Journal of Constructional Steel Research 62 (2006) 424–433

www.elsevier.com/locate/jcsr

Analytical model for predicting axial capacity and behavior of concrete

encased steel composite stub columns
Cheng-Chih Chen ∗ , Nan-Jiao Lin
Department of Civil Engineering, National Chiao Tung University, Hsinchu 30010, Taiwan

Received 15 April 2005; accepted 16 April 2005

Abstract

The axial compressive capacity and force–deformation behavior of concrete encased steel stub columns were analytically investigated. An
analytical model was developed for predicting the force–deformation response for composite stub columns with various structural steel sections
and volumetric lateral reinforcement. Constitutive relationships were established for materials used in the composite cross section, which included
unconfined concrete, partially and highly confined concrete, structural steel section, and longitudinal reinforcing bar. The axial capacity of
composite stub columns can be determined from strengths contributed from each material component following the stress–strain relationship.
Analytical results show that the axial load-carrying capacity and force–deformation behavior measured in the experiments can be accurately
predicted. In addition to the lateral reinforcement, the structural steel section can provide a confinement effect on the concrete and enhance the
axial capacity and post-peak strength.
c 2005 Elsevier Ltd. All rights reserved.

Keywords: Concrete encased; Composite column; Confined concrete

1. Introduction steel section. However, other shapes of steel section such as

cross- or T-shaped are generally used in composite buildings.
Concrete encased steel columns are one type of composite The composite column with cross-shaped steel section is widely
columns used in composite structures. The concrete encased used in an interior column to connect four steel beams in
steel composite column consists of structural steel section orthogonal directions. The cross-shaped steel section is usually
encased in reinforced concrete. The structural steel is rolled or fabricated by welding two H-shaped steel sections together.
built-up shape. Deriving benefits from combining the structural The composite column with T-shaped steel section is usually
steel and reinforced concrete, the composite columns possess designed for an outer column. There is very little research
great load-carrying capacity and stiffness owing to composite regarding the effect of various shapes of steel section on the
action. Further, the concrete encasement can serve for fire axial compressive behavior of concrete encased steel columns.
protection. Therefore, the use of the composite columns The concrete confinement of composite columns with various
in medium-rise or high-rise buildings has been increased shapes of steel section is not well understood yet.
significantly in recent decades [1,2]. Research has been conducted to study the confinement
Numerous experimental investigations have been carried effect of the concrete in concrete-filled steel tube (CFT)
out to study the ultimate strength of concrete encased steel composite columns [11–16]. From previous work, it is clear
composite columns [3–8]. Research has also been carried out on that the steel tube, particularly the circular tube, can provide
cyclic behavior of composite beam–columns [9,10]. Although confinement on concrete and lead to the enhancement of
the behavior of concrete encased steel composite columns strength and ductility of the CFT columns. It is of important
has been extensively studied, many of the research works to investigate the concrete confinement effect in the concrete
emphasized the composite columns with H-shaped structural encased steel composite columns. In this study, an analytical
approach is developed for determining the axial compressive
∗ Corresponding author. Tel.: +886 3 571 2121x54915; fax: +886 3 572 7109. load–deformation relationship for concrete encased composite
E-mail address: chrischen@mail.nctu.edu.tw (C.-C. Chen). stub columns. The emphasis of the proposed modeling is the

0143-974X/$ - see front matter 

c 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcsr.2005.04.021
 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433 425

Nomenclature
Ach area of the highly confined concrete;
Acp area of the partially confined concrete;
Acu area of the unconfined concrete;
Ar cross-sectional area of the longitudinal bars;
As cross-sectional area of the structural steel section;
Ec tangent modulus of elasticity of concrete;
E sec secant modulus of the confined concrete at peak
stress;
fc longitudinal concrete stress;

f cc compressive strength of the confined concrete;
f ch stress of the highly confined concrete;

