Enhancement of Seismic Performance of Reinforced Concrete Columns with BucklingRestrained Reinforcement

P. Lukkunaprasita,*, T. Tangbunchooa, K. Rodsinb

Department of Civil Engineering, Chulalongkorn University Bangkok, Thailand

Department of Civil and Environmental Engineering Technology, King Mongkuts Univesity

of Technology North Bangkok Bangkok, Thailand

Abstract
Reinforced concrete (RC) columns with light confinement prevalent in developing countries
exhibit low ductility with brittle shear failure, especially when buckling of longitudinal rebars
takes place. This study applies the buckling restraining concept widely used in seismic
resistant steel structures to reinforcing bars. Two RC columns 270mm x 300mm in cross
section with a height of 1200 mm and minimum (non-seismic) transverse reinforcement were
tested under cyclic lateral loading. Buckling-restrained reinforcement was provided over the
critical zone. The buckling-restraining casing effectively prevented buckling of slender
vertical bars under a substantially high axial load level, resulting in a more ductile mode of
failure with evident formation of plastic hinge at the base of the column. Prior to gravity load
collapse, the drift capacities and the degraded concrete shear capacities of the specimens were
significantly increased compared to their counterparts without casings. Thus, it is important
to prevent longitudinal bar buckling in investigation of degraded concrete shear capacity of
RC columns.
Keywords: Seismic performance; Gravity load collapse; Reinforced column tests; Light
transverse reinforcement; Longitudinal bar buckling; Buckling-restrained reinforcement;
Degraded concrete shear capacity

1. Introduction
Reinforced concrete columns with light longitudinal and transverse reinforcement are
prevalent in existing low rise buildings in regions of low or even moderate seismicity,
especially in developing countries. These structures are vulnerable to damage or even
collapse in the event of a strong earthquake. Unfortunately, research work on lightly
reinforced concrete columns is quite limited. RC columns with light transverse steel
subjected to cyclic lateral load exhibit rapid loss of lateral load resistance soon after attaining
the peak capacity. Shear mode of failure often prevails with small drift capacity [1],[2].
Under moderate to high axial load ratios, longitudinal bars tend to buckle, with the
consequence of abrupt shear failure as reported by Wibowo et al. [3]. Sezen and Moehle [4]
earlier reported that for columns with light axial load, shear failure would be triggered due to
apparent strength degradation after development of the flexural strength whereas columns
with high axial load would suffer abrupt shear compression failure.

*Corresponding author. Tel.: +66 2 2186571; fax: +662 2186571

E-mail address: [email protected] (P. Lukkunaprasit).

It was speculated that preventing longitudinal bar buckling would greatly enhance the
seismic performance of RC columns since it eliminates the transfer of gravity load from the
steel to concrete which would otherwise be caused by bar buckling, thereby reducing the
shear demand on concrete on the diagonal crack plane. The buckling restraining concept
successfully applied in seismic resistant steel structures was adopted to provide buckling
restrained reinforcement (BRR). RC column specimens with BRR were cyclically loaded in
the horizontal direction under constant axial load, and their performances compared with
their counterparts without BRR.

2. Performance of control columns without buckling-restraining casing

The specimens S2 and S3 tested by Wibowo et al. [3] serve as the control specimens. The
columns, 270mm300mm in cross section, were reinforced with four 16 mm Grade 400
MPa longitudinal steel bars. Hoop ties, 6 mm in diameter, were provided at 300mm spacing
corresponding to a transverse reinforcement ratio H of 0.0007. The nominal concrete
compressive strength fc' was 20 MPa. The column was loaded in single curvature under cyclic
loading with the lateral load applied at a height of 1200 mm from the base. A constant axial
load of 20% the axial load capacity based on fc'Ag was applied to specimen S2, while that for
S3 was 40%.

