Accelerat ing t he world's research.

CONFINEMENT REINFORCEMENT
DESIGN CONSIDERATIONS FOR
DUCTILE HSC COLUMNS
Putu Pradnyanita

Cite this paper Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers Download a PDF Pack of t he best relat ed papers 

Duct ilit y Design of Reinforced Very-High St rengt h Concret e Columns (100-150 MPa) Using Cu…
Shanaka Krist ombu Baduge

Novel energy-based rat ional for nominal duct ilit y design of very-high st rengt h concret e columns (> 10…
Shanaka Krist ombu Baduge

EXPERIMENTAL PERFORMANCE OF HSC COLUMNS UNDER CYCLIC LOADS

Mohamed S Mohamed
 CONFINEMENT REINFORCEMENT DESIGN CONSIDERATIONS FOR
DUCTILE HSC COLUMNS

By Oguzhan Bayrak1 and Shamim A. Sheikhz

ABSTRACT: This paper presents results from a continuing research program that aims to study confinement of
concrete by lateral reinforcement. The current work deals with the experimental behavior of high-strength con-
crete (HSC) and ultrahigh-strength concrete (UHSC) column behavior. Realistically sized columns (305 X 305
X 1,473 mm) with heavy stubs (508 X 762 X 813 mm) were tested under moderate to high axial load levels
and reversed cyclic displacement excursions. The behavior' of UHSC and HSC columns were examined and
compared to normal strength concrete column behavior. The variables studied in this research program are the
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

concrete strength, axial load level, steel configuration, amount of lateral steel, and the presence of a heavy stub.
A performance-based design procedure for the design of confinement reinforcement in HSC and UHSC columns
is presented and compared with current American Construction Institute (United States) and New Zealand
Standards (New Zealand) codes.

INTRODUCTION (Sheikh and Khoury 1993). This comparison allows a direct

evaluation of the effect of concrete strength. The effects of
According to current seismic design philosophy, the ability reinforcement configuration, amount of lateral reinforcement,
of concrete-framed structures to withstand strong ground mo- and axial load level on the behavior of columns are also eval-
tions depends mainly on the formation of plastic hinges and uated. The validity of a recently proposed performance-based
their capacities to absorb and dissipate energy without signif- confinement reinforcement design procedure (Sheikh and
icant loss of strength. For this reason, most building codes Khoury 1997) for NSC columns is checked against HSC and
attempt to ensure hinging in the beams rather than the columns UHSC test data. To make these design equations applicable to
to guarantee stability. However, recent earthquakes and ana- HSC and UHSC columns, a modification to this design pro-
lytical investigations (Paulay 1977; Mitchell and Paultre 1994; cedure is suggested. The suggested procedure is compared
Bayrak 1995) show that formation of plastic hinges in columns with NZS 3101:1995 code's requirements which essentially
of a framed structure, at locations other than the column bases have a very similar form with the design equations derived by
at the foundation level, is still possible as a result of a severe Watson et al. (1994) and Watson and Park (1994).
earthquake despite the application of the "strong column-weak
beam" concept, as recommended by various design codes. RESEARCH SIGNIFICANCE
During the last two decades, the use of high-strength concrete
(HSC) has become widespread adding another dimension to In most concrete design codes (ACI 318-95, CAN3-A23.3-
the design problem. The equations that exist in the current M94, and NZS3101:1995) provisions for the design of con-
design codes [ACI 318-95 ("Building" 1995), CAN3-A23.3- finement reinforcement and calculation of moment capacity
M94 ("Code" 1994), and NZS3101:1995 ("Code" 1995a,b)] contain empirical constants. In the derivation of these empir-
were originally derived based on experimental results in which ical constants experimental data available in the literature were
normal strength concrete (NSC) was used. Therefore the ap- used. The majority of the data utilized was from tests in which
plicability of current code provisions for the design of con- NSC was used. Only recently, a very limited amount of ex-
finement reinforcement in HSC columns and for the calcula- perimental data on HSC columns has become available. Ex-
tion of moment capacity of HSC column sections has to be perimental data on UHSC column behavior can rarely be
investigated. found in the literature, especially on realistically sized columns
The work presented here, is part of a comprehensive re- tested under moderate to high axial load levels and subjected
search program (Sheikh and Uzumeri 1980; Sheikh and to large inelastic displacement excursions.
Khoury 1993; Sheikh and Toklucu 1993; Bayrak and Sheikh Design engineers have no option but to use the equations
1996) that aims to study confinement of concrete by circular provided in concrete design codes. There is still a lot to be
as well as rectilinear lateral reinforcement. The current work known about HSC and UHSC column behavior; due to their
deals with the experimental behavior of concrete columns con- appealing mechanical properties HSC and UHSC are being
fined by rectilinear ties and subjected to moderate to high lev- used regularly by design engineers. This necessitates the eval-
els of axial load and cyclic lateral displacements simulating uation of applicability of relevant requirements of concrete
earthquake effects. The main focus of the current work is on design codes to HSC and UHSC and is partially accomplished
the earthquake performance of ultrahigh-strength concrete using the experimental data obtained from this research.
(UHSC) specimens with concrete strength of 102 MPa. UHSC A more rational design procedure for the design of confine-
specimen behavior is compared with HSC specimen behavior ment reinforcement in normal strength columns has been sug-
(Bayrak and Sheikh 1996) and NSC specimen behavior gested by Khoury and Sheikh (1991) and Sheikh and Khoury
(1997). Concrete strengths up to 55 MPa were used in the
'PhD Candidate, Dept. of Civ. Engrg., Univ. of Toronto, Toronto, On- experimental work upon which the design equations were
tario, M5S lA4, Canada. based. Test data on HSC and UHSC column behavior from
1Prof., Dept. of Civ. Engrg., Univ. of Toronto, Toronto, Ontario, M5S this research are used to evaluate the validity of the suggested
IA4, Canada. design procedure and necessary modifications are suggested.
Note. Associate Editor: Walter H. Gerstle. Discussion open until Feb- Watson et al. (1994) reported another rational design proce-
ruary I, 1999. To extend the closing date one month, a written request
must be tiled with the ASCE Manager of Journals. The manuscript for
dure that forms the basis of the current New Zealand code's
this paper was submitted for review and possible publication on Novem- requirement. These equations were mainly based on experi-
ber 21, 1996. This paper is part of the Jour1l4l of Structural Engineer- mental data reported by Watson and Park (1994). All of the
ing, Vol. 124, No.9, September, 1998. ©ASCE, ISSN 0733-9445/98/ 11 columns reported in Watson and Park's paper were made
0009-0999-1010/$8.00 + $.50 per page. Paper No. 14626. of concrete with strengths ranging from 39 to 47 MPa.
JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998/999

J. Struct. Eng. 1998.124:999-1010.

 Gauged ties:

liB Strain Gauges on Outer Ties

a
TA 3
TA4
1111111111 TA2
I I I 1111111
All dimensions in mm TA1
Strain Gauges on Inner Ties
TAS
76 76 127 TA7-a-
32-H--432-r-356 ~ 1'1 I 1175 r Bars L2 and L6 TA6 JA5
iI )(t- otJ"I- otJ !IJ tI.)(~)(t-otil t- - __

~S19 ,1----- I
1429 - - - - - : Longitudinal Bars
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

II 22::r-S-L-1.-L-3-.-L4-,-L-5-'L-7-.-L-S- L~P;[s
Strain Gauges on Longitudinal Bars L3~ L2
L1

