns with dense
lateral steel spacing. Here c stands for axial strain corresponding to the first peak i.e. beginning of spalling and
co is peak strain of unconfined concrete column. The
load ratio Pcc/ Pocc ranges from 1.19 to 1.68 for circular
columns and 1.10 to 1.68 for square columns. The ratio
of peak confined strain cc to co varies from 1.32 to 3.5
for circular columns and 1.09 to 3.0 for square columns.
The strain ratio c50c/co ranges from 3.12 to 13.92 for
circular columns and 1.61 to 13.50 for square columns.
Where, c50c is axial strain corresponding to 0.5 Pcc in the
descending part. The values of ratios Pcc/Pocc, cc /co and
c50c/co prove that the important gains in deformability
of core concrete can be achieved following the spalling
of the cover concrete depending upon the level of confinement.
To quantify the effect of test variables on the postpeak deformability of confined high strength concrete
columns more effectively I10 ductility index can be used
(Foster 1999; Foster & Attard 2001). The definition of
I10 ductility index is adopted from ASTM C1018 (26) for
the measurement of toughness. The I10 index is the ratio
of the area under the load versus strain curve up to a
strain of 5.5 times the yield strain to the area under the
curve for a strain equal to the yield strain. The yield
strain is taken as 1.33 times the strain corresponding to a
load of 0.75 times the maximum load on the ascending
side. The significance of taking area under the curve up
to 5.5 times the yield strain is that for a perfectly elasto-
P c/P oc, P c/P occ
Poc = 0.85 f ' c Ac
P c /P occ
1.8
1.6
1.4
Behavior of
confined concrete
1.2
P c /P oc
0.8
A : B eginning o f spalling
B : A fter spalling o f co ver
C : M aximum co nfined stress and beginning o f
0.6
0.4
0.2
co re crushing.
0
0
0.01
0.02
0.03
Concrete axial strain, c
Fig.7 Nondimensionalized concrete contribution curve.
plastic material I10 = 10, while for a perfectly elasticbrittle material I10 = 1. The I10 values were computed for
all the specimens of the study and the same are given in
Table 3.
5. Effect of test variables

5.1 Concrete compressive strength
The concrete compressive strength was the most important and one of the primary variables investigated extensively in the test program. The behavior of specimens
with same volumetric ratio, spacing, configuration and
yield strength of lateral steel but with different concrete
strengths was compared to quantify the effects of this
parameter. Fig. 8 shows comparisons of both circular
and square columns with different strength concretes.
The post peak curves of the higher strength concrete mix
columns are steeper indicating faster rate of strength
decay as compared to the lower strength concrete specimens. The decreasing trend of strength enhancement
(Pcc/ Pocc) and I10 ductility values for both circular and
square columns from lower strength concrete mix
specimens to higher strength mix specimens show that
effectiveness of confinement decreases as the concrete
strength increases (Table 3). For example, if similarly
confined CC (61.85 MPa) and CH (81.80 MPa) specimens are compared, the percentage increase in strength
reduced from 68 to 41% and I10 ductility decreased from
8.58 to 7.64 from former to latter. Therefore, if the same
levels of strength and ductility enhancements are desired,
higher strength concrete columns shall require more confinement than lower strength concrete columns.
5.2 Volumetric ratios and spacing of lateral steel
The importance of the amount of lateral confining steel
as a factor that affects the behavior of confined concrete
is well recognized. An increase in the volumetric ratio of
confinement steel may be directly translated into a proportional increase in lateral confining pressure. Also, the
spacing of transverse reinforcement is an important parameter that affects the distribution of confinement pressure on the confined core in addition to the stability of
274
longitudinal bars. Fig. 9 demonstrates the effect of the

volumetric ratio of confining reinforcement on the behavior of a few confined high strength concrete columns
of the study. These columns also have different spacing
of lateral steel spirals or ties. As expected, the larger the
volumetric ratio or closer the spacing of lateral steel, the
more ductile is the behavior of columns. The columns
with low volumetric ratio or increased spacing of lateral
steel exhibit brittle behavior, showing faster rate of
CA
62.20 MPa
1.6
1.4
Pc/Poc, Pc/Pocc
Relative Concrete Load ,
1.8
CD
63.35 MPa
CB
62.80 MPa
1.2
1
0.8
0.6
CF
81.75 MPa
0.4
0.2
CI
82.55 MPa
CG
83.15 MPa
0
0
0.02
0.04
0.06
0.08
0.1
Axial Strain
Axial
Strain (mm/mm)
(m m /m m )
(a)
(a)
1.8
Relative Concrete Load ,

