Title no.

86-S1

Response of Reinforced Concrete Columns to Simulated

Seismic Loading

by Murat Saatcioglu and Guney Ozcebe

Response of reinforced concrete columns to seismic loading was in- Hospital, damaged during the 1971 San Fernando
vestigate.6 experimentally. Full-scale columns were tested under slowly earthquake, are additional examples of structural fail-
applied lateral load reversals. Both unidirectional and bidirectional
loadings were included. Columns with and without axial loads, in-
ures attributed to poor column behavior.
cluding variable axial tension and compression, were tested. The test An extensive experimental investigation was con-
parameters included: axial load, shear/confinement reinforcement, ducted by the authors to study the response of rein-
and deformation path. forced concrete columns to simulated seismic loading.
Test results indicate that axial loads have a significant influence on The research program included tests of 14 full-size col-
hysteretic response of columns. Column capacity changes with the
concurrent level of axial load, axial compression reduces column re-
umns subjected to uniaxial and biaxial load reversals.
sponse, and bidirectional load cycles can produce severe strength and The column geometry was selected such that, while
stiffness degradation. premature shear failure prior to flexural yielding was
prevented as required by the current practice, the col-
Keywords: axial loads; biaxial loads; columns (supports); cyclic loads; defor- umns were subjected to relatively high shear stresses.
mation; ductility; earthquake-resistant structures; reinforced concrete; struc-
tural design; tests.
The shear span-to-depth ratio was 2.85. The test pa-
rameters included: axial load, shear/confinement rein-
Reinforced concrete frame structures generally de- forcement, and deformation path. The results are pre-
velop inelastic deformations when subjected to strong sented and discussed in this paper.
earthquakes. Earthquake-induced energy is dissipated
through the formation of plastic hinges, preferably in RESEARCH SIGNIFICANCE
the beams. However, it may not be possible to prevent The experimental research presented in this paper
hinging of the first-story columns, especially in the provides design information for earthquake-resistant
event of a strong ground motion. It is therefore essen- concrete columns. The results illustrate the effects of
tial to design and detail earthquake-resistant columns axial load, transverse reinforcement, and bidirectional
such that their design strengths are maintained during a loading on column ductility. Conditions that make col-
large number of inelastic deformation cycles. umn response unfavorable under simulated seismic
Structural response during recent earthquakes has loading are discussed. Improvements in column behav-
indicated that the majority of column failures was ior resulting from proper design and detailing of trans-
caused by high shear stresses, lack of concrete confine- verse reinforcement are illustrated.
ment, and bidirectional load effects. The failure of
short columns at the Macuto-Sheraton Hotel during the EXPERIMENTAL PROGRAM
1967 Caracas, Venezuela, earthquake is a good exam- Test specimens
ple of column failures caused by high shear stress re- Test specimens were representative of a first-story
versals. Many column failures were observed after the column between the foundation and the inflection
I985 Mexico City earthquake due to insufficient hoop point. Fig. 1 illustrates the specimen geometry. A sum-
and tie reinforcement. The columns of the Imperial mary of column properties is presented in Table 1.
County Services Building, damaged during the 1979
Imperial Valley earthquake, provide a good example of
Received May 4, 1987, and reviewed under Institute publication policies.
failures due to bidirectional loading effects. Further- Copyright (<;;) 1989, American Concrete Institute. All rights reserved, including
more, the Hachinohe City Library, damaged during the the making of copies unless permission is obtained from the copyright propri-
etors. Pertinent discussion will be published in the November-December 1989
1968 Tokachi-Oki earthquake in Japan, and Olive View ACI Structural Journal if received by July I, 1989.

ACI Structural Journal I January-February 1989 3

AC!member /\Jurat Saatcioglu is an associate professor of civil engineering at
Three groups of specimens, U, D, and B, were tested.
the University of Ollawa, Ontario, Canada. He received his PhD in structural Each group was labelled according to the deformation
engineering in 1981 from Northwestern University, Evanston, lllinois. Dr. path imposed. U specimens were loaded uniaxially in a
Saatcioglu was previously on staff at the Portland Cement Association, Sko-
kie, lll., and was on the faculty of the University of Toronto, Ontario, Can-
direction parallel to one of the principal axes of the
ada. He is a member of joint ACJ-ASCE Commillees 441, Reinforced Con- column section. D specimens were loaded uniaxially
crete Columns; and 442, Response of Concrete Buildings to Lateral Forces; and along the section diagonal. B specimens were loaded si-
is a member of ACJ Commillee 340, Design Aids for ACI Building Codes. His
research interests include experimental investigation of reinforced concrete
multaneously in two orthogonal directions, following a
members and seismic response of concrete structures. bidirectional deformation path.
