See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/328457104

Analysis of mechanical behaviour of circular concrete filled steel tube

columns using high strength concrete

Article · October 2018

CITATIONS READS

5 316

1 author:

HAO DINH PHAN

The University of Danang - University of Science and Technology
6 PUBLICATIONS 19 CITATIONS

SEE PROFILE

All content following this page was uploaded by HAO DINH PHAN on 21 September 2020.

The user has requested enhancement of the downloaded file.

 Mechanics of Structures and Materials: Advancements and Challenges – Hao & Zhang (Eds)
© 2017 Taylor & Francis Group, London, ISBN 978-1-138-02993-4

Analysis of mechanical behaviour of circular concrete filled steel tube

columns using high strength concrete

H.D. Phan & H.H. Trinh

Faculty of Civil Engineering, University of Science and Technology—The University of Danang,
Danang City, Vietnam

ABSTRACT: This paper presents a study on mechanical behaviour of circular Concrete Filled Steel
Tube (CFST) columns using high strength concrete ranging between 65 and 85 MPa. The finite element
method was used to simulate short CFST columns under concentric compression loading. These columns
include the hollow steel tube and CFST columns loaded on the concrete core only, the steel tube only, and
the entire column section. The study investigated influences of the bond strength between concrete and
steel as well as the confinement effect on the loading capacity of CFST columns. The numerical results
showed that the mechanical behaviour of the columns depends on the ways to apply loading. In which, the
bond strength has no influence on the behaviour when the steel and concrete sections are loaded simulta-
neously. The bond strength highly affects the confinement effect for the columns loaded on the concrete
core only, which results in the highest loading capacity of these columns.

1 INTRODUCTION ing (Le & Pham 2016). However, these studies

were still decremented and the application of these
Concrete Filled Steel Tube (CFST) columns were structures in Vietnam was still limited. Therefore,
studied and utilized for high-rise buildings and the deeper research about load bearing capacities
long-span bridges in some developed countries such of CFST columns is really necessary, especially in
as United States of America, United Kingdom, the case of using high strength concrete filled.
Australia, Japan, China, Sweden, etc. Therefore, The main purpose of this study is to analyze the
the CFST columns with square and circular sec- mechanical behaviour and estimate the load bear-
tions have been analyzed based on numerical mod- ing capacity of CFST columns under concentric
eling and experimental investigation. Previous compression loading conditions. The concrete filled
studies related to determining the loading capac- in the steel tub was high strength concrete namely
ity, the local stability of steel tubes, flexural behav- 65, 75 and 85 MPa. The steel tubes were chosen in
iour of CFST columns were conducted by some ranging of yield strength from 433 to 511 MPa.
researchers such as: Ge & Usami (1992, 1994); Uy Meanwhile, the study also investigates the distribu-
(1998); Han (2004); Fujimoto et al. (2004); Yu et al. tion of axial internal forces between the concrete
(2007); Han et al. (2008). Most of such studies were core and the steel tube, and estimates the increase
performed by using experimental methods and of compression strength of the concrete core due to
analyzing the mechanical behaviour obtained from the confinement effects offered by the steel tube.
testing specimens. The number of studies based
on numerical modeling was still limited and the
2 MODELING PROGRAM OF CFST
numerical analyses were not very deep due to the
COLUMNS
complex properties of this composite structure.
In Vietnam, there were some studies about
2.1 Finite element modeling
CFST structures such as the theoretical research
and numerical modeling study on analyzing non- Finite Element Method (FEM) was used in order
linear behaviour of CFST structures (Phan et al. to simulate the CFST columns subjected to com-
2012), experimental study on the compressive pressive loading. The mechanical properties of con-
behaviour of short circular CFST columns using stituent materials, the interaction working between
large specimens (Ngo-Huu et al. 2016), and study the steel tube and the concrete core, as well as the
on strengthening of slip resistance between con- increment of concrete compressive strength due to
crete core and steel tube surfaces for slender CFST confinement effects need to be considered carefully
columns subjected to eccentric compression load- in the simulation progress.

