You are on page 1of 16

Engineering Structures 179 (2019) 268–283

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Ultra-high strength circular short CFST columns: Axisymmetric analysis, T


behaviour and design

V.I. Patela, , M.F. Hassaneinb, Huu-Tai Thaic, H. Al Abadid, M. Elchalakanie, Y. Baif
a
School of Engineering and Mathematical Sciences, La Trobe University, Bendigo, VIC 3552, Australia
b
Department of Structural Engineering, Faculty of Engineering, Tanta University, Tanta, Egypt
c
Department of Infrastructure Engineering, The University of Melbourne, Parkville, VIC 3010, Australia
d
School of Engineering and Mathematical Sciences, La Trobe University, Bundoora, VIC 3086, Australia
e
School of Civil, Environmental and Mining Engineering, The University of Western Australia, Crewley, WA 6009, Australia
f
Institute of Steel, Lightweight and Composite Structures, Department of Civil and Environmental Engineering, Ruhr-Universität Bochum, Bochum D-44801, Germany

A R T I C LE I N FO A B S T R A C T

Keywords: A nonlinear axisymmetric simulation of axially loaded ultra-high strength circular short Concrete-Filled Steel
UHSC Tubular (CFST) columns is reported in this paper. An efficient material law of Ultra-High Strength Concrete
CFST (UHSC) is employed in this axisymmetric simulation to consider the confining effects which increase the ultimate
Axial performance strain, stress and axial ductility. The peak loads and load-strain responses of short CFST columns predicted by the
Ductility
axisymmetric model are verified with the independent test data and those predicted by the three-dimensional
Confinement effects
(3D) model. The structural characteristics of circular short CFST columns having different material and geo-
Design
metric parameters are subsequently examined by the axisymmetric analysis. The numerical results indicate that
the column size can approximately be reduced by 50% when UHSC is used in the composite construction instead
of Normal Strength Concrete (NSC). Comparisons of the current test data are made with the current structural
codes including AS/NZS 5100.6, Eurocode 4, AISC and GB 50936. The comparisons demonstrate that the
Australian/New Zealand and European design codes provide consistent predictions for the ultimate axial
strengths of ultra-high strength circular short CFST columns. On the other hand, the North American design code
is conservative, whereas the Chinese code is unconservative.

1. Introduction Although extensive research on CFST circular columns with high


strength concrete and steel has been undertaken, the numerical analysis
CFST columns are considered as an economical option for the is very limited for the simulation of circular short CFST columns with
composite structural members since most of the concentrically com- UHSC. Therefore, the numerical investigation on ultra-high strength
pressive force is carried out by the concrete which is cheaper than the short circular CFST columns is needed [3].
steel. Accordingly, further economies could be obtained by utilizing Many researchers have experimentally assessed the characteristics
UHSC in CFST columns. The use of UHSC in CFST columns considerably of CFST columns with NSC and High Strength Concrete (HSC) [4–14].
decreases the cross-sectional size, and thus saves the workspace area. The constructional applications of CFST columns were recently re-
However, the use of UHSC in CFST columns significantly reduces their viewed by Han et al. [15]. A comprehensive experimental study was
ductility [1]. To overcome this problem, steel tubes with high yield first performed by Liew and Xiong [1] and Xiong et al. [3] to investigate
stress and small diameter-to-thickness D / t ratio should be employed to the axial performance of concentrically compressed ultra-high strength
ensure the ductility of CFST columns with UHSC. The workability of the CFST short columns. Their specimens were made from UHSC with
UHSC is lower than that of the NSC [1]. The prefabricated columns compressive strength up to 193 MPa and hollow steel tube having yield
should be used to avoid the workability problems at the construction stress up to 428 MPa. The results of their experiments demonstrated
sites. The current structural codes restrict the use of UHSC in CFST that the ductility is decreased because of the brittleness of UHSC, but it
columns [1]. This is due to the limited research on ultra-high strength increases by utilizing the small D / t ratios. They concluded that Euro-
CFST columns. Techno station in Tokyo Japan [2] is one of the well- code 4 [16] method safely estimates the maximum capacities of axially
known example in adopting the application of UHSC in CFST columns. loaded ultra-high strength short circular CFST columns. Moreover, the


Corresponding author.
E-mail addresses: v.patel@latrobe.edu.au (V.I. Patel), mostafa.fahmi@f-eng.tanta.edu.eg (M.F. Hassanein).

https://doi.org/10.1016/j.engstruct.2018.10.081
Received 6 December 2017; Received in revised form 15 October 2018; Accepted 29 October 2018
Available online 06 November 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

3000

2500

Axial load P (kN)


2000

1500
Experiment (C6)
1000 Friction angle ( ) = 0.01°
Friction angle ( ) = 20°
Friction angle ( ) = 30°
500
Friction angle ( ) = 35°
Friction angle ( ) = 40°
0
0 0.03 0.06 0.09 0.12
Fig. 1. Cross-section of CFST columns. Axial strain ε

(a) Effect of friction angle

3000

2500

Axial load P (kN)


2000

1500

Experiment (C5)
1000
Flow stress ratio (K) = 0.8
Flow stress ratio (K) = 0.9
500
Flow stress ratio (K) = 1.0

0
Fig. 2. Stress-strain model for confined concrete in circular CFST columns. 0 0.03 0.06 0.09 0.12
Axial strain ε
required ductility of short circular CFST columns with UHSC occurred (b) Effect of flow stress ratio
at a steel contribution ratio of 0.3 or more. The steel contribution ratio
is defined as the ratio of ultimate strength of steel section to the ulti-
3000
mate strength of CFST column section. They also concluded that there is
a limited research performed on CFST columns with UHSC.
Nowadays, the concrete core may be reinforced by fibres or steel 2500
bars to enhance the ductility of the column. Australian Standard,
AS3600-2009 [17], allows the use of fibre reinforced concrete in the
Axial load P (kN)

2000
structural design of bridges. The influences of high performance con-
structional materials on the characteristics of axially and eccentrically
compressed slender CFST beam-columns were examined by Portolés 1500
et al. [18]. They showed that the filled UHSC generally decreases the Experiment (C5)
ductility but it increases by using the steel fibres. They carried out 1000 Dilation angle ( ) = 0.01°
experiments on 24 beam-column specimens with NSC, HSC and UHSC.
The results indicate that the contribution of UHSC on concentrically Dilation angle ( ) = 20°
loaded CFST columns is more significant than that of eccentrically 500 Dilation angle ( ) = 30°
loaded ones. Xiong et al. [19] experimentally assessed the character- Dilation angle ( ) = 40°
istics of ultra-high strength slender CFST beam-columns with circular 0
sections. The test results showed that pin-ended slender CFST beam- 0 0.03 0.06 0.09 0.12
columns are mainly failed by the overall buckling. Axial strain ε
Considerable numerical research on the structural characteristics of
CFST columns has been devoted over the past two decades [20–29]. (c) Effect of dilatation angle
However, very limited numerical study on CFST columns with UHSC Fig. 3. Effects of Drucker-Prager parameters on load-strain responses of ultra-
can be found in the literature. Wang et al. [30] employed the finite high strength CFST columns.
element (FE) simulation code Abaqus to assess the structural char-
acteristics of concentrically the compressed ultra-high strength short
circular CFST columns. A new stress-strain relationship for the con-
finement mechanism of UHSC is proposed. Xiong et al. [19]

269
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

3000

Axial load P (kN) 2500

2000

1500

1000 Experiment (C7)

500 Unconfined concrete


model
Confined concrete model
0
0 0.03 0.06 0.09 0.12
Axial strain ε
Fig. 5. Stress-strain model for hot-rolled steel in CFST columns.
Fig. 4. Effects of concrete confinement on axial load-strain curves.

2. Constitutive stress-strain models


numerically investigated the performance of eccentrically compressed
slender circular CFST beam-columns with UHSC. The material law 2.1. Confined concrete
suggested by Eurocode 2 [31,32] for the filled concrete was im-
plemented in the FE model based on Abaqus. Le Hoang and Fehling As the lateral strain of the concrete component in CFST columns
[33,34] recently developed a FE model using Atena-3D for the simu- becomes more than that of the outer tube, the radial pressure develops
lation of concentrically compressed circular ultra-high strength CFST at the contacting surfaces between the concrete and steel. A steel tube
columns. An equation was reported to estimate the peak capacities of in circular CFST column is bi-axially stressed in the hoop and axial
ultra-high strength circular CFST short columns. The results indicated directions. The hoop stress component induces the confinement to the
that the proposed model safely estimates the maximum load of the inner concrete while the axial stress accelerates the steel tube buckling.
ultra-high strength circular short CFST columns. The confinement mechanism offered by the steel tube to the concrete
As mentioned, the research study on UHSC is very limited in com- component increases both the strain ductility and axial strength. The
parison with the investigation on NSC and HSC. An axisymmetric si- constitutive concrete model given by Patel et al. [35] considering the
mulation is undertaken in this paper to model the characteristics of confinement effects is employed in this axisymmetric analysis. The
axially loaded circular CFST short columns with UHSC as illustrated in material law for the concrete component is illustrated in Fig. 2. The
Fig. 1. The proposed axisymmetric model is verified with the specimens concrete exhibits elastic behaviour up to stress of 0.4fc′ which is reported
tested by Xiong et al. [3]. The verification study demonstrates that the in EN 1992-1-1 [31]. The plasticity curve of concrete in compression
proposed axisymmetric model is not only much simpler than detailed starts at the stress of 0.4fc′. The material law of the confining concrete
3D model, but also predicts well the behaviour of axially loaded circular suggested by Mander et al. [36] is considered to compute the nonlinear
CFST short columns with UHSC including elastic stiffness, ultimate curve A-B of the stress-strain response, as given by:
load-carrying capacity and post-peak response. A parametric study is
fcc′ λ (εc / εcc′ )
also performed on ultra-high strength short circular CFST columns for a σc =
wide range of parameters including D / t ratio, concrete compressive λ − 1 + (εc / εcc′ ) λ (1)
strength and steel yield stress. Finally, the design of ultra-high strength
Ec
circular short CFST columns using the existing code provisions and λ=
Ec − (fcc′ / εcc′ ) (2)
proposed model is discussed in details.
Ec = 4700 γc fc′ (MPa) (3)