f co compressive strength of the unconfined concrete; Fig. 1. Stress–strain curves for unconfined and confined concrete proposed by
f cp stress of the partially confined concrete; Mander et al. [18].
f cu stress of the unconfined concrete;
generated for the composite cross section. Several assumptions
fl effective lateral confining stress;
considered in this analytical model are as follows: (1)
fs stress of the structural steel;
uniform distribution of compressive strain is assumed on the
f sr stress of the longitudinal bar;
cross section; (2) stresses of the materials are calculated
f ys yield strength of the structural steel;
based on corresponding uniaxial stress–strain relations; (3) a
f yr yield strength of the longitudinal bar;
confinement effect caused by the lateral reinforcement and
Kh confinement factor for highly confined concrete;
Kp confinement factor for partially confined con- elements of the structural steel on the concrete is considered;
crete; and (4) local buckling of the longitudinal bars and elements of
PAnaly analytical load; the structural steel is assumed.
PSquash squash load;
2.1. Constitutive model for concrete
PTest experimental load;
ε axial compressive strain;
The confinement effect of concrete by lateral reinforcement
εc longitudinal concrete strain;
in a reinforced concrete column has been recognized because
εcc strain at maximum confined concrete stress;
εcc, p strain at maximum partially confined concrete the lateral reinforcement can provide confining pressure to
stress; the concrete core [17,18]. The confining pressure results
εco strain at maximum unconfined concrete stress. in an enhancement in the strength and ductility of the
concrete, depending on the degree of the confining pressure.
In addition to the lateral reinforcement, the confinement
establishment of stress–strain relations for concrete confined is also affected by other factors, such as distribution of
by the lateral reinforcement and various structural steel the longitudinal reinforcement, cross section configuration,
sections. The predicted axial compressive capacity and axial and loading type. Analytical models to predict the uniaxial
load–deformation relationship were compared with available stress–strain behavior for confined concrete have been proposed
experimental results to validate the analytical modeling and by researchers [17–19].
investigate the effect of design variables. Mander et al. [18] proposed a unified stress–strain model,
shown in Fig. 1, for confined concrete for members with
2. Analytical modeling different cross sections under various loading conditions.
The longitudinal compressive stress–strain ( f c –εc ) curve for
The cross section of the concrete encased steel composite confined concrete is given by
column comprises three materials, i.e., concrete, structural  xr
steel, and longitudinal reinforcing bar. For a stub column, fcc
fc = (1)
the axial compressive capacity and axial load–deformation r − 1 + xr
response can be determined based on the strain compatibility with
on the composite cross section. When a uniform axial εc
compressive strain is assumed, the stress of each material of x= (2)
εcc
the composite column can be obtained through the constitutive Ec
model established for each material. Consequently, the axial r= (3)
E c − E sec
load can be calculated by adding the axial force from each
material, while the axial force is computed by multiplying where fcc  is the compressive strength (peak stress) of confined

the stress of material by the corresponding cross-sectional concrete; εcc is the strain at maximum confined concrete stress;
area. Furthermore, the axial load versus strain curve can be E c is the tangent modulus of elasticity of the concrete; and E sec
426 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433

Fig. 2. Materials in a concrete encased steel composite column.

is the secant modulus of confined concrete at peak stress and is

expressed as

f cc Fig. 3. Stress–strain relation for longitudinal reinforcing bar in compression.
E sec = . (4)
εcc
The strain at maximum confined concrete stress εcc is suggested
as

  
f cc
εcc = εco 1 + 5 
− 1 (5)
fco
where fco  is the compressive strength of unconfined concrete

and εco is the strain at maximum unconfined concrete stress.

Proposed by Mander et al., the compressive strength of
confined concrete f cc  is determined by the compressive

strength of unconfined concrete f co  and the effective lateral

  is given
confining stress fl . The confined concrete strength f cc
by
  
  7.94 f l fl
f cc = f co −1.254 + 2.254 1 + 
−2  . (6)
f co f co

The effective lateral confining stress fl is dependent on the Fig. 4. Stress–strain relation for structural steel section in compression.
volumetric ratio of lateral reinforcement, configuration of the
lateral and longitudinal reinforcement, and area of effectively for unconfined concrete is determined by Eq. (1), substituting
confined concrete core. Full equations can be found in Mander fl = 0 in Eq. (6) and εco = 0.002 in Eq. (5). The strength f co
 is