ACJ member Guney Ozcebe is an assistant professor of civil engineering at the
Middle East Technical University, Ankara, Turkey. Dr. Ozcebe received his
Test setup and procedure
PhD in 1987 from the University of Toronto. Two test setups were used. A 2700 kN (600 kip) test
frame was used in the first setup to apply the axial load
that included variable axial tension and compression.
~350mm-=;-f_ Another test setup was used for specimens where the

'~[ 13]·.·.:<::··.
lateral deformation capacities were expected to be
hi~her than that permitted by the test frame. In this

. i..; •••
..
setup, the axial load was applied by pulling the load
transfer assembly down by means of two 500 kN (112

25mm bars
l ':. ·. ':
10mm bars
'

6.4mm bars
(double layer)
.
6.4mm bars
(double layer)
kip) hydraulic jacks. The lateral load was applied by
two 250 kN (56 kip) capacity actuators. Fig. 2 illus-
trates the test setups. An additional 500 kN (112 kip)
TYPE A TYPE 8 TYPE C
hydraulic jack was mounted on the load transfer as-
Loading direclions _ y Loading beam sembly in the orthogonal direction for bidirectional
loading. Type B specimens were tested using this addi-
Spacer block 1 -:--~n. -· T--=:IiOomm
Shear key plate/ f==H l tional hydraulic jack.
Each column footing was post-tensioned to the lab-
oratory strong floor to achieve fixity at the base. Lat-
900mm
eral loads were applied through the load transfer as-
sembly at the top. Shear keys and matching steel plates
were used to transfer the load. The load was applied to
the center region of the top column section and was
then distributed to the specimen through a steel plate
I
ISOOmm
and the shear keys.
The specimens were instrumented for deflection, ro-
~~~~j_ tation, and steel strain measurements. DC-operated
1 - - - - - 1 2 SOmm ------1 linear variable deflection transformers (L VDTs) were
used for the horizontal deflection measurements. Ad-
Fig. ]-Geometric details of test specimens ditional L VDTs were placed vertically on opposite faces

Table 1 - Properties of test specimens

--
Concrete Longitudinal steel Transverse steel Axial
Test strength, J.. J.. S, load,
specimen MPa MPa MPa Percent mm Configuration kN
Ul 43.6 430 470 0.85 150 Type A 0
U2 30.2 453 470 0.85 !50 Type A 600
U3 34.8 430 470 1.69 75 Type A 600
U4 32.0 438 470 2.54 50 Type A 600
U5 49.3 430 470 0.85 !50 Type A Variable
U6 37.3 437 425 1.95 65 Type B 600
U7 39.0 437 425 1.95 65 Type C 600
01 40.3 453 470 0.85 !50 Type A 0
02 30.2 453 470 0.85 150 Type A 600
03 34.8 430 470 1.69 75 Type A 600
04 43.6 430 470 2.54 50 Type A 600
05 49.3 430 470 0.85 !50 Type A Variable
Bl 32.0 438 470 2.54 50 Type A 600
B2 39.5 437 470 1.69 75 Type A 600
I MPa ; 145 pst; I mm ; 0.0394 m.; I kN ; 0.225 ktp.
Note: The longitudinal reinforcement ratio in all columns is 3.27 percent. For confinement configurations and
other geometric details, see Fig. I.

4 ACI Structural Journal I January-February

 Hydraulic jack
in lateral loading set-up

------N
Loading direction

Hydraulic jack
in vertical loading set-up

Fig. 2(a)-Test setup 1

of the columns to measure rotations. Steel strains were

Hollow square section
measured using electrical resistance strain gages. An
automatic data acquisition system and a microcompu-
ter were used to record the data. Fig. 2(b)-Test setup 2
The specimens were subjected to displacement-con-
trolled lateral load reversals. Fig. 3 (a) illustrates the
6~N--------------------------------------,
unidirectional displacement history followed in testing
5
Type U and D specimens. Each deformation level of 29 31 33
4
the load history was expressed as an increment of yield
displacement. The yield displacement was defined as ,., 3 .