P c/P oc, P c/P occ
1.6
SD
63.35 M Pa
1.4
SA
62.20 M Pa
1.2
SB
62.80 M Pa
1
0.8
SI
82.55 M Pa
0.6
0.4
SF
81.75 M Pa
0.2
0
0
S G 8 3 .15 M P a
0.02
0.04
0.06
0.08
0.1
Axial Strain (m m /m m )
(b)
Fig. 8 Effect of concrete strength on the behavior of confined concrete: (a) Circular columns & (b) Square columns.
strength decay after the peak. An increase of 32 % in

strength, 270% in strain ductility (c50c/co) and 58.27 %
in I10 ductility was noticed for the higher concrete
strength mix circular columns as the volumetric ratio of
spirals increased from 2.2% to 5.5%. Similarly 25% gain
in strength, 236% enhancement in strain ductility and
90% increase in I10 ductility were observed for higher
strength concrete mix square columns due to the increase
in volumetric ratio of lateral ties from 2.2% to 5.5%.
5.3 Yield strength of transverse steel
The effect of yield strength of transverse confining steel
was investigated by comparing results of two pairs each
of circular and square columns. The compared specimens of each pair had the same concrete strength as well
as the same volumetric ratio, arrangement, and spacing
of lateral steel but different yield strengths (Fig. 10). In
general, it is observed that as the yield strength is increased from 412 MPa to 520 MPa, the strength and
deformability of confined concrete get improved. However, for the range of steel grades used in this study, the
results show that the effect of increasing yield strength
of lateral steel has not very significant effect on the behavior of high strength concrete columns. The maximum
enhancements of 5% in strength and 10% in I10 ductility
were observed. This may be due to the reason and as
mentioned in the earlier section also, that the transverse
steel of only well confined specimens reached its yield
value at peak-confined stress of concrete and that too for
mostly lower concrete strength circular columns only.
The values of stress, fhcc, in lateral steel at peak confined
stress of different specimens are given in Table 3. The
passive confinement pressure is generated from the tensile forces that develop in the lateral confining steel.
However, tensile stress in transverse steel is generated as
a result of lateral expansion of the core concrete, which
in turn is dependent upon the characteristics of concrete.
The higher strength concrete exhibits less lateral expansion than lower strength concrete under axial loads due
to its higher modulus of elasticity and its lower internal
micro cracking. Therefore, the advantages of providing
higher yield strength lateral steel in high strength concrete columns can be realized only if considerably higher
2
1.6
SE
s = 30 mm
s = 5 .5 %
1.4
1.2
1
Relative Concrete Load

P c/P oc, P c/P occ

P c/P oc, P c/P occ
CE
s = 30 mm
s = 5 .5 %
1.8
SF
s = 50 mm
s = 3 .3 %
0.8
0.6
0.4
CG
s = 75 mm
s = 2 .2 %
0.2
0
0
0.01
CF
s = 50 mm
s = 3 .3 %
0.02
SG, s = 75 mm
s = 2 .2 %
0.03
0.04
0.05
0.06
1.8
CC
1.6
f yh = 520 MPa
1.2
0.07
Fig. 9 Effect of volumetric ratio and spacing of lateral steel.
SC
SH
f y h = 520 MPa
f yh = 520 MPa
1
0.8
CA
f yh = 412 MPa
0.6
0.4
SA
CF
0.2
Axial Strain (m m/mm )
CH
f yh = 520 M Pa
1.4
f y h = 412 MPa
f y h = 412 MPa
0.01
0.02
0.03
0.04
0.05
SF
f yh = 412 MPa
0
0.06
0.07
0.08
0.09
0.1
0.11
Fig. 10 Effect of yield strength of lateral steel.
0.12
275
5.5 Amount of longitudinal reinforcement

Fig. 12 shows the two different pairs of circular specimens, and for each matched pair, two specimens varying
only in their ratios of longitudinal reinforcement are
compared. It may be mentioned here that this type of
comparison was not possible in case of square columns
because, in addition to amount of longitudinal steel,
volumetric ratios of lateral ties were also varying for any
two comparable specimens. The results show (Table 3)
that the strength and deformability gains are equal to
0.4% and 6.4% for the pair CA-CD and 1.2% and 11%
for the pair CF-CI, respectively. Therefore, based upon
the results it can be said that the amount of longitudinal
steel has only little effect on the behavior of confined
concrete, which might be because of the reason that larger longitudinal bar diameters prevent their premature
buckling.
5.6 Section geometry
It is well established now that circular spirals are more
effective in confining concrete than rectilinear ties owing
to their better uniform distribution of lateral confining
pressure around the core compared to the case of square
or rectangular ties (Mander 1988; Razvi & Saatcioglu
1996). In the present study this variable was investigated
extensively to quantify its effect in case of confinement
1.6

Pc/Poc, Pc/Pocc
1.4
1.2
SI
1
0.8
SE
0.6
0.4
0.2
0
0
0.01
0.02
0.03
0.04
Fig. 11 Effect of configuration of lateral ties.

