experimental evaluation of the effects of steel fiber
TRANSCRIPT
Jordan Journal of Civil Engineering, Volume 12, No. 3, 2018
- 484 - © 2018 JUST. All Rights Reserved.
Experimental Evaluation of the Effects of Steel Fiber Percentage on
the Mechanical Behavior of Reinforced SCC Beams Subjected to
ATC-24 Loading Protocol
H. R. Tavakoli 1), Pedram Jalali 2) and Sina Mahmoudi 3)
1),2),3) Department of Civil Engineering, Babol University of Technology, Babol, Iran. * Corresponding Authoor. E-Mail: [email protected]
2) E-Mail: [email protected] 3) E-Mail: [email protected]
ABSTRACT
Beams containing steel fiber with different fractions (0.1%, 0.2% and 0.3% of volume ratio), as well as without
fiber as reference beam, were fabricated and tested in the laboratory. Monotonic and cyclic loadings were
applied to 12 specimens with different mix designs. Reinforcement details and constituents except for fraction
of fiber were constant in all specimens. The loading protocol which specimens were subjected to was adapted
from ATC-24 guideline. The results achieved from this investigation indicated that adding more fiber in
concrete would improve load bearing capacity in the monotonic and cyclic behavior of the specimens. By
increasing fraction of fiber in concrete, the amount of dissipated energy would increase.
KEYWORDS: Steel fiber, SCC beam, Cyclic loading, ATC-24 loading protocol, Dissipated energy.
INTRODUCTION
Self-compacting concrete (SCC) was presented for
the first time in 1988 in order to achieve a concrete with
stable structure. The early studies on the efficiency of
SCC were conducted by Ozawa and Okamura at Tokyo
University (Okamura, 1997; Okamura and Ozawa,
1994; Ozawa et al., 1996). According to ACI-237R-07
definition, SCC is a highly flowable, non-segregating
concrete that can spread into place, fill the form work
and encapsulate the reinforcement without any
mechanical consolidation (Tavakoli et al., 2017).
Due to the brittle nature of concrete and its limitation
in strain capacity, fiber-reinforced concrete (FRC) has
been pronounced as a construction material to improve
the behavior of concrete in tension, in addition to other
aspects. The reason for using discrete fiber in concrete
is to improve the results after cracking with controlling
the opening and distribution of cracks (Tavakoli et al.,
2018). The mechanical behavior of FRC highly depends
on material, geometry, shape and the physical and
chemical properties of the fiber, but generally, using
fiber is considered as an important step in avoiding
cracks and micro-crack distribution and compensation
of tensile strength weakness (Balaguru and Shah, 1992).
In the past few decades, using fiber as an admixture
in concrete and reinforced concrete members was the
main topic of many investigations (Masud and
Chorzepa, 2015; Song et al., 2004; Valipour et al., 2015;
Wang et al., 2015). Studies have been conducted on
many aspects of using fiber in cementitious materials,
but only a few of them have been carried out on the
mechanical behavior of structures under seismic loads,
Received on 27/11/2017. Accepted for Publication on 14/4/2018.
Jordan Journal of Civil Engineering, Volume 12, No. 3, 2018
- 485 -
which have been modeled via cyclic loading (Shaheen
et al., 2003; Yun et al., 2008). The results of past studies
indicated that adding fiber to RC members enhances the
energy dissipation capacity and strength.
For evaluating the behavior of structures under
major earthquakes in experimental and analytical
investigations, a series of loading regimes based on the
world's most reliable codes and standards was used. The
main aim was to present an appropriate loading history,
which is statistically representative of the full range of
characteristics. This issue becomes possible by adopting
an increasing magnitude loading history. In most
practical cases (except for near-fault ground motions),
the actual sequence of cycles (excursions) occurring in
an earthquake has to be rearranged into cycles of
increasing amplitude in order to avoid commitment to a
single maximum amplitude that may only have meaning
for a specific combination of ground motion and
structural configuration. There is no unique and best
loading history, because no two earthquakes are alike.
The main issue is to account for cumulative damage
effects through cyclic loading. These protocols may not
be completely representative of the earthquakes, but
they evaluate the mechanical behavior of a structure
under cyclic or dynamic loading.