21 23 25

~ 13 15 17 27
the level at which column section yielded as a whole. It <l 2
was obtained experimentally during the test as the be- z
Q o!L-l.,...P.+'>r-/--J,-i.--H4-1~+++4-f-lf+-H+f+-,f-l-j~+++4-rl
ginning of excessive increase in deflection under a rela- 1-
0
tively constant load. This value was observed to be w
...J
somewhat higher than the level at which the first yield- LL.
w
2 141618 28
ing of extreme tension steel was recorded. It corre- 0 3
22 24 26
sponded approximately to a displacement at which the 4 30 32 34
strain at the centroid of the five tension bars was equal 5 LOAD STAGE
to the yield strain. 6L-------------------------------------_...J
s
Type B specimens were subjected to the bidirectional
deformation path shown in Fig. 3 (b). This loading his-
Fig. 3(a)-Unidirectionalload history
tory was intended to simulate a major seismic action in
one direction while a minor action was occurring si-
multaneously in the orthogonal direction. It consisted WEST
of an elliptical deformation path, where maximum de-
3
flection in one direction was always twice the maxi-
mum deflection in the orthogonal direction. Each de- 2
flection level was expessed as an increment of the yield
deflection, as in the case of the unidirectional path.
Further discussion of deformation paths is provided in O SOUTH NORTH
the "Effect of Deformation Path" section.

Observed behavior and test results

All the specimens developed their full flexural yield 2
strengths prior to shear failure. Little damage was ob- 3
served upon the application of load cycles at yield.
Horizontal flexural cracks on faces perpendicular to the
6 54 3 21012 3 4 5 6
direction of loading and diagonal tension cracks on the
DEFLECTION, t::.lt::.y
side faces were observed. While this crack pattern was
typical of Series U columns, Series D and B columns
developed flexural and shear cracks on all faces. Very Fig. 3(b)-Bidirectionalload history
ACI Structural Journal I January-February 1989 5
 400
Fore<> (kN)

-100

100
D<>f' 1"'et 1 on (mm)

-200

(c)
-400

400
ForCQ(kN)

-100 -50

100
D<>f1<>etion <mm)

-200

(d)

-400
Fig. 4-Progressive damage in Specimen U3: (a) at .::1y,· Fig. 5(c) and (d)-Latera/load-top deflection hysteretic
(b) at 2.:::1y; (c) at 3.:::1); and (d) at the end of test relationships in the direction of loading: (c) Specimen
U2; (d) Specimen D2

400 few flexural cracks were formed beyond the initial

Fore"' <kNl cracking at yield displacement. Unlike the flexural
cracks, new diagonal cracks appeared every time the
previous magnitude of deformation was exceeded. Fig.
4 illustrates a typical crack pattern and progressive
damage observed in a test specimen.
-10~00- --i~~J1~~2~100 Concrete crushing was first observed in the compres-
D<>f'l<>etion <mm) sion zone near the column base. This was followed by
crushing of diagonal compression struts near the base,
causing concrete spalling within the plastic hinge re-
gion. Disintegration of core concrete led to rapid and
(a) excessive strength degradation. Yielding of tie steel was
-400 observed during the inelastic cycles, mostly toward the
400 end of testing. The extent of stiffness and strength deg-
radation varied with the test parameters considered.
Fore a <kNl
Fig. 5 through 7 illustrate the force deformation hys-
teretic relationships for all the columns tested. Table 2
provides a summary of observed and computed
-100
strengths.
