Engineering Structures 219 (2020) 110608

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: [Link]/locate/engstruct

Axial compressive performance of steel tube columns filled with steel fiber- T
reinforced high strength concrete containing tire aggregate after exposure to
high temperatures
Amirhossein Karimi, Mahdi Nematzadeh
⁎

Department of Civil Engineering, University of Mazandaran, Babolsar, Iran

ARTICLE INFO ABSTRACT

Keywords: The performance of concrete-filled steel tube (CFST) columns including scrap-tire rubber under compression and
Scrap-tire rubber in particular changes in their behavior after thermal treatment has remained mostly understudied. This has led
Steel fiber-reinforced concrete (SFRC) to the lack of the use of rubberized concrete in these columns. This study attempted to examine the post-heating
High temperatures performance of CFST columns subject to several temperatures (20, 250, 500, and 750 °C) with regard to the
High-strength concrete
amount of scrap-tire rubber aggregate used as a volume fraction of sand (0, 5, and 10%), quantity of steel fibers
Concrete-filled steel tube (CFST)
Confinement
in volume (0, 1, and 1.5%), and the ratio of the outer diameter to thickness of the steel tube (43 and 25.4). The
specimens had high-strength concrete cores and seamless steel tubes. The total number of manufactured con-
fined and unconfined cylindrical specimens was 114, which were then exposed to heat and axial compressive
testing. Through these tests, the loading capacity, ductility, toughness, and compressive load-strain curves of
different specimens were explored. Based on the results, although including the scrap-tire aggregate (by 10%) in
the concrete mixture lowered the strength of the confined specimens (by 12%), thermal treatment had no
particular increasing effect on this decline. Furthermore, as the quantity of steel fibers and the thickness of the
steel tube in the heated and non-heated CFST specimens increased, toughness increased in all the specimens, and
the ductility had an ascending trend in most specimens. In addition, the optimum quantity of steel fibers used
here was 1%. In this regard, although the fibers affected the compressive strength of the CFST specimens neg-
ligibly, they demonstrated a considerable improving effect on the ductility and absorbed energy in all specimens
particularly the heated ones. Up to 250 °C, the decline in strength was negligible, while above this temperature,
this decline in strength was much more considerable, and also, the post-peak slope of all the curves decreased.
Finally, due to the importance of predicting the strength of CFST columns after exposure to heat, a relationship
for predicting the loading capacity of these columns under heat was proposed and the associated results were
compared with the experimental results of this work as well as the experimental results reported by others. Good
consistency was seen between the predicted and experimental results.

1. Introduction reinforcement, which provides both axial resistance and confining

pressure, resulting in the strength and ductility improvement of the
The use of concrete-filled steel tube (CFST) columns has become columns. Moreover, the confining steel tube eliminates the need for
commonplace in many modern structures, such as high-rise buildings, formwork, which in turn reduces construction time and saves formwork
residential buildings, industrial buildings, arch bridges, overhead materials [7]. In CFST columns, both the steel tube and concrete par-
walking bridges, subway stations, offshore structures, and protective ticipate in carrying axial loads, and thus, they interact to significantly
structures [1]. This is due to high strength and stiffness as well as large improve the capacity to carry static and seismic loads [8].
ductility and energy absorption capacity that the CFST columns provide In CFSTs, to reduce the dimensions and weight, it is beneficial to use
through their composite action [2,3]. In these columns, due to the high-strength concrete [9,10]. However, the use of high-strength con-
supporting role of the infill concrete, the inward buckling of the steel crete in these columns is also detrimental for confinement and load
tube can be prevented, resulting in greater resistance to local buckling transfer, since it leads to a severe drop in the post-peak axial load
[4–6]. In addition, the steel tube acts as the longitudinal and transverse [11–13]. To obtain ductile behavior, either a high-strength steel tube

⁎
Corresponding author.
E-mail address: [Link]@[Link] (M. Nematzadeh).

[Link]
Received 9 October 2019; Received in revised form 8 March 2020; Accepted 2 April 2020
0141-0296/ © 2020 Elsevier Ltd. All rights reserved.
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

must be used or the diameter-to-thickness ratio (D/t) of the steel tube rubberized concrete-filled steel tube columns are scarce. Duarte et al.
must be reduced, with the latter not being cost-effective [14]. In this [26] focused their study on the short steel columns filled with rubber
regard, to improve concrete behavior (its inherent brittleness), fibers aggregate concrete. Their main finding was that this column type
can be helpful. Fibers enhance the post-cracking behavior of concrete provides a better ductility relative to the corresponding columns filled
by bridging across the cracks and transferring the applied load to the with normal concrete and that its use may prove beneficial in seismic
concrete matrix [15]. Fibers also improve the flexural and tensile regions where energy dissipation is an important criterion. Silva et al.
properties of CFSTs and increase resistance against impacts [16]. Liew [27] addressed the flexural behavior of CFST members filled with
and Xiong [17] conducted an experimental study on the performance of rubberized concrete. Their results suggested a significant effect of the
CFST columns with an ultra-high-strength concrete core, where the rubberized concrete on the ductile performance of these members
compressive strength of concrete reached up to 200 MPa. The results under both monotonic and cyclic flexural loads. Also, the slenderness of
indicated that the stub columns with concrete cores of ultra-high- the cross-section had an insignificant effect on the monotonic and cyclic
strength with steel fibers had greater ductility and strength. Further- behavior of these specimens. Elchalakani et al. [22] investigated the
more, the minimum quantity of fibers for improved compressive be- compressive behavior of double-skin tube columns filled with concrete
havior of the CFST columns was 1%. Portolés et al. [18] tested slender containing rubber particles replacing 15 and 30% of the coarse and fine
steel tubes filled with normal-, high-, and ultra-high-strength concrete aggregate by weight in the concrete mix, respectively. It was reported
under eccentric load. In addition, each concrete class was prepared in that the use of double skin columns not only compensated for the in-
three plain, fiber-reinforced, and steel bar-reinforced types. The ob- herent weakness of the recycled concrete in compression but also led to
tained results showed that among the columns with normal-strength increased ductility.
concrete cores, the one including fibers had a better performance In this paper, the compressive behavior of the CFST columns with
compared to the one with the plain concrete core and almost the same high-strength rubberized concrete core reinforced with steel fibers after
performance as the column with the steel bar-reinforced core. Never- thermal treatment is explored. The presence of rubberized concrete in
theless, introducing fibers in the columns with high-strength and ultra- these columns and its effect on the compressive performance of heated
high-strength cores had no noticeable effect on their mechanical and non-heated specimens have remained understudied until now. In
properties, since in the slender columns, the general behavior controls this regard, the volume percentage of scrap-tire rubber aggregate re-
the section behavior. placing sand (0, 5, and 10%), the volume quantity of steel fibers (0, 1,
Fire is a major destructive phenomenon for structures, which brings and 1.5%), the diameter-to-thickness ratio of steel tube (43 and 25.4),
about billions of dollars’ worth of damage annually. As a result of its and the exposure temperature (20, 250, 500, and 750 °C) were con-
high thermal conductivity, steel experiences sudden loss of strength and sidered as the variables in this study. The reason behind selecting these
stiffness when exposed to heat. Therefore, the necessity to investigate variables is briefly explained above, based on which various parameters
the performance of CFST members under high temperatures to con- consisting of the loading capacity, ductility, toughness, elastic modulus,
struct safer structures is clear. and load-strain curves of the CFST stub columns were investigated.
The fire resistance of the CFST columns with high-strength steel and Furthermore, per every confined specimen, a corresponding unconfined
various concrete grades under constant axial load was investigated by specimen was also tested to assess the role played by the steel tube and
Wang and Young [19]. They found that the diameter and strength of the concrete core in the compressive performance of the CFST columns
concrete core had a greater effect compared to the steel strength on the separately. Ultimately, some relationships to predict the compressive
fire resistance time of the CFST columns. In addition, the fire resistance strength of CFST columns exposed to high temperatures were pre-
experienced a general decreasing trend with an increase in the strength sented.
of concrete and steel. Lu et al. [20] reported on the performance of the
CFST stub columns with high-strength self-consolidating concrete ex- 2. Experimental work
posed to standard fire and showed the thermal behavior of these col-
umns to be similar to that of the column with normal concrete. Li et al. 2.1. Material properties and mixture proportions
[21] investigated the compressive behavior of the recycled aggregate
concrete-filled steel tube stub columns under several high temperatures The high-strength concrete mixture was designed based on ACI 211
in an experimental study. They reported that under the same tem- [28], where Portland cement of type I, as well as natural fine and coarse
perature, the compressive strength and elastic modulus of the speci- aggregates were used. The coarse aggregate had a maximum grain size
mens including recycled aggregate were lower compared to the corre- of 9.5 mm and a specific gravity of 2.65 (according to ASTM C127
sponding specimens including natural aggregate, with this discrepancy [29]). The fine aggregate had a specific gravity and a fineness modulus
becoming larger with increase in the recycled aggregate percentage and of 2.56 and 2.70, respectively (according to ASTM C128 [30] and ASTM
temperature; however, a higher maximum strain was reported for the C33 [31], respectively). As a result of the presence of fibers in the
specimens including recycled aggregate relative to those including concrete, a superplasticizer was used to reach a proper workability,
natural aggregate. whose percentage changed for different mixture designs and increased
Recycling scrap tires seems extremely important due to their high with increasing fiber quantity (according to [15,32,33]).
global production rate. Although using scrap tires as aggregate in In the concrete mixtures, sand was replaced with tire aggregate by
concrete reduces the compressive strength, it provides some positive different percentages. For the tire aggregate used here, 47 percent of
properties such as higher acoustic and energy dissipation, lower weight, weight was between the sieve sizes of 4.75 and 2.36 mm, 36 percent of
higher impact resistance, proper crack distribution, and reduced cost weight was between 2.36 and 1.18 mm, with the remaining particles
[22]. Furthermore, rubberized concrete has a higher plastic energy being smaller (according to ASTM C136 [34]). The particle size dis-
capacity due to the presence of scrap-tire aggregate and a greater tribution curve of tire aggregate is presented in Fig. 1. The tire ag-
ductility due to the elasticity of rubber particles in comparison with gregate had zero water absorption and a specific gravity of 0.92. By
normal concrete. These properties make this concrete a good option for using thermal gravimetric analysis (TGA), the point of degradation
road construction where high strength is not needed, nailing concrete, (decomposition temperature) of the tire was obtained at around 360 °C.
wall panels where low unit weight is necessary, structural elements and Hot-rolled seamless steel tubs with circular sections were used in this
Jersey Barriers exposed to impact, and railroads to fix rails to the study, with the same length and internal diameter of 188 and 82 mm,
ground. respectively. However, their external diameter (and thus thickness) was
Despite many studies addressing concrete including scrap-tire variable, being 89 mm (thickness = 3.5 mm) and 86 mm (thick-
rubber particles [23–25], investigations devoted to the performance of ness = 2 mm). The tubes were of the same material, and through

2
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

Fig. 1. Crumb rubber gradating curve.