.A .4"
•. A•. fl-

./'.
"-

51mm

LVOTs on north face L VOTs on south face

FIG. 1. Location of Strain Gauges and LVDTs

TABLE 1. Member and Section Ductility Parameters

Lateral Reinforcement Axial Load Ductility Factor Ductility Ratios Energy Indicators
Size @
P
f~
Specimen RA/p (MPa)
spacing
(mm)
P. fyh ~
('¥o) (MPa) Aoh(ACI)
-f~Ag P Po
IJ." at IJ.. at IJ.. at
0.8Pm• x 0.8Mmox 0.9Mmox N"80 N"'t N.80 N. t W80 ~ E80 Et
(1 ) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11 ) (12) (13) (14) (15) (16) (17) (18) (19) (20)
AS-5HT 2.3 101.8 10M @ 90 4.02 542.0 1.09 0.45 0.48 4.0 9.6 5.6 14 35 27 49 23 98 144 311
2.3 101.8 15M @ 90 4.02 463.0 1.09 0.45 0.48 4.0 9.6 5.6 14 35 27 49 23 98 144 311
AS-6HT 3.5 101.9 15M @ 76 6.74 463.0 1.70 0.46 0.49 6.3 14.0 10.4 23 68 54 114 49 230 450 1,243
AS-7HT 1.7 102.0 10M @ 94 2.72 542.0 0.80 0.45 0.48 3.1 7.2 5.5 9 12 14 24 11 17 25 57
ES-8HT 2.2 102.2 15M @ 70 4.29 463.0 1.08 0.47 0.50 3.6 6.7 5.1 11 16 14 22 18 32 33 99
ES-1HT 2.3 72.1 15M @ 95 3.15 463.0 1.13 0.50 0.50 4.6 6.6 5.9 15 20 19 25 33 57 80 105
AS-2HT 3.3 71.7 10M @ 90 2.84 542.0 1.19 0.36 0.36 6.2 15.8 13.6 18 61 53 113 41 313 631 1,412
AS-3HT 2.4 71.8 10M @ 90 2.84 542.0 1.19 0.50 0.50 5.0 10.1 9.1 15 28 20 42 36 102 161 396
AS-4HT 3.7 71.9 15M @ 100 5.12 463.0 1.83 0.50 0.50 7.0 21.2 17.7 25 69 84 151 231 354 997 1,688

EXPERIMENTAL PROGRAM specimens P/Po varied between 0.36 and 0.50 to evaluate the
effect of axial load on HSC specimen behavior. Test specimens
Results from the tests of four large-scale UHSC column were subjected to considerably high axial loads because most
specimens made from 102-MPa concrete are presented. All of of the concrete design codes allow high axial loads. Only a
the column-stub specimens were tested under constant mod- very limited amount of test data are available for large-size
erate to high axial loads and large cyclic inelastic displace- UHSC specimens tested under these levels of axial load, es-
ments. These results are compared with results from recent pecially when they are subjected to very large inelastic dis-
tests (Bayrak and Sheikh 1996) on HSC concrete specimens placement cycles.
made of 72-MPa concrete and earlier tests (Sheikh and Khoury
1993; Sheikh et al. 1994) on similar specimens having con- Specimens
crete strengths between 30 and 55 MPa.
The level of axial load, as measured by the index P/Po, was Specimens consisted of 305 X 305 X 1,473 mm columns
roughly equal to 0.50 for all of the currently tested UHSC cast integrally with 508 X 762 X 813 mm stubs (Fig. 1). The
specimens. On the other hand, for recently tested HSC column column part of the specimen represents the part of a column
1000 / JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998

J. Struct. Eng. 1998.124:999-1010.

 800
in a regular building frame between the section of maximum
moment and the point of contraflexure. The stub represents a 700 _____________~.Q~ _______ 1OM
discontinuity like a beam-column joint or a footing adjacent ,...----15M
to the section of maximum moment. Table I includes details 600 /' / /
of the test specimens. "
,,----,/ Bar Size 10M 15M 20M
:- 500
,/ Nominal 11.3 16.0 19.5
Concrete and Reinforcement !. Diameter (mm)
iii' 400 I y (MPa) 542 463 454
Ul
Ready-mix normal weight concrete was used. The concrete W
€y 0.0026 0.0024 0.0023
II: 300
strength of each specimen (Table 1) was obtained by averaging In E(MPa) 210080 193250 197580
three standard cylinder tests' results. Three different types of 200 0.0191 0.0207 0.0067
€.h
reinforcing bars were used to construct specimens. 10M (100 lu(MPa) 683 648 700
mrn2) and 15M (200 mrn2) bars were used for the ties and 100 0.1550 0.1130 0.1290
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

€u
20M (300 mrn2 ) size was used for the longitudinal bars. Stress-
strain curves of steel in tension are given in Fig. 2. Each curve 0.05 0.10 0.15 0.20
represents an average of three test results. STRAIN, (mmlmm)
FIG. 2. Stress-Strain Behavior of Steel
Reinforcing Cages
2
The reinforcement for the stub consisted of 10M (100 mm )
lateral and longitudinal stirrups at 64 mm spacing. The lon- justment was required, but when necessary the specimen was
gitudinal bars in columns were extended through the stub to unloaded for adjustment. After the final positioning, the align-
20 mm from the ends in all specimens. The ties were placed ment was checked to the maximum predetermined axial load,
at predetermined spacing within the 91O-mrn-long test regions and the lateral load actuator was connected to the specimen
of the columns adjacent to the stubs. Beyond that point, the after applying the full axial load. The specimen was then sub-
spacing was reduced to half of the predetermined value to jected to predetermined displacement excursions (Fig. 4). All
provide extra confinement and to reduce chances of failure. specimens were first loaded downward (with respect to the test
This was an extra precaution, and in none of the tests were frame shown in Fig. 3). In the first cycle the specimen was
the stresses outside the test region close to being critical. No subjected to 75% of the elastic or yield displacement a lt which
cracks were observed outside of the test region, meaning can be defined as the lateral deflection corresponding to the
strains were very small and confinement steel was not effec- estimated lateral load-carrying capacity (Vrna.) on a straight line
tive. Two different reinforcement configurations (Fig. I) were joining origin and a point ~65% of Vmax on the lateral load-
used in the test regions of the columns. Minimum anchorage displacement curve (Fig. 5). In all the test specimens the as-
of ties conformed to the ACI building code's requirements cending parts of the lateral load-displacement envelopes were
(1995). No anchorage failure was observed in any specimen. linear up to at least 70% of the maximum lateral load, and
hence the method used was essentially the initial tangent ap-
Instrumentation proach. It should be recognized that both a 1 and Vmax were
calculated using the theoretical sectional response of the un-
Concrete and reinforcement strains at various locations, de- confined column and integrating curvatures along the length
flections along the specimen length, and axial and lateral loads of the specimen. Moment generated by the axial load was not
were monitored during each test through the use of extensive considered in the calculation of a 1 since this effect is minimal
instrumentation. Fig. 1 shows the locations of strain gauges in the earlier stages of column response. The specimen was
on longitudinal and lateral reinforcement, as well as the linear subjected to increasing reversed cyclic displacement excur-
variable differential transducer (LVDT) arrangement. Two sets sions until it was unable to maintain the originally applied
of ties in the column closest to the stub were instrumented axial load.
with strain gauges. Longitudinal concrete strains in the core
were measured by using LVDTs over gauge lengths that
ranged from 51 to 102 mrn and covered a length of -460 mm TEST RESULTS
from the column-stub interface. Longitudinal strains at the top
and bottom of the concrete core were monitored in a segmental Behavior of each specimen is presented graphically in the
form to locate the section showing the largest deformations. form of column shear force-versus-tip deflection and moment-
The first segment from the stub's face was 51 mm wide and versus-curvature relationships. Fig. 6 shows the idealization of
the subsequent four segments were 102 mm wide (Fig. 1). a specimen and the definitions of shear force V and tip de-
Lateral deflections at six locations along the length of the spec- flection a, used in Figs. 7-14. The deflection at the failed
imens were measured using LVDTs. Shear deformations in the section was determined from the measured deflected shape of
plastic hinge region were also measured through the use of the column and was used to calculate secondary moment due
two diagonally placed LVDTs. In each test, a total of 24 to axial load. The moments plotted in Figs. 7-14 are those at
LVDTs were used to monitor axial as well as transverse de- the failed sections of the columns, and they include secondary
formations. moments caused by the axial load. The curvature was calcu-
lated from the deformation readings measured by the upper
Testing and lower LVDTs located in the most damaged region within
the hinging zone. The gauge lengths were kept constant in all
All of the specimens were tested under constant axial load the specimens. Spalling of top and bottom concrete covers,
and reversed cyclic displacement excursions in the test frame yielding of inner and outer ties, and buckling of top and bot-
illustrated in Fig. 3. A 4,450-kN hydraulic jack and a load cell tom longitudinal bars are marked on the graphs in Figs. 7-
of similar capacity were used to apply and measure the axial 10. In all of the specimens failure did not occur at the column-
load. Optical measurement devices were used for the align- stub connection, although this section was subjected to the
ment of the specimen. To check the alignment, axial load was maximum moment. Due to the additional confinement pro-
applied in 250-kN intervals and the strain gauges and LVDTs vided by the stub to the adjacent column section, the failure
were monitored regularly. In most specimens, very little ad- shifted away from the stub.
JOURNAL OF STRUCTURAL ENGINEERING 1 SEPTEMBER 1998/1001