Columns were instrumented to measure components
100 of total horizontal displacement. Fig. 8 illustrates the
O<>f' 1<>et ion (mm)
significance of shear, flexure, and reinforcement ex-
-200
tension within the anchorage zone in a typical test
specimen. The results indicate that, for the column ge-
ometry and the shear span considered, all three defor-
(b)
-400
mation components are significant within the hinging
region. Deformations due to shear and reinforcement
Fig. 5(a) and (b)-Latera/load-top deflection hyster- extension increase with increasing inelasticity. The shear
etic relationships in the direction of loading: (a) Speci- component of top displacement appears to be less sig-
men Ul; (b) Specimen DJ nificant than the other components, mainly because the
6 ACI Structural Journal I January-February 1989
 400 400
ForcliO'(N)
TENSION Force <kN)

-100

-100

(e) COMPRESSION
-400

400
Forca(kN)
(c) -400
TENSION
400
Forc<>(kNl

-100 -50

-100

(f) COMPRESSION

(d)
Fig. 5(e) and (f)-Latera/load-top deflection hyste:etic -400
relationships in the direction of loadmg: (e) Spec1men
U5; (f) Specimen D5 Fig. 6(c) and (d)-Latera/load-top deflection hysteretic
relationships in the direction of loading: (c) Specimen
U4; (d) Specimen D4
400
400
Force<kNl

-100

(a)
-400
400 (e) -400
Force <kN)
400

-100
-100

(b)
-400
(f)
-400

Fig. 6(a) and (b)-Latera/load-top deflection hyster- Fig. 6(e) and (f)-Latera/load-top deflection hyster-
etic relationships in the direction of loading: (a) Speci- etic relationships in the direction of loading: (e) Speci-
men U3; (b) Specimen D3 men U6; (f) Specimen U7
ACI Structural Journal I January-February 1989 7
 400 Table 2 - Summary of specimen strengths
ForcQ(kN)
Applied ACI 318-83
Test vert vmoxt Mm.x• v:. v" v•• M.,
-·
specimen kN kN kN.m kN kN kN kN.m
Ul 125 275 275 Ill !59 270 233
-100 U2 165 270 270 102 !59 261 287
U3 166 268 268 99 317 416 279
U4 170 326 326 96 476 572 271
U5 Compression !54 255 255 122 !59 281 298
Tension !52 250 250 0 !59 !59 149
U6 175 343 343 101 319 420 286
(a) U7 170 342 342 103 319 422 289
-400 01 180 285 285 107 159 266 231
400
02 180 270 270 102 159 261 247
ForcQ(kN) 03 140 275 275 416
99 317 242
04 210 300 300 119 476 595 272
-I-·
05 Compression 200 254 254 122 !59 281 274
Tension !50 205 205 0 !59 159 178
-100 Bl N-S 175 316 316 96 476 572 283
E-W 228 228 96 476 572 283
50 100
o.. f l .. ction (mm) B2 N-S 180 307 307 103 317 420 314
E-W 202 202 103 317 420 314
-200 I kN = 0.225 kip; I m = 3.3 ft.
Notes: Shear capacity calculations for D specimens are assumed to be the
same as those for U specimens.
(b) Capacity reduction factor q, = 1.0 is used in all calculations.
Actual material strengths given in Table I are used in capacity calculations.
-400

Fig. 7(a) and (b)-Latera/load-top deflection hyster-

etic relationships: (a) Specimen Bl in N-S direction; (b)
Specimen Bl in E- W direction

400 25 Specimen US

e
e 20
Flexure

~/
c:
-~ 15
-100 ....
u //f
o.. f!ect}On
100
Cmm)
-
~
cu
......
10 ,•"'/
,/Extension
of Reinf.

_,..,--
, .. / ,.., t
...J.·'
5
(c)
400
Fore" CkN)
-400

·.--
17
,• y-'

1
---·

2
--·--·

4
Shear

5 6
200

Fig. 8-Deflection components of Specimen U6 at 250

-100 -so mm (hinge length)
50 100

other displacement components result from rotations

and are magnified with the column height. This is
shown in Fig. 9. Hysteretic force-deformation relation-
(d)
ships for each deformation component of Specimen U6
-400
are shown in Fig. 10 (a) through 10 (c). These relation-
Fig. 7(c) and (d)-Latera/load-top deflection hysteretic ships provide a representative case for hysteretic re-
relationships: (c) Specimen B2 in N-S direction; (d) sponses of deformation components in a concrete col-
Specimen B2 in E- W direction umn.