Fig. 2. Stress-strain curves for the heated and non-heated specimens.

mechanical practices, their thickness changed from 3.5 to 2 mm.
To calculate fy and extract steel properties, 16 dumbbell-shaped
scrap-tire aggregate in the concrete mixture with the following number
steel coupons cut from the steel tubes’ material (4 coupons per each
giving its percentage replacing the fine aggregate.
temperature) were made and tested under tensile loading in accordance
with ASTM E8 [35]. The yielding strength ( fy ) and elastic modulus (E )
of the non-heated steel tubes used in this study were 342 MPa and 213 2.2. Test specimens
GPa, respectively. The yielding strength and elastic modulus of the
specimens heated at different temperatures were not very different For this study, 114 confined specimens were built in 19 different
from those of the non-heated specimens, due to conducting the tension groups. Each group consisted of six specimens, three of which stayed
test after the cooling of steel. In this regard, after heating at 750 °C, this confined, while for the other three, after exposure to 20, 250, 500, and
difference was obtained as 4 and 8%, respectively, for the two above 750 °C, the steel tubes were cut and removed to evaluate the uniaxial
parameters. Table 1 presents the properties of the heated and non-he- compressive strength of the concrete core alone. This was done to in-
ated steel, where it is seen that the ultimate strain ( su ) and thus the vestigate the actual behavior of concrete attached to the confining steel
ductility of the steel are significantly affected by temperature. Stress- tube after heating, not that of concrete heated without the presence of
strain curves of the steel specimens exposed to 20 and 750 °C are given the steel tube. In the unconfined and confined specimens, the mean
in Fig. 2 for comparison. value of the results of three similar specimens was considered as the
Other researchers have suggested a proper length-to-external dia- final result. In addition to temperature, the variables included factors
meter (L/D) ratio of the steel tube to be in the 2–3 range, and in the deemed to be susceptible to heat, namely the volume quantity of steel
present study, this ratio was considered as 2.13 and 2.21 to minimize fibers, percentage of tire rubber, concrete strength, and the diameter-to-
end effects and total buckling risk, so that the specimens could be re- thickness ratio of steel tube. A summary of specimen details can be seen
garded as stub columns. The fibers were of hooked-end steel type with a in Table 3. To identify the specimens, the number following letter T
length and diameter of 30 and 0.6 mm, respectively (L/D = 50). represents the temperature the specimen has been exposed to and the
Moreover, the tensile strength of these fibers was reported by the numbers seen after letters R and F express the quantity of fibers and tire
manufacturer as 1100 MPa, which was well above the tensile strength rubber in the mixture design, with the number following the hyphen
lower limit of 350 MPa given by ASTM A820 [36]. The melting point of indicating the steel tube thickness in mm.
the steel fibers was 1516 °C. Of the 200 mm initial length of the steel tubes, 190 mm was filled
To evaluate different mechanical properties of hardened concrete, with concrete, and the remaining 10 mm was filled with two 5 mm-
the quantity of fibers and percentage of tire aggregate in the mix design thick steel caps at the top and bottom for proper curing and preventing
were regarded as variables, which would lead to change in the strength the mixture’s water from evaporation. After the end of curing, the two
of the concrete. The proportions of constituents in 1 m3 of the high- concrete-free ends were cut with a lathe machine, so that the final
strength mixture design are listed in Table 2. In the names assigned to length of all specimens became 188 mm, in order to not only remove
the mixtures, the number after letter C represents the mean compressive the additional tube length and provide an even level for concrete and
strength of the unconfined non-heated cylindrical specimens; letter F steel but also create a smooth surface for loading. To measure the
shows the presence of fibers with the following number giving the temperature inside the specimens, thermocouples were placed at the
quantity of fibers in volume; and finally, letter R shows the presence of center of the concrete section at mid-height of the specimens, and then
the temperature variation of the center of the specimens and outside of
the specimens (furnace environment) was obtained and plotted for
Table 1 comparison. This was done for the specimens under different tem-
Mechanical properties of steel tube. peratures (250, 500, and 750 °C) and different thicknesses (2 and
3.5 mm).
T(°C) f y (MPa) fsu (MPa) y p su E (GPa)

20 342 480 0.00161 0.0136 0.149 0.29 213 2.3. Test set-up and instrumentation
250 342 480 0.00164 0.0133 0.135 0.29 209
500 335 470 0.00164 0.0122 0.112 0.29 205
The specimens were heated in an electric furnace. Before applying
750 330 450 0.00168 0.0108 0.089 0.29 196
the target temperatures, all specimens were dried under 60 °C for 24 h,
Note: as suggested by ISO 834 [37], to remove the internal moisture from the
fy = yield stress, fsu = ultimate stress, y = yield strain, p = strain at the be- specimens gradually without damaging them. There exist generally two
ginning of strain hardening, su = ultimate strain, = Poisson’s ratio, approaches to thermal loading application to structural elements and
E = Young’s modulus evaluation of the behavior of structures under fire. In the first approach,

3
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

Table 2
Mixture proportions of high-strength concrete.
Mix ID Cement Gravel Sand Water SP Steel fiber Crumb rubber Micro silica
(kg/m3) (kg/m3) (kg/m3) (kg/m3) (kg/m3) (kg/m3) (kg/m3) (kg/m3)

C85F1R0 468 977.4 734.1 161.2 5.2 78.5 – 52

C80F1R5 468 977.4 685.4 161.2 5.04 78.5 12.7 52
C75F1.5R10 468 977.4 649.8 161.2 6.6 117.75 25.45 52
C65F0R10 468 977.4 649.8 161.2 3.79 – 25.45 52
C75F1R10 468 977.4 649.8 161.2 5.98 78.5 25.45 52

thermal loading and axial compressive loading are applied simulta- would have an almost constant residual strength during the compres-
neously to a structural member, which corresponds to the service state sion testing. Based on the results reported by other researchers [42], the
of the structure, to determine the response (fire endurance time) of the residual strength of the concrete specimens stays almost constant after
structural member during a fire. This method is a common practice remaining at the target temperature for more than one hour. For in-
conducted in some studies [38,39]. In the second approach, the axial stance, for the concrete specimens exposed to 800 °C, the strength loss
compression is applied to a structural member after thermal treatment difference between one hour and four hours of exposure was obtained
and cooling; a practice that is also common in some other studies as only 5% [42]. With regard to these observations, all the CFST spe-
[4,21]. In the second method that was employed in the present study, cimens were kept under the target temperature for one hour, similar to
the main objective is to determine the residual strength of structural the work of Peng et al. [43]. Finally, in phase three, the specimens
elements after a fire incident, so that by knowing the post-fire behavior reached room temperature naturally through opening a small hatch
and strength data of the damaged structural members, an optimal and installed above the furnace (according to [21]). Although the concrete
safe design for retrofitting or strengthening the structure after the fire compressive strength experiences reduction during the cooling phase
can be carried out. In the second method (used in this study), the ap- and depending on the geometry of the specimens, the highest reduction
plication of the heating regime to the confined and unconfined speci- level is obtained after around one week [44], the researchers have
mens generally consists of three phases, including in order, increasing rarely investigated the concrete structures damaged in fire to determine
the temperature up to the target temperature, keeping the specimens at the lowest residual loading capacity. Therefore, in this study, all com-
the target temperature for a specified time, and cooling. In the first pression tests were conducted at least one week after the heating of the
phase, similar to the works of Li and Kodur [40] and Wan et al. [41], CFST specimens.
the furnace reached the target temperature with maximum capacity to To measure the longitudinal and lateral deformation of the speci-
be as consistent as possible with the standard curve proposed by the mens, two vertical and two horizontal LVDTs (linear variable differ-
codes. The standard temperature–time curve presented by ISO 834 [37] ential transformers) mounted symmetrically at the two sides of the
is based on Eq. (1), specimens were used, and the averages of the measured values were
reported. In addition, to measure the longitudinal and hoop strains
Tt T0 = 345 log(8t + 1) (1) more accurately, 4 strain gauges were installed at the mid-height of the
CFST specimens in two pairs with an angle of 180° relative to each
where t is the heating time in minutes, and Tt and T0 are the furnace
other, with the strain gauges in each pair being perpendicular to each
temperature at t and the initial furnace temperature, respectively, in °C.
other. Fig. 3 shows the setup of the test and also the location of strain
The second phase was performed so that the center temperature of the
gauges and LVDTs on a CFST specimen.
specimens at the end of phase one which was lower than the furnace
temperature would reach the target temperature and the specimens

Table 3
Details of specimens and test results.
Specimen ID D/t D (mm) fc (MPa) fcc (MPa) f yc (MPa) f yc fcc VR (%) Vf (%)

T20F0R10-2 43 86 63 95 61 0.64 10 0 0.54

T20F1R0-2 43 86 85.5 120.6 107 0.89 0 1 0.4
T20F1R5-2 43 86 80.1 111 58 0.52 5 1 0.42
T20F1R10-2 43 86 73.3 107.7 68 0.63 10 1 0.46
T20F0R10-3.5 25.4 89 63 122.9 104 0.85 10 0 0.97
T20F1R10-3.5 25.4 89 73.3 132.6 64 0.48 10 1 0.83
T20F1.5R10-3.5 25.4 89 74.1 129.4 100 0.77 10 1.5 0.82
T250F0R10-2 43 86 54.1 97.5 64 0.66 10 0 0.63
T250F1R10-2 43 86 56 103.5 77 0.74 10 1 0.61
T250F0R10-3.5 25.4 89 51.5 117 88 0.75 10 0 1.18
T250F1R10-3.5 25.4 89 53.5 122.7 70 0.57 10 1 1.14
T500F0R10-2 43 86 31.3 86.3 55 0.64 10 0 1.09
T500F1R0-2 43 86 51.1 114.2 61 0.53 0 1 0.67
T500F1R5-2 43 86 41.6 108.2 60 0.55 5 1 0.82
T500F1R10-2 43 86 41.5 102.3 51 0.50 10 1 0.82
T500F0R10-3.5 25.4 89 27.5 109.6 77 0.70 10 0 2.21
T500F1R10-3.5 25.4 89 35.2 118 64 0.54 10 1 1.73
T500F1.5R10-3.5 25.4 89 38.3 122 67 0.55 10 1.5 1.59
T750F1R10-2 43 86 16.7 70.5 31 0.44 10 1 2.04

Notes:
D = Diameter of steel tube, t = thickness of steel tube, fc = Compressive strength of unconfined concrete, fcc = Compressive strength of CFST, fyc = CFST’s stress at
steel tube yielding, VR = tire aggregate volume percentage, Vf = Steel fiber volume quantity, = Confinement factor

4
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

Fig. 3. (a) Test setup; (b) location of strain gauges and LVDTs.