231
 It is important to know that the steel tube, the The preliminary tension tests of steel specimens
concrete core and the loading plates are separated were conducted according to Swidish Standard.
objectives and they interact with each other under For the first type of steel tube, average values
loading conditions. In order to exactly simulate of five tensile specimens include: yield strength,
the real working of CFST columns, the constitu- fy = 433 MPa, ultimate strength, fu = 568 MPa, hard-
ent components need to be modeled using appro- enning strain, εah = 0.029, strain at ultimate stress,
priate elements. The solid three-dimension element εau = 0.136, and elastic module, Ea = 206 GPa. Ten-
was established for both steel tube and concrete sile test were carried out similarly for the remain-
core. In which, to simulate the steel tube, the solid ing two types of steel.
8-node (C3D8) element with full integrated points
was used. Meanwhile, for concrete core simulation,
2.3 Steel-concrete interaction modeling
both the solid 8-node and 6-node (C3D6) elements
within the decreasing the number of integrated The interaction between the steel tube surface and
points would be used. For the loading plates, the concrete core would be considered carefully. The sur-
rigid 4-node (R3D4) element was used. The interac- face-to-surface contact model was used to simulate
tion surfaces between the steel tube, the concrete this interaction. While the inner surface of the steel
core and the loading plates were modeled using tube and the concrete core have contact each other
surfaced interaction based on the Coulomb fric- under compressive loading, the shear force and the
tion model. force in perpendicular directions on their contact
surface. The bonding degree between the steel tube
and the concrete core was simulated based on the
2.2 Constituent material modeling
surface interaction using the pressure-overclosure
2.2.1 Concrete contact model and the Coulomb friction model
The Damage Plasticity Model was used to simulate in perpendicular and tangential directions of the
the behaviour of concrete in the CFST columns. contact surface, respectively. As using these models,
This model has ability to predict both compression the surfaces could separate or slide over each other
and tension behaviour of concrete material under as well as spread contact pressure and shear stress
the confined pressures. The stress-strain curve for between the concrete core and steel tube.
uniaxial compression of concrete used in FE mod- In the basic forms of Coulomb friction model,
eling was the result obtained from testing standard two constituent surfaces can receive the same shear
cylinder specimens. The tensile strength would be stress in their contact surface up to a determined
used in order to determine the cracking proper- stress before they commence to slide over each
ties and post-cracking of concrete. This model other. The Coulomb friction model defines this
were assumed that the direct stress across a crack criticle shear stress (τcrit) to be the stress in which
reduces to zero when the crack width opens. The the slip between surfaces begins. The critical shear
Poisson coefficient of concrete in elastic deforma- stress, τcrit, is determined as part of contact pres-
tion field is calculated as ϑc = 0.2. sure (p) between surfaces as below.
For 65 MPa specimen, according to Swidish
Standard, the cylinder and cube concrete specimens τcrit = µp (1)
with the dimensions of D = 150 mm, H = 300 mm,
and 150 × 150 × 150 mm, respectively, were tested Where µ is the friction coefficient between the
at the 28-day age to obtain compression strength, concrete core and steel tube. Baltay & Gjelsvik
including fc,cyl = 65 MPa and fc,cube = 79.4 MPa. (1990) proposed the values of the friction coeffi-
The elastic module of concrete is Ec = 38.5 GPa. cient between concrete and steel as 0.2–0.6. In this
The cracking energy of concrete is GF = 157 N/m study, the friction coefficient was chosen as 0.2 for
that was determined in accordance with RILEM’s all models because the columns were concentrically
(1985) recommendation. The tests were carried out loaded in compression. With the Coulomb friction
similarly for the remaining two types of concrete. model, while the concrete and steel surfaces have
contact each other then the contact pressure will be
2.2.2 Steel spread between them. In the contrast, the contact
An elastic-plastic model with plastic criteria Von- pressure reduces to zero if the surfaces separate.
Mises related to the flow rule and stiff isotropic
strain was used to describe the basic behaviour of
2.4 Boundary conditions and loading apply
the steel (HKS 1997). The stress-strain relationship
obtained from uniaxial tensile testing of standard Figure 1 shows that two loading plates were used
steel specimens would be used in Finite Element at top and bottom of the column. The compressive
Analysis (FEA). The Poisson coefficient of steel in loads would be applied to columns via the top load-
elastic deformation field is calculated as ϑa = 0.3. ing plate. Therefore, the top loading plate would be

232
 and steel tube simultaneously (SFE), respectively.
In addition, the empty (hollow) steel tube (SES)
was also loaded to failure for the references.