1.8

P c/P oc, P c/P occ
5.4 Configuration of lateral steel

The transverse steel configuration and the resulting distribution of the longitudinal steel play a significant role
in the confinement of concrete. If the lateral confining
pressure applied by the transverse reinforcement on concrete is well distributed around the perimeter of the core
concrete, the efficiency of confinement is improved. In
this study it was possible to observe the effect of varying
the lateral steel configuration by comparing the behavior
of square column specimens SE and SI, which have almost, equal volumetric ratio of lateral steel (Fig.11). The
specimen SE was constructed using single perimeter tie
(4 bar arrangement), and SI specimen had double tie
configuration (8 bar arrangement). The ductility indices
(I10) of the specimens SE and SI were 6.44 and 7.86,
respectively indicating a 22% improvement in ductility
of specimen SI which has better steel configuration.
However, the ratio Pcc/Pocc is equal to 1.382 and 1.43 for
specimens SE and SI respectively showing strength enhancement of only 3.5%. This indicates that the improvements in deformability are more marked than in
strength due to better configuration of transverse steel.
Therefore, according to the test results, transverse steel
configuration and the resulting distribution of longitudinal steel has a considerable effect on the ductility of the
confined concrete but insignificant effect on its strength.
of high strength concrete columns. However, only a few

representative comparisons are shown in Fig. 13 with
respect to this parameter. The specimens of each
matched pair have similar concrete strength and volumetric ratios of lateral steel. It is clear from this figure
that circular confinement is substantially more effective
than rectilinear confinement in case of high strength concrete also as evidenced by the margin of increase in
1.6
CD
l = 3.84
1.4
1.2
CI
l = 3.84 %
CA
l = 1.7 %
0.8
0.6
0.4
CF
0.2
l = 1.7 %
0
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
Fig. 12 Effect of amount of longitudinal steel.

2
1.8

P c/P oc, P c/P occ
amount of confinement is provided to generate enough

lateral expansion of core so that yield strength of high
yield strength lateral steel is fully developed prior to
appreciable post-peak strength decay.
1.6
1.4
CC
1.2
CE
1
0.8
SC
0.6
SE
0.4
0.2
0
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
Fig. 13 Effect of section geometry on confined concrete.
strength and ductility from square to circular columns.

The square columns SC and SE failed in a comparatively
brittle fashion at 0.0042 and 0.0051 axial strains respectively, while the matched circular columns CC and CE
were able to develop axial strains of 0.0081 and 0.0074
respectively without any strength decay. The maximum
gains of 28% in confined strength (CC vs SC) and 66%
in I10 ductility (CB vs SB) were noticed due to the
change in section geometry. Therefore, if the same percentages of strength and ductility enhancements are desired, square or rectilinearly confined columns are required to be confined more vigorously than circular columns.
6. Design implications for ductility

As a result of some field observations regarding the performance of columns after major earthquakes, it is now
well established that despite following the strongcolumn weak-beam concept in design, damage could
occur at ends of the columns. Therefore, ductile detailing
of these potential plastic hinge regions of columns becomes even more important when high strength concrete
is employed. One of the ways in which most of the
building codes of various countries ensure ductility in
concrete columns, is by specifying the amount of confining transverse reinforcement in critical regions of columns. However, these code equations specifying the
minimum amount of transverse steel for columns are
empirical and have been developed based upon the experimental data obtained from the testing of normal
strength concrete columns. An attempt has been made in
the present study to work out the confinement requirements of high strength concrete columns, based upon the
present test results. Ideally, the behavior of columns
should be studied under combined axial compression
and bending to assess the ductility performance of confined reinforced concrete and the effect of cyclic loading
should be taken in to account. But, in this study uniaxial
load-strain data of confined concrete column sections
under monotonic loading has been analyzed, and this
should give a reasonable assessment in the first instance
considering it to be an extreme case.
A number of past studies (Sheikh & Uzmeri 1980;
Mander et al 1988, Razvi & Saatcioglu 1994, Foster &
Attard 2001) have indicated that ductility is a function of
effective confinement index k e s f yh / f ' c , where ke is a
confinement effectiveness parameter, which accounts for
configuration of lateral steel and resulting longitudinal
steel distribution. However, the coefficient ke has been
ignored by the present seismic code requirements as indicated by the following relevant expressions of ACI 318
code for rectilinearly confined columns:
s f yh / f ' c 0.18
s f yh / f ' c 0.6(
(4)
Ag
Ac
1)
(5)
In Fig. 14, the I10 ductility index for all the specimens
is plotted against the effective confinement parameter k e s f yh / f ' c . The ke for both spirally confined circular columns and tie confined square columns was
computed using the procedure suggested by Mander et al
(1988). The computed values of confinement index
k e s f yh / f ' c for the various confined specimens are
given in Table 1. In a few columns (CG, SB, SF and SG)
tested in this study, a faster rate of strength decay was
observed after peak. For these columns, an actual value
of I10 ductility could not be obtained. Therefore, for such
cases average values of I10 were calculated by obtaining
the upper and lower bounds (Foster & Attard 2001). A
best fit relationship between I10 and k e s f yh / f ' c was
found from the plot and is given by:
I 10 = 2.89 ln(1000k e s f yh / f ' c ) 0.45
(6)
Foster & Attard (2001) analyzed the test data of 40

