Malaysian Journal of Civil Engineering 22(2) : 264-290 (2010)

SEISMIC PERFORMANCE OF BEAM-COLUMN JOINTS IN REINFORCED

CONCRETE BUILDINGS SUBJECTED TO
REVERSIBLE VERTICAL CYCLIC LOADING

Nor Hayati Abdul Hamid1

1
Faculty of Civil Engineering, Univesiti Teknologi MARA, 40450, Shah Alam, Selangor,
Malaysia

Abstract: Three identical half-scale of exterior beam-column joints of RC buildings with

different arrangement of connections detailing are designed, constructed and tested under
reversible vertical cyclic loading. Each set consists of reinforced concrete beams, column and
foundation beam which represents the exterior joints of the ground floor of RC buildings. The
objectives of this study are to examine seismic performance, to observe visual damages and to
classify different types of damages which occur in beam-column joints. The first set is designed
beam-column joint in accordance to BS 8110 where no anchorages of bars are tied between beam
and column. Second set comprises of three anchorage bars from column are placed and tied with
top of reinforcement bars in the beam. Third set is designed with cross-bracing of reinforcement
bars at the intersection of beam-column joint. Results showed that the beam-column connections
which designed according to BS 8110 without anchorage bars suffer the most severe damages as
compared to connections with anchorage bars and cross-bracing bars. Beam-column joint with
cross-bracing bars experienced less damage as compared to two types of joints. Visual structural
damages of beam-column joints include cracks prorogation, the spalling of concrete shell,
buckling of longitudinal reinforcement bars and fractured of shear reinforcement bars. Therefore,
it is an appropriate time to revise and replace the current code of practice to seismic code so that
reinforced concrete buildings which are constructed in Malaysia are save under long-distant
earthquake excitation and local earthquake.

Keywords: beam-column joint, vertical cyclic loading, damages, spalling of concrete, visual
observation

1.0 Introduction

The seismic performance of reinforced concrete buildings in high, medium and

low seismic regions need to be examined carefully. Detailing of the point of
intersection in RC buildings where the load is transferred from one structural
component to another structural component such as beam-column joint, wall-
foundation joint and slab-beam joint must be carefully designed. Particular
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 265

attention should be focused on arrangement of reinforcement bars at joints,

spacing and arrangement of the links. Compressive strength of concrete,
detailing of beam-column joints and workmanship play an important role in
assessing the seismic performance under seismic loading (Paulay et al., 1978).
Beam-column joint is defined as the zone of intersection between beams and
columns with the functional requirement which enable the adjoining members to
develop and sustain their ultimate capacity. The joint should provide sufficient
strength and endurance to resist the internal forces transferred by the framing
members.
Basically, there are three types of beam-column joints which are interior
joint, exterior joint and corner joint (Uma and Meher, 2002). This study focuses
on the design, construction and testing exterior beam-column joint at ground
level under quasi-static vertical cyclic loading. At ground level of beam-column
joint, all the loads which include dead, imposed, wind and earthquake loadings
are meet and transfer to the foundation beam. Therefore, the exterior beam-
column joints at ground floor are considered critical joints and require
investigating their performance under loading combination. Consequently, three
identical half-scale beam-column joints with three types of joint arrangements
are constructed, calibrated and tested in heavy structural laboratory. These types
of connections together with their individual foundation beam are clamped to
strong floor before testing under vertical cyclic loading. Mode of failures,
spalling of concrete, buckling of longitudinal reinforcement bars and fractured of
shear reinforcement bars are among the damages that can be observed during
experimental work. Identification and classification of damages are based on the
visual damages, percentage of drift, cracking pattern, buckling of reinforcement
bars and spalling of concrete cover. Further nonlinear dynamic analysis can be
modeled based on the load versus displacement (hysteresis loop), stress-strain
relationship and crack propagation.

2.0 Structural Damages Following Past Earthquakes

A beam-column joint is the crucial zone in reinforced concrete buildings where

vertical loading and lateral loading are met and transfer their load to the
foundation. This type of joint has high risk of failures as compared to others
structural components when an earthquake strikes at any areas in seismic regions
where this is the possibility of occurrence of plastic-hinge zone mechanism.
Failure of beam-column joints in RC buildings was identified as one of the
leading causes of collapse during earthquake which classified them as soft-story
mechanism (strong beam and column weak). Figure 1 shows the joint failure due
to poor detailing and workmanship which lead to the collapse of the structure
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 266

during the 2003 Boumerdes Earthquake. The damages are due to no diagonal
reinforcement bars in the joint and spacing between links are wider which
contribute to the failure of structures under earthquake attack (Alcocer and
Carranza, 2002).

