Journal of Constructional Steel Research 213 (2024) 108414

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

journal homepage: www.elsevier.com/locate/jcsr

Axial compression behavior of square section concrete-filled steel tubes

reinforced with internal latticed steel angles
Jun Wang a, Zi-Ming Yang b, Xiang-Long Zheng a, Yang Ding a, *
a
Department of Civil Engineering, Hangzhou City University, Hangzhou 310015, Zhejiang, China
b
College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, Zhejiang, China

A R T I C L E I N F O A B S T R A C T

Keywords: The axial compression behavior of square CFSTs (concrete-filled steel tubes) reinforced with internal latticed
Axial compression behavior steel angles was investigated in this study, and the test results including load-displacement and load-strain
Latticed steel angles behavior, failure modes, and ultimate load were obtained. The experimental results demonstrated that the
Finite element model
steel angles exerted an additional confining effect on the enclosed concrete, and altered the stress distributions of
Axial compression mechanisms
Superposition method
the adjacent concrete. The corresponding finite element model was established subsequent to the experiment,
Reliability analysis and the analysis and discussion of the axial compression mechanisms was conducted. The parametric analysis
was performed to evaluate the influence of main parameters on the load-displacement behavior, and the analysis
results revealed that the concrete made the highest contribution to the ultimate load. Additionally, the formulas
for calculating the ultimate load were proposed according to the superposition method, and the calculation
results demonstrated that the design formula derives from the unified theory of CFSTs exhibit enhanced accuracy
and reduced deviation. Finally, a reliability analysis was conducted to evaluate the feasibility of the proposed
design formula, and the analysis results indicated the calculated reliability index decreases with an increase in
the steel ratio, and for specimens with a higher steel ratio, a lower yield strength of steel tube leads to a lower
reliability index.

1. Introduction arranging longitudinal stiffeners [11]. However, these methods are

relatively time-consuming and labor-intensive [12,13]. In recent years,
Concrete filled steel tube (CFST) has been widely used in practical some scholars have incorporated various steel components with
engineering because of its excellent mechanical properties, and the different cross sections, such as steel tubes [14,15], reinforcement cages
current prevalent section forms for CFST are circular and square [1–4]. [16–18], spiral stirrups [12,13,19,20], and lattice steel angles [21,22]
The square steel tube exerts a relatively lower confining effect on the into CFST to enhance its stiffness and ductility. As result of these ad­
core concrete compared to the circular steel tube, assuming an equal vancements, the Steel Reinforced Concrete-Filled Steel Tubes (SRCFST)
amount of steel is utilized, however, it still effectively enhances the have been developed [23–25]. Additionally, with the extensive appli­
strength and ductility of the core concrete [2–5]. The utilization of cation of Fiber Reinforced Polymer (FRP), several scholars have
square CFST in modern building structures is relatively extensive due to attempted to integrate the pultruded FRP profiles into CFST in order to
its simplified beam-column joint connections and enhanced bending enhance its ductility and strength, and a favorable enhancement effect
stiffness [6–8]. In fact, the circular steel tube provides a relatively uni­ was also achieved [26–28].
form confining effect on the core concrete, while the flat position of the Among the SRCFSTs, there are various cross-sectional forms of the
square steel tube hardly provides any confining effect, hence the local inserted steel bones, and the studies on square CFSTs reinforced with
buckling of square steel tube is more prone to appear, and the ductility of different forms of steel bones have been carried out by related scholars.
square CFST is obviously weaker than that of circular CFST [1–3]. In Specifically, for the square CFSTs reinforced with I-shaped profile steel
order to delay the local buckling and enhance the ductility of square or the crossed I-shaped profile steel, Ji et al. [29] studied the hysteretic
CFST, some researchers have implemented various approaches such as properties, Shi et al. [30] analyzed the torsional behavior, Wang et al.
incorporating binding bars [9], welding steel tie-bars [10], and [6] investigated the bending properties, Jia et al. [31] researched the

* Corresponding author.
E-mail address: [email protected] (Y. Ding).

https://doi.org/10.1016/j.jcsr.2023.108414
Received 14 October 2023; Received in revised form 2 December 2023; Accepted 11 December 2023
Available online 21 December 2023
0143-974X/© 2023 Elsevier Ltd. All rights reserved.
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

mechanical properties under long-term axial compression load. For the photo of specimen's cross-section and the schematic diagram of designed
square CFSTs reinforced with steel tube, the axial and eccentric latticed steel angles are presented in Fig. 1. The dimensions and
compression behavior of square section SRCFST with inner square steel nomenclature of the designed specimens are presented in Table 1.
tube were studied by Huang and Zhao et al. [32,33], and the axial
compression behavior of square section SRCFST with inner circular steel 2.2. Mechanical properties of material
tube was studied by Wang et al. [34]. For the square CFSTs reinforced
with reinforcement, the axial and eccentric compression properties of The axial compressive strength test on concrete was conducted on 6
square section SRCFST with inner spiral stirrups were researched by concrete cubes (150 × 150 × 150 mm) in accordance with the GB/T
Yuan et al. [12,13] and Hu et al. [19], and the axial compression 50081–2019 [40], and the average experimental compressive strength
behavior of square section SRCFST with inner reinforcement cages was of concrete was 37.3 MPa.
analyzed by Chen et al. [20] and Xiamuxi et al. [35]. Regarding the The uniaxial tensile test was conducted on steel, with 3 standard
square CFSTs strengthened with latticed steel angles, Wang et al. [36] samples tested in each group. The testing procedure followed the
researched the bending behavior through experiment and numerical guidelines specified in GB/T228.1–2021 [41]. The measured material
method. The aforementioned studies collectively demonstrate that the properties of steel are presented in Table 2.
incorporation of steel bones enhances the stiffness and strength of CFSTs
under various loading conditions, while significantly improving their 2.3. Loading procedure
ductility performance when subjected to compression loads.
As for the mechanical behavior of square CFSTs reinforced with The axial compression load was applied through the displacement
pultruded FRP profiles, some scholars have also carried out relevant control at a loading speed of 1.0 mm/min [21]. The loading diagram and
research work. For example, Li et al. [37–39] investigated the me­ schematic illustration of the axial compression test are depicted in Fig. 2.
chanical properties of square CFSTs reinforced with I-shaped CFRP In order to measure the deformation of the steel tube, several longitu­
profiles under both axial and eccentric compression, the results dinal and transverse strain gauges were arranged on the flat and corner
demonstrated that the insertion of CFRP profiles enhanced the ductility, sections of steel tube at the mid-heigh of specimen, as shown in Fig. 2.
capacity, and stability of the composite members, while also delaying
the local buckling process of the steel tube. Feng et al. [26] and Cheng 2.4. Experimental results
et al. [27] investigated the axial compression behavior of square CFSTs
reinforced with circular GFRP profiles, and developed a three-stage 2.4.1. Failure modes
load-strain relationship model that takes into account the thickness of The experimental failure modes are illustrated in Fig. 3. As shown in
the inner GFRP tube and the outer steel tube, and the concrete strength, Fig. 3, the occurrence of significant local buckling near the mid-height is
the research findings confirmed that the presence of inner GFRP tube observed, and the local buckling is predominantly concentrated in the
can enhance ductility, post-yield stiffness, and residual load. Lai et al. vicinity of the mid-height of the specimens. For specimens Y2504-L505,
[28] also investigated the axial compression behavior of square CFSTs Y3005-L404 and Y3506-L505, the outer steel tubes were removed after
reinforced with circular GFRP profiles and confirmed that the inserted test, and the concrete located in the local buckling position of the steel
GFRP tubes both enhance the strength and ductility of the composite tube is observed to be crushed, as shown in Fig. 3. Moreover, the steel
members, while also exerting a secondary confining effect on the core angles also experienced obvious local buckling, which occurred
concrete. approximately at the same locations where the steel tubes buckled. The
In summary, the aforementioned findings collectively demonstrate aforementioned observations demonstrate the effectiveness of collabo­
that incorporating diverse stiffening elements into square CFSTs leads to rative interaction between the internal steel angles and the CFSTs.
favorable improvements in their compression properties. Therefore, this Besides, it was found in Fig. 3 that the concrete enclosed by the steel
study aims to insert latticed steel angles into square CFSTs in order to angles remained uncrushed, indicating an additional confining effect
investigate their potential for enhancing the axial compression proper­ exerted by the steel angles on the enclosed concrete.
ties. In this paper, the load-displacement and load-strain behaviors of
the tested specimens were obtained through experiment, and the impact 2.4.2. Load-displacement behavior
of latticed steel angles on experimental results such as yield load, ulti­ Fig. 4 shows the load-displacement curves. In Fig. 4, the experi­
mate load, and ductility was analyzed. Subsequently, the axial mental results demonstrate that the compression load exhibits a
compression mechanisms were discussed using the developed finite decrease after reaching the ultimate load, regardless of whether the
element model, and the main influencing parameters on axial specimens are reinforced with steel angles or not, which indicates that
compression properties were also analyzed. Finally, design formulas for the inclusion of steel angles has minimal influence on the overall trend
calculating the ultimate load of square CFSTs reinforced with latticed of the load-displacement curves of square CFSTs. Furthermore, the
steel angles were proposed based on the parametric analysis results. compression load of the specimen group with widths of 250 mm and
300 mm exhibit a gradual decrease after N85%, while the specimen group
2. Experimental investigations with a width of 350 mm demonstrates a pronounced decline after N85%,
as depicted in Fig. 4. The observed phenomenon can be attributed to the
2.1. Description of specimens fact that specimens with a larger width of steel tube tend to exhibit an
increased magnitude of local buckling, thereby resulting in a premature
The test included a total of 9 specimens, each with a net length of tearing of the weld.
1000 mm, excluding the two end plates. The exterior width of steel tube Moreover, Fig. 4 illustrates the increase of cross-sectional area of
(Bo) of the designed specimens was classified into three options: 250 internal steel angles has a certain enhancing effect on the ultimate load,
mm, 300 mm, and 350 mm. The designed latticed steel angles were and the enhancing effect decreases for specimens with larger size of steel
fabricated by welding 4 equal-edge steel angles together using the tube, indicating that the compressive load is mainly contributed by the
splicing plates, the dimensions for the equal-edge steel angles were set as concrete.
40 × 40 × 4 mm and 50 × 50 × 5 mm. The spacing between all splice
plates in this study was uniformly set at 200 mm. For specimens with a 2.4.3. Ultimate load and ductility
Bo of 250 mm, the dimensions of the splice plates were specified as 90 × The ultimate load (Nu) and ductility index (DI) for all specimens are
50 × 4 mm; for specimens with a Bo of 300 mm and 350 mm, the di­ provided in Table 3. In Table 3, DI represents the specimens' capacity for
mensions of the splice plates were specified as 120 × 50 × 4 mm. The undergoing plastic deformation without experiencing intrinsic strength

