Thin-Walled Structures 97 (2015) 22–34

Contents lists available at ScienceDirect

Thin-Walled Structures
journal homepage: www.elsevier.com/locate/tws

Mechanical performances of concrete-filled steel tubular stub columns

with round ends under axial loading
Ding Faxing a, Fu Lei a, Yu Zhiwu a,b,n, Li Gang a
a
School of Civil Engineering, Central South University, Changsha, Hunan Province 410075, PR China
b
National Engineering Laboratory for High Speed Railway Construction, Changsha, Hunan Province 410075, PR China

art ic l e i nf o a b s t r a c t

Article history: An experimental study of 22 concrete-filled round-ended steel tubular (CFRT) stub columns under axial
Received 17 March 2015 compression is conducted compared with 4 circular concrete-filled steel tubular (CFT) stub columns. The
Received in revised form influences of width–thickness ratio, concrete strength, steel yield strength and wall-thickness of steel
16 July 2015
tube on the ultimate bearing capacity of the CFRT columns are discussed. The 3D finite element (FE)
Accepted 25 July 2015
model is also developed to analyze the behavior of the CFRT columns under axial compression. From the
Available online 19 September 2015
results, local buckling of the round-ended steel tube associated with shear failure of in-filled concrete
Keywords: could be observed. With the increasing width–thickness ratio, the corresponding load–strain curves have
Concrete-filled round-ended steel tubular a shorter elastic–plastic stage. The parametric studies indicate that the concrete strength, tube thickness
stub column
and width–thickness ratio of the steel tube also have a great effect on the ultimate bearing capacity. The
Ultimate bearing capacity
numerical results also show that the confinement effect of the stub columns decreases with the in-
Width–thickness ratio
Confinement coefficient creasing width–thickness ratio. A practical calculation formula for the bearing capacity of the CFRT stub
columns is proposed, which is well in agreement with the experimental results.
& 2015 Published by Elsevier Ltd.

1. Introduction thickness of the column and t is the wall-thickness of steel tube.

Concrete-filled steel tube (CFT) columns have been increasingly
used in the construction of bridges, high-rise building, transmis-
sion tower and warehouses etc. It is well known that CFT columns
have been frequently used in numerous engineering structures
owing to their excellent structural behavior in terms of high
strength, high ductility, high stiffness and full usage of construc-
tion materials. In the past, many studies on CFT columns have been
conducted, such as Schneide [1], Hassanein et al. [2–4], Chang
et al. [5–8], Ellobody [9], Yu et al. [10], Thayalan et al. [11], Mo-
hamed [12], Tu et al. [13], Uenaka [14], Ren et al. [15], Ding et al.
[16], Han [17], Tao et al. [18], Lam and Williams [19], Sakino et al.
[20] etc.
With the development of heavy haul railway and highway
bridges, requirements for the bearing capacity and ductility of the
pier become higher. In order to meet the requirement of capacity
and ductility, the concrete-filled round-ended steel tubular (CFRT)
bridge pier column is proposed and has been applied in actual
engineering [21,22]. Fig. 1 shows the cross-section of the CFRT
column, where B is the sectional width, D is the sectional
n
Corresponding author at: School of Civil Engineering, Central South University,
Changsha, Hunan Province 410075, PR China. Fax: þ 86 731 2655536.
E-mail address: zhwyu@csu.edu.cn (Y. Zhiwu).
According to [23], the CFRT column has not only high bearing
capacity, plastic toughness, convenient construction, and low cost
advantages but also has the following advantages:
(1) The CFRT column has a good external model, and can meet
the requirement of architectural esthetics.
(2) The CFRT column has strong axis and the weak axis, and the
strong axis has strong stability.
(3) The circular arc section of the CFRT column is more smooth
than any other section of the form of concrete-filled steel tube,
which can effectively resist running water of other liquid impact.
For example in the river bridge across the river, it can reduce the
impact of water load and prolong the service life of the pier.
Thanks to the excellent features of CFRT columns, therefore,
CFRT columns have a good application prospect in bridge struc-
ture. In recent years, there is also some research reported on CFRT
stub columns. Wang et al. [23] reported that the experimental
research of concrete-filled round-end steel tubular stub columns
under axial compression was carried out with different sectional
aspect ratio, steel ratios and the structural measures. It can be
found from experiments that core concrete can be confined by the
steel tube. Xie et al. [24] studied the behavior of round-ended
concrete-filled steel tube (CFST) coupled column. Han [25] in-
vestigated the behavior of round-end stainless steel-concrete-
carbon steel double-skin tubular stub columns; the test results
showed this type of hybrid column with high bearing capacity and

http://dx.doi.org/10.1016/j.tws.2015.07.021
0263-8231/& 2015 Published by Elsevier Ltd.
 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34 23

Nomenclature Nu,0 ultimate bearing capacity of CFRT stub columns from

experimental results
Ac cross-sectional area of core concrete t wall thickness of steel tube
As cross-sectional area of steel tube s axial stress
Ac1 the constrained concrete area sL,c axial stress of concrete
As1 the constrained steel tube area sL,c1 the axial compressive stress of constrained concrete
Ac2 the unconstrained concrete area sL,c2 the axial compressive stress of unconstrained concrete
As2 the unconstrained steel tube area sr,c1 radial constrained concrete stress caused by steel tube
B sectional width of CFRT stub column sL,s axial stress of steel tube
D sectional thickness of CFRT stub column sL,s1 the round-ended steel tube's axial compressive stress
DI ductility index sL,s2 the rectangular steel tube's axial compressive stress
Ec elastic modulus of concrete sr,c2 unconstrained concrete stress
Es elastic modulus of steel sθ,st tensile transverse stress of steel tube
fc uniaxial compressive strength of concrete sθ,s1 constrained transverse stress of the steel tube
fcu cubic compressive strength of concrete sθ,s2 unconstrained transverse stress of the steel tube
fs yielding strength of steel εL axial strain
K coefficient of confinement effect εθ,s tensile transverse strain of steel tube
L height of specimen εI equivalent strain of steel
N axial load εy yielding strain of steel
Nu axial ultimate bearing capacity εst strengthening strain of steel
Nu,1 ultimate bearing capacity of CFRT stub columns from εu ultimate strain of steel
FE results Ф confinement index
Nu,2 ultimate bearing capacity of CFRT stub columns from ρ steel ratio
Eq. (11)