concentrically loaded columns tested by Razvi & Saatcioglu (1996) and computed I10 values for all the specimens. For the specimens with I10 < 8 sudden changes in
the load strain data and faster rate of post-peak strength
decay were recorded. However, columns with I10 > 8
exhibited ductile post-peak curves. Therefore, the authors based upon the evaluation of data concluded that
for the regions of moderate seismicity desirable ductility
could be achieved if I10 is more than 8. An analysis of
present test data also reveals that the columns with I10 >
8 showed ductile post-peak load-strain curves with
slower rate of strength decay. So if the same yardstick is
applied on the present test data to achieve a moderate
seismic resistance in terms of ductility, the following
relationship could be derived after substituting the value
of I10 as 8 in equation (6):
k e s f yh / f ' c 0.136
(7)
The equation (7) can be used to design the critical

10
9
8
Ductlity Index I 10
276
7
6
5
4
3
2
1
0
1
10
30
100
Effective Confinem ent Index, k e s f yh /f' c (%)
Fig. 14 I10 ductility index versus effective confinement index.
hinge regions of high strength concrete columns. It may

be mentioned here that equation (7) specifies the minimum effective confinement index or ratio for high
strength concrete columns, which includes confinement
effectiveness coefficient (ke <1) unlike equations (4) and
(5) of ACI 318 provisions.
7. Assessment of existing stress-strain

confinement models
The analysis of structural members requires an analytical
model for the full stress-strain relationship of concrete in
compression both in confined and unconfined states. The
analytical models for confined normal strength concrete
based on extensive experimental data are well established. These models, based on the test results of normal
strength concrete columns, might be inadequate for high
strength concrete columns, which possess a less ductile
stress-strain behavior. So the normal strength concrete
models if applied to high strength concrete shall overestimate the ductility. As a result a number of such models
specifically for high strength concrete columns have also
been proposed in the recent past. The following section
provides an overview of these analytical confinement
models that cover high strength concrete with strengths
more than 60 MPa.
Yong, Nour and Nawy (1988)
Yong, Nour and Nawy proposed a model for rectilinearly confined high strength concrete columns. They
tested 24 square prisms that were made of high strength
concrete with compressive strength ranging from 83.6 to
93.5 MPa, confined with square ties with yield strength
of 496 MPa. The variables considered were volumetric
ratio of lateral ties, concrete cover and distribution of
lateral steel. A three-part stress-strain relation was proposed to predict the constitutive behavior of confined
high strength concrete.
Bjerkli, Tomaszewicz and Jansen (1990)
Bjerkli, Tomaszewicz and Jansen proposed a threepart stress-strain curve for high strength concrete columns for both circular and rectilinear cross sectional
shapes based upon their test results. They tested a large
number of plain and confined high strength concrete
columns with compressive strength ranging from 65 to
115 MPa. Both cross sectional shapes i.e. cylinders (150
mm diameter and 500 mm high) and prisms
(150x150x500mm and 300x500x2000mm) were included. The test specimens contained longitudinal steel
but no concrete cover.
Nagashima, Sugano, Kimura and Ichikawa (1992)
Twenty-six prism specimens (225x716mm) of high
strength concrete of strengths 59 and 118 Mpa and laterally reinforced with ties of yield strengths 784 and 1374
MPa were tested by Nagashima, Sugano, Kimura and
Ichikawa. Based on the test results a two-part stressstrain relationship was proposed for confined high
strength concrete columns. The variables taken in to
account were concrete strength, yield strength of lateral
277
steel, tie configuration and spacing of lateral steel.

