Numerical Study on the Axial Compression Behavior of Composite Columns with High-Strength Concrete-Filled Steel Tube and Honeycombed Steel Web Subjected to Freeze–Thaw Cycles

Abstract

To investigate the axial compression behavior of composite columns with high-strength concrete-filled steel tube flanges and honeycombed steel web (STHHC) under load during freeze–thaw cycles, 48 full-scale composite column specimens were designed with different parameters: the restraint effect coefficient (ξ), concrete strength (f_cu), number of freeze–thaw cycles (n_d), slenderness ratio (λ), space–height ratio (s/h_w), and hole–height ratio (d/h_w). The finite element models of STHHC composite columns were simulated using ABAQUS finite element software (Version: 2021). The modeling method’s rationality was verified by comparing simulation results with experimental outcomes. Based on the finite element model, a parametric analysis of the composite columns under freeze–thaw cycles was conducted, analyzing their failure modes and load-bearing processes. The results indicate that the bearing capacity of the STHHC increased with increases in ξ and f_cu, and decreased with a rise in λ. In contrast, the influence of s/h_w and d/h_w on the ultimate bearing capacity of the composite columns was relatively minor. An equation for calculating the axial bearing capacity of the STHHC composite columns under freeze–thaw cycles was derived using statistical regression methods and considering the impact of different parameters on the axial compressive performance of the composite columns, laying the foundation for the promotion and application of this type of composite column in practical engineering projects.

1. Introduction

To meet the needs of modern society for building structures in the aspects of long span, high height, wind resistance, and earthquake resistance [1,2,3,4], concrete-filled steel tube structures came into being [5]. A concrete-filled steel tube structure has many advantages, such as lightweight, large span, strong bearing capacity, short construction period, good seismic performance, etc.; it is widely used in long-span, super-high-rise, and heavy-duty structural systems [6]. According to the demand for modern construction and the advantages of concrete-filled steel tube structures, a new type of composite column (the rectangular steel tube high-strength concrete flange–honeycomb steel web composite column) is proposed [7]. The composite column is mainly composed of two rectangular steel tube high-strength concrete columns connected by a honeycomb steel web: the steel tube can structure the core concrete in a three-way compression state, thereby improving the bearing capacity and stability of the column, and the concrete can reduce or delay the local buckling of the steel tube, extending the service life of the column. The use of honeycomb steel web can not only connect the two flanges [8,9,10] but also reduce the structural weight.

As we all know, building structures in cold regions are significantly affected by freeze–thaw action, and traditional reinforced concrete structures are prone to spalling under freeze–thaw conditions [11], which seriously affects the structures’ performance and bearing capacity [12,13]. Given the lack of research on the mechanical properties of rectangular high-strength concrete-filled steel tube flange–honeycomb steel web H-section composite columns after freeze–thaw cycles and the failure mechanism of such composite columns after freeze–thaw cycles, it is of great theoretical value to study the mechanical properties of this new type of composite columns after freeze–thaw cycles. It can provide technical support for practical engineering applications.

Many scholars have conducted much experimental research and many FEM simulation analyses on the axial compression properties of concrete-filled steel tubes. A one-factor numerical parameter study was conducted on 144 finite element models [14]. The effects of the volume ratio of basalt fiber, the reinforcement ratio of basalt FRP bars, and the stirrup spacing on the bending moment redistribution of basalt FRC continuous beams reinforced with BFRP bars have been discussed in depth. The results show that the parameter that has the greatest influence on the bending moment redistribution is the longitudinal reinforcement configuration. Sabih [15] used the finite element method to evaluate the axial load behavior of concrete-filled steel tubular slender columns strengthened with CFRP. The results showed that single-layer carbon fiber cloth with a thickness of 1.2 mm increases the axial bearing capacity of the composite column by 8.5%. The total thickness of the carbon fiber cloth doubled to 2.4 mm, and the resistance increased to 23.5%, while three carbon fiber cloths with a total thickness of 3.6 mm and four carbon fiber cloths with a total thickness of 4.8 mm produce 35.1% and 44.5% of the axial load resistance, respectively. Xie [16] studied the influence of the length to length-to-width ratio of steel tubes on the mechanical properties of short columns. The results showed that the outer tube’s section size and the inner tube’s local buckling greatly influence the specimen’s axial compressive capacity. Ji [17,18] performed nonlinear buckling analysis on 30 STHCC specimens with different hoop coefficients, section sizes, and other parameters. Based on the simulation results, a calculation equation for the stable bearing capacity of STHCC composite long columns with initial defects was established, and design suggestions for such axial compression long columns were put forward based on the analysis results. Elzeadani [19] studied the axial compression behavior of steel tubes filled with rubber alkali-activated concrete experimentally and numerically. The numerical results were validated against the experimental results, and parameter studies involving 315 finite element models were performed to cover a wide range of concrete and steel properties and different steel tube sizes. Wei [20] studied the bias mechanical properties of concrete-filled steel tube members with round ends. Based on the experimental research and numerical simulation results, equations for calculating the flexural bearing capacity of these components have been presented.

