Professional Documents
Culture Documents
Scientific paper
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 col-
umns 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.
tested in duplicate for each case in order to get the aver- faces of the specimens to prevent direct loading of the
age of two results thus making 22 independent column bars. In all the specimens, the ratio of gross area of the
designs. All the details regarding the various column section, Ac, to the core area, Acc, measured to the outside
specimens are illustrated in Table 1 and Fig. 1. Most of of the lateral reinforcement was 1.33. Longitudinal rein-
the results reported here in this study are the averages of forcement was provided in all the confined column
two tests with a few exceptions of SG, SC, CH and SF specimens. The lateral ties of all square specimens had
columns where duplicate results could not be taken due 1350 hooks and a development length as per the ACI 318
to the following reasons. One of the SG specimens failed code provisions. Fig.1 and Table1 show the various
rather more suddenly compared to the other and all the properties of longitudinal and transverse reinforcement.
measuring instruments like LVDTs and strain gauges got The notation fc shows standard cylinder compressive
dislodged and further readings after peak could not be strength on the day of testing of columns, l and fy are
taken. In case of one of the CH specimens the top end respectively longitudinal steel ratio and yield strength of
concrete cover of the column got crushed and the loads longitudinal steel, s, s and fyh are respectively spacing,
came directly on the end steel collars which got dam- volumetric ratio and yield strength of lateral steel. Two
aged and therefore the test was stopped. Whereas one of different specified concrete strengths, two lateral steel
the results of each of SC and SF specimens were rejected yield strengths and two longitudinal steel ratios and re-
due to the reason that in spite of exercising extra care to sulting lateral steel configurations were employed. The
avoid any eccentric loading of the specimens, these col- spacing of spirals and ties were varied from approxi-
umns seemed to suffer from this problem as evidenced mately one fourth to half of the total lateral dimension of
by their considerable tilting and entire spalling on one column in order to give varying volumetric ratios of lat-
side only. eral confining steel.
The specimens were cast in total eleven sets. Each set Failure of the specimens was forced in the test region,
consisted of two duplicate circular and two duplicate which was equal to 300 mm in the middle of the speci-
square columns of one design. For example two speci- men height by reducing the spacing of the lateral rein-
mens of CA and SA each were cast in one set. The col- forcement out side the test region to half of the specified
umns were designed to investigate the parameters of spacing in the test region. However, in the case of the
confinement, including volumetric ratio and spacing of specimens where specified spacing of lateral steel was
transverse reinforcement, yield strength of transverse 30 mm, the spacing in the end zones was reduced to only
reinforcement, longitudinal reinforcement ratio, and lat- three fourth of that in test region to facilitate proper plac-
eral steel configuration in addition to shape of cross sec- ing of concrete in the cover of columns in these end re-
tion and concrete compressive strength. Concrete cover gions. All the specimens were also externally confined
of 10 mm was provided in all the confined specimens. A by 10mm thick mild steel collars in the end regions of
concrete cover of 15 mm was also provided between the 150 mm to further prevent premature failure.
ends of the longitudinal bars and the top and bottom sur-
150mm 150 mm
A B
250 mm
LVDT Cover 10 mm
600 mm 600 mm Cover 10mm
Gauge Length
8 Nos
4 Nos
12 mm 150 mm 12 mm dia.
150 mm
dia.
150 mm
End Region 150 mm 150 mm
A B
Fig. 1 Details of column specimens, reinforcement arrangement and location of strain gauges.
U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005 269
The fabricated steel cages were fixed in the steel higher strength concrete mixes respectively. The slump
moulds and the casting was carried out in the laboratory. value ranging from 100-120mm was maintained to en-
After 24 hours, the specimens were removed from the sure that the concrete could be placed through the dense
moulds and submerged in a water tank for curing. The reinforcement cages.