3. Experimental results and discussions made here indicated that the failure mode of the heated specimens was
not different from that of the non-heated specimens, similar to [21,45].
3.1. Failure modes and visual observations Furthermore, after adding fibers and increasing their quantity, the
failure mode still remained the same. Since fibers act after the ap-
After heating the CFST specimens, the surface color of the steel pearance of cracks, this mechanism plays a role in preventing the
tubes experienced some changes, as shown in Fig. 4. After exposure to widening, growth, and propagation of cracks, thus delaying the for-
250 °C, the surface color of the steel tubes became brighter than that at mation of bulges in the steel tubes and decreasing the lateral confining
20 °C (room temperature). However, when the temperature reached pressure.
500 °C, the steel tubes became visibly darker. The steel tubes became Fig. 5c gives a view of the concrete core after the removal of the
almost black after 750 °C, and a thin brick red layer covered their outer steel tube, where a 45-degree shear plane can be seen. A bulge appeared
wall. This layer could be easily removed from the tubes’ surface, and in the CFST specimens in the vicinity of this shear plane, and then
also, during loading and before reaching the ultimate load, a portion of failure occurred. From observing the failed CFST specimens and re-
this layer gradually fell from the surface of the tubes. garding the slippage lines on the steel tube surface, as well as 45-degree
The failure pattern of the specimens under axial compression is cracks after the removal of the steel tubes from the concrete core, it can
shown in Fig. 5. During the compressive loading of the CFST specimens be concluded that all specimens experienced the shear failure mode.
previously exposed to 750 °C, shear slippage lines with a 45-degree
angle relative to the horizontal line could be clearly seen on the surface
of the specimens, as shown in Fig. 5a. This phenomenon was seen on a 3.2. Temperature variation
smaller extent in the specimens heated at 250 and 500 °C. The occur-
rence of this phenomenon near the two ends increases the risk of shear To determine the core center temperature of the CFST specimens
failure in the specimens. Moreover, at 750 °C, aggregate particles lo- and compare it to the furnace temperature during the heating process at
cated on the top and bottom ends of the concrete core became loose all three temperatures, i.e. 250, 500, and 750 °C, the thermocouples
before compressive loading and spalled off. placed at the center of the cross-section at mid-height of the specimens
Fig. 5b demonstrates the failure mode of some of the CFST speci- were used. Fig. 6 compares the temperature variation of the center of
mens after the compression tests. According to this figure, an outward the CFST specimens to that of the furnace. Moreover, to compare the
folding as a result of the local buckling of steel occurred at the top or effect of the tube wall thickness on the temperature variation inside the
bottom of the specimens (near supports). In addition, the observations specimens at 250 °C and 500 °C, the tubes with thicknesses of 2 and
3.5 mm were used (see Fig. 6a and b). In the figure, the number written
after letters T and t represents the target temperature of the furnace in
°C and steel tube thickness in mm, respectively. The internal tempera-
ture of the specimens increases with increasing furnace temperature,
and continues when the furnace temperature remains constant; how-
ever, when the temperature inside the furnace decreases, the tem-
perature at the center of specimens sees a decreasing trend. The max-
imum temperature that the center of the specimens reached for the
furnace temperature of 250, 500, and 750 °C was 178, 437, and 690 °C,
respectively. The discrepancy between the temperatures inside the
specimens and furnace is attributed to a low heat conductivity of con-
crete. However, due to their high heat conductivity, the temperature of
the steel tube is considered almost the same as the furnace.
Fig. 6 shows that during heating at the target temperatures of 250
and 500 °C, the center temperature of the CFST specimens with the
Fig. 4. Color change of the steel tube surface of CFST specimens. 3.5 mm-thick steel tube was slightly higher than that of the specimens

5
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

Fig. 5. Failure mode of the specimens.

with the 2 mm-thick steel tube. This may be due to higher heat transfer becomes greater, as shown in Fig. 7. In addition, at 500 °C, contrary to
in the steel tubes with the higher thickness. This led to greater damage 250 °C, the concrete core is also affected by the dehydration of calcium
to the CFST specimens with the thicker steel tube for the same concrete hydroxide and the consequent higher level of deterioration; hence, the
mixture design. As shown in Table 3, in the series of specimens in- difference seen between the concrete core strengths due to different
cluding T250F0R10, T250F1R10, T500F0R10, and T500F1R10, by tube thicknesses becomes more pronounced. It is also seen in Fig. 6 that
changing the thickness of the steel tube from 2 to 3.5 mm, the com- the temperature at the center of the CFST specimens and that of the
pressive strength of their heated concrete core declined by 4.8, 4.5, furnace become closer at higher target temperatures, such that the
12.1, and 15.2%, respectively. These results show that by increasing the specimens reached 71, 87, and 92% of the furnace temperature at 250,
target temperature, the concrete core strength difference between the 500, and 750 °C, respectively. This led to more serious damage to the
specimens with 2-mm steel tubes and those with 3.5-mm steel tubes concrete core at higher temperatures.

6
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

concrete recovers a fraction of the lost strength through rehydration

[46]. In addition, the confining steel tube loses only a small fraction of
its strength and elastic modulus when exposed to temperatures below
300 °C (see Table 1). Similar observations were reported in [47,48].
However, at 400–600 °C, the calcium hydroxide decomposes and micro-
cracks developed as a result of the escape of the hydration water and
the consequent additional stresses known as “vapor pressure me-
chanism” spread, leading to significant strength reduction of concrete
[43]. The cooling phase extending one week after exposure to these
temperatures involved the absorption of water by calcium hydroxide
and its resulting expansion, resulting in the widening of cracks and
further deterioration of the concrete. Furthermore, this group of spe-
cimens included 10% scrap-tire aggregate, which started to burn in this
temperature range (due to the degradation point of tire rubber being
360 °C), leading to increased weight loss. In this temperature range,
negligible reduction occurred in the yielding strength and elastic
modulus of the steel tube, as shown in Table 1. In the 600–800 °C range,
the mechanism involving the dehydrated of hydrated calcium silicate
(because of its amorphous structure and wide dehydration temperature
range), release of water pressure, and internal stresses led to the ac-
celeration of the deterioration of the concrete [44]. On the other hand,
the decomposition of the tire aggregate had already completed, thus its
weight loss (according to TGA test) reached its maximum value. In this
temperature range, the confining steel tube experienced negligible
strength loss but more noticeable elastic modulus loss (see Table 1).
Fig. 7a shows the loading capacity of the F1R10-2 specimen series at
different temperatures together with the loading capacity of the con-
crete core alone, steel tube alone, and the composite action contribu-
tion. As can be seen, although the capacity reduction of the concrete
core at high temperatures is significant, the capacity reduction of the
composite sections is not significant particularly up to 500 °C. This
reason for this is two-fold. First, the steel tube as the main component
participating in carrying axial compression and the concrete core’s
confining element shows insignificant strength change after heating.
Second, the concrete core in this specimen series is reinforced with 1%
steel fibers, which improves the composite action with the confining
steel tube under axial compression, due to significant strain capacity
and residual strength at high temperatures in this core type (according
to the axial and lateral load-strain curves of the concrete core, shown in
Fig. 8a). In fact, steel fibers increase the confining effect of the steel
tube on the concrete core, such that concrete strength reduction due to
heat becomes less intense. After the cracking, the steel fibers in the
concrete act as bridges and delay the crack opening by transferring
stress across the crack faces [15]. In the steel post-yielding phase, where
Fig. 6. Temperature versus time curves for concrete core and furnace under
the concrete core continues to participate in the load-carrying capacity
high temperatures.
to reach the ultimate strength of the composite cross section, the fiber
bridging mechanism significantly delays the concrete core crushing.
3.3. Compressive strength Given that during this delay the confining pressure of the steel tube is
applied to the concrete core (composite action), the load-carrying ca-
Table 3 summarizes the compressive strength of all non-heated and pacity is improved. Therefore, it can be concluded that the presence of
heated CFST specimens. In addition, Fig. 8 demonstrates the loading steel fibers in concrete increases the contribution of the composite ac-
capacity of the composite sections, steel tube alone, and concrete core tion in the load-carrying capacity of CFST columns. It is also seen in
alone, as well as the contribution of the composite action. In this figure, Fig. 7a that the contribution of the composite action in the CFST spe-
the average of experimental results of three identical specimens is given cimens increases significantly with temperature, with the maximum
and the variation of the results of the three identical specimens is contribution obtained at 500 °C. This contribution increase may be due
presented as an error bar on each graph bar. As per the table, for the to cracks developed in the heated concrete core, leading to faster di-
F1R10-2 specimen series, after the application of temperatures 250, lation of the damaged concrete during the compressive loading, thus a
500, and 750 °C, the compressive strength declined by 3.8, 5, and 35%, significant capacity of the steel tube is used for the confinement. Also, it
respectively, according to which, up to 500 °C, decline in the strength seems that the heated specimens obtain a level of pretensioning after
was small, while at 750 °C, this decline increased significantly. This is cooling, since the lateral deformation of the heated steel tube is re-
due to the fact that for temperature under 300 °C, the concrete core versible while the lateral deformation of the cracked concrete core is
experiences a decline in strength because of the initial dehydration of irreversible, which in turn improves the composite action contribution.
the hydrated calcium silicate together with the escape of chemical In the confined specimens belonging to the T20F1-2 series, as the
water, on one hand, and the thermal expansion of aggregate particles tire aggregate percentage in the concrete core increased from 0 to 5 and
together with the resulting internal stresses, on the other hand. How- then to 10%, the compressive strength decreased by 8 and 11%, re-
ever, after the specimens are returned to ambient temperature, the spectively. In addition, in the confined specimens from the T500F1-2

