Engineering Structures 171 (2018) 730–746
Contents lists available at ScienceDirect
Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct
Experimental investigation of rubberised concrete-filled double skin square
tubular columns under axial compression
Mohamed Elchalakania, M.F. Hassaneinb, Ali Karrecha, Bo Yangc,
T
⁎
a
School of Civil, Environmental and Mining Engineering, Faculty of Engineering, Computing and Mathematics, The University of Western Australia, Australia
Department of Structural Engineering, Faculty of Engineering, Tanta University, Tanta, Egypt
c
School of Civil Engineering, Chongqing University, Chongqing 400045, China
b
A R T I C LE I N FO
A B S T R A C T
Keywords:
Rubberised concrete
Concrete-filled double skin steel tubes
Stub column
Confinement
Axial compression
Design model
Waste tyres are among the largest and most problematic sources of waste in modern society due to their durability and high rate of dumping in landfills. One possible recycling alternative is to incorporate waste tyre
rubber as an aggregate replacement in concrete to promote sustainability and utilise the elastic properties of
rubber. Rubberised concrete has not reached its full potential because of the decrease in compressive strength
and a lack of research to solve such challenge. Recent research suggests that combining rubberised concrete with
confinement increases ductility and energy absorption. Specifically, confined rubberised concrete using single
skin or double skin square hollow section tubular columns present higher ductility than those made of normal
concrete. This study explored experimentally the use of rubberised concrete filled single skin and double skin
steel tubes under concentric axial compression. The experimental investigation included changing the confinement of the outer and inner square hollow sections and explored how confinement affected normal concrete
compared to rubberised concrete. Four variations of double skin steel tubes with a total of twelve 300 mm long
columns of 0%, 15%, and 30% rubber replacement were created and tested concentrically. Three single skin
short columns with 0%, 15%, and 30% rubber content were also tested and compared. The compressive
strengths were determined theoretically and compared against those measured experimentally. An interesting
spring back phenomenon occurred where the infill rubberised concrete moved upwards after testing due to the
large confinement of the core and elasticity of the rubber. This study examined the use of rubberised concrete
filled double skin steel tubular columns as a promising construction technique for applications such as columns
in buildings located in seismic active zones, security bollards and flexible road side barriers.
1. Introduction
1.1. Development of rubberised concrete (RuC)
Currently, waste tyres are among the largest and most problematic
sources of waste for modern society due to their durability and high rate
of dumping in landfills [1]. In the USA, the total amount of tyre rubber
waste is 20.53 million ton/year and as large as 87% of such amount is
recycled every year [2]. In Europe, the total amount of tyre rubber
waste is 28.92 million ton/year and only 69% of such amount is recycled. In Australia, 50 million tyres are wasted every year [2]. Tyre
landfills can be harmful to the environment and surrounding areas by
providing a breeding ground for mosquitos, rats and other animals.
Additionally, if a fire started in a tyre landfill, it becomes hard to distinguish, and it gives rise to harmful smoke and noxious emissions.
Accordingly, waste tyre management and disposal is a major
⁎
environmental concern in many countries because waste tyres are becoming a significant environmental, health, and aesthetical problem
that cannot be easily solved. A disposal alternative is to incorporate
tyres into the manufacture of the so called Rubberised Concrete (RuC)
as a way to conserve natural resources and reduce the amount of tyres
entering landfills. RuC is a relatively new and innovative field of research aiming at providing a sustainable way of disposing tyres as well
as complementing concrete properties [3,4]. For example, the partial
replacements of sand and cement by rubber enhance the mechanical
characteristics of concrete in terms of its fracture properties, ductility,
impact and seismic resistances [5–7]. Additionally, Liu et al. [8] found
that the ratio of flexural strength to compressive strength of RuC increases relative to normal concrete, indicating that the rubber was
better in anti-cracking performance. Furthermore, Liu et al. [8] found
that increasing the rubber volume content increases the toughness of
the concrete. Hassanli et al. [9] observed that as the rubber content
Corresponding author.
E-mail addresses: mostafa.fahmi@f-eng.tanta.edu.eg (M.F. Hassanein), yang0206@cqu.edu.cn (B. Yang).
https://doi.org/10.1016/j.engstruct.2018.05.123
Received 10 June 2017; Received in revised form 21 May 2018; Accepted 31 May 2018
Available online 18 June 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
filled column
Ppl, Rd, Mod currently modified plastic resistance to axial compression
of the RuCFDST column
PtheoryConc compressive strength of the sandwiched concrete according to Zhao and Grzebieta [18]
PtheorySHS compressive strength of the empty hollow sections according to Zhao and Grzebieta [18]
Pul, EC 4
compressive strength of the CFDST columns with inner
SHSs according to EC4 [16]
Pul, EC 4, Mod currently modified compressive strength of the RuCFDST
based on EC4 [16]
Pul, Zh
compressive strength of the CFDST columns with inner
SHSs according to Zhao and Grzebieta [18]
Pul, Tao
compressive strength of the CFDST columns with inner
SHSs according to Tao and Han [35]
λ
the slenderness parameter of the column
σyf
the yield stress at the flat portions of the cross-sections
σyc
the yield stress at corners of the cross-sections
ζ
the confinement factor used in the calculations by Tao and
Han [35]
χ
reduction factor calculated by using the European strut
curves to account for the overall-buckling
ρs
the ratio of the cross-sectional area of the steel tube to that
of the concrete core
Nomenclature
Ac
Ac, no min al
Asi
Asc
Asc
Aso
D
(EI )e
fck
fcu
fsyi
fsyo
KL
L
ry
Pcr
Pi,u
Posc, u
Ppl, Rd
cross-sectional area of the concrete
the nominal cross-sectional area of the concrete
cross-sectional area of the inner steel tube
the cross-sectional areas of the sandwiched concrete following Tao and Han [35]
the cross-sectional areas of the outer steel tube following
Tao and Han [35]
cross-sectional area of the outer steel tube
specimen width
the effective elastic flexural stiffness of the member
the characteristic concrete strength
the characteristic cube strength of concrete
yield stress of the inner steel tube
yield stress of the outer steel tube
the effective length of the member
specimen length
smallest radius of gyration of the cross-section
the critical buckling load of the column
the compressive strength of the inner tube computed, Tao
and Han [35]
the compressive strength of the outer tube with the
sandwiched concrete following Tao and Han [35]
the plastic resistance to axial compression of the concrete-
improve the workability of the RuC, from which it has been found that
the NaOH pre-treatment of rubber increases the adhesion of rubber to
cement paste and hence it improves the mechanical properties of the
RuC. Another important reason to the lower strengths of the RuC is the
Poisson’s ratio of rubber which is twice that of concrete and the Young’s
modulus which is about 1/3 that of concrete [10]. According to Youssf
et al. [10], this leads to large relative deformations between rubber and
concrete leading to early cracking. Additionally, there are high internal
tensile stresses perpendicular to the direction of the compression load
attributable to the low modulus of elasticity of the rubber particles
[10]. This insight by Youssf et al. [10] leads to the importance of understanding confinement of rubber concrete as a way of reducing stress
increases, the compressive strain capacity of the members increases.