Type of material and its behavior under excitation
are so effective in selecting a loading history for
application to the specimens. As an example, for a
material such as wood, due to different nature of
behavior than that of steel or reinforced concrete, a
specific loading history is more suitable for evaluating
its cyclic behavior. Because a parameter such as yielding
point, neither can be calculated by theoretical
formulations (due to lack of behavioral theories’
improvement) nor can be specified or observed by
laboratory tests, another parameter should be used to
control loading amplitudes. For this reason, loading
histories corresponding to this kind of material are
offered by different guidelines. However, in reinforced
concrete beams, similar to steel, due to the ability of
determination of yielding point and reliable behavior
prediction, it would be a proper choice to use this
parameter to control the amplitudes. Therefore, ATC-24
loading history is employed in this study, in which
amplitudes in different steps are defined by means of
yielding point. ATC-24 recommendations for
experimental work are intended to produce reliable
information that can serve as a basis for analytical
modeling or the development of rational design criteria
or that will demonstrate the integrity and safety of
specific structural configurations under various levels of
ground motion (ATC-24, 1992).
In the present study, the specimens are subjected to
ATC-24 loading protocol and the effects of fiber
percentage (1%, 2% and 3% of volume fraction) are
studied on the mechanical properties of the beams.
Hysteresis loops, envelope curves and energy
dissipation capacity are analyzed and some valuable
results are achieved.
Experimental Program
Materials
For this investigation, hooked steel fibers with 36
mm length, 0.7 mm diameter, aspect ratio of 50 and
yielding strength of 2100 Mpa were added to the
concrete in different volume ratios. P10-3R super-
plasticizer based on modified polycarboxylate ether
with bulk density of 1.1 ⁄ was used. Coarse
aggregate and fine aggregate with maximum dimensions
of 12.5 and 4.75 mm were adopted for this test. Type 2
Portland cement and limestone powder with bulk
density of 2.6 ⁄ were also used for this study.
Mixture Proportions and Casting
In this study, 4 mixture proportions including one
mix design without fiber as reference concrete and 3 mix
designs with different steel fiber volume fractions (0.1,
0.2 and 0.3%) were fabricated and tested. Table 1
demonstrates details of mixture proportions. In all four
mix designs, the amounts of concrete components are
constant, except the fiber volume percentage. Water to
cement ratio was 0.39.
Experimental Evaluation of.... H. R. Tavakoli, Pedram Jalali and Sina Mahmoudi
- 486 -
Table 1. Mixing proportions ( )
Mix design name
(%)
Coarse
aggregate
Fine
aggregate
Limestone
powder Cement Water
Super-
plasticizer
Cont --- 722 826 288.9 413.1 162 7
ST10 0.1 722 826 288.9 413.1 162 7
ST20 0.2 722 826 288.9 413.1 162 7
ST30 0.3 722 826 288.9 413.1 162 7
In Table 1, refers to the fiber volume percentage.
A conventional rotary concrete mixer was used for
constructing concrete. Dry aggregate, limestone powder
and cement were mixed at first and after that water was
added gradually. Finally, the fiber was added to prevent
balling. After the mixing procedure, the concrete was
put into design frameworks and kept for 24 hours in
laboratory situation. After that, the specimens were put
into water for 28 days.
In the process of making, it should be considered that
adding fiber with different volume percentages to SCC
can decrease its performance (Fernandes et al., 2013).
To be sure about this subject, L-box, slump-flow and T50
tests can be carried out and the results are compared with
the European guideline for self-compacting concrete
specifications.
Details of Specimens
Twelve rectangular SCC reinforced beams with
similar geometry (cross-section of 200 × 150 mm and
height of 1300 mm) and reinforcement detailing were
fabricated in the laboratory for this test. The dimensions
and distributed steel bars of specimens are demonstrated
in Figure 1. Two ribbed bars with 8mm diameter were
applied for longitudinal reinforcement of top and bottom
of the specimens. Yielding stress and ultimate stress of
longitudinal reinforcement are 405 Mpa and 600 Mpa,
respectively. The dimensions of the samples were
selected based on the maximum dimensions and
capacity of the testing machine.
Test Setup and Apparatus
A UTM (Universal Testing Machine) device was
used in this study for applying cyclic load and is
demonstrated in Figure 2. As is shown, the supports at
the end of beams are hinged, so that no bending moment
is formed at supports and the beam can easily rotate.