et al. [18]. taken as the concrete compressive strength measured from the
To determine the area of effectively confined concrete cylinder test. The strain εco of 0.002 is generally acceptable for
core, parabolic arching was assumed to occur between the unconfined concrete [17,18]. Considering the confining stress
reinforcing bars in the cross section [17,18]. Accordingly, a contributed from the lateral reinforcement, the stress–strain
similar parabolic arching is assumed herein for the concrete curve for partially confined concrete can be determined as
further confined by the structural steel section. Therefore, proposed by Mander et al.
as presented in Fig. 2, the concrete in a concrete encased In addition to the lateral reinforcement, the confining stress
steel composite section can be divided into three regions: (1) in the highly confined concrete region is enhanced by the
an unconfined concrete region outside the parabolic arching structural steel section. Li et al. [9] analyzed the flexural
formed by the longitudinal bars; (2) a highly confined region strength of concrete encased steel composite beam–columns
inside the perimeter of the structural steel section and the without considering the highly confined concrete region.
arching formed by the steel section; and (3) a partially confined However, Li et al. found that one of the reasons to
region inside the unconfined concrete region and outside the underestimate the ultimate strength was that the confining effect
highly confined concrete region. Mirza and Skrabek [6], and El- of structural steel section on concrete was not accounted for in
Tawil and Deierlein [7] adopted a similar but simple subdivision their analysis. Therefore, the structural steel section is taken
by simplifying the arching formed for the partially confined into account in this study to determine the confining stress,
and a higher compressive strength f cc  is obtained for the
concrete region.
Based on these three distinct regions, different stress–strain highly confined concrete region. By considering the effect of
relations are assumed for the concrete. The stress–strain curve unstiffened elements of the steel section on the confinement,
 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433 427

Fig. 5. Cross sections of concrete encased steel composite columns with I-, H- and cross-shaped steel sections [21,22].

Fig. 6. Cross sections of concrete encased steel composite columns with T-shaped steel sections [23].

one-half of the corresponding steel area is taken to calculate the plateau. The stress of the bar begins to degrade when the axial
confining stress. In order to discuss the confinement effect, the strain of the bar reaches the strain εco , corresponding to the
concrete strength f cc for partially and highly confined concrete peak compressive stress f co of the unconfined concrete. It is
  , respectively, as follows:
is defined as K p f co and K h f co assumed that the bar will buckle and lose its strength caused
for partially confined concrete by the spalling of the concrete cover when the concrete cover
  reaches the peak strength. The stress of the bar will drop to 20%
fcc = K p fco (7)
of its yield strength and maintain constant afterwards.
for highly confined concrete
 
2.3. Constitutive model for structural steel section
fcc = K h f co (8)
where K p and K h are defined as confinement factors for A similar stress–strain relation as that adopted for the
partially and highly confined concrete, respectively. longitudinal bar is assumed for the structural steel section, as
shown in Fig. 4. The local buckling of the elements, particularly
2.2. Constitutive model for longitudinal bar the flanges, of the structural steel section is likely to occur
after the crushing of the partially confined concrete. Therefore,
In general, identical behavior under tension and compression stress degradation is assumed after the axial strain reaches the
is assumed for modeling the longitudinal bars in the reinforced strain, εcc, p , representing the crushing of the partially confined
concrete or composite members subjected to flexural bending concrete. Post-peak strength of 20% of the yield strength is
moment [6,7]. However, buckling of the longitudinal bars assumed when the axial strain reaches four times of the strain
occurred at large inelastic deformation when those members of εcc, p .
were subjected to axial compressive force. The buckling of the
longitudinal bars greatly influences the strength and ductility of 3. Analytical results and discussion
the member. As a result, load-carrying capacity and ductility of
the bars decrease when the bars buckle, as observed in the tests 3.1. Composite stub column tests
of reinforcing bars by Bayrak and Sheikh [20].
A simple constitutive model considering the inelastic The analytical predictions of axial compressive behavior
buckling of the longitudinal bars under compression is and capacity of composite stub columns are compared to
suggested herein, as presented in Fig. 3. The following experimental results. Three series of tests, matching the purpose
assumptions are used to generate the model. The longitudinal of this study, are considered: the tests conducted by Chen
bar under compression will reach the yield strength with a yield and Yeh [21], by Tsai et al. [22], and by Chen et al. [23].
428 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433