in detail. Especially, when width–thickness ratio (B/D) equals to 1,

the hybrid column is reduced to the circular concrete-filled steel
tubular (CFT) stub column. Further, 4 circular CFT ( B /D = 1) stub
columns are also examined for comparison. As an attempt to give a
further insight on the performances of the proposed columns, a
numerical analysis based on the ABAQUS/standard is also carried
out. A practical calculation formula for the bearing capacity of
CFRT stub columns is also proposed.

Fig. 1. Section of the CFRT stub column.
good ductility. However, more attention has been paid to the
smaller width–thickness ratio(1 o B/Do3), the round-end stain-
less steel-concrete-carbon steel double-skin tubular stub column
and round-ended CFST coupled column. In addition, Wang et al.
[23] then calculated the bearing capacity of the CFRT stub column
based on a practical calculation formula for the bearing capacity of
the concrete-filled rectangular steel tube columns. However, any
formula concerning the capacity of CFRT stub column is still in
absence.
In this paper, in order to comprehensively study the mechanical
behavior of CFRT stub columns and the confining effect of width–
thickness ratios on core concrete. Compressive tests are conducted
on 22 CFRT stub columns, with various width–thickness ratios
(from 1 to 4), concrete strengths, the steel thicknesses and steel
yield strengths. The effects of such parameters on the load–strain
cures, load-bearing capacities and column ductility are discussed
2. Experimental program
2.1. Specimens and materials
Three types of steel are adopted in the test. For each type of
steel, three tension coupons are cut with dimensions in ac-
cordance with the Chinese Standard GB2975 [26] to determine the
steel material properties. The average yield strengths are found to
be 327.7 MPa, 311 MPa and 299.5 MPa. The cube strength and
elastic modulus of concrete are measured in accordance with the
Chinese standard GBJ81-85 [27]. Four types of concretes are used
and their cube strengths are 39.3 MPa, 40.4 MPa, 50.4 MPa and
57.4 MPa. The fine aggregate used is silica based sand; the coarse
aggregate is carbonate stone and the mix ratio is summarized in
Table 1.
2.2. Test specimens
A total of 26 specimens are designed for this study and the
details are shown in Table 2, where t is the wall thickness of steel

Table 1
Concrete mix proportions.

Concrete strength level Mix ratio Water (kg/m3) Cement (kg/m3) Sand (kg/m3) Aggregates (kg/m3) water reducer (kg/m3)

C30 0.38 175 420 512 1252 4.10

C40 0.34 168 432 320 1201 4.32
C50 0.38 164 461 686 1118 5.11
24 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34

Table 2
Specimen details, test and calculated results.

Specimen ID B × D × t (mm) L (mm) fcu (MPa) fs (MPa) Nu,0 (kN) DI

CFT1-A 251  251  3.70 750 40.4 327.7 3023 5.96

CFT1-B 250  250  3.69 3265
5.81
CFT2-A 251  251  5.51 299.5 3556 6.53
CFT2-B 252  252  5.74 3661
6.79
WST1-A 299  252  3.75 327.7 3429 5.46
WST1-B 302  249  3.75 3338 5.44
WST2-A 299  255  5.84 299.5 4162 5.64
WST2-B 300  251  5.80 4168 5.58
WST3-A 350  255  3.72 900 327.7 3929 5.28
WST3-B 351  252  3.76 4158 5.32
WST4-A 352  251  5.90 299.5 4492 5.39
WST4-B 349  251  5.92 50.4 5530 5.29
WST5-A 394  260  3.79 1000 327.7 5620 5.21
WST5-B 396  264  3.80 5500 5.18
WST6-A 405  197  3.75 800 39.3 311 3240 4.95
WST6-B 402  196  3.72 2993 4.86
WST7-A 608  186  3.75 1200 4826 2.67
WST7-B 605  194  3.77 4944 2.76
WST8-A 805  190  3.74 1600 6521 2.61
WST8-B 806  191  3.67 6493 2.36
WST9-A 405  198  3.75 800 57.4 4203 3.21
WST9-B 405  198  3.70 4180 3.29
WST10-A 610  196  3.80 1200 7201 2.32
WST10-B 606  189  3.77 6905 2.21
WST11-A 805  190  3.68 1600 9065 1.56
WST11-B 805  194  3.80 8799 1.78

Note: Nu,0 is the experimental ultimate bearing capacity, DI is the ductility index.

tube, L is the height of the specimens, fcu is the cube strength of a b

concrete, and fs is the yield strength of steel. For a better ob-
servation of deformation of the specimens, red paint is sprayed on
the outer surface of the steel tubes and grids of 50 mm  50 mm
are drawn on the painted surface. Before the placement of con-
crete, the cover plate is welded to one end of the steel tube and
then the steel tube is erected. Concrete is placed from the top of
the specimens and carefully vibrated using a vibrator to evenly
distribute the concrete inside the steel tube. Finally the top and
bottom surfaces are smoothened to form the CFRT column cross-
sections. In company with the specimen preparation, concrete
standard cubes and prism specimens are made and cured at the
same condition as the CFRT columns. After one month of curing,
concrete surface is rubbed smooth with concrete grinding ma-
chine and then a small amount of concrete mixed with epoxy resin
binder is used to fill the gap to ensure that both surfaces are at the
same level. And then the cover plate is bonded to the surface of
the column. In this way, it can be guaranteed that the steel tube
and core concrete shared the stress during the loading process.