water-curing period lasted for 28 days, after which the Three standard concrete cylinders (100 x 200 mm)
specimens were left in the laboratory at ambient tem- were cast along with each set of four columns. They
perature until the time of testing. The testing was com- were tested at various ages to determine the strength of
menced after 90 days of ageing. concrete fc on the day of testing of columns. The prop-
erties of unconfined concrete were obtained by testing
2.2 Material properties plain unreinforced column specimens (CPL, CPH, SPL
Two high strength concrete mixes were employed in the & SPH). It was observed that concrete strength of col-
study. The materials consisted of Ordinary Portland Ce- umn specimens were generally lower than the concentric
ment, natural river sand, crushed stone aggregate of strength measured on standard cylinders. The average
maximum size 10mm, tap water for mixing and curing, unreinforced specimen concrete strength was measured
silica fume and superplasticizer admixture to maintain as 88% and 90% of the average concrete cylinder
adequate workability of mix. To obtain the desired strength for the lower high strength (CPL & SPL) and
strength, several preliminary mix designs were practiced upper high strength (CPH and SPH) concretes respec-
and then optimal mix designs were determined after test- tively. The commonly used ratio of 0.85 was then used
ing 28 days strength. The final mixes had 28 days aver- for evaluating the concrete section capacity of the
age cube (150x150x150mm) compressive strengths of specimens tested in the present study.
68.4 & 87.5 MPa and average cylinder (100x200mm) The longitudinal reinforcement consisted of 8 mm and
strengths of 58.03 and 76.80 MPa. Table 2 shows the 12mm deformed bars for circular columns and 12 mm
mix proportions and 28 days compressive strengths for deformed bars for square columns. Two different grades
both the mixes. The measured tangent elastic moduli (412 MPa and 520 MPa) of reinforcing steel with a di-
(Ec) were 31645 MPa and 34465 MPa for lower and ameter of 8 mm were used as lateral reinforcement.
270 U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005
Stress-strain relationships, established by performing at taken to complete each test ranged from 30 minutes to 1
least three coupon tests for each reinforcement, are illus- hour depending upon the degree of confinement in the
trated in Fig. 2. The yield strength of the lateral rein- specimen. To ensure concentric loading, an initial load
forcement is defined as an offset strain of 0.2%. of approximately 20% of the total ultimate load was ap-
plied and the readings of the four LVDTs were moni-
2.3 Instrumentation and testing procedure tored. If the readings of the LVDTs were not approxi-
Longitudinal and lateral steel strains were measured by mately equal, this signals that the load is not concentric.
electrical resistance strain gauges glued to the steel bars. The column was then unloaded and the additional pack-
The strain gauges were pasted to the two opposite longi- ing was given under the wooden plies unless the load
tudinal steel bars at their middle lengths. Similarly two becomes almost concentric. In spite of the rigorous pro-
gauges were glued at the two locations on the lateral cedure followed for aligning the specimens, some eccen-
steel at approximately middle length of the specimen as tricities were unavoidable.
shown in the Fig. 1. The axial displacement of the
specimens was recorded using four linear variable dif-
ferential transducers (LVDTs). Two LVDTs were at- 800
tached on the opposite faces to a top and bottom steel 8 m m ( f y h = 5 2 0 M P a , y h = 0 .0 0 2 6 )
700
clamps to the specimens to give a gauge length of 250
mm. Whereas, two LVDTs were mounted on the other 600
two faces of the specimen and were held by the attach- 500
Stress (MPa)
ments provided to the end steel collars. If a sudden fail-
8 m m ( f y h = 4 12 M P a , y h = 0 .0 0 2 4 )
ure occurs or due to spalling, the steel clamps get dis- 400
The second peak was observed for many circular col- confined core than the other specimens of the study as
umns CA, CC, CD, CE, CH and CI, and a few square implied by their respective volumetric ratios, spacing
columns SA, SD, SE and SI which have better efficiently and distribution of lateral steel. In the case of CE and SD
specimens the load values at second peak were in fact
1.2 more than those at respective first peaks. This may be
Relative Column Load (P/P o)
1
attributed due to the fact that CE specimen has a very
CB
CD high volumetric ratio and closer spacing of spirals.
0.8 CC
Square column SD had also high volumetric ratio and
0.6 CA the lateral tie arrangement of eight bars, which provided
0.4 better confinement than an arrangement of four bars.
0.2
However, the SG, CG and SB specimens failed immedi-
ately after the first maximum load was reached, indicat-
0
0 0.03 0.06 0.09 0.12 0.15
ing that the spacing of the lateral steel was too wide to
Colum n Axial Strain (m m /mm )
provide lateral confinement. Overall, it was observed
(a) that for the similar confinement circular columns had a
better post peak behavior than the square tie confined
1.2
columns and the lower high strength mix columns failed
slowly in a better ductile fashion than the similarly con-
Relative Column Load (P/P o)
1
CH
fined higher concrete strength mix columns.