Muguruma, Nishiyama and Watanabe (1993)
Muguruma, Nishiyama and Watanabe proposed a
three-part stress-strain model for confined concrete
based on their own previous studies. A wide range of
concrete strength ranging from 40 to 130 MPa was covered. They tested small square specimens confined laterally by square helix hoops of different yield strengths
and with various volumetric ratios. The yield strengths
of the hoops ranged from 161 to 1353 MPa.
Li, Park and Tanaka (1994)
Li, Park and Tanaka proposed a three-part stress-strain
model for confined high strength concrete based on their
experimental results. Forty reinforced concrete short
columns of both cylindrical (240x720 mm) and square
(240x240x720mm) cross sectional shapes were tested.
The main parameters were in place concrete strength
(35.2 to 82.5 MPa) and lateral steel grade (445 and 1318
MPa) in addition to other parameters like spacing, volumetric ratio and configuration of lateral steel.
Cusson and Paultre (1995)
Cusson and Paultre developed a confinement model
for high strength concrete on the basis of test results of
50 large-scale high strength concrete tied columns tested
under concentric loading. Out of them, 30 HSC tied columns (235x235x1400 mm) were tested by authors themselves and 20 HSC tied columns (225x225x715 mm)
were tested by Nagashima et al. (1992). The concrete
compressive strengths of the specimens ranged from 60120 MPa. The ties with yield strength from 400 to 800
MPa were used. The proposed model takes in to account
tie yield strength, tie configuration, transverse reinforcement ratio, tie spacing, and longitudinal reinforcement ratio. The two-part stress-strain relationship included separate expressions for ascending and descending parts.
Razvi and Saatcioglu (1999)
Razvi and Saatcioglu proposed a model for confined
normal and high strength concrete columns using extensive test data of authors own test results as well as the
experimental results of other research studies. This included the test results of nearly full size specimens of
different shapes, sizes, reinforcement configurations, tie
yield strength (400 to 1387 MPa) and concrete strengths
(30 to 130 MPa). The parameters incorporated in the
model were type, volumetric ratio, spacing, yield
strength, and arrangement of transverse reinforcement,
distribution and amount of longitudinal steel as well as
concrete strength and section geometry. The two part
stress-strain model proposed by the authors was in the
form of the ascending parabolic branch up to peak and a
linear descending branch up to 20% of the peak stress.
Legeron and Paultre (2003)
Legeron and Paultre proposed a stress-strain confinement model for normal and high strength concrete columns based on the large number of test results of circular,
square and rectangular columns tested under various
research studies that included the studies undertaken by
cular specimens. In the case of square columns Razvi &

Saatcioglu (1999); Li et al. (1994); Bjerkli et al. (1990)
and Yong et al. (1988) models always overestimated the
experimental stress-strain curves by a considerable margin for all the specimens. The model proposed by Muguruma et al. (1993) produced considerably steeper stressstrain curves for all the specimens that were far below
the actual test curves. Whereas, the models proposed by
Cusson & Paultre (1995) and Nagashima et al. (1992)
continuously underestimated the test curves though; the
predictions were quite close in the latter case for few
specimens. As in the case of circular columns, Legeron
& Paultre (2003) model was also able to produce close
predictions of stress-strain behavior for almost all the
square columns except for few cases like SH specimen
where it slightly overestimated the experimental stressstrain curve. The ability of Legeron & Paultre (2003)
model to consistently follow the actual test behavior for
both circular and square columns undoubtedly proves its
superiority over the other models of the study. Therefore,
the present study concludes that this model can be employed to predict the uniaxial response of high strength
concrete columns with a reasonable degree of accuracy.
120
Razvi
Legero n
B jerkli
100
Experimental.
Li
80
Stress (MPa)
themselves and a number of other researchers. The concrete compressive strengths ranged from 20 to 140 MPa
and tie yield strengths ranged from 300 to 1400 MPa.
The model incorporates almost all the parameters of confinement. The stress strain relationship was basically
same as proposed by Cusson and Paultre (1995), but the
parameters of the model were recalibrated on the basis
of large number of test data collected by the authors.
A critical review of the above models indicate that
most of them have limited validity in terms of concrete
strengths, column geometry, transverse reinforcement
yield strength and loading conditions. With the exception
of the models proposed by Razvi & Saatcioglu (1999);
Legeron & Paultre (2003); Li et al. (1994) and Bjerkli et
al. (1990), which cover both circular and rectilinear sections, all other models are applicable to only square or
rectilinear shapes. The models proposed by Yong et al.
(1988); Nagashima et al. (1992); Li et al. (1994) have
limited applicability for wide concrete strength ranges. It
has been proved experimentally in the present study and
in many earlier studies (Cusson & Paultre 1994; Razvi &
Saatcioglu 1996; Foster 1999) that for high strength concrete columns, lateral-confining ties may not yield when
peak of confined concrete stress-strain is reached. But,
most of the models use lateral steel yield strength to calculate lateral confining pressure at peak-confined
strength. Only Cusson & Paultre (1995); Razvi & Saatcioglu (1999) and Legeron & Paultre (2003) have incorporated this fact into their respective models by proposing procedures to calculate actual tie stress at peak of
confined stress-strain response. Li et al. (1994) has also
accounted for this indirectly by suggesting different expression for confined strength when higher grades of
lateral steel are to be used, but no explicit expression for
finding the actual tie stress at peak was proposed. There
is hardly any one model, which takes into account all the
loading conditions namely monotonic, cyclic, strain rates
and eccentric loading into account.
To investigate the relative performance of the various
proposed analytical models (as listed above) with regards to their capabilities of predicting the experimentally observed stress-strain profile, stress-strain curves of
the test specimens of present study were compared with
the ones predicted by the various models. It may again
be mentioned here that all the eight confinement models
of the study are applicable to square sections whereas
only four namely Razvi & Saatcioglu (1999); Legeron &
Paultre (2003); Li et al. (1994) and Bjerkli et al. (1990)
can be applied to circular columns. Figures 15 to 22
illustrate the comparisons of the experimental and predicted stress-strain curves of a few representative test
specimens only. The comparisons for circular columns
indicate that Razvi & Saatciouglu (1999) and Li et al.
(1994) models consistently overestimate the actual test
behavior, while Bjerkli et al. (1990) model underestimates the test curves. Legeron & Paultre (2003) model
closely follows the experimental stress-strain curves
though, with a slight overestimation, for most of the cir-
60
40
20
CA
0
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
Strain (m m /m m )
Fig. 15 Comparison of confined axial stress- strain