Damage of beam-column
joint

Figure 1: Seismic deficiency weak beam-column joint during the 2003 Boumerdes Earthquake.

Figure 2 shows two examples of damages occurred in beam-column joints

after earthquake disasters. Figure 2(a) illustrates the cracks and spalling of
nominal concrete cover at beam-column joints of at ground floor of five-storey
reinforced concrete building during the 2001 Bhuj Earthquake, India. This is due
to disorganization of beam-column joint and insufficient of longitudinal and
shears reinforcement in the concrete. Figure 2(b) shows the severe damage of
exterior beam-column joints during the 2001 Bhuj Earthquake due to instability
of column where the spalling of concrete and buckling of reinforcement bars
(Telford, 1996). Poor workmanship and lacking of detailing are among the
factor that causes this type of damage on reinforced concrete structures when
earthquake strike.
Figure 3 shows the spalling of nominal concrete cover in beam-column
joints following an earthquake. The damage is due to insufficient stirrups which
produce unconfined concrete and buckling of reinforcement at the intersection of
beam-column joints. However, this damage can be repaired and strengthened by
jacketing method around the column and joint.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 267

Exterior beam-
column joint damage

dislodge of beam-column joint

at ground beam

Figure 2: Beam-column joint damages during the 2001 Bhuj Earthquake India ; (a)
Disorganization of a beam to column joint is inadequate; and (b) Overloading of exterior beam-
column joints.and lacking of stirrups.

Figure 3: Spalling of concrete cover at beam-column joints in reinforced concrete building.

 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 268

Based the pictures‟ illustrations and problem arise in beam-column joint, the
main objective of this paper is to improve the seismic performance of beam-
column joint by using different types of arrangement of reinforcement bars at
beam-column joints. Three sets of specimens were designed, constructed and
tested under quasi-static vertical cyclic loading until collapse. Anchorage
reinforcement bars are arranged in different orientation in order to improve the
structural performance and strengthening the connection by avoiding soft-storey
mechanism which expected to occur in the ground column. Three set sub-
assemblage of half-scale beam-column joint together with foundation beam were
designed in accordance to BS8110, constructed and tested in Heavy Structural
Laboratory, Faculty of Civil Engineering, Universiti Teknologi MARA, Shah
Alam, Selangor. Figure 4 shows the prototype of the exterior beam-column joint
of reinforced concrete building of Block 1, Faculty of Civil Engineering, UiTM,
Shah Alam, Selangor. However due to the space limitation, only one beam
connected to column is constructed in the laboratory.

Prototype exterior beam-

column joint

Figure 4: A prototype of exterior beam-column joint of ground floor of Block 1, Faculty of Civil
Engineering, Universiti Teknologi MARA, Shah Alam, Selangor.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 269

3.0 Monolithic Beam Analogy

In monolithic beam analogy, the member compatibility condition is used where

the force in tension is equivalent to force in the compression by considering the
stress-strain relationship in concrete and reinforcement bars. The stress-strain
relationship is adopted for the concrete by assuming the simplified stress block
by solving two unknowns which are the position of neutral axis and the strain in
the concrete. The relationship between these global parameters is defined by
introducing an analogy in term of global displacement by equating the
displacement in precast connection with monolithic connection.

Point of
Lcant contraflexure

Precast imp
ductile imp
L
imp cant
connection
el

Plastic Rotation
Lp p pLp
tot(precast)= tot(monolithic)

Monolithic
Connection
p
L
p cant

el
Plastic Hinge Region

Figure 5: Monolithic beam analogy

Figure 5 shows the monolithic beam analogy for the beam-column

connection. By assuming the point of contraflexure occurring in the beam due to
lateral loads at mid-span and the end displacement of two different cantilever
schemes are compared with each other. In precast ductile connection, the
opening gap of the beam-column interface will result in an imposed rigid
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 270

rotation ( imp ) . The contribution to the total beam-edge displacement is due to a

rigid rotation which can be derived as follows:

imp imp Lcant (1)