Eleswhere, Ali [21] investigated the structural behavior of double-skin columns and utilized double steel tubes and concrete to enhance the load-carrying capacity and ductility beyond conventional double-skin hollow tubular (DSHT) columns, employing a combination of finite element model and machine learning techniques. Shi [22] studied the eccentric compression properties of concrete-filled steel tubes (CFSTs) with the hollow sandwich. The numerical simulation results showed that the eccentricity and slenderness ratio have great influence on the bearing capacity and stiffness of the biased members. On the basis of the analysis results, the calculation method for the flexural bearing capacity of such components was proposed. On the basis of an experimental study on the eccentric compression column, Ma [23] analyzed the influence of various parameters on the mechanical properties of the eccentric compression column and proposed the practical calculation equation for the bearing capacity of the eccentric compression column. Li [24] analyzed the deformation properties of concrete-filled steel tube columns under long-term load and conducted long-term load-holding tests on 16 concrete-filled steel tube short columns and two concrete-filled steel tube short columns, and the results showed that the long-term deformation increases with the increase in the stress ratio of the core concrete, and decreases with a rise in the total steel content and the ratio of the inner steel content to the outer steel tube. Chen [25] studied the axial compression properties of short columns filled with ultra-high-performance concrete (UHPC) in high-strength steel tubes. They designed 22 short columns of UHPC in high-strength steel tubes and carried out axial compression tests. The results showed that the axial compression properties of short columns of UHPC were affected by the ratio of diameter to thickness and mixing. The influence of soil strength is excellent. The bearing capacity of UHPC specimens increases with the increase in steel tube wall thickness. Still, the law for the same steel tube diameter and the increase in different steel tube diameters is different. Ji [26], to explore the influence of various control parameters on the axial compressive ultimate bearing capacity of such short column specimens, designed and conducted axial compressive test studies on 16 STHCC specimens and established a calculation equation of the axial compressive bearing capacity of the short column.

And many studies have investigated the axial compression properties of the concrete-filled steel tube members undergoing freeze–thaw cycles. Gao [27] studied the effects of various concrete grades and freeze–thaw cycles on concrete properties. The results showed that with the concrete strength rising, the axial strength deterioration of CFST specimens increases under different freeze–thaw cycles. The freeze–thaw cycles have little influence on concrete-filled steel tube specimens’ ultimate displacement and ductility. Li [28], taking the number of freeze–thaw cycles, steel tube wall thickness, and concrete strength as variables, analyzed the changing laws of bond strength, load slip, and steel tube strain of circular steel tube concrete columns under the action of freeze–thaw damage. The results showed that the interfacial bond performance of the cylindrical steel tube concrete column affected by the freeze–thaw cycles decreases, and the bond strength is oppositely proportional to the time of freeze–thaw cycles. Accordingly, an equation for calculating the bond strength of concrete columns with round steel tubes was presented, which considers the time of freeze–thaw cycles and the coefficient of the hoop. Zhou [29] conducted axial compression tests on specimens of circular steel tube concrete short columns subjected to freeze–thaw cycles, taking concrete strength and the number of freeze–thaw cycles as the main parameters to analyze the specimens’ failure forms and force–displacement relationships. The results showed that the number of freeze–thaw cycles has no noticeable effect on the failure form of the specimen, and the ultimate bearing capacity and elastic section length of the specimen decrease gradually with the increase in the number of freeze–thaw cycles.