Table 1
Comparison of ultimate load of ultra-high strength circular CFST short columns.
Specimen D (mm) t (mm) D/t L (mm) fc′ (MPa) f y (MPa) fu (MPa) Es (GPa) Pu. exp (kN) Pu.FE (kN) Pu .FE
Pu . exp

C3 114.3 3.6 32 250 173.5 403 460 213 2422 2365 0.96
C4 114.3 3.6 32 250 173.5 403 460 213 2340 2364 1.00
C5 114.3 3.6 32 250 184.2 403 460 213 2497 2461 0.97
C6 114.3 3.6 32 250 184.2 403 460 213 2314 2457 1.04
C7 114.3 6.3 18 250 173.5 428 520 209 2610 2666 1.02
C8 114.3 6.3 18 250 173.5 428 520 209 2633 2697 1.01
C10 219.1 5.0 44 600 185.1 380 460 205 7837 7970 1.01
C11 219.1 5.0 44 600 193.3 380 460 205 8664 8201 0.94
C13 219.1 10.0 22 600 185.1 381 460 212 9085 8956 0.97
C14 219.1 10.0 22 600 193.3 381 460 212 9187 9195 0.99
C15 219.1 6.3 35 600 163.0 300 430 202 6915 7017 1.01
C16 219.1 6.3 35 600 175.4 300 430 202 7407 7425 0.99
C17 219.1 6.3 35 600 148.8 300 430 202 6838 6614 0.95
C18 219.1 6.3 35 600 174.5 300 430 202 7569 7390 0.96

Mean 1.00
Standard deviation (SD) 0.03
Coefficient of variation (CoV) 0.03

270
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

(a) Axisymmetric model (b) 3D model


Fig. 6. Modelling of CFST columns.

in which σc represents the axial concrete stress under compression, εc model is incorporated in the axisymmetric analysis conducted in this
denotes the axial strain at σc , fcc′ stands for the compressive strength of paper for ultra-high strength circular short CFST columns, which is
confined concrete, εcc′ represents the axial strain at fcc′ , λ is a material estimated by
constant, fc′ is the compressive strength of unconfined concrete and Ec
denotes the concrete elastic modulus [37]. The reduction factor γc al-
frp =
( )
⎧ 0.02663 − 0.0002504 Dt f y D
for 17⩽ t < 47
lows the specimen scale effect. The factor γc is implemented to de- ⎨ 0.01588 − 0.0000149 f
termine the concrete strength, which can be expressed by [38]: ⎩
( ) D
t y for 47 ⩽
D
t
⩽ 221
(8)

γc = 1.85Dc−0.135 (0.85 ⩽ γc ⩽ 1.0) (4) where f y represents the steel yield stress. The post-peak performance of
the concrete model represented by two linear trends (BC and CD) as
whereDc = D − 2t represents the diameter of the filled concrete, in shown in Fig. 2 can be obtained from the following equations [35]:
which D denotes the diameter and t stands for the tube thickness as
depicted in Fig. 1.
The uniaxial compressive strength fcc′ and strain εcc′ of the confined σc =
⎧ f′ +
⎪ cc ( ′
εc − εcc

εe − εcc ) (f e − fcc′ ) for εcc' < εc ⩽ εe

concrete are higher than those of the unconfined concrete. The equa-
tions given by Liang and Fragomeni [22] are adopted herein to accu-
⎪ fe +

( εc − εe
εcu − εe ) (f
cu − fe ) for εe < εc ⩽ εcu
(9)
rately predict the ultimate concrete stress fcc′ and strain εcc′ as where εe and εcu are the concrete axial strains at Points C and D as
illustrated in Fig. 2. The compressive strain εe is taken as 10εc′ corre-
fcc′ = γc fc′ + k1 frp (5)
sponding to Point C [40], and the compressive strain εcu at Point D is
taken as 30εc′ [40]. In Eq. (9), fe denotes the compressive stress at strain
frp ⎞
εcc′ = εc′ ⎛⎜1 + k2 ⎟
εe and fcu stands for the compressive stress at εcu . The stresses fe and fcu
⎝ γc f c′ ⎠ (6) are given by [35]:

in which frp is the confining pressure around concrete. The constants fe = α c fcc′ (10)
k1 = 4.1 and k2 = 20.5 are used in the axisymmetric analysis [39]. The
fcu = βc fcc′ (11)
axial strain εc′ is expressed by [22]
in which the factors α c and βc represent the confining influence on the
⎧ 0.002 for γc fc′ ⩽ 28 (MPa)
⎪ concrete ductility in the post-peak range. These factors depend on the
γc f c′ − 28
εc′ = 0.002 + for 28< γc fc′ ⩽ 82 (MPa) concrete compressive strength [40] and were proposed by Patel et al.
⎨ 54, 000
⎪ [35] as
0.003 for γc fc′ > 82 (MPa) (7)

⎧ 0.8987 − 0.00122γc fc′ for γc fc′ < 50 MPa
The confinement mechanism offered by the steel tube in circular αc =
⎨ 0.774 - 0.0016γc fc′ for γc fc′ ⩾ 50 MPa (12)
CFST columns was given by Patel et al. [35]. This confining pressure ⎩

271
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

2500 Since the steel tube provides the confining pressure to infilled
concrete, the core concrete exhibits the three-dimensional stress, the
linear Drucker-Prager model was employed to consider the concrete
2000 behaviour in such multiple stress state. The yield surface of the
Drucker-Prager model expands by an increase in the hydrostatic pres-
Axial load P (kN)

sure. The Drucker-Prager model was utilized for simulating the concrete
1500 behaviour. Sensitivity analysis is undertaken to examine the influences
of friction angle φ, flow stress ratio K and dilation angle ψ on the load-
strain curve of ultra-high strength circular CFST short columns tested
1000 by Xiong et al. [3]. Fig. 3 illustrates the structural characteristic of short
Experiment (C4) CFST columns is not affected by the flow stress ratio K and dilation
angle ψ . Therefore, the flow stress ratio K and dilation angle ψ were
500 3D FE model-CPU time:337 sec
taken as 0.8 and 30°, respectively. As appear from the figure, the fiction
Axisymmetric model-CPU time:18 sec angle φ does not influence the initial stiffness of the load-strain re-
sponse. The peak load slightly increases as the friction angle φ in-
0
creases. The 20° friction angle φ gives the accurate prediction of the
0 0.03 0.06 0.09 0.12 0.15
post-peak response. The Poisson’s ratio is taken as 0.25 for the elastic
Axial strain ε
behaviour of UHSC. Liew and Xiong [1] suggested that the Poisson’s
(a) Axial load-strain response ratio of 0.25 should be utilised for UHSC instead of 0.2 which is gen-
erally used for HSC.
The use of only linear Drucker-Prager model causes the dis-
crepancies between the predicted results and experimental results. In
Fig. 4, the specimen C7 (Table 1) is analyzed using the linear Drucker-
Prager model without incorporating the confining pressure given in Eq.
(8). It can be seen from Fig. 4 that the linear Drucker-Prager model does
not accurately predict the ultimate axial strengths and axial load-strain
behaviour in the post-peak range. Therefore, the concrete confining
pressure given in Eq. (8) needs to be used in Abaqus with the linear
Drucker-Prager model, which has been successfully used by Hassanein
et al. [41].

2.2. Hot-rolled steel

The structural steel tube of CFST columns is bi-axially stressed be-


cause of the confining effects. The developed hoop tension in the steel
component decreases the yield stress along the axial direction [22].
This influence is considered in the steel stress-strain characteristics. A
typical material behaviour for the hot-rolled steel is illustrated in Fig. 5,
(b) von Mises stress predicted by 3D model in which σs represents the axial stress in the steel component, εs is the
strain at the stress σs , εy is the steel yield strain, εt stands for the hard-
ening strain which is taken as 0.005 and fu and εu denote the ultimate
stress and strain of steel, respectively. The ultimate strain εu of hot-
rolled steel is taken as 0.1. The behaviour of steel is considered to be
linear up to 0.9f y . The stress-strain curve beyond 0.9f y stress is defined
using the plastic material properties. The Poisson's ratio for the hot-
rolled steel is assumed as 0.3 for the elastic region of stress-strain re-
sponse.