3 MODELING ANALYSIS RESULTS AND

DISCUSSION

3.1 Loading capacity of CFST columns

The numerical results of ultimate loading capacity
of all columns were shown in Table 2. In general,
Figure 1. Different types of loading application. the ultimate loading capacity of the CFST columns
increases when using higher concrete compressive
Table 1. Matrix of columns loaded. strength and steel yield strength. It also indicated
that the CFST column applied load to the concrete
Types of loading applied on core only (SFC) has the highest ultimate loading
the column capacity, and following by SFS and SFE cases,
Loaded Loaded Loaded respectively.
Concrete on the on the on the
compressive concrete steel entire Steel yield 3.2 Verification of numerical results
strength core tube CFST strength,
(MPa) only only column fy (MPa) The numerical results of loading capacity for the
CFST columns using the steel tube with yield
65 SFC1-65 SFS1-65 SFE1-65 433 strength of 433 MPa were verified with test results
75 SFC1-75 SFS1-75 SFE1-75 of the previous study (Johansson & Gylltoft 2002)
85 SFC1-85 SFS1-85 SFE1-85 and calculation results using Eurocode 4 in this
None-concrete filled SES1-0 study as can be seen in Table 3.
65 SFC2-65 SFS2-65 SFE2-65 475 There was a good agreement between the load-
75 SFC2-75 SFS2-75 SFE2-75 ing capacity received from FEA and from the test
85 SFC2-85 SFS2-85 SFE2-85 for CFST columns using 65-MPa concrete filled,
None-concrete filled SES2-0 including SES1, SFS1 and SFE1 with the compres-
65 SFC3-65 SFS3-65 SFE3-65 511 sive force ratio of 1.1, 1.05, and 1.09, respectively.
75 SFC3-75 SFS3-75 SFE3-75 However, the numerical result for SFC1 was not
85 SFC3-85 SFS3-85 SFE3-85 very good agreement with the test result (force ratio
None-concrete filled SES3-0 of 1.31). Maybe because the material model used
for filled concrete was not exact in three-dimension
stress tensor, which leads to the increase of con-
restrained displacements by five freedom degrees,
it could move in direction along the column axis.
Meanwhile, the bottom plate would be restrained Table 2. FEA ultimate loading capacity (kN).
all deflections by six freedom degrees.
The loading was applied using the displacement Concrete compressive strength Steel
increment at the central node of the top loading (MPa) yield
Load strength,
plate because the final purpose is to investigate the
case 65 75 85 fy (MPa)
mechanical behaviour of columns up to failure. For
evenly distributing loads at the top of columns, all SFC1 2914 3159 3280 433
remain nodes located on the surface of the loading SFS1 994-2692 994-2951 994-3065
plate need to have the same displacement with the SFE1 2334 2580 2688
central node. In addition, the Newton-Raphson SES1 1008 1008 1008
method was used to reach the balance for each SFC2 3117 3371 3482 475
loading increment. SFS2 1082-2965 1082-3209 1082-3320
SFE2 2534 2780 2892
2.5 Matrix of simulated columns SES2 1190 1190 1190
SFC3 3295 3576 3690 511
The types of columns in this study are shown as in SFS3 1174-3161 1174-3407 1174-3516
the Table 1. Compressive loads were concentrically SFE3 2690 2935 3047
applied to the concrete core only (SFC), to the SES3 1288 1288 1288
steel tube only (SFS) and to both the concrete core

233
Table 3. Loading capacity comparison of numerical results with test and calculation results.

Strength of concrete Pmax PFEA Pmax/ PFEA/ PFEA/

Load case (MPa) from test (kN) (kN) Pu,cal Pmax Pu,cal

SES1 – 920 1008 – 1.1 –

SFC1 65 2220 2914 1.03 1.31 1.36
SFS1 65 950 994-2692 0.44 1.05-2.83 0.46-1.25
SFE1 65 2150 2334 1 1.09 1.09
SFC1 75 – 3159 – – 1.36
SFS1 75 – 994–2951 – – 0.43-1.27
SFE1 75 – 2580 – – 1.11
SFC1 85 – 3280 – – 1.31
SFS1 85 – 994–3065 – – 0.40-1.23
SFE1 85 – 2688 – – 1.08

crete compressive strength compared to that in

the reality. In addition, the difference between the
testing setup and FEA modeling can result in this
gap. Namely, for testing case, the load was applied
to the concrete core at both end of the column.
Meanwhile, for FEA modeling case, the load was
applied only to the concrete core at the top of the
column and the entire section at the bottom of the
column was restrained for all six freedom degrees.
There was a good agreement in SFE1 cases for all
three levels of concrete compressive strength rang-
ing from 65 to 85 MPa, with the appropriate force
ratio of 1.09, 1.11 and 1.08, respectively.