Also, they found that adding rubber to concrete increases the viscous
damping ratio and kinetic energy [9].
1.2. Methods used to enhance the mechanical properties of the RuC
Despite of the above mentioned advantages, the RuC are characterised by a significant reduction in its compressive, tensile and
flexural strengths [3,5,10]. Experimental testing [3] showed that the
lower workability of the RuC, caused by loss of adherence between the
surface of rubber particles and the cement, is one reason of such lower
strengths. Therefore, several investigations [11–14] were undertaken to
Fig. 1. Cross-section of square CFDST columns.
731
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
The concrete mixes for RuC15% and RuC30% is provided in Table 2.
and deformation perpendicular to the direction of the compression
load. This encouraged Duarte et al. [15] to conduct large scale tests on
rubberised concrete-filled (RuCFST) columns with outer steel confinement under static compression. Duarte et al. [15] indicated that the
decrease in axial strength with confinement was not as large as that
taking place without confinement due to the contribution of the steel
tube to the column’s capacity. Positively, the short steel tubes with
rubber concrete presented a higher ductility. They, additionally, discovered that Eurocode 4 [16] provided good but slightly conservative
estimates of ultimate strength for the confined square columns. Moreover, Youssf et al. [12,13] studied crumb RuC confined by fibre reinforced polymer tubes as a means of overcoming material deficiencies
such as decreased compressive strength. In conclusion, it has been
realised that RuC with outer confinement can be a major benefit for
structures in seismic areas where energy dissipation requirements are
mandatory [12,13,15].
2.1.2. Steel tubes
Cold-formed steel manufactured to AS1163 [23] was used in the
construction of the specimens, which was delivered by Midalia Steel in
Bibra Lake, Western Australia. The 100 mm × 100 mm × 5 mm square
hollow sections (SHS1005) and 50 mm × 50 mm × 5 mm square
hollow sections (SHS505) were epoxy painted sections grade DuragalPlus C350LO. The 100 mm × 100 mm × 2 mm square hollow sections (SHS1002) and 50 mm × 50 mm × 2 mm square hollow sections
(SHS502) were galvanised sections grade DuragalPlus C350L0. The
SHSs with a width of 100 mm is used as the outer tubes of the CFDST
columns, while those with a width of 50 mm are the inner ones. It is
worth pointing out that the steel material properties, based on standard
tensile coupon tests, are provided in Section 4.2.
2.1.3. Rubber particles
In order to fit in the SHS annulus, a 7 mm maximum aggregate was
required. Fig. 2 shows the relative size of the rubber particles used in
the concrete mixes. Rubber replacement of 0%, 15% and 30% by weight
of coarse aggregates were selected to show significant results. The
rubber was obtained from Tyrecycle in New South Wales, which is a
leading national tyre recycler. The rubber was delivered in bags consisting of sizes 2–5 mm and 5–10 mm. The 5–10 mm aggregate was
sieved through a 6.75 mm sieve to be replaced with the 7 mm aggregate.
The sieve test results are shown on the particle size distribution
graph illustrated in Fig. 3. It is seen that the sieved 5–7 mm rubber
proved as a good replacement for the 7 mm aggregate with a similar
particle size distribution.
1.3. Double-skin tubular (CFDST) column
It has been widely accepted that the central concrete, in the CFST
columns, closing to the neutral axis has insignificant contribution to the
flexural strength [17]. Accordingly, the central part of the concrete core
of the CFST column can effectively be replaced by another smaller
hollow steel tube with similar axial, flexural and torsional strengths
maintained. This form of column construction is known as the concretefilled double-skin tubular (CFDST) column, which is available in four
different combinations by using the square and circular hollow sections
(SHS and CHS, respectively) [18–21]. Fig. 1 provides the basic crosssectional form of the CFDST columns previously tested by Zhao and
Grzebieta [18] using normal concrete. The results of such columns [18]
showed that the CFDST columns are characterised by increased ductility and energy absorption under compression compared with bare steel
tubes. Accordingly, these CFDST columns have already been implemented in bridge piers in Japan to reduce total bridge weight whilst
maintaining large absorption capacity against seismic loading [18].
2.2. Concrete compression tests
Concrete cylinders 100 mm diameter and 200 mm long were prepared for 0%, 15% and 30% rubber content and tested in a 600 kN
capacity Baldwin Machine to AS1012.9 [24], at 28 days after the cylinders were poured. The stress-strain response from the standard cylinder tests was obtained to be used in the discussion of the results;
especially when the energy absorption and ductility of the current
columns are discussed. The density of the concrete cylinders of the
normal concrete (NC) is 2615 kg/m3, from which the density of the mix
decreased by approximately 8.1% and 14.4% at 15% and 30% rubber
replacements, respectively. These reductions in density are consistent
with those obtained for normal strength rubberised concrete [1,2]. The
present compressive cylinder test results are relatively low compared
with those tested by Zhao and Grzebieta [18] on ordinary CFDST columns which were equal to 71.3 MPa. Accordingly, the comparison between the ordinary and rubberised CFDST columns will focus on those
tested in the current investigation not elsewhere.
1.4. Research significance
This study explores experimentally the use of rubberised concrete
filled double skin cold-formed steel tubes as possible alternatives for
applications such as columns in buildings in seismic active zones, security bollards and flexible road side barriers. In particular, it is devoted
to the experimental investigation of the CFDST stub columns filled with
RuC under the axial compression, with the main aim of combining the
advantages of both the RuC and the CFDST columns into one structural
element. The current experiments involved 0, 15, 30% by weight
rubber replacing the fine and coarse aggregate. This experimental
campaign focuses on the confinement mechanism of the RuC in the
CFDST columns and how it can potentially negate the compressive
strength loss whilst maintaining positive rubber characteristics. It is
worth pointing out that there are currently no experimental results on
literature investigating the square CFDST filled with RuC, which shows
the importance of this paper within the RuC research field.
2.3. Rubber pre-treatment
Conformed to previous investigations [11–14], the rubber used in
this investigation had to be pre-treated in order to remove the oil and
dirt from the outer surface and to improve the overall strength of the
concrete. The oil and dirt on the surface could have created an unwanted layer between the cement paste and rubber surface, which stops
a strong adhesion between rubber aggregate and cement matrix. The
2. Materials and methods
2.1. Material properties
2.1.1. Concrete
General purpose Portland cement to AS3972 [22] was acquired
from Swan Cement pty ltd in Western Australia and used as the binder
material in the normal and rubberised concrete mixes. The chemical
composition of the cement is shown in Table 1. The control mix had
213 kg/m3 water, 426 kg/m3 cement, 750 kg/m3 of 7 mm crushed rock
coarse aggregate, 130 kg/m3 of 4 mm crushed rock coarse aggregate,
and 843 kg/m3 of fine sand. The water/cement ratio was w/c = 0.5.