2
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 1. Diagrams of the cross-section and latticed steel angles of specimen.

Table 1
The dimensions and key parameters of the tested specimens.
Specimen Steel tube Steel angles αo (Aso/Ac) αi (Asi/Ac) L/mm

Bo/mm to/mm R/mm Type b/mm ti/mm h/mm

Y2504 250 4 24 / / / / 6.72% 0 1000

Y2504-L404 250 4 24 L40 ×4 40 4 130 6.86% 2.12% 1000
Y2504-L505 250 4 24 L50 ×5 50 5 130 6.95% 3.35% 1000
Y3005 300 5 24 / / / / 7.02% 0 1000
Y3005-L404 300 5 24 L40 ×4 40 4 160 7.12% 1.47% 1000
Y3005-L505 300 5 24 L50 ×5 50 5 160 7.18% 2.31% 1000
Y3506 350 6 24 / 7.23% 0 1000
W3506-L404 350 6 24 L40 ×4 50 5 160 7.30% 1.08% 1000
W3506-L505 350 6 24 L50 ×5 50 5 160 7.35% 1.69% 1000

Note: The nomenclature of a specimen was determined in a specific sequence based on the dimensions of the steel tube and steel angles. For instance, the specimen
labeled as “Y3005-L505” indicated the dimension of outer steel tube (Bo × to) was 300 mm × 5 mm, while the dimension of steel angles (b × ti) was 50 mm × 5 mm.

effect on the yield load (Ny), but the enhancing effect decreases as the
Table 2
exterior diameter of the steel tube increases. In Fig. 6(b) and Fig. 6(c), it
Material properties of steel.
is clear that the internal steel angles can obviously improve the ultimate
Material Bo to fy/MPa fu/MPa Es/GPa load (Nu) and ductility coefficient (DI), with a pronounced enhancing
Steel tube 250 4 351 486 231 effect observed on DI. Similarly, the enhancing effect of internal steel
300 5 369 529 238 angles on ultimate load (Nu) and ductility index (DI) also exhibit a
350 6 498 583 222 decreasing trend as the exterior diameter of the steel tube increases. This
Steel angles b ti fy/MPa fu/MPa Es/GPa
40 4 330 482 226
is mainly because when the size of steel tube increases, the value of the
50 5 299 441 227 steel angles' area to concrete's area (i.e., αi in Table 1) diminishes, and
the contribution of steel angles to the axial compression load is conse­
Note: In Table 2, fy is the yield strength, fu is the ultimate strength, and Es is the
quently diminished. In summary, the enhancing effect of steel angles on
elastic modulus.
the yield load (Ny), ultimate load (Nu) and ductility index (DI) decrease
with the increase of steel tube's size.
failure, which is defined as follows [42]:
In Fig. 6(c), although the steel angles have a certain influence on the
Δ85% ductility index (DI), there is no strict linear relationship between the
DI = (1)
Δu increment of steel angles' area and the increment of DI. This phenome­
non suggests that the ductility can be optimally improved by achieving
Where, Δ85% represents the displacement at the point where the
an optimal value of αi, and the value of αi could be influenced by factors
specimen's load decreases to 85% of the ultimate load, and Δu represents
such as the size of steel angles and the size of steel tube.
the displacement associated with the ultimate load [42].
Additionally, Table 3 also presents the yield load (Ny) and yield
2.4.4. Load-strain behavior
displacement (Δy) of the tested specimens, and the definitions of Ny and
To analyze the deformation behavior of the steel tubes, Fig. 7 pre­
Δy are derived from the “Farthest point method” as described in refer­
sents the load-strain curves at the mid-height section for two specimen
ence [42], and the detailed definition of the “Farthest point method”
groups: Y2504 and Y3005. The strain values were obtained by averaging
[42] is shown in Fig. 5. Moreover, the elastic stiffness is taken as the
the measurements taken at both the mid and corner positions at the mid-
scant stiffness corresponding to 0.7Ny [28], as shown in Fig. 5.
height section, and the strain measurement results exhibit a negative
To analyze the influence of internal steel angles on the yield load Ny,
value for compressive strain and a positive value for tensile strain.
ultimate load Nu and ductility index DI, the normalized values of Ny, Nu,
In Fig. 7, the strain distortion at the corner and plate position of steel
and DI were obtained (assuming the vales of Ny, Nu, and DI of the
tube is evident for specimen groups Y2504 and Y3005 under an axial
counterpart CFSTs are all equal to 1). The comparisons of the increment
load ranging from 0.8Nu to Nu, which indicates that the local buckling
in Ny, Nu, and DI of the tested specimens are illustrated in Fig. 6.
occurs at both of the corner and the plate position. As depicted in Fig. 7,
In Fig. 6(a), the internal steel angles have a moderate improving
when the axial compressive load is <0.8Nu, the longitudinal and

3
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 2. Loading diagram of the axial compression test.