Although research on honeycomb columns and STHHC columns is comparatively mature, research on the mechanical behavior of STHHC after freeze–thaw cycles has not been reported, the contribution of concrete and steel tubes to the axial compression capacity of STHHC is not clear, and there is a lack of in-depth understanding of the failure mechanism and axial compression behavior of STHHC under freeze–thaw cycles. Therefore, it is of great theoretical significance and practical engineering value to study the mechanical properties and failure mechanism of STHHC composite columns after freeze–thaw cycles under axial pressure. In this study, ABAQUS finite element software was used to numerically analyze the axial compression performance of 48 full-scale STHHC specimens and obtain the load–displacement curve, ductility coefficient, failure pattern, and the stress–strain distribution law of the section element of the combined column under axial compression. Finally, considering the influence of different parameters on the mechanical properties of composite columns, 1stOpt software (Version 1stOpt 15) was used for regression analysis, and the equation for calculating the axial compressive capacity of STHHC composite columns under freeze–thaw cycles was obtained, which lays a foundation for the application of such composite columns in practical engineering.

2. Analysis Process of the Paper

The analysis process of this paper mainly includes specimen design, the establishment of the finite element model, the validation of the model’s rationality, expanded parametric analysis, force mechanism, the proposed equation for the ultimate bearing capacity of the novel composite columns, and conclusions. The detailed procedure of this paper is shown in Figure 1.

3. Specimen Design

To investigate the axial compression performance of full-scale STHHC composite columns after experiencing freeze–thaw cycles, finite element analysis was conducted using parameters such as the confinement effect coefficient (ξ), concrete strength (f_cu), number of freeze–thaw cycles (n_d), slenderness ratio (λ), aspect ratio (s/h_w), and aspect ratio of holes (d/h_w) [30,31]. A total of 48 full-scale specimens were designed, with specific parameters and detailed dimensions of the specimens provided in

Table 1. Dimensions of full-scale axial compression specimens.