This machine contains a loading cell with a loading
capacity of 120 kN and a displacement capacity of ±500
mm. In the process of loading, mid-span displacement
of the specimen was measured via LVDT (Linear
Variable Differential Transducer) and two data per
second were recorded. To accurately apply the load to
the specimen, steel plates are fixed on top and bottom of
the beam via bolts and the load is applied through 180
mm steel plate in the middle of the beam to avoid local
bearing crush failure caused by point load pressure.
Figure (1): Dimensions and distributed steel bars
Jordan Journal of Civil Engineering, Volume 12, No. 3, 2018
- 487 -
Figure (2): Testing apparatus
Loading Program As demonstrated in Figure 3, loading cycle
amplitudes in ATC-24 loading protocol are based on ∆ . Therefore, for this study, because of using fiber
with different ratios in concrete mixture and as a result
changing the concrete's properties, ∆ is obtained from
monotonic loading of the beam specimens for each
design proportion. For this purpose, two specimens from
each mix design were tested monotonically and from the
average results, the yielding point for each of them was
determined separately. Afterwards, ∆ ∆⁄ ratio was
calculated and given to the loading device for
application to the specimens. ATC-24 loading protocol
consists of 7 steps of loading amplitudes, in which for
the first five steps, 3 cycles with repetitive amplitudes
and for the last two, 2 cycles with repetitive amplitudes
are applied.
A low and constant loading rate of 2 mm/min was
adopted for this test to minimize the dynamic effects of
high loading rate and to allow observing the events and
variations during the tests.
Figure (3): ATC-24 loading protocol
RESULTS AND DISCUSSION
Monotonic Response
To study the monotonic behavior of the beams with
different mix designs, one specimen from each design
was tested monotonically. To prove the similarity of the
specimens (similarity in fabrication, casting and
loading), another specimen was tested for each mix
design (repeatability test). The average monotonic
behavior of two specimens of each mix design is
demonstrated in Figure 4 as representative of that
design. The displacement and force values
corresponding to yielding points for all mix designs are
listed in Table 2. Furthermore, displacement amplitudes
for each mix design and every step of ATC-24 are
calculated and presented in Table 3.
Cyclic Response
During major earthquakes, it is expected that the
structure enters into elasto-plastic zone and hysteresis
loops can provide a proper understanding for elasto-
plastic response analysis (Xue et al., 2008). Force-
displacement hysteresis response demonstrates the
energy dissipation capacity of the structure by
considering the area enclosed by the hysteresis loops.
Figure 5 shows the force-displacement hysteresis curves
for all four mix designs.
-600
-400
-200
0
200
400
600
Δ/ Δ
yiel
d(%
)
Number of cycles
5 75200
300400
500
100
3 6 9 12 15 17 19
50
Experimental Evaluation of.... H. R. Tavakoli, Pedram Jalali and Sina Mahmoudi
- 488 -
Table 2. Force and displacement values of yielding points
Yielding Displacement Yielding Force
CONT ST10 ST20 ST30 CONT ST10 ST20 ST30
Mon1 1.4359 1.8499 2.2973 2.6606 23797.8 30598.79 36360.80 37983
Mon2 1.3829 1.8125 2.2977 2.5582 23252.3 30164.40 36309.20 37574.8
Average 1.4094 1.8312 2.2975 2.6094 23525.05 30381.60 36335 37778.9
Table 3. Loading amplitudes of ATC-24
Mix Design
Yield Displacement
(mm)
Loading Amplitude (Percentage of ∆ ) (mm)
50% 75% 100% 200% 300% 400% 500%
Cont 1.45 0.725 1.0875 1.45 2.9 4.35 5.8 7.25
ST10 1.85 0.925 1.3875 1.85 3.7 5.55 7.4 9.25
ST20 2.3 1.15 1.725 2.3 4.6 6.9 9.2 11.5
ST30 2.6 1.3 1.95 2.6 5.2 7.8 10.4 13
Figure (4): Monotonic behavior for various mix designs
In the presented hysteresis curves for each mix
design, the corresponding monotonic curves are also
placed in the diagram to show that the monotonic curves
always cover hysteresis loops, and because of
cumulative damage increment with the increase of
loading amplitude and cycles, the gap between the two
curves becomes more obvious.