Where Lcant is the distance between the interface and point of contraflexure. By
adding the contribution due to elastic deformation, the total displacement of the
precast beam is given by the following equation:

total ( precast ) imp elastic

(2)

In the monolithic cantilever beam, the total displacement is the sum of elastic
and plastic contribution where the latter approximated is given by the rigid
rotation about the end of the beam (or the plastic hinge centroid).

total ( monolithic ) plastic elastic

(3)

By assuming the two beams are identical in terms of geometry and

reinforcement, the elastic deformations would be the same and when imposing
the same total displacement, the „plastic‟ contributions can be equated. In ductile
precast connections, the inelastic deformation is localized at the interface while
in monolithic beam is distributed along the plastic hinge. Hence, the equation
becomes:

total ( precast ) total ( monolithic )

(4)

imp elastic plastic elastic

(5)

By utilizing the monolithic case for ultimate and yielding curvature concepts
introduced by Paulay and Priestley (1975):

Lp Lp
p ( Lcant ) ( y ) L p ( Lcant )
plastic u
(6)
2 2
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 271

Where L p is the plastic hinge length in the monolithic beam. By combining

Equation 5 and Equation 6, the above equation becomes:
Lp
p Lcant ( u y ) L p ( Lcant ) (7)
2
( Lcant )
imp
( u y
) (8)
Lp
( Lcant )Lp
2
c
( imp Lcant )
u y (9)
c Lp
( Lcant )Lp
2
( imp Lcant )
c
( y
).c (10)
Lp
( Lcant )Lp
2
By introducing additional condition on the global member displacement satisfies
the member compatibility and results in a simple relationship between concrete
strain and neutral axis position. For each guessed value of concrete strain, the
member compatibility relationship provides a unique value of concrete strain,
which should satisfy section equilibrium considerations. If the local equilibrium
is violated a different value of neutral axis depth c should guessed, resulting in a
trial and error procedure to find the real value. The whole procedure of plotting
the graph of moment-rotation and load-deflection are described in the following
section:

4.0 Procedure For Moment-Rotation Analysis For Ductile Connections

In the first stage of procedure moment-rotation analysis of beam-column

connections is to perform trial and error guess by assuming the simplified
hypotheses on the stress-strain distribution of concrete, stress-strain of
reinforcement bars and member compatibility condition. The concept of
“monolithic beam analogy” is used to evaluate the moment-rotation and load-
displacement of beam-column joints in reinforced concrete buildings. The
procedures for moment-rotation analysis for ductile connections in beam-column
joints for reinforced concrete buildings are as follows:
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 272

Step 1: Fix the beam end rotation

The effective rotation ( b ) developed at the beam-column interface due to the

opening of the crack and it can be related to the drift frame system with simple
geometric considerations as shown in Figure 6. The relationship between
effective rotation ( b ) and column rotation ( ) is given by the following
equation:

b (11)
h
1 c
L

Step 2: Guess an initial neutral axis depth c for the beam

center joint

beam
column

b
L

hc

Figure 6: Interstorey drift-beam rotation relationship

Step 3: Evaluate the strain in reinforcement bars

The increase in the strain in the reinforcement bars is due to the beam
deformation is taken into account the length and elongation of reinforcement
bars. The strain of the reinforcement bar is given below:
n. pt
pt (12)
lub
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 273

where n number of total openings along the beam (at beam-column

interfaces), lub length of reinforcement bars in the beam-column interfaces,
pt
elongation (elastic + plastic) at the level of the reinforcement bars.

h
pt
. c (13)
2
h
where c is the relative position of the reinforcement bars (assumed to be
2
at mid-height of the section) and h is the beam height.