3. Analysis of circular CFST columns

3.1. General description

Abaqus [42] is used for the computational simulation of con-


centrically loaded short circular CFST columns with UHSC. The ax-
isymmetric model and corresponding three dimension model are shown
in Fig. 6. The initial geometric imperfection was employed in the
hollow steel tube and considered in the axisymmetric simulation to
allow for the outwards local buckling of steel tubes. The shape of the
(c) von Mises stress predicted by axisymmetric model imperfection of steel tubes is based on the first buckling mode obtained
from the eigenvalue analysis of hollow steel tubes, and the magnitude
Fig. 7. Comparison between 3D model and axisymmetric model. of the imperfection is taken as 0.01t with t being the steel tube thickness
[43]. A Python script over the conventional CAE technique is re-
commended for the FE simulation of axially loaded CFST columns. The
⎧ α c − 0.1 for γc fc′ < 50 MPa
βc = Python script can generate parts, materials, loads, steps, submit analysis
⎨ 0.4 for γc fc′ ⩾ 50 MPa (13)
⎩ jobs, write output database and view the analysis results for the mul-
tiple times using the programming loop, which can be downloaded

272
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

2500 3000 3000

2500 2500
2000

Axial load P (kN)


Axial load P (kN)

2000 2000

Axial load P (kN)


1500

1500 1500

1000
Experiment (C3) 1000 Experiment (C5) 1000 Experiment (C7)

Experiment (C4) Experiment (C6) Experiment (C8)


500
500 500
Axisymmetric analysis Axisymmetric analysis Axisymmetric analysis

0 0 0
0 0.03 0.06 0.09 0.12 0.15 0 0.03 0.06 0.09 0.12 0 0.04 0.08 0.12 0.16

Axial strain ε Axial strain ε Axial strain ε

8000 9000 10000

8000 9000
7000
8000
6000 7000
7000

Axial load P (kN)


Axial load P (kN)

Axial load P (kN)

6000
5000
6000
5000
4000 5000
4000
3000 4000
3000
Experiment (C10) Experiment (C11) 3000 Experiment (C13)
2000
2000 2000
Axisymmetric analysis Axisymmetric analysis Axisymmetric analysis
1000
1000 1000

0 0 0
0 0.03 0.06 0.09 0.12 0 0.03 0.06 0.09 0.12 0 0.03 0.06 0.09 0.12
Axial strain ε Axial strain ε Axial strain ε

10000 8000 8000

9000
7000 7000
8000
6000 6000
7000
Axial load P (kN)

Axial load P (kN)

Axial load P (kN)

6000 5000 5000

5000 4000 4000


4000
3000 3000
3000 Experiment (C14) Experiment (C15) Experiment (C16)
2000 2000
2000
Axisymmetric analysis Axisymmetric analysis Axisymmetric analysis
1000 1000 1000

0 0 0
0 0.015 0.03 0.045 0.06 0.075 0.09 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0 0.03 0.06 0.09 0.12 0.15
Axial strain ε Axial strain ε Axial strain ε

8000 8000

7000 7000

6000 6000
Axial load P (kN)
Axial load P (kN)

5000 5000

4000 4000

3000 3000
Experiment (C17) Experiment (C18)
2000 2000

Axisymmetric analysis Axisymmetric analysis


1000 1000

0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.04 0.08 0.12 0.16
Axial strain ε Axial strain ε

Fig. 8. Comparison between axisymmetric analysis and experimental results [3].

273
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

Fig. 9. Effects of friction angle on the post-peak comparison of Specimens C10, C15, C16, C17 and C18 [3].

from the link given in Appendix. preventing all displacements and rotations is imposed on the bottom
reference point as shown in Fig. 6(a).
3.2. Element type and mesh size
3.4. Axisymmetric model versus 3D model
The experimental observations indicate that the concrete core in
circular CFST columns does not rotate [3]. Therefore, CAX4R element The geometry, loadings, boundary conditions and materials of axi-
which prevents the rotation is used in the modelling of the steel and ally loaded circular short CFST columns are axisymmetric to the vertical
concrete components as shown in Fig. 6(a). This is a 4-node reduced- axis. Therefore, these columns can be simulated by using an economical
integration axisymmetric quadrilateral element. The element size in axisymmetric analysis instead of a full 3D model. The 3D finite element
lateral and longitudinal directions was chosen as t/2 for the steel and t model is shown in Fig. 6(b). In axisymmetric analysis, the input geo-
for the concrete, in which t is the tube thickness as illustrated in Fig. 1. metry is 2D, but the predicted results are 3D because of the rotational
The axisymmetric model consists of a relatively fine mesh. This mesh is symmetry. Fig. 7 compares the axial load-strain responses and de-
deemed fine enough to accurately estimate the correct characteristics of formation shape of ultra-high strength circular short CFST columns
ultra-high strength circular CFST columns. The element size in the ax- predicted by axisymmetric model and full 3D model. A close agreement
isymmetric analysis is depicted in Fig. 6(a). between the predictions indicates that the axisymmetric model is not
only accurate but also very efficient compared with full 3D simulation
3.3. Contact interaction, loading and boundary conditions since the computational time of the axisymmetric analysis is less than
approximate 19 times compared with that of the 3D model.
The contact interaction between steel tubes and infilled concrete is
modelled by the surface-to-surface contact algorithm with Hard contact 4. Verification of axisymmetric model
condition for the normal behaviour and Coulomb friction model for the
tangent behaviour. The characteristics of CFST columns are not affected 4.1. Peak load of concentrically loaded CFST short columns
by the friction coefficient used for the Coulomb friction model [43].
Therefore, the rough surface of the concrete is modelled by specifying a The dimension of tested columns and concrete and steel material
friction coefficient of 0.3. parameters are listed in Table 1. The experimental investigation carried
The top and bottom surfaces of CFST columns are connected to out by Xiong et al. [3] consists of 14 ultra-high strength circular short
reference points located at the centroid of the section using a rigid body CFST columns with D / t ratio varied from 18 to 44. The column length
constraint. With this constraint, the end sections are planar during the was equal to either 250 mm or 600 mm. The experimental specimens
analysis and hence there is no need to include the end plates or stif- were made from the hot-rolled steel with yield stress from 300 MPa to
feners in the model [25]. The loading and boundary conditions are then 428 MPa. The compressive strength of UHSC is ranged between
directly applied to reference points. In the axisymmetric model, the top 149 MPa and 193 MPa. The predicated and experimental peak loads of
end is free to displace in the loading direction and the displacement ultra-high strength circular short CFST columns under axial loads are
loading is applied at the top reference point. A boundary condition presented in Table 1, in which Pu. exp denotes the peak load obtained

274
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

Table 2
Parametric study of circular short CFST columns (D = 600 mm, L/D = 3).
Specimens t (mm) D/t f y (MPa) fu (MPa) fc′ (MPa) Ps.FE (kN) Pu.FE (kN) Pr.FE (kN) δ=
Ps .FE Pr .FE
Pu .FE Pu .FE

S1 30 20 380 460 190 21,393 66,842 48,685 0.32 0.73


S2 20 30 380 460 190 14,009 62,307 41,678 0.22 0.67
S3 15 40 380 460 190 10,471 59,348 37,928 0.18 0.64
S4 12 50 380 460 190 8409 57,601 35,511 0.15 0.62
S5 30 20 420 500 60 23,551 44,697 36,464 0.53 0.82
S6 30 20 420 500 70 23,551 46,631 38,221 0.51 0.82
S7 30 20 420 500 80 23,551 48,468 39,896 0.49 0.82
S8 30 20 420 500 90 23,551 50,552 41,610 0.47 0.82
S9 30 20 420 500 100 23,551 52,398 42,893 0.45 0.82
S10 30 20 420 500 110 23,551 54,457 43,946 0.43 0.81
S11 30 20 420 500 120 23,551 56,349 44,990 0.42 0.80
S12 30 20 420 500 130 23,551 58,279 45,870 0.40 0.79
S13 30 20 420 500 140 23,551 60,252 46,953 0.39 0.78
S14 30 20 420 500 150 23,551 62,036 47,871 0.38 0.77
S15 30 20 420 500 160 23,551 64,007 48,828 0.37 0.76
S16 30 20 420 500 170 23,551 66,091 49,857 0.36 0.75
S17 30 20 420 500 180 23,551 67,870 50,820 0.35 0.75
S18 30 20 420 500 190 23,551 69,842 51,728 0.34 0.74
S19 20 30 420 500 60 15,472 37,365 27,538 0.41 0.74
S20 20 30 420 500 70 15,472 39,410 29,352 0.39 0.74
S21 20 30 420 500 80 15,472 41,615 31,216 0.37 0.75
S22 20 30 420 500 90 15,472 43,757 32,993 0.35 0.75
S23 20 30 420 500 100 15,472 45,826 34,542 0.34 0.75
S24 20 30 420 500 110 15,472 47,862 35,642 0.32 0.74
S25 20 30 420 500 120 15,472 49,953 36,729 0.31 0.74
S26 20 30 420 500 130 15,472 52,088 37,835 0.30 0.73
S27 20 30 420 500 140 15,472 54,220 38,865 0.29 0.72
S28 20 30 420 500 150 15,472 55,894 39,938 0.28 0.71
S29 20 30 420 500 160 15,472 58,328 40,985 0.27 0.70
S30 20 30 420 500 170 15,472 60,449 42,015 0.26 0.70
S31 20 30 420 500 180 15,472 62,566 43,066 0.25 0.69
S32 20 30 420 500 190 15,472 64,698 44,076 0.24 0.68
S33 15 40 420 500 60 11,616 32,955 22,825 0.35 0.69
S34 15 40 420 500 70 11,616 35,172 24,723 0.33 0.70
S35 15 40 420 500 80 11,616 37,386 26,603 0.31 0.71
S36 15 40 420 500 90 11,616 39,557 28,401 0.29 0.72
S37 15 40 420 500 100 11,616 41,768 29,944 0.28 0.72
S38 15 40 420 500 110 11,616 43,815 31,028 0.27 0.71
S39 15 40 420 500 120 11,616 46,027 32,113 0.25 0.70
S40 15 40 420 500 130 11,616 48,294 33,201 0.24 0.69
S41 15 40 420 500 140 11,616 50,424 34,281 0.23 0.68
S42 15 40 420 500 150 11,616 52,622 35,380 0.22 0.67
S43 15 40 420 500 160 11,616 54,816 36,485 0.21 0.67
S44 15 40 420 500 170 11,616 56,881 37,585 0.20 0.66
S45 15 40 420 500 180 11,616 58,941 38,666 0.20 0.66
S46 15 40 420 500 190 11,616 61,280 39,649 0.19 0.65
S47 12 50 420 500 60 9285 30,294 20,005 0.31 0.66
S48 12 50 420 500 70 9285 32,553 21,868 0.29 0.67
S49 12 50 420 500 80 9285 34,810 23,712 0.27 0.68
S50 12 50 420 500 90 9285 37,057 25,567 0.25 0.69
S51 12 50 420 500 100 9285 39,091 27,100 0.24 0.69
S52 12 50 420 500 110 9285 41,491 28,265 0.22 0.68
S53 12 50 420 500 120 9285 43,727 29,398 0.21 0.67
S54 12 50 420 500 130 9285 45,894 30,506 0.20 0.66
S55 12 50 420 500 140 9285 48,042 31,642 0.19 0.66
S56 12 50 420 500 150 9285 50,262 32,704 0.18 0.65
S57 12 50 420 500 160 9285 52,483 33,787 0.18 0.64
S58 12 50 420 500 170 9285 54,772 34,843 0.17 0.64
S59 12 50 420 500 180 9285 56,999 35,879 0.16 0.63
S60 12 50 420 500 190 9285 59,220 37,143 0.16 0.63
S61 30 20 250 410 60 16,230 31,945 27,325 0.51 0.86
S62 30 20 260 410 60 16,490 32,143 27,812 0.51 0.87
S63 30 20 280 410 60 16,889 33,282 27,890 0.51 0.84
S64 30 20 300 430 60 17,868 35,165 29,888 0.51 0.85
S65 30 20 310 430 60 18,179 35,999 30,449 0.50 0.85
S66 30 20 320 430 60 18,608 36,805 30,867 0.51 0.84
S67 30 20 330 450 60 19,167 37,613 31,645 0.51 0.84
S68 30 20 340 450 60 19,535 38,386 32,046 0.51 0.83
S69 30 20 350 450 60 20,038 39,165 32,451 0.51 0.83
S70 30 20 360 450 60 20,438 39,847 32,716 0.51 0.82
S71 30 20 380 480 60 21,632 41,410 34,199 0.52 0.83
S72 30 20 400 480 60 22,455 43,088 34,976 0.52 0.81
S73 30 20 420 500 60 23,551 44,697 36,464 0.53 0.82
(continued on next page)