Hysteresis curves indicate that cyclic behavior of the
0
10000
20000
30000
40000
50000
60000
70000
0 10 20 30 40 50
For
ce (
N)
Mid Displacement (mm)
CONT
ST10
ST20
ST30
Jordan Journal of Civil Engineering, Volume 12, No. 3, 2018
- 489 -
specimens is linear before cracking and after that, in non-
linear zone, the behavior of the specimens begins to
change. Until the yielding point, stiffness and strength
degradation caused by cyclic loading are negligibles, but
after cracking, slope of the curves (secant stiffness)
decreases with the increase of displacement. Results
achieved based on Figure 5 may be summarized as follows:
In the specimen using reference concrete design,
cracking of the concrete's cover was severe, while in
specimens containing fiber, this phenomenon was rarely
observed. In fibrous specimens, fiber causes integration
in cement matrix and results in less cracking in concrete
cover.
Even though more than two cracks were observed in
all beam specimens, with increasing the amplitude of
displacement, two major cracks propagate on top and
bottom of the beams in a distance equal to the width of
loading plates representing two plastic hinges.
It is assumed that by using fiber in concrete mixture,
instead of the occurrence of a few wide cracks, the
number of cracks increases while their width
diminishes. The width of cracks observed in FRC beam
specimens was less than in the specimen without fiber.
This difference in crack width is more obvious in
loading cycles with higher amplitude than the yielding
point.
In similar and consecutive loading cycles, stiffness
and strength degradations were observed. With
increasing loading cycles, the cumulative damage
effects become more obvious and the behavioral
parameters of the specimens confront more degradation.
Figure (5): Hysteresis curves for different mix
designs: (a) CONT; (b) ST10; (c) ST20; (d) ST30
Envelope Curve
To study the effects of adding fiber on the bearing
capacity of the specimens, the maximum force values
obtained from both positive and negative bending cycles of
each specimen are considered. The resulting curve from
connecting the maximum points is presented as the envelope
curve and is demonstrated in Figure 6 for all 4 mix designs.
-40000
-30000
-20000
-10000
0
10000
20000
30000
40000
-10 -5 0 5 10
For
ce (
N)
Mid Displacement (mm)
Cyclic
Mono 1
Mono 2
(a)
-50000
-40000
-30000
-20000
-10000
0
10000
20000
30000
40000
50000
-15 -10 -5 0 5 10 15
For
ce (
N)
Mid displacement (mm)
Cyclic
Mono 1
Mono 2
(b)
-60000
-40000
-20000
0
20000
40000
60000
-15 -10 -5 0 5 10 15
For
ce (
N)
Mid displacement (mm)
Cyclic
Mono 1
Mono 2
(c)
-60000
-40000
-20000
0
20000
40000
60000
-15 -10 -5 0 5 10 15 20
Fro
ce (
N)
Mid displacement (mm)
Cyclic
Mono 1
Mono 2
(d)
Experimental Evaluation of.... H. R. Tavakoli, Pedram Jalali and Sina Mahmoudi
- 490 -
Figure (6): Envelope curves
The difference observed in the force-displacement
behavior between positive and negative bending
positions is because of two main reasons. In the stage of
pouring concrete into the framework, the fibers tend to
sediment within the concrete's mixture. Therefore, the
distribution of fibers differs from top to bottom (more in
the bottom) and accordingly the bending capacity of the
section in positive bending position would be more.
Another reason for this difference is the crack
occurrence. In the elastic zone, the force values for
similar displacements in both positive and negative
bending positions are approximately equal. But, in the
inelastic behavior zone, with the specimen cracking in
positive bending position, a drop in strength parameters
and force capacity would result; therefore, with
changing the bending position to negative, for the same
displacement amplitude, less force would be achieved.
It is essential to note that because of the difference in
the yielding point of each mix design, the displacement
amplitude of loading for each of them varies. However,
to compare the results, a same displacement can be used
and as demonstrated in Figure 6, with the increase of
fiber dosage in concrete, the force capacity of the
specimen increases.
Dissipated Energy
The energy entering into structures has two forms:
dissipated energy and recoverable energy. Total
absorbed energy of the structure is equal to the sum of
dissipated energy and recoverable energy. Recoverable
energy includes the element capacity to retain its
primary conditions (Elmenshawi and Brown, 2010 a).