curves for CA specimen.
Razvi
Legero n
B jerkli
120
Experimental.
Li
100
Stress (MPa)
278
80
60
40
CC
20
0
0
0.01
0.02
0.03
0.04
Strain (m m /m m )
Fig. 16 Comparison of confined axial stress-strain curves

for CC specimen.
279
120
160
Razvi
Legero n
B jerkli
140
Experimental
Li
120
Cusso n
Experimental
Li
M uguruma
80
Stress (MPa)
Stress (MPa)
Nagashima
Razvi
Legero n
B jerkli
100
100
80
60
60
40
40
20
SC
CE
20
0
0
0.01
0.02
0.03
0.04
0.005
0.01

for CE specimen.
120
120
0.02
0.025
0.03

for SC specimen.
140
Razvi
Legeron
B jerkli
0.015
Strain (m m /m m )
Strain (m m /m m )
Experimental
Li
100
Nagashima
Cusson
Razvi
Experimental
Legeron
Li
Bjerkli
M uguruma
Yong
80
Stress (MPa)
Stress (MPa)
100
80
60
40
60
40
20
20
SD
CI
0.01
0.02
0.03
0.04
0.01
Strain (m m /m m )
0.02
0.03
0.04
Strain (m m /mm )

for CI specimen.

for SD specimen.
120
100
Nagashima
Razvi
Legero n
B jerkli
80
Cusso n
Experimental
Li
M uguruma
100
Nagashima
Cusson
Razvi
Experiment al
Bjerkli
Legeron
M uguruma
Li
Yong
Stress (MPa)
Stress (MPa)
80
60
40
20
60
40
20
SA
SH
0
0
0.005
0.01
0.015
0.02
0.025
Strain (m m /m m )
0.005
0.01
0.015
0.02
0.025
Strain (m m /m m )

for SA specimen.

for SH specimen.
8. Conclusions
umns subjected to concentric axial compression. The test

variables are volumetric ratio and spacing of transverse
reinforcement, yield strength of transverse reinforcement,
longitudinal reinforcement ratio, lateral steel configura-
This paper presents the results of spirally confined circular and tie confined square high strength concrete col-
280
tion, shape of cross section and concrete compressive