CF
0.8
CG
CI At the stage of cover spalling, in all the specimens
CE
0.6
longitudinal steel yielded but, the level of stress in lateral
confining steel was considerably less than their respec-
0.4
tive yield strengths. Similar observations were made in
0.2 some earlier studies also (Cusson & Paultre 1994; Foster
0
1999). Lateral steel yielded either at the second peak or
0 0.03 0.06 0.09 0.12 0.15 even after that in the descending part of load-
Colum n Axial Strain (mm /m m )
(b)
deformation behavior. In fact yielding of lateral steel at
second peak was noticed only for well-confined columns
Fig. 5 (a) & (b) Column axial load versus axial strain and mostly for lower high strength mix specimens. They
curves (circular columns). 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
1.2
tie fracture was invariably at the moment when the load
Relative Column Load (P/P o)
1
SD dropped to approximately 40-50% of the maximum
SA
0.8 SB SC value in the descending segment. It was observed that
the fracture of lateral steel caused a sudden drop in load.
0.6
In the case of columns with 75 mm spacing of transverse
0.4 steel, buckling of longitudinal steel was observed prior
0.2 to the fracture of lateral steel, which ultimately lead to
the fracture of ties and crushing of core.
0
0 0.02 0.04 0.06 0.08 0.1
1 SE
curves of confined concrete, as suggested by Sheikh &
SF
SG SI Uzumeri (1980) and Cusson & Paultre (1994). The con-
0.8 SH
crete contribution Pc at a certain deformation was deter-
0.6 mined by subtracting the contribution of longitudinal
0.4
steel from the applied load P. The load carried by the
longitudinal steel was determined from the stress-strain
0.2
curves obtained from the tension test. In computing the
0 load carried by the longitudinal steel, the strain harden-
0 0.02 0.04 0.06 0.08 0.1 0.12
ing regions of the curves were assumed to be a straight
Colum n Axial Strain (mm /m m ) line that resulted in a maximum error of less than 5% in
(b)
the computed concrete force. The concrete contribution
Fig.6 (a) and (b) Column axial load versus axial strain curves were nondimensionalized with respect to gross
curves (square columns). concrete area force Poc and core concrete area force Pocc
U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005 273
longitudinal bars. Fig. 9 demonstrates the effect of the strength decay after the peak. An increase of 32 % in
volumetric ratio of confining reinforcement on the be- strength, 270% in strain ductility (c50c/co) and 58.27 %
havior of a few confined high strength concrete columns in I10 ductility was noticed for the higher concrete
of the study. These columns also have different spacing strength mix circular columns as the volumetric ratio of
of lateral steel spirals or ties. As expected, the larger the spirals increased from 2.2% to 5.5%. Similarly 25% gain
volumetric ratio or closer the spacing of lateral steel, the in strength, 236% enhancement in strain ductility and
more ductile is the behavior of columns. The columns 90% increase in I10 ductility were observed for higher
with low volumetric ratio or increased spacing of lateral strength concrete mix square columns due to the increase
steel exhibit brittle behavior, showing faster rate of in volumetric ratio of lateral ties from 2.2% to 5.5%.
1.4 63.35 MPa was investigated by comparing results of two pairs each
CB
1.2
of circular and square columns. The compared speci-
62.80 MPa
Pc/Poc, Pc/Pocc
1.4 SD
SA 63.35 M Pa
steel of only well confined specimens reached its yield
1.2
62.20 M Pa SB value at peak-confined stress of concrete and that too for
P c/P oc, P c/P occ
SE 1.6 CH
1.4 s = 30 mm
f yh = 520 M Pa SC
1.4 SH
Relative Concrete Load
s = 5 .5 %
P c/P oc, P c/P occ
1 SF 1
0.8 s = 50 mm
s = 3 .3 % 0.8
CA
0.6
0.6 f yh = 412 MPa
0.4 CG CF
s = 50 mm 0.4
0.2 s = 75 mm SG, s = 75 mm CF SA
s = 2 .2 % s = 3 .3 % s = 2 .2 % 0.2 f y h = 412 MPa SF
f y h = 412 MPa
0 f yh = 412 MPa
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12
Axial Strain (m m/mm ) Axial Strain (m m /m m )
Fig. 9 Effect of volumetric ratio and spacing of lateral steel. Fig. 10 Effect of yield strength of lateral steel.