J. Struct. Eng. 1998.124:999-1010.

Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

FIG. 3. Schematic of Test Setup

8r--------------------" cracks formed first in the hinging zone at a distance of

6 ~175-30 mm from the face of the stub and extended further
in later stages toward the stub. The most extensive damage
4 was concentrated at ~20-3o mm from the face of the stub
and extended toward the stub in later stages. Spalling of the

LATERAL LOAD V

-6

LATERAL V~1rI U =t( Vt + vj - )

DISPLACEMENT, ~
Kj =t(Kt+Kj-)
Test Observations
area =
Wj

As in NSC and HSC specimens, the first signs of distress

in all of the tested specimens were the cracks in the top and
bottom concrete covers. On the other hand, the number of (a) Member Ductility Parameters
cracks formed in the first three cycles seemed to be slightly
lower than those in the HSC specimens. In HSC specimens

i
O.8MIrIIU or O.9Mma.r
MOMENT M
having a concrete strength of 72 MPa, the average distance 5, ~ Envelope curve
M IrIIU - ~ .......( (average of both
between flexural cracks ranged between 60 and 100 mm. In
_ ",>~Y-dlrectonS)
UHSC specimens this spacing varied between 80 and 140 mm.
For Specimens AS-5HT, AS-6HT, and AS-7HT top concrete
M/
,I "~
/1 I
" cl>2
Jl</>=-
spalled off suddenly just before the first downward peak of cI>, /" I : cI>,
the fourth cycle (Ii. = 21i.\), and the bottom concrete spalled A.. I I ~/' cP. Jp +
.;,"~-_:Lr2J!j _ cI>/ =t(cI>t + cl>j-)
off at the upward (with respect to the test frame shown in Fig.
3) peak of the same cycle. For Specimen ES-8HT, top concrete CURVATURE 4>
spalled off at the first peak of the fourth cycle (Ii. = 21i.\), and
the bottom concrete severely cracked at the upward peak of
the same cycle and spalled off in the next cycle. The concrete
strain at the time of spalling of cover concrete was 0.0029,
0.0027, 0.0030, and 0.0029 for Specimens AS-5HT, AS-6HT,
AS-7HT, and ES-8HT, respectively. After the sixth cycle (Ii. =
(b) Section Ductility Parameters
3Ii. t ), cracking propagated to the sides of the column followed
by cover spalling at the sides of the specimen. Lateral flexural FIG. 5. Definitions of Ductility Parameters

1002/ JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998

J. Struct. Eng. 1998.124:999-1010.

 tudinal bars after the yielding of both inner and perimeter ties.
A tie spacing-to-Iongitudinal bar diameter ratio of 6 was rec-
ommended as the design limit by Mander et al. (1988a,b) to
avoid premature longitudinal bar buckling. Mau and EI-Mab-
sout (1989) reported that, for cases where the tie spacing-to-
bar diameter ratio is smaller than or equal to 6, the stress-
strain behavior of the steel was identical in tension and
compression. For the specimens reported herein, the tie spac-
ing-to-Iongitudinal bar diameter ratio ranged between 3.5 and
4.5, and hence premature buckling of the longitudinal bars was
not a problem. However, at very large inelastic curvatures it
is impossible to prevent buckling of the longitudinal bars, a
conclusion similar to that drawn by Scribner (1986).
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

ANALYSIS OF RESULTS
Because the behavior of reinforced concrete sections and
200 members is not elastic-perfectly plastic, several definitions for
ductility and deformability are available in the literature. In
150
this study the ductility parameters suggested by Sheikh and
Z V Khoury (1993) are used to evaluate the performance of the
~ 100
~ .1, specimens, which makes the comparison of the current test
~

z·
:E
::I
50 Ddt results to previous test results (Sheikh and Khoury 1993;
Sheikh et al. 1994; Bayrak and Sheikh 1996) more rational.
..J 0 Fig. 5 illustrates these ductility parameters. The definitions for
0
0 most of the ductility parameters illustrated in Fig. 5 are self
~ -50 explanatory; others however need further explanation. NtJ, and

*
II:
« N4> are cumulative displacement and curvature ductility ratios,
w -100 respectively. These ratios can be used to assess the cumulative
Dat
J:
en amount of inelastic deformations, normalized with respect to
-150 V .1,
yield deformations, experienced by a section or a member. W
-200 '--_'--_'--_'--_'--_L.-_L.-_L.-_L.-_L---.J and E are work damage and energy damage indicators, re-
-100 -80 -60 -40 -20 0 20 40 60 80 100
TIP DEFLECTION, L\t (mm) 200 r;:===;::==;==::;:::;:::::::::;-"""'F''i'1Y:::==='=iP--:----,
f; Ps P Ash
400 r - - - - - - - - - - . . - r . < - - - - - - - - - - - - , 150 (MPa) (%) p;, )~
AS-5HT M i 349 kN.m = Z 101.96.740.49 1.70
~

Igzf>2~kNm
300 100

200 Z 50
E :E
::I
~ 100 6 Ol-------/---#'--,I-I-.{-,
~
o
01- - - -H- fJHf-~1l_- - - - l ~ -50
t-='
z o Spalling of top cover II:
«
~ -100 <D Spalling of bot. cover w -100
J:
o en
:E ® Yielding of outer tie
-200 -150
@ Yielding of inner tie
@ Buckling of top bars -200 '- _.l.-_.l.-_.l.-_.l.-_.l.-_L-_L-_L-_L. ~
-300
1>y • Buckling of bot. bars -100 -80 -60 -40 -20 0 20 40 60 80 100
-400 L-_L----.J'-----l_----L_..J..LL--L_...L_....L_...L.---l TIP DEFLECTION, L\t (mm)
-200 -160 -120 -80 -40 0 40 80 120 160 200
CURVATURE, rJ> (xl0- 6 rad/mm) 400 r - - - - - - - - - - , " ' : : - : : : : - - - - - - - - - - ,