transverse strains at the corner position of most specimens exhibit a Where, x = ε/εo; y = σ/fc’; fc’ and fck are the cylinder and charac­
lower magnitude compared to that observed at the plate position, and teristic compressive strength of concrete, respectively; εo is the strain
the phenomenon can be attributed to a higher confining effect on con­ corresponding to the peak stress of fc’.
crete exerted by the steel tube at the corner position, thereby resulting in Moreover, the compressive plastic damage factor of concrete (dc) was
a pronounced restraint effect on the deformation of steel tube within this also considered based on the energy equivalent hypothesis [44–47],
specific region. which is expressed as follows:
As shown in Fig. 7, with the exception of specimen Y3005, the ⎧ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
magnitudes of longitudinal and transverse strain for all other specimens ⎪
⎪ f ′c
⎪
⎪ 1 − ⋅(2 − x) (x ≤ 1)
at the ultimate load are below 0.002 and 0.001 respectively, and the ⎨ E0 εo
results presented in Table 3 demonstrate that the overall axial strain (i. dc = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ (7)
⎪
⎪
e., Δu/L) of all specimens exceeds 0.002 at the ultimate load. The ⎪
⎪
⎩1 − f ′c 1
⋅ (x > 1)
aforementioned observation suggests that despite the local buckling of E0 εo β0 (x − 1)1.6+1.5/x + x
steel tube occurs before the ultimate load, the concrete is still capable of
bearing the axial compression load until reaching the ultimate load.
2) For steel, it is acknowledged that the mechanical properties of the
However, when the ultimate load is reached, a noticeable decrease in the
plate and corner positions of square steel tube are different [2],
axial compression load can be observed due to concrete crushing.
hence the stress-strain relationship of the steel at the plate and corner
positions are defined separately.
3. Finite element analysis
For the plate position of square steel tube, the “five-stage” stress-
3.1. Overview of the model strain model based on the Von-Mises yield criterion [2,45] is adopted
for the constitutive model of steel, which is expressed as follows:
For a precise understanding of the working mechanism and the ⎧ ( / )
impact of key parameters on the compression behavior, the ABAQUS ⎪
⎪ Es ε
⎪
⎪
ε ≤ εe , εe = 0.8fy Es
6.14 software [43] was employed for subsequent analysis, and the finite ⎪
⎪
⎪
⎪ − Aε 2 + B ε + C (εe < ε ≤ εe1 , εe1 = 1.5εe )
element model was established according to the following 5 principles. ⎪
⎪
⎪
⎨
fy (εe1 < ε ≤ εe2 , εe2 = 10εe1 )
σ= (8)
1) For concrete, the stress-strain relationship proposed by Han et al. [2] ⎪
⎪
⎪
[
ε − εe2
]
⎪
⎪
for the square steel tube confined concrete is adopted, which is ⎪
⎪
⎪
fy 1 + 0.6⋅
εe3 − εe2
(εe2 < ε ≤ εe3 , εe3 = 100εe1 )
⎪
expressed as follows: ⎪
⎪
⎩ 1.6f
y (ε > ε )
e3
⎧
⎪
⎨ 2x − x2 (x ≤ 1)
/
y= x (2) A = 0.2fy (εe1 − εe )2 (9)
⎪
⎩ (x > 1)
β0 (x − 1)1.6+1.5/x + x
B = 2Aεe1 (10)
(fc′)0.1
β0 = √̅̅̅̅̅̅̅̅̅̅̅ (3) C = 0.8fy + Aε2e − Bεe (11)
1.2 1 + ξ
For the corner position of square steel tube, the constitutive model of
Aso f yo
ξ= (4) steel is expressed as follows [2,36]:
Ac fck

εo = εc + 800 • ξ0.2 • 10− 6

(5)

εc = (1300 + 12.5 • fc ’ ) × 10− 6

(6)

4
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 3. Failure modes of tested specimens.

⎧ ( / )
⎪ Es ε ε ≤ εe , εe = 0.75⋅fyl Es of the steel before cold-forming, respectively. Moreover, Es is the elastic
⎪ ( / )
⎨
0.75fyl + Es1 (ε − εe ) ε < ε ≤ εe1 , εe1 = εe + 0.125⋅fyl E/s1 ) modulus of the steel before cold-forming, while Es1, Es2 and Es3 are equal
σ= ( e
⎪ 0.875fyl + Es2 (ε − εe1 ) εe1 < ε ≤ εe2 , εe2 = εe1 + 0.125⋅fyl Es2
⎪ to Es/2, Es/10 and Es/200 [2], respectively.
⎩
fyl + Es3 (ε − εe2 ) (εe2 < ε)
(12) 3) The specimen described in this paper consists of 4 components:
concrete, a steel tube, steel angles, and two end plates. The defini­
[ ]
Bc tions of the interactions [46,47] between each component are
fyl = 0.6⋅ m + 0.4 ⋅fy (13)
(r/to ) depicted in Fig. 8.
4) The mesh selection and division [46,47] of each part of a specimen
( / ) ( / )2
Bc = 3.69⋅ fu fy − 0.819⋅ fu fy − 1.79 (14) are depicted in Fig. 8. Prior to the analysis, a comprehensive mesh
sensitivity analysis was conducted following the methodology out­
( / )
m = 0.192⋅ fu fy − 0.068 (15) lined in references [46,47].
5) In this paper, only half of the axial compression specimen is analyzed
Where, r is the inner radius of the chamfer; fyl is the yield strength of due to its symmetry property [46,47], and a symmetric boundary
the cold-formed steel; fu and fy are the tensile strength and yield strength

5
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 4. Experimental load-displacement curves.

Table 3
Ultimate load and ductility.
1
Specimen Nu/kN Δu/mm N85% kN Δ85%/mm Ny/kN Δy/mm Δe/mm K/kN•mm− DI Ny/Nu

Y2504 2982 3.64 2535 7.11 2427 1.59 1.02 1674 1.95 81.4%
Y2504-L404 3348 3.25 2846 12.63 2887 1.54 1.11 1829 3.89 86.2%
Y2504-L505 3579 2.60 3042 9.17 3095 1.30 0.96 2269 3.53 86.5%
Y3005 4475 3.27 3804 6.41 3909 2.07 1.48 1849 1.96 87.4%
Y3005-L404 4877 3.69 4145 11.12 3957 1.68 1.13 2462 3.01 81.1%
Y3005-L505 4905 5.61 4169 16.65 4351 3.07 2.21 1378 2.97 88.7%
Y3506 7131 2.35 6061 7.13 6013 1.49 1.01 4188 3.03 84.3%
W3506-L404 7566 4.05 6431 15.98 6212 2.71 1.93 2259 3.95 82.1%
W3506-L505 7628 4.07 6484 14.32 5874 1.91 1.20 3427 3.52 77.0%

condition is applied on the symmetry surface [46,47]. The two ends

of the specimen are subjected to different boundary conditions, with
one end being fixed and the other end being hinged [47]. Specif­
ically, the hinged end serves as the point where axial compression
load is applied [47].

The established finite element model using the aforementioned

method is illustrated in Fig. 8.

3.2. Verifications of the model

The comparison of failure modes and load-displacement curves are

presented in Fig. 9 and Fig. 10, the comparison of ultimate load is given
in Table 4.
In Fig. 9, it can be observed that the finite element model accurately
replicates the failure modes observed in the tested specimens, including
local buckling of steel tubes near the mid-height section, crushing of
concrete near the mid-height section (the crushed concrete described in
FE model indicates that DAMAGEC exceeds 0.95 [47,48]), as well as
Fig. 5. Definition of the “Farthest point method”.
deformation of steel angles. In Fig. 10, except for the specimen group of
Y3506, the simulated load-displacement curves depicted in Fig. 10

Fig. 6. Comparison of the increment in Ny, Nu and DI.

6
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 7. Load-strain curves.

Fig. 8. Schematic diagram of the finite element model [2,46,47].

accurately capturing the ascending, descending, and plastic stages of the 1.018, accompanied by a coefficient of variation (Cov) of 1.46%, indi­
experimental curves, which demonstrate a good agreement with the cating that the proposed FE model can provide a relatively accurate
experimental results. For the specimen group of Y3506, an obvious prediction for the ultimate load with minimal deviation.
discrepancy between the simulated results and test results is observed, In conclusion, the comparison results revealed that the established
and the reason of this phenomenon shown in Fig. 10(c) is that specimens finite element model can accurately predict the axial compression
with larger width of steel tube tend to exhibit an increased magnitude of behavior of the tested specimens. Consequently, the proposed FE model
local buckling, thereby resulting in a premature tearing of the weld, thus in this paper can be effectively employed for the subsequent analysis.
resulting in a relatively shorter plastic stage of the specimen group of
Y3506. Furthermore, Table 4 reveals that the average value of NFE/Nue is

7
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 9. Comparison of failure modes.

Fig. 10. Comparison of load-displacement curves.