Table 1
Chemical composition of cement (w%).
732
Cement type
SiO2
CaO
Al2O3
Fe2O3
MgO
SO3
LOI
Na2O
Swan Grey Cement Type
GP AS3972 [22]
20.6
63.5
5.2
3.0
1.3
2.6
1.8
0.5
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
workable so that it could fit it into the SHS annulus of 21 mm minimal
gap. Given the rubber was partially soaked in water beforehand, it was
required to account for the water in the rubber. The difference in rubber
weight before and after the full pre-treatment process was deducted
from the free mixing water. This was chosen to keep uniformity across
all the mixes. The mix design also included replacing coarse aggregate
with rubber by weight up to 30%, to ensure a high replacement of
aggregate to show an opportunity for large amounts of rubber waste to
be used in RuC.
Table 2
Mix quantities of this paper.
Mix
NC
RuC15
RuC30
Water
Mix proportion
(kg/m3)
Coarse
aggregates
(7mm)
Mix proportion
(kg/m3)
Tyre rubber
aggregates
(2–7 mm)
Mix proportion
(kg/m3)
Mix proportion
(kg/m3)
973
827
681
750
638
525
0
112
224
213
213
213
Fine aggregate
(0–4 mm)
3. Test program
NaOH pre-treatment was aligned with the previous research by Elchalakani [1,2], which suggested treating the rubber in 10% NaOH solution for 24 h. This roughened the rubber surface to the optimal level,
allowing a stronger bond between cement paste and rubber. A shorter
time did not alter the surface of the rubber and a longer time roughened
the surface too much, allowing small air pockets to appear on the
surface of the rubber [11–14]. In addition to this, Zinc stearate is an
additive which is added to tyre rubbers to make them more resistant to
oxidation. Zinc stearate makes rubber more hydrophobic, but is turned
soluble in NaOH solution. The rubber was semi saturated through a
water soaking process which allowed the now formed soluble Sodium
stearate to wash off and wash the NaOH off the rubber surface. The
water soaking also increased the specific gravity of the rubber in the
concrete mix, preventing the rubber from floating during the curing
stage.
3.1. Specimens
Twelve CFDST stub columns in addition to three CFST stub columns
were tested in this investigation. As stated in Section 2.1.2, the steel
tubes of 100 mm width are used as the outer tubes (termed as O5 and
O2, respectively for those with 5 and 2 mm thickness) of the CFDST
columns, while the inner tubes are those formed from SHSs with 50 mm
width (termed as I5 and I2 for those of 5 and 2 mm thickness, respectively). At the end of the column's designation, the weight of the rubber
replacement is given. For example, SHS-O5I2-15 belongs to the CFDST
column which is filled with RuC with 15% rubber replacement. This
column is formed from outer tube of 100 and 5 mm width and thickness, respectively, and the inner tube has a width and thickness of 50
and 2 mm, respectively. The three CFST columns are formed from the
steel tubes O5, and hence the designation of these columns does not
include the letter I. The tubes were tack welded onto a 10 mm thick
mild steel base plate to allow the annulus to be filled and to ensure
concentricity. The specimens were prodded to compact the normal and
rubberised concrete. The specimens were placed in a mist curing room
(90% humidity and 21 °C) for 21 days to limit drying shrinkage, then
removed and placed inside the laboratory for another 7 days. There
were still small amounts of shrinkage in the concrete, so the top of each
specimen was levelled using non-shrink grout to achieve simultaneous
loading on the steel and concrete. In the discussion, specimens with a
thinner outer steel tubes (relative to the inner ones) is called Type A
specimens, while those with thicker exterior skins are called Type B
specimens.
The height of the specimens was selected on the basis of being a stub
column length for a cold-formed shape. This means, according to
Galambos [26], that the height should not be less than three times the
largest dimension of cross section and not more than 20 times the least
radius of gyration. Currently, the specimen overall width (D) was
100 mm, the length (L) was 300 mm and the smallest radius of gyration
(ry) was 31.65 mm; 300 mm = 3D ⩽ L ⩽ 20ry = 633 mm . The thicknesses of the steel hollow sections were selected based on the following
criteria: (1) to allow for the maximum load on the specimens to be less
than 2000 kN, which is capacity of the Amsler UTM, and (2) to achieve
2.4. Concrete mix procedure
The mixing method of the RuC is of great importance because the
rubber has a lower specific gravity than concrete hence due to the vibration process, the rubber migrates to the top section resulting in a
nonhomogeneous mix and reduction in strength [10]. Accordingly, this
investigation did not use any vibration which is the method of removing air voids in concrete and instead was compacted with a steel
rod 12 mm in diameter so that there was a limited chance for segregation. The concrete mixing procedure followed that suggested by Elchalakani [1,2], which could be summarised as (1) mix the dry fine and
coarse aggregates for 1 min, (2) add 10% of the water and mix for
1 min, if RuC, add rubber with 10% of the water, (3) add cement and
mix for 1 min, (4) add half of the remaining water and mix for 1 min,
(5) add the remaining amount of water and mix for 1 min, and (6) add
super very small amount of general purpose super plasticiser, and mix
for 1 min, (7) check slump to AS1012.3.1 [25], if the slump is less than
150 mm then add more super plasticiser until a 150–175 mm slump is
achieved. This target slump was important to successfully fill the
narrow annulus between the steel tubes with segregation or bleeding.
The mix design had a water cement ratio of 0.5, to be more
Fig. 2. Rubber aggregate sizes.
733
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
Fig. 3. Particle size distribution graph.
4. Material properties
different confinement on the concrete, different thickness variations
were chosen, i.e., 2 mm and 5 mm for both inner and outer tubes.
4.1. Rubberised concrete (RuC)
Commonly in past experimental research, rubber has been seen
floating to the surface of the concrete [10]. Through the pre-treatment
process and saturation of the rubber prior to putting it in the mix
(discussed above in Section 2.4), the rubber in the current investigation
appeared to be evenly distributed across the vertical direction; see
Fig. 4. The cylinders were cut vertically into two pieces through the
diameter (of the 100 mm length). From this figure (representing merely
the upper part of the cylinders with a 100 mm depth), it can be seen
that the RuC mix showed evenly dispersed rubber throughout the
100 mm × 100 mm samples cut out from 100 mm × 200 mm test cylinders.