U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005 275
amount of confinement is provided to generate enough of high strength concrete columns. However, only a few
lateral expansion of core so that yield strength of high representative comparisons are shown in Fig. 13 with
yield strength lateral steel is fully developed prior to respect to this parameter. The specimens of each
appreciable post-peak strength decay. matched pair have similar concrete strength and volu-
metric ratios of lateral steel. It is clear from this figure
5.4 Configuration of lateral steel that circular confinement is substantially more effective
The transverse steel configuration and the resulting dis- than rectilinear confinement in case of high strength con-
tribution of the longitudinal steel play a significant role crete also as evidenced by the margin of increase in
in the confinement of concrete. If the lateral confining
pressure applied by the transverse reinforcement on con- 1.6
crete is well distributed around the perimeter of the core
1.4
concrete, the efficiency of confinement is improved. In
Pc/Poc, Pc/Pocc
the lateral steel configuration by comparing the behavior 1
SI
of square column specimens SE and SI, which have al-
0.8
most, equal volumetric ratio of lateral steel (Fig.11). The SE
specimen SE was constructed using single perimeter tie 0.6
(4 bar arrangement), and SI specimen had double tie 0.4
configuration (8 bar arrangement). The ductility indices
0.2
(I10) of the specimens SE and SI were 6.44 and 7.86,
respectively indicating a 22% improvement in ductility 0
of specimen SI which has better steel configuration. 0 0.01 0.02 0.03 0.04
However, the ratio Pcc/Pocc is equal to 1.382 and 1.43 for Axial Strain (m m /m m )
specimens SE and SI respectively showing strength en-
hancement of only 3.5%. This indicates that the im- Fig. 11 Effect of configuration of lateral ties.
provements in deformability are more marked than in
strength due to better configuration of transverse steel. 1.8
Therefore, according to the test results, transverse steel
1.6
configuration and the resulting distribution of longitudi- CD
Relative Concrete Load
1.4 l = 3.84
nal steel has a considerable effect on the ductility of the
confined concrete but insignificant effect on its strength. 1.2
P c/P oc, P c/P occ
CI
1 l = 3.84 %
5.5 Amount of longitudinal reinforcement 0.8 CA
Fig. 12 shows the two different pairs of circular speci- 0.6 l = 1.7 %
mens, and for each matched pair, two specimens varying 0.4 CF
only in their ratios of longitudinal reinforcement are 0.2 l = 1.7 %
compared. It may be mentioned here that this type of 0
comparison was not possible in case of square columns 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
because, in addition to amount of longitudinal steel,
volumetric ratios of lateral ties were also varying for any Axial Strain (m m /m m )
two comparable specimens. The results show (Table 3) Fig. 12 Effect of amount of longitudinal steel.
that the strength and deformability gains are equal to
0.4% and 6.4% for the pair CA-CD and 1.2% and 11%
2
for the pair CF-CI, respectively. Therefore, based upon
the results it can be said that the amount of longitudinal 1.8
1.2
buckling. 1 CE
0.8
5.6 Section geometry 0.6 SC
It is well established now that circular spirals are more 0.4 SE
effective in confining concrete than rectilinear ties owing 0.2
to their better uniform distribution of lateral confining
0
pressure around the core compared to the case of square 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
or rectangular ties (Mander 1988; Razvi & Saatcioglu Axial Strain (m m /m m )
1996). In the present study this variable was investigated
extensively to quantify its effect in case of confinement Fig. 13 Effect of section geometry on confined concrete.