Step 4: Estimate in the mild steel and concrete

A strain compatibility cannot be adopted in order to relate the strain in the mild
steel and high yield steel which should be separately evaluated referring to the
deformation of all the beam member (member compatibility). In this case the
concentrate of the rotation at the beam-column interface due to the opening of
the crack simplifies the procedure. The strain in the steel can be estimates as:

Unbonded length

Mild steel

d-c
M Unbonded tendons
M
pt Beam h

c
Column

Figure 7: Gap opening mechanism

 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 274

2 sp
s
(14)
lub
where elongation at the level of the mild steel due to the opening of crack,
sp
displacement due to strain penetration. lub unbonded length of the mild
steel. For the mild steel, the strain penetration is assumed to occur at both ends
of the unbonded region. The total extension reinforcement at the interface crack
is:

.(d c) (15)

where d beam section depth. Contribution in displacement due to strain

penetration may be obtained using the procedure outlined in Sritharan (1998):

2
sp lsp e lsp p
(16)
3

where lsp strain penetration taken as 0.15 f y d bl ; f y yield strength of

reinforcement; d bl diameter of reinforcement bar; e elastic strain in the
beam reinforcement p
plastic strain in beam reinforcement. By substituting

p s e
and e y
in Equation 14, the value of s can be found using
the following equation:

(4 / 3l sp y 2l sp y )
s (17)
lub 2l sp

which can be simplified into:

(2 / 3l sp y )
s (18)
lub 2l sp
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 275

stress (fs)

elastic plastic strain (

s

Figure 8: Elastic and plastic strain components

Following the estimation of the strain in the steel, the strain in the concrete
should be derived with an accurate relationship that can no longer rely on the
classical linear distribution hypothesis. If assuming a correct complete stress-
strain relationship for the concrete, the problem would consist of a system of the
two unknowns, namely the neutral axis depth c and the concrete strain c . Two
equations should thus be introduced by the section equilibrium and a sort of
member compatibility. At this stage, a triangular or rectangular stress-block
assumption can be provided an acceptable approximation. There is no
calculation of c and the procedure is reduced to a trial and error iteration on the
unique unknown, c.

Step 5: Section equilibrium: new value of neutral axis depth ‘c’

The compression resultant in the concrete is calculated from equilibrium

considerations as the section as specify in Equation 19 :

C Ts C's Tpt (19)

where the compressive reinforcement force acting as an external force (Tpt) is

given by the sum of the initial reinforcement bars (Tin) and the increment due to
the deformation of the beam.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 276

Tpt Tin f( pt ) Apt (20)

where pt
the elongation of the reinforcement bars, f ( pt ) the tensile stress
in the reinforcement bars sue to their elongation. The neutral axis depth (c) is
then derived from the compression resultant in the concrete, C, depending on the
hypotheses on the stress-strain behaviour of the concrete. A simplified approach
can be derived by assuming an equivalent stress block to represent distribution of
stress in concrete.

Step 6: Iterative procedure until convergence

The iteration on the neutral axis depth (c) is therefore carried out until
convergence. Rigorously, the initial hypotheses on the relation between the
elastic and plastic components in the reinforcement strain ( pt ) based on the
values of the constant ( e y ) should cross-checked. Therefore few
“double” iteration on (c) and should be performed to reach convergence on
both the parameters. It should be noted that updating is important at small
level of strains (thus at small level of rotation-drift) and becomes negligible at
higher levels when entering the plastic domain.

Step 7: Evaluate the Moment Capacity

The moment resistance capacity of the section M, corresponding to the fixed

rotation can be obtained by taking moment about an axis (such as through the
mid-height of the section). The graph of moment capacity versus rotation can be
plotted in order to compare with the experimental results.

Step 8: Determine the Load versus Displacement

The final step is to plot the graph load versus displacement based on the moment
versus rotation as described in Step 7. Then, experimental result will be
compared with theoretical results in order to find the correlation between them.
In order to validate between theoretical results and experimental results, the sub-
assemblage 3 sets of half-scale beam-column connections were constructed and
tested under vertical cyclic loading. The next section will describe the
construction of half-scale beam-column connections in heavy structures
laboratory.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 277