275
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

Table 2 (continued)

Specimens t (mm) D/t f y (MPa) fu (MPa) fc′ (MPa) Ps.FE (kN) Pu.FE (kN) Pr.FE (kN) δ=
Ps .FE Pr .FE
Pu .FE Pu .FE

S74 30 20 250 410 120 16,230 43,065 34,425 0.38 0.80


S75 30 20 260 410 120 16,490 43,679 34,895 0.38 0.80
S76 30 20 280 410 120 16,889 45,195 35,964 0.37 0.80
S77 30 20 300 430 120 17,868 46,991 37,308 0.38 0.79
S78 30 20 310 430 120 18,179 47,742 37,844 0.38 0.79
S79 30 20 320 430 120 18,608 48,521 38,259 0.38 0.79
S80 30 20 330 450 120 19,167 49,311 39,174 0.39 0.79
S81 30 20 340 450 120 19,535 50,114 39,633 0.39 0.79
S82 30 20 350 450 120 20,038 50,875 40,148 0.39 0.79
S83 30 20 360 450 120 20,438 51,550 40,673 0.40 0.79
S84 30 20 380 480 120 21,632 53,257 42,334 0.41 0.79
S85 30 20 400 480 120 22,455 54,768 43,345 0.41 0.79
S86 30 20 420 500 120 23,551 56,349 44,990 0.42 0.80
S87 30 20 250 410 190 16,230 56,685 39,967 0.29 0.71
S88 30 20 260 410 190 16,490 57,477 40,651 0.29 0.71
S89 30 20 280 410 190 16,889 59,038 41,968 0.29 0.71
S90 30 20 300 430 190 17,868 60,602 43,215 0.29 0.71
S91 30 20 310 430 190 18,179 61,400 43,907 0.30 0.72
S92 30 20 320 430 190 18,608 62,163 44,523 0.30 0.72
S93 30 20 330 450 190 19,167 62,953 45,441 0.30 0.72
S94 30 20 340 450 190 19,535 63,740 45,922 0.31 0.72
S95 30 20 350 450 190 20,038 64,487 46,557 0.31 0.72
S96 30 20 360 450 190 20,438 65,291 46,883 0.31 0.72
S97 30 20 380 480 190 21,632 66,864 49,145 0.32 0.73
S98 30 20 400 480 190 22,455 68,402 50,040 0.33 0.73
S99 30 20 420 500 190 23,551 69,842 51,728 0.34 0.74

from the experiments, Pu.FE represents the peak load obtained from the on the residual strength ratio, ductility, column diameter and axial
axisymmetric model and Es denotes the elastic modulus of steel com- load-strain behaviour of ultra-high strength circular short CFST col-
ponent. It can be seen that the axisymmetric model can accurately umns. Table 2 presents the geometric and material properties of 99
predict the ultimate load-carrying capacity of ultra-high strength cir- columns used in this parametric study with a yield stress f y varying
cular short CFST columns with the mean and standard deviation values from 250 MPa to 420 MPa, compressive strength fc′ varying from
of 1.00 and 0.03, respectively. 60 MPa to 190 MPa, and D / t ratio varying from 20 to 50. The column
diameter is taken as 600 mm, whilst the column length L is taken as
1800 mm. The steel elastic modulus of 200 GPa was used for all ax-
4.2. Load-strain responses of axially compressed CFST short columns
isymmetric analyses of the parametric study.
The axial load-strain behaviours obtained from the axisymmetric
5.1. Residual strength ratio
model and measured in the experiment [3] are compared in Fig. 8. As
appear from the figure, the axisymmetric model results in good agree-
The residual strength ratio can be utilized for describing the ducti-
ment with the experimental results for ultra-high strength circular CFST
lity of ultra-high strength circular short CFST members under axial
stub columns. The initial stiffness predicted by the axisymmetric model
loads. The residual strength Pr.FE is predicted from the load-strain re-
is somewhat stiffer than the experimental one for specimens C10, C11,
sponse at the shortening limit ϕlimit which is taken as 15f y L/ Es [1] with L
C13 and C14. This might be due to the use of the average concrete
being the column length. To ensure the ductility of the columns, the
strength in the axisymmetric simulation. It points out that the concrete
residual strength ratio should be maintained above 0.7 [1]. The residual
cube tested is rather nonuniform in strength properties in spite of the
strength ratio depends on the steel contribution ratio, concrete com-
care taken to measure the concrete strength and column capacity on the
pressive strength and D / t ratio. The steel contribution ratio was de-
same day [3]. It can be concluded that discrepancies of this magnitude
termined by taking the ratio of steel tube strength (Ps.FE) to the max-
cannot be eliminated in practical cases, and the design derived from
imum load (Pu.FE) of CFST columns [22].
this axisymmetric analysis must allow for them.
The influences of steel contribution, concrete strength and steel
For specimens C10, C15, C16, C17 and C18, the numerical results
yield stress on the residual strength ratio are illustrated in Fig. 10. It can
differ from the experimental ones in the post-peak range as shown in
be seen that the residual strength increases with an increase in the steel
Fig. 8. It was found that the different friction angle and Drucker-Prager
contribution ratio (Fig. 10a), but it slightly decreases as the concrete
parameter affects the predictions of axial load in the post-peak range.
strength increases (Fig. 10b). At the same concrete strength level, de-
The effects of friction angle on specimens C10, C15, C16, C17 and C18
creasing the D / t ratio, which corresponds to an increase in the steel
are illustrated in Fig. 9. It can be seen from the figure that the axial load
contribution ratio, can lead to a higher ductility (Fig. 10b). It should be
in the post-peak region increases with an increase in the friction angle.
noted that both ultimate axial strength Pu . FE and residual strength Pr . FE
The friction angle less than 20° improves the comparison between ex-
increase with an increase in the steel yield strength f y as shown in
perimental and numerical results in the post-peak region for specimens
Fig. 11. Therefore, the residual strength ratio Pr . FE / Pu . FE is not affected
C10, C15, C16, C17 and C18. For the parametric study, the 20° friction
by the steel yield strength f y as shown in Fig. 10(c). However, the high
angle was used.
strength steel should be utilized to achieve the confining pressure for
UHSC for maintaining the sufficient ductility of CFST columns. As a
5. Parametric study general remark, for ultra-high strength CFST columns, a smaller D/t
ratio (less than 30) and a higher steel contribution ratio (greater than
The parametric study based on the verified axisymmetric model is 0.3) could lead to a sufficient ductility. This confirms the experimental
undertaken to examine the effects of material and geometric parameters observation reported by Liew and Xiong [1]. It should be noted that AS/

276
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

1.00 60000

0.90
Residual strength ratio PFE,r /PFE,u

50000
0.80

Residual strength Pr.FE (kN)