The specimens exhibited flexure dominated inelastic behavior with well distributed
flexural cracks up to the peak strength at about 1.5% and 1.0% drifts for specimens S2 (20%
axial load ratio) and S3 (40% axial load ratio), respectively.1.0 %. Soon after the peak load,
previously developed vertical cracks widened, indicating impending vertical bar buckling.
The drift capacity was (1.75 strain data)2.0% with sustainable lateral load capacity of
(98??)about97 % of the peak value for S2. The corresponding values for S3 are 1.25% drift
and 97 % of the peak capacity, respectively. Note that these drifts are very close to those at
peak loads. At impending failure upon visible longitudinal bar buckling (2.5% and 1.5%
drifts for S2 and S3, respectively), vertical bars buckled in the potential buckling zone near
the base followed by an abrupt transfer of the force carried by steel to the concrete core with
a consequence of significant increase in shear along the cracked shear plane. This triggered
an abrupt shear failure (due to deterioration of cyclic shear resistance) and loss of gravity
load capacity, which resulted in a sharp drop in the descending branch of the envelope curve.
Furthemore, the failure shear plane cut through the concrete core at roughly 45 in between
the ties. Thus, practically no shear resistance was provided by the transverse reinforcement at
the failed section.
( Please read the first 7 line of this paragraph again)

3. Buckling-restrained reinforcement
The concept of buckling restrained brace successfully used in steel structures for enhancing
seismic performance of steel buildings was adopted. In our pioneer study, 28 mm steel
tubes with 4mm-thickness were used to encase the 16 mm longitudinal bars. The deformed
bars were coated with silicone sealant compound to a practically smooth surface and
subsequently wrapped with plumbers tape. Non-shrink grout was then injected into the void
between the wrapping and the steel casing. The stress-strain relationships of the steel bar with
and without casing under compressive loading are shown in Fig. 1.

Fig. 1: Compressive stress-strain relationships of 16 mm diameter bar with and without

casing.

The length of the bars between the grips was 300 mm. For BRR, the length of the casing
was 250mm. As expected, right after peak load, the capacity of the bare steel bar sharply
drops due to bar buckling. In contrast, the buckling-restrained reinforcing bar could sustain
almost constant load after yielding since the lateral restraint provided by the casing prevents
the bar from buckling. Interestingly, even at a large axial strain of 4%, BRR could sustain a
stress as high as 90% of the peak value.
4. Test specimens
The specimens, designated by SC1and SC2, were practically identical to S3 and S2, in
dimensions and reinforcement details, respectively, except for the concrete compressive
strengths and provision of BRR. Due to some prior mistake in quality control, the contractor

Table 1
Details of Column Specimen
Specime
n

Main

Reinforcmen
t

Ties
(@mm)

Axial
load
ratio

fc'

Remark

(MPa)

S2

1.0 %

4 16

0.07 %

6@300

0.2

21.0

S3

1.0 %

4 16

0.07 %

6@300

0.4

18.4

SC1

1.0 %

4 16

0.07 %

6@300

0.4

27.7

With
BRR

SC2

1.0 %

4 16

0.07 %

6@300

0.2

29.8

With
BRR

Notation : V is the longitudinal reinforcement ratio; axial load ratio is the ratio of the
applied axial load to axial load-carrying capacity based on fc' and gross concrete area.

Fig. 2. (a) Geometry and reinforcement details of column specimens; (b) Buckling-restrained
reinforcement (BRR).

somehow modified their mix design which resulted in about 40% increase in concrete
strength over the specified value of 20 MPa. As for the reinforcement, buckling restraining
casings were provided over the critical failure zone as previously experienced by specimens
S2 and S3, i.e. over the lowest tie spacing of 300mm. The yield strengths of the rebars were
530 MPa and 240 MPa for the longitudinal and transverse reinforcement, respectively.
Relevant details of the specimens are presented in Table 1 and Figure 2.

5. Test setup
Figure 3 shows the test setup. The specimens were tested in the testing frame developed by
Warnitchai and Rodsin [personal communication]. The axial load was applied by means of a
vertical hydraulic jack constrained to move horizontally on guided rollers, thereby ensuring
verticality of the axial load. The lateral load was applied using an actuator with 1000 kN
capacity. During testing, special care was taken to ensure that the axial load was maintained
constant to within 10% of the initial load. The drifts as well as the axial displacement of the
columns were measured using linear variable displacement transducers (LVDTs). Electrical
strain gauges were installed above and below the casings at a distance as close as practicable
to the casings to measure the strains in the longitudinal rebars.
The displacement controlled loading sequence consisted of drift-controlled mode with two
increments of 0.125% drift followed by increments of 0.25% up to 2% drift, after which the
drift increments were 0.5%. Two cycles of loading were repeated in each drift ratio to ensure

that stable response could be maintained. The test was performed until the vertical load
capacity was practically lost.
Fig. 3. Test setup.