1R:;f "26"Nm
AS-6HT
FIG. 7. Behavior of Specimen AS-5HT 300

cover extended from close to the stub for a distance that 200
E
ranged between about 350 and 600 mm in different specimens.
~
z 100
In all of the specimens, during the last cycles, buckling of
longitudinal bars was observed after yielding of diamond- ~
shaped inner ties and perimeter ties, which was an indication t-='
Of-----7f7fJW-f--mlfHH-f-I.({L--/---------1
z
of the commencement of failure. The initiation of buckling of ~ -100
longitudinal bars was determined by visual inspection during o
:E
each test. At this point, maximum concrete compressive strain -200
in the core at the level of longitudinal reinforcement was
0.010, 0.012, 0.009, and 0.008 for Specimens AS-5HT, AS- -300
6HT, AS-7HT, and ES-8HT, respectively. The maximum con- _4001.---IL-......l_----L_.....J..._.LLL-....L_....L_..L._..L....--.J
crete compressive strain in the core just before the failure was -200 -160 -120 -80 ·40 0 40 80 120 160 200
0.016,0.019,0.011, and 0.012 for Specimens AS-5HT, AS- CURVATURE, rJ> (xl0- 6 rad/mm)
6HT, AS-7HT, and ES-8HT, respectively. In each test speci-
men, the failure was accompanied by buckling of the longi- FIG. 8. Behavior of Specimen AS·6HT

JOURNAL OF STRUCTURAL ENGINEERING 1 SEPTEMBER 1998/1003

J. Struct. Eng. 1998.124:999-1010.

 200
AS-3HT, AS-6HT can be compared to AS-2HT, and ES-8HT
l--- P .1
to ES-lHT (Figs. 7, 8, and 10-13). Specimens ES-lHT, AS-
150 '"
Y Effect V
~ Al 2HT, and AS-3HT were tested in an earlier stage of the on-

E?d*
'
Z 3
going experimental research program (Bayrak and Sheikh
::£ 100
;,. 1996). Details of these test specimens along with Specimen
Z 50 AS-4HT are given in Table 1. The variable RA1P used in Table
:::; 1 was defined by Bayrak (1995) as follows:
::;)
..J 0
0
u -.:,hIA~(;c=!)
A
~ -50
R AlP -
-
- PIPo (I)
a:
«

D4*
w -100 It is believed that specimens having the same RAtp ratios and
:I:
en the same type of lateral reinforcement configuration should
-150 V Al
behave in a similar manner. In the expression for RAIP the level
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

-200 "--_'--_.l...-_.l...-_.l...-_.l...-_.l...-_.l...-_.l.-_.l.------l of axial load is presented by the index variable PIPo rather
-100 -80 -60 -40 -20 0 20 40 60 80 100 than PIf;A g • For columns with similar f;, both of these indices
TIP DEFLECTION, At (mm)
provide a similar comparison; whereas for diferent f; values
in columns the comparison may not remain valid with index
400 ,- - - - - - ,- - - - - ~
Plf:A g • The required amount of lateral reinforcement was ob-
AS-7HT M i =351 kN.m 1
served to be proportional to the strength of concrete for a
300
M/=266 kN.m certain column performance if the axial load is measured as a
200
fraction of Po rather than f; A g (Sheikh et al. 1994).
~ Specimens AS-5HT and AS-3HT were tested under very
z 100 similar levels of axial load (PIP o = 0.48 and 0.50, respectively)
::£
and they satisfied the ACI 318-95 requirements for the amount
~
0 of confinement reinforcement approximately to the same de-
t-'
z gree (A'hIA'h(ACt) = 1.09 and 1.19, respectively). As a result,
w
:::; -100 o Spalling of top cover the RAtp ratios are very similar in these two specimens. Both
0 CD Spalling of bot. cover
:::; the sectional and member performances of these two speci-
-200 @ Yielding of outer tie
@ Yielding of inner tie
mens are very similar (Figs. 7 and 13; Table 1). To conclude
-300 @ Buckling of top bars this discussion, the comparison of behavior of Specimen AS-
• Buckling of bot. bars 6HT to that of Specimen AS-2HT is of great significance. Un-
-400 "--_'--_'--_'--_'------1..1.1-_-'--_-'--_-'--_-'-------'
-200 -160 -120 -80 -40 0 40 80 120 160 200 200
CURVATURE, ¢ (xl0· 6 rad/mm)
150
FIG. 9. Behavior of Specimen AS·7HT
Z
::£ 100
spectively. By using these two parameters energy absorption ;,.
and dissipation characteristics of the test specimens were eval- Z 50
:::;
uated. These four parameters can be defined up to the end of ::;)
..J 0
a test (in this case subscript t is used in Table I) or up to the 0
cycle in which the strength loss is ::520% (in this case subscript u
~ -50
80 is used in Table 1). The moment plotted in the moment- a:
curvature diagram included in Fig. 5 includes the secondary «
w -100 E?4l. 3
moments, and hence the reduction of Mm • x to 80 or 90% is
due to section strength degradation. In the equations presented
:I:
en
-150
IT---- ..... ' . . . . . . .
V L\.(
in Fig. 5 LJ is the length of the extensively damaged region,
-200 L . ~ L - . . . . J _ . . . J . _ - L _ - L _ - L _ . . L _ - L - _ . L - _
and h is the height of the column section.
-100 -80 -60 -40 -20 0 20 40 60 80 100
A general study of Figs. 7 -10 indicates that UHSC columns
TIP DEFLECTION. At (mm)
having concrete strengths around 102 MPa can be made to
behave in a ductile manner under high levels of axial load, 400,------------,-----------
provided that a sufficient amount of confining reinforcement ES-8HT
is used in an efficient configuration. AS-6HT, which had 70% 300
more reinforcement than the ACI 318-95 requirements, be-
haved in a very ductile manner, showing a curvature ductility
factor fL<l>80 of 14.0 and a displacement ductility factor fLd of
6.3. The noticeable differences between the responses of the
test specimens (Figs. 7 -1 0) indicate that confinement is af-
E
~
~
200

100

01------7"""'"-1--
El
fected greatly by different variables. It should be recognized t-'
z
that sectional behavior represented by the M-<j> relationship is ~ -100
o Spalling of top cover
of primary concern here because the deformations concentrate o CD Spalling of bot. cover
:::; @ Yielding of outer tie
at the plastic hinge once the column is loaded in the postelastic -200
@ Buckling of top bars
range. Further lateral displacements will take place mainly as
-300 • Buckling of bot. bars
a result of plastic hinge rotation.
-400 L-_.l...-_.l...-_.l...-_.l...-----L.L.L.._.l...-_.l...-_.l...-_.l...---'
Effect of Concrete Strength -200 -160 -120 -80 ·40 0 40 80 120 160 200

To evaluate the influence of concrete strength on column CURVATURE, ¢ (xl0· 6 rad/mm)

behavior, Specimen AS-5HT can be compared to Specimen FIG. 10. Behavior of Specimen ES·8HT

10041 JOURNAL OF STRUCTURAL ENGINEERING I SEPTEMBER 1998

J. Struct. Eng. 1998.124:999-1010.

 200
specimens contained 8-9% more reinforcement than the ACI
150
318-95 code's requirements and were tested under very similar
axial load levels. It can be observed from Figs. 7 and 10 that
Z AS-5HT having an N<j>so value of 27 and an N<j>, value of 49
~ 100 V
;::.
Z
:E
50
T~
tt behaved in a considerably more ductile manner than ES-8HT
(N<j>so = 14, N<j>, = 22). Furthermore, the curvature ductility fac-
:::l
tor !J..<j>so of AS-5HT is 43% larger than that of ES-8HT. Similar
..I 0 comparisons can be made using all of the member and section
0
U ductility parameters listed in Table 1; however, the most im-
~ -50 portant evidence that shows weaker behavior of Specimen ES-
II:
8HT is its poor energy absorption and dissipation capacity. The
<C
w -100
J:
(J)
EP-4-L '--
------ tT --- energy dissipated in Specimen AS-5HT, measured by the en-
-150 ergy damage indicator E so , is 4.4 times as much as the energy
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

V.1, dissipated in ES-8HT. Better distribution of reinforcement and

-200 '--_'--_'--_'--_'--_'--_'---_'---_'---_.L.---.l better lateral support to the longitudinal bars provided tougher
-100 -80 -60 -40 ·20 0 20 40 60 80 100 response of UHSC columns, an effect similar to the ones ob-
TIP DEFLECTION, .1 t (mm) served for NSC and HSC columns. At this point it should be
400,-----------,--------------, recognized that the current ACI code's (1995) requirements
ES-1HT for the design of confinement reinforcement do not relate the

D
300 amount of lateral reinforcement to reinforcement configura-
tion. The use of configurations A and E is allowed without
200 any differentiation between the efficiency of confinement in
E
~
z 100
each configuration.