3.3. Analysis of the axial compression mechanism towards the periphery, with the smallest concrete stress located near the
flat position of the square tube. The observations suggest that the corner
3.3.1. Concrete stress position of the square tube acts as a region of strong confinement, while
The distributions of the concrete stress at mid-height section for the flat position of the square tube exhibits weaker confinement.
typical specimens under the point of ultimate load are illustrated in In Fig. 11, for the square CFST specimens reinforced with steel an­
Fig. 11. The black region in Fig. 11 represents the area where the lon­ gles, the longitudinal stress of the concrete located at the inner corners
gitudinal stress of concrete exceeds the cylinder compressive strength of of the steel angles is also relatively larger than at other locations when
concrete (fc’). the ultimate load is reached, indicating that the inner corners of the steel
In Fig. 11, for the square CFST specimens, the maximum longitudinal angles also belong to the strong confining area.
stress of the concrete in the mid-height section is located at the corner Moreover, by comparing the core regions in CFST specimens with
position of the square steel tube when the ultimate load is reached. and without internal steel angles, it is demonstrated that the presence of
Except for the corner position of the square tube, the concrete stress is steel angles altered the stress distributions in the concrete near the steel
greatest in the core region of the square section and gradually decreases angles, which is primarily caused by the mutual interactions between

8
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Table 4
Comparison of ultimate load.
Specimen fyo/MPa fyi/MPa b/mm ti/mm fc’/MPa Nue/kN NFE/kN Nu1/kN Nu2/kN NFE/ Nue Nu1/ Nue Nu2/ Nue

Y2504 351 0 0 0 29.5 2982 3025 2939 2833 1.014 0.986 0.950
Y2504-L404 351 330 40 4 29.5 3348 3390 3283 3175 1.013 0.981 0.948
Y2504-L505 351 299 50 5 29.5 3579 3537 3418 3309 0.988 0.955 0.925
Y3005 369 0 0 0 29.5 4475 4574 4431 4120 1.022 0.990 0.921
Y3005-L404 369 330 40 4 29.5 4877 4900 4772 4462 1.005 0.978 0.915
Y3005-L505 369 299 50 5 29.5 4905 5084 4905 4595 1.036 1.000 0.937
Y3506 498 0 0 0 29.5 7131 7377 7358 6229 1.034 1.032 0.874
W3506-L404 498 330 40 4 29.5 7566 7701 7686 6564 1.018 1.016 0.868
W3506-L505 498 299 50 5 29.5 7628 7886 7812 6694 1.034 1.024 0.878
Ave 1.018 0.996 0.913
Cov 1.46% 2.33% 3.32%

Note: In Table 4, NFE represents the simulated ultimate load, while Nue represents the experimental ultimate load.

Fig. 11. Longitudinal stress distributions of concrete at the mid-height section.

the steel angles and surrounding concrete. Therefore, the aforemen­ 3.3.2. Axial compression load of individual components
tioned analysis also suggests that the steel angles impose a certain To understand the working mechanism of the tested specimens, the
confining effect on the enclosed concrete and consequently modify the axial loads borne by each component (steel angles, steel tube, and
stress distributions in the adjacent concrete. concrete) [47] are analyzed throughout the entire loading process.

Fig. 12. Axial compression loads borne by each component.

9
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 12 illustrates the relationship between the axial compression loads dimensional aspects [36,47].
borne by each component (steel angles, steel tube, concrete) and the From the material aspect: 1) For steel tube, the analysis on its yield
corresponding axial strain for the typical specimens of Y3005, Y3005- strength (fyo) and thickness (to) are conducted; 2) For steel angles, the
L4, and Y3005-L5. analysis on its yield strength (fyi) and cross-section area (Asi) are con­
In Fig. 12, it can be observed that the components such as steel angles ducted; 3) For concrete, the analysis on its cylindrical compressive
and steel tubes of the specimens maintain a constant axial load after strength (fc’) is conducted.
reaching the ultimate load, without experiencing any decrease From the dimensional aspect: the external width of the steel tubes
throughout the loading process, which is mainly attributed to the (Bo) in the specimens to be analyzed is divided into four specifications:
favorable plastic behavior exhibited by the steel material, coupled with 250 mm, 300 mm, 350 mm and 400 mm.
the restraining effect provided by concrete on steel tube and steel angles
against the local buckling. 3.4.1. Influence of fyo and to
Additionally, as shown in Fig. 12, the steel tube and concrete play a In this section, the influence of yield strength (fyo) and thickness (to)
major role in bearing the axial compression load of the specimens. of steel tube on the axial compressive properties is analyzed. The steel
However, once the ultimate load is reached, there is a noticeable tubes are divided into two groups with dimensions of 250 × 4 mm and
decrease in the contribution of concrete owing to the relatively weaker 300 × 5 mm. The steel angles have a fixed dimension of L50 × 5 mm,
confining effect of the square steel tube. while their yield strength (fyi) remains constant at 300 MPa. Addition­
ally, the concrete strength (fc’) is maintained at a fixed value of 29.5
3.3.3. Variations in the contact pressure MPa. The load-displacement curves of the analyzed specimens are pre­
In order to investigate the confining effect of steel tube and steel sented in Fig. 15 and Fig. 16, respectively.
angles on concrete, Fig. 13 presents the distributions of contact pressure To analyze the influence of fyo and to on the ultimate load, the re­
between concrete and steel tube [47], as well as between concrete and lationships between the increase rate of ultimate load (i.e., δ, defined as
steel angles [47] for specimen Y3005-L5. the ratio of the incremental ultimate load to the ultimate load of a
As depicted in Fig. 13(a), the contact pressure between concrete and specific specimen) and the values of fyo and to are plotted in Fig. 17. As
steel tube at the corners of square steel tube consistently maintains its shown in Fig. 17, the δ value exhibits a nearly linear correlation with fyo
maximum value throughout the entire loading process, while the and to. Additionally, under the assumption of all other factors remaining
negligible contact pressure is observed at other locations. The finding constant, it can be found that the increase of to has a more pronounced
indicates that the confining effect on concrete primarily originates from impact on the ultimate load of specimens with smaller Bo.
the corner positions, while the confining effect of the other positions is
almost negligible. 3.4.2. Influence of fyi and Asi
As shown in Fig. 13(b), the contact pressure at the inner corners of In this section, the influence of yield strength (fyi) and area (Asi) of
the steel angles (i.e., point 3 in Fig. 13(b)) is the highest. Furthermore, steel angles on the axial compressive properties is analyzed. The steel
although the contact pressures at points 2 and 4 are smaller than that at tubes are divided into two groups with dimensions of 250 × 4 mm and
point 3, they are slightly larger than those at points 1 and 5. The 300 × 5 mm. The steel tubes have a fixed yield strength of 300 MPa,
aforementioned findings suggest that the concrete located at the inner while the concrete strength (fc’) is maintained at a fixed value of 29.5
corners of steel angles experiences the most pronounced confining ef­ MPa. The load-displacement curves of the analyzed specimens are pre­
fect, and this effect decreases from the inner corner to the edge of steel sented in Fig. 18 and Fig. 19, respectively.
angles. To analyze the influence of fyi and Asi on the ultimate load, the re­
In summary, considering the findings proposed in the previous lationships between δ and the values of fyi and Asi are plotted in Fig. 20.
research [1–3,49–52], the effective confining regions on the concrete of As depicted in Fig. 20, the δ value also exhibits a nearly linear correla­
square CFSTs, reinforced with or without internal steel angles, are tion with fyi and Asi.
summarized in Fig. 14. Additionally, it can be found in Fig. 20 that: 1) when to is fixed, the
increase of Asi has a more pronounced enhancing effect on the ultimate
load of specimens with smaller Bo; 2) when Bo is fixed, the increase of Asi
3.4. Parametric analysis
has a more pronounced enhancing effect on the ultimate load of speci­
mens with smaller to. In summary, the influence of Asi on the ultimate
In this section, the impact of main parameters on the axial
load is more pronounced for specimens with smaller Bo and to.
compressive properties is analyzed, considering both material and

Fig. 13. Distributions of the contact pressure of typical specimen Y3005-L5.

10
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 14. Distributions of the effective confining regions.

Fig. 15. Influence of fyo on load-displacement curves.

t t

Fig. 16. Influence of to on load-displacement curves.

3.4.3. Influence of fc’ 4. Calculation of the ultimate load

In this section, the impact of concrete's cylinder compressive strength
(fc’) on the axial compressive properties is analyzed. The steel tubes are In this paper, the superposition method [53–55] is used to estimate
divided into two groups with dimensions of 250 × 4 mm and 300 × 5 the ultimate load of square CFSTs reinforced with internal latticed steel
mm. Both the steel tubes and steel angles have a fixed yield strength of angles. The superposition method considers the axial ultimate load (N)
300 MPa, while the steel angles have a fixed dimension of L50 × 5 mm. of the tested specimens as the summation of the ultimate load of square
The load-displacement curves of the analyzed specimens are shown in CFSTs (No) and the ultimate load contributed by the internal steel angles
Fig. 21. The observation from Fig. 21 reveals that an increase in concrete (Ni), and the calculation formula is expressed as follows:
strength leads to higher brittleness in the load-displacement curves, as
N = No + Ni (16)
evidenced by a more pronounced decline of the compression load in the
post-ultimate load stage. The determination of ultimate load of the steel angles (Ni) disregards
To analyze the influence of fc’ on the ultimate load, the relationship the issue of local stability, as they are connected by the splicing plates
between δ and the value of fc’ is plotted in Fig. 22. As depicted in Fig. 22, and encased in concrete [36], therefore, it can be assumed that the
the concrete strength has a significant influence on the ultimate load, strength of steel angles has been fully developed under the ultimate
which is mainly because the concrete contributes significantly to the condition. The expression for the ultimate load of the steel angles is as
axial compressive load of the specimens (see in Fig. 12). follows:

11
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Nt Nt
Nt

Nf Nf
Nt Nt
Nf
Nt

Fig. 17. The relationship between δ and fyo, δ and to.