276 U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005
strength and ductility from square to circular columns. In Fig. 14, the I10 ductility index for all the specimens
The square columns SC and SE failed in a comparatively is plotted against the effective confinement parame-
brittle fashion at 0.0042 and 0.0051 axial strains respec- ter k e s f yh / f ' c . The ke for both spirally confined cir-
tively, while the matched circular columns CC and CE cular columns and tie confined square columns was
were able to develop axial strains of 0.0081 and 0.0074 computed using the procedure suggested by Mander et al
respectively without any strength decay. The maximum (1988). The computed values of confinement index
gains of 28% in confined strength (CC vs SC) and 66% k e s f yh / f ' c for the various confined specimens are
in I10 ductility (CB vs SB) were noticed due to the given in Table 1. In a few columns (CG, SB, SF and SG)
change in section geometry. Therefore, if the same per- tested in this study, a faster rate of strength decay was
centages of strength and ductility enhancements are de- observed after peak. For these columns, an actual value
sired, square or rectilinearly confined columns are re- of I10 ductility could not be obtained. Therefore, for such
quired to be confined more vigorously than circular col- cases average values of I10 were calculated by obtaining
umns. the upper and lower bounds (Foster & Attard 2001). A
best fit relationship between I10 and k e s f yh / f ' c was
6. Design implications for ductility found from the plot and is given by:
As a result of some field observations regarding the per- I 10 = 2.89 ln(1000k e s f yh / f ' c ) 0.45 (6)
formance of columns after major earthquakes, it is now
well established that despite following the strong- Foster & Attard (2001) analyzed the test data of 40
column weak-beam concept in design, damage could concentrically loaded columns tested by Razvi & Saat-
occur at ends of the columns. Therefore, ductile detailing cioglu (1996) and computed I10 values for all the speci-
of these potential plastic hinge regions of columns be- mens. For the specimens with I10 < 8 sudden changes in
comes even more important when high strength concrete the load strain data and faster rate of post-peak strength
is employed. One of the ways in which most of the decay were recorded. However, columns with I10 > 8
building codes of various countries ensure ductility in exhibited ductile post-peak curves. Therefore, the au-
concrete columns, is by specifying the amount of confin- thors based upon the evaluation of data concluded that
ing transverse reinforcement in critical regions of col- for the regions of moderate seismicity desirable ductility
umns. However, these code equations specifying the could be achieved if I10 is more than 8. An analysis of
minimum amount of transverse steel for columns are present test data also reveals that the columns with I10 >
empirical and have been developed based upon the ex- 8 showed ductile post-peak load-strain curves with
perimental data obtained from the testing of normal slower rate of strength decay. So if the same yardstick is
strength concrete columns. An attempt has been made in applied on the present test data to achieve a moderate
the present study to work out the confinement require- seismic resistance in terms of ductility, the following
ments of high strength concrete columns, based upon the relationship could be derived after substituting the value
present test results. Ideally, the behavior of columns of I10 as 8 in equation (6):
should be studied under combined axial compression k e s f yh / f ' c 0.136 (7)
and bending to assess the ductility performance of con-
fined reinforced concrete and the effect of cyclic loading The equation (7) can be used to design the critical
should be taken in to account. But, in this study uniaxial
load-strain data of confined concrete column sections 10
under monotonic loading has been analyzed, and this
9
should give a reasonable assessment in the first instance
considering it to be an extreme case. 8
A number of past studies (Sheikh & Uzmeri 1980; 7
Ductlity Index I 10
Ag
s f yh / f ' c 0.6( 1) (5) Fig. 14 I10 ductility index versus effective confinement in-
Ac dex.
U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005 277
hinge regions of high strength concrete columns. It may steel, tie configuration and spacing of lateral steel.
be mentioned here that equation (7) specifies the mini- Muguruma, Nishiyama and Watanabe (1993)
mum effective confinement index or ratio for high Muguruma, Nishiyama and Watanabe proposed a
strength concrete columns, which includes confinement three-part stress-strain model for confined concrete
effectiveness coefficient (ke <1) unlike equations (4) and based on their own previous studies. A wide range of
(5) of ACI 318 provisions. concrete strength ranging from 40 to 130 MPa was cov-
ered. They tested small square specimens confined later-
7. Assessment of existing stress-strain ally by square helix hoops of different yield strengths
confinement models and with various volumetric ratios. The yield strengths
of the hoops ranged from 161 to 1353 MPa.
The analysis of structural members requires an analytical Li, Park and Tanaka (1994)
model for the full stress-strain relationship of concrete in Li, Park and Tanaka proposed a three-part stress-strain
compression both in confined and unconfined states. The model for confined high strength concrete based on their
analytical models for confined normal strength concrete experimental results. Forty reinforced concrete short
based on extensive experimental data are well estab- columns of both cylindrical (240x720 mm) and square
lished. These models, based on the test results of normal (240x240x720mm) cross sectional shapes were tested.
strength concrete columns, might be inadequate for high The main parameters were in place concrete strength
strength concrete columns, which possess a less ductile (35.2 to 82.5 MPa) and lateral steel grade (445 and 1318
stress-strain behavior. So the normal strength concrete MPa) in addition to other parameters like spacing, volu-
models if applied to high strength concrete shall overes- metric ratio and configuration of lateral steel.