D/t=20 D/t=17
Meet ductility requirement
0.70 40000
D/t=190
0.60 D/t=30
D/t=40
0.50 30000 Pr.FE=26221 kN
D/t=50 Pr.FE=36774 kN
0.40 D/t=60
20000
0.30 D/t=70 Concrete strength = 60 MPa
D/t=130
0.20 10000 Concrete strength = 120 MPa

0.10 Concrete strength = 190 MPa


0
0.00 0 100 200 300 400 500
0.00 0.10 0.20 0.30 0.40
Steel yield strength fy (MPa)
Steel contribution ratio ɷ
(a) Effect of steel yield stress on residual strength
(a) Effect of steel contribution ratio
80000
1

0.9 70000

Ultimate axial strength Pu.FE (kN)


Residual strength ratio PFE,r /PFE,u

0.8 60000

0.7 50000
0.6
40000 Pu.FE=31945 kN
0.5
30000 Pu.FE=44697 kN
0.4
Concrete strength = 60 MPa
0.3 Diameter-to-thickness ratio = 20 20000
Concrete strength = 120 MPa
Diameter-to-thickness ratio = 30
0.2 10000
Diameter-to-thickness ratio = 40 Concrete strength = 190 MPa
0.1
Diameter-to-thickness ratio = 50 0
0 100 200 300 400 500
0
0 40 80 120 160 200 240 Steel yield strength fy (MPa)
Concrete compressive strength fc' (MPa) (b) Effect of steel yield stress on ultimate axial strength
(b) Effect of compressive strength of concrete Fig. 11. Effects of steel yield strength on residual and ultimate strengths.

1
Pr.FE /Pu.FE NZS 5100.6:2017 [44] requires the steel contribution ratio in the range
Pr.FE /Pu,FE =36774/44697
0.9 =26221/31945 of 0.2 and 0.9 for CFST columns with HSC. The steel contribution ratio
Residual strength ratio Pr.FE /Pu.FE

=0.82
=0.82 should be increased to 0.3 for the practical application of UHSC in CFST
0.8
columns.
0.7
Residual strength ratio=0.7 for
0.6 ductility requirements 5.2. Column size

0.5
Liew and Xiong [1] recommended that the use of UHSC in CFST
0.4 columns reduces its cross-sectional area, which offers the significant
economic benefits. This recommendation is numerically investigated
0.3 Concrete strength = 60 MPa using a calculation model given by Liang and Fragomeni [22]. For this
0.2 Concrete strength = 120 MPa purpose, the ultimate strength Pu.CAL = 48, 503 kN was determined
using the axisymmetric model for CFST column with 500 mm diameter,
0.1 Concrete strength = 190 MPa 25 mm tube thickness, 190 MPa concrete compressive strength and
0 420 MPa steel yield stress. A calculation equation given by Liang and
0 60 120 180 240 300 360 420 480 Fragomeni [22] is employed to predict diameter of circular short CFST
Steel yield strength fy (MPa) columns as
π
(c) Effect of steel yield stress Pu .CAL = [(γ f ′ + 4.1frp ) (D − 2t )2 + 4γs f y t (D − t )]
4 c c (14)
Fig. 10. Residual strength ratios of ultra-high strength circular short CFST
columns. D −0.1
γs = 1.458( ) (0.9 ⩽ γs ⩽ 1.1)
t (15)

The confining pressure frp can be determined from Eq. (8). The
calculation model given by Liang and Fragomeni [22] is verified later
for its feasibility in evaluating the maximum capacity of circular short

277
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

1200 80000
Steel strength fy = 300 MPa
70000
1000
Steel strength fy = 420 MPa
Column diameter D (mm)

60000
800

Axial load P (kN)


50000

600 40000
D/t = 20
D/t = 30
400 30000

20000 S5 = 60 MPa
200 = 120 MPa
S11
10000
S18 = 190 MPa
0
0 40 80 120 160 200 0
0 0.02 0.04 0.06 0.08 0.1 0.12
Concrete compressive strength fc' (MPa)
Axial strain ε
(a) Effect of compressive strength of concrete (a) Effect of compressive strength of concrete

800 80000

700 = 120 MPa


70000
Column diameter D (mm)

600
60000

Axial load P (kN)


500
50000
= 190 MPa
400
40000
300
30000
200 S1 D/t = 20
20000
Steel strength fy = 300 MPa
100 S2 D/t = 30
Steel strength fy = 420 MPa 10000
S3 D/t = 40
0
0 40 80 120 160 200 240 0
0 0.02 0.04 0.06 0.08 0.1 0.12
Diameter-to-thickness ratio D/t
Axial strain ε
(b) Effect of diameter-to-thickness ratio
(b) Effect of diameter-to-thickness ratio
800
80000
700
= 120 MPa 70000
Column diameter D (mm)

600
= 190 MPa 60000
500
Axial load P (kN)

50000
400
40000
300
30000
200
Diameter-to-thickness ratio = 200 S90 = 300 MPa
20000
100
Diameter-to-thickness ratio = 20 S95 = 350 MPa
10000
0 S99 = 420 MPa
0 100 200 300 400 500
0
Steel yield strength fy (MPa) 0 0.02 0.04 0.06 0.08 0.1 0.12
(c) Effect of yield stress of structural steel tubes Axial strain ε

Fig. 12. Diameter of circular short CFST columns.


(c) Effect of yield stress of structural steel tubes
Fig. 13. Axial load-strain behaviour of circular short CFST columns.
CFST columns with UHSC. Fig. 12(a) demonstrates the column dia-
meter as a function of the concrete compressive strength. It can be seen column using the steel yield stress of 420 MPa and D / t ratio of 20, its
that the diameter of CFST columns can approximately be reduced by diameter can be reduced from 802 mm diameter to 498 mm when the
50% if UHSC is used instead of NSC. For example, for the case of CFST compressive strength of concrete increases from 20 MPa to 190 MPa.

278
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

axisymmetric to the longitudinal axis, only the contour plot of a quarter


section is given. The stress distribution is shown for Specimen S4 given
in Table 2. The CFST column reaches its ultimate load-carrying capacity
at the strain of 0.008. The confined concrete stress at the strain of 0.008
is higher than concrete compressive strength fc′. This is attributed to the
confining mechanism induced by the steel tube. The axial stress in the
concrete then reduces when the strain greater than 0.008. The stress-
(a) İ = 0.002 (b) İ = 0.008
strain behaviours of steel and concrete components in the circular CFST
column are illustrated in Fig. 15. Unlike concrete, the stress in steel is
increased when the axial strain increases greater than 0.013. When the
strain in the concrete component is less than 0.008, the axial stress is
increased with increasing the strain.

6. International standards for constructional practice

(c) İ = 0.015 (d) İ = 0.04 6.1. Design codes


Fig. 14. Axial stress distribution in concrete at different strain levels.
The current design methods limit the material strengths for the
design of CFST columns. The international standards of AS/NZS 5100.6
600
[44], Eurocode 4 [16], AISC [45] and GB 50956 [46] are used to check
their feasibility for ultra-high strength short CFST columns. These de-
500 sign models are compared to the FE predictions and the test data given
by Xiong et al. [3].
Axial stress σ (MPa)

400
6.1.1. AS/Nzs 5100.6
AS/NZS 5100.6 [44] for the composite bridges and buildings is re-
300 vised in 2017 to allows the use of high strength CFST columns with the
concrete compressive strength up to 100 MPa and steel yield stress up to
Concrete component
200
690 MPa. The effects of the concrete confinements are also included in
AS/NZS 5100.6 [44] as expressed in the following equation

100 η1 tf y ⎞
Pu .AS = φkf As η2 f y + φAr fry + φc Ac fc′ ⎛⎜1 + ⎟
⎝ do fc′ ⎠ (16)
0 in which φ represents the capacity reduction factor for steel, kf is the
0 0.01 0.02 0.03 0.04 0.05
form factor, As denotes the area of steel component, Ar represents the
Axial strain ε reinforcement area which is taken as zero, f yr stands for the re-
Fig. 15. Stress-strain responses of steel and concrete in ultra-high strength CFST inforcement yield strength, φc is the capacity factor for concrete in
columns. compression, Ac is the concrete area, do represents the column dia-
meter, which is taken as D in this study, and η1 and η2 are coefficients
which are determined from the following equations:
The effect of D / t ratio on the column diameter is not significant as
shown in Fig. 12(b). The column diameter as a function of the steel η1 = 4.9 − 18.5λr + 17(λr )2 (η1 ⩾ 0) (17)
yield stress is represented in Fig. 12(c). As shown in the figure, the
column diameter slightly decreases when the steel yield strength in- η2 = 0.25(3 + 2λr ) (η2 ⩽ 1) (18)
creases. This study revealed that the column size should be reduced by where λr is the column slenderness, which is determined by the fol-
increasing the concrete compressive strength rather than D / t ratio and lowing equations:
steel yield strength. However, the D / t ratio should be smaller than 30 to
satisfy the ductility requirement. The high strength steel can be used to Nus
λr =
increase the confining pressure leading to increasing the ductility. Ncr (19)

Nus = As f y + Ac fc′ (20)


5.3. Axial load-strain behaviour

π 2 (EI )e ⎞
The effects of D / t ratio, concrete compressive strength and steel Ncr = ⎜⎛ 2 ⎟

yield stress on the axial load-strain behaviour of circular short CFST ⎝ Le ⎠ (21)
columns are illustrated in Fig. 13. It can be seen that the load-carrying
(EI )e = EIs + EIr + Ec Ic (22)
capacity of circular short CFST columns increases with increasing
concrete compressive strength and steel yield stress as expected. in which Nus represents the cross-sectional maximum strength, Ncr de-
However, it decreases as D / t ratio increases. The initial axial stiffness is notes the critical elastic load, (EI )e stands for the effective elastic flex-
not affected by D / t ratio and steel yield stress. It is also observed that ural stiffness, Le stands for the effective length, E represents the elastic
load-strain curves exhibit the strain softening behaviour due to the modulus for the steel and reinforcement, Is denotes the steel second
brittleness of UHSC. moment of area, and Ir represents the reinforcement second moment of
area. In Eq. (16), the values of φ , φc and kf are taken as 1.0. All para-
5.4. Stress distribution meters related to the reinforcement are also taken as zero.