8 ACI Structural Journal I January-February 1989
 8001 800

\
1
~i ~! ~ \
700 ~ .:Ej 4J
Ulltl): ~ 700
\
jl "'
+I +
I.,.: E \
~ 600 ~I"
/dJ:~
E 600 i
j~I'E
.s:~
,:!~ _....,
\
\
jj
_...., 50C
.c
0'1 I $1: ~ 500 f i ~:I

i
£:"1
Ill
:r ~oc It' \
f \
c
E Ii 400
\
:::J
~
0
300 0
u
300 ~ \
u \
200 200
\

100
* 1'otal
Measured
!00
(a) (b)
0 0
-4 -2 0 2 4 -80 -60 -40 -20 0 20 40 60 80
Deflection( mm ) Deflection ( mm )

Fig. 9-Deflection components along the height of Specimen U6 computed from

experimental data-(a) at Lly; (b) at 4 .1)

400 400
FORCE (kN) FORCE (kN)
300

02 .04 .06

11
ROTATION(rad)

'/
(a)
-400
fl I
(b)
-200

-300

-400
'

Fig. JO(a) and (b)-Hysteretic relationships for deformation components of Speci-

men U6-(a) Lateral force-hinge rotation relationship; (b) Lateral force-base rota-
tion relationship due to extension of anchored reinforcement within the support

ANALYSIS OF TEST RESULTS (135 kips) of constant axial compression. This level of
Effect of axial load axial load corresponded approximately to 20 percent of
Four columns with identical design strengths and re- concentric design axial load capacity. The strength loss
inforcement arrangements were tested to investigate the resulting from the ·three cycles at 211, was approxi-
effect of constant axial compression on column re- mately 50 percent. The specimens deteriorated very
sponse. Among these specimens, U 1 and D I were tested rapidly and the test was stopped as the load resistance
under zero axial load. These columns initially showed approached zero prior to the completion of cycles at
little strength and stiffness degradation. The strength 311,..
loss at 211v• when the displacement was approximately The hysteretic responses of the four specimens can be
equal to 3 percent of the column height, was about 18 compared by examining Fig. 5(a) through 5(d). The re-
percent. The specimens showed progressively increas- sults indicate that the effect of axial compression on
ing strength and stiffness degradation. The test was column response is to accelerate strength and stiffness
stopped when the load resistance dropped below 20 degradation under repeated inelastic load cycles.
percent of the recorded peak resistance. The remaining Two additional specimens, U5 and D5, were tested to
two columns, U2 and D2, were tested under 600 kN study the effect of variable axial load on column re-
ACI Structural Journal I January-February 1989 g
 the same reinforcement, and hence are difficult to sep-
FORCE (kN) arate.
A series of column tests was conducted to investigate
the effect of transverse reinforcement on column re-
sponse. Fig. 1 illustrates the reinforcement configura-
tions considered in the test program. Columns U3 and
U6 were reinforced using Type A and Type B configu-
- - - Z.Gauge
rations, respectively. The shear capacity provided by
- - LVDT
transverse reinforcement was identical in these two col-
umns. Column U3 was designed to conform to Appen-
dix A requirements of the ACI Building Code (ACI
318-83) 1 for ductile structures in seismically active re-
gions.
The results, shown in Fig. 6(a) and 6(e), indicate that
Column U6 with the crossties shows a superior re-
-200 sponse with stable hysteresis loops, while Column U3
shows a poor behavior with rapid strength and stiffness
degradation. The superiority of the Type B configura-
-300
tion, used in U6, can be explained by the effectiveness
of longitudinal column reinforcement in confining the
-400 core concrete when supported by the crossties. This was
first observed by Sheikh and Uzumeri 2 in column tests
Fig. JO(c)-Lateral force-shear displacement relation- under monotonically increasing axial compression.
ship Column U7 was tested to investigate the effective-
ness of crossties with a 90-deg hook at one end. There-
sponse. These specimens were prepared to have prop- fore, a Type C configuration was used in Column U7.
erties identical to Specimens U 1, U2, D 1, and D2. The The results indicate that, for the level of axial load and
applied load varied linearly between 500-kN (112-kips) column geometry considered, the Type C configuration
tension when the column was displaced horizontally 6 was as effective as Type B in confining the core con-
percent of the column height in one direction and 500- crete, up to a lateral displacement equal to 9 percent of
kN (112-kips) compression when the specimen was dis- column height.
placed the same amount in the opposite direction. This The significance of tie spacing on column response
loading was selected to simulate an axial load couple was also investigated. Specimens U2 through U4 and
resulting from a lateral loading. D2 through D4 were reinforced with a Type A trans-
Hysteretic responses of Specimens U5 and D5 are verse reinforcement configuration. The tie spacing in
shown in Fig. 5(e) and (f). The results indicate that the U2 and D2 was equal to 150 mm (6 in). The lateral re-
yield moment is affected by the level of concurrent ax- inforcement in these two columns was designed to con-
ial load, as expected. The specimens showed early form to the ACI Building Code (ACI 318-83) 1 require-
yielding when accompanied by axial tension. However, ments for an ordinary structural member that does not
axial tension produced a significant delay in strength form part of a ductile moment-resisting frame in a seis-
degradation beyond the initial yield. mically active region. The columns failed by disintegra-
On the compression side, both specimens showed tion of core concrete resulting from lack of concrete
higher yield strengths, resulting from the accompanied confinement and yielding of transverse reinforcement.