timate the ductility. As a result a number of such models Cusson and Paultre (1995)
specifically for high strength concrete columns have also Cusson and Paultre developed a confinement model
been proposed in the recent past. The following section for high strength concrete on the basis of test results of
provides an overview of these analytical confinement 50 large-scale high strength concrete tied columns tested
models that cover high strength concrete with strengths under concentric loading. Out of them, 30 HSC tied col-
more than 60 MPa. umns (235x235x1400 mm) were tested by authors them-
Yong, Nour and Nawy (1988) selves and 20 HSC tied columns (225x225x715 mm)
Yong, Nour and Nawy proposed a model for rectiline- were tested by Nagashima et al. (1992). The concrete
arly confined high strength concrete columns. They compressive strengths of the specimens ranged from 60-
tested 24 square prisms that were made of high strength 120 MPa. The ties with yield strength from 400 to 800
concrete with compressive strength ranging from 83.6 to MPa were used. The proposed model takes in to account
93.5 MPa, confined with square ties with yield strength tie yield strength, tie configuration, transverse rein-
of 496 MPa. The variables considered were volumetric forcement ratio, tie spacing, and longitudinal reinforce-
ratio of lateral ties, concrete cover and distribution of ment ratio. The two-part stress-strain relationship in-
lateral steel. A three-part stress-strain relation was pro- cluded separate expressions for ascending and descend-
posed to predict the constitutive behavior of confined ing parts.
high strength concrete. Razvi and Saatcioglu (1999)
Bjerkli, Tomaszewicz and Jansen (1990) Razvi and Saatcioglu proposed a model for confined
Bjerkli, Tomaszewicz and Jansen proposed a three- normal and high strength concrete columns using exten-
part stress-strain curve for high strength concrete col- sive test data of authors own test results as well as the
umns for both circular and rectilinear cross sectional experimental results of other research studies. This in-
shapes based upon their test results. They tested a large cluded the test results of nearly full size specimens of
number of plain and confined high strength concrete different shapes, sizes, reinforcement configurations, tie
columns with compressive strength ranging from 65 to yield strength (400 to 1387 MPa) and concrete strengths
115 MPa. Both cross sectional shapes i.e. cylinders (150 (30 to 130 MPa). The parameters incorporated in the
mm diameter and 500 mm high) and prisms model were type, volumetric ratio, spacing, yield
(150x150x500mm and 300x500x2000mm) were in- strength, and arrangement of transverse reinforcement,
cluded. The test specimens contained longitudinal steel distribution and amount of longitudinal steel as well as
but no concrete cover. concrete strength and section geometry. The two part
Nagashima, Sugano, Kimura and Ichikawa (1992) stress-strain model proposed by the authors was in the
Twenty-six prism specimens (225x716mm) of high form of the ascending parabolic branch up to peak and a
strength concrete of strengths 59 and 118 Mpa and later- linear descending branch up to 20% of the peak stress.
ally reinforced with ties of yield strengths 784 and 1374 Legeron and Paultre (2003)
MPa were tested by Nagashima, Sugano, Kimura and Legeron and Paultre proposed a stress-strain confine-
Ichikawa. Based on the test results a two-part stress- ment model for normal and high strength concrete col-
strain relationship was proposed for confined high umns based on the large number of test results of circular,
strength concrete columns. The variables taken in to square and rectangular columns tested under various
account were concrete strength, yield strength of lateral research studies that included the studies undertaken by
278 U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005
themselves and a number of other researchers. The con- cular specimens. In the case of square columns Razvi &
crete compressive strengths ranged from 20 to 140 MPa Saatcioglu (1999); Li et al. (1994); Bjerkli et al. (1990)
and tie yield strengths ranged from 300 to 1400 MPa. and Yong et al. (1988) models always overestimated the
The model incorporates almost all the parameters of con- experimental stress-strain curves by a considerable mar-
finement. The stress strain relationship was basically gin for all the specimens. The model proposed by Mugu-
same as proposed by Cusson and Paultre (1995), but the ruma et al. (1993) produced considerably steeper stress-
parameters of the model were recalibrated on the basis strain curves for all the specimens that were far below
of large number of test data collected by the authors. the actual test curves. Whereas, the models proposed by
A critical review of the above models indicate that Cusson & Paultre (1995) and Nagashima et al. (1992)
most of them have limited validity in terms of concrete continuously underestimated the test curves though; the
strengths, column geometry, transverse reinforcement predictions were quite close in the latter case for few
yield strength and loading conditions. With the exception specimens. As in the case of circular columns, Legeron
of the models proposed by Razvi & Saatcioglu (1999); & Paultre (2003) model was also able to produce close
Legeron & Paultre (2003); Li et al. (1994) and Bjerkli et predictions of stress-strain behavior for almost all the
al. (1990), which cover both circular and rectilinear sec- square columns except for few cases like SH specimen
tions, all other models are applicable to only square or where it slightly overestimated the experimental stress-
rectilinear shapes. The models proposed by Yong et al. strain curve. The ability of Legeron & Paultre (2003)
(1988); Nagashima et al. (1992); Li et al. (1994) have model to consistently follow the actual test behavior for
limited applicability for wide concrete strength ranges. It both circular and square columns undoubtedly proves its
has been proved experimentally in the present study and superiority over the other models of the study. Therefore,
in many earlier studies (Cusson & Paultre 1994; Razvi & the present study concludes that this model can be em-
Saatcioglu 1996; Foster 1999) that for high strength con- ployed to predict the uniaxial response of high strength
crete columns, lateral-confining ties may not yield when concrete columns with a reasonable degree of accuracy.