Fig. 14 displays the distribution of the concrete axial normal stresses 6.1.2. Eurocode 4
for CFST section at different strain levels. Since the analysis results are Eurocode 4 [16] allows the design of CFST columns with the limited

279
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

Table 3
Comparison of the peak loads obtained from axisymmetric model and design codes.
Specimens AS/NZS 5100.6 [44] Eurocode 4 [16] AISC [45] GB 50936 [46] Design model [22]

Pu.AS (kN) Pu .AS Pu.EC4 (kN) Pu .EC4 Pu.AISC (kN) Pu .AISC Pu.GB (kN) Pu .GB Pu.CAL (kN) Pu .CAL
Pu .FE Pu .FE Pu .FE Pu .FE Pu .FE

S1 70,237 1.05 69,586 1.04 61,753 0.92 73,811 1.10 66,761 1.00
S2 65,070 1.04 64,414 1.03 58,305 0.94 74,292 1.19 61,483 0.99
S3 62,326 1.05 61,710 1.04 56,535 0.95 73,057 1.23 58,378 0.98
S4 60,637 1.05 60,070 1.04 55,458 0.96 71,931 1.25 56,533 0.98
S5 44,836 1.00 44,408 0.99 35,617 0.80 29,302 0.66 44,588 1.00
S6 46,968 1.01 46,508 1.00 37,793 0.81 32,526 0.70 46,535 1.00
S7 49,104 1.01 48,615 1.00 39,969 0.82 35,801 0.74 48,482 1.00
S8 51,246 1.01 50,730 1.00 42,144 0.83 39,110 0.77 50,428 1.00
S9 53,393 1.02 52,851 1.01 44,320 0.85 42,442 0.81 52,375 1.00
S10 55,545 1.02 54,978 1.01 46,496 0.85 45,792 0.84 54,322 1.00
S11 57,701 1.02 57,111 1.01 48,671 0.86 49,154 0.87 56,268 1.00
S12 59,862 1.03 59,250 1.02 50,847 0.87 52,527 0.90 58,215 1.00
S13 62,027 1.03 61,395 1.02 53,023 0.88 55,907 0.93 60,162 1.00
S14 64,196 1.03 63,544 1.02 55,199 0.89 59,293 0.96 62,108 1.00
S15 66,369 1.04 65,698 1.03 57,374 0.90 62,684 0.98 64,055 1.00
S16 68,546 1.04 67,856 1.03 59,550 0.90 66,080 1.00 66,002 1.00
S17 70,725 1.04 70,018 1.03 61,726 0.91 69,479 1.02 67,948 1.00
S18 72,908 1.04 72,185 1.03 63,901 0.91 72,881 1.04 69,895 1.00
S19 36,322 0.97 35,869 0.96 29,345 0.79 30,360 0.81 36,552 0.98
S20 38,644 0.98 38,159 0.97 31,685 0.80 33,707 0.86 38,645 0.98
S21 40,972 0.98 40,459 0.97 34,025 0.82 37,073 0.89 40,739 0.98
S22 43,307 0.99 42,767 0.98 36,365 0.83 40,453 0.92 42,832 0.98
S23 45,647 1.00 45,083 0.98 38,704 0.84 43,842 0.96 44,926 0.98
S24 47,993 1.00 47,406 0.99 41,044 0.86 47,238 0.99 47,020 0.98
S25 50,344 1.01 49,736 1.00 43,384 0.87 50,639 1.01 49,113 0.98
S26 52,700 1.01 52,072 1.00 45,724 0.88 54,045 1.04 51,207 0.98
S27 55,061 1.02 54,413 1.00 48,064 0.89 57,453 1.06 53,300 0.98
S28 57,425 1.03 56,760 1.02 50,404 0.90 60,864 1.09 55,394 0.99
S29 59,793 1.03 59,111 1.01 52,744 0.90 64,276 1.10 57,487 0.99
S30 62,165 1.03 61,467 1.02 55,083 0.91 67,691 1.12 59,581 0.99
S31 64,540 1.03 63,827 1.02 57,423 0.92 71,107 1.14 61,674 0.99
S32 66,918 1.03 66,191 1.02 59,763 0.92 74,523 1.15 63,768 0.99
S33 31,758 0.96 31,314 0.95 26,123 0.79 29,159 0.88 31,988 0.97
S34 34,186 0.97 33,713 0.96 28,548 0.81 32,543 0.93 34,157 0.97
S35 36,621 0.98 36,122 0.97 30,972 0.83 35,937 0.96 36,326 0.97
S36 39,063 0.99 38,540 0.97 33,396 0.84 39,339 0.99 38,495 0.97
S37 41,510 0.99 40,966 0.98 35,820 0.86 42,746 1.02 40,664 0.97
S38 43,964 1.00 43,400 0.99 38,244 0.87 46,156 1.05 42,833 0.98
S39 46,422 1.01 45,839 1.00 40,668 0.88 49,569 1.08 45,002 0.98
S40 48,885 1.01 48,285 1.00 43,093 0.89 52,985 1.10 47,171 0.98
S41 51,351 1.02 50,736 1.01 45,517 0.90 56,402 1.12 49,340 0.98
S42 53,822 1.02 53,191 1.01 47,941 0.91 59,820 1.14 51,509 0.98
S43 56,296 1.03 55,651 1.02 50,365 0.92 63,239 1.15 53,678 0.98
S44 58,773 1.03 58,115 1.02 52,789 0.93 66,659 1.17 55,847 0.98
S45 61,254 1.04 60,583 1.03 55,213 0.94 70,080 1.19 58,016 0.98
S46 63,737 1.04 63,054 1.03 57,638 0.94 73,502 1.20 60,185 0.98
S47 28,923 0.95 28,499 0.94 24,163 0.80 27,985 0.92 29,260 0.97
S48 31,420 0.97 30,970 0.95 26,638 0.82 31,385 0.96 31,475 0.97
S49 33,924 0.97 33,451 0.96 29,114 0.84 34,792 1.00 33,690 0.97
S50 36,434 0.98 35,940 0.97 31,589 0.85 38,203 1.03 35,905 0.97
S51 38,950 1.00 38,438 0.98 34,065 0.87 41,617 1.06 38,120 0.98
S52 41,472 1.00 40,942 0.99 36,540 0.88 45,034 1.09 40,335 0.97
S53 43,998 1.01 43,452 0.99 39,016 0.89 48,452 1.11 42,550 0.97
S54 46,528 1.01 45,967 1.00 41,491 0.90 51,872 1.13 44,764 0.98
S55 49,062 1.02 48,488 1.01 43,967 0.92 55,293 1.15 46,979 0.98
S56 51,599 1.03 51,013 1.01 46,442 0.92 58,714 1.17 49,194 0.98
S57 54,140 1.03 53,541 1.02 48,918 0.93 62,136 1.18 51,409 0.98
S58 56,683 1.03 56,074 1.02 51,393 0.94 65,559 1.20 53,624 0.98
S59 59,229 1.04 58,610 1.03 53,869 0.95 68,982 1.21 55,839 0.98
S60 61,778 1.04 61,149 1.03 56,344 0.95 72,406 1.22 58,054 0.98
S61 32,889 1.03 32,656 1.02 26,485 0.83 30,202 0.95 31,268 0.98
S62 33,612 1.05 33,368 1.04 27,022 0.84 30,316 0.94 32,052 1.00
S63 35,050 1.05 34,784 1.05 28,096 0.84 30,482 0.92 33,619 1.01
S64 36,477 1.04 36,188 1.03 29,171 0.83 30,563 0.87 35,186 1.00
S65 37,187 1.03 36,887 1.02 29,708 0.83 30,573 0.85 35,969 1.00
S66 37,894 1.03 37,583 1.02 30,245 0.82 30,562 0.83 36,753 1.00
S67 38,598 1.03 38,276 1.02 30,782 0.82 30,530 0.81 37,536 1.00
S68 39,301 1.02 38,967 1.02 31,319 0.82 30,477 0.79 38,320 1.00
S69 40,001 1.02 39,655 1.01 31,857 0.81 30,403 0.78 39,104 1.00
S70 40,698 1.02 40,341 1.01 32,394 0.81 30,308 0.76 39,887 1.00
S71 42,086 1.02 41,706 1.01 33,468 0.81 30,056 0.73 41,454 1.00
(continued on next page)

280
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

Table 3 (continued)

Specimens AS/NZS 5100.6 [44] Eurocode 4 [16] AISC [45] GB 50936 [46] Design model [22]

Pu.AS (kN) Pu .AS Pu.EC4 (kN) Pu .EC4 Pu.AISC (kN) Pu .AISC Pu.GB (kN) Pu .GB Pu.CAL (kN) Pu .CAL
Pu .FE Pu .FE Pu .FE Pu .FE Pu .FE