Concrete cylinder strengths of the six cylinders tested to AS1012.9
[24] are listed in Table 3. The compressive cylinder test results indicated that the mix created was about 50.3 MPa, while by adding the
rubber by replacing 15% aggregate resulted in approximately half the
compressive strength of 24.95 MPa. A further increase in rubber content
to 30% replacement of aggregate gave a mix with a compressive
strength of 14.4 MPa. These results confirm with previous investigations that showed the deterioration effect of the rubber on the compressive strength of the concrete [3,10,27,28]. Comparing the 15%
specimens, CT-15-02 showed a significant defect probably due to uneven loading or smaller contact area, therefore the compressive
3.2. Test procedure
The composite columns have been tested at the 28th day of the
concrete pouring; the same day of testing the concrete cylinders. This is
to ensure that the concrete strength is compatible with the standard
cylinder tests. A displacement control procedure was used at a constant
rate of 2 mm/minute. A data logger attached to the Amsler Universal
Testing Machine (UTM) was used to transfer load, displacement and
strain gauge data to the computer. The specimen was set up concentrically with the flat plates of the Amsler levelled horizontally; noticing that an angled plate on the specimen would cause the machine to
load unevenly on the section and will thus will not produce the composite action required. A camera was set up capturing a photo every
30 s across the duration of the test to associate certain visual aspects of
buckling with the load/displacement/strain data. To assess the behaviour of the Normal/Rubberised CFDST, the four variations of steel
hollow sections must be assessed alone. Therefore, empty hollow sections were also tested in the 2000 kN Amsler UTM to determine axial
compression strength and failure mechanisms.
Fig. 4. Inside visual inspection of the RuC of 100 mm × 100 mm cut-outs from test specimens with (a) 0%; (b) 15% and (c) 30% rubber content.
734
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
cold-formed sections grade C35L0 have less yield stress but higher
ductility when compared to galvanised sections [29,30]; and (2) the
effects of residual stress distribution at the weld seam found in small
size cold-formed tubes [29,30]. Note, CT-50-5 has member slenderness
KL/r = 15.43 < 20, which is considered as a short column [26]. Fig. 6
shows the failure modes of the four hollow steel tubes. As can be seen,
all the specimens, except CT-50-5, failed by outwards and inwards local
buckling across the one horizontal plane. This failure mechanism is so
called roof mechanism which is a common failure mode for SHS stub
columns [15,18].
Table 3
Mass and density of compressive cylinders.
Specimen
name
Density
(kg/m3)
Average
density (kg/
m3)
Concrete
cylinder
strength (MPa)
Average
cylinder
strength (MPa)
CT-00-01
CT-00-02
CT-00-03
2256.8
2279.1
2275.9
2271
47.4
52.2
51.2
50.27
CT-15-01
CT-15-02
CT-15-03
2088.1
2084.9
2084.9
2086
24.9
21.1*
25.0
24.95
CT-30-01
CT-30-02
CT-30-03
1948.1
1941.7
1938.5
1943
13.7
14.4
15.0
14.37
5. Test results for rubberised CFDST columns
5.1. Fundamental behaviour
* The top surface of this specimen was excessively polished.
Table 4 summarises the maximum forces of the current experimental campaign, from which the designation system described previously was used to label the specimens. As can be seen, the label firstly
refers to SHS as the hollow section type, then denotes the outer thickness (O) and inner thickness (I) (since the width of outer and inner were
constant through all specimens), then the rubber content is denoted as
00%, 15% or 30%. From the table, it can be noticed that the RuCFDST
specimens had less axial strength compared to normal CFDST specimens; this is due to the lower compressive strength of RuC shown in
Table 3. This also may be attributed to the lower steel capacity taking
place in case of the RuC compared with the ordinary concrete. This is
because the steel becomes under a bi-axial stress state much earlier
because of the high Poisson's ratio of the rubber [10]. The compressive
strength of the RuC30 was less than the RuC15 but in most cases they
performed similarly. This could have been due to the strength of the
steel, accounting for a large portion of the overall strength of the CFDST
column. The thinner inner column showed more evident reduction in
strength between normal CFDST and RuCFDST. This was due to the
inward collapse mechanism of the interior tube, given the RuC effectively had voids inside it, the thicker internal tubes can resist more and
the concrete can condense before the inner tube fails. Additionally, as
discussed in the next section, the 30% RuCFDST showed segregated
sand, aggregate and rubber around the top surface of the specimen,
something that has not been obvious for the 15% RuCFDST. Hence, it is
recommended, in future research, to compact the cement paste of the
strength of the 15% mix was considered as 24.95 MPa by neglecting the
cylinder CT-15-02.
4.2. Empty square hollow sections
The properties of the steel material used in forming the current
CFDST columns, empty cold-formed hollow tubes conforms to AS1163
[23] were tested under axial compression. The results were additionally
used to quantify the effect of the bare steel tubes on the CFDST columns. Furthermore, they were used to assess the suitability of Zhao’s
CFDST axial load theoretical calculations for empty hollow section
predictions [18].
The axial load-axial displacement relationships for the tubes with
different thickness and tube dimensions are shown in Fig. 5, which
shows that increasing the thickness (for a specific tube size) provides
higher strength. Generally, the load-displacement curves show typical
ascending and descending branches with some strain hardening parts
after the maximum load was reached, particularly for compact tubes
with 5 mm thickness. On the opposite, relationship of the fully effective
compact specimen CT-50-5 exhibits a considerable flat plateau just
above 400 kN. This is attributed to the overall buckling failure mechanism which may have resulted from: (1) the effect of surface finish
where this specimen is epoxy coated whereas the 50x2SHS was galvanised (Fig. 6d). Previous experimental research found that epoxy coated
Fig. 5. Empty SHS compression tests - load vs. displacement.
735
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
Fig. 6. Failure mechanism of the empty hollow sections after the tests (a) 100 × 5 SHS, (b) 100 × 2 SHS, (c) 50 × 5 SHS, (d) 50 × 2 SHS.
30% rubber. This is an example for thin outer-thick inner RuC CFDST.
Fig. 8 represents those for the specimens with outer 5 mm and inner
2 mm (O5I2) tubes. This is an example for thick outer-thin inner RuC
CFDST.
In Fig. 7, it can generally be observed that Type A specimens with a
thinner outer steel tube yielded a flatter load-displacement relationship
at the post-peak loading stage. The flatter curve could be because both
tubes have fairly similar area, hence they are contributing almost
equally to the axial compressive strength of the specimen. A compression between Figs. 7 and 8 shows that the outer skin thickness dictated
the strength. In general Type A provided a larger ductility, (will be
discussed in Section 5), whereas Type B provided a larger strength.
The post-peak wave like response in Fig. 8 could be due to a repeated process of deep plastic collapse of the folds formed in the inner
and outer skins associated with drop in the load followed by a significant strain hardening and full flattening and contact of such folds.
Figs. 7 and 8 show that the rubber did not have significant effect at
Table 4
Maximum experimental forces for CFDST/CFST.
Rubber content
Specimen type
0%
Force (kN)
15%
Force (kN)
30%
Force (kN)
SHS-O2I2
SHS-O2I5
SHS-O5I2
SHS-O5
SHS-O5I5
657
810
1302
1318
1555
483 (26.5%)
804 (0.7%)
1190 (8.6%)
1143 (13.3%)
1450 (6.7%)
492 (25.1%)
691 (14.7%)
1191 (8.5%)
1035 (21.5%)
1430 (8.0%)
The numbers in brackets are the percent reductions in strength due to rubber
current RuC by adding by-products such as fly ash and silica fume, as
suggested [31].