2.3. Experimental setup and instrumentation Fig. 2. (a) Sketch for test setup and (b) test setup for all specimens.

For each specimen, six strain rosettes are installed at the increased at a step of 1/10 of the ultimate load in the elastic stage
middle height of two opposite side surfaces (i.e. A2 and A3 in Fig. 1) and at a step of 1/20 of the ultimate load in the elastic–plastic stage.
and two LVDTs are installed at the same positions of the other two Each loading step takes 3 5 min which is similar to a slow con-
opposite side surfaces (i.e. A1 from Fig. 1(a)) to accurately measure tinuous loading mode and data is acquired by different classifica-
the deformation. The test setup is shown in Fig. 2(b). Specimens
tions. When the ultimate load is approached, specimens are loaded
are placed within the loading frame and sat directly on the strong
slowly and continuously until final failure and data is recorded
floor. The load–strain curves are recorded by a DH3818 static strain
continuously. The loading of each specimen lasts about 1.5 h.
measurement system and load–deformation curves are recorded
by the electronic transducers and data acquisition system. More
details of boundary conditions can be found in Yu et al. [9].
All specimens are tested under a universal testing machine with
a capacity of 20,000 kN. The compressive load is applied from the
top of the specimen through a load control mode. The load is
 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34 25

a b

c d

Fig. 3. Typical failure modes for test specimens: (a) B /D = 1; (b) B /D = 2; (c) B /D = 3; and (d) B /D = 4 .

3. Test results and discussion

3.1. Failure modes

Fig. 3(a)–(d) shows the typical failure modes of the specimens,

in which B /D ¼1, 2, 3 and 4, respectively. As can be seen from these
figures, local buckling of the round-ended steel tubes in the bigger
width-thickness ratio direction is associated with shear failure of
in-filled concrete, which appears from the upper to the lower parts
of the specimens. The observed failure mode is identical to the
results obtained for ordinary CFT stub column test under centric
compressive loading.

3.2. Load–deformation responses

The typical load against the corresponding axial strain curves

are shown in Fig. 4. The curves can be generally characterized by
three processes: elastic stage, elastic–plastic stage and post peak Fig. 4. Typical load–strain response for B /D = 1, 2, 3, 4 .
stage. As indicated in Fig. 4, the linear relationship is up to about
60–70% of the failure load. The response then becomes non-linear that of specimens with smaller ratios. The specimen stiffness is
and the stiffness of the specimen deviates from its initial value.
also higher and the bearing capacity decreases rapidly after ulti-
The comparison of the load–axial strain curves indicates that big
mate one. This is because higher value of B /D can significantly
difference can be observed for the width-thickness ratio ranging
weaken the confining effect of the steel tube. Since the mechanical
from 1 to 4. As shown in Fig. 4, the specimens with bigger B /D
property of the concrete core is closely related to such a confining
ratios seem to experience a shorter elastic–plastic stage than the
specimens with smaller ratios. The corresponding strain at peak effect, and hence a bigger value of B /D can lead to a shorter
load for the specimens with bigger value of B /D is also lower than elastic–plastic stage.
26 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34

b c

Fig. 5. Comparison of stress–axial strain curves for different thicknesses of steel tube, concrete strength and width–thickness ratio (a) The different thickness of steel tube;
(b) The different strength of concrete and (c) The different width-thickness ratio.

3.3. Parametric study
3.3.1. Steel tube wall-thickness
For CFT columns, the wall-thickness of steel tube is an im-
portant parameter which has been extensively investigated in the
previous research work. In this paper, the tube wall thickness on
the mechanical performances of the concrete-filled steel tubular
stub columns with round ends is also discussed. The wall-thick-
ness of steel tube is the variable and other parameters are kept the
same as mentioned above. Fig. 5(a) shows the effect of wall-
thickness on the load–strain responses, where the wall-thickness
of steel tube for specimens WST1-A, WST1-B, WST2-A, and WST2-
B is 3.75, 3.75, 5.84, and 5.80 mm, respectively. It is indicated that
mechanical performances of a CFRT stub column increase by in-
creasing the steel tube wall-thickness.
3.3.2. Concrete strength
Fig. 5(b) shows the effect of concrete strength on load–strain
response, where all parameters for the specimens are the same
and the concrete strengths for specimen WST6 and specimen
WST9 are 39.3 MPa and 57.4 MPa, respectively. The column
strength in the test increases from 2985 kN (WST6) to 4203 kN
(WST9). It can be seen that a higher concrete strength leads to a
higher bearing capacity, which is similar to results of other spe-
cimens (WST3 and WST4, WST7 and WST10, WST8 and WST11).
3.3.3. Width–thickness ratio
The influence of width–thickness ratio on load–strain response
is shown in Fig. 5(c). The load–axial strain curves with B /D = 2, 3
and 4 are compared and it can be found that a bigger B /D ratio
leads to a shorter elastic–plastic stage but a higher peak axial load.
However, the peak load for specimen with bigger B /D ratio de-
clines quickly after the peak load. It is indicated that the CFRT
specimens can significantly weaken the confining effect of the
steel tube on concrete core with a bigger B /D .
3.3.4. Ductility
To investigate the effect of various parameters on the ductility
of the CFRT stub columns, a ductility index (DI) proposed by Han
et al. [17] is also adopted in this study, which is defined as follows:
ε0.85
DI =
εb (1)
where ε0.85 is the axial strain when the load falls to 85% of the
ultimate load, and εb is equal to ε0.75/0.75, where ε0.75 is the axial
strain when the load attains 75% of the ultimate load in the pre-
peak stage.
Fig. 6 shows the ductility indexes from Eq. (1) for the test
specimens, where a larger value of DI indicates a slower process of
load reduction after the peak load. Fig. 6(a) shows the effect of
tube wall-thickness on the ductility index. It is indicated that a
thicker steel tube leads to a better ductility. Fig. 6(b) shows the
influences of B /D . It can be seen that a bigger B /D can lead to a
weaker ductility. The influences of concrete strength on the duc-
tility index are shown in Fig. 6(c). It can be found that a higher
concrete strength leads to a weaker ductility.
 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34 27