Fig. 18. Influence of fyi on load-displacement curves.

[ ]
Ni = Asi fyi = 4 • b2 − (b − ti )2 • fyi (17) Where, fck is concrete's characteristic compressive strength, and the
expressions for the parameters of a0 and b0 in Eqs. (21) and (22) can be
Where, b and ti are the width and thickness of single equal-edge steel referred from reference [2].
angle, respectively.

4.1. Design equations based on the unified theory of CFSTs 4.2. Design equations based on the confined concrete theory

Referring to the design equations for the ultimate load (No) of square Referring to the design equations for the ultimate load (No) of square
CFSTs proposed based on the unified theory [2], the design equation for CFSTs proposed based on the confined concrete theory [52,55,56], the
determining the ultimate load capacity of square CFSTs reinforced with design equation for determining the ultimate load capacity of square
internal latticed steel angles is presented as follows: CFSTs reinforced with internal latticed steel angles is presented as fol­
lows [52,55,56]:
Nu1 = (Aso + Ac )⋅fsc + Asi fyi (18)
Nu2 = fcc Ac + Asi fyi (23)
Where, fsc is the combined compressive strength of the square CFSTs,
and it is expressed as follows [2]: ( √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ )
fr,s fr,s
( )
fsc = 1.212 + Bξ + Cξ2 ⋅fck (19) fcc = − 1.254 + 2.254 1 + 7.94 − 2 ⋅fco (24)
fco fco

Aso f yo 2σh to
ξ= (20) fr,s = Ke ⋅ (25)
Ac fck Bo -2to
( )a
fyo 0 σ h = βfyo (26)
B = 0.1381⋅ + 0.7646 (21)
235
Where, fcc is the compressive strength of confined concrete; fco is the
( )b0 compressive strength of unconfined concrete; and fr,s is the effective
fck
C = − 0.0727⋅ + 0.0216 (22) confining stress generated by the square steel tube; β is the reduction
20
factor related with the yield strength of steel tube, according to

12
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

b t

b t

b t
b t

Fig. 19. Influence of Asi on load-displacement curves.

Nf Nf
Nf

N N
N

Fig. 20. The relationship between δ and fyi, δ and Asi.

references [52,55,56], the formula for calculating the coefficient of β is

Aw
as follows: Ke = 1 − (28)
Ac
√̅̅̅̅̅̅̅̅̅̅̅
β = 0.1⋅ Bo /to (27) Where, Aw is the area of the concrete in the ineffectively confined
The application range of coefficient β is: 50 ≤ Bo/to ≤ 100 [49]. zone of square steel tube. The formula for calculating Aw provided in
In Eq. (25), Ke is the effective confinement coefficient for the square reference [51] is as follows:
steel tube, which is expressed as follows [51]:
Aw = 0.571 • (Bo − 2R − 6to )2 (29)
Where, R is the chamfer radius of square steel tube.

13
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 21. Influence of fc’ on load-displacement curves.

theory of CFST [1], and a lot of reliability analyses have been conducted
by Han et al. [1] to substantiate its safety in structural design, and the
analysis results indicate that the relevant design formula is both feasible
and reliable. Therefore, the reliability analysis on Eq. (23) is only con­
ducted in this paper.

5.1. Calculation method

The reliability analysis on Eq. (23) is calculated using the First-order

Second-moment (FOSM) method as described in the Chinese standard
GB50068–2018 [57].
The resistance of structure is denoted as R, the load effect is denoted
as S, including the dead load SG and live load SQ, and the limit state
function Z is expressed as follows [57]:
( )
Z = R − S = K0 RP − γ 0 γ G SGk + γQ SQk (30)

Where, SGk and SQk are the standard values of dead load and live
load, respectively; γ G and γQ are the partial coefficients of dead load and
live load, respectively, which are taken as 1.3 and 1.5 according to
GB50068–2018 [57].
Fig. 22. The relationship between δ and fc’.
γ0 is the structural importance factor, and the second safety level of
general industry and civil buildings is chosen in the reliability analysis,
hence, γ 0 is set as 1.0 [58].
4.3. Comparison of calculation results
K0 is the uncertainty of structural resistance calculation model, and
the value of K0 is determined based on the data provided in Table 5. RP is
The ultimate loads of Nu1 and Nu2, which are calculated using Eqs.
the structural resistance calculated by Eq. (23), which hinges on the
(18) and (23) respectively, are compared with the experimental results
geometrical configurations and material properties [59], and the spe­
(Nue) and finite element analysis results (NFE). The comparison data are
cific calculation methods of the mean value and coefficient of variation
presented in Table 4 and Table 5, and the distributions of these data
of RP are referred from GB50068–2018 [57].
points are illustrated in Fig. 23.
According to the definitions provided in references [57, 58, 60, 61],
The mean values of Nu1/Nue and Nu1/NFE are 0.996 and 0.950,
the reliability index β can be determined by the following formula:
respectively, with corresponding coefficients of variation (Cov) of
2.33% and 1.01%, as indicated by the data presented in Table 4 and μ*R − μ*SG − μ*SQ
Table 5. Similarly, the mean values of Nu2/Nue and Nu2/NFE are 0.913 β = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
̅ (31)
σ*R + σ*SG + σ *SQ
2 2 2
and 0.930, respectively, accompanied by coefficients of variation of
3.32% and 3.57%. Furthermore, Fig. 23 demonstrates that the predic­ Where, μ* and σ * are the mean value and the standard deviation of
tion error of Eq. (18) remains within a range of <10%. In summary, both the corresponding variables, respectively.
formula (18) and formula (23) can conservatively predict the ultimate
load of square CFSTs reinforced with internal latticed steel angles. 5.1.1. Combinations of load
However, comparative results indicate that the formula proposed based According to references [59,62], this paper considers three pre­
on the unified theory of CFSTs demonstrates superior accuracy and a dominant forms of load combination, that is: a) SG + SQO; b) SG + SQR; c)
smaller deviation. SG + SQW. The specific meanings and statistical values of symbols SG,
SQO, SQR and SQW are shown in Table 6. Moreover, the ratio of SQ to SG is
5. Reliability analysis determined as 3 based on the research provided by Zhao et al. [58] and
An et al. [60], and the target reliability index of β0 is chosen as 3.2
In this paper, two design equations (i.e., Eq. (18) and Eq. (23)) are [58,60] for the second safety level of general industry and civil buildings
presented based on the superposition method. The design equation for is chosen in the reliability analysis.
the ultimate load of CFST in Eq. (18) was proposed based on the unified In Table 6, K = mean value/standard value; COV is the coefficient of

14
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Table 5
Comparison of calculation results of the ultimate load.
Bo × to/mm fyo/MPa fyi/MPa b/mm ti/mm h/mm fc’/MPa NFE/kN Nu1/kN Nu2/kN Nu1/NFE Nu2/NFE