6. Experimental results
6.1. Effectiveness of BRR
Figure 4 shows the hysteretic loops of specimens SC1 and SC2 together with S3 and S2 for
comparison. Specimens SC1 and SC2 with BRR experienced relatively less cracking than
their counterparts at the same drift ratio as illustrated in Fig. 5 for SC1 and S3 . As expected,
the buckling-restraining casing effectively prevented buckling of rebars in the zone reinforced
with BRR. Specimens SC1 and SC2 could sustain the gravity load with stable hysteretic
loops past the peak load with relatively more gradual drop in lateral load capacity as evident
in the envelope curves depicted in Fig. 6. While specimens S3 and S2 practically lost their
load carrying capacities at 1.5% and 2.5% drift ratio, respectively, their counterparts with
BRR could still sustain more than 90% of the peak capacities at these levels. Prior to gravity
load collapse, the drift capacity of specimen SC1 was 1.75%, a 40% increase compared to the
specimen without casing. The improvement was much greater for specimen SC2 (with half
the axial load ratio of 0.2) which sustained the gravity load up to a drift of 4.5% whereas the
specimen without BRR could do so up to 2% drift only.
An investigation of the strains in the longitudinal rebars reveals that, soon after the peak
loads, they reached compression yield strain at about 1.25% and 2.5% drift ratios for
specimens SC1 and SC2, respectively (see Fig. 7). Interestingly, the rebars in their
counterparts S3 and S2 buckled soon after reaching the compressive yield strain. It should be
noted that these results are indicative only because some of the strain gauges got damaged
soon after the peak load. Furthermore, the strains in specimens SC1 and SC2 were obtained
from interpolations of the readings from strain gauges above and below the casings, whereas
those in specimens S2 and S3 (reported by Wibowo et al.) were values from strain gauges
close to the lowest tie. Nevertheless, the results clearly indicate that the rebars in the
specimens with buckling-restrained casings, being highly stressed to the compressive yield
strain level, would have buckled if they were bare bars, with subsequent trigger of abrupt
failure not far from the drift levels when the compression yield was reached in a similar
manner as in specimens S2 and S3.

Fig. 4. Hysteretic curve of specimen (a) S3; (b) SC1; (c) S2; (d) SC2.

Fig. 5: Comparison of damage at same drift ratio of 1.5%(cycle 1) (a) specimen S3(note sign
of impending collapse) ; (b) specimen SC1.

Fig. 6. Normalized envelope curves of specimens

S2,S3(without BRR) and SC1,SC2 (with BRR).

Fig. 7. Compressive strains in the longitudinal rebars versus lateral drift.

Figures 8 and 9 show the modes of failure of all specimens for comparison. Clearly, the
specimens without BRR failed in an undesirable shear mode of failure, whereas the ones with
BRRs could develop plastic hinges at the bases of the columns. However, splitting as well as
shear distress in SC1 and SC2 also prevailed at impending collapse, but failure was more
gradual than the case without BRRs. It should be noted that the tie spacing/diameter ratio of
the vertical rebars was well over 18, and yet BRRs worked effectively and remained
practically straight without buckling even under a significantly high axial load ratio of 40%
(see Fig. 8 (d)).

6.2. Degraded shear capacity

The ductility of Seismic performance It is highly desirable to have Under earthquake

loading, structural components need to possess sufficient post-peak shear capacity in order to
perform satisfactorily. As depicted in the lateral load - deflection envelopes in Fig. 10, due to
bar buckling, the shear capacity of specimen S3 at 1.5% drift dropped to almost zero, and it
would naturally be reported as total loss of shear capacity at this drift level. In contrast, at this
drift, specimen SC1, the counterpart of S3 with (almost) the same column configurations
except for the provision of buckling-restrained casings and a somewhat difference in concrete
compressive strength, could still sustain a lateral load of more than 90% the peak shear lateral
load capacity. It should be noted that, due to the large tie spacing much in excess of half the
section depth, the contribution of transverse reinforcement is negligible. Therefore, the lateral
load capacity derives solely from contribution of concrete.
It is important to note that concrete shear degradation models in the literature do not take into
account the influence of longitudinal bar buckling, and hence they could much underestimate
the actual shear capacity of concrete near impending collapse. This is because previous
researchers did not have an effective way to separate the effect of bar buckling, nor were they
able to prevent buckling from taking place for a given reinforcement configuration in slender
longitudinal bars. Thus, BRR makes it possible evaluation of actual concrete shear capacity
with practically no interference from longitudinal bar buckling.