3.3 Force-displacement diagrams

The axial force and displacement relationship of
the CFST columns using the steel tubes with dif-
ferent levels of yield strengths in tension and con- Figure 2. Axial force—displacement relationship for
crete compressive strengths, and loading cases were the 433-MPa steel tube.
shown as in Figures 2–4.
It is clear that the force-displacement curves of
CFST columns were similar for the same load-
ing case when increasing the concrete strength.
The increase of concrete compressive strength as
well as steel yield strength leads to the significant
increase of loading capacity of CFST columns.
The curves in Figures 2–4 show that SFE col-
umns have harder mechanical behaviour than SFC
columns in the elastic stage. The concrete core and
the steel tube bear simultaneously the loading at the
beginning of loading process for SFE columns, hence
the load is distributed to both the steel tube and con-
crete core. Meanwhile, the concrete core bears most
of loading at this stage for SFC columns. Such curves
also indicated that the mechanical behaviour of SFS
columns are similar to SES columns in the elastic
stage. However, the concrete core subjects to loading
along with the steel tube when the tube reaches the
yield strength. Then, the loading capacity of such
columns is recovered until the concrete core reaches Figure 3. Axial force—displacement relationship for
its ultimate strength. Moreover, the ductility of the the 475-MPa steel tube.

234
 Figure 5. Mechanical failures of CFST columns for the
433-MPa steel tube and 65-MPa concrete filled.

Figure 6. Mechanical failures of CFST columns for the

Figure 4. Axial force—displacement relationship for 475-MPa steel tube and 65-MPa concrete filled.
the 511-MPa steel tube.

SFS and SES columns is much higher than that of

the SFE and SFC columns. This is because load is
applied to the steel tube for the two former columns
so the steel tube contributes significantly to the duc-
tility of these columns.

3.4 Mechanical failures of CFST columns Figure 7. Mechanical failures of CFST columns for the
To have a better understand about the mechanical 511-MPa steel tube and 65-MPa concrete filled.
behaviour of CFST columns using the steel tube
with various yield strengths ranging from 433 to
511 MPa under different loading cases, the failure
modes of these columns are shown in Figures 5–7.
The above images showed that the local buckling
of steel tube in SES1-0 column is different from
SFS1-65 column. To be more specific, the local
buckling of the steel tube in SES1-0 is to deform
inwardly and increase horizontal deflection and
deformation along with the vertical deflection of
the column (see Fig. 5a). In contrast, the SFS1-65 Figure 8. Axial force distribution in constitutive materials
column illustrated that the local buckling of the for the 433-MPa steel tube and 65-MPa concrete filled.
steel tube is to deform outwardly due to there is
the resistance of the concrete core. Therefore, the
3.5 Axial force distribution in the concrete core
local buckling of the steel tube in this case have
and steel tube
wavy shapes as can be seen in Figure 5c.
For SES2-0 and SES3-0 columns, the local In order to determine the axial force distribution in
buckling of the steel tube is to deform at the bot- the concrete core and steel tube of SFC1 and SFE1
tom of the column. Meanwhile, the local buckling columns, the force-displacement diagrams for these
of the SFS2-65 and SFS3-65 columns have a slight columns with concrete compressive strength ranging
difference comparing to the SFS1-65 column due from 65 to 85 MPa were drawn in Figures 8–10.
to the increase of steel yield strength. In addition, The axial force in concrete core for SFC1 and
the failure modes of the remain columns (SFC & SFE1 columns was much higher than the nominate
SFE) are due to the cracking of the concrete core compressive loading capacity of concrete according
and the yielding failure of the steel tube without Eurocode 4. This pointed out that the confinement
the local buckling. effect offered by the steel tube increases the loading

235
 Confining effect on concrete core offered by the steel
tube creates three-dimension compression stress. It
leads to a significant increase of compressive strength
of the concrete core. The CFST columns loaded on
the concrete core only have the best confining effect.