Table 1
Geometrical and material properties of composite stub columns

Author Specimen Cross section Length Structural steel Longitudinal bar Lateral tie Concrete strength
(mm) (mm) Shape Size spacing (mm) f c (MPa)
SRC1 280 × 280 1200 H H150 × 150 × 7 × 10 12 No. 5 140 29.5
SRC2 280 × 280 1200 H H150 × 150 × 7 × 10 12 No. 5 75 28.1
SRC3 280 × 280 1200 H H150 × 150 × 7 × 10 12 No. 5 35 29.8
SRC4 280 × 280 1200 Cross Two H175 × 90 × 5 × 8 12 No. 5 140 29.8
Chen and SRC5 280 × 280 1200 Cross Two H175 × 90 × 5 × 8 12 No. 5 75 29.8
Yeh [21] SRC6 280 × 280 1200 Cross Two H175 × 90 × 5 × 8 12 No. 5 35 29.5
SRC7 280 × 280 1200 I H150 × 75 × 5 × 7 12 No. 5 140 28.1
SRC8 280 × 280 1200 I H150 × 75 × 5 × 7 12 No. 5 75 26.4
SRC9 280 × 280 1200 I H150 × 75 × 5 × 7 12 No. 5 140 28.1
SRC10 280 × 280 1200 I H150 × 75 × 5 × 7 12 No. 5 75 29.8
src1 280 × 280 1200 Cross Two H175 × 90 × 5 × 8 4 No. 5 140 23.9
src2 280 × 280 1200 Cross Two H175 × 90 × 5 × 8 4 No. 5 100 23.5
src3 280 × 280 1200 Cross Two H175 × 90 × 5 × 8 12 No. 5 100 21.8
src4 280 × 280 1200 Cross Two H175 × 90 × 5 × 8 12 No. 5 100 25.3
src5 280 × 280 1200 Cross Two H160 × 50 × 3.2 × 4.5 4 No. 5 190 26.0
Tsai et al. [22]
src6 280 × 280 1200 Cross Two H160 × 50 × 3.2 × 4.5 4 No. 5 140 26.3
src7 280 × 280 1200 Cross Two H160 × 50 × 3.2 × 4.5 12 No. 5 140 25.0
src8 280 × 280 1200 Cross Two H160 × 50 × 3.2 × 4.5 4 No. 5 100 26.6
src9 280 × 280 1200 Cross Two H160 × 50 × 3.2 × 4.5 12 No. 5 100 24.6
src10 280 × 280 1200 Cross Two H160 × 50 × 3.2 × 4.5 12 No. 5 100 24.3
CL-TE 300 × 300 1000 T H100 × 50 × 5 × 7, H125 × 60 × 6 × 8 4 No. 6 100 22.9
CL-TO 300 × 300 1000 T H100 × 50 × 5 × 7, H125 × 60 × 6 × 8 4 No. 6 100 22.9
CL-HO 300 × 300 1000 Cross H100 × 50 × 5 × 7, H125 × 60 × 6 × 8 4 No. 6 100 22.9
Chen et al. [23]
CH-TE 300 × 300 1000 T H150 × 100 × 6 × 9, H175 × 90 × 5 × 8 4 No. 6 100 31.4
CH-TO 300 × 300 1000 T H150 × 100 × 6 × 9, H175 × 90 × 5 × 8 4 No. 6 100 31.4
CH-HO 300 × 300 1000 Cross H150 × 100 × 6 × 9, H175 × 90 × 5 × 8 4 No. 6 100 31.4