b c

Fig. 6. Comparison of ductility index DI for different thicknesses of steel tube, width–thickness ratio and concrete strength (a) The different thickness of steel tube; (b) The
different width-thickness ratio and (c) The different strength of concrete.

Table 3
Comparisons of FE results, predicted results based on Eq. (11) and experimental
results of ultimate load bearing capacity of columns.

Specimen Experimental va- Finite ele- Eq. (11) Nu,0/Nu,1 Nu,0/Nu,2

number lues (kN) ment (kN) (kN)

CFT1-A 3023 2962 3011 1.021 1.004

CFT1-B 3265 2933 2977 1.113 1.097
CFT2-A 3556 3634 3520 0.979 1.010
CFT2-B 3661 3623 3621 1.010 1.011
WST1-A 3429 3392 3436 1.011 0.998
WST1-B 3338 3377 3419 0.988 0.976
WST2-A 4162 4107 4101 1.013 1.015
WST2-B 4168 4011 4046 1.039 1.030
WST3-A 3929 3832 3881 1.025 1.012
WST3-B 4158 3830 3879 1.085 1.072
WST4-A 4492 4492 4548 1.000 0.988
WST4-B 5530 4953 5127 1.116 1.079
Fig. 7. FE models (a) FE model with a half of height; (b) loading plate element; (c) WST5-A 5620 5484 5107 1.025 1.100
steel tube element and (d) concrete element. WST5-B 5500 5557 5196 0.990 1.059
WST6-A 3240 3215 3247 0.998 1.008
WST6-B 2993 3205 3237 0.922 0.931
WST7-A 4826 4915 4776 1.010 0.982
4. Finite element (FE) modeling WST7-B 4944 5059 4913 1.006 0.977
WST8-A 6321 6534 6282 1.038 0.998
WST8-B 6293 6527 6275 1.034 0.995
4.1. FE models
WST9-A 4203 4465 4423 0.950 0.941
WST9-B 4180 4461 4405 0.949 0.937
A nonlinear finite element (FE) analysis model for the simula- WST10-A 7101 7021 6758 1.066 1.026
tion of CFT stub columns subjected to axial compression was de- WST10-B 6905 7063 6513 1.030 0.978
WST11-A 9065 8696 8853 1.059 1.042
veloped based on the works presented by Tao et al. [17,28], Chang
WST11-B 8799 8683 8761 1.004 1.013
et al. [5,6] and Ding et al. [29]. In this paper, according to Ding et al.
[29] 4-node reduced integral format shell elements (S4R) were shell elements to ensure the calculation accuracy. 8-node reduced
adopted for the round-ended steel tubes of all the specimens and integral format 3D solid elements(C3D8R) were used to model the
9-node Simpson integration was adopted along the thickness of core concrete and loading plate for all the specimens. The contact
28 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34

Fig. 8. Comparisons of stress–axial strain curves between calculated and experimental results for circular CFT and CFRT stub column specimens.
 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34 29

Fig. 8. (continued)