Dissipated energy in a cycle is represented by the area
enclosed by the corresponding load-deflection curve of
the specimen (Figure 7) and is representative of the
structural element capacity to extinguish the earthquake
effects through inelastic behavior of reinforcement steel
bars, which causes excessive fractions and perdurable
deformations (Elmenshawi and Brown, 2010a;
Gençoğlu and Eren, 2002; Maekawa and Dhakal, 2001;
Qiu et al., 2002; Xue et al., 2008). Energy of oscillatory
systems could be dissipated via different mechanisms
and mostly there would be more than one mechanism at
the same time. In vibration of a reinforced concrete
structure, these mechanisms comprise: cracking,
reinforcement yielding and friction between concrete
and steel bars during the slippage. In reinforced fibrous
concrete, friction between fiber and matrix, fiber
yielding, inelastic behavior, fraction and pulling out
fiber from matrix are additional mechanisms of
dissipating energy (Chopra, 2007; Laurent and Ahmed,
2002). Non-linear static methods for design use
parameters dependent on energy dissipation capacity for
assessing non-linear seismic response of structures and
for describing strength and stiffness deterioration in
reinforced concrete elements which are subjected to
cyclic loading (Eom and Park, 2010; Rodrigues et al.,
2012). More dissipated energy represents better
performance of the element (Elmenshawi and Brown,
2010 b).
Because of the different input excitations for each
mix design, to normalize the dissipated energy values,
yielding force and displacement parameters were
adopted from the corresponding monotonic loading
curves. The results achieved from the normalized
dissipated energy in the first cycle of every step of cyclic
loading are presented in Figure 8. As demonstrated, with
the increase of loading amplitude, the dissipated energy
values increase as well.
Also, adding fiber increases the dissipated energy.
As far as in the 6th step of loading, the percentage of
-50000
-40000
-30000
-20000
-10000
0
10000
20000
30000
40000
50000
-15 -10 -5 0 5 10 15
For
ce (
N)
Mid Displacement (mm)
CONT
ST10
ST20
ST30
Jordan Journal of Civil Engineering, Volume 12, No. 3, 2018
- 491 -
dissipated energy increment in comparison to the
reference concrete for 0.1%, 0.2% and 0.3% of fiber
steel percentage is 26%, 28% and 42%, respectively.
The effect of adding fiber is more obvious in higher
amplitudes, where the specimen has cracked and the
bond effect of fiber becomes more significant.
Figure (7): Dissipated energy loop with definitions
Figure (8): Normalized dissipated energy
0
0.5
1
1.5
2
2.5
3
3.5
1 2 3 4 5 6 7
Nor
mal
ized
Dis
sipa
ted
ener
gy
Loading step
CONT
ST10
ST20
ST30
Experimental Evaluation of.... H. R. Tavakoli, Pedram Jalali and Sina Mahmoudi
- 492 -
CONCLUSIONS
An experimental study is accomplished on 12
reinforced self-compacting concrete beams containing
different fractions of steel fiber. Monotonic and cyclic
loadings were used to assess the mechanical behavior of
specimens in each mix design. Effects of the amount of
steel fiber on various parameters are investigated and the
results are summarized as follows:
1. By adding more fiber in concrete mix design,
yielding point in monotonic behavior would improve
and the specimen would undergo more force and
displacement elastically.
2. In cyclic loading, by increasing steel fiber in
concrete, force capacity of the members would
increase in an identical displacement, as it is obvious
from the envelope curves.
3. By increasing steel fiber in concrete, more dissipated
energy can be observed from the hysteresis curves.
In elastic amplitudes, the amounts of dissipated
energy for different mix designs are close to each
other and negligible. However, for amplitudes more
than the yield point, differences would appear.
REFERENCES
ATC-24. (1992). "Guidelines for cyclic seismic testing of
components of steel structures". ATC-24, Redwood
City, CA.
Balaguru, P., and Shah, S. (1992). "Fiber-reinforced cement
composites". McGraw-Hill, Inc., New York, USA.
Chopra, A.K. (2007). "Dynamics of structures: theory and
applications to earthquake engineering". Prentice-Hall.
Elmenshawi, A., and Brown, T. (2010 a). "Hysteretic energy
and damping capacity of flexural elements constructed
with different concrete strengths". Engineering
Structures, 32 (1), 297-305.