Specimens	b × h1 × hw × t2 × t1/mm	L/mm	λ	nd	ξ	fyk/MPa	fcu/MPa	d/hw	s/hw
STHHC-1	500 × 320 × 400 × 10 × 04	1600	11.08	0	0.23	345	70	0.6	0.25
STHHC-2	500 × 320 × 400 × 10 × 04	1600	11.08	100	0.23	345	70	0.6	0.25
STHHC-3	500 × 320 × 400 × 10 × 04	1600	11.08	200	0.23	345	70	0.6	0.25
STHHC-4	500 × 320 × 400 × 10 × 04	1600	11.08	300	0.23	345	70	0.6	0.25
STHHC-5	500 × 320 × 400 × 10 × 06	1600	11.08	0	0.35	345	70	0.6	0.25
STHHC-6	500 × 320 × 400 × 10 × 06	1600	11.08	100	0.35	345	70	0.6	0.25
STHHC-7	500 × 320 × 400 × 10 × 06	1600	11.08	200	0.35	345	70	0.6	0.25
STHHC-8	500 × 320 × 400 × 10 × 06	1600	11.08	300	0.35	345	70	0.6	0.25
STHHC-9	500 × 320 × 400 × 10 × 08	1600	11.08	0	0.48	345	70	0.6	0.25
STHHC-10	500 × 320 × 400 × 10 × 08	1600	11.08	100	0.48	345	70	0.6	0.25
STHHC-11	500 × 320 × 400 × 10 × 08	1600	11.08	200	0.48	345	70	0.6	0.25
STHHC-12	500 × 320 × 400 × 10 × 08	1600	11.08	300	0.48	345	70	0.6	0.25
STHHC-13	500 × 320 × 400 × 10 × 10	1600	11.08	0	0.60	345	70	0.6	0.25
STHHC-14	500 × 320 × 400 × 10 × 10	1600	11.08	100	0.60	345	70	0.6	0.25
STHHC-15	500 × 320 × 400 × 10 × 10	1600	11.08	200	0.60	345	70	0.6	0.25
STHHC-16	500 × 320 × 400 × 10 × 10	1600	11.08	300	0.60	345	70	0.6	0.25
STHHC-17	500 × 320 × 400 × 10 × 04	1600	11.08	0	0.27	345	60	0.6	0.25
STHHC-18	500 × 320 × 400 × 10 × 04	1600	11.08	100	0.27	345	60	0.6	0.25
STHHC-19	500 × 320 × 400 × 10 × 04	1600	11.08	200	0.27	345	60	0.6	0.25
STHHC-20	500 × 320 × 400 × 10 × 04	1600	11.08	300	0.27	345	60	0.6	0.25
STHHC-21	500 × 320 × 400 × 10 × 04	1600	11.08	0	0.20	345	80	0.6	0.25
STHHC-22	500 × 320 × 400 × 10 × 04	1600	11.08	100	0.20	345	80	0.6	0.25
STHHC-23	500 × 320 × 400 × 10 × 04	1600	11.08	200	0.20	345	80	0.6	0.25
STHHC-24	500 × 320 × 400 × 10 × 04	1600	11.08	300	0.20	345	80	0.6	0.25
STHHC-25	500 × 320 × 400 × 10 × 04	1700	11.78	0	0.23	345	70	0.6	0.25
STHHC-26	500 × 320 × 400 × 10 × 04	1700	11.78	100	0.23	345	70	0.6	0.25
STHHC-27	500 × 320 × 400 × 10 × 04	1700	11.78	200	0.23	345	70	0.6	0.25
STHHC-28	500 × 320 × 400 × 10 × 04	1700	11.78	300	0.23	345	70	0.6	0.25
STHHC-29	500 × 320 × 400 × 10 × 04	1800	12.47	0	0.23	345	70	0.6	0.25
STHHC-30	500 × 320 × 400 × 10 × 04	1800	12.47	100	0.23	345	70	0.6	0.25
STHHC-31	500 × 320 × 400 × 10 × 04	1800	12.47	200	0.23	345	70	0.6	0.25
STHHC-32	500 × 320 × 400 × 10 × 04	1800	12.47	300	0.23	345	70	0.6	0.25
STHHC-33	500 × 320 × 400 × 10 × 04	1600	11.08	0	0.23	345	70	0.5	0.25
STHHC-34	500 × 320 × 400 × 10 × 04	1600	11.08	100	0.23	345	70	0.5	0.25
STHHC-35	500 × 320 × 400 × 10 × 04	1600	11.08	200	0.23	345	70	0.5	0.25
STHHC-36	500 × 320 × 400 × 10 × 04	1600	11.08	300	0.23	345	70	0.5	0.25
STHHC-37	500 × 320 × 400 × 10 × 04	1600	11.08	0	0.23	345	70	0.7	0.25
STHHC-38	500 × 320 × 400 × 10 × 04	1600	11.08	100	0.23	345	70	0.7	0.25
STHHC-39	500 × 320 × 400 × 10 × 04	1600	11.08	200	0.23	345	70	0.7	0.25
STHHC-40	500 × 320 × 400 × 10 × 04	1600	11.08	300	0.23	345	70	0.7	0.25
STHHC-41	500 × 320 × 400 × 10 × 04	1600	11.08	0	0.23	345	70	0.6	0.15
STHHC-42	500 × 320 × 400 × 10 × 04	1600	11.08	100	0.23	345	70	0.6	0.15
STHHC-43	500 × 320 × 400 × 10 × 04	1600	11.08	200	0.23	345	70	0.6	0.15
STHHC-44	500 × 320 × 400 × 10 × 04	1600	11.08	300	0.23	345	70	0.6	0.15
STHHC-45	500 × 320 × 400 × 10 × 04	1600	11.08	0	0.23	345	70	0.6	0.35
STHHC-46	500 × 320 × 400 × 10 × 04	1600	11.08	100	0.23	345	70	0.6	0.35
STHHC-47	500 × 320 × 400 × 10 × 04	1600	11.08	200	0.23	345	70	0.6	0.35
STHHC-48	500 × 320 × 400 × 10 × 04	1600	11.08	300	0.23	345	70	0.6	0.35

and Figure 2. Due to the harsh climate of Northeast China, where winters are cold and prolonged with a high number of daily freeze–thaw cycles, causing significant frost damage to building structures, the number of freeze–thaw cycles for the specimens in this study was determined to be 0, 100, 200, and 300 cycles [32,33]. This decision was made to more closely simulate the impact of freeze–thaw cycles on building structures under realistic conditions, while also referring to the design methods of freeze–thaw cycle experiments from universities both domestically and internationally. If ξ is the steel tube confinement effect coefficient, then the specific equation is $\xi =A_{\mathrm{s}}{f}_{\mathrm{yk}}/\left(A_c{f}_{\mathrm{cu}}\right)$; f_yk is the yield strength of the steel tube; A_s and A_c are the cross-sectional area of the flange steel tube and the cross-sectional area of the concrete; and λ is the slenderness ratio of the specimen, $\lambda =2\sqrt{3}L/b$.