7
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

Fig. 7. Loading capacity of CFST, unconfined concrete core, steel tube alone, and contribution of composite action.

series, when the tire aggregate increased from 0 to 5 and then to 10%, phenomenon makes the concrete porous and lowers the strength, while
the compressive strength decreased by 7 and 12%, respectively. This at the same time considerably decreases the pore pressure of water
shows that the strength loss of the heated CFST specimens due to the vapor and reduces cracking in the heated concrete. However, in the
increase in the tire percentage was not much different from that of the confined specimens, since the steel tube suffered little damage from
non-heated specimens. In the F1R5-2 and F1R10-2 specimen series, heat and also due to the presence of the concrete-steel composite action,
when temperature increased from 20 to 500 °C, the compressive the strength loss is much less than the concrete core. As shown in
strength of the confined specimens declined by 2.5 and 5% and that of Fig. 7b, the composite action contribution in the heated CFST speci-
their concrete core by 48 and 43%, respectively. This high strength loss mens (at 500 °C) changes negligibly with the addition of 5% tire ag-
of the unconfined concrete core in these two specimen series is a result gregate, while it sees a noticeable change when 10% tire aggregate is
of the reasons discussed above as well as the initiation of tire de- added. Furthermore, the contribution of composite action increases
gradation (at 360 °C) and its increasing weight loss (up to 500 °C). This significantly in the heated specimens compared to the non-heated ones,

8
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

with the maximum increase in the specimen containing 5% tire ag- regardless of the improvement seen in the compressive and tensile
gregate being around 5.5 times. In addition, note that in Fig. 7b, for all concrete behavior. Fig. 7c demonstrates that increasing the quantity of
the heated CFST specimens with different tire percentages, the com- fibers from 0 to 1 and then to 1.5% in the heated CFST specimens in-
posite action contribution is higher than the loading capacity con- creases the composite action contribution by a small amount. However,
tribution of the steel tube alone. in the non-heated CFST specimens (at room temperature), the compo-
Steel fibers generally present their highest efficiency when used by 1 site action contribution became maximum by a small amount for 1%
to 1.5% in concrete volume [16], while above this range, these fibers fibers. Nevertheless, in both heated and non-heated CFST specimens,
have an insignificant role in the improvement of concrete performance the strength of the concrete core alone constantly increased with in-
[49]. In the confined specimens from the T20R10-3.5 series, as the creasing fiber quantity in the volume, indicating the enhancing effect of
quantity of fibers increased from 0 to 1 and then to 1.5%, the com- adding fibers and increasing their volume quantity on the surface
pressive strength of the composite section increased by 8 and 4%, re- bonding characteristics and the matrix strength of the concrete [50].
spectively. It seems that at 20 °C (room temperature), adding fibers by Fig. 7c also shows that in the specimens exposed to 500 °C, the com-
1% is better and more effective than 1.5% for improving the com- posite action contribution is always higher than that of the corre-
pressive strength of the CFST specimens. In addition, in the T500R10- sponding specimens at room temperature per all the steel fiber quan-
3.5 series of specimens, by increasing the quantity of fibers in the tities, the reason for which was discussed above. In addition, the point
concrete from 0 to 1 and then to 1.5%, the strength increased by 8 and worth noting in this figure is that in all the heated CFST specimens with
11%, which shows the complementary effect of increasing the fiber 10% tire aggregate and different fiber quantities, the composite action
quantity from 1 to 1.5% on the strength improvement of the heated contribution is considerably higher than the loading capacity of the
composite section. It is generally found that adding fibers has no sig- steel tube alone. A similar but less pronounced trend is also seen in the
nificant effect on the compressive strength of the CFST specimens, non-heated CFST specimens.

Fig. 8. Load-strain curves of CFST specimens and corresponding concrete cores.

9
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

Fig. 8. (continued)

10
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

In the confined specimens of the T20F0R10 series, when thickness 250 °C, the slope of the CFST curves becomes slightly steeper, while as
increased from 2 to 3.5 mm, the compressive strength increased by temperature further increases to 500 and 750 °C, this slope declines.
26%, and in the specimens of the T20F1R10 series, this thickness However, at 750 °C, considering the great deformation of the specimen
change increased the strength by 20%. However, the compressive occurring at the peak point of the curve, the descending portion cannot
strength of the concrete core in the first and second series was 63 and be identified easily, and after the peak, the curve follows a flattened
73.3 MPa, respectively. Regarding the results obtained for the descending path. In addition, the strain of the CFST specimens at the
T20F0R10 and T20F1R10 specimen series, it can be seen that the peak point of the curves (strain corresponding to ultimate strength)
strength difference of the CFST specimens due to change in the strength increases with temperature. The reason behind the increase in the peak
of the concrete core is less than their strength variation due to change in strain, especially at 750 °C, is the degradation of the mechanical
the D/t ratio. This is because when D/t ratio remained constant, with an properties of the rubberized concrete core, as can be deduced from the
increase in the concrete strength, the ultimate strength of the CFST compressive curves of the unconfined concrete in Fig. 8a. In addition, as
specimens slightly increased, while when the concrete core strength shown in Fig. 8a for the unconfined concrete, the post-peak slope of the
was constant, as D/t decreased, increase in the ultimate strength of the curve for the non-heated specimen is very steep (due to the brittle be-
CFST specimens was more pronounced [36]. In addition, by comparing havior of high-strength concrete) in the form of a sudden drop. This
the first series containing 10% tire aggregate without fibers to the drop is not observed in the heated specimens such that the post-peak
second series containing the same tire aggregate content with 1% fibers, slope is softened at 250 and 500 °C and becomes almost parallel to the
it is found that adding steel fibers with volume quantity of 1% to the horizontal axis at 750 °C.
concrete mix increases the strength of the concrete core at room tem- The load-strain curves of the T20F1-2 and T500F1-2 specimen series
perature by 16%, while it increases the compressive strength of the in which tire aggregate percentage is 0, 5, and 10% are plotted in
composite section with 2-mm- and 3.5-mm-thick steel tubes by 13 and Fig. 8b. At room temperature, as the tire percentage in the CFST spe-
8%, respectively. This shows that adding fibers to CFST columns with cimens increases to 5%, the compressive strength declines, while for the
thicker steel tubes is less effective, since the improvement in the be- tire aggregate percentage of 10%, the strength loss is small. In addition,
havior of these columns is provided mainly by the steel tube, with the all three curves have almost the same slope in the descending portion.
effect of fibers being small. Fig. 7d demonstrates the effect of the steel At 500 °C, the descending branches of the curves of all the CFST spe-
tube thickness on the loading capacity of the CFST specimens, and gives cimens are almost coincident. The same slopes for the descending
the capacity of each component of the composite section in two series of portion of the curves may be due to the presence of 1% fibers in all the
specimens including 10% tire aggregate, with and without fibers. As CFST specimens under consideration. It is also found from Fig. 8b that
can be seen, as the thickness and temperature increase, the composite as the tire aggregate increases, the peak point strain in the curves of the
action contribution in all the specimens with and without fibers in- heated and non-heated CFST columns increases. It can also be seen in
creases, with this contribution in the fiber-reinforced specimens always the curves of the unconfined concrete core specimens shown in Fig. 8b
being higher than the fiber-lacking specimens. (on the right) that increase in the rubber aggregate content results in a
significant increase in the peak strain of the non-heated specimens,
3.3.1. Load-strain relationships while having no significant effect on the post-peak behavior of the
The load-strain curves of the CFST and unconfined concrete core concrete in both heated and non-heated specimens. Fig. 8c gives the
specimens in the longitudinal and lateral directions after the compres- load-strain curves of the T20R10-3.5 and T500R10-3.5 specimen series,
sive test are given in Fig. 8. In this figure, each curve belongs to one where the fiber quantity is 0, 1, and 1.5%. As can be seen, at 20 and
specimen from the three identical specimens. Here, letter C at the be- 500 °C, as the fiber quantity increases, a higher strain at peak stress is
ginning of the names of the specimens indicates that the curve belongs obtained for the CFST specimens, and also, the ultimate lateral strain
to the unconfined concrete core. The axial and lateral strains are ob- experiences a significant improvement particularly in the heated spe-
tained from the average values of shortening measured by the two cimens. This improvement can be clearly seen in the concrete core
vertical and two horizontal LVDTs. Here, the axial strains are presented curve of these specimens. Moreover, when fibers are added to the
in the positive section of the horizontal axis, and the lateral strains are concrete mixture, the compressive strength of the CFST specimens after
shown in the negative section of the horizontal axis. exposure to 20 and 500 °C increases. This improvement in the perfor-
Fig. 8a shows the load-strain curves of the CFST specimens of the mance of the CFST specimens at and after the peak point is attributed to
series F1R10-2 exposed to temperatures of 20, 250, 500, and 750 °C. It the bridging action of steel fibers across cracks, which leads to en-
is seen that with increasing the temperature to 500 °C, the variation of hancement in the performance of the concrete core, as can be seen in
the compressive strength is not notable, while at 750 °C, the compres- the load-strain curves of the unconfined concrete in Fig. 8c. It is worth
sive strength sees a noticeable reduction. Furthermore, by increasing noting that the improvement in post-peak behavior of the concrete core
the temperature to which the CFST specimens are exposed, the elastic (especially in the case of the lateral strain) was more pronounced in the
portion of the curves (strain up to the yielding of the steel tube) de- heated specimens than in the non-heated specimens. This change in the
creases, while their elastoplastic portion (strain between the yielding brittleness of high-strength concrete as a result of adding steel fibers
and peak point) increases. The higher the temperature, the greater the makes the addition of fibers in this concrete type a necessity.
ductility and post-yielding deformation of the specimens. One reason Fig. 8d shows load-strain curves of the CFSTs in which the only
for the good ductility of the CFST specimens is the presence of both variable is the tube thickness, so that the effect of change in the steel
fibers and tire in their concrete core. Their positive synergic effect tube thickness on the performance of the heated and non-heated spe-
works in a way that the tire aggregate provides a high strain capacity cimens can be assessed. The curves cover two fiberless and fiber-re-
while fibers significantly improve the residual post-peak strength [51]. inforced series of specimens. According to this figure, when the tube
In addition, propagation of micro-cracks in the heated concrete core thickness increases, a considerable increase is seen in the compressive
induced by the evaporation of intermolecular water generally leads to strength of the specimens. In addition, peak strain in the specimens
higher ductility and deformation capacity [52,53], which in turn im- with 3.5-mm tube walls increases significantly compared to those with
proves the properties of the CFST specimens. Up to 250 °C, no sig- 2-mm tubes, especially for the fiber-reinforced specimens. With the
nificant change is seen in the curves and the mechanical properties of increase in the steel tube thickness, the linear portion of the load-strain
the CFST specimens are relatively constant. However, when tempera- curve becomes longer and also, the slope of the descending branch and
ture reaches 500 and 750 °C, the initial slope of the curves experiences a thus the axial load loss after the peak point decrease. This improved
considerable reduction, leading to lower stiffness of the specimens. In performance in the CFST specimens with the thicker tubes is a result of
the descending portion of the curves, as temperature increases to higher lateral confinement and enhanced composite action. In the