For the tests by Chen and Yeh, and Tsai et al., three shapes
of the structural steel section used in the specimens were
I-, H- and cross-shaped sections, as illustrated in Fig. 5. The
H-shaped steel section is more like the wide-flange section,
while the I-shaped section has a narrow flange. For the tests
carried out by Chen et al., a T-shaped steel section was used in
the specimens. Fig. 6 depicts the cross section configurations
of the composite columns. Table 1 summarizes geometrical
properties of the composite stub columns in these three tests.
A total of 26 composite stub columns are included in this
study. The main variables are the shape of the structural steel
section, longitudinal reinforcing bar and lateral tie. Four types
of the structural steel section were used. The regions defined for
unconfined, partially and highly confined concrete for various
composite cross sections used in these tests are illustrated
in Fig. 7. The cross section with cross-shaped steel section
results in the largest area of highly confined concrete while
I-shaped has the smallest area. Table 2 tabulates all the
measured material properties used for analytical prediction.
According to the analytical procedure, the stress–strain Fig. 7. Regions for unconfined, partially confined, and highly confined
curves for unconfined, partially and highly confined concrete concrete in various composite cross sections.
are calculated and shown in Fig. 8, which presents the
curves for specimens with H- and cross-shaped steel sections. especially for composite cross section with cross-shaped steel
Strength increase for partially confined concrete is primarily section, as shown in Fig. 8(b).
due to the lateral reinforcement. However, besides the lateral
reinforcement, strength increase for highly confined concrete 3.2. Comparison of analytical and experimental results
is further contributed by the confining stress owing to the steel
section. As indicated in the figures, the highly confined concrete On the basis of strain compatibility and constitutive
provides higher strength than partially confined concrete, relationships for material components at a given axial
 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433 429

Table 2
Material properties used for analytical modeling

Specimen Area of Area of Area of Area of Area of Yield Yield Concrete Confinement factor Confinement factor
steel bar unconfined partially highly strength of strength of strength for partially for highly confined
As (mm2 ) Ar (mm2 ) concrete confined confined steel f ys bar f yr  (MPa)
f co confined concrete concrete K h
Acu (mm2 ) concrete concrete (MPa) (MPa) Kp
Acp (mm2 ) Ach (mm2 )
SRC1 4014 2400 29 955 29 074 12 957 296 350 29.5 1.08 1.23
SRC2 4014 2400 29 955 29 074 12 957 296 350 28.1 1.22 1.24
SRC3 4014 2400 29 955 29 074 12 957 296 350 29.8 1.50 1.50
SRC4 4585 2400 32 086 19 580 19 749 345 350 29.8 1.08 1.87
SRC5 4585 2400 32 086 19 580 19 749 345 350 29.8 1.20 1.90
SRC6 4585 2400 32 086 19 580 19 749 345 350 29.5 1.48 1.97
SRC7 1785 2400 32 086 38 774 3 355 303 350 28.1 1.09 1.10
SRC8 1785 2400 32 086 38 774 3 355 303 350 26.4 1.24 1.24
SRC9 1785 2400 32 086 38 774 3 355 303 350 28.1 1.09 1.10
SRC10 1785 2400 32 086 38 774 3 355 303 350 29.8 1.21 1.21
src1 4585 800 49 753 3 538 19 724 274 453 23.9 1.08 1.86
src2 4585 800 49 753 3 538 19 724 274 453 23.5 1.14 1.88
src3 4585 2400 31 390 20 301 19 724 274 453 21.8 1.25 1.96
src4 4585 2400 31 390 20 301 19 724 274 453 25.3 1.22 1.86
src5 1856 800 49 753 12 657 13 334 271 453 26.0 1.04 1.34
src6 1856 800 49 753 12 657 13 334 271 453 26.3 1.07 1.35
src7 1856 2400 31 390 29 420 13 334 271 453 25.0 1.25 1.37
src8 1856 800 49 753 12 657 13 334 271 453 26.6 1.13 1.35
src9 1856 2400 31 390 29 420 13 334 271 453 24.6 1.22 1.39
src10 1856 2400 31 390 29 420 13 334 271 453 24.3 1.42 1.42
CL-TE 2869 1136 54 674 24 681 6 640 333 388 22.9 1.26 1.26
CL-TO 2869 1136 54 674 24 681 6 640 333 388 22.9 1.26 1.26
CL-HO 2839 1136 54 674 24 521 6 830 333 388 22.9 1.26 1.34
CH-TE 4989 1136 54 674 12 360 16 841 320 388 31.4 1.19 1.31
CH-TO 4989 1136 54 674 12 360 16 841 320 388 31.4 1.19 1.31
CH-HO 4959 1136 54 674 11 943 17 288 320 388 31.4 1.19 1.65