S72 43,466 1.01 43,062 1.00 34,543 0.80 29,721 0.69 43,021 1.00
S73 44,836 1.00 44,408 0.99 35,617 0.80 29,302 0.66 44,588 1.00
S74 45,950 1.07 45,616 1.06 39,539 0.92 50,512 1.17 42,948 1.00
S75 46,657 1.07 46,309 1.06 40,076 0.92 50,606 1.16 43,732 1.00
S76 48,065 1.06 47,687 1.06 41,150 0.91 50,728 1.12 45,299 1.00
S77 49,465 1.05 49,058 1.04 42,225 0.90 50,763 1.08 46,866 1.00
S78 50,162 1.05 49,740 1.04 42,762 0.90 50,748 1.06 47,650 1.00
S79 50,857 1.05 50,419 1.04 43,299 0.89 50,711 1.05 48,433 1.00
S80 51,549 1.05 51,097 1.04 43,837 0.89 50,653 1.03 49,217 1.00
S81 52,240 1.04 51,773 1.03 44,374 0.89 50,573 1.01 50,000 1.00
S82 52,930 1.04 52,447 1.03 44,911 0.88 50,472 0.99 50,784 1.00
S83 53,617 1.04 53,119 1.03 45,448 0.88 50,349 0.98 51,567 1.00
S84 54,985 1.03 54,457 1.02 46,523 0.87 50,037 0.94 53,134 1.00
S85 56,347 1.03 55,788 1.02 47,597 0.87 49,639 0.91 54,701 1.00
S86 57,701 1.02 57,111 1.01 48,671 0.86 49,154 0.87 56,268 1.00
S87 61,385 1.08 60,967 1.08 54,769 0.97 74,408 1.31 56,575 1.00
S88 62,075 1.08 61,640 1.07 55,306 0.96 74,494 1.30 57,359 1.00
S89 63,452 1.07 62,981 1.07 56,380 0.95 74,599 1.26 58,926 1.00
S90 64,821 1.07 64,315 1.06 57,455 0.95 74,618 1.23 60,493 1.00
S91 65,504 1.07 64,980 1.06 57,992 0.94 74,594 1.21 61,276 1.00
S92 66,185 1.06 65,642 1.06 58,529 0.94 74,548 1.20 62,060 1.00
S93 66,864 1.06 66,304 1.05 59,066 0.94 74,480 1.18 62,843 1.00
S94 67,542 1.06 66,963 1.05 59,604 0.94 74,390 1.17 63,627 1.00
S95 68,218 1.06 67,621 1.05 60,141 0.93 74,278 1.15 64,410 1.00
S96 68,892 1.06 68,278 1.05 60,678 0.93 74,145 1.14 65,194 1.00
S97 70,237 1.05 69,586 1.04 61,753 0.92 73,811 1.10 66,761 1.00
S98 71,575 1.05 70,888 1.04 62,827 0.92 73,390 1.07 68,328 1.00
S99 72,908 1.04 72,185 1.03 63,901 0.91 72,881 1.04 69,895 1.00

Mean 1.03 1.02 0.88 1.02 0.99


SD 0.03 0.03 0.05 0.16 0.01
CoV 0.03 0.03 0.06 0.15 0.01

Table 4
Comparison of the peak loads obtained from design codes and test [3].
Specimens Ecm [3] (GPa) Pu. exp [3] (kN) AS/NZS 5100.6 [44] Eurocode 4 [16] AISC [45] GB 50936 [46] Design model [22]

Pu.AS (kN) Pu .AS Pu.EC4 (kN) Pu .EC4 Pu.AISC (kN) Pu .AISC Pu.GB (kN) Pu .GB Pu.CAL (kN) Pu .CAL
Pu . exp Pu . exp Pu . exp Pu . exp Pu . exp

C3 63 2422 2297 0.95 2277 0.94 1989 0.82 2491 1.03 2337 0.96
C4 63 2340 2297 0.98 2277 0.97 1989 0.85 2491 1.06 2337 1.00
C5 63 2497 2389 0.96 2369 0.95 2081 0.83 2624 1.05 2432 0.97
C6 63 2314 2389 1.03 2369 1.02 2081 0.90 2624 1.13 2432 1.05
C7 63 2610 2706 1.04 2685 1.03 2254 0.86 2367 0.91 2710 1.04
C8 63 2633 2706 1.03 2685 1.02 2254 0.86 2367 0.90 2710 1.03
C10 66 7837 8106 1.03 8030 1.02 7316 0.93 9457 1.21 7831 1.00
C11 67 8664 8380 0.97 8303 0.96 7584 0.88 9832 1.13 8084 0.93
C13 66 9085 9162 1.01 9080 1.00 7978 0.88 9693 1.07 8925 0.98
C14 67 9187 9406 1.02 9323 1.01 8220 0.89 10,066 1.10 9157 1.00
C15 62 6915 7213 1.04 7152 1.03 6450 0.93 8457 1.22 6948 1.00
C16 58 7407 7605 1.03 7545 1.02 6844 0.92 9023 1.22 7322 0.99
C17 54 6838 6738 0.99 6682 0.98 5998 0.88 7809 1.14 6520 0.95
C18 56 7569 7571 1.00 7512 0.99 6816 0.90 8982 1.19 7295 0.96

Mean 1.01 1.00 0.88 1.10 0.99


SD 0.03 0.03 0.04 0.10 0.03
CoV 0.03 0.03 0.04 0.09 0.03

compressive strength and yield stress of 60 MPa and 460 MPa, respec- Npl.Rk
λ¯ =
tively. Eurocode 4 [16] considers the confinement effect of concrete in Ncr (26)
the following equations
Npl.Rk = As f y + Ac fc′ (27)
t fy ⎤
Pu . EC 4 = ηa As f y + Ac fc′ ⎡
⎢1 + ηc d f ′ ⎥
⎣ c⎦ (23) π 2 (EI )eff
Ncr =
L2 (28)
ηa = 0.25(3 + 2λ¯ ) ⩽ 1.0 (24)
(EI )eff = Es Is + 0.6Ecm Ic (29)
ηc = 4.9 − 18.5λ¯ + 17λ¯ 2 ≥ 0 (25) where (EI )eff denotes the effective flexural stiffness, ηa and ηc denotes

281
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

the concrete confining factors, λ̄ represents the column slenderness, comparison between the design codes and test results as shown in
Npl.Rk denotes the section plastic strength, and Ncr represents the elastic Table 4. It can further concluded that current AS/NZS 5100.6 and
critical normal force. Eurocode 4 can accurately predict the ultimate strength of ultra-high
strength circular short CFST. In addition, the design equations by Liang
6.1.3. American code AISC and Fragomeni [22] can also be used as the calculation model for these
The recommendation given in AISC [45] does not consider the columns.
composite actions in CFST columns. The ultimate load of composite
column is obtained by adding the individual strength of steel-concrete 7. Conclusions
components. This means that AISC [45] ignores the influences of the
confinement mechanism on the axial capacity of circular CFST columns. An efficient axisymmetric simulation for circular short CFST col-
The material limitations applied by AISC [45] in the design are 69 MPa umns with UHSC has been presented in this paper. The axisymmetric
concrete and 525 MPa steel. The equation enclosed in the AISC [45] is simulation considers the influences of initial geometric imperfection,
given as: material and geometric nonlinearities. The developed axisymmetric
models were verified by the independent test data. The axisymmetric
Pu .AISC = As f y + C2 fc′ Ac (30) results agree with the experimental results including elastic stiffness,
in which C2 represents the safety factor specified in American code AISC peak load and load-strain response. The ratio of the peak loads pre-
[45], which is taken as 0.95. dicted by the axisymmetric model and experiment is in the range from
0.94 to 1.04 with the mean value of 1.00 and SD value of 0.03. The use
6.1.4. Chinese code GB 50936-2014 of axisymmetric model can significantly save computational time (over
The design formula recommended in the technical specification GB 19 times) when compared with the full 3D model.
50936-2014 code [46] also considers the increased strength from the A parametric study undertaken using the validated axisymmetric
confining mechanism in circular CFST columns. The equation given in model reveals that the steel contribution ratio greater than 0.3 and D / t
GB 50936-2014 [46] is expressed by: ratio less than 30 should be maintained to ensure sufficient ductility of
CFST columns with UHSC. The residual strength ratio increases as the
Pu .GB = Asc fsc (31) steel contribution ratio increases, but it decreases as concrete com-
pressive strength increases. The application of UHSC in CFST column
fsc = (1.212 + Bθ + Cθ 2) fc (32) approximately reduces the cross-sectional size by 50% when compared
As f y with the use of NSC. The peak load increases as concrete compressive
θ= strength and steel yield stress increase, but it decreases as D / t ratio
Ac fc′ (33) increases.
0.176 The design equations in the existing code provisions were utilized to
B= + 0.974 evaluate the peak loads of ultra-high strength short circular CFST col-
fy (34)
umns. It was found that the Australian/New Zealand, European and
0.104fc′ Liang and Fragomeni’s design equations provide consistent predictions
C=− + 0.031 of the peak loads of ultra-high strength circular short CFST columns.
14.4 (35)
The North American design code is conservative, whereas the Chinese
in which Asc represents the cross-sectional area of a CFST column, fsc design code is unconservative.
denotes the compressive design strength of the confining concrete in
CFST column, B and C are influential factors of cross-sectional shapes Acknowledgements
on confinement effect and θ stands for the confinement factor.
This work is supported by the School of Engineering and
6.1.5. Liang and Fragomeni’s [22] model Mathematical Sciences at La Trobe University with the startup research
A design question given by Liang and Fragomeni [22] is developed fund. This financial support is gratefully acknowledged.
using the regression analysis of the test data and numerical results of
circular short CFST columns with HSC. This equation accurately pre- Appendix
dicts the ultimate strengths of CFST columns with 120 MPa concrete
and 690 MPa steel. Liang and Fragomeni [22] design concept is ad- The Python script, Matlab code and numerical results in excel
ditionally checked here for ultra-high strength CFST columns. The de- format can be downloaded from the following link:
sign concept presented by Laing and Fragomeni [22] considers the https://figshare.com/s/afb3cc2f4cd7d831ed8b.
confinement mechanism and steel strain hardening as depicted in Eq.
(14). References