11
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

curves corresponding to the concrete core specimens, shown in Fig. 8d point which is a point corresponding to 15% loss in compressive
(on the right), it can be seen that in similar heated specimens which are strength in this work [15]. In Fig. 9, the vertical axis represents the
different only in the thickness of the removed steel tube, different post- residual toughness ratio (T) parameter and subscripts T, R, F, and t
peak behaviors are observed due to different thermal properties of the indicate temperature, tire aggregate, steel fibers, and steel tube thick-
steel tubes with 2 and 3.5 mm wall thicknesses. However, the curves of ness, respectively. The residual toughness ratio values are given in
the non-heated specimens perfectly match, and the tube thickness does Table 4 for all CFST specimens.
not affect the load-strain relationship of the concrete core. Moreover, at As can be seen in Table 4 and Fig. 9a, the ratio of energy absorption
500° C, due to greater damage to the concrete core specimens which of the F1R10-2 specimen series exposed to 250, 500, and 750 °C to that
had 3.5 mm-thick steel tubes, in addition to the peak strength, their of the non-heated specimens is 0.89, 1.61, and 3.24, respectively. These
stiffness is also lower. results show that at higher temperatures, energy absorption increases,
Note that in all load-strain curves of the CFST specimens, above an despite the compressive strength reduction, since the peak strains and
axial strain of around 0.002, the stiffness of the specimens experiences a strain at load corresponding to 85% of strength increase (see Table 4).
sudden increase, due to the cracking of the concrete core and the in- By increasing the replacement percentage of tire aggregate from 0 to
itiation of the confinement by the steel tube. However, the closing of 5 and then to 10%, residual toughness ratio in the T20F1-2 specimen
cracks affects the stiffness and shape of the curve in this initial region. series is obtained as 0.82 and 0.85, and in the T500F1-2 series as 1.03
Also, note that the parameters obtained in the following from the load- and 1.10, respectively. It is found from the results that by adding tire
strain curves of the CFST specimens are mean values of 3 similar spe- aggregate, energy absorption decreases in the non-heated specimens,
cimens. while it slightly increases in the heated specimens (see Fig. 9b), due to
the compressive strength loss of the CFST specimens and the downward
3.3.2. Yielding strength shift of their load-strain curve when tire aggregate is present (see
The values of the axial stress of the CFST specimen at the steel tube Fig. 8b). Furthermore, 10% tire aggregate was always more desirable
yielding point ( fyc ) as well as its ratio to the corresponding compressive than 5% tire aggregate in maintaining or improving the toughness.
strength are given in Table 3. In addition, all the yielding points have With respect to Table 4, when fiber quantity increases from 0 to 1
been shown in the load-strain curves of the CFST specimens in Fig. 8. and then to 1.5%, the residual toughness ratio in the T20R10-3.5 series
According to this table, in the confined specimens of the series F1R10-2, reaches 1.33 and 1.94, and in the T500R10-3.5 series, it reaches 1.03
as the temperature increases from 20 to 250, 500, and 750 °C, the and 1.06, respectively. Based on these results, adding steel fibers to the
yielding ratio, fyc fcc , changes from 0.63 to 0.74, 0.50, and 0.44, re- concrete mixture and increasing its quantity can significantly improve
spectively. The variation trend of this ratio is generally descending with the residual toughness ratio of the non-heated CFST specimens (see
temperature, showing that the yielding point goes away from the peak Fig. 9c). It is the result of the smaller slope of the descending portion of
point of the load-strain curves, leading to a shorter length of the elastic the load-strain curve in the non-heated CFST specimens with fibers
region and a longer length of the elasto-plastic region (see Fig. 8a). compared to the corresponding fiberless specimens (see Fig. 8c). This
In the confined specimens of the series T20F1-2, as the percentage improvement is much less pronounced in the heated CFST specimens.
of tire particles in the concrete core increased from 0 to 5 and then to According to Table 4 and Fig. 9d, as the steel tube thickness in-
10%, the fyc fcc ratio changed from 0.89 to 0.52 and then to 0.63, re- creased from 2 to 3.5 mm, the residual toughness ratio in the fiberless
spectively. It is observed that in this series of specimens, given the re- specimens under 20, 250, and 500 °C is calculated as 1.98, 2.11, and
duction in the axial stiffness of the CFST specimens due to the addition 3.32, and in the fiber-reinforced specimens as 2.15, 5.04, and 4.04,
of tire particles to the concrete mix, the fyc fcc ratio declined. In the non- respectively. Based on these findings, increasing the thickness of the
heated specimens, as the tire percentage rose from 5 to 10%, the tube has a significant improving effect on the residual toughness ratio,
Poisson’s ratio of the concrete declined (concluded from the axial and which is attributed to the considerable increase in the compressive
lateral stress–strain relationships of the concrete core), which decreased strength as well as a smaller slope of the descending portion of the load-
the hoop stresses relative to the yielding biaxial stresses of the steel tube strain curve. Moreover, according to Fig. 9d, the specimens having 1%
and compensated for a portion of the lost fyc fcc ratio. steel fibers always have a higher residual toughness ratio compared to
In the confined specimens of the series T20R10-3.5, as the volume the corresponding fiberless specimens. This suggests the significant ef-
quantity of fibers increased from 0 to 1 and then to 1.5%, fyc fcc fect of increasing the steel tube thickness and adding steel fibers in the
changed from 0.85 to 0.48 and then to 0.77, respectively. It is observed concrete core at the same time in enhancing the residual toughness
that adding fibers to the confined specimens decreased fyc and fyc fcc . By ratio.
adding steel fibers to the concrete, the Poisson’s ratio increased, which
in turn increased the hoop stress of the steel tube, accelerated yielding, 3.3.4. Ductility index
and finally lowered fyc fcc . One way to quantify the ductility of CFST members is through the
It is also found from the results that increasing the steel tube ductility index which can be defined as Eq. (2) [55–57],
thickness in the specimens without steel fibers increased fyc fcc , con-
85%
trary to the case for most of the specimens with steel fibers. In the µ=
specimens without steel fibers, the highest increase in fyc fcc due to the p (2)
increase in the steel tube thickness occurred in the non-heated speci- where 85% is the strain in the descending portion of the axial load-
mens. In the specimens with fibers, the combined effects of fibers (faster strain at a point corresponding to 85% of the maximum load, and p is
yielding in the composite section) and the steel tube thickness increase the strain at the peak point of the curve associated with the maximum
(increased compressive strength of the composite section and decreased axial load. Table 4 lists the ductility index of all the CFST specimens,
fyc fcc ratio) lowered the fyc fcc ratio. ranging from 1.57 to 6.98. In some of the heated specimens (such as
T750F1R10-2), the loss of axial load in the descending portion of the
3.3.3. Absorbed energy load-strain curve is less than 15% of the maximum axial load. In these
Energy absorption (toughness) of CFST members is an important specimens, the ultimate strain is regarded as the strain at a point cor-
structural parameter that depends on their deformation capacity and responding to 85% of the maximum axial load. The ductility index in
strength. As a result of this property of the structural members, energy the F1R10-2 series under 20, 250, 500, and 750 °C is obtained as 2.02,
is dissipated through the natural damping of the constituent materials 1.86, 1.57, and 1.49, respectively. The highest ductility index occurred
of the element during an earthquake [54]. Energy absorption is defined in the non-heated specimens, and it decreased with temperature (see
as the area under the load-axial shortening curve up to the ultimate Fig. 10a). Although the heated CFST specimens have a high 85% , the

12
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

Fig. 9. Absorbed Energy of CFST specimens: a) Effect of high temperatures, b) Effect of rubber percentage, c) Effect of steel fiber quantity, d) Effect of steel tube
thickness.