::?1 Effect of Amount of Lateral Reinforcement

Of-7#',JLlm".~!
t-='
z Specimens AS-5HT and AS-6HT can be compared to eval-
~ -100 o Spalling of top cover uate the effect of the amount of lateral reinforcement on the
o CD Spalling of bot. cover behavior of UHSC columns. An increase in the amount of
:E @ Yielding of outer tie
-200 lateral reinforcement significantly improved the cyclic behav-
@ Buckling of top bars
• Buckling of bot. bars
ior of the specimen (Figs. 8 and 9). An increase of 67% in the
-300
volumetric ratio of tie reinforcement-to-concrete core resulted
-400 '--_'---_'---_'---_'----L..L.L_'---_'---_.L.-_.L.---.l
-200 -160 -120 -80 -40 0 40 80 120 160 200

FIG. 11.
CURVATURE, ¢ (xl0- 6 rad/mm)

Behavior of Specimen ES·1 HT

z
150
------T
'11..-7_
p~
---i-Effect
84*v.1,
~ 100
like Specimens AS-3HT and AS-5HT that were tested under ;::. ~UJ!sk~-
similar levels of axial load and obeyed the code's requirements z 50
:E
for confinement reinforcement to the same degree, Specimens :::l
..I
AS-6HT and AS-2HT were tested under different levels of o
U
axial load and contained different amounts of lateral reinforce-
~ -50
ment as indicated by AshIAsh(ACI)' Specimen AS-2HT was tested II:
under a moderate axial load level (PIPo = 0.36) and had 19% <C
w -100
J:
more reinforcement than the code's requirements. Specimen (J)

AS-6HT, on the other hand, was tested under a higher level -150
of axial load (PIPo = 0.49) and contained 70% more reinforce-
ment than the ACI 318-95 requirements for confinement re- -200 '--_'---_'--_'-----''----..JL..---..JL..---..J_--..J_--..J_........J
-100 ·80 -60 -40 -20 0 20 40 60 80 100
inforcement. As a result the RA1P ratios are very similar in these
TIP DEFLECTION, .1 1 (mm)
two specimens. An examination of the moment-curvature be-
havior of these two specimens (Figs. 8 and 12) and a com- 400,-------------.,,-------------.,
parison of ductility parameters in Table 1 indicate that despite AS-2HT

~
the differences in their concrete strength, these specimens dis- 300
played a very similar behavior. An examination of member
and section ductility parameters for these specimens (Table I) 200
E
indicates that the higher strength concrete specimens have
lower deformability, energy absorption, and dissipation capac- ~ 100
ities initially, but during the latter part of the displacement ::?1
excursions, these properties improve rapidly and the total val- t-='
z
ues are comparable to those of lower strength concrete spec- ~ -100
imens. Comparison of sectional and member behavior of Spec- o
:E
imens ES-lHT and ES-8HT (Figs. 10 and 11; Table 1) yields -200
to the same conclusion.
-300
Effect of Reinforcement Configuration -400 '---_.L.-_.L.-_.L.-_.L.--LLl-_'---_'---_'---_'---..J
-200 -160 -120 -80 -40 0 40 80 120 160 200
The effect of reinforcement configuration on the cycle be-
CURVATURE, ¢ (xlO- 6 rad/mm)
havior of UHSC columns can be evaluated by comparing the
behavior of Specimens AS-5HT and ES-8HT. Both of these FIG. 12. Behavior of Specimen AS-2HT

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998/ 1005

J. Struct. Eng. 1998.124:999-1010.

 200

150 150

Z z
:::. 100 :::. 100
;;.. ;:..
Z 50 Z 50
::;; ::;;
:::l
...J 0 :3o ~+- l,-(.r - lo
0
() ()

~ -50 ~ -50
a: a:
<I: <I:
w -100 w -100
J: J:
en en
-150 -150
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

-200 .200 '--_'---..J_---l_.....L_--L_....L_-l...._..l-_..L.----l

-100 -80 -60 -40 -20 0 20 40 60 80 100 -100 -80 -60 -40 -20 0 20 40 60 80 100
TIP DEFLECTION, fl., (mm) TIP DEFLECTION, fl., (mm)

400 400 r - - - - - - - - - - - , - - - - - - - - - - - ,
AS-3HT AS-4HT
300

~ ~
300

200 200
E E
z ~
:::. 100 o Spalling of top cover 100

~ ~
0
.-: .-:
z z
w
::;; -100 ~ -100 o Spalling of top cover
0 o CD Spalling of bot. cover
::;; ::;;
-200 -200 @ Yielding of outer tie
@ Yielding of inner tie
-300 -300 @ Buckling of top bars
• Buckling of bot. bars
-400 L - _ ' - _ ' - _ ' - _ ' - - L l . . . . L _ ' - _ ' - _ ' - _ L . . - - - - - l -400 L---l_--..l_......l._-l.._.L..UL-...L_...L._..L._.L.-----l
-200 -160 -120 -80 -40 0 40 80 120 160 200 -200 -160 -120 -80 -40 0 40 80 120 160 200
CURVATURE, 1> (x10- 6 rad/mm) CURVATURE, 1> (x1O- 6 rad/mm)

FIG. 13. Behavior of Specimen AS-3HT FIG. 14. Behavior of Specimen AS-4HT

in pronounced improvements of energy dissipation capacity,

deformability, and ductility (Table 1, Specimens AS-5HT and
AS-6HT). Both the stiffness degradation and strength reduc-
tion rate with every displacement cycle were lower for AS-
6HT compared to AS-5HT. Specimen ES-1HT Specimen AS-2HT
Specimens AS-3HT and AS-4HT can be compared to eval-
uate the effect of the amount of lateral reinforcement on the
behavior of HSC columns. Volumetric ratio of tie reinforce-
ment-to-concrete core Ps for AS-4HT is 80% greater than that
of AS-3HT. An increase in the amount of lateral reinforcement
significantly improved the behavior of the specimen (Table 1,
Specimens AS-3HT and AS-4HT). The section moment ca-
pacity was stilI increasing after the spaIIing of cover concrete Specimen AS-3HT Specimen AS-4HT
in Specimen AS-4HT, an indication of the excellent confine-
ment of concrete core. It is important to recognize that the
observations made for HSC and UHSC columns are similar to
those that were made for both NSC columns (Sheikh and
Khoury 1993).