axial compression. However, early strength degrada- The level of tie spacing used was not adequate to pro-
tion was evident under cyclic loading. The behavior vide sufficient concrete confinement to prevent early
under increasing axial compression was similar to that and rapid strength degradation.
of the specimens tested under constant axial compres- The tie spacing was reduced to 75 mm (3 in.) in
sion. Specimens U3 and D3 to conform to the Appendix A
Both U5 and D5 experienced similar magnitudes of requirements of the ACI Building Code. There was very
stiffness degradation during reloading, irrespective of little improvement in response; both columns showed
the sign of the accompanying axial load. An increased early strength degradation.
number of deformation cycles at higher increments of The tie spacing was reduced further to 50 mm (2 in.)
yield displacement produced progressively increasing in Columns U4 and D4. The results showed a signifi-
strength loss. cant improvement in column response. This improve-
ment was attributed, to a great extent, to increased
Effect of shear/confinement reinforcement confinement of the core concrete resulting from close
Transverse reinforcement in columns has two pri- spacing of the ties. Hysteretic responses of the columns
mary functions that affect column strength. One is to considered in this series are presented in Fig. 5 and 6.
provide shear resistance and the other is to confine the The improvement in column response, produced by
core concrete. Often these two functions are served by a reduction in tie spacing from 75 to 50 mm (3 to 2 in.)
10 ACI Structural Journal I January-February 1989
was close to, but not as much as, the improvement re-
sulting from the use of Type B and Type C configura-
tions. This suggests that a proper choice of transverse
steel arrangement may be more feasible than reducing
the tie spacing to achieve the same level of confine-
ment. Further details on column confinement, includ-
ing analytical predictions, can be found in Reference 3.
Columns considered in this investigation were de-
signed to prevent premature shear failure prior to flex-
ural yielding. Consequently, the shear capacity under
monotonic loading was either approximately equal to
or greater than the shear force corresponding to flex-
ural failure. In columns with closely spaced ties, the
·400 L._.....___L~___J-~--'---~--'--~_J__~--L.~----,L-~_J
improvement in column response is attributed in major -!00 -75 -50 -25 25 50 75 100
part to concrete confinement. The increase in shear ca-
pacity associated with increased transverse steel, over Fig. 11-Comparison of U4 and D4 responses in N-S
and above the shear capacity corresponding to the peak direction
shear force, is believed to have little effect on improved direction. Series B specimens were designed and con-
response. Comparison of the hysteretic responses of U4
structed to have the same capacity in the N-S direction
and U6, given in Fig. 6 (c) and 6 (e), supports this view. as in the E-W direction. Their responses in the two or-
Specimen U6, with 50 percent less shear reinforcement, thogonal directions are shown in Fig. 7. Deformations
showed a more favorable response than Specimen U4 prior to yielding in one direction do not affect noticea-
mainly because of the difference in transverse rein- bly the response in the other direction. However, in-
forcement arrangement and the associated characteris- elastic cycles in one direction drastically reduce column
tics of confined concrete. strength in the other direction. This reduction was 20 to
Effect of deformation path 30 percent in Column B 1 and 30 to 50 percent in Col-
Three different deformation paths were followed umn B2. Beyond the initial yielding, if the columns had
during the test program. Series U columns were sub- experienced inelasticity in both directions, a sudden
jected to unidirectional deformation cycles, parallel to strength decay was inevitable.
a principal axis. Series D columns were loaded parallel The main difference between Columns B1 and B2
to a section diagonal, producing equal components of was the tie spacing. Bl, with closely spaced ties, had
bending and shear in two principal directions. Series B better concrete confinement and showed a more favor-
columns were subjected to an elliptical bidirectional able response than B2. The improvement in response
deformation path. Fig. 3 illustrates the deformation due to concrete confinement was less pronounced in
paths employed. Series B than in Series U and D.