models and boundary conditions were described by Ding et al. axial stress–strain relationship. For a concrete-filled steel tubular
[29]. The resulting FE model is shown in Fig. 7. stub column, parameter α1 can be taken as 0.15. The Poisson ratio
According to Ding et al. [29], a tie constraint may couple two vc of concrete is assumed to be 0.2. Details of the values of these
separate surfaces together so that no relative motion occurs be- parameters were presented in [29].
tween them. Therefore, tie option is adopted for the constraint An elasto-plastic model, considering Von-Mises yielding cri-
between concrete and loading plate. teria, Prandtl–Reuss flow rule and isotropic strain hardening, is
The following stress–strain relationship of concrete under used to describe the constitutive behavior of steel. The expression
uniaxial compression proposed in [29] is adopted in this paper: for the stress–strain relationship of steel is described as follows
[29]:
⎧ qx + (m − 1) x2
⎪ x≤1 ⎧ Es εi
⎪ 1 + (q − 2) x + mx2 εi ≤ εy
y=⎨ ⎪
⎪ x ⎪ fs εy < εi < εst
⎪ x>1 σi = ⎨
⎩ α1 (x − 1)2 + x (2) ⎪ fs + ζEs (εi − εst ) εst < εi ≤ εu
⎪
⎩ fu εi > εu (3)
where q is the ratio of the initial tangent modulus to the secant
modulus at peak stress. m is a parameter that controls the de- where si is the equivalent stress of steel; fs is the yield strength; fu
crease in the elastic modulus along the ascending branch of the is the ultimate strength and fu = 1.5 fs; Es is the elastic modulus,
30 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34
Fig. 9. The average stress–axial strain curves of concrete.
Es ¼2.06  105 MPa; Est is the strengthening modulus, which is
described by Est ¼ ζEs; εi is the equivalent strain; εy is the yield
strain; εst is the strengthening strain; and εu is the ultimate strain,
which is described by εu ¼ εst þ0.5fs/(ζEs), where εst ¼12εy,
εu ¼120εy and ζ ¼ 1/216.
Because of the symmetry of cross sections, only half of a section
is analyzed for the axially-loaded CFRT stub columns. In order to
model the decrease of load-bearing capacity of specimens, load is
applied through the increments of displacement and both material
and structural nonlinearities are considered and solved using the
incremental-interactive method in ABAQUS.
a b
c d
Fig. 10. Comparisons of stress–axial strain curves of steel tube at the different points.
4.2. Validation
The comparison of experimental and modeling ultimate bear-
ing capacity is shown in Table 3. A good agreement is found in
general and the discrepancies between experimental and pre-
dicted ultimate load bearing capacities are less than 9.1% for all the
specimens. In addition, the comparisons of the experimental and
the predicated load–axial strain curves are shown in Fig. 8, which
also shows a good agreement.
4.3. Confining effect analysis
In order to give a clear understanding of the mechanical per-
formances of the axially loaded CFRT stub columns, the confining
effect of steel tube on the concrete core is conducted in this sec-
tion. Load–strain curves of axially-loaded circular CFT stub col-
umns and CFRT stub columns with different width–thickness ra-
tios are discussed to study the confining effect from steel tube to
concrete. Four typical values for B /D ratios, 1, 2, 3, and 4, are
adopted and the corresponding axial stress–strain curves for the
concrete are compared in Fig. 9. It is obvious that a higher B /D
ratio can lead to a lower confining effect.
The axial and transverse stress–strain curves for the steel tube
at these critical points (A1, A2 and A3 as shown in Fig. 1) are shown
in Fig. 10. According to Ding et al. [29], the stress–strain curve in
the steel tube can reflect the confining effect. If the axial stress–
strain curve for the steel tube crosses the transverse one the steel
tube can provide a high confining effect on the concrete. Therefore,
there exists obviously confining effects for round section (Fig. 10
(a)). Fig. 10(b) and (c) shows the axial stress–strain curves and
transverse ones at points A1 and A2 for B /D = 2, 3, 4 . It is indicated
 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34
Fig. 11. Typical stress contours at mid-section for CFRT stub columns (B/D ¼ 2,3,4). (For interpretation of the references to color in this figure, the reader is referred to the web
version of this article.)
Fig. 12. Simplified stress distribution within mid-section for CFRT stub column.
Table 4
The relationship of coefficient with B/D.
B /D σL, s1 σθ, s1 As1/As As2 /As
2 0.79fs 0.36fs π/(πþ 2) 2/(π þ 2)
3 0.85fs 0.28fs π/(π þ4) 4/(π þ 4)
4 0.86fs 0.25fs π/(πþ 6) 6/(π þ 6)
that for each case the axial stress–strain curve separates, which
means that the steel tubes of CFRT specimens provide a weaker
confining effect on concrete core at points A1 and A2. For A3 point
in Fig. 10(d), the axial stress–strain curves and transverse ones
cross each other only for B /D = 2. This indicates that a bigger value
of width–thickness ratio can lead to weaker confining effect of the
steel tube on concrete core.
5. Design approach
5.1. Model simplification
According to Ding et al. [29], the similar derived way can be
adopted in this section. In ultimate state, the FE modeling result of
the typical stress distribution at mid-section for the CFRT stub
column is shown in Fig. 11. It is indicated that the mid-section area
of CFRT column around semicircular and straight edge is respec-
tively the strong (the green part) and weak (the blue part) con-
finement region of the core concrete. Therefore, for the different
width–thickness ratios of CFRT column, the core concrete area can
be divided into constrained and unconstrained area. According to
the stress contour in Fig. 11, the stress distribution of the CFRT stub
column can be simplified as Fig. 12. This is completely based on
stress distribution and the superposition principle of the concrete
section at the ultimate state, where Ac1 is the constrained concrete
area, Ac2 is the unconstrained concrete area, As1 is the constrained
steel tube area, As2 is the unconstrained steel tube area, B is the
length of the section, D is the width of the section, and t is the
31
thickness of the steel tube. Based on the extensive numerical
works, the relationships between As1 and As2 with different B /D
ratios are proposed, as described in Table 4. In this paper, ac-
cording to the stress distribution indicated in Fig. 12, the following
relationships can be obtained:
⎧ As1 + As2 = As
⎪ π
⎪ As1 = As
⎪ π + 2 (B/D − 1)
⎨
⎪ A = 2 (B/D − 1) A
⎪ s2 π + 2 (B/D − 1)
s
⎪
⎩ A c1 + A c2 = A c (4)
2
where Ac1 = π (D/2−t ) , Ac2 = (B − D)(D − 2t ), As1 = πDt ,
As2 = 2 (B − D) t .
K
5.2. Formulation
1.25
1.14 Based on the numerical model, parameters such as width–
1.09 thickness ratio ( B /D ), concrete strength, steel ratio and steel
yielding strength are extensively investigated. 24 FE models are
analyzed, and the above parameters are described as follows: B /D
ratio ranging from 2 to 4; steel yielding strength ranging from
Q235 to Q420; concrete strength covering from C40 to C100 and
steel ratio from 0.05 to 0.1(changing the thickness of the steel
tube). According to Ding et al. [29], the similar derived way can be
adopted in this section. When the numerical load–strain responses
reach the ultimate state (i.e. maximum load-bearing capacity), the
axial stresses of steel tube ( σL, s1) at two points of the steel tube (A2,
A3 as shown in Fig. 1) are obtained, as well as the axial stress ( σL, s1)
to yield strength ratio ( fs ) and the specimen's ultimate strength
( fsc = Nu/Asc , Asc = As + Ac ). The relationship between the axial
stress to yield strength ratio and the specimen's ultimate strength
is shown in Fig. 13 ( α = σL, s1/fs ), where σL, s1 is the axial stress of the
round-ended steel tube. The axial stress to yield strength ratios for
various B /D ratios are listed in Table 4. Based on Von Mises yield
criterion, the corresponding tensile transverse stress ( σθ, s1) of the
round-ended steel tube can also be obtained, as listed in Table 4.
As shown in Fig. 13, it can be found that a bigger B /D ratio leads
to a higher axial stress of steel tube, and the corresponding axial
stress of steel tube is 0.79fs, 0.85fs and 0.87fs. However, the cor-
responding tensile transverse stress of steel tube declines as B /D
ratio increases, which are 0.36fs, 0.28fs and 0.25fs as listed in Ta-
ble 4. This is because with the increase of the value of B /D , the
cross-sectional area of the CFRT stub column increases, and the
ultimate bearing capacity of CFRT stub column increases, then the
axial stress of steel tube becomes greater. However, with the in-
crease of the value of B /D , a higher value of Mechanical Perfor-
mances of Concrete-Filled Steel Tubular Stub Columns With Round
Ends Under Axial Loading can significantly weaken the confining
effect of the steel tube, and the tensile transverse stress of steel
tube becomes smaller.
32 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34