Fig. 8. Stress–strain curves for unconfined and confined concrete: (a) H-shaped steel section; (b) cross-shaped steel section.

compressive strain ε, the analytical axial load PAnaly is given sectional area of the longitudinal bars; Acu is the area of the
by unconfined concrete; Acp is the area of the partially confined
concrete; and Ach is the area of the highly confined concrete.
PAnaly = f s As + f sr Ar + f cu Acu + f cp Acp + fch Ach (9)
The axial load and axial strain relationships for all specimens
where fs is the stress of the structural steel; f sr is the stress were calculated by the proposed model. The analytical axial
of the longitudinal bars; f cu is the stress of the unconfined load–strain curves for specimen SRC4 are presented in Fig. 9 in
concrete; f cp is the stress of the partially confined concrete; fch which the curves for each material are included. It is observed
is the stress of the highly confined concrete; As is the cross- that the unconfined concrete reaches its maximum strength at
sectional area of the structural steel section; Ar is the cross- the strain of 0.002. The structural steel and longitudinal bars
430 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433

concrete confinement is also not taken into account for nominal

axial compressive strength in various codes, such as ACI
building code [24] and AIJ standards [25].
As presented in Table 3, the squash loads for all stub
columns are also compared to the maximum experimental
loads. The squash loads were calculated using measured
material strengths as presented in Table 2. The average ratios of
the experimental loads to squash loads, PTest/PSquash , for three
series tests range from 1.11 to 1.15, which are higher than those
of PTest/PAnaly . The errors of the squash loads compared to the
experimental results are generally more than 10% and even up
to 21%. The reason for these errors may be attributed to the
confinement effect of the concrete. Because of neglect of the
concrete confinement, the ratios of PTest /PSquash become large
for the columns with large concrete area ratio, as in the case of
Fig. 9. Analytical axial load–strain curves for specimen SRC4. the columns SRC7∼10, src5∼10, and CL-series. Nevertheless,
the analytical predictions give a fair ratio of PTest /PAnaly for
have yielded at this strain. After that, the axial resistance of the columns with different concrete area ratios.
column still increases due to the strength enhancement from
partially and highly confined concrete. 3.4. Confinement factors for partially and highly confined
Fig. 10 shows the comparison of analytical and experimental concrete
axial load–strain curves for specimens with different structural
steel sections. The predicted curves are in good agreement with The strength of the confined concrete is influenced by
the experimental results. In addition to the pre-peak behavior, the tie spacing, volumetric ratio of the lateral reinforcement,
an accurate prediction for the post-peak behavior is obtained. and the distribution of the longitudinal reinforcing bar [18].
The post-peak behavior results from the use of the degrading According to the model for partially confined concrete, the
model for the longitudinal bars and structural steel as well as partial confinement factor is highly dependent on the lateral tie
the post-peak behavior of the unconfined and partially confined spacing for the specimens used in this study. The effect of the
concrete. tie spacing on the axial load–deformation behavior is shown
Table 3 presents the maximum axial compressive loads of in Fig. 11, which confirms that reduction of the tie spacing
the experimental tests and analytical predictions for all the enhances post-peak behavior. Fig. 12 shows the relations of the
specimens. The analytical expressions can accurately predict tie spacing versus confinement factor K p for partially confined
the experimental axial compressive loads. The average ratios of concrete for columns tested by Chen and Yeh. It is apparent that
the experimental loads to analytical capacities, PTest/PAnaly , are the effectiveness of confinement by the lateral reinforcement is
1.01, 1.02 and 1.00 for three series tests, and the corresponding shown in the figure. Small tie spacing results in the increase of
coefficients of variation are 0.02, 0.06 and 0.02. Among the K p .
the specimens, the axial capacities of the stub columns are The confinement factor K h for highly confined concrete is
well overestimated in only two columns, src3 and src4, with influenced by the shape of the structural steel section which
testing to predicted load ratios of 0.94 and 0.90, respectively. provides confining stress on the core concrete. Fig. 13 presents
However, the testing axial loads were suspected because of the the steel shape versus K h relations and shows the effect
concrete defect during casting the columns, as reported in the of the steel shape on the effectiveness of confinement. The
literature [22]. confining effect is enhanced by the structural steel, particularly
by the cross-shaped section. It is because the cross-shaped steel
3.3. Squash load of composite stub columns section can provide confining pressure in both directions and
result in a higher value of the confinement factor K h than the
A squash load calculation is a simplified method to predict I- or H-shaped steel section which confines the concrete in only
the axial compressive capacity for a stub column. Assuming one direction.
that each material reaches its ultimate strength, the squash load
of a composite stub column, PSquash , is defined by: 4. Conclusions