peak of confined concrete stress-strain is reached. But,
most of the models use lateral steel yield strength to cal- 120
Razvi Experimental.
culate lateral confining pressure at peak-confined Legero n Li
B jerkli
strength. Only Cusson & Paultre (1995); Razvi & Saat- 100
80
only four namely Razvi & Saatcioglu (1999); Legeron &
Paultre (2003); Li et al. (1994) and Bjerkli et al. (1990) 60
can be applied to circular columns. Figures 15 to 22
illustrate the comparisons of the experimental and pre- 40
dicted stress-strain curves of a few representative test
specimens only. The comparisons for circular columns 20 CC
indicate that Razvi & Saatciouglu (1999) and Li et al.
0
(1994) models consistently overestimate the actual test 0 0.01 0.02 0.03 0.04
behavior, while Bjerkli et al. (1990) model underesti-
Strain (m m /m m )
mates the test curves. Legeron & Paultre (2003) model
closely follows the experimental stress-strain curves Fig. 16 Comparison of confined axial stress-strain curves
though, with a slight overestimation, for most of the cir- for CC specimen.
U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005 279
120
160 Nagashima Cusso n
Razvi Experimental Razvi Experimental
Legero n Li 100 Legero n Li
140
B jerkli B jerkli M uguruma
120
80
Stress (MPa)
100
Stress (MPa)
60
80
60 40
40
20
SC
20 CE
0
0
0 0.005 0.01 0.015 0.02 0.025 0.03
0 0.01 0.02 0.03 0.04
Strain (m m /m m ) Strain (m m /m m )
Fig. 17 Comparison of confined axial stress-strain curves Fig. 20 Comparison of confined axial stress-strain curves
for CE specimen. for SC specimen.
140 120
Nagashima Cusson Razvi
Razvi Experimental
Experimental Legeron Li
Legeron Li
120 Bjerkli M uguruma Yong
B jerkli 100
100
80
Stress (MPa)
Stress (MPa)
80
60
60
40
40
20 20
SD
CI
0 0
0 0.01 0.02 0.03 0.04 0 0.01 0.02 0.03 0.04
Strain (m m /m m ) Strain (m m /mm )
Fig. 18 Comparison of confined axial stress-strain curves Fig. 21 Comparison of confined axial stress-strain curves
for CI specimen. for SD specimen.
100 120
Nagashima Cusso n Nagashima Cusson Razvi
Razvi Experimental Experiment al Legeron Li
Legero n Li 100 Bjerkli M uguruma Yong
80 B jerkli M uguruma
80
Stress (MPa)
Stress (MPa)
60
60
40
40
20
SA 20
SH
0 0
0 0.005 0.01 0.015 0.02 0.025 0 0.005 0.01 0.015 0.02 0.025
Strain (m m /m m ) Strain (m m /m m )
Fig. 19 Comparison of confined axial stress-strain curves Fig. 22 Comparison of confined axial stress-strain curves
for SA specimen. for SH specimen.