250 × 3 300 300 50 5 130 29.5 3078 2936 3034 0.954 0.986
250 × 4 300 / / / / 29.5 2857 2718 2671 0.951 0.935
300 300 40 4 130 29.5 3184 3030 2981 0.952 0.936
300 200 50 5 130 29.5 3180 3015 2965 0.948 0.932
300 300 50 5 130 29.5 3370 3205 3155 0.951 0.936
300 400 50 5 130 29.5 3560 3395 3345 0.954 0.939
300 500 50 5 130 29.5 3750 3585 3535 0.956 0.943
200 300 50 5 130 29.5 2954 2810 2801 0.951 0.948
400 300 50 5 130 29.5 3781 3634 3449 0.961 0.912
500 300 50 5 130 29.5 4190 4086 3699 0.975 0.883
300 300 50 5 130 20.6 2842 2725 2605 0.959 0.917
300 300 50 5 130 38.4 3917 3626 3584 0.926 0.915
300 300 50 5 130 47.3 4458 4077 4015 0.915 0.901
300 300 60 6 130 29.5 3597 3419 3367 0.951 0.936
250 ×5 300 300 50 5 130 29.5 3667 3472 3249 0.947 0.886
250 ×6 300 300 50 5 130 29.5 3963 3736 3326 0.943 0.839
300 ×3 300 300 50 5 160 29.5 3992 3813 3992 0.955 1.000
300 ×4 300 / / / / 29.5 3835 3647 3673 0.951 0.958
300 300 50 5 160 29.5 4350 4140 4161 0.952 0.956
300 300 60 6 160 29.5 4580 4357 4375 0.951 0.955
300 300 70 7 160 29.5 4847 4614 4629 0.952 0.955
300 × 5 300 / / / / 29.5 4189 3980 3813 0.950 0.910
300 200 50 5 160 29.5 4511 4276 4107 0.948 0.910
300 300 50 5 160 29.5 4701 4466 4297 0.950 0.914
300 400 50 5 160 29.5 4892 4656 4487 0.952 0.917
300 500 50 5 160 29.5 5082 4846 4677 0.954 0.920
200 300 50 5 160 29.5 4076 3868 3788 0.949 0.929
400 300 50 5 160 29.5 5318 5113 4721 0.961 0.888
500 300 50 5 160 29.5 5929 5795 5082 0.977 0.857
300 300 50 5 160 20.6 3930 3769 3505 0.959 0.892
300 300 50 5 160 38.4 5493 5076 4916 0.924 0.895
300 300 50 5 160 47.3 6283 5733 5540 0.912 0.882
300 300 60 6 160 29.5 4929 4680 4510 0.949 0.915
300 300 70 7 160 29.5 5194 4932 4761 0.950 0.917
300 × 6 300 300 50 5 160 29.5 5050 4789 4410 0.948 0.873
350 × 4 300 / / / / 29.5 4943 4705 4811 0.952 0.973
300 300 40 4 200 29.5 5272 5023 5126 0.953 0.972
300 300 50 5 200 29.5 5463 5202 5303 0.952 0.971
300 300 60 6 200 29.5 5692 5421 5520 0.952 0.970
300 300 70 7 200 29.5 5953 5679 5775 0.954 0.970
350 × 5 300 / / / / 29.5 5359 5094 4997 0.951 0.932
300 300 40 4 200 29.5 5683 5408 5309 0.952 0.934
300 300 50 5 200 29.5 5875 5585 5484 0.951 0.934
300 300 60 6 200 29.5 6100 5801 5699 0.951 0.934
300 300 70 7 200 29.5 6370 6056 5953 0.951 0.935
350 × 6 300 / / / / 29.5 5767 5481 5153 0.950 0.894
300 300 40 4 200 29.5 6097 5792 5463 0.950 0.896
300 300 50 5 200 29.5 6283 5966 5637 0.950 0.897
300 300 60 6 200 29.5 6505 6180 5850 0.950 0.899
300 300 70 7 200 29.5 6778 6432 6102 0.949 0.900
400 × 4 300 / / / / 29.5 6204 5890 6086 0.949 0.981
300 300 50 5 200 29.5 6712 6390 6581 0.952 0.980
300 300 60 6 200 29.5 6940 6610 6798 0.952 0.980
300 300 70 7 200 29.5 7212 6870 7056 0.953 0.978
300 300 80 8 200 29.5 7514 7170 7353 0.954 0.979
400 × 5 300 / / / / 29.5 6660 6336 6319 0.951 0.949
300 300 50 5 200 29.5 7163 6831 6810 0.954 0.951
300 300 60 6 200 29.5 7397 7048 7026 0.953 0.950
300 300 70 7 200 29.5 7674 7306 7282 0.952 0.949
300 300 80 8 200 29.5 7974 7603 7576 0.953 0.950
400 × 6 300 / / / / 29.5 7120 6780 6518 0.952 0.915
300 300 50 5 200 29.5 7650 7270 7006 0.950 0.916
300 300 60 6 200 29.5 7875 7485 7221 0.950 0.917
300 300 70 7 200 29.5 8142 7740 7474 0.951 0.918
300 300 80 8 200 29.5 8451 8033 7767 0.951 0.919
Average 0.950 0.930
Cov 1.01% 3.57%

variation. uncertainties of calculation model (i.e., K0) [63]. Referring to the defi­
nitions provided by Zhao et al. [58] and An et al. [60], the mean values
5.1.2. Uncertainties of resistance of Aso, Asi, Ac and other geometric parameters are taken as 1.0 with a
The uncertainties of resistance include the uncertainties of material corresponding coefficient of variation (COV) of 0.05 to represent the
properties, the uncertainties of geometric parameters and the uncertainties of geometric parameters. Moreover, the variables of

15
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

Fig. 23. Distributions of the comparison data points of the ultimate load.

Table 6 Table 8
Statistical parameters of load variables [59,62]. Statistical parameters of concrete.
Load type Dead Office floor Residential floor Wind load Concrete C30 C40 C50 C60 C70 C80 C90
load SG live load SQO live load SQR SQW
fck /MPa 20.1 26.8 32.4 38.5 44.5 50.2 55.7
Distribution Normal Extreme type I Extreme type I Extreme fcd /MPa 14.3 19.1 23.1 27.5 31.8 35.9 39.8
type type I K 1.374 1.342 1.337 1.332 1.292 1.262 1.232
K 1.060 0.700 0.859 0.999 COV 17.2% 15.6% 14.9% 14.1% 12.1% 10.1% 9.1%
COV 7.0% 29.0% 23.3% 19.3%

In addition, it can also be found in Fig. 24 that: 1) Under the three

geometric parameters all obey the normal distribution [64]. predominant forms of load combination, the order of the calculated
Referring to references [58,59,64–66], the valued statistical pa­ value of β is as follows: SG + SQO > SG + SQR > SG + SQW; 2) The
rameters of steel are listed in Table 7. calculated value of β under the load combinations of SG+ SQO and SG+
In Table 7, K = mean value/standard value; COV is the coefficient of SQR are >3.2 (the target reliability index); 3) Under the load combina­
variation; fyk and fyd are the standard value and design value of steel tion of SG + SQW, the calculated value of β of specimens with normal-
strength, respectively. strength concrete is basically larger than 3.2, while the calculated
Referring to reference [64], the valued statistical parameters of value of β of some specimens with high-strength concrete is lower than
concrete are listed in Table 8. 3.2, and the minimum calculated value of β is 3.14. Therefore, it can be
In Table 8, K = mean value/standard value; COV is the coefficient of concluded that the calculation model of Eq. (23) meets the reliability
variation; fck and fcd are the standard value and design value of the prism requirements for specimens with ordinary concrete.
compressive strength of concrete, respectively.
6. Conclusions
5.2. Reliability analysis results
The axial compression behavior of square CFSTs reinforced with
According to the calculation method and specific parameters pre­ internal latticed steel angles was investigated through experiment and
sented above, the reliability analysis on the square CFSTs reinforced finite element analysis, and the reliability analysis was also conducted to
with internal latticed steel angles studied in this paper is carried out, and evaluate the feasibility of the proposed design equation. Based on the
the obtained distribution diagrams of the reliability index β under the findings presented above, the following conclusions can be drawn:
three predominant forms of load combination are shown in Fig. 24. The
ranges of parameters of the examples are: Bo = 350 mm, b × ti = 50 × 5 1) The collaborative interaction between the internal steel angles and
mm, fyo = 235–500 MPa, fck = 30–90 MPa, α = 0.04–0.22 (α is the steel CFST was observed during testing, the steel angles exerted an addi­
ratio, and α = Aso/Ac). tional confining effect on the enclosed concrete and also altered the
As can be seen from Fig. 24, the following three rules can be stress distributions of the adjacent concrete;
observed: 1) The reliability index β decreases with an increase in the 2) The inclusion of internal steel angles had minimal influence on the
steel ratio α; 2) For specimens with normal-strength concrete, the reli­ overall trend of load-displacement curves of square CFSTs, because
ability index β generally increases with an increase in the concrete the compressive load was mainly contributed by concrete, however,
strength, while for specimens with high-strength concrete, the reliability the increase of Asi still had a certain enhancing effect on the yield
index β decreases with an increase in the concrete strength; 3) For load (Ny), ultimate load (Nu) and ductility coefficient (DI), and the
specimens with a higher steel ratio α, a lower yield strength of steel leads enhancing effect was more pronounced for DI;
to a lower reliability index β. 3) The corner position of square steel tube acted as a region of strong
confining effect and the flat position of square steel tube exhibited
Table 7 almost no confining effect on concrete, and the concrete located in
Statistical parameters of steel (t ≤ 16 mm). the inner corners of steel angles experienced the highest level of
confining effect;
Steel Q235 [64] Q345 [64] Q390 [66] Q420 [59] Q500 [58]
4) Among the main influencing parameters of ultimate load, the in­
fyk /MPa 235 345 390 420 500 crease of to had a more pronounced enhancing effect on the ultimate
fyd /MPa 215 305 345 375 445
load for specimens with smaller Bo, and the increase of Asi had a more
K 1.080 1.090 1.097 1.099 1.103
COV 8.0% 7.0% 7.1% 5.1% 4.7%