PSquash = 0.85 f c Ac + f ys As + f yr Ar (10) An analytical model for determining the axial behavior
where f ys is the yield strength of the structural steel; f yr and capacity of concrete encased composite stub columns is
proposed. The following conclusions can be made based on the
is the yield strength of the longitudinal bar; Ac is the total
analytical investigation.
area of the concrete. It should be noted that 0.85 f c is used
to represent the concrete strength in a structural column. 1. Based on strain compatibility, the analytical model can
Of course, concrete strength increase owing to the confinement reasonably simulate the axial compressive load–deformation
effect is not considered for calculating the squash load. The relations of composite stub columns with various steel
 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433 431

Fig. 10. Comparison of experimental and analytical axial load–strain curves: (a) specimen SRC2; (b) specimen SRC4; (c) specimen SRC7; (d) specimen CL-TO.

Fig. 11. Effect of tie spacing on axial load–strain curves: (a) H-shaped steel section; (b) cross-shaped steel section.
432 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433

Table 3
Comparison between experimental and analytical results
PTest PTest
Author Specimen Test PTest Squash PSquash Analysis PAnaly PSquash PAnaly
(kN) (kN) (kN)
SRC1 4220 3832 4247 1.10 0.99
SRC2 4228 3748 4180 1.13 1.01
SRC3 4399 3832 4381 1.15 1.00
SRC4 4441 4230 4459 1.05 1.00
SRC5 4519 4230 4491 1.07 1.01
Chen and Yeh [21]
SRC6 4527 4212 4535 1.07 1.00
SRC7 3788 3154 3758 1.20 1.01
SRC8 3683 3047 3582 1.21 1.03
SRC9 3630 3154 3530 1.15 1.03
SRC10 3893 3260 3693 1.19 1.05
Average 1.13 1.01
Coefficient of variation 0.05 0.02
src1 3602 3101 3486 1.16 1.03
src2 3502 3076 3462 1.14 1.01
src3 3836 3666 4062 1.05 0.94
src4 3854 3818 4304 0.99 0.90
src5 3063 2539 2877 1.21 1.06
Tsai et al. [22]
src6 3009 2559 2905 1.18 1.04
src7 3696 3166 3539 1.17 1.04
src8 3088 2578 2934 1.20 1.05
src9 3748 3141 3506 1.19 1.07
src10 3744 3122 3528 1.15 1.06
Average 1.15 1.02
Coefficient of variation 0.06 0.06
CL-TE 3452 3070 3433 1.12 1.01
CL-TO 3448 3070 3433 1.12 1.00
CL-HO 3514 3060 3428 1.15 1.03
Chen et al. [23]
CH-TE 4652 4277 4732 1.09 0.98
CH-TO 4718 4277 4732 1.10 1.00
CH-HO 4676 4269 4766 1.10 0.98
Average 1.11 1.00
Coefficient of variation 0.02 0.02

Fig. 12. Confinement factor for partially confined concrete.

Fig. 13. Confinement factor for highly confined concrete.

sections, tie spacings, and distributions of the longitudinal 2. The axial compressive loads for 26 composite stub columns
bars. can be accurately predicted by the analytical model. The
 C.-C. Chen, N.-J. Lin / Journal of Constructional Steel Research 62 (2006) 424–433 433