Table 4 the ductility index. At 500 °C, as the tire replacement percentage in-
Parameters of CFST specimens. creased from 0 to 5 and then to 10%, the ductility index decreased from
Specimens Peak strain TT TR TF Tt µ 1.75 to 1.71 and then to 1.57, respectively, This suggests the reduction
of the ductility index with increase in the tire replacement percentage
T20F0R10-2 0.0096 – – – 1 1.19 in the heated CFST specimens, similar to the case in the non-heated
T20F1R0-2 0.0081 1 1 – – 2.31 specimens. The lower ductility index at 500 °C compared to that at
T20F1R5-2 0.0091 – 0.82 – – 1.97
20 °C (see Fig. 10b) is attributed to the considerable increase in the
T20F1R10-2 0.0094 – 0.85 – 1 1.80
T20F0R10-3.5 0.0085 – – 1 1.98 1.17 peak strain of the specimens under thermal treatment.
T20F1R10-3.5 0.0089 – – 1.33 2.15 3.54 As the volume quantity of fibers in the CFST specimens exposed to
T20F1.5R10-3.5 0.0163 – – 1.94 – 2.99 20 °C increased from 0 to 1 and then to 1.5%, their ductility index
T250F0R10-2 0.0094 – – – 1 2.75
reached 1.17, 3.54 and 2.99, respectively. These results show that using
T250F1R10-2 0.0109 0.89 – – 1 1.86
T250F0R10-3.5 0.0106 – – – 2.11 6.98
steel fibers with a quantity of 1% in the concrete volume not only in-
T250F1R10-3.5 0.0226 – – – 5.04 3.47 creased the ductility index by almost three times but also was the op-
T500F0R10-2 0.0170 – – – 1 4.94 timum fiber quantity in terms of improving the ductility index of the
T500F1R0-2 0.0162 1.61 1 – – 1.75 CFST specimens. In addition, the reason for the lack of further increase
T500F1R5-2 0.0165 – 1.03 – – 1.71
in the ductility index via adding 1.5% fibers is the significant increase
T500F1R10-2 0.0203 – 1.10 – 1 1.57
T500F0R10-3.5 0.0192 – – 1 3.32 4.42 seen in the peak strain; however, the slope of the descending portion of
T500F1R10-3.5 0.0225 – – 1.03 4.04 3.64 the load-strain curve becomes smaller (flatter descending portion), as
T500F1.5R10-3.5 0.0223 – – 1.06 – 3.81 shown in Fig. 8c. At 500 °C, as the volume quantity of fibers in the CFST
T750F1R10-2 0.0561 3.24 – – – 1.49 specimens increased from 0 to 1 and then to 1.5%, the ductility index
was obtained as 4.42, 3.64, and 3.81, respectively. It can be understood
that in the heated CFST specimens, adding steel fibers in the concrete
significant increase in the peak strain due to heat (see Table 4) leads to
has adverse effects on the ductility index (see Fig. 10c), since after the
the reduction of their ductility index with temperature.
addition of steel fibers in the concrete, despite the increase seen in the
As the tire percentage in the CFST specimens exposed to 20 °C in-
compressive strength and peak strain of the heated CFST specimens, the
creased from 0 to 5 and then to 10%, their ductility index decreased
post-peak slope of their load-strain curve increases (becomes steeper),
from 2.32 to 2.07 and then to 2.02, respectively. The main reason be-
according to Fig. 8c. Furthermore, the ductility index of the heated
hind the reduction in the ductility index with an increase in the tire
CFST specimens containing fibers increased compared to that of the
aggregate replacement percentage is increase in peak strain, as pre-
corresponding non-heated specimens, since the axial load reduction in
viously discussed in Section 3.3.3. On the other hand, the presence of
the post-peak stage of the heated specimens was insignificant.
tire aggregate in concrete lowers the bond among the concrete core
Fig. 10d shows the variation of the ductility index of the CFST
constituents, which consequently leads to faster reduction of the load in
specimens with the increase in the steel tube thickness from 2 to
the descending portion of the load-strain of the concrete core alone and
3.5 mm. In this figure, in the F0R10 specimen series, when the thickness
relatively greater loss of load in the descending portion of the CFST
increases from 2 to 3.5 mm, the ductility index at 20 °C changes from
specimens’ curves (see Fig. 8b). This in turn decreases 85% and finally

13
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

Fig. 10. Ductility index of CFST: a) Effect of elevated temperature, b) Effect of rubber content, c) Effect of steel fiber content, d) Effect of steel tube thickness.

1.96 to 2.78 and at 250 °C from 2.75 to 6.98, respectively, indicating 42 Table 5
and 154% increase, respectively; while this parameter changed from Strength prediction formulas from other researchers.
4.94 to 4.42 at 500 °C with change in the thickness, showing 11% re- Author(s) Formula
duction. In the F1R10 series, by increasing the thickness from 2 to
3.5 mm, the ductility index for 20 °C changed from 2.02 to 3.54, for Lu et al. [15]
Nu = A c fc 1 + 1.55 + 7Vf + 1000 V 2f
250 °C from 1.86 to 3.47, and for 500 °C, from 1.57 to 3.64, demon-
strating 75, 87, and 131% improvement, respectively. It is clear that by Li et al. [21] P (T ) 1
=
increasing the thickness of the steel tube in the specimens including 1% p(T = 20°C ) 1+ 0.54 (T 20)2.6 10 8

fibers (previously obtained as the optimum fiber quantity), the ductility Wan et al. [41] N0T = (1.212 + B +c 2) f A k T
ck sc sc
index always experienced high improvement, contrary to the case for Guo et al. [58] P = (P core + P tube )
the fiberless specimens. Therefore, to improve the ductility of CFST =
0.15 + 0.5 if >1
specimens, in addition to increasing the steel tube thickness, adding 1% 1 if 1

steel fibers seems necessary to achieve an economic design with the

maximum increase in ductility.
useful. Some of these equations are provided in Table 5. Lu et al. [15]
who focused on the effect of steel fibers on the compressive strength of
4. Predicting loading capacity of CFST specimens CFSTs considered parameters Vf and in their proposed formula, which
As fy
represent the volume ratio of fibers and the confinement factor (= ),
Regarding the fact that structures damaged by fire must be retro- Ac fc

fitted, evaluating the post-fire behavior and the residual strength of the respectively. Li and Lu [21] proposed a prediction relationship to de-
members of these structures is necessary for strengthening calculations termine the compressive strength ratio of the heated CFST columns to
and obtaining an optimal design. In all existing design codes, the effect corresponding non-heated ones as a function of temperature and con-
of temperature is not addressed for predicting the loading capacity of finement factor. They ignored the effect of replacement level of re-
composite sections, and this deficiency lowers the accuracy of the cycled aggregate on the proposed relationship. In their research re-
strength predictions of these columns. In addition, research by others in garding the fire resistance of CFST columns, Guo et al. [58] not only
this area is very limited, and the proposed equations deliver low ac- developed a finite element model of the experimental confined speci-
curacy in predicting the compressive strength of heated composite mens exposed to heat but also presented an equation for predicting the
sections due to being independent of temperature. Therefore, devel- loading capacity of these columns under high temperatures in their
oping an accurate equation to predict the strength of CFST columns parametric study, which demonstrated good consistency with the ex-
exposed to high temperature seems necessary. Finding a formula for perimental results. Wan et al. [41] carried out extensive research into
predicting the compressive strength of the CFST specimens investigated the strength of the CFST and CFDST (concrete-filled double skin steel
in this study seems challenging due to different factors affecting this tubular) columns exposed to thermal loading and proposed a new for-
parameter including the volume ratio of steel fibers, tire rubber ag- mula to properly predict the strength of a wide range of heated CFST
gregate content, exposure temperature, and the diameter-to-thickness and CFDST columns.
ratio of the steel tube. Therefore, to obtain a proper equation, in- By reviewing the results and predictions of other studies
vestigating the findings of other researchers who took one or more of [15,21,41,58], different equations provided in Table 5, with variables
these variables in their prediction equations into account would be used in this study were assessed. After conducting a sensitivity analysis

14
 Table 6
Comparison between the predicted and experimental cross-sectional strength.
A. Karimi and M. Nematzadeh

(Present study) Lu et al. [15] Abed et al. [59] Li et al. [21] Xiong et al. [62]

Specimens ID Pexp Pcal Pexp/Pcal Specimens ID Pexp Pcal Pexp/Pcal Specimens ID Pexp Pcal Pexp/Pcal Specimens ID Pexp Pcal Pexp/Pcal Specimens ID Pexp Pcal Pexp/Pcal

T20F0R10-2 552 561 0.98 C50-3-0 1068 1001 1.07 f60D167t3.1 1873 1821 1.03 NACFST(20°c) 620 537 1.16 C1 2677 2677 1.07
T20F1R0-2 700 673 1.04 C50-3-0.6 1093 1015 1.08 f60D114t3.6 1095 1044 1.05 RAC50FST(20 °c) 630 542 1.16 C2 2677 2677 0.97
T20F1R5-2 644 646 1.00 C50-3-0.9 1118 1023 1.09 f60D114t5.6 1365 1347 1.01 RAC100FST(20°c) 565 493 1.15 C3 2196 2196 1.10
T20F1R10-2 625 612 1.02 C50-3-1.2 1162 1044 1.11 f44D167t3.1 1710 1516 1.13 NACFST(200°c) 595 560 1.06 C4 2196 2196 1.07
T20F0R10-3.5 745 773 0.96 C50-4-0 1280 1201 1.07 f44D114t3.6 1042 922 1.13 RAC50FST(200°c) 600 566 1.06 C5 2289 2289 1.09
T20F1R10-3.5 806 819 0.98 C50-4-0.6 1360 1214 1.12 f44D114t5.6 1314 1256 1.05 RAC100FST(200°c) 560 510 1.10 C6 2289 2289 1.01
T20F1.5R10-3.5 778 822 0.95 C50-4-0.9 1403 1222 1.15 Mean 1.06 NACFST(500°c) 495 541 0.92 C7 2677 2677 0.97
T250F0R10-2 566 592 0.96 C50-4-1.2 1459 1241 1.18 COV 0.048 RAC50FST(500°c) 490 546 0.90 C8 2677 2677 0.98
T250F1R10-2 600 605 0.99 C5-5-0 1454 1417 1.03 RAC100FST(500°c) 445 493 0.90 C9 3499 3499 0.89
T250F0R10-3.5 728 776 0.94 C50-5-0.6 1470 1429 1.03 NACFST(700°c) 400 439 0.91 C10 7883 7883 0.99
T250F1R10-3.5 763 787 0.97 C50-5-0.9 1518 1436 1.06 RAC50FST(700°c) 395 441 0.90 C11 8159 8159 1.06
T500F0R10-2 500 504 0.99 C50-5-1.2 1569 1454 1.08 RAC100FST(700°c) 350 422 0.83 C12 5483 5483 0.96
T500F1R0-2 675 693 0.97 C60-3-0 1095 1062 1.03 Mean 1.00 C13 9044 9044 1.00
T500F1R5-2 628 599 1.05 C60-3-0.6 1128 1075 1.05 COV 0.121 C14 9285 9285 0.99
T500F1R10-2 593 598 0.99 C60-3-0.9 1147 1089 1.05 Yang and Hou [61] C15 6995 6995 0.99
T500F0R10-3.5 681 702 0.97 C60-3-1.2 1192 1108 1.08 C16 7401 7401 1.00
T500F1R10-3.5 733 747 0.98 C60-4-0 1320 1258 1.05 Specimens ID Pexp Pcal Pexp/Pcal C17 6537 6537 1.05