Stub Effect Specimen AS-5HT Specimen AS-6HT

Although the maximum moment occurred at the column-

stub interface in alI the test specimens, damage initiated away
from this section. It seemed that the strength of the section
at the column-stub connection was higher than that of the
Failed section
failed section because of the additional confinement provided
by the stub. The additional confinement provided by the stub Specimen AS-7HT Specimen ES-8HT
caused a delay in the spreading of cracks in concrete and re-
duced the tendency of lateral expansion. As a result, the mo- All dimensions in mm
ment capacity of the critical section increased and failure FIG. 15. Extensively Damaged Regions of Test Specimens

10061 JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998

J. Struct. Eng. 1998.124:999-1010.

 s a b
shifted to a nearby section. Fig. 15 shows the most damaged
regions of the columns in all of the specimens.
Experimental moment capacities Mexp of the most critical I
I~damge
k~:nes1 v
sections in the extensively damaged regions are presented in
Table 2. M i is the theoretical moment capacity calculated using --------1'
the HSC stress-strain relationship (Collins and Mitchell 1991)
and actual steel properties. M ACI is the moment capacity cal- p
culated using the ACI code's (1995) provisions for stress block
and actual steel properties. It was observed during testing that
cover concrete spalled off suddenly in most of the specimens
and that the concrete strain at the time of spalling of the top
cover was in the range of 2.4 X 10- 3 -3.0 X 10- 3 • To evaluate
the effect of confinement after the spalling of the cover, M i (a) Column Idealization
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

and M ACI values are calculated both in the presence and in the
absence of the top cover. The Mexp values with intact top con- Moment capacity
crete cover generally represent the unconfined section capac- envelope
ities. Before spalling of the cover, Mexp values are very similar
to M; values. M ACI values, however, are significantly higher
(10-29%) than the experimental capacities. After the spalling
of the cover, Mexp values represent the confined section capac-
ities, and they are 11-45% higher than the M: values without ---.L
C'Mi'
--------J------
I
M'i
T t~C
a top cover.
Fig. 16 illustrates an idealization made for a qualitative eval- M '= (1+C+S)M',.T r- :; ~I ~- ~- l
s ['-First crack
uation of the stub effect on the moment capacity of the adja- +
cent column section, and moment capacity envelope diagrams.
Ignoring the effects of confinement as well as stub, the mo-
ment capacity of the unconfined concrete section without a top M; =~ (M;''"+M;') c =J..(c++c)
2
concrete cover may be assumed to be equal to the theoretical = ~ (s+ +.5")
Mee = ~ (Me/+Mcc') s
capacity M:. Due to confinement, the section moment capacity
would increase to M ee = (1 + C)M;, where C is the confine-
(b) Moment capacity envelope diagrams
ment effect factor due to ties. The additional enhancement in
moment capacity SM; due to stub would vary from zero to FIG. 16. Quantitative Evaluation of Stub Effect
maximum at the stub's face as suggested by Khoury and
Sheikh (1991). Therefore S is the additional confinement effect TABLE 3. Degree of Enhancement in Moment Capacities
factor due to stubs (Fig. 16). At this point it should be rec-
ognized that the method used to evaluate Sand C is not a Stub
precise analysis technique. It is rather, a technique used to Moment Enhancement Confinement Effect
evaluate and compare relative magnitudes of the confining ef- Ratios Effect Factor Factor
fects arising from two different sources, i.e., the lateral rein- Specimen MjM; M./M; MilM; C S C+S
forcement and the stub. (1 ) (2) (3) (4) (5) (6) (7)
It may be assumed that the failure is approximately initiated AS-5HT 1.517 1.411 L189 0.189 0.328 0.517
at a distance La + L b from the stub, where the stub effect AS-6HT 1.511 1.444 1.306 0.306 0.205 0.511
diminishes (Section bob in Fig. 16). It may also be assumed AS-7HT 1.350 1.309 1.067 0.067 0.283 0.350
that the additional restraint provided by the stub may reduce ES-8HT 1.444 1.380 L190 0.190 0.254 0.444
gradually after the failure is initiated and that the zone of fail- ES-lHT 1.507 1.380 1.180 0.180 0.327 0.507
ure may extend from Section b-b towards the stub with an AS-2HT 1.398 1.348 1.208 0.208 0.190 0.398
AS-3HT 1.561 1.481 1.282 0.282 0.279 0.561
extensive damage zone concentrated between Sections a-a and AS-4HT 1.588 1.506 1.304 0.304 0.284 0.588
b-b, as shown in Fig. 16. The distance La varied from 50 to Average 1.485 1.407 1.206 0.216 0.269 0.485
150 mm, and L b varied between 250 and 400 mm in HSC and
UHSC specimens (Fig. 15).
The degree of enhancement in section moment capacities of
the tested specimens is analyzed and the results are presented
TABLE 2. Moment Capacities of Specimens in Table 3. Enhancement in the moment capacity due to con-
finement is in the range of 7 - 31 % of the theoretical moment
Top Concrete Cover in Top Concrete Cover capacity M;. Stub restraint increased the capacity further by
Place Spalled Off at least 19-33%. The maximum moment experienced by the
Speci- M{ MACl M.xp M; MACl M.xp M: columns at the column-stub interface ranged from 1.35 to 1.59
men (kN· m) (kN· m) (kN· m) (kN'm) (kN· m) (kN· m) (kN· m) times the theoretical capacity M;. This moment did not cause
(1 ) (2) (3) (4) (5) (6) (7) (8) failure in any of the specimens. It should be noted that for this
AS-5HT 349 419 367 265 322 318 402 analysis Section b-b (Fig. 16) was considered to be the sec-
AS-6HT 349 416 361 262 316 356 396 tion where failure initiated, and the effect of stub restraint was
AS-7HT 351 419 326 266 323 296 359 minimal. Furthermore, Specimens AS-2HT and AS-3HT were
ES-8HT 350 416 342 261 315 321 377 analyzed using a plasticity-based finite-element analysis tech-
ES-IHT 274 310 272 205 237 274 309 nique. This analysis was performed using ABAQUS (ABAQUS
AS-2HT 287 314 286 231 260 282 323
274 310 205
1992), and the material model was coded as user-defined sub-
AS·3HT 279 238 274 320
AS-4HT 274 310 266 204 237 296 324 routine. Predictions for La and L b were 98 and 103% of the
values shown in Fig. 15 and predictions C and S were in rea-
"Maximum experimental moment at section adjacent to stub as an av-
erage of both load directions.
sonable agreement with the values listed in Table 3. Further
discussion about this part of the work is omitted due to space
JOURNAL OF STRUCTURAL ENGINEERING 1 SEPTEMBER 1998/1007

J. Struct. Eng. 1998.124:999-1010.

 TABLE 4. Comparison of Predicted and Observed Curvature Ductility Factors
Properties Curvature Ductility Factors, fJ.<I>SO

~f Aoh P Prediction by HSC fJ.<I>SOHSC Prediction by NSC fJ.<I>SONSC

Specimen (MPa) Aoh(ACI) Po Observed fJ.<I>SOEXP formulation, !-480HSC fJ.<I>soexp formulation, fJ.<I>80NSC fJ.<I>8OEXP
(1 ) (2) (3) (4) (5) (6) (7) (8) (9)
AS-5HT 101.8 1.09 0.48 9.6 10.1 1.05 15.7 1.64
AS-6HT 101.9 1.70 0.49 14.0 16.8 1.20 22.6 1.61
AS-7HT 102.0 0.80 0.48 7.2 6.9 0.96 12.0 1.67
ES-8HT 102.2 1.08 0.50 6.7 6.7 1.00 6.7 1.00
ES-1HT 72.1 1.13 0.50 6.6 6.6 1.00 7.0 1.06
AS-2HT 71.7 1.19 0.36 15.8 14.5 0.92 20.4 1.29
AS-3HT 71.8 1.19 0.50 10.1 10.5 1.04 16.2 1.60
AS-4HT 71.9 1.83 0.50 21.2 17.7 0.83 23.5 1.11
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

Average - - - 11.4 11.2 1.00 15.5 1.37

limitations and is beyond the scope of this paper. However, it AS-7HT, and ES-8HT. As suggested by Sheikh and Khoury
should be appreciated that this technique was used to evaluate (1997) the required amount of lateral reinforcement in tied
the stub effect of the specimens tested in this study and may columns was assumed to be
not provide accurate results for other specimens.
Ash = [Ash(Acn]' a . Y p ' Y<j> (2)