tion, shape of cross section and concrete compressive Acc = core concrete area
strength. A comparative study of existing stress-strain Ast = sectional area of longitudinal reinforcement
models for high strength concrete columns is also re- Ec = tangent modulus of elasticity of concrete
ported. The following conclusions can be made based fc = cylinder compressive strength of concrete
upon this study. fyh = yield strength of tie steel
1. High strength concrete columns show brittle behavior fy = yield strength of longitudinal steel
unless the columns are confined with sufficient lateral fhcc = stress in lateral steel at peak stress of confined
reinforcement that can provide adequately high lateral concrete
confining pressure. A consistent decrease in strength Pmax = maximum applied load on the column
enhancement and deformability of columns is ob- Pc = concrete load corresponding to first peak
served with increasing concrete strength. Therefore, a Pcc = peak confined concrete load (corresponding to
higher degree of confinement is required in columns second peak)
with higher concrete strength than in a column with Po = theoretical load carrying capacity of column in-
lower concrete strength to achieve similar advantages. cluding longitudinal bars
2. High strength concrete columns suffer from the prob- Poc = theoretical capacity of concrete in the column
lem of premature cover spalling which reduces the Poc c = theoretical capacity of concrete core
column load carrying capacity to even less than un- s = spacing of lateral steel
confined strength if sufficient confinement is not pro- s = volumetric ratio of lateral steel
vided. The post spalling behavior depends very much l = volumetric ratio of longitudinal steel
on the degree and efficiency of confinement. co = unconfined strain of concrete column
3. Among the test variables studied, volumetric ratio and c50o = axial strain at which the stress drops to 50% of
spacing of lateral steel has a more pronounced effect peak in unconfined concrete
on the behavior of confined columns than the other = axial strain corresponding to the first peak, Pc
parameters like yield strength of lateral steel, longitu- cc = axial strain at peak confined load, Pcc.
dinal steel ratio and configuration of lateral steel
c50c = axial strain at which the stress drops to 50% of
though, improvement in each of the variables consid-
peak in confined concrete
ered, translated into enhancements in strength and
yh = yield strain of reinforcement.
ductility. However, increasing the yield strength of
lateral steel seems to give benefits only when a col-
References
umn has high volumetric ratio and efficient arrange-
Bjerkli, L., Tomaszewicz A. and Jansen J. J. (1990).
ment of lateral steel.
Deformation properties and ductility of high strength
4. The ductility of columns is shown to be dependent on
concrete. Proceedings of Second International
effective confinement index k e s f yh / f ' c . It is con-
Symposium on Utilization of High Strength Concrete,
cluded that for regions of moderate seismicity, the
Berkley, California, 215-238.
critical hinge regions of high strength concrete col-
Cusson, D. and Paultre, P. (1994). High strength
umns should have effective confinement index of
concrete columns confined by rectangular ties. ASCE
more than 0.136 to achieve adequate ductility.
Journal of Structural Engineering, 120 (3), 783-804.
5. A comparative study is undertaken to evaluate the
Cusson, D. and Paultre, P. (1995). Stress-strain model
capabilities of the various confinement models of high
for confined high strength concrete. ASCE Journal
strength concrete columns to predict the actual ex-
of Structural Engineering, 121 (3), 468-477.
perimental behavior. The study indicated that almost
Foster, S. J. (1999). Design and detailing of high
all the models are able to estimate correctly ascending
strength concrete columns. Research Report No. R-
part of stress-strain curve. But, there are wide varia-
375, University of New South Wales, Sydney,
tions in the prediction of the post-peak part of stress-
Australia.
strain curves, with a few models underestimating and
Foster, S. J. (2001). On behavior of high strength
a few overestimating the test behavior except the
concrete columns: cover spalling, steel fibers, and
Legeron and Paultre (2003) model, which consistently
ductility. ACI Structural Journal, 98 (4), 583-589.
predicted experimental results with least amount of
Foster, S. J. and Attard, M. M. (2001). Strength and
discrepancy. Therefore, the present study concludes
ductility of fiber reinforced high strength concrete
that the Legeron and Paultre (2003) model can be
columns. ASCE Journal of Structural Engineering,
used to analytically predict the uniaxial response of
127 (1), 281-289.
high strength concrete columns with a reasonable de-
Joint ACI-ASCE Committee (1997). High Strength
gree of accuracy.
Concrete Columns: State of the Art. Report No. ACI-
441R-96, ACI Structural Journal, 94 (5), 323-335.
Notations
Li, B. Park, R. and Tanaka, H. (1994). Strength &
Ag = gross area of column cross section including area
ductility of reinforced concrete members and frames
of longitudinal steel bars.
constructed using high strength concrete. Research
Ac = gross area of concrete in the section
report No. 94-5, University of Canterbury,
U.K. Sharma, P. Bhargava and S.K. Kaushik / Journal of Advanced Concrete Technology Vol. 3, No. 2, 267-281, 2005 281