Fig. 7 presents the axial load-displacement curves for Type A specimens with outer 2 mm and inner 5 mm (O2I5) tubes of 0%, 15% and
Fig. 7. Load-displacement curves for Type A CFDST specimens with outer 2 mm and inner 5 mm (O2I5) tubes of 0%, 15% and 30% RuCFDST columns.
736
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
Fig. 8. Load-displacement curves for Type B CFDST specimens with outer 5 mm and inner 2 mm (O5I2) tubes of 0%, 15% and 30% RuCFDST columns.
inside the folds formed in the outer tube. Such confinement enhances
the ductility of the composite specimen and allows it to maintain large
residual strength after failure.
Topcu [32] stated that in RuC high internal tensile stresses were
found perpendicular to axial load direction because of the low modulus
of elasticity of rubber particles and its higher Poisson’s ratio. The failure
mode in Fig. 12 shows that confining the concrete provides lateral restraint for the internal tensile forces of the concrete that delays cracking
which in turns enhances the strength in the direction where the axial
load is applied. Despite this, after opening the steel specimen (Fig. 11),
it confirms that the concrete has segregated inside the steel tube and is
worse with higher rubber replacement. The aggregate appears loose
inside the 15% and 30% RuCFDST specimens but still morphs to the
inside of the outer steel section. The concrete shows little rigidity inside
the concrete specimen showing poor compaction due to the high rubber
small deformation prior to the ultimate load (i.e. the elastic stages of
the load-displacement curves shown in Figs. 7 and 8.), but its effect
became more obvious in the post-peak regime, where increasing the
rubber content reduces the residual strength after failure (with respect
to any axial displacement value before the load increases again). On the
other hand, since strain gauges were not mounted on the inner tube, its
behaviour during the test was not clearly understood.
5.2. Deformed shapes of the RuCFDST columns
Fig. 9(a) shows the CFDST specimens prior to concrete pouring. The
deformed shapes of the current columns are illustrated in Fig. 9(b),
from which the outward buckling of the outer steel tubes becomes
obvious. Appendix A shows the progressive axial crushing of SHS-O2I230 with 30% RuC, 2 mm outer, and 2 mm inner CFDST Specimen. It can
be noticed that the failure commonly began by forming one fold slightly
below the top of the specimen, and then progressively propagated down
the specimen with continuing axial crushing. Total of 4 folds along the
full length have completely formed by the end of the test. Frames 17
and 18 show the spring back of the concrete core phenomenon occurred
upon unloading of the specimen. The specimen exhibited an upward
movement of the concrete above the level of the steel tubes. This expansion (see Fig. 10) is obvious with the RuC30 specimen but showed
only a slight expansion in the 15% RuC specimen. The expansion occurred because of the elastic properties of the rubber within the concrete matrix and to a less extent due to confinement of the concrete
core. It is worth noting that the current specimens without rubber as
well as those tested by Zhao and Grzebieta [18] did not show such
phenomenon. The 30% RuCFDST showed segregated sand, aggregate
and rubber around the top surface of the specimen (Fig. 12), the concrete above the steel surface appears to have little structural capacity
and can be removed with minimal force by hand.
5.3. Concrete and outer steel interface zone
During axial compression the confined concrete pushes laterally on
the inner and outer tube attributing to the failure of the steel sections.
The outer steel tube was removed to show the local buckling failure
(shown in Fig. 11). The concrete without rubber bonded extremely well
to the inside of the outer steel and remained on the removed strip as
seen in Fig. 11. The cut out strip separated from the rest of the concrete
on a vertical shear plane and appears to be structural. As the concrete
has nowhere else to go, it behaved more ductile and instead of failing
through cracking vertically down the specimen, it plastically deformed
Fig. 9. CFDST specimens (a) before and (b) after testing.
737
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
Fig. 10. Rubber concrete expansion (0%, 15%, and 3% rubber).
further buckling occurs until the buckled section has strain hardened
[20]. The increase in axial load (after collapse) with further displacement could be associated the external folds are in contact and the
concrete core is progressively crushing further as well as the inner steel
inward folds and touching to form a hardened metal core down the
centre of the specimen. In past experimental research [18] tests are
stopped at small axial displacement, however in the present tests the
axial deflection more than 60 mm were reached. This is more than 20%
axial strain. Note, the test data for SHS-O5I5-30 specimen was lost after
30 mm deflection.
It has previously been found that CFDST columns filled with ordinary concrete have similar performance to traditional CFST columns
of the same dimensions of outer steel tube and strength of materials
[33]. Fig. 13, generally, proves this similar behaviour between the
CFDST and CFST columns only for inner thin tubes (I2); see columns
SHS-O5-00 and SHS-O5I2-00. On the other hand, by adding the rubber
as shown in Figs. 14 and 15, even thin inner tubes (I2) seem to share a
higher load contribution compared with the equivalent CFST columns.
This, however, becomes obvious for 30% rubber ratio; see columns
SHS-O5-30 and SHS-O5I2-30.
Fig. 11. 15% RuCFDST left, 30% RuCFDST right.
content compared to the control normal CFDST specimen shown in
Fig. 12. Self compact concrete will be better for future construction of
RuC CFDST to avoid segragation and poor compaction in the narrow
annulus.
5.4. Load-displacement relationships
5.5. Energy absorption and ductility
For the entire program, Figs. 13–15 show the load-displacement
relationships for 0%, 15% and 30%, respectively. As expected, the
thicker exterior skins (Type B) provided larger strength given their
larger amount of steel area. The notable wave-like CFDST response was
evident during axial displacement after the first peak. The post peak
reduction in load is due to the plastic collapse of the steel tube where no
The energy absorbed by a specimen can be determined by the area
under the load (kN) vs. displacement (mm) curve. The energy absorption for the RuCFDST 2 mm outer, 5 mm inner and 15% RuCFDST was
determined. The components of the composite section were separately
examined to determine the individual energy absorption. The results
Fig. 12. Concrete and outer steel bonding zone of 0% CFDST specimen (inside of CFDST left, strip removed right).
738
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
Fig. 13. Load-displacement relationships – 0% RuCFDST.
nine 100 × 5 mm outer sections inclusive of the three CFST specimens
(the remaining 9 specimens). This leads to the conclusion that CFDST
columns filled with RuC constructed using thin outer sections are more
ductile than their corresponding counterparts with thick outer sections.