4. Finite Element Model

4.1. Constitutive Model of Materials

4.1.1. Steel

The stress–strain curve obtained using the standard tensile test of steel is the constitutive model of steel. Through the different stress–strain relationships, the stress–strain curve of steel is divided into five stages: elastic stage, elastic–plastic stage, plastic stage, strengthening stage, and the secondary plastic flow stage. The influence of the plastic hardening effect of steel was considered; this study employed an ideal elastoplastic constitutive model as the constitutive model for steel materials [34], as shown in Figure 3.

In the figure, E_s is the elastic modulus of steel, f_py is the yield strength of steel, and ε_p is the strain at the yield strength.

4.1.2. Concrete

The constitutive model of concrete can reflect the stress characteristics of concrete in the whole process of compression and tension. It is an indispensable part of the stress analysis of concrete members and the basis of nonlinear analysis. In the process of ABAQUS modeling in this study, the influence of the plastic damage of the concrete on the concrete performance is considered. This study utilized the Han constitutive model [35], which considers the confinement effect of concrete, to simulate the stress–strain behavior of concrete in specimens before undergoing freeze–thaw cycles, as depicted in Figure 4.

4.1.3. Concrete under Freeze–Thaw Cycles

This study adopted the confined concrete constitutive model proposed by Han Lin-Hai and adjusted by Cao Kai [33], considering the effect of freeze–thaw cycles on the peak strain (ε₀) and peak stress (σ₀) of the core concrete. A uniaxial σ-ε relationship model for core concrete under the action of freeze–thaw cycles is proposed as follows:

(1) Compression analysis:

$$y=\left\{\begin{array}{c}2x-x^2\hspace{1em}\hspace{1em}\hspace{1em} \hspace{1em}\hspace{1em}(x\le 1)\\ {\displaystyle \frac{x}{\beta {(x-1)}^{\eta }+x}}\hspace{1em}\hspace{1em}\hspace{1em}(x>1)\end{array}\right.$$

(1)

of the form $x=\epsilon /{\epsilon }_0$; $y=\delta /{\delta }_0$

$${\delta }_0=f_{\mathrm{c}}^{\prime }(1-0.065n)$$

(2)

$${\epsilon }_0=(1300+12.5f_c^{\prime }+800{\xi }^{0.2})\times {10}^{-6}$$

(3)

$$\eta =1.6+{\displaystyle \frac{1.5}{x}}$$

(4)

$$\beta ={(f_{\mathrm{c}}^{\prime })}^{0.1}/[1.2\sqrt{1+\xi }]$$

(5)

where f_c′ represents the compressive strength of the concrete prism, and n = n_d/100, where n_d is the number of freeze–thaw cycles the specimen has experienced.

(2) Tensile Analysis:

In ABAQUS, for the simulation of concrete in tension, the stress–crack energy relationship (GFI) can be used to describe the tensile softening behavior of concrete. The cracking stress is taken as one-tenth of the concrete’s ultimate compressive stress, and the fracture energy is calculated using Equation (6).

$$G_f=a\cdot {\left({\displaystyle \frac{f_c^{\prime }}{10}}\right)}^{0.7}\times {10}^{-3}({\mathrm{N}/\mathrm{mm}}^2)$$

(6)

In the equation, a = 1.25 d_max + 10, where d_max is the maximum size of the coarse aggregate.

When calculating, the elastic modulus of the concrete is determined according to Equation (7).

Poisson’s ratio for concrete in the elastic stage is taken as 0.2.

$$E_c=4700\sqrt{f_c^{\prime }}(\mathrm{MPa})$$

(7)

The constitutive curve of the model calculated using the above formula is shown in Figure 5.

4.2. Establishment of Finite Element Model

4.2.1. Element Type and Contact Method

In the modeling of the composite short column with rectangular concrete-filled steel tube flanges and honeycomb steel webs, the steel tube, steel webs, and concrete were all modeled using C3D8R elements. General contact was used to simulate the contact between the steel tube and concrete, with normal contact set as “hard” and tangential contact set as relative sliding, with a friction coefficient of 0.3. This type of contact can simulate the actual stress conditions between the two materials to the greatest extent.