15
T500F1.5R10-3.5 758 768 0.99 C60-4-0.6 1389 1270 1.09 C18 7372 7372 1.03
T750F1R10-2 409 427 0.96 C60-4-0.9 1430 1283 1.11 C-NC-20 711 697 1.02 Mean 1.01
Mean 0.99 C60-4-1.2 1452 1301 1.12 C-NC-300 683 794 0.86 COV 0.052
COV 0.028 C60-5-0 1548 1469 1.05 C-NC-600 460 515 0.89
C60-5-0.6 1581 1481 1.07 C-NC-800 388 418 0.93
C60-5-0.9 1609 1493 1.08 C-RCC1-20 704 668 1.05
C60-5-1.2 1661 1509 1.10 C-RCC1-300 631 754 0.84
C70-3-0 1266 1139 1.11 C-RCC1-600 367 405 0.91
C70-3-0.6 1287 1155 1.11 C-RCC1-800 353 377 0.94
C70-3-0.9 1301 1182 1.10 C-RCC2-20 695 649 1.07
C70-3-1.2 1356 1198 1.13 C-RCC2-300 605 654 0.92
C70-4-0 1469 1330 1.10 C-RCC2-600 357 420 0.85
C70-4-0.6 1482 1346 1.10 C-RCC2-800 329 370 0.89
C70-4-0.9 1496 1371 1.09 C-RCF1-20 680 640 1.06
C70-4-1.2 1542 1387 1.11 C-RCF1-300 556 538 1.03
C70-5-0 1659 1537 1.08 C-RCF1-600 331 402 0.82
C70-5-0.6 1673 1551 1.08 C-RCF1-800 306 369 0.83
C70-5-0.9 1712 1575 1.09 C-RCF2-20 621 640 0.97
C70-5-1.2 1774 1590 1.12 C-RCF2-300 494 535 0.92
Mean 1.08 C-RCF2-600 285 340 0.84
COV 0.030 C-RCF2-800 241 271 0.89
Mean 0.93
COV 0.088
Engineering Structures 219 (2020) 110608
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

on the existing variables, it was found that the effect of parameters As fy

=
including the volume ratio of steel fibers and rubber aggregate content Ac fc
on the compressive strength results of the CFST specimens was negli-
gible (as seen in Section 3.3), while the effect of the temperature where T is the temperature experienced by the CFST specimen, and fc is
parameter on the compressive strength results was significant. There- the compressive strength of the corresponding concrete core. Moreover,
fore, in the proposed formula, the temperature parameter with a value is the confinement factor defined as the ratio of the plastic capacity of
ranging from 20 to 750 °C and the confinement factor were considered the empty steel tube to that of the unconfined concrete core; the ob-
as the variables. tained values for this factor are listed in Table 3. Table 6 and Fig. 11
In this research, 19 series of CFST specimens comprising of 12 series give the estimated values for the compressive strength of the CFST
of heated and 7 series of non-heated specimens with different mixture specimens using Eq. (3) together with the corresponding values mea-
designs of concrete core and steel tube thicknesses were tested. Through sured here. Via a comparison, high accuracy of the proposed equation
performing the nonlinear regression analysis on the results obtained for the heated and non-heated specimens is seen, with less than 6%
from these specimens, Eq. (3) was proposed to predict the compressive error and mean ratio of the experimental to computed results of 0.99.
strength as follows, Furthermore, as Table 6 shows, by heating the specimens, no decline is
seen in the prediction accuracy.
1.4
8 0.5 T 20 3 1.5 To evaluate the proposed equation and investigate its accuracy in
Pcal = As fy + Ac fc 1 + +
5 1000 8 (3) predicting the compressive strength of CFST specimens, the

Fig. 11. Comparison between experimental and calculated results of axial compressive bearing capacity.

16
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

compressive strength values were predicted for the specimens tested in 2. To improve the compressive strength and ductility of CFST, adding
five other studies and then compared with their experimental results in steel fibers by 1% in the concrete mix design was more effective
Table 6 and Fig. 11. In selecting these studies, the variables of tem- than adding them by 1.5%.
perature, concrete core compressive strength, steel tube strength, dia- 3. In the non-heated confined specimens, by increasing the tire content
meter and thickness of the steel tube, volume quantity of steel fibers, from 0 to 5 and then to 10%, the compressive strength decreased by
and the concrete type containing natural and waste aggregates were 8 and 11%, respectively. Moreover, the strength loss of the heated
considered. In the work of Lu et al. [15], the compressive performance CFST specimens due to the increase in the tire content was not much
of CFST columns with concrete reinforced with steel fibers was in- different from that of the non-heated specimens.
vestigated. In the names assigned to the specimens, the number after 4. As temperature increased, the elastic part of the axial load-strain
letter C indicates the cubic compressive strength of the concrete core curve of CFST specimens decreased, the elastoplastic part increased,
(50–70 MPa), and the other two numbers represent the thickness of the and the slope of the descending portion in the post-peak stage de-
confining tube (3–5 mm) and volume quantity of fibers used in the creased independent of tire content and fiber volume fraction.
concrete core (0–1.2%). Abed et al. [59] conducted an experimental 5. By increasing the volume quantity of steel fibers in the concrete
and numerical study on the compressive performance of CFSTs. In their mixture, the ductility of the non-heated rubberized CFST specimens
work, the number following letter f in the names given to the specimens improved, contrary to the case for the heated specimens, and the
shows the cylindrical compressive strength of the concrete core (44 and toughness of heated and non-heated rubberized specimens in-
60 MPa), with the numbers following letters D and t representing the creased.
steel tube’s outer diameter (11 and 16 mm) and thickness 6. The toughness of the CFST specimens heated at different tempera-
(3.1–5.6 mm), respectively. Li et al. [21] reported on the mechanical tures except 250 °C increased compared to that of the non-heated
behavior of CFST stub columns with recycled aggregate concrete core specimens (approximately by 3 times at 750 °C). Adding tire ag-
exposed to high temperatures. Since the compressive strength of the gregate to the non-heated specimens reduced toughness and to the
heated unconfined concrete was not reported in their paper, to predict heated specimens slightly increased toughness.
fcT , ACI 216 code [60] was used. In this work, expressions NAC and 7. An accurate equation as a function of temperature was proposed to
RAC seen in the names of the specimens indicate the use of natural and estimate the compressive strength of the CFST specimens, which
recycled aggregate in concrete mixture, respectively; numbers 50 and showed good agreement with the experimental results considering
100 after RAC give the percentage of recycled aggregate replacing the variables including temperature, diameter-to-thickness ratio, con-
coarse aggregate; and also, numbers in parentheses, i.e. 20, 500, and crete strength, and replacement percentage of tire aggregate.
700 °C, are temperatures the specimens experienced. Yang and Hou
[61] examined the compressive performance of the CFST columns CRediT authorship contribution statement
containing recycled aggregate concrete core after thermal treatment. In
this research, due to the absence of a report on the strength of the Amirhossein Karimi: Data curation, Formal analysis, Investigation,
heated unconfined concrete, ACI 216 [60] was also used to predict fcT . Methodology, Project administration, Resources, Visualization, Writing
The variables included the temperature (20, 300, 600, and 800 °C) and - original draft, Writing - review & editing. Mahdi Nematzadeh:
the percentage of recycled aggregate replacing natural fine and coarse Conceptualization, Formal analysis, Funding acquisition, Supervision,
aggregates (0, 50, and 100%). Xiong et al. [62] explored the axial be- Validation, Writing - review & editing.
havior of stub columns made with high- and ultra-high-strength mate-
rials. In their specimens, the yielding stress of steel varied from 300 to Declaration of Competing Interest
428 MPa and the concrete compressive strength from 50 to 175 MPa.
In Table 6 and Fig. 11, the prediction results of 111 different spe- The authors declare that they have no known competing financial
cimens together with their experimental compressive strength are interests or personal relationships that could have appeared to influ-
given. By comparing the experimental and prediction results, it is found ence the work reported in this paper.
that for around 70% of specimens, the prediction error is less than 10%,
and for the rest of specimens, it varies from 10 to 20%. These results References
indicate that the proposed equation was still capable of predicting the
compressive strength of the specimens with high accuracy even in cases [1] Tao Z, Han LH, Wang DY. Strength and ductility of stiffened thin-walled hollow
the effective parameters of the CFST specimens varied greatly by, for steel structural stub columns filled with concrete. Thin-walled Struct
example, increasing the material strength to a great extent (Xiong et al. 2008;46(10):1113–28.
[2] Elchalakani M, Zhao XL. Concrete-filled cold-formed circular steel tubes subjected
[62]), heating the specimens and using recycled aggregate (Yang and to variable amplitude cyclic pure bending. Eng Struct 2008;30(2):287–99.
Hou [61] and Li et al. [21]), simultaneously changing the strength and [3] Bambach MR, Jama H, Zhao XL, Grzebieta RH. Hollow and concrete filled steel
diameter-to-thickness ratio of the steel tube (Abed et a. [59]), and hollow sections under transverse impact loads. Eng Struct 2008;30(10):2859–70.
[4] Li W, Wang T, Han LH. Seismic performance of concrete-filled double-skin steel
adding steel fibers and changing their volume quantity in specimens tubes after exposure to fire: Experiments. J Constr Steel Res 2019;154:209–23.
with different steel tube strength and thickness values (Lu et al. [15]). [5] Uy B. Local and post-local buckling of concrete filled steel welded box columns. J
Constr Steel Res 1998;47(1–2):47–72.
[6] Ellobody E, Young B. Nonlinear analysis of concrete-filled steel SHS and RHS col-
5. Conclusions umns. Thin-walled Struct 2006;44(8):919–30.
[7] Nematzadeh M, Fazli S, Naghipour M, Jalali J. Experimental study on modulus of
In this study, the compressive strength of the CFST stub columns elasticity of steel tube-confined concrete stub columns with active and passive
confinement. Eng Struct 2017;130:142–53.
with high-strength rubberized concrete core reinforced with steel fibers
[8] Uy B, Das S. Wet concrete loading of thin-walled steel box columns during the
was explored after exposure to high temperatures. The main results of construction of a tall building. J Constr Steel Res 1997;42(2):95–119.
this research are listed below. [9] Wee TH, Chin MS, Mansur MA. Stress-strain relationship of high-strength concrete
in compression. J Mater Civ Eng 1996;8(2):70–6.
[10] Ho JCM, Lam JYK, Kwan AKH. Effectiveness of adding confinement for ductility
1. The compressive strength loss of the rubberized concretes confined improvement of high-strength concrete columns. Eng Struct 2010;32(3):714–25.
with steel tube was not noticeable (less than 5%) up to 500 °C; [11] Zhang SM, Wang YY. Failure modes of short columns of high-strength concrete
however, when temperature further increased to 750 °C, the com- filled steel tubes. China Civil Eng J 2004;37(9):1–10.
[12] Lai MH, Chen MT, Ren FM, Ho JCM. Uni-axial behaviour of externally confined
pressive strength experienced a loss of around 30% compared to that UHSCFST columns. Thin-Walled Struct 2019;142:19–36.
at 500 °C. [13] Lai MH, Liang YW, Wang Q, Ren FM, Chen MT, Ho JCM. A stress-path dependent