Columns Ul through U5 were companions to D1
through D5. Their responses can be compared in Fig. 5 CONCLUSIONS
and 6. The results indicate that the overall hysteretic The following conclusions can be made based on the
characteristics of U and D companion specimens are experimental investigation reported in this paper:
similar in terms of strength, stiffness, and ductility. 1. The effect of constant axial compression on hys-
When column responses in a direction parallel to one teretic response of reinforced concrete columns is to re-
of the principal axes are compared, the Series D col- duce ductility and accelerate strength and stiffness deg-
umns show a significantly reduced capacity. This is due radation. Columns subjected to variable axial tension
to the biaxial interaction effects and is expected. Re- and compression exhibit a different response than those
sponses of U4 and D4 in the N-S direction are com- subjected to constant axial load. Axial tension pro-
pared in Fig. II to illustrate this point. duces a reduction in flexural yield strength. However,
The deformation path imposed on Series B columns strength degradation under cyclic loading is delayed.
simulated a seismic response in which the structure was Axial compression produces an increase in flexural
subjected to major deformation cycles in one direction yield strength, followed by a more rapid strength deg-
while simultaneously being subjected to minor cycles in radation.
the orthogonal direction. Column Bl was a companion 2. Ductility of columns, under combined axial
to U4 and D4, whereas Column B2 was the companion compression and bending moment reversals, is im-
to U3 and D3. proved very significantly with the use of a proper con-
Hysteretic responses of Columns B 1 and B2 are finement configuration. A column confined by trans-
shown in Fig. 7. The results indicate that the columns verse and longitudinal reinforcement, where every lon-
show similar response initially. However, Series B col- gitudinal bar is supported by either a hoop or a
umns exhibit progressively increasing strength and crosstie, exhibits a response better than that of a col-
stiffness degradation due to bidirectional effects. The umn where every other longitudinal bar is supported.
amount of degradation in one direction is directly re- ACI 318-83 requirements for confinement steel are not
lated to the degree of inelasticity in the perpendicular adequate for columns with unsupported longitudinal
ACI Structural Journal I January-February 1989 11
 bars. Selection of a proper confinement configuration Graduate and undergraduate students Mr. B. C. K. Lee, Mr. R.
is a more feasible approach than a reduction in hoop Scott, and Mr. S. L. Yip prepared and tested seven of the fourteen
specimens. The experiments were conducted at the Structures Labo-
spacing alone to achieve the same level of ductility. ratory of the University of Toronto.
3. Within the deformation and axial load ranges
considered in this investigation, crossties with a 90-deg
hook at one end, extending 10 bar diameters, perform
as satisfactorily as those with 135-deg hooks at both NOTATION
ends and extending 10 bar diameters. J: concrete cylinder strength
4. Biaxial bending reduces column capacity in each f. yield stress
direction. If the lateral load producing biaxial bending M"'"' = maximum applied moment
M, nominal moment capacity
follows a straight line path, the overall characteristics v nominal shear force resisted by concrete
of hysteretic response in terms of strength and stiffness V,, shear force at which the first diagonal crack was observed
degradation is not significantly affected by biaxial V, nominal shear force resisted by transverse reinforcement
bending. V," maximum applied shear force
V, nominal shear capacity of a member
5. Columns subjected to simultaneously varying bi- .d, yield displacement
directional load reversals show a different response () rotation
than those subjected to unidirectional load reversals. <P capacity reduction factor
The level of damage in one direction adversely effects s spacing of tie steel
the column in the other direction. If the deformation in
one direction is less than the yield deformation, bidi-
rectional effects on response in the orthogonal direc- REFERENCES
tion are small. However, if post-yield deformations are I. "Building Code Requirements for Reinforced Concrete (ACI
experienced in both directions, severe strength and 318-83)," American Concrete Institute, Detroit, 1983, Ill pp.
stiffness degradation are observed. 2. Sheikh, Shamin A., and Uzumeri, S.M., "Strength and Ductil-
ity of Tied Columns," Proceedings, ASCE, V. 106, ST5, May 1980,
pp. 1079-1102.
ACKNOWLEDGMENTS 3. Ozcebe, Guney, and Saatcioglu, Murat, "Confinement of Con-
This investigation was sponsored by the Natural Sciences and En- crete Columns for Seismic Loading," ACI Structural Journal, V. 84,
gineering Research Council of Canada under Grant No. U0040. No.4, July-Aug. 1987, pp. 308-315.

12 ACI Structural Journal I January-February 1989