predictions are better than the squash load. The axial load- concrete beam–columns. In: Proceedings of the international conference
carrying capacity of a composite stub column is developed on advances in structures. 2003, p. 1201–7.
owing to the confinement effect of the concrete. [10] Ricles JM, Paboojian SD. Seismic performance of steel-encased
composite columns. Journal of Structural Engineering 1994;120(8):
3. The concrete confinement effect due to the confining stress 2474–94.
from the lateral reinforcement as well as different shapes [11] Susantha KAS, Ge H, Usami T. Uniaxial stress–strain relationship of
of the structural steel section is established for partially concrete confined by various shaped steel tubes. Engineering Structures
and highly confined concrete. The concrete confinement is 2001;23(10):1331–47.
confirmed from the comparisons of the predictions with [12] Han LH. Tests on stub columns of concrete-filled RHS sections. Journal
of Constructional Steel Research 2002;58:353–72.
experimental results. The cross-shaped steel section leads to [13] O’Shea MD, Bridge RQ. Design of circular thin-walled concrete filled
the highest confinement while the I-shaped has the lowest steel tubes. Journal of Structural Engineering 2000;126(11):1295–303.
one. [14] Sakino K, Nakahara H, Morino S, Nishiyama I. Behavior of centrally
loaded concrete-filled steel-tube short columns. Journal of Structural
Engineering 2004;130(2):180–8.
Acknowledgment [15] Giakoumelis G, Lam D. Axial capacity of circular concrete-filled tube
columns. Journal of Constructional Steel Research 2004;60:1049–68.
The authors would like to thank the National Science [16] Schneider SP. Axially loaded concrete-filled steel tubes. Journal of
Council of Taiwan for partially supporting this research. Structural Engineering 1998;124(10):1125–38.
[17] Sheikh SA, Uzumeri SM. Analytical model for concrete confinement in
References tied columns. Journal of the Structural Division, ASCE 1982;108(12):
2703–22.
[1] Roeder CW. Overview of hybrid and composite systems for seismic [18] Mander JB, Priestley MJN, Park R. Theoretical stress–strain model
design in the United States. Engineering Structures 1998;20(4–6): for confined concrete. Journal of Structural Engineering 1988;114(8):
355–63. 1804–26.
[2] Morino S. Recent developments in hybrid structures in Japan — research, [19] Kent DC, Park R. Flexural members with confined concrete. Journal of
design and construction. Engineering Structures 1998;20(4–6):336–46. the Structural Division, ASCE 1971;97(7):1969–90.
[3] Brettle HJ. Ultimate strength design of composite columns. Journal of the [20] Bayrak O, Sheikh SA. Plastic hinge analysis. Journal of Structural
Structural Division, ASCE 1973;99(9):1931–51. Engineering 2001;127(9):1092–100.
[4] Bridge RQ, Roderick JW. Behavior of built-up composite columns. [21] Chen CC, Yeh SC. Ultimate strength of concrete encased steel composite
Journal of the Structural Division, ASCE 1978;104(7):1141–55. columns. In: Proceedings of the third national conference on structural
[5] Furlong RW. Concrete encased steel columns — design tables. Journal of engineering. 1996, p. 2197–206 [in Chinese].
the Structural Division, ASCE 1974;100(9):1865–82. [22] Tsai KC, Lien Y, Chen CC. Behavior of axially loaded steel reinforced
[6] Mirza SA, Skrabek BW. Statistical analysis of slender composite concrete columns. Journal of the Chinese Institute of Civil and Hydraulic
beam–column strength. Journal of Structural Engineering 1992;118(5): Engineering 1996;8(4):535–45 [in Chinese].
1312–32. [23] Chen CC, Weng CC, Lin IM, Li JM. Seismic behavior and strength
[7] El-Tawil S, Deierlein GG. Strength and ductility of concrete encased of concrete encased steel stub columns and beam–columns. Report no.
composite columns. Journal of Structural Engineering 1999;125(9): MOIS 881012-1. Architecture and Building Research Institute; 1999 [in
1009–19. Chinese].
[8] Mirza SA, Hyttinen V, Hyttinen E. Physical tests and analyses [24] American Concrete Institute. Building code requirements for structural
of composite steel–concrete beam–columns. Journal of Structural concrete (ACI 318-02) and Commentary (ACI 318R-02). Detroit; 2002.
Engineering 1996;122(11):1317–26. [25] Architectural Institute of Japan. AIJ standards for structural calculation of
[9] Li L, Sakai J, Matsui C. Seismic behavior of steel encased reinforced steel reinforced concrete structures. 1991.