5.0 Construction Of The Specimens

After predicting the theoretical result of the beam-column connection, it is

important to validate the experimental result with the theoretical results. Initially,
the beam-column joints together with their foundation beams were assembled
and constructed in heavy structural laboratory, Faculty of Civil Engineering,
Universiti Teknologi MARA, Shah Alam, Selangor. The reinforcement bars
cages for beam-column connections and foundation beam are prepared
separately. Figure 9(a) shows the reinforcement bar cages for beam-column
joints with 300mm spacing of stirrup within the six longitudinal reinforcement
bars. Figure 9(b) shows the reinforcement bars cages for foundation beam with
PVC pipes were inserted into bottom of the plywood. This hole is used to insert
highly threaded rod to clamp the foundation to the strong floor. The beam-
column cage is placed at the center of the foundation beam and tied it to the
bottom reinforcement bar in foundation beam. This connection is similar as pile
cap as designed for medium high-rise buildings.
Figure 10 shows the schematic arrangement of reinforcement bars in beam
and detail connection between beam and column. Accurate measurement and
arrangement of reinforcement bars in beam-column and column-foundation
connection are very important before pouring the concrete. Figure 10(a) shows
the bottom parts of column reinforcement bars which are overlapping to the
foundation beam. Proper connections work and good workmanship contribute to
the strength and stability of the buildings during earthquake attack. Figure 10(b)
shows the location of PVC pipes attached to bottom plywood before pouring of
concrete take place.

(a) (b)

Figure 9: Construction of beam-column connection; (a) schematic arrangement of links in beam;

and (b) location of PVC pipes in foundation beam before pouring concrete.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 278

Figure 11 shows the detail connection between beam-column connections

which will be tested in the laboratory under vertical cyclic loading to imitate the
earthquake motion in vertical direction. Figure 11(a) shows detail connection
using BS 8110 where there is no additional links of stirrup in the connection.
The overlapping of reinforcement bars are placed in side the links of the column.
This type of connection is normally used in most of the construction of
reinforced concrete buildings in Malaysia. Figure 11(b) shows the cross-bracing
of reinforcement bars inside the links and these links are tied together with
longitudinal reinforcement bars of the column. Additional reinforcement bars
were connected between top of beam to the longitudinal bars in column. This
type of connection can transmit the lateral force from the beam to the column
and finally to the foundation.

(a) (b)
Figure 10: Detailing of reinforcement bars in each specimen; (a) connections detailing between
column and foundation ; and (b) beam and column caging of reinforcement bar.

Figure 12 shows the process of concreting for foundation beam using ready-mix
concrete with compressive strength of 30MPa. The caging of foundation beam is
placed into formwork before concreting take place. Figure 12(a) shows the
ready-mix concrete is poured into the foundation beam and vibrator is used to
make sure that the concrete has a proper compaction before hardening and avoid
the formation of honeycomb in the concrete. Figure 12(b) shows the process of
pouring concrete up to top level of foundation beam. The wet concrete was
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 279

poured up to the level of PVC pipe and the hole of PVC must be closed with
plastic to avoid the concrete running through the hole. When concreting finished,
the top part of the foundation beam must be covered with wet rugs for curing
process take place otherwise creep, shortening and the surface crack will occur.

(a) Specimen 1 (c) Specimen 3

(a) Specimen 1 (b) Specimen 2 (c) Specimen 1

Figure 11: Connection detailing in beam-column connection; (a) Connection of Specimen 1; (b)
Connection of Specimen 2; and (c) Connection of Specimen 3.

(a) (b)
Figure 12: Process of concreting for foundation beam; (a) pouring concrete into the formwork of
foundation beam; and (b) leveling top part of foundation beam up to PVC pipe.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 280

After 3 days of curing, formwork can be dismantled and it is ready for

testing. Figure 13 shows the isometric view of the specimen before testing.
Foundation beam is clamped to strong floor using high-yield threaded rods so
that it did not move during testing under vertical cyclic loading test. This
specimen is ready for instrumentation and testing under vertical cyclic loading.

Figure 13: Foundation beam is clamped to strong floor using high-yield threaded rod.

6.0 Experimental Set-Up

Before experimental testing, five linear potentiometers were installed along the
beam and column to measure the deformation under vertical cyclic loading. A
total number of six strain gauges were attached to longitudinal reinforcement
bars in beam and column to measure stress-strain relationship under cyclic
loading. Figure 14 shows the five schematic locations of LVDT and six strain
gauges in the beam and column.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 281

Figure 14: Schematic locations of strain gauges and LVDT along the beam and column.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 282

7.0 Visual Observation On Beam-Column Damage

Figure 15 exhibits the visual observation of Specimen 1 which designed

according to BS 8110 without any anchorages of reinforcement bars in the joints
at 3% drift. Most of the diagonal cracks concentrated at joints and spalling of
concrete cover at top surface of column (see Figure 15(a)). Diagonal cracks were
occurred at middle of intersection of beam-column starting from top to bottom
of the joints and spread down to the bottom of column (see Figure 15(b) and (c)).