Effect of Axial Load where ~sh(ACI) = total cross-sectional area of rectilinear ties as
suggested by the ACI 318-95 code; a = configuration effi-
Effect of axial load is evaluated using HSC columns. Spec- ciency factor [a = 1 when all longitudinal bars are laterally
imens AS-2HT and AS-3HT (Figs. 12 and 13) were identical, supported by tie bends, a > 1 when some of the longitudinal
containing the same amount of lateral reinforcement that was bars are not supported by tie bends, e.g., for NSC columns
19% more than the ACI code's (1995) requirements. Specimen having E-type reinforcement configuration a was suggested to
AS-2HT was tested under a moderate level of axial load (PI be 2.5 (Sheikh and Khoury 1997)]; Yp = parameter to take into
Po = 0.36); whereas AS-3HT was tested under a higher axial account the effect of axial load; and Y<j> = parameter to take
load level (PIPo = 0.50). Hence, the axial load level is the only into account section ductility demand.
variable for the comparative analysis of these two HSC spec- The expressions suggested for Yp and Y<j> are kept in the
imens. An increase in axial load caused substantial reductions same form as the original model as follows:
in the curvature ductility factors, JoL<j>so and JoL<j>90' Moreover, the
cumulative curvature ductility ratios showed significant reduc-
tions, from 53 to 20 for N<j>so and 113 to 42 for N<j>" as a result
Yp = al + a2' (:J a3 (3)

of increased load. Similar conclusions can be drawn from the

comparison of the energy damage indicators, E so and E4 ; in
Y<jo = a4' (JoL<I>8o)as (4)

fact, energy damage indicators appear to be affected the most where a, to as = constants that were evaluated performing a
compared to the other parameters. The energy dissipated in regression analysis using results from test series in which f;
Specimen AS-2HT measured by E so and E, is 3.6-4 times as ranged from 30 to 55 MPa.
much as the energy dissipated in Specimen AS-3HT. Reduc- In the application to specimens having concrete strengths
tions of 19, 54, and 67% were observed in displacement duc- varying between 72 and 102 MPa, it was found that constants
tility factor JoL.1.' cumulative displacement ductility ratio N.1." ai' a2, and a3 remained the same and constants a4 and as
and work damage indicator W" respectively, due to the in- changed. In other words, the effect of axial load on the lateral
crease in axial load. A higher axial load resulted in an increase reinforcement demand is not influenced significantly by con-
in the rate of stiffness degradation with every load cycle and crete strength. Section ductility demand measured by JoL<j>80, on
adversely affected the cyclic performance of HSC columns the other hand, is influenced by concrete strength. This is be-
(Bayrak and Sheikh 1996). These results underlined the need cause higher strength concrete specimens have lower deform-
to incorporate the level of axial load in computing the required ability and energy absorption capacities initially. These prop-
amount of confining reinforcement. erties improve considerably during the latter part of the
displacement excursions as discussed previously. Curvature
CONFINEMENT REINFORCEMENT REQUIREMENTS ductility factor JoL<j>so does not reflect the total behavior of a
FOR HSC COLUMNS column section. Instead, it focuses only upon the part of a
moment-curvature relationship in which the strength loss is
Sheikh and Khoury (1997) proposed a procedure for the smaller than or equal to 20% of the maximum bending mo-
design of confinement reinforcement, for a given ductile per- ment. For NSC columns a4 and as were found to be equal to
formance that takes into account parameters such as reinforce- 1/29 and 1.15, respectively. Using the results of four HSC
ment distribution and axial load level. This method is sug- specimens and four UHSC specimens a least-squares analysis
gested for the design of columns that have concrete strengths yielded in a4 = 1/8.12 and as = 0.82. With these new constants,
up to 55 MPa. Accuracy of this procedure is checked against the average of the predicted curvature ductility ratios is
the results from this experimental program. It is found that the roughly equal to the average of the experimental values, and
average of all eight predictions made by the design equations the standard deviation from the mean is 10%. Furthermore,
suggested by Sheikh and Khoury (1997) (for Specimens ES- the value of configuration efficiency factor a for E-type HSC
IHT, AS-2HT, AS-3HT, AS-4HT, AS-5HT, AS-6HT, AS-7HT, and UHSC columns is calculated to be equal to 1.35. At this
and ES-8HT) is 37% larger than the average experimental val- point it should be recognized that regression analysis for a4
ues (Table 4). Standard deviation from the mean was 29% for and as were performed only on eight columns having concrete
the HSC and UHSC specimens. To improve the predictions, a strengths ranging from 72 to 102 MPa. Constants determined
regression analysis was performed using results from Speci- in this study by regression analysis may be refined with the
mens ES-IHT, AS-2HT, AS-3HT, AS-4HT, AS-5HT, AS-6HT, availability of more experimental data, but it is believed that
1008/ JOURNAL OF STRUCTURAL ENGINEERING 1 SEPTEMBER 1998

J. Struct. Eng. 1998.124:999-1010.

 the general form of the equation would remain unchanged. Propertle. of 500mm x 500mm column MCtlon:
Regression analyses for 01> 02' and 03 were performed on 17 1,,'= 110 MPa ~
columns having concrete strengths ranging from 30 to 102 6 1,= 460 MPa A
MPa. The following equation summarizes the suggested de- Concrete cover" 25 mm
A-type .teel configuration
sign procedure for the design of confinement reinforcement Longhudlnal reinforcement: 840M
5
for HSC columns with/; in the range of 55 MPa to over 100 =
Tie spacing 100 mm
MPa:
Maximum perml••lble - - - -.
4
axial load ratio, AC1318-96
P )'] [(,",,<1>80)0,82]
A.h = [A.h(ACnl· a' [ 1 + 13· ( Po . 8.12 (5) 4>P,.(IllU} lp,

As a first impression (5) may look complicated to a design

practitioner, but it requires little extra effort than using the ACI
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