17
A. Karimi and M. Nematzadeh Engineering Structures 219 (2020) 110608

stress-strain model for FRP-confined concrete. Eng Struct 2020;203:109824. [39] Liu JQ, Han LH, Zhao XL. Performance of concrete-filled steel tubular column-wall
[14] Wang W, Ma H, Li Z, Tang Z. Size effect in circular concrete-filled steel tubes with structure subjected to ISO-834 standard fire: analytical behaviour. Thin-Walled
different diameter-to-thickness ratios under axial compression. Eng Struct Struct 2018;129:28–44.
2017;151:554–67. [40] Kodur VKR, Lie TT. Evaluation of fire resistance of rectangular steel columns filled
[15] Lu Y, Li N, Li S, Liang H. Behavior of steel fiber reinforced concrete-filled steel tube with fibre-reinforced concrete. Can J Civ Eng 1997;24(3):339–49.
columns under axial compression. Constr Build Mater 2015;95:74–85. [41] Wan CY, Zha XX, Dassekpo JBM. Analysis of axially loaded concrete filled circular
[16] Wang HT, Wang LC. Experimental study on static and dynamic mechanical prop- hollow double steel tubular columns exposed to fire. Fire Saf J 2017;88:1–12.
erties of steel fiber reinforced lightweight aggregate concrete. Constr Build Mater [42] Mohamedbhai GTG. Effect of exposure time and rates of heating and cooling on
2013;38:1146–51. residual strength of heated concrete. Mag Concr Res 1986;38(136):151–8.
[17] Liew JR, Xiong DX. Ultra-high strength concrete filled composite columns for multi- [43] Peng GF, Yang WW, Zhao J, Liu YF, Bian SH, Zhao LH. Explosive spalling and
storey building construction. Adv Struct Eng 2012;15(9):1487–503. residual mechanical properties of fiber-toughened high-performance concrete sub-
[18] Portoles JM, Serra E, Romero ML. Influence of ultra-high strength infill in slender jected to high temperatures. Cem Concr Res 2006;36(4):723–7.
concrete-filled steel tubular columns. J Constr Steel Res 2013;86:107–14. [44] Hertz KD. Concrete strength for fire safety design. Mag Concr Res
[19] Wang K, Young B. Fire resistance of concrete-filled high strength steel tubular 2005;57(8):445–53.
columns. Thin-Walled Struct 2013;71:46–56. [45] Han LH, Huo JS, Wang YC. Compressive and flexural behaviour of concrete filled
[20] Lu H, Zhao XL, Han LH. Fire behaviour of high strength self-consolidating concrete steel tubes after exposure to standard fire. J Constr Steel Res 2005;61(7):882–901.
filled steel tubular stub columns. J Constr Steel Res 2009;65(10–11):1995–2010. [46] Cree D, Green M, Noumowé A. Residual strength of concrete containing recycled
[21] Li W, Luo Z, Tao Z, Duan WH, Shah SP. Mechanical behavior of recycled aggregate materials after exposure to fire: a review. Constr Build Mater 2013;45:208–23.
concrete-filled steel tube stub columns after exposure to elevated temperatures. [47] Han LH, Tao Z, Liu W. Concrete filled steel tubular structures from theory to
Constr Build Mater 2017;146:571–81. practice. J Fuzhou University (Natural Sci Ed) 2001:6.
[22] Elchalakani M, Hassanein MF, Karrech A, Fawzia S, Yang B, Patel VI. Experimental [48] Yao Y, Hu XX. Cooling behavior and residual strength of post-fire concrete filled
tests and design of rubberised concrete-filled double skin circular tubular short steel tubular columns. J Constr Steel Res 2015;112:282–92.
columns. Structures 2018;15:196–210. [49] Lau A, Anson M. Effect of high temperatures on high performance steel fibre re-
[23] Aiello MA, Leuzzi F. Waste tyre rubberized concrete: Properties at fresh and har- inforced concrete. Cem Concr Res 2006;36(9):1698–707.
dened state. Waste Manage 2010;30(8–9):1696–704. [50] Tao Z, Han LH, Wang DY. Experimental behaviour of concrete-filled stiffened thin-
[24] Toutanji HA. The use of rubber tire particles in concrete to replace mineral ag- walled steel tubular columns. Thin-Walled Struct 2007;45(5):517–27.
gregates. Cem Concr Compos 1996;18(2):135–9. [51] Nguyen TH, Toumi A, Turatsinze A. Mechanical properties of steel fibre reinforced
[25] Abendeh R, Ahmad HS, Hunaiti YM. Experimental studies on the behavior of and rubberised cement-based mortars. Mater Des 2010;31(1):641–7.
concrete-filled steel tubes incorporating crumb rubber. J Constr Steel Res [52] Behnood A, Ziari H. Effects of silica fume addition and water to cement ratio on the
2016;122:251–60. properties of high-strength concrete after exposure to high temperatures. Cem
[26] Duarte APC, Silva BA, Silvestre N, De Brito J, Júlio E, Castro JM. Tests and design of Concr Compos 2008;30(2):106–12.
short steel tubes filled with rubberised concrete. Eng Struct 2016;112:274–86. [53] Nematzadeh M, Baradaran-Nasiri A. Residual properties of concrete containing
[27] Silva, A., Jiang, Y., Castro, J. M., & Monteiro, R. (2017). 08.38: Experimental recycled refractory brick aggregate at elevated temperatures. J Mater Civ Eng
characterisation of the flexural behaviour of rubberized concrete‐filled steel tubular 2017;30(1):04017255.
members. ce/papers, 1(2-3), 2147-2156. [54] Ghadami J, Nematzadeh M. Effect of bond shear stress on compressive behaviour of
[28] ACI 211.1-91 Standard Practice for Selecting Proportions for Normal, Heavyweight, steel tube-confined concrete with active and passive confinement. European J
and Mass Concrete—Procedure for Mix Design. Environ Civil Eng 2018;22(7):783–810.
[29] ASTM C127 – 15 Standard Test Method for Relative Density (Specific Gravity) and [55] Tao Z, Han LH, Wang ZB. Experimental behaviour of stiffened concrete-filled thin-
Absorption of Coarse Aggregate. walled hollow steel structural (HSS) stub columns. J Constr Steel Res
[30] ASTM C128 - 15 Standard Test Method for Relative Density (Specific Gravity) and 2005;61(7):962–83.
Absorption of Fine Aggregate. [56] Yang YF, Han LH. Experiments on rectangular concrete-filled steel tubes loaded
[31] ASTM C33-18 Standard Specification for Concrete Aggregate. axially on a partially stressed cross-sectional area. J Constr Steel Res
[32] Lai MH, Binhowimal SAM, Hanzic L, Wang Q, Ho JCM. Cause and mitigation of 2009;65(8–9):1617–30.
dilatancy in cement powder paste. Constr Build Mater 2020;236:117595. [57] Liao FY, Han LH, Tao Z. Seismic behaviour of circular CFST columns and RC shear
[33] Lai M, Hanzic L, Ho JC. Fillers to improve passing ability of concrete. Struct wall mixed structures: experiments. J Constr Steel Res 2009;65(8–9):1582–96.
Concrete 2019;20(1):185–97. [58] Guo H, Long X, Yao Y. Fire resistance of concrete filled steel tube columns subjected
[34] ASTM C 136-01, Standard Test Method for Sieve Analysis of Fine and Coarse to non-uniform heating. J Constr Steel Res 2017;128:542–54.
Aggregates. [59] Abed F, AlHamaydeh M, Abdalla S. Experimental and numerical investigations of
[35] ASTM E8/E8M – 16a Standars Test methods for Tension Testing of Metalic the compressive behavior of concrete filled steel tubes (CFSTs). J Constr Steel Res
Materials. 2013;80:429–39.
[36] ASTM A820/A820m -16 Standard Specification for Steel Fibers for Fiber-Reinforced [60] ACI 216.1-07 Code Requirements for Determining Fire Resistance of Concrete and
Concrete. Masonry Construction Assemblies.
[37] ISO-834, Fire resistance tests-Elements of building construction-Part 1: general re- [61] Yang YF, Hou R. Experimental behaviour of RACFST stub columns after exposed to
quirements, ISO-834-1, Intemational Standard Organization, Switzerland, 1999. high temperatures. Thin-Walled Struct 2012;59:1–10.
[38] Zhou K, Han LH. Experimental performance of concrete-encased CFST columns [62] Xiong MX, Xiong DX, Liew JR. Axial performance of short concrete filled steel tubes
subjected to full-range fire including heating and cooling. Eng Struct with high-and ultra-high-strength materials. Eng Struct 2017;136:494–510.
2018;165:331–48.

18