REFERENCES

Figure 9. Axial force distribution in constitutive materials Baltay, P. & Gjelsvik, A. 1990. Coefficient of friction for
for the 433-MPa steel tube and 75-MPa concrete filled. steel on concrete at high normal stress. J. Mater. Civ.
Eng. 2(1):46–49.
Eurocode 4. Design of composite steel and concrete struc-
ture. Part 1.1, General rules and rules for buildings.
Fujimoto, T., Mukai, A., Nishiyama, I. & Sakino, K.
2004. Behavior of Eccentrically Loaded Concrete-
Filled Steel Tubular Columns. Journal of Structural
Engineering. 130(2):203–212.
Ge, H. & Usami, T. 1992. Strength of concrete-filled
thin-walled steel box columns- experiment. Journal of
Structural Engineering. 118(11):3036–3054.
Ge, H. & Usami, T. 1994. Strength analysis of concrete-
filled thin-walled steel box columns. Journal of Con-
Figure 10. Axial force distribution in constitutive materi- structional Steel Research. 30:259–281.
als for the 433-MPa steel tube and 85-MPa concrete filled. Han, L.-H. 2004. Flexural behaviour of concrete-filled
steel tubes. Journal of Constructional Steel Research.
60:313–337.
Han, L.-H., Liu, W. & Yang, Y.-F. 2008. Behaviour of
capacity of concrete core, and this effect was much concrete-filled steel tubular stub columns subjected to
stronger for SFC1 columns. For SFE1 columns, the axially local compression. Journal of Constructional
concrete core contributes about 59–65% the entire Steel Research. 64:377–387.
column loading capacity. Meanwhile, the concrete HKS 1997. ABAQUS/Standard User’s Manual, version
core contributes around 84% the entire column 5.7. Hibbit, Karlsson & Sorensen, Pawtucket, R.I.
loading capacity for SFC1 columns. Therefore, Johansson, M. & Gylltoft, K. 2002. Mechanical Behavior
loading on concrete core only (SFC) creates the of Circular Steel-Concrete Composite Stub Columns.
highest confinement effect in the concrete core. Journal of Structural Engineering. 128(8):1073–1081.
Le, X.D. & Pham, M. 2016. Investigation into relative sliding
effect between steel tube and concrete core on eccentric
compression resistance of concrete filled steel tube col-
4 CONCLUSIONS umns. Journal of Science and Technology. 1(98):15–21.
Ngo-Huu, C., Nguyen-Minh, L., Ho-Huu, C., Kamura, H.,
Based on the numerical analysis results of this Nanba, T. & Nakagawa, K. 2016. Experimental study
study and the comparison with previous studies, of circular stub CFT columns under axial compression
the following conclusions can be drawn: loads. Proceedings of 14th East Asia-Pacific Conference
Mechanical behaviour and ultimate compres- on Structural Engineering and Construction: 994–1002.
sive strength of CFST columns depend on the way Phan, D.H., Ngo, H.C., Ngo T.L.V. & Nguyen, H.H. 2012.
Nonlinear analysis of concrete filled steel tube struc-
applying loads to the column. The case of loading
tures. Journal of Science and Technology. 8(57):1–6.
on the concrete core only (SFC) leads to the high- RILEM 50-FMC Committee. 1985. Determination of
est ultimate compressive strength, and followed by the fracture energy of mortar and concrete by means
SFS and SFE columns, respectively. of three-point bend tests on notched beams. Mater.
Compressive strength of concrete filled signifi- Struct. 18(106):285–290.
cantly affects the ultimate loading capacity of CFST Starossek, U., Falah, N. & Lohning, T. 2008. Numerical Anal-
columns. When increasing the concrete strength leads yses of the Force Transfer in Concrete-Filled Steel Tube
to an increase of the column loading capacity. More- Columns. The 4th International Conference on Advances in
over, the force-displacement curves are similar for the Structural Engineering and Mechanics: 2651–2666.
Uy, B. 1998. Concrete-filled fabricated steel box columns
columns with the same way of loading while chang-
for multi-storey buildings: behaviour and design.
ing the compressive strength of concrete filled. Progress in Structural Engineering and Materials. 1(2):
Increasing steel yield strength of the steel tube 150–158.
makes an increase of the column loading capacity, Yu, Z. 2007. Experimental behavior of circular concrete-
in general, and changes the failure modes of both filled steel tube stub columns. Journal of Construc-
the hollow steel tube and CFST columns. tional Steel Research. 63:165–174.

236

View publication stats