6.2. Design verification [1] Liew JYR, Xiong DX. Ultra-high strength concrete filled composite columns for
multi-storey building construction. Adv Struct Eng 2012;15:1487–503.
[2] Liew JYR, Xiong MX, Xiong DX. Design of concrete filled tubular beam-columns
The peak loads of ultra-high strength short circular CFST columns with high strength steel and concrete. Structures 2016;8:213–26.
predicted by design codes and design model are compared with those [3] Xiong MX, Xiong DX, Liew JYR. Axial performance of short concrete filled steel
obtained from the axisymmetric analysis in Table 3 and experimental tubes with high- and ultra-high- strength materials. Eng Struct 2017;136:494–510.
[4] O’Shea MD, Bridge RQ. Design of circular thin-walled concrete filled steel tubes. J
tested by Xiong [3] in Table 4. As shown in Table 3, the AS/NZS 5100.6 Struct Eng 2000;126:1295–303.
and Eurocode 4 approaches and design model provide very close pre- [5] Sakino K, Nakahara H, Morino S, Nishiyama I. Behavior of centrally loaded con-
dictions of peak loads compared with the axisymmetric analysis with crete-filled steel-tube short columns. J Struct Eng 2004;130:180–8.
[6] Giakoumelis G, Lam D. Axial capacity of circular concrete-filled tube columns. J
mean values from 0.99 to 1.03 and SD values less than 0.03. The AISC
Constr Steel Res 2004;60:1049–68.
approach gives a conservative prediction of ultra-high strength circular [7] Han LH, Huang H, Tao Z, Zhao XL. Concrete-filled double skin steel tubular (CFDST)
short CFST columns with a mean value of 0.88, whilst the GB 50936 beam-columns subjected to cyclic bending. Eng Struct 2006;28:1698–714.
[8] Varma AH, Ricles JM, Sause R, Liu LW. Seismic behavior and design of high-
method provides an unconservative design with a mean value of 1.02
strength square concrete-filled steel tube beam columns. J Struct Eng
with the SD value of 0.16. The recommendations about the code pre- 2004;130:169–79.
diction of CFST columns with UHSC are also supported by the [9] Wang Q, Zhao D, Guan P. Experimental study on the strength and ductility of steel

282
V.I. Patel et al. Engineering Structures 179 (2019) 268–283

tubular columns filled with steel-reinforced concrete. Eng Struct 2004;26:907–15. 2016;16(2):1–16.
[10] Elchalakani M, Zhao XL, Grzebieta R. Concrete-filled steel circular tubes subjected [28] Wang ZB, Tao Z, Han LH, Uy B, Lam D, Kang WH. Strength, stiffness and ductility of
to constant amplitude cyclic pure bending. Eng Struct 2004;26:2125–35. concrete-filled steel columns under axial compression. Eng Struct
[11] Young B, Ellobody E. Experimental investigation of concrete-filled cold-formed high 2017;135:209–21.
strength stainless steel tube columns. J Constr Steel Res 2006;62:484–92. [29] Al Abadi H, Abo El-Naga H, Shaia H, Paton-Cole V. Refined approach for modelling
[12] Gupta PK, Sarda SM, Kumar MS. Experimental and computational study of concrete strength enhancement of FRP-confined concrete. Constr Build Mater
filled steel tubular columns under axial loads. J Constr Steel Res 2007;63:182–93. 2016;119:152–74.
[13] Uy B, Tao Z, Han LH. Behaviour of short and slender concrete-filled stainless steel [30] Wang YB, Liew JYR. Constitutive model for confined ultra-high strength concrete in
tubular columns. J Constr Steel Res 2011;67:360–78. steel tube. Constr Build Mater 2016;126:812–22.
[14] Khan M, Uy B, Tao Z, Mashiri F. Concentrically loaded slender square hollow and [31] EN 1992-1-1. Eurocode 2 – design of concrete structures. Part 1-1: General rules and
composite columns incorporating high strength properties. Eng Struct rules for buildings. European Committee for Standardization (CEN); 2004.
2017;131:69–89. [32] EN 1992-1-2, Eurocode 2 – design of concrete structures. Part 1-2: General rules-
[15] Han LH, Li W, Bjorhovde R. Development and advanced applications of concrete- structural fire design. European Committee for Standardization (CEN); 2004.
filled steel tubular (CFST) structures: members. J Constr Steel Res [33] Le Hoang A, Fehling E. Analysis of circular steel tube confined UHPC stub columns.
2014;100:211–28. Steel Compo Struct 2017;23:669–82.
[16] EN 1994-1-1. Eurocode 4 – design of composite steel and concrete structures, part 1. [34] Le Hoang A, Fehling E. Numerical analysis of circular steel tube confined UHPC stub
1: general rules and rules for building. London (UK): British Standards Institution; columns. Comput Concr 2017;19:263–73.
2004. [35] Patel VI, Uy B, Prajwal KA, Aslani F. Confined concrete model of circular, elliptical
[17] AS 3600-2009. Concrete structures, Sydney, New South Wales, Australia: Standards and octagonal CFST short columns. Steel Compo Struct 2016;22:497–520.
Australia; 2009. [36] Mander JB, Priestly MNJ, Park R. Theoretical stress-strain model for confined
[18] Portolés JM, Serra E, Romero ML. Influence of ultra-high strength infill in slender concrete. J Struct Eng 1988;114:1804–26.
concrete-filled steel tubular columns. J Constr Steel Res 2013;86:107–14. [37] American Concrete Institute. Building code requirements for structural concrete
[19] Xiong MX, Xiong DX, Liew JYR. Behaviour of steel tubular members infilled with (ACI 318-11) and commentary. MI, USA: Farmington Hills; 2011.
ultra high strength concrete. J Constr Steel Res 2017;138:168–83. [38] Liang QQ. Performance-based analysis of concrete-filled steel tubular beam-col-
[20] Hajjar JF, Molodan A, Schiller PH. A distributed plasticity model for cyclic analysis umns, I. Part Theory and algorithms. J Constr Steel Res 2009;65:363–72.
of concrete-filled steel tube beam-columns and composite frames. Eng Struct [39] Richart FE, Bradtzaeg A, Brown RL. A study of the failure of concrete under com-
1998;20:398–412. bined compressive stresses. Bull. Champaign (III): University of Illionis, Engineering
[21] Shanmugam NE, Lakshmi B, Uy B. An analytical model for thin-walled steel box Experimental Station; 1928. p. 185.
columns with concrete in-fill. Eng Struct 2002;24:825–38. [40] Dai X, Lam D. Numerical modelling of the axial compressive behaviour of short
[22] Liang QQ, Fragomeni S. Nonlinear analysis of circular concrete-filled steel tubular concrete-filled elliptical steel columns. J Constr Steel Res 2010;66:931–42.
short columns under axial loading. J Constr Steel Res 2009;65:2186–96. [41] Hassanein MF, Patel VI, Bock M. Behaviour and design of hexagonal concrete-filled
[23] Bai Y, Kawano A, Odawara K, Matsuo S. Constitutive models for hollow steel tubes steel tubular short columns under axial compression. Eng Struct 2017;153:732–48.
and concrete filled steel tubes considering the strength deterioration. J Struct [42] ABAQUS. ABAQUS standard user’s manual, version 6.13. Providence, RI (USA):
Constr Eng (Trans AIJ) 2012;77(677):1141–50. Dassault Systèmes Corp.; 2013.
[24] Tao Z, Wang ZB, Yu Q. Finite element modelling of concrete-filled steel stub col- [43] Theofanous M, Chan TM, Gardner L. Structural response of stainless steel oval
umns under axial compression. J Constr Steel Res 2013;89:121–31. hollow section compression members. Eng Struct 2009;31:922–34.
[25] Thai HT, Uy B, Khan M, Tao Z, Mashiri F. Numerical modelling of concrete-filled [44] AS/NZS 5100.6. Bridge design Part 6: steel and composite construction, Sydney,
steel box columns incorporating high strength materials. J Constr Steel Res New South Wales, Australia: Standards Australia; 2017.
2014;102:256–95. [45] AISC 360-16. Specification for structural steel buildings. Chicago (IL, USA):
[26] Hassanein MF, Kharoob OF, Gardner L. Behaviour and design of square concrete- American Institute of Steel Construction; 2016.
filled double skin tubular columns with inner circular tubes. Eng Struct [46] GB 50936-2014. Technical code for concrete filled steel tubular structures. Beijing
2015;100:410–24. (China): Ministry of Housing and Urban-Rural Development of the People’s
[27] Bai Y, Lin X, Mou B. Numerical modeling on post-local buckling behavior of circular Republic of China, (MOHURD); 2014 [in Chinese].
and square concrete filled steel tubular beam columns. Int J Steel Struct

283

You might also like