strength. A comparative study of existing stress-strain
models for high strength concrete columns is also reported. The following conclusions can be made based
upon this study.
1. High strength concrete columns show brittle behavior
unless the columns are confined with sufficient lateral
reinforcement that can provide adequately high lateral
confining pressure. A consistent decrease in strength
enhancement and deformability of columns is observed with increasing concrete strength. Therefore, a
higher degree of confinement is required in columns
with higher concrete strength than in a column with
lower concrete strength to achieve similar advantages.
2. High strength concrete columns suffer from the problem of premature cover spalling which reduces the
column load carrying capacity to even less than unconfined strength if sufficient confinement is not provided. The post spalling behavior depends very much
on the degree and efficiency of confinement.
3. Among the test variables studied, volumetric ratio and
spacing of lateral steel has a more pronounced effect
on the behavior of confined columns than the other
parameters like yield strength of lateral steel, longitudinal steel ratio and configuration of lateral steel
though, improvement in each of the variables considered, translated into enhancements in strength and
ductility. However, increasing the yield strength of
lateral steel seems to give benefits only when a column has high volumetric ratio and efficient arrangement of lateral steel.
4. The ductility of columns is shown to be dependent on
effective confinement index k e s f yh / f ' c . It is concluded that for regions of moderate seismicity, the
critical hinge regions of high strength concrete columns should have effective confinement index of
more than 0.136 to achieve adequate ductility.
5. A comparative study is undertaken to evaluate the
capabilities of the various confinement models of high
strength concrete columns to predict the actual experimental behavior. The study indicated that almost
all the models are able to estimate correctly ascending
part of stress-strain curve. But, there are wide variations in the prediction of the post-peak part of stressstrain curves, with a few models underestimating and
a few overestimating the test behavior except the
Legeron and Paultre (2003) model, which consistently
predicted experimental results with least amount of
discrepancy. Therefore, the present study concludes
that the Legeron and Paultre (2003) model can be
used to analytically predict the uniaxial response of
high strength concrete columns with a reasonable degree of accuracy.
Notations
Ag = gross area of column cross section including area
of longitudinal steel bars.
Ac = gross area of concrete in the section
Acc = core concrete area

Ast = sectional area of longitudinal reinforcement
Ec = tangent modulus of elasticity of concrete
fc = cylinder compressive strength of concrete
fyh = yield strength of tie steel
fy
= yield strength of longitudinal steel
fhcc = stress in lateral steel at peak stress of confined
concrete
Pmax = maximum applied load on the column
Pc
= concrete load corresponding to first peak
Pcc = peak confined concrete load (corresponding to
second peak)
Po = theoretical load carrying capacity of column including longitudinal bars
Poc = theoretical capacity of concrete in the column
Poc c = theoretical capacity of concrete core
s
= spacing of lateral steel
s = volumetric ratio of lateral steel
l = volumetric ratio of longitudinal steel
co = unconfined strain of concrete column
c50o = axial strain at which the stress drops to 50% of
peak in unconfined concrete
= axial strain corresponding to the first peak, Pc

cc = axial strain at peak confined load, Pcc.
c50c = axial strain at which the stress drops to 50% of
peak in confined concrete
yh = yield strain of reinforcement.
References
Bjerkli, L., Tomaszewicz A. and Jansen J. J. (1990).
Deformation properties and ductility of high strength
concrete. Proceedings of Second International
Symposium on Utilization of High Strength Concrete,
Berkley, California, 215-238.
Cusson, D. and Paultre, P. (1994). High strength
concrete columns confined by rectangular ties. ASCE
Journal of Structural Engineering, 120 (3), 783-804.
Cusson, D. and Paultre, P. (1995). Stress-strain model
for confined high strength concrete. ASCE Journal
of Structural Engineering, 121 (3), 468-477.
Foster, S. J. (1999). Design and detailing of high
strength concrete columns. Research Report No. R375, University of New South Wales, Sydney,
Australia.
Foster, S. J. (2001). On behavior of high strength
concrete columns: cover spalling, steel fibers, and
ductility. ACI Structural Journal, 98 (4), 583-589.
Foster, S. J. and Attard, M. M. (2001). Strength and
ductility of fiber reinforced high strength concrete
columns. ASCE Journal of Structural Engineering,
127 (1), 281-289.
Joint ACI-ASCE Committee (1997). High Strength
Concrete Columns: State of the Art. Report No. ACI441R-96, ACI Structural Journal, 94 (5), 323-335.
Li, B. Park, R. and Tanaka, H. (1994). Strength &
ductility of reinforced concrete members and frames
constructed using high strength concrete. Research
report No. 94-5, University of Canterbury,
Christchurch, New Zealand.