4.2.2. Boundary Conditions and Loading Method

To facilitate the application of loads and the setting of boundary conditions, reference points RP-1 and RP-2 were set at the center of the top and bottom surfaces of the composite column, respectively, and the upper and lower surfaces were coupled with the reference points to ensure that the specimen surface underwent coordinated deformation during the application of axial load, avoiding biased pressure. Fixed boundary conditions were applied to the upper surface of the composite column, and fully constrained boundary conditions were applied to the lower surface, as shown in Figure 6. The loading method was displacement loading, with a displacement load P applied at the top reference point of the column, the load size being 1/10 of the column height.

4.2.3. Mesh Division

A hexahedral mesh was used for dividing the model in finite element modeling. In the finite element modeling process, the smaller the mesh size, the higher the computational accuracy of the finite element model. However, the rise in computational volume slows down the calculation speed [37]. To ensure computational accuracy while improving the calculation speed, this study took two examples from [17], set up meshes of different sizes for the finite element model calculations, and compared the computational results. It was found that a grid size of 20 mm yields results that are closer to the experimental values, as shown in Figure 7.

5. Validation of Finite Element Model

5.1. Validation of Finite Element Model for Concrete-Filled Steel Tube Flange–Honeycomb Steel Web Composite Short Columns

The research group led by Ji [18] conducted axial compression tests on 14 rectangular concrete-filled steel tube flange–honeycomb steel web composite short columns. The specific test parameters of the specimens are shown in Table 1.

In using the ABAQUS finite element software, the load–displacement curves of the 14 specimens under axial compression were obtained. Through comparing the experimental and finite element results, it can be seen that the curves obtained from the finite element simulation match well with those obtained from the experiments, with the comparison results shown in Figure 8 and

Table 2. Parameters of the 14 STHCC composite short column specimens.

Specimens	h1 × b × hw × t2 × t1/mm	λ	l/mm	ξ	fyk/MPa	fcu/MPa	${\mathit{N}}_{\mathbf{U}}^{\mathbf{T}}$/kN	${\mathit{N}}_{\mathbf{U}}^{\mathbf{S}}$/kN	$\frac{\left|{\mathit{N}}_{\mathbf{U}}^{\mathbf{T}}-{\mathit{N}}_{\mathbf{U}}^{\mathbf{S}}\right|}{{\mathit{N}}_{\mathbf{U}}^{\mathbf{T}}}$
STHHC-1	50 × 100 × 100 × 6 × 1.7	12.82	370	0.60	269	49.50	740.0	763.4	3.17%
STHHC-2	50 × 100 × 100 × 6 × 2.3	12.82	370	0.88	282	49.50	871.7	864.5	0.83%
STHHC-3	50 × 100 × 100 × 6 × 3.8	12.82	370	1.60	286	49.50	1175.6	1150.8	2.11%
STHHC-4	50 × 100 × 100 × 6 × 2.3	12.82	370	0.82	282	53.17	1070.4	1057.4	1.21%
STHHC-5	50 × 100 × 100 × 6 × 2.3	12.82	370	0.66	282	55.69	1172.7	1152.3	1.73%
STHHC-6	50 × 100 × 100 × 6 × 2.3	12.82	370	0.78	282	65.60	1015.9	1020.1	0.41%
STHHC-7	50 × 100 × 100 × 8 × 1.7	12.82	370	0.60	269	49.50	752.6	789.6	4.91%
STHHC-8	50 × 100 × 100 × 11 × 2.7	12.82	370	0.60	269	49.50	796.7	834.1	4.69%
STHHC-9	50 × 100 × 100 × 6 × 2.3	9.35	270	0.88	282	49.50	947.0	936.9	1.06%
STHHC-10	50 × 100 × 100 × 6 × 2.3	16.28	470	0.88	282	49.50	841.1	829.0	1.36%
STHHC-11	50 × 100 × 100 × 6 × 1.7	9.35	270	0.60	269	49.50	772.9	787.8	1.93%
STHHC-12	50 × 100 × 100 × 6 × 1.7	16.28	470	0.60	269	49.50	741.1	762.8	2.92%
STHHC-13	50 × 100 × 100 × 6 × 3.8	9.35	270	1.60	286	49.50	1276.0	1299.1	1.81%
STHHC-14	50 × 100 × 100 × 6 × 3.8	16.28	470	1.60	286	49.50	1069.0	1080.5	1.07%

Note: The physical meanings of the other variables can be found in reference [18].