(a) (b) (c)

Figure 15: Visual observation on the damage of Specimen 1; (a) spalling of concrete on surface
of column ; (b) diagonal cracks from top to bottom of column; and (c) spalling and cracking of
concrete cover of column.

Figure 16 illustrates the visual damage on Specimen 2 with three anchorage

reinforcement bars are placed at top of beam and connected to the reinforcement
bar in column. Specimen 2 has lesser damage as compared to the connection
which designed according to BS 8110, namely Specimen 1. Based on visual
observation, the upper part of beam is stronger than bottom part of beam due
increase of percentage of steel in concrete. Figure 16(a), (b) and (c) show the
progress development of cracks at the intersection starting from 1% drift to 3%
drift. It can be concluded that this type joint in Specimen 2 has better seismic
performance as compared to the joint in Specimen 1.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 283

Figure 17 shows the visual damages which occurred on Specimen 3 with cross-
bracing of anchorage bars from top to bottom of the link. This type of connection

(a) (b) (c)

Figure 16: Visual damages occurred in Specimen 2 with additional of three reinforcement bars
starting from 1% to 3% drift under vertical cyclic loading; (a) some diagonal minor cracks
occurred in column and beam-column interfaces; (b) surface cracks occurred on reinforced
concrete beam; and (c) spalling of concrete cover on front surface of column .

(a) (b) (c)

Figure 17: Minor damages occurred in Specimen 3 with additional of cross-bracing in the joint;
(a) no cracks at beam-column interface at 0.2% drift; (b) diagonal crack at bottom column; and
(c) diagonal and top diagonal cracks in the column and beam.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 284

has minimum damage and crack as compared to the other two types of
connection. Only hairline cracks were observed on top surface of he connections
and no spalling of concrete occurred either at top or bottom of the connection. It
can be concluded that the vertical cyclic load are transferred to the column
through the cross-bracing of reinforcement bars in the joints. Therefore, it is
suggested that cross-bracing connection should be adopted in current code of
practice to avoid the damage of beam-column connection under long distant-
earthquake excitation.

8.0 Experimental Results And Analysis

Figure 18 shows the hysteresis loops for load versus displacement measured
using linear potentiometers marked as LVDT 1 and LVDT 2 for beam-column
connection which designed according BS 8110 without additional reinforcement
bars in the joints. Initially, the beam-column connections behave linearly and
then non-linearly. Figure 18(a) shows hysteresis loop for LVDT 1 which placed
at end of beam with maximum load of 30kN and displacement of 55mm. Figure
18(b) shows the hysteresis loop for LVDT 2 located at the center of beam. It
seem that the lower part of hysteresis loop has less value of load and
displacement because the foundation beam is not properly clamped to the strong
floor and during testing the foundation beam did not stay stationary and it also
moving upward and downward. The darker line represents the theoretical value
of load versus displacement using the eight steps of moment-rotation and
equations in Section 3. There are closed relationship between the theoretical
values and experimental values of beam-column joints.

(a) (b)
Figure 18: Hysteresis loops of beam-column connection located on Specimen 1; (a) hysteresis
loops for LVDT 1 and (b) hysteresis loops for LVDT 2.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 285

Figure 19 shows the experimental and theoretical values of hysteresis loops (load
versus displacement) for beam-column connection with three anchorage bars at
top of beam and connected to bottom of column. The red line represents the
theoretical values of load versus displacement and the blue lines correspond to
the experimental values of the hysteresis loops of beam-column interface under
vertical cyclic loadings. Figure 19(a) shows the hysteresis loops for LVDT 1
closed to the load cell. There some discrepancies between theoretical value and
experimental value in terms of load and displacement. One of the main reasons is
that while applying the vertical cyclic loading on the beam, the column is
moving and it is not proper clamping to the strong floor. While the beam is
moving, the column is also moving with the same direction between them. In
order to get the best fit line, only the beam is allows to move while the other
structures components such as beam and foundation must be in the stationary
position. Figure 19(b) shows the hysteresis loops for LVDT 2 which located
500mm from LVDT 1. Both of these figures show how much energy dissipated
during ground shaking. The hysteresis loops are closed to each other because the
foundation beam is not properly clamped to strong floor.