An important conclusion that may be drawn from the figure is that the
ductility increases with the increased outer steel section slenderness,
from which similar results were found by Zhao and Grzebieta [18] after
testing CFDST columns filled with ordinary concrete. It is also seen that
O2I2-30 (the 3rd one from the left) showed lower ductility results when
compared with the other thin outer section specimens (the first 6 specimens from left). The exact reason for this is unknown to the authors,
but a possible reason is that the elastic energy (We) was relatively large
compared to other specimens with thin outer section. Fig. 18 also shows
the very slight variations in the ductility indexes across all the rubber
contents. Note, because data was lost after 30 mm deflection, DI3 was
not determined for SHS-O5I5-30.
are shown in Fig. 16, which show the difference between the composite
response and the response of the individual components. The concrete
strength was accounted for by using the stress-strain response from the
standard cylinder test (15% RuC) but using the actual area of the
concrete core, which produced the concrete strength in the 15%
RuCFDST specimen.
Fig. 17 shows a schematic for the method of determination for the
ductility indexes based on the energy absorbed. As shown in the figure,
the elastic energy (We = Area ABG) was determined for the specimens
at a displacement (Δ75) corresponding to 75% of the ultimate load, thus
We = 0.5 × PΔ75 × Δ75 [34]; from which PΔ75 is the load corresponding
to Δ75. Three ductility indices were determined from the energy absorbed (area under P-δ curve) up to 15 mm (point D), 25 mm (point F),
and 60 mm(point J). Through dividing the energy absorption at displacements 15 mm, 25 mm, and 60 mm by the absorbed elastic energy
We, the three Ductility Indexes DI1, DI2, DI3 were determined. Thus, the
ductility indexes were calculated as, DI1 = AABCD/AABG; DI2 = AABCF/
AABG; and DI3 = AABCJ/AABG.
Fig. 18 shows that the thinner six 100x2mm outer sections (the first
6 specimens from left) have higher DI1, DI2, and DI3 than the remaining
6. Strength calculations
In this section, experimental strengths for both the empty tubes and
Fig. 14. Load-displacement relationships – 15% RuCFDST.
739
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
Fig. 15. Load-displacement relationships – 30% RuCFDST.
Fig. 16. Composite Energy Absorption for 15% RuCFDST O2I5-15.
the CFDST columns are compared with the available design predictions.
Firstly, the strengths of the SHSs are compared with those calculated by
the method introduced by Zhao and Grzebieta [18]. This is followed by
the comparison of the strengths of the CFDST columns with the predictions of Refs. [16,18,35].
be = B−2rext if λ e < 40
40
be = (B−2rext ) × λ if λ e > 40 ,
e
λe =
×
σyf
250
.
The comparison results are presented in Table 5, from which it can
be noticed that the adaptation of Zhao’s CFDST axial load theoretical
calculations [18] for empty hollow section predictions showed good
results. The largest difference between theoretical calculations and
experimental results was for the SHS1005 hollow section where the
difference is 8.34%. However, the difference of 8.34% for the specimen
SHS1005 is due to neglecting the parts in the flat portions of the crosssections that own higher yield strengths (similar to the corner).
6.1. Strength prediction of empty hollow sections
The Zhao’s CFDST axial load theoretical calculation method [18]
was used in this section to predict the strengths of the empty hollow
sections. According to this method, the strength is to be calculated as
(see Fig. 1):
2
2 2
PtheorySHS = σyc × π × (rext
−rint
) + 4 × (σyf × be × t )
B − 2rext
ti
(1)
6.2. Strength prediction of the CFDST columns
where σyf and σyc are the yield stresses at the flat and corner portions of
the cross-sections. In the calculations, the following are considered:
6.2.1. Design model by Zhao and Grzebieta [18]
The CFDST axial load theoretical calculation method by Zhao and
Grzebieta [18] is used again in this section to predict the strengths
(Pul, Zh ) of rubberised CFDST columns. According to this method, the
rint = t if t < 3.0 and rint = 1.5t if t > 3.0 ,
rext = 2t if t < 3.0 and rext = 2.5t if t > 3.0 ,
740
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
Table 5
Zhao’s Hollow Steel Tubes axial load theoretical predictions [18]
J
F
K
15
δ
25
L
Specimen No.
SHS1005
SHS1002
SHS505
SHS502
Width B (mm)
Thickness ti (mm)
rint (mm)
rext (mm)
λe
be
Length L (mm)
Yield Stress σyf (MPa)
Yield Stress σyc (MPa)
Area An (mm2)
P Squash Load (kN)
Experimental Squash Load (kN)
Difference %
100
5
7.5
12.5
20.46
75.00
300
465
567
1810
875.63
955.3
8.34
100
2
2
4
62.74
58.66
300
465
567
774
239.59
226.4
5.82
50
5
7.5
12.5
6.82
25
300
465
567
814
410.63
417.3
1.60
50
2
2
4
28.64
42
300
465
567
374
177.62
171.4
3.63
60
follows:
Fig. 17. Schematic showing the method of determination of the Ductility
Indexes (DI1, DI2, DI3) for CFDST/CFST Specimens.
Pul, Tao = Posc, u + Pi, u
where Posc, u is the compressive capacity of the outer tube with the
sandwiched concrete and Pi,u is the capacity of the inner tube computed
as ( Asi fsyi ), where Asi and fsyi are the cross-sectional area and the yield
strength of the inner CHSs, respectively. To determine the capacity
Posc, u , the following equation was put forward:
strength is to be calculated as:
Pul, Zh = PtheorySHS + PtheoryConc
(5)
(2)
in which the calculation of the PtheorySHS is provided in detail in Section
4.2, while Ptheoryconc is given as:
Posc, u = fscy Asco with Asco = Aso + Asc
(6)
PtheoryConc = Ac × 0.85fc′
(3)
π 2 ⎞
π 2
2
2
Ac = ⎛(Bo−2to)2−4 × ⎛rinto− rint
o⎞ −⎛Bo −4rexti− rexti ⎞
4
4
⎠
⎝
⎠
⎝
⎝
⎠
in which Asc and Aso are the cross-sectional areas of the sandwiched
concrete and the outer steel tube, respectively. The strength fscy , defined
in MPa, was given as:
(4)
fscy = C1 χ 2 fsyo + C2 (1.18 + 0.85ζ ) fck
Bo2 − (Bi − 2ti)2
12
(7)
where
.
where the radius of gyration is given as r =
The 0.85 factor in Eq. (3) suggests low or no confinement in square
double skin construction. This could be very true from observations in
the current or previous tests [18] due to the ability of the concrete to
force the inner tube inwards, releasing the confinement pressure on the
concrete and exert a biaxial stress state on the steel skins.
fck is the characteristic concrete strength in MPa (0.67fcu ), where fcu
is the characteristic cube strength of concrete in MPa,
f yo is the yield strength of the outer SHS in MPa
ζ is the confinement factor (
Aso fso
Ac, nominal fck
), and
Ac, nominal is the nominal cross-sectional area of the concrete, given
by D 2−Aso .
6.2.2. Design model by Tao and Han [35]
Proposals to predict the strength of short square CFDST columns
(i.e. the cross-section resistance) with inner SHSs (Pul, Tao ) have been
made by Tao and Han [35]. The predicted strength (Pul, Tao ) is given as
Fig. 18. Ductility index (DI1, DI2, DI3) of CFDST/CFST specimens.