Legeron, F. and Paultre, P. (2003). Uniaxial
confinement model for normal and high strength
concrete columns. ASCE Journal of Structural
Engineering, 29(2), 241-252.
Mander, J. B., Priestly, M. J. N. and Park, R. (1988).
Theoretical stress-strain model for confined
concrete. ASCE Journal of Structural Engineering,
114 (8), 1804-1826.
Muguruma, H., Nishiyama, M. and Watanabe, F. (1993).
Stress strain curve for concrete with a wide range of
compressive strength. Proceedings of Symposium on
High Strength Concrete, California, 314-321.
Nagashima, T., Sugano, S., Kimura, H. and Ichikawa, A.
(1992). Monotonic axial compression test on ultrahigh strength concrete tied columns. Proceedings of
10th World Conference on Earthquake Engineering.
Balkema, Rotterdam , The Netherlands, 2983-2988.
Park, R., Priestly, M. J. N. and Gill, W. D. (1982).
Ductility of square confined concrete columns.
ASCE Journal of Structural Division, 108 (4), 929951.
Pessiki, S. and Pieroni, A. (1997). Axial load behavior
of large scale spirally reinforced high strength
281
concrete column. ACI Structural Journal, 94 (3),

304-314.
Richart, F. E., Brandtzaeg, A. and Brown, R. L. (1928).
A study of failure of concrete under
combined
compressive stresses. Engineering Experiment
Station Bulletin No. 185, University of Illinois,
Urbana.
Razvi, S. and Saaticioglu, M. (1994). Strength and
deformability of confined high strength concrete
columns. ACI Structural Journal, 91 (6), 678-687.
Razvi, S. and Saatcioglu, M. (1996). Tests of high
strength concrete under concentric loading. Report
No.
OCEERC
96-03,
Carleton
Earthquake
Engineering Research Centre, Ottawa, Canada.
Razvi, S. R. and Saatcioglu, M. (1999). Confinement
model for high strength concrete. ASCE Journal of
Structural Engineering, 125 (3), 281-289.
Sheikh, S. A. and Uzmeri, S. M. (1980). Strength and
ductility of tied concrete columns. ASCE Journal of
Structural Division, 106 (5), 1079-1102.
Yong, Y. K., Nour, M. G. and Nawy, E. G. (1988).
Behavior of laterally confined high strength concrete
under axial loads. ASCE Journal of Structural
Engineering, 114 (2), 332-351.

Short Columns

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Short Columns

Diunggah oleh

Hak Cipta:

Format Tersedia

Journal of Advanced Concrete Technology Vol. 3, No.

2, 267-281, June 2005 / Copyright 2005 Japan Concrete Institute

Behavior of Confined High Strength Concrete Columns under Axial

Research Scholar, Department of Civil Engineering,

strength concrete columns in the recent past indicate that

faces of the specimens to prevent direct loading of the

Strain Gauges - Longitudinal Bar

Table 1 Properties of column specimens.

The fabricated steel cages were fixed in the steel

higher strength concrete mixes respectively. The slump

taken to complete each test ranged from 30 minutes to 1

Stress-strain relationships, established by performing at

Fig 2 Stress-strain curves for reinforcing bars.

Fig. 3 Experimental setup.

Table 2 Mix proportions.

velopment of strains associated with the failure of unconfined concrete specimens.

All columns initially behaved in a similar manner and

Fig. 4 Appearance of specimen CC at various

Table 3 Test results.

-- Not applicable, * Could not be measured.

Relative Column Load (P/P o)

Colum n Axial Strain (m m /mm )

Relative Column Load (P/P o)

Colum n Axial Strain (mm /m m )

Relative Column Load (P/P o)

Relative Column Load (P/P o)

Colum n Axial Strain (mm /m m)

Colum n Axial Strain (mm /m m )

confined core than the other specimens of the study as

(Fig. 7), where

Pocc = 0.85 f ' c Acc

While the gross concrete area Ac represented column

Po = 0.85 f ' c ( Ag Ast ) + f y Ast

The ratio Pmax/ Po ranges from 0.95 to 1.10 for circular

P c/P oc, P c/P occ

Poc = 0.85 f ' c Ac

Concrete axial strain, c

Fig.7 Nondimensionalized concrete contribution curve.

5. Effect of test variables

longitudinal bars. Fig. 9 demonstrates the effect of the

Relative Concrete Load ,

Relative Concrete Load ,

strength decay after the peak. An increase of 32 % in

Relative Concrete Load

Relative Concrete Load

Fig. 9 Effect of volumetric ratio and spacing of lateral steel.

Axial Strain (m m/mm )

Fig. 10 Effect of yield strength of lateral steel.

5.5 Amount of longitudinal reinforcement

Relative Concrete Load

Fig. 11 Effect of configuration of lateral ties.

Relative Concrete Load

5.4 Configuration of lateral steel

of high strength concrete columns. However, only a few

Fig. 12 Effect of amount of longitudinal steel.

Relative Concrete Load

amount of confinement is provided to generate enough

Fig. 13 Effect of section geometry on confined concrete.

strength and ductility from square to circular columns.

6. Design implications for ductility

Foster & Attard (2001) analyzed the test data of 40

The equation (7) can be used to design the critical

Effective Confinem ent Index, k e s f yh /f' c (%)