16
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

(a) Normal-strength concrete (b) High-strength concrete

(1) Load combination of SG+ SQO

(a) Normal-strength concrete (b) High-strength concrete

(2) Load combination of SG+ SQR

(a) Normal-strength concrete (b) High-strength concrete

(3) Load combination of SG+ SQW

Calculated reliability index β under the three predominant forms of load combination

Fig. 24. Calculated reliability index β under the three predominant forms of load combination.

pronounced enhancing effect on the ultimate load for specimens with Writing – original draft, Writing – review & editing. Zi-Ming Yang:
smaller Bo and to; Conceptualization, Formal analysis, Investigation, Methodology,
5) The load combinations of SG + SQO and SG + SQW yielded the highest Writing – original draft, Writing – review & editing. Xiang-Long Zheng:
and lowest reliability index of β, and the value of β decreases as the Data curation, Funding acquisition, Project administration, Supervision,
steel ratio α increases, and for specimens with a higher α, a lower Visualization. Yang Ding: Conceptualization, Data curation, Funding
yield strength of steel tube leads to a decreased β. acquisition, Project administration, Resources, Supervision, Validation,
Visualization, Writing – review & editing.
CRediT authorship contribution statement
Declaration of Competing Interest
Jun Wang: Conceptualization, Data curation, Formal analysis,
Funding acquisition, Investigation, Methodology, Project administra­ We declare that we have no financial and personal relationships with
tion, Resources, Software, Supervision, Validation, Visualization, other people or organizations that can inappropriately influence our

17
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

work, there is no professional or other personal interest of any nature or [24] J.M. Cai, J.L. Pan, Y.F. Wu, Mechanical behavior of steel-reinforced concrete-filled
steel tubular (SRCFST) columns under uniaxial compressive loading[J], Thin-
kind in any product, service and/or company that could be construed as
Walled Struct. 97 (2015) 1–10.
influencing the position presented in, or the review of, the manuscript [25] X. Wei, W.S. Chen, H. Hong, W.D. Wang, Experimental and numerical studies on
entitled “Axial Compression Behavior of Square Section Concrete-filled square steel-reinforced concrete-filled steel tubular (SRCFST) members subjected
Steel Tubes Reinforced with Internal Latticed Steel Angles”. to lateral impact[J], Thin-Walled Struct. 160 (2021), 107409.
[26] P. Feng, S. Cheng, Y. Bai, et al., Mechanical behavior of concrete-filled square steel
tube with FRP-confined concrete core subjected to axial compression[J], Compos.
Data availability Struct. 123 (2015) 312–324.
[27] S. Cheng, P. Feng, Y. Bai, et al., Load-strain model for steel-concrete-FRP-concrete
columns in axial compression[J], J. Compos. Constr. 20 (5) (2016) 04016017.
Data will be made available on request. [28] Z. Lai, J. Yan, Y. Wang, et al., Axial compressive behavior and design of high-
strength square concrete-filled steel tube short columns with embedded GFRP
Acknowledgements tubes[J], J. Constr. Steel Res. 207 (2023), 107955.
[29] S.H. Ji, W.D. Wang, W. Xian, et al., Cyclic and monotonic behaviour of steel-
reinforced concrete-filled steel tubular columns[J], Thin-Walled Struct. 185
This work was supported by the Guangdong Basic and Applied Basic (2023), 110644.
Research Foundation [grant number: 2021A1515110804], the National [30] Y.L. Shi, Z.L. Jia, W.D. Wang, et al., Experimental and numerical study on torsional
behaviour of steel-reinforced concrete-filled square steel tubular members[J],
Natural Science Foundation of China [grant numbers: 51908016, Structures 32 (2021) 713–730.
52208217]. [31] Z.L. Jia, W.D. Wang, Y.L. Shi, et al., Performance of steel-reinforced concrete-filled
square steel tubular members under sustained axial compression loading[J], Eng.
Struct. 264 (2022), 114464.
References
[32] Y. Huang, H. Liang, Y. Lu, et al., Axial behaviour of square stub CFST columns
strengthened with square steel tube and HPC jacket[J], J. Constr. Steel Res. 186
[1] L.H. Han, W. Li, R. Bjorhovde, Developments and advanced applications of (2021), 106896.
concrete-filled steel tubular (CFST) structures: members[J], J. Constr. Steel Res. [33] P. Zhao, Y. Huang, Y. Lu, et al., Eccentric behaviour of square CFST columns
100 (2014) 211–228. strengthened using square steel tube and high-performance concrete jackets[J],
[2] L.H. Han, Concrete Filled Steel Tubular Structures-Theory and Practice (Third Eng. Struct. 253 (2022), 113772.
Edition) [M], Science Press, Beijing, 2016 (in Chinese). [34] W.H. Wang, W. Li, Q.Y. Song, et al., Performance of steel-reinforced square CFST
[3] L.H. Han, Y.F. Yang, Modern Concrete-Filled Steel Tubular Structures Technology stub columns loaded in axial compression: experiments[J], J. Constr. Steel Res. 191
(Second Edition) [M], China Architecture & Building Press, 2007 (in Chinese). (2022), 107197.
[4] L.H. Han, G.H. Yao, X.L. Zhao, Tests and calculations for hollow structural steel [35] A. Xiamuxi, Z. Wang, B. Yang, Optimum reinforcement ratio of axially loaded
(HSS) stub columns filled with self-consolidating concrete (SCC)[J], J. Constr. Steel reinforced concrete-filled square steel tube column[J], J. Constr. Steel Res. 203
Res. 61 (9) (2005) 1241–1269. (2023), 107805.
[5] X. Yang, C. Tang, Y. Chen, et al., Compressive behavior of steel-reinforced [36] J. Wang, J.R. Li, H. Li, et al., Behaviour of square concrete-filled steel tubes
concrete-filled square steel tubular stub columns after exposure to elevated reinforced with internal latticed steel angles under bending[J], Structures 48
temperature[J], Eng. Struct. 204 (2020), 110048. (2023) 1436–1454.
[6] W.D. Wang, X. Wei, C. Hou, Y.L. Shi, Experimental investigation and FE modelling [37] G. Li, Y. Yang, Z. Yang, et al., Mechanical behavior of high-strength concrete filled
of the flexural performance of square and rectangular SRCFST members[J], high-strength steel tubular stub columns stiffened with encased I-shaped CFRP
Structures 27 (2020) 2411–2425. profile under axial compression[J], Compos. Struct. 275 (2021), 114504.
[7] Y. Du, Y. Zhang, Z. Chen, et al., Axial compressive performance of CFRP confined [38] G. Li, X. Sun, Z. Yang, et al., Structural performance of concrete-filled square steel
rectangular CFST columns using high-strength materials with moderate tubular columns encased with I-shaped CFRP under eccentric compression[J], Eng.
slenderness[J], Constr. Build. Mater. 299 (2021), 123912. Struct. 248 (2021), 113254.
[8] Y. Du, M. Fu, Z. Chen, et al., Axial compressive behavior of CFRP confined [39] G. Li, X. Sun, Z. Yang, et al., Buckling behavior of slender square concrete filled
rectangular CFT columns using high-strength materials: numerical analysis and steel tubular columns strengthened with CFRP profile under combined
carrying capacity model[J], Structures 36 (2022) 997–1020. compression and bending[J], J. Build. Eng. 53 (2022), 104563.
[9] J. Cai, Z.Q. He, Axial load behavior of square CFT stub column with binding bars [40] GB/T 50081–-2019, Standard for Test Methods of Concrete Physical and
[J], J. Constr. Steel Res. 62 (5) (2006) 472–483. Mechanical Properties, China Architecture & Building Press, Beijing, China, 2019
[10] C.S. Huang, Y.K. Yeh, G.Y. Liu, et al., Axial load behavior of stiffened concrete- (in Chinese).
filled steel columns[J], J. Struct. Eng. 128 (9) (2002) 1222–1230. [41] GB/T 228.1–-2021, Metallic materials-Tensile testing-Part 1: Method of test at
[11] Z. Tao, L.H. Han, D.Y. Wang, Experimental behaviour of concrete-filled stiffened room temperature, Standards Press of China, Beijing, China, 2021 (in Chinese).
thin-walled steel tubular columns[J], Thin-Walled Struct. 45 (5) (2007) 517–527. [42] P. Feng, H.L. Qiang, L.P. Ye, Discussion and definition on yield points of materials,
[12] F. Yuan, L. Cao, H. Li, Axial compressive behaviour of high-strength steel spiral- members and structures[J], Eng. Mech. 34 (3) (2017) 36–46 (in Chinese).
confined square concrete-filled steel tubular columns[J], J. Constr. Steel Res. 192 [43] Abaqus 6.14, Abaqus Analysis User's Guide. USA, 2014.
(2022), 107245. [44] Y.J. Xiao, Z.S. Chen, J.T. Zhou, Y.L. Leng, R.C. Xia, Concrete plastic-damage factor
[13] F. Yuan, C. Hu, H. Li, et al., Effect of high-strength spirals on eccentric compression for finite element analysis: concept, simulation, and experiment [J], Adv. Mech.
behavior of square CFST columns[J], J. Constr. Steel Res. 196 (2022), 107372. Eng. 9 (9) (2017) 1–10.
[14] T. Ekmekyapar, B.J.M. Al-Eliwi, Concrete filled double circular steel tube [45] H. Lu, L.H. Han, X.L. Zhao, Analytical behavior of circular concrete-filled thin-
(CFDCST) stub columns[J], Eng. Struct. 135 (2017) 68–80. walled steel tubes subjected to bending[J], Thin-Walled Struct. 47 (3) (2009)
[15] V. Albero, C. Ibañez, A. Piquer, D.H. Figueiridoa, Behaviour of slender concrete- 346–358.
filled dual steel tubular columns subjected to eccentric loads[J], J. Constr. Steel [46] J. Wang, X. Cheng, L. Yan, et al., Numerical study on I-section steel-reinforced
Res. 176 (2021), 106365. concrete-filled steel tubes (SRCFST) under bending[J], Eng. Struct. 225 (2020),
[16] A. Xiamuxi, A. Hasegawa, A study on axial compressive behaviors of reinforced 111276.
concrete filled tubular steel columns[J], J. Constr. Steel Res. 76 (2012) 144–154. [47] J. Wang, J. Chen, W.J. Qiu, et al., Analysis of CFRP-confined CFST columns
[17] J. Moon, D.E. Lehman, C.W. Roeder, H.E. Lee, Strength of circular concrete-filled reinforced with steel angles under axial compression[J], J. Constr. Steel Res. 199
tubes with and without internal reinforcement under combined loading[J], (2022), 107644.
J. Struct. Eng. 139 (12) (2013) 04013012. [48] Z.H. Chen, S.H. Dong, Y.S. Du, Experimental study and numerical analysis on
[18] M.R. Hamidian, M.Z. Jumaat, U.J. Alengaram, N.H.R. Sulong, P. Shafigh, Pitch seismic performance of FRP confined high-strength rectangular concrete-filled steel
spacing effect on the axial compressive behaviour of spirally reinforced concrete- tube columns[J], Thin-Walled Struct. 162 (2021), 107560.
filled steel tube (SRCFT)[J], Thin-Walled Struct. 100 (2016) 213–223. [49] H. Qi, L. Guo, J. Liu, et al., Axial load behavior and strength of tubed steel
[19] H.S. Hu, L. Xu, Z.X. Guo, et al., Behavior of eccentrically loaded square spiral- reinforced-concrete (SRC) stub columns[J], Thin-Walled Struct. 49 (9) (2011)
confined high-strength concrete-filled steel tube columns[J], Eng. Struct. 216 1141–1150.
(2020), 110743. [50] J.B. Mander, M.J.N. Priestley, R. Park, Theoretical stress-strain model for confined
[20] Z. Chen, J. Zhou, C. Jing, et al., Mechanical behavior of spiral stirrup reinforced concrete[J], J. Struct. Eng. 114 (8) (1988) 1804–1826.
concrete filled square steel tubular columns under compression[J], Eng. Struct. 226 [51] H.Y. Tang, L.J. Hou, Z.J. Yuan, Y.G. Jia, Y.Z. Wang, Eccentric compressive
(2021), 111377. behavior of square concrete-filled stainless steel tube (CFSST) stub columns[J],
[21] F. Xu, J. Chen, W.L. Jin, Experimental investigation of thin-walled concrete-filled Structures 55 (2023) 1920–1935.
steel tube columns with reinforced lattice angle [J], Thin-Walled Struct. 84 (2014) [52] Q.Y. Qiao, R.L. Ding, W.L. Cao, Experimental investigation on axial local
59–67. compressive behavior of square concrete-filled double tubular stub columns[J],
[22] F. Xu, J. Chen, T.M. Chan, Numerical investigation on compressive performance of Structures 55 (2023) 2157–2174.
CFST columns with encased built-up lattice-angles[J], J. Constr. Steel Res. 137 [53] Z. Tao, L.H. Han, X.L. Zhao, Behaviour of concrete-filled double skin (CHS inner
(2017) 242–253. and CHS outer) steel tubular stub columns and beam-columns[J], J. Constr. Steel
[23] X. Chang, Y.Y. Wei, Y.C. Yun, Analysis of steel-reinforced concrete-filled-steel Res. 60 (8) (2004) 1129–1158.
tubular (SRCFST) columns under cyclic loading[J], Constr. Build. Mater. 28 (1)
(2012) 88–95.