741
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
some highly unsafe results of about 23% (for SHS-O2I2-15). Hence, it is
currently recommended to use the modified EC4 [16] prediction for
future calculation of the compressive strength of the RuCFDST short
columns.
An interesting observation that can be noticed from Table 6 is that
the modified EC4 [16] prediction provides slightly conservative results
for the columns with outer tube thickness of 5 mm (D / t = 20 ), compared with those columns formed by using thin tubes of 2 mm
(D / t = 50 ). This may be attributed to the consideration of unconfined
concrete strengths on the design model by the EC4 [16], while it is a
common fact [36,37] that square columns with D / t ⩽ 29.2 contain some
confinement effects that raise the original concrete cylinder strength.
Therefore, the EC4 [16] provided strengths less than the experimental
values with about 9% in average for the columns with D / t ⩽ 29.2 .
6.2.3. Design model by Eurocode 4 [16]
The EC4 [16] does not contain until now a design resistance model
for the CFDST columns filled even with normal concrete. Instead, it
contains a compressive strength formula for the CFST columns, which is
given as:
Pul, EC 4 = χPpl, Rd
(8)
As can be noticed, Pul, EC 4 is based on the plastic resistance to axial
compression (Ppl, Rd ) accounting for the contribution of different elements. To be able to check the strength of the CFDST columns by using
the EC4 [16] (Pul, EC 4, Mod ), a modification for the Ppl, Rd expression is
currently made to consider the inner tube contribution as presented in
Eq. (9).
Ppl, Rd, Mod = f yo Aso + fc′ Ac + f yi Asi
(9)
7. Conclusions
It is worth pointing out that the effective areas of the steel tubes are
employed in the case of slender cross-sections. The reduction factor ( χ )
is calculated using the European strut curves as:
χ = 1/(φ +
φ2−λ 2 ) ⩽ 1.0
This paper presents an experimental investigation of CFDST/CFST
confined and unconfined rubberised concrete. The results of this paper
are summarised in the following points:
(10)
φ = 0.5(1 + α (λ −0.2) + λ 2)
(11)
(1) The rubber pre-treatment process was successful in creating vertically uniform specimens and avoided rubber particles floating to
the top surface.
(2) The mix design produced compressive strength of 25 MPa 15% RuC,
which is the minimum strength concrete for applications made of
composite structures [16].
(3) The available methods of prediction for the axial strength of CFDST
specimens filled with normal concrete produce close approximations to the present experimental results with RuC.
(4) Energy absorption for composite material is significantly larger
than the components that make it up, showing the positive effects of
composite action.
(5) The phenomenon of concrete core spring back upward movement
was observed in this project and has not been previously researched
in CFDST/CFST. Rubber elasticity and lateral confinement allowed
for this to occur.
(6) The interface zone of the normal concrete and inside of the outer
steel section was extremely bonded showing the concrete behaving
like a ductile material. This shows clearly the significant benefits of
CFST and CFDST as a method of avoiding brittle failure found in
plain concrete.
(7) The ductility index for thinner outer steel specimens was higher
than that of thicker outer steel specimens. Analysis of the ductility
index of the range of specimens showed that the ductility index is
fairly constant across the three rubber replacements.
(8) It is important that the preliminary results of this study of
with α = 0.34 (buckling curve (b)) for 3% < ρs ⩽ 6% which is the case
for the current models where ρs is the ratio of the cross-sectional area of
the steel tube to that of the concrete core. The critical buckling load
(Pcr ), used in the calculation of slenderness parameter (λ ) according to
EC4 [16], is calculated from:
Pcr =
π 2 (EI )e
(KL)2
(12)
where KL is the effective length of the member and (EI )e is the effective
elastic flexural stiffness.
6.2.4. Calculated strengths and discussion
Table 6 shows the theoretical calculation of axial compressive
strength of the RuCFDST columns with different rubber ratios by using
the above three design models. Fig. 1 shows a labelled example of a
CFDST specimen with the radius specified. From this table, it can be
seen that the design model by Zhao and Grzebieta [18] provides the
most conservative results among others, with an average and standard
deviation of 0.87 and 0.131, respectively. On the opposite, the design
models by Tao and Han [35] and the modified EC4 [16] predict the
strengths of the current RuCFDST columns much better compared with
the experimental values (Pul, Exp ). As can be noticed, the average and
standard deviation of 1.03 and 0.108, respectively, by using the design
model by Tao and Han [35], while they are 0.95 and 0.076 by using the
EC4 [16] formula. However, the method by Tao and Han [35] provides
Table 6
Comparisons between predicted and experimental strengths.
Specimen
Pul, Exp [kN]
Pul, Zh [kN]
Pul, Tao [kN]
Pul, EC 4, Mod [kN]
Pul, Zh
Pul, Exp
Pul, Tao
Pul, Exp
Pul, EC 4, Mod
Pul, Exp
SHS-O2I2-0
SHS-O2I5-0
SHS-O5I2-0
SHS-O5I5-0
SHS-O2I2-15
SHS-O2I5-15
SHS-O5I2-15
SHS-O5I5-15
SHS-O2I2-30
SHS-O2I5-30
SHS-O5I2-30
SHS-O5I5-30
657
810
1302
1555
483
804
1190
1450
492
691
1191
1430
666
856
1041
1208
523
711
922
1087
466
652
875
1038
773
963
1328
1545
593
783
1157
1374
521
711
1089
1306
690
879
1239
1456
522
711
1099
1316
454
644
1043
1260
1.01
1.06
0.80
0.78
1.08
0.88
0.77
0.75
0.95
0.94
0.73
0.73
1.18
1.19
1.02
0.99
1.23
0.97
0.97
0.95
1.06
1.03
0.91
0.91
1.05
1.09
0.95
0.94
1.08
0.88
0.92
0.91
0.92
0.93
0.88
0.88
0.87
0.131
1.03
0.108
0.95
0.076
Average
Standard deviation (SD)
742
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
rubberised concrete filled single skin and double skin steel tubular
columns came promising to encourage further research devoted for
such members as a feasible construction method for applications
such as columns in buildings located in seismic active zones, security bollards and flexible road side barriers.
Acknowledgments
The authors would like to deeply thank Liam O’keefe from Tyres
Stewardship Australia and Adrian Jones from Tyrecycle. Thanks are
given to Andrew Sarkady and Anup Chakrabortty from BASF for kindly
donating the superplasticizer required for all the specimens. Thanks are
given the following technicians Matt Arpin, Malcolm Stafford, Jim
Waters and Brad Rose for assisting the students in performing the experiments. Thanks are given to Cameron Marshal and Armin Hosseini,
David Pegrum and Aarin Ryan, former students of UWA for performing
the tests and processing the test data.
Finally, to allow for conceptual design recommendations to be put
forward, additional results should be obtained accounting for different
cross-section sizes, inner-to-outer thickness ratios, and rubber contents
with respect to tube thickness. This is under consideration by the authors at the moment through finite element analyses.