HYSTERESIS LOOPS LVDT 1

HYSTERESIS LOOPS LVDT 2
35
35

20
20
LOAD (kN)
LOAD (kN)

5
5
-65 -55 -45 -35 -25 -15 -5 5 15 25 35 45 55 65

-10
-40 -20 -10 0 20 40

-25 -25

-40 -40
DISPLACEM ENT (mm)
DISPLACEMENT (mm)

(a) (b)

Figure 19: Hysteresis loops for Specimen 2 with three anchorage bars ; (a) hysteresis loops for
LVDT 1; and (b) hysteresis loops for LVDT 2.

Figure 20 shows experimental results and theoretical result of hysteresis loops

(load versus displacement) of beam-column interface with additional cross-
bracing reinforcement bars. Figure 20(a) shows the hysteresis loops for LVDT
1. The black darker line represents the theoretical value and maroon lines
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 286

correspond to the experimental results. Both graphs show a close relationship

between them with very minimum percentage differences. Figure 20(b) shows
the hysteresis loops (load versus displacement) for LVDT 2 with theoretical and
experimental results. From these graphs it shows that this type of joint has higher
strength capacity which is 45kN than the connection which designed according
to BS 8110 which is only 32kN. These graphs are stiffer than the graph as shown
in Figure 3.20 where beam-column joints as designed according to BS 8110. In
other words, this type of connection can resist higher seismic loading with
maximum displacement of 60mm. Therefore, it can be concluded that additional
cross-bracing reinforcement bars in beam-column interfaces contribute to better
stiffener, high ductility and more stability. Additional bars and cross-bracing
anchorage reinforcement bars can also increase the strength and shear capacity
of the connections and simultaneously improved the seismic performance of the
system.

(a) (b)

Figure 20: Hysteresis loops for Specimen 3 with cross-bracing reinforcement bars; (a) hysteresis
loops for LVDT 1 and (b) hysteresis loops for LVDT 2.

9.0 DISCUSSION

The seismic performance of beam-column joints under quasi-static vertical

cyclic loading depends on a few parameters such as the arrangement of
reinforcement bars especially at the joints, the percentage of reinforcement bars,
concrete cover, compressive strength and confined concrete. These parameters
have a significant effect on energy absorption, equivalent viscous damping, the
base shear, shear strength, moment resistance, stiffness and ductility of beam-
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 287

column joints. The amount of energy absorbed during earthquake can be

measured by determine the equivalent viscous damping of the system. The
percentage of equivalent damping can be calculated by dividing the area under
the hysteresis loop (load versus displacement) over average area under maximum
strength in forward and reverse loading directions. Figure 21 shows graph in
terms of percentage equivalent viscous damping for three different types of
joints which denoted as Specimen 1, Specimen 2 and Specimen 3. Specimen 1
absorbed a lot of energy as compared to Specimen 2 and Specimen 3. It is
meaning to say that Specimen 1 suffered a lot of damage, followed by Specimen
2 and finally Specimen 3. The first cycle of Specimen 1 has the greatest
percentage equivalent viscous damping which is 14.78% as compared to the
second cycle (12%). In earthquake, the first strike will destroy most of the
buildings, infrastructures and lifeline as compared aftershock. Normally, the first
strike of earthquake released a lot of energy and this energy will destroy and
damage the buildings and infrastructures as compared to aftershock.

EQUIVALENT VISCOUS DAMPING VERSUS DRIFT

16
1 cycle-Specimen 1
14
2 cycle-Specimen 1
EQUIVALENT VISCOUS DAMPING

1 cycle - Specimen 2
12
2 cycle - Specimen 2
1 cycle - Specimen 3
10
2 cycle - Specimen 3
8

0
0 0.5 1 1.5 2 2.5
DRIFT(%)

Figure 21: Equivalent viscous damping versus drift for Specimen 1, Specimen 2 and Specimen 3
at 1 cycle and 2 cycle under vertical cyclic loading.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 288

Table 1 shows the experimental result for maximum loading and displacement
for Specimen 1, 2 and 3. Based on the experimental results, the beam-connection
which designed using cross-bracing reinforcement bars has the highest
maximum loading capacity which is 45kN as compared to conventional method
(34kN) and overlapping (30kN). It is evident that the specimen possessed
bracing connection type in beam-column joint recorded the highest maximum
loading capacity among the three samples. Overlapping samples has recorded the
lowest maximum loading capacity during the experiment. It is observed that
overlapping sample has recorded the highest deflection reading in y-axis
direction (LVDT 1, 2, and 3). Hence, it is reveals that the overlapping type is
less effective connection of beam-column joint among the three samples that had
tested in the laboratory. Therefore, it can be concluded that the bracing joint of
the beam-column connection is the greatest type of connection among the three
samples of connection since it showed the well performance under seismic
condition.