code equation. The designer would generally know the level

of axial load measured by PIPo for the column under consid-
eration and should be able to assess the curvature ductility
factor needed, e.g., based on the response reduction or the
modification factor used in the calculation of the design forces.
Factor a, on the other hand, can be taken as unity for columns
in which a minimum of three longitudinal bars are effectively
supported by tie corners on each face of the column and hooks FIG. 17. Confinement Reinforcement Requirements for HSC
are anchored into the core concrete (Sheikh and Khoury 1997). Columns (f~ > 55 MPa)
It is recommended that for HSC and UHSC columns that need
to exhibit highly ductile performance, configurations that will it is obvious that the behavior of compressed concrete would
produce 0: = 1 should be used. improve significantly if it is confined properly regardless of
To compare the suggested design procedure to the existing the level of axial load. Fig. 17 shows that even for axial load
design codes, confinement reinforcement design for a 500 X levels around O. 2APIPo), the required amount of lateral rein-
500 mm column is shown in Fig. 17. The concrete strength is forcement is larger than that required by the ACI code, if
taken to be equal to 90 MPa; the yield strength of both lon- highly ductile performance is needed.
gitudinal and lateral reinforcement is taken to be equal to 450
MPa. An A-type reinforcement configuration that provides ex- CONCLUDING REMARKS
cellent lateral support to longitudinal bars by the tie bends is
used for this particular column. Spacing of tie sets is selected Four UHSC column-stub specimens made of 102-MPa con-
to be equal to 100 mm. The area of confinement reinforcement crete were tested under high axial load levels and reversed
required by the ACI 318-95 code, the NZS 3101:1995 code, cyclic lateral displacement excursions. Results from earlier
and the suggested design procedure are calculated. The results tests on four similar HSC specimens made of 72-MPa concrete
are normalized with respect to ACI code's requirements. At are also included and used in evaluating the effect of concrete
this point it should be recognized that the ACI 318-95 provi- strength on the response of tied columns. The following con-
sions for confinement reinforcement results in a constant re- clusions can be drawn from the current research reported in
inforcement area for different levels of axial load and different this paper:
section performance as indicated by the curvature ductility fac-
tor ,",,<1>80' • UHSC columns that are made with 102-MPa concrete can
It can be observed from Fig. 17, that the column designed be made to behave in a ductile manner under high levels
according to ACI 318-95 code's provisions is able to supply of axial load, provided that a sufficient amount of lateral
a curvature ductility factor of -8 if the axial load on this reinforcement is used in an efficient configuration. A
column is 0.5Po• It is stated in the NZS 3101:1995 code that UHSC column specimen made with 102-MPa concrete
the aim is to obtain a curvature ductility factor of at least 10. and having 70% more reinforcement than the code's re-
Sheikh and Khoury (1993) suggested that columns that are quirements displayed a very ductile behavior showing a
able to supply a curvature ductility factor of at least 16 can curvature ductility factor of 14.0 and a displacement duc-
be considered as highly ductile columns. At this point it should tility factor of 6.3. As in NSC and HSC specimens, be-
be understood that the ductility factor used in the New Zealand havior of UHSC columns subjected to constant axial load
code uses a secant approach to determine the yield curvature, and reversed cyclic lateral load is substantially influenced
whereas in the study conducted by Sheikh and Khoury (1993) by the confinement provided by rectilinear ties.
a different approach is used to calculate the yield curvature. • An increase in axial load reduces the column's deforma-
A curvature ductility factor of 10 calculated using NZS bility and ductility and accelerates strength and stiffness
3101: 1995 definitions is approximately equivalent to a cur- degradation with every load cycle. To compensate for this
vature ductility factor of 16-20 calculated using the defini- effect a larger amount of lateral reinforcement is required.
tions presented in Fig. 5. If the confinement reinforcement in Therefore, the axial load level should be incorporated as
the sample HSC column under consideration were designed a design parameter in the design of confinement reinforce-
according to ACI provisions, it would not be able to provide ment.
a curvature ductility factor of 16 for small PIPo values. Hence, • An examination of member and section ductility param-
columns designed according to the current ACI code can dis- eters for comparable specimens indicates that the higher
play a wide range of behaviors from very ductile to brittle, strength concrete specimens have lower deformability and
depending on the level of axial load and the efficiency of the energy absorption and dissipation capacities initially, but
reinforcement configuration. during the latter part of the displacement excursions, these
All of the eight columns used in this study were tested under properties improve rapidly and the total values are com-
moderate to high axial load levels. Therefore it is believed that parable to those of lower strength concrete specimens.
the equation suggested here should be checked against data • UHSC columns having E-type reinforcement configura-
from columns tested under lower axial load levels. However, tion (middle longitudinal bars are not supported by tie
JOURNAL OF STRUCTURAL ENGINEERING I SEPTEMBER 1998/1009

J. Struct. Eng. 1998.124:999-1010.

 bends) have lower energy dissipation and deformation ca- "Code of practice for the design of concrete structures," (l995b). Part
pacities than those having A-type reinforcement configu- II: Commentary on the design of concrete structures, NZS 3101:1995,
New Zealand Standards, Wellington, New Zealand, 264.
ration (all eight longitudinal bars are supported by tie Collins, M. P., and Mitchell, D. (1991). Prestressed concrete structures.
bends). Better distribution of reinforcement and better lat- Prentice-Hall, Inc., Englewood Cliffs, N.J., 766.
eral support to the longitudinal bars provided tougher re- Khoury, S. S., and Sheikh, S. A (1991). "Behavior of normal and high
sponse of UHSC columns, an effect similar to the ones strength confined concrete columns with and without stubs." Res. Rep.
observed for NSC and HSC columns. ACI code's (1995) No. UHCEE 91-4, Dept. of Civ. and Envir. Engrg., Univ. of Houston,
Tex., 345.
requirements for the design of confinement reinforcement Mander, J. B., Priestley, M. J. N., and Park, R. (l988a). "Observed stress-
do not relate the amount of lateral reinforcement to re- strain behavior of confined concrete," J. Struct. Engrg., ASCE, 114(8),
inforcement configuration and the level of axial load. 1827-1849.
However, current experimental results suggest that rein- Mander, J. B., Priestley, M. J. N., and Park, R. (l988b). "Theoretical
forcement configuration and axial load level must be con- stress-strain model for confined concrete," J. Struet. Engrg., ASCE,
sidered in the design of confinement reinforcement. 114(8), 1804-1826.
Mau, S. T., and EI-Mabsout, M. (1989). "Inelastic buckling of reinforcing
Downloaded from ascelibrary.org by UNIVERSIDADE DE BRASILIA on 05/11/14. Copyright ASCE. For personal use only; all rights reserved.

bars." J. Engrg. Mech., ASCE, 115(1), 1-17.

ACKNOWLEDGMENTS Mitchell, D., and Paultre, P. (1994). "Ductility and overstrength in seis-
mic design of reinforced concrete structures," Can. J. Civ. Engrg.,
The research for this study was supported by grants from the Natural Ottawa, Canada, 21(6), 1049-1060.
Sciences and Engineering Council of Canada. The experimental work was Paulay, T. (1977). "Seismic design of ductile moment resisting reinforced
performed in the Structural Testing Laboratories of the University of To- concrete frames, columns: Evaluation of actions." Bull. New Zealand
ronto, Toronto, Ontario, Canada. Nat. Soc. for Earthquake Engrg., Waikanae, New Zealand, 10(2),
85-94.
Scribner, C. F. (1986). "Reinforcement buckling in reinforced concrete
APPENDIX. REFERENCES flexural members." ACI J., 83(6),966-973.
Sheikh, S. A, and Khoury, S. S. (1993). "Confined concrete columns
ABAQUS User's Manual Version 5.2. (1992). Hibbit, Karlsson & Sor- with stubs." ACI Struet. J., 90(4), 414-431.
ensen Inc. Sheikh, S. A., and Khoury, S. S. (1997). "A performance-based approach
Bayrak, O. (1995). "High strength concrete columns subjected to earth- for the design of confining steel in tied columns," ACI Struet. J., 94(4),
quake type loading," MASc thesis, Dept. of Civ. Engrg., Univ. of To- 421-431.
ronto, Toronto, Ont., Canada, 239. Sheikh, S. A., Shah, D. V., and Khoury, S. S. (1994). "Confinement of
Bayrak, 0., and Sheikh, S. A (1996). "Confinement steel requirements high-strength concrete columns," ACI Struct. J., 91(1), 100-111.
for high strength concrete columns." Proe., 11 th World Conj. on Earth- Sheikh, S. A., and Toklucu, M. A. (1993). "Reinforced concrete columns
quake Engrg., 8. confined by circular spirals and hoops." ACI Struct. J., 90(5), 542-
"Building code requirements for reinforced concrete and commentary." 553.
(1995). ACI 318-95/ACI 381R-95, Am. Concrete Inst., Detroit, Mich., Sheikh, S. A., and Uzumeri, S. M. (1980). "Strength and ductility of tied
369. concrete columns," J. Struet. Div., ASCE, 106(5), 1079-1102.
"Code for design of concrete structures for building," (1994). CAN3- Watson, S., and Park, R. (1994). "Simulated seismic load tests on rein-
A23.3-M94, Can. Standards Assn., Rexdale, Ont., Canada, 199. forced concrete columns," J. Struct. Engrg., ASCE, 120(6), 1825-
"Code of practice for the design of concrete structures," (l995a). Part 1849.
1: The design of concrete structures, NZS 3101:1995, New Zealand Watson, S., Zahn, F. A., and Park, R. (1994). "Confining reinforcement
Standards, Wellington, New Zealand, 256. for concrete columns," J. Struct. Engrg., ASCE, 120(6), 1798-1824.

1010/ JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998

J. Struct. Eng. 1998.124:999-1010.