18
J. Wang et al. Journal of Constructional Steel Research 213 (2024) 108414

[54] L.H. Han, Z. Tao, H. Huang, X.L. Zhao, Concrete-filled double skin (SHS outer and [61] L.H. Han, C. Hou, Y.X. Hua, Concrete-filled steel tubes subjected to axial
CHS inner) steel tubular beam-columns[J], Thin-Walled Struct. 42 (9) (2004) compression: life-cycle based performance[J], J. Constr. Steel Res. 170 (2020),
1329–1355. 106063.
[55] Y.L. Wang, G.H. Cai, A.S. Larbi, D. Waldmann, K.D. Tsavdaridis, J.H. Ran, [62] W.D. Wang, X.J. Li, J.H. Fan, Design method research and reliability analysis of
Monotonic axial compressive behaviour and confinement mechanism of square concrete-filled steel tubular columns with inner steel under axial compression[J],
CFRP-steel tube confined concrete[J], Eng. Struct. 217 (2020), 110802. J. Architect. Civ. Eng. 37 (6) (2020) 11–20 (in Chinese).
[56] J. Liu, X. Zhou, Behavior and strength of tubed RC stub columns under axial [63] D.Y. Ma, Z.C. Yang, Mechanism and reliability evaluation for concrete-encased
compression[J], J. Constr. Steel Res. 66 (1) (2010) 28–36. CFST structures under axial compression [J], China Civ. Eng. J. (2023) 1–17 (in
[57] GB50068–-2018, Unified Standard for Reliability Design of Building Structures, Chinese), 10.15951/j.tmgcxb.22111143.
China Architecture & Building Press, Beijing, China, 2018 (in Chinese). [64] F. Ding, S. Xiong, H. Zhang, et al., Reliability analysis of axial bearing capacity of
[58] H.Y. Zhao, L.H. Han, C. Hou, et al., Performance of recycled aggregate concrete- concrete filled steel tubular stub columns with different cross sections[J],
filled high-strength steel tubes under axial compression, tension and torsion[J], Structures 33 (2021) 4193–4202.
Thin-Walled Struct. 184 (2023), 110478. [65] GB50017–-2017, Standard for Design of Steel Structures, China Architecture &
[59] W. Chen, J. Xu, Z. Li, et al., Load-carrying capacity of circular recycled aggregate Building Press, Beijing, China, 2017 (in Chinese).
concrete-filled steel tubular stub columns under axial compression: reliability [66] Zhu Xi, Shi Gang, Statistics and analysis of material properties of domestic steel
analysis and design factor calibration[J], J. Build. Eng. 66 (2023), 105935. plates for building structures[J], Build. Struct. 45 (21) (2015) 16–20 (in Chinese).
[60] Y.F. An, L.H. Han, X.L. Zhao, Behaviour and design calculations on very slender
thin-walled CFST columns[J], Thin-Walled Struct. 53 (2012) 161–175.

19