Table 1: Summary of experimental results on the three types of connections

LVDT No. Connection Max Deflection Max load (kN)

Type (mm)
Conventional 63 34
1 Overlapping 75 30
(y-axis) Bracing 70 45
Conventional 32 34
2 Overlapping 36 30
(y-axis) Bracing 35 45
Conventional 3.5 34
3 Overlapping 8 30
(y-axis) Bracing 5.5 45
Conventional 10 34
4 Overlapping 1.5 30
(x-axis) Bracing 2.5 45
Conventional 1.5 34
5 Overlapping 12 30
(z-axis) Bracing 32 45
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 289

10. Conclusion And Recommendation

Based on visual observation, experimental results and discussion as mentioned

above, the conclusion and recommendation can be drawn as follows:

1. Under seismic design, structural engineers should design adequate

percentage of reinforcement bars particular at the intersection between beam-
column. Sufficient anchorage length must be provided to connect the
longitudinal bar in beam to the longitudinal bar in the column.
2. The joints should have adequate strength and stiffness to resist the internal
forces induced by the framing members and external force like earthquake
and wind loadings.
3. In order to prevent the diagonal cracking and concrete crushing, the close
space closed-loop steel ties can be provided around the column bars in the
joint region. It can resist the shear force, thereby reducing the cracking and
crushing the concrete.
4. Based on the experimental results, the cross-bracing of reinforcement bars in
the beam-column connections provided the higher loading capacity and has
better seismic performance as compared to the other two types of
connections.
5. It is recommended that the current code of practice BS 8110 should be
changed to another seismic code of practice such as Eurocode 8 so that
reinforced concrete buildings in Malaysia are safe under long-distant
earthquake attack.

References

Alcocer, S. M., R. Carranza, D. Perez - Navaratte, and R.Martinez. (2002), Seismic

Tests of Beam to Column Connections in a Precast Concrete Frame. Precast
Concrete Institute Journal, V. 47:30-45.
ACI 318-05, 2005), Building Code Requirement for Structural Concrete and
Commentary. ACI Committee 318, 2005, 430pp.
Dodd, L.L., and Restrepo-Posada, J.I (1995), Model for Predicting Cyclic Behaviour of
Reinforcing Steel. Journal of Structural Engineering, ASCE, Vol 34, No.4: 433-
445.
Hamid, N.H.A and Iwan Surdano (September 2006), Dynamic Response of
Slender/Thin Reinforced Precast Concrete Walls Using Shaking Table. Journal-
The Institution of Engineers, Malaysia, Vol.67, No.3:48-55.
Hanson, N.W., and Conner, H.W., (1967), Seismic Resistance of Reinforced Concrete
Beam-Column Joints. Journal of the Structural Division, ASCE, 93:533-560.
 Malaysian Journal of Civil Engineering 22(2) : 264-290(2010) 290

Park, R. and Paulay, T., (1975), Reinforced Concrete Structures, John Wiley & Sons,
United States of America, 758pp.
Paulay, T., Park, R. and Preistley, M.J.N. (1978), Reinforced Concrete Beam-Column
Joints under Seismic Actions. Structural Journal, ASCE Journal, 75:585-593.
Paulay, T., Park, R. and Preistley, M.J.N. (1992), Seismic Design of Reinforced
Concrete and Masonry, John Wiley Interscience, United States of America, 768pp.
Sritharan, S., Priestley, M.J.n, Seible, F. and Igarashi, A. (1998), A Five Storey Precast
Concrete Test Building for Seismic Conditions-an Overview. Proceedings of the
Twelfth World Conference on Earthquake Engineering, Auckland, paper No: 1299.
Uma, S.R. and Meher, P.A., (2002), Seismic Behavior of Beam Column Joints in
Reinforced Concrete Moment Resisting Frames, http://www.iitk.ac.in/nicee/IITK-
GSDMA/EQ31, Accessed 27 August 2007.