743
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
Appendix A. Progressive axial loading of specimen SHS-O2I2-30
744
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
construction of sustainable road side barriers. Structures 2015;1:20–38.
[2] Elchalakani M, Aly T, Abu-Aisheh E. Mechanical properties of rubberised concrete
for road side barrier. Aust J Civil Eng 2017;14(1):1–12.
[3] Elchalakani M, Basarir H, Karrech A. Green concrete with high-volume fly ash and
References
[1] Elchalakani M. High strength rubberised concrete containing silica fume for the
745
Engineering Structures 171 (2018) 730–746
M. Elchalakani et al.
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20] Elchalakani M, Zhao X-L, Grzebieta R. Tests on concrete filled double-skin (CHS
outer and SHS inner) composite short columns under axial compression. ThinWalled Struct 2002;40:415–41.
[21] Hassanein MF, Kharoob OF. Compressive strength of circular concrete-filled double
skin tubular short columns. Thin-Walled Struct 2014;77:165–73.
[22] AS 3972-2010. General purpose and blended cements, Australian Standards,
Sydney, Australia.
[23] AS1163. Structural steel hollow sections, Standards Australia, Sydney; 1991.
[24] AS 1012.9. Methods of testing concrete – Compressive strength tests – Concrete,
mortar and grout specimens, Standards Australia; 2014.
[25] AS 1012.3.1. Methods of testing concrete – Determination of properties related to
the consistency of concrete – Slump test, Standards Australia; 2014.
[26] Galambos TV. Guide to stability design criteria for metal structures. John Wiley &
Sons; 1998.
[27] Bompa DV, Elghazouli AY, Xu B, Stafford PJ, Ruiz-Teran AM. Experimental assessment and constitutive modelling of rubberised concrete materials. Constr Build
Mater 2017;137:246–60.
[28] Aslani F, Ma G, Wan DLY, Le VXT. Experimental investigation into rubber granules
and their effects on the fresh and hardened properties of self-compacting concrete. J
Cleaner Prod 2018;172:1835–47.
[29] Elchalakani M, Zhao X-L, Grzebieta RH. Bending tests to determine slenderness
limits for cold-formed circular hollow sections. J Constr Steel Res
2002;58(11):1407–30.
[30] Elchalakani M, Zhao X-L, Grzebieta RH. Plastic slenderness limit for cold-formed
circular steel hollow sections. J Struct Eng 2002;3(3):127–39.
[31] Raffoul S, Garcia R, Escolano-Margarit D, Guadagnini M, Hajirasouliha I, Pilakoutas
K. Behaviour of unconfined and FRP-confined rubberised concrete in axial compression. Constr Build Mater 2017;147:388–97.
[32] Topcu IB. The properties of rubberized concretes. Cem Concr Res
1995;25(2):304–10.
[33] Zhao X-L, Han LH. Double skin composite construction. Prog Struct Mat Eng
2006;8(3):93–102.
[34] Elchalakani M, Ma G, Aslani F, Duan W. Design of GFRP-reinforced rectangular
concrete columns under eccentric axial loading. Mag Concr Res 2017. http://dx.doi.
org/10.1680/jmacr.16.00437.
[35] Tao Z, Han L-H. Behaviour of concrete-filled double skin rectangular steel tubular
beam-columns. J Constr Steel Res 2006;62:631–46.
[36] Mander JB, Priestley MJN, Park R. Theoretical stress-strain model for confined
concrete. J Struct Eng, ASCE 1988;114(8):1804–26.
[37] Hu HT, Huang CS, Wu MH, Wu YM. Nonlinear analysis of axially loaded concretefilled tube columns with confinement effect. J Struct Eng, ASCE
2003;129(10):1322–9.
slag with recycled aggregate and recycled water to build future sustainable cities. J
Mater Civ Eng 2017;29(2). http://dx.doi.org/10.1061/(ASCE)MT.1943-5533.
0001748. 12pp.
Khatib ZK, Bayomy FM. Rubberized Portland cement concrete. J Mater Civ Eng
1999;11(3):206–13.
Ho AC, Turatsinze A, Hameed R, Vu DC. Effects of rubber aggregates from grinded
used tyres on the concrete resistance to cracking. J Cleaner Prod
2012;23(1):209–15.
Al-Tayeb MM, Bakar BH, Md Akil H, Ismail H. Effect of partial replacements of sand
by waste rubber on the low impact resistance of concrete. Adv Mater Res
2013;626:696–700.
Xue J, Shinozuka M. Rubberized concrete: a green structural material with enhanced energy-dissipation capability. Constr Build Mater 2013;42:196–204.
Liu F, Zheng W, Li L, Feng W, Ning G. Mechanical and fatigue performance of
rubber concrete. Constr Build Mater 2013;47:711–9.
Hassanli R, Youssf O, Mills JE. Experimental investigations of reinforced rubberized
concrete structural members. J Build Eng 2017;10:149–65.
Youssf O, ElGawady MA, Mills JE, Ma X. Prediction of crumb rubber concrete
strength. In: Smith, ST, editor. 23rd Australasian Conference on the Mechanics of
Structures and Materials (ACMSM23), Byron Bay, Australia, 9–12 December, 2014.
Pelisser F, Zavarise N, Longo TA, Bernardin AM. Concrete made with recycled tire
rubber: effect of alkaline activation and silica fume addition. J Cleaner Prod
2011;19(6):757–63.
Youssf O, ElGawady MA, Mills JE, Ma X. An experimental investigation of crumb
rubber concrete confined by fibre reinforced polymer tubes. Constr Build Mater
2014;53:522–32.
Youssf O, Hassanli R, Mills JE. Mechanical performance of FRP-confined and unconfined crumb rubber concrete containing high rubber content. J Build Eng
2017;11:115–26.
Li LJ, Tu GR, Lan C, Liu F. Mechanical characterization of waste-rubber-modified
recycled-aggregate concrete. J Cleaner Prod 2016;124:325–38.
Duarte APC, Silva BA, Silvestre N, de Brito J, Júlio E, Castro JM. Tests and design of
short steel tubes filled with rubberised concrete. Eng Struct 2016;112:274–86.
Eurocode 4. Design of composite steel and concrete structures- Part 1. 1: General
Rules and Rules for Buildings. British Standard Institution, London, ENV 1994-1-1;
2004.
Ho JCM, Dong CX. Improving strength, stiffness and ductility of CFDST columns by
external confinement. Thin-Walled Struct 2014;75:18–29.
Zhao X-L, Grzebieta R. Strength and ductility of concrete filled double skin (SHS
inner and SHS outer) tubes. Thin-Walled Struct 2002;40(2):199–213.
Hassanein MF, Kharoob OF, Gardner L. Behaviour and design of square concretefilled double skin tubular columns with inner circular tubes. Eng Struct
2015;100:410–24.
746