Elmenshawi, A., and Brown, T. (2010b). "Seismic behavior
of 150 MPa (22 ksi) concrete flexural elements". ACI
Structural Journal, 107 (3), 311.
Eom, T.-S., and Park, H.-G. (2010). "Evaluation of energy
dissipation of slender reinforced concrete members and
its applications". Engineering Structures, 32 (9), 2884-
2893.
Fernandes, C., Melo, J., Varum, H., and Costa, A. (2013).
"Cyclic behavior of sub-standard reinforced concrete
beam-column joints with plain bars". ACI Structural
Journal, 110 (1), 137.
Gençoğlu, M., and Eren, I. (2002). "An experimental study
on the effect of steel fiber-reinforced concrete on the
behavior of the exterior beam-column joints subjected to
reversal cyclic loading". Turkish Journal of Engineering
and Environmental Sciences, 26 (6), 493-502.
Laurent, D., and Ahmed, L. (2002). "Behavior of high-
strength fiber-reinforced concrete beams under cyclic
loading". Structural Journal, 99 (3).
Maekawa, K., and Dhakal, R. (2001). "Post-peak cyclic
response analysis and energy dissipation capacity of RC
columns".
Masud, M., and Chorzepa, M. G. (2015). "Impact resilience
of multi-scale fibre-reinforced composites". Magazine
of Concrete Research, 68 (8), 379-390.
Okamura, H. (1997). "Self-compacting high-performance
concrete".
Okamura, H., and Ozawa, K. (1994). "Self-compactable
high-performance concrete in Japan". International
Workshop on High-Performance Concrete (Edited by
Paul Zia).
Ozawa, K., Maekawa, K., and Okamura, H. (1996). "Self-
compacting high-performance concrete". Collected
Papers (University of Tokyo: Department of Civil
Engineering), 34, 135-149.
Jordan Journal of Civil Engineering, Volume 12, No. 3, 2018
- 493 -
Qiu, F., Li, W., Pan, P., and Qian, J. (2002). "Experimental
tests on reinforced concrete columns under biaxial
quasi-static loading". Engineering Structures, 24 (4),
419-428.
Rodrigues, H., Varum, H., Arêde, A., and Costa, A. (2012).
"A comparative analysis of energy dissipation and
equivalent viscous damping of RC columns subjected to
uniaxial and biaxial loading". Engineering Structures,
35, 149-164.
Shaheen, H., Rakib, T., Hashem, Y., Shaaban, I., and
Abdelrahman, A. (2003). "Behaviour of RC columns
retrofitted by fiber-reinforced polymers under cyclic
loads". Proceedings of the Sixth International
Symposium on FRP Reinforcement for Concrete
Structures (FRPRCS‐6), 10.
Song, P., Hwang, S., and Sheu, B. (2004). "Statistical
evaluation for impact resistance of steel-fibre-reinforced
concretes". Magazine of Concrete Research, 56 (8), 437-
442.
Tavakoli, H. R., Jalali, P., and Mahmoudi, S. (2018). "A
comparative study of the effect of adding steel fiber on
the mechanical behavior of RSCC beams under
displacement and force control conditions".
Construction and Building Materials, 166, 315-323.
Tavakoli, H. R., Mahmoudi, S., Goltabar, A. R., and Jalali,
P. (2017). "Experimental evaluation of the effects of
reverse cyclic loading rate on the mechanical behavior
of reinforced SCC beams". Construction and Building
Materials, 131, 254-266.
Valipour, H., Vessali, N., and Foster, S. (2015). "Fibre-
reinforced concrete beam assemblages subject to
column loss". Magazine of Concrete Research, 68 (6),
305-317.
Wang, J., Morikawa, H., and Kawaguchi, T. (2015).
"Experimental and analytical investigations of bonding
ultrahigh-strength fibre-concrete panels on reinforced
concrete structures with low concrete strength".
Magazine of Concrete Research, 67 (24), 1329-1339.
Xue, W., Li, L., Cheng, B., and Li, J. (2008). "The reversed
cyclic load tests of normal and pre-stressed concrete
beams". Engineering Structures, 30 (4), 1014-1023.
Yun, H.-D., Kim, S.-W., Jeon, E., Park, W.-S., and Lee, Y.-
T. (2008). "Effects of fibre-reinforced cement
composites' ductility on the seismic performance of
short coupling beams". Magazine of Concrete Research,
60 (3), 223-233.