a b

Fig. 13. Relationship between ratio of various stresses and yield stress and ultimate bearing strength.

Fig. 15. Comparisons of those FE modeling ( Nu,1) to calculated results of ultimate

Fig. 14. Relationship between confinement effect coefficient K and width–thick- bearing capacity ( Nu,2 ) using Eq. (11).
ness ratio B /D .
tube can also be obtained, as listed in Table 4.
Therefore, considering the confining stresses, the axial com-
As shown in Fig. 12, according to Ding et al. [16], the relation- pressive stress of core concrete in the two regions (see Fig. 12) can
ship of the constrained concrete confining stress (σθ, c1) caused by be expressed as
round-ended steel and the tensile transverse stress of the round-
ended steel at the ultimate state can be given as ⎧ σ L, c1 = f + kσr,
⎪ c1
c
⎨
⎪
ρ ⎩ σ L, c2 = fc (6)
σr,c1 = σθ,s1
2 (1 − ρ) (5)
where ρ is the steel ratio, k = 3. 4 according to [16].
The tensile transverse stress (σθ, s1) of the round-ended steel The axial compressive stress in the plate steel σL, s2 (see Fig. 12)
 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34 33

Table 5
Calculated strengths for CFRT columns by various design codes.

Specimen ID Nu,2 by DL/T (kN) Nu,2 by EC4 (kN) Nu,2 by Eq. (11) (kN) Experimental values Nu,0 (kN) Nu,0 / DL/T Nu,0 / EC4 Nu,0 / Eq.(11)

CFT1-A 3046 1932 3011 3023 1.008 0.639 0.996

CFT1-B 3023 1918 2977 3265 0.926 0.588 0.912
CFT2-A 3415 2233 3520 3556 0.960 0.628 0.989
CFT2-B 3496 2395 3621 3661 0.955 0.654 0.989
WST1-A 3409 2316 3436 3429 0.994 0.675 1.002
WST1-B 3419 2323 3419 3338 1.024 0.696 1.024
WST2-A 3837 2739 4101 4162 0.922 0.658 0.985
WST2-B 3803 2714 4046 4168 0.912 0.651 0.971
WST3-A 4050 2717 3881 3929 1.031 0.692 0.988
WST3-B 4042 2717 3879 4158 0.972 0.653 0.933
WST4-A 4505 3186 4548 4492 1.003 0.709 1.012
WST4-B 5174 3525 5127 5530 0.936 0.637 0.927
WST5-A 5510 3557 5107 5620 0.980 0.633 0.909
WST5-B 5603 3614 5196 5500 1.019 0.657 0.944
WST6-A 3817 2592 3247 3240 1.178 0.8 1.002
WST6-B 3766 2556 3237 2993 1.258 0.854 1.082
WST7-A 5538 3728 4776 4826 1.148 0.772 0.989
WST7-B 5687 3815 4913 4944 1.150 0.772 0.994
WST8-A 7461 4975 6282 6321 1.180 0.787 0.994
WST8-B 7462 4961 6275 6293 1.186 0.788 0.997
WST9-A 8464 5219 4423 4203 2.014 1.242 1.052
WST9-B 8446 5201 4405 4180 2.021 1.244 1.054
WST10-A 18181 10861 6758 7101 2.560 1.530 0.952
WST10-B 17940 10716 6513 6905 2.598 1.552 0.943
WST11-A 30683 17978 8853 9065 3.385 1.983 0.977
WST11-B 30771 18062 8761 8799 3.497 2.053 0.996