Fig. 8. (a) Specimen S3 at gravity load collapse at 1.5% drift; (b); (c) Specimen SC1 at 1.75%
drift with sustainable gravity load and 2% drift with loss of gravity load capacity, resp. Note
no buckling of BRR as evident in (d).

Fig. 9. (a) Specimen S2 at 2% drift; (b); (c) Specimen SC2 at 4.5% drift with sustainable
gravity load and 5% drift with loss of gravity load capacity, resp.

7. Conclusions

From the limited number of tests conducted, the buckling-restrained reinforcement developed has
demonstrated its potential in preventing buckling of slender vertical bars under a significantly high
axial load level. Shear failure, witnessed in specimens when bar buckling takes place, can be deferred
or even eliminated when BRR is provided in the critical zone to prevent bar buckling. This results in a
more ductile mode of failure with evident formation of plastic hinge at the base of the column. Prior
to gravity load collapse, the drift capacities and the degraded concrete shear capacities of the
specimens are significantly increased compared to their counterparts without casings. Thus, the
deteriorated concrete shear capacity reported in the literature without distinction of longitudinal bar
buckling can be significantly in error, especially for lightly reinforced columns. atdisplacement
ductility close to gravity load collapse. Furthermore, BRR would be a useful device in

experimental investigation of actual degraded concrete shear capacity without influence from
longitudinal bar buckling.
BRR makes it possible evaluation of degraded shear capacity contributed by concrete with
practically no interference from longitudinal bar buckling. This could be a significant factor
to consider in investigation of actual degraded shear capacity of concrete in the future.

Clearly more extensive experiments are needed to fully investigate the effectiveness of BRR
over a wide range of applications.
Buckling of longitudinal bars with large tie spacing (L/db >18 !) leads to an abrupt
transfer of axial load from the steel bars to the concrete, thereby triggering shear failure
due to deterioration of concrete shear strength during cyclic loading.
Past concrete shear strength degradation models without accounting for longitudinal bar
buckling can be in error, especially for lightly reinforced columns.
Obviously, the practicality and the effectiveness of BRR over a wide range of applications are
subject to further extensive investigations.

Acknowlegements

The authors are grateful for the funding from the Commission on Higher Education,
Ministry of Education and Chulalongkorn University.

References

[1] Lynn AC, Moehle JP, Mahin SA, Holmes WT. Seismic Evaluation of Existing
Reinforced Concrete Columns. Earthquake Spectra, Earthquake Engineering Research
Institute 1996; 12(4):715-739.
[2] Sezen H, Moehle JP. Shear strength model for lightly reinforced concrete columns.
Journal of Structural Engineering, ASCE 2004; 130(11): 1692-1703.
[3] Wibowo A, Wilson JL, Fardipour M, Lam NTK, Rodsin K, Lukkunaprasit P, Gad EF.
Seismic Performance Assessment of Lightly Reinforced Concrete Column, The 21st
Australasian Conference on the Mechanics of Structures and Materials 2010.
[4] Sezen H, Moehle JP. Seismic tests of concrete columns with light transverse
reinforcement. ACI Journal 2006; 103(6):842-849.
[5] Ruangrassamee A, Sawaroj A. Seismic Enhancement of Reinforced-Concrete Columns
by Rebar-Restraining Collars. Earthquake Engineering and Structural Dynamics 2010;
under review.
[5] Sawaroj A, Ductility enhancement of reinforced-concrete columns by rebar-restraining
collars, Master Thesis, Department of Civil Engineering, Faculty of Engineering,
Chulalongkorn University 2010.

[5] Priestley MJN, Verma R, Xiao Y. Seismic shear strength of reinforced concrete column.
Journal of Structural Engineering, ASCE 1994; 120(8): 2310-2329.
[6] FEMA-273. NEHRP guideline for the seismic rehabilitation of buildings. Washington
DC (USA): Federal Emergency Management Agency; 1997.