can be expressed as 5.3. Evaluating the calculation method of CFRT column

σ L, s2 = fs
The ultimate bearing capacity Nu,2 of axially-loaded CFRT col-
umns can therefore be expressed as
Nu, 2 = σ L, c1A c1 + σ L, s1A + σ L, c2 A c2 + σ L, s2 As2
Substituting Eqs. (4)–(7) into Eq. (8), based on the existing
design method for the circular CFT stub columns [16], the bearing
capacity under axial compressive can be described as
Nu,2 = fc A c (1 + KΦ)
where Ф is the confinement index and is determined by
Ф = fs As /fc Ac .
Fig. 14 shows the influence of width–thickness ratio on the
coefficient K . And the relationship between K and B /D can be
described as
(7) Wang et al. [23] adopted the design rules specified by EC4 [30],
BS [31], ACI [32] and DL/T [33] to calculate the test strength for the
CFRT stub column. It was indicated that EC4 [30] and DL/T [33]
were closer to the test column strength than those by the other
(8) design rules. Table 5 lists the comparisons between the load
bearing capacities calculated by EC4 [30], DL/T [33] and Eq. (11)
(i.e. Nu,2) and the experimental ones ( Nu,0 ) for all the CFRT speci-
mens. The average value of the ratio of Nu,0 to Nu,2 (by EC4 [30], DL/
T [33] and Eq. (11)) is 1.264, 0.839 and 1.020 with a dispersion
coefficient of 0.293, 0.317 and 0.034, respectively. In general, the
(9) predicted strengths of CFRT columns by Eq. (11) are more appro-
priate to the test column strength than those by EC4 [30] and DL/T
[33]. Therefore, Eq. (11) is adopted as the basic form to predict the
column strength of CFRT specimen.
6. Conclusions

K = 0.8 + 0.9D/B (10) This paper presents an experimental study on 22 axially loaded
CFRT stub columns compared with 4 circular CFT stub columns. A
Therefore, the ultimate bearing capacity for CFRT stub columns further analysis based on finite element method is also carried out.
under axial compression Nu,2 can be obtained as The following conclusions can be drawn based on this work:

Nu,2 = A c fc [1 + (0.8 + 0.9D/B) Φ]
Fig. 15 shows the comparison between the bearing capacities
from FE modeling ( Nu,1) and those from Eq. (11) ( Nu,2). The average
ratio of Nu,1 to Nu,2 is 1.011 with a dispersion coefficient of 0.026,
which suggests that the proposed formula is in good accordance
with FE modeling results.
Table 3 lists the comparisons between the load bearing capa-
cities calculated by Eq. (11) (i.e. Nu,2) and the experimental ones
( Nu,0 ) for all the CFRT specimens. The average value of the ratio of
Nu,0 to Nu,2 is 1.020 with a dispersion coefficient of 0.034. Com-
pared with experimental results, the load bearing capacities from
Eq. (11) are on the safer side.
(11) (1) Experimental results suggest that the typical failure modes of
steel tubes are local outward buckling. Local buckling of the
round-ended steel tube is associated with shear failure of in-
filled concrete. For CFRT stub columns, with the increase of
width–thickness ratio, the ductility of the specimen decreases.
The CFRT stub columns with bigger width–thickness ratio seem
to provide insufficient confining effect on the concrete core
inside it.
(2) The FE analysis of axially-loaded CFRT stub columns shows a
good agreement with experimental results. FE modeling results
also confirm that the confinement effect increases as a function
of the width–thickness ratio. The specimen with a lower width–
thickness ratio can lead to a higher confinement effect.
 34 D. Faxing et al. / Thin-Walled Structures 97 (2015) 22–34
(3) The parametric studies indicate that the concrete strength, tube
thickness and wall-thickness of the steel tube also have a great
effect on the ultimate bearing capacity as well as the column
ductility. With the increasing concrete strength, the ultimate
bearing capacity of the specimen increases while the specimen
ductility decreases. The ultimate bearing capacity and ductility
of the specimen increases with the increasing tube thickness.
However, a higher width–thickness ratio can lead to a lower
ultimate bearing capacity and a lower column ductility.
(4) A simplified approach is proposed to predict the ultimate
bearing capacity of CFRT stub columns based on the existing
design method for the CFT stub columns. The predicted results
from the proposed method agree well with those from
experiments.
Acknowledgments
This research work was financially supported by the Program
for Changjiang Scholars and Innovative Research Team in Uni-
versity (PCSIRT), Grant no. IRT1296, the Program for New Century
Excellent Talents in University, Grant no. NCET-11-0508, and the
National Key Technology R&D Program, Grant no. 2011BAJ09B02.
References
[1] S.P. Schneide, Axially loaded concrete-filled steel tubes, ASCE J. Struct. Eng. 124
(10) (1998) 1125–1138.
[2] M.F. Hassanein, Numerical modeling of concrete-filled lean duplex slender
stainless steel tubular stub columns, J. Constr. Steel Res. 66 (2010) 1057–1068.
[3] M.F. Hassanein, O.F. Kharoob, Q.Q. Liang, Circular concrete-filled double skin
tubular short columns with external stainless steel tubes under axial com-
pression, Thin-Walled Struc. 73 (12) (2013) 252–263.
[4] M.F. Hassanein, O.F. Kharoob, Analysis of circular concrete-filled double skin
tubular slender columns with external stainless steel tube, Thin-Walled Struct.
79 (6) (2014) 23–37.
[5] X. Chang, Z.L. Ru, W. Zhou, et al., Study on concrete-filled stainless steel carbon
steel tubular (CFSCT) stub columns under compression, Thin-Walled Struct. 63
(2) (2013) 125–133.
[6] X. Chang, Y.Y. Wei, Y.C. Yun, Analysis of steel-reinforced concrete-filled steel
tubular (SRCFST) columns under cyclic loading, Constr. Build. Mater. 28 (1)
(2012) 88–95.
[7] X. Chang, C.K. Huang, Y.J. Chen, Mechanical performance of eccentrically loa-
ded pre-stressing concrete filled circular steel tube columns by means of ex-
pansive cement, Eng. Struct. 31 (11) (2009) 2588–2597.
[8] X. Chang, X.L. Luo, C.X. Zhu, et al., Analysis of circular concrete-filled steel tube
support in high ground stress conditions, Tunnel. Undergr. Space Technol. 43
(6) (2014) 41–48.
[9] E. Ellobody, Nonlinear behavior of concrete-filled stainless steel stiffened
slender tube columns, Thin-walled Struct. 45 (3) (2007) 259–273.
[10] Z.W. Yu, F.X. Ding, C.S. Cai, Experimental behavior of circular concrete-filled
steel tube stub columns, J. Constr. Steel Res. 63 (2) (2007) 165–174.
[11] P. Thayalan, T. Aly, I. Patnaikuni, Behaviour of concrete-filled steel tubes under
static and variable repeated loading, J. Constr. Steel Res. 65 (4) (2009)
900–908.
[12] Mohamed Meih, On the performance of circular concrete-fill high strength
steel columns under axial loading, Alexandria Eng. J. 51 (2) (2012) 109–119.
[13] Y.Q. Tu, Y.F. Shen, P. Li, Behavior of multi-cell composite T-shaped concrete-
filled steel tubular columns under axial compression, Thin-Walled Struct. 85
(12) (2014) 57–70.
[14] K. Uenaka, Experimental study on concrete filled elliptical/oval steel tubular
stub columns under compression, Thin-Walled Struct. 78 (5) (2014) 131–137.
[15] Q.X. Ren, L.H. Han, D. Lam, et al., Experiments on special-shaped CFST stub
columns under axial compression, J. Constr. Steel Res. 98 (6) (2014) 123–133.
[16] F.X. Ding, Z.W. Yu, Y. Bai, et al., Elasto-plastic analysis of circular concrete-filled
steel tube stub columns, J. Constr. Steel Res. 67 (10) (2011) 1567–1577.
[17] L.H. Han, Tests on stub columns of concrete-filled RHS sections, J. Constr. Steel
Res. 58 (3) (2002) 353–372.
[18] Z. Tao, B. Uy, F.Y. Liao, et al., Nonlinear analysis of concrete-filled square
stainless steel stub columns under axial compression, J. Constr. Steel Res. 67
(2) (2011) 1719–1732.
[19] D. Lam, C.A. Williams, Experimental study on concrete filled square hollow
sections, Steel Compos. Struct. 4 (2) (2004) 98–112.
[20] K. Sakino, H. Nakahara, S. Morino, et al., Behavior of centrally loaded concrete-
filled steel-tube short columns, ASCE J. Struct. Eng. 130 (2) (2004) 180–188.
[21] X. XieJ, Z.A. Lu, P. Tang, et al., Modal analysis and experimental study on
round-ended CFST coupled column cable stayed bridge, in: Proceedings of the
Second International Conference on Mechanic Automation and Control En-
gineering (MACE2011), Inner Mongolia, China, 2011, vol. 6, pp. 15–17.
[22] J.X. Xie, Z.A. Lu, Numerical simulation and test study on non-uniform areas of
round-ended CFST tubular tower, in: Proceedings of the Third International
Conference on Information and Computing (ICIC2010), Wu xi, Jing su, China,
2010, vol. 6, pp. 4–6.
[23] Z.B. Wang, J. Chen, T.P. Xie, et al., Behavior of concrete-filled round-end steel
tubular stub columns under axial compression, J. Build. Struct. 35 (7) (2014)
123–130 (in Chinese).
[24] J.X. Xie, Z.A. Lu, Test study and contact analysis of round-ended CFST coupled
column, in: Proceedings of the Second International Conference on Modeling
And Simulation (ICMS2009), Manchester, England, UK, 2009, vol. 5, pp. 21–22.
[25] L.H. Han, Q.X. Ren, W. Li, Tests on stub stainless steel-concrete-carbon steel
double-skin tubular (DST) columns, J. Constr. Steel Res. 67 (3) (2011) 437–452.
[26] GB2975, Test Standard for Steel Material, Chinese Planning Press, Peking, 1982
(in Chinese).
[27] GBJ81-85, Test Standard for Concrete Material, Chinese Planning Press, Peking,
1985 (in Chinese).
[28] Z. Tao, B. Uy, L.H. Han, et al., Analysis and design of concrete-filled stiffened
thin-walled steel tubular columns under axial compression, Thin-Walled
Struct. 47 (12) (2009) 1544–1556.
[29] F.X. Ding, C.J. Fang, Y. Bai, et al., Mechanical performance of stirrup-confined
concrete-filled steel tubular stub columns under axial loading, J. Constr. Steel
Res. 98 (11) (2014) 146–157.
[30] Eurocode 4, Design of Steel and Concrete Structures, Part 1.1, General Rules
and Rules DOE Building, DD ENV 1994-1-1-1996, British Standards Institution,
London, 1996.
[31] BS 5400, Steel, Concrete and Composite Bridges: Part 5: Code of Practice for
Design of Composite Bridges, British Standards Institution, London, 2005.
[32] ACI 318-95, Building Code Requirements for Structural Concrete and Com-
mentary, American Concrete Institution, Detroit, MI, USA, 1999.
[33] Standard of Economic and Trade Committee of China Electric Power Pub-
lishing House, Peking, 1999 (in Chinese).