. Figure 9 shows that the maximum error (Error_max) between the finite element results and the experimental data is 4.91%. This indicates that the finite element model established in this study can accurately simulate the load-bearing process of this new type of composite column.

5.2. Validation of Finite Element Model for High-Strength Concrete-Filled Square Steel Tube under Freeze–Thaw Cycles

Fifteen square concrete-filled steel tube short columns designed by Yang [38] and Cao [33] for axial compression performance tests were selected to verify the rationality of the model above. The specific data for the 15 specimens are presented in

Table 3. Specific parameters of the specimens.

Specimens	D/mm	L/mm	t/mm	fcu/MPa	α	fy/MPa	${\mathit{N}}_{\mathbf{U}}^{\mathbf{T}}$/kN	${\mathit{N}}_{\mathbf{U}}^{\mathbf{S}}$/kN	$\frac{\left|{\mathit{N}}_{\mathbf{U}}^{\mathbf{T}}-{\mathit{N}}_{\mathbf{U}}^{\mathbf{S}}\right|}{{\mathit{N}}_{\mathbf{U}}^{\mathbf{T}}}$
SC2-0	100	300	2.06	63.7	0.113	338.4	808.38	834.96	3.34%
SC2-50	100	300	2.06	63.7	0.113	338.4	803.00	785.81	2.13%
SC2-100	100	300	2.06	63.7	0.113	338.4	783.77	802.04	2.32%
SC2-200	100	300	2.06	63.7	0.113	338.4	808.99	832.09	2.83%
SC2-300	100	300	2.06	63.7	0.113	338.4	749.10	775.16	3.41%
S2-0	100	300	2.00	83.9	0.085	184.7	861.67	814.23	5.83%
S2-50	100	300	2.00	83.9	0.085	184.7	804.47	793.24	1.42%
S2-100	100	300	2.00	83.9	0.085	184.7	783.92	778.60	0.68%
S2-200	100	300	2.00	83.9	0.085	184.7	777.55	754.89	3.00%
S2-300	100	300	2.00	83.9	0.085	184.7	663.06	715.48	7.91%
S3-0	100	300	3.00	83.9	0.132	176.5	891.58	856.82	4.06%
S3-50	100	300	3.00	83.9	0.132	176.5	873.91	852.65	2.49%
S3-100	100	300	3.00	83.9	0.132	176.5	775.53	800.38	3.20%
S3-200	100	300	3.00	83.9	0.132	176.5	754.48	815.90	8.14%
S3-300	100	300	3.00	83.9	0.132	176.5	722.31	772.20	6.91%

Note: The physical meanings of the other variables can be found in reference [38].

.

The load–strain curves (N-ε) for 15 square steel tube high-strength concrete short columns under axial compression were obtained through finite element simulation, as shown in Figure 10. It can be observed that the load–displacement curves exhibit good consistency. The comparison results between the simulation results ($N_{\mathrm{U}}^{\mathrm{S}}$) and the experimental results ($N_{\mathrm{U}}^{\mathrm{T}}$) are shown in Table 3 and Figure 9, with the maximum error between $N_{\mathrm{U}}^{\mathrm{S}}$ and $N_{\mathrm{U}}^{\mathrm{T}}$ being 8.14%. This validates the rationality of the finite element method and demonstrates that the finite element model can accurately predict the ultimate load capacity of the specimens, indicating that the high-strength concrete constitutive model under these freeze–thaw cycles can reflect the actual conditions of the experiments quite well.

6. Parametric Analysis

In the numerical simulation, six design parameters were considered to comprehensively understand the performance of this type of column under axial compression: confinement effect coefficient (ξ), concrete strength (f_cu), number of freeze–thaw cycles (n_d), slenderness ratio (λ), aspect ratio (s/h_w), and aspect ratio of holes (d/h_w). The effects of different parameters on the load–displacement (N-Δ) curves, ductility, and stiffness of the composite columns were obtained. This study used the energy equivalent method to determine the yield displacement of the specimens. It analyzed the ductility of the composite columns after freeze–thaw cycles through the displacement ductility coefficient (μ). The expression of μ is shown in Equation (8):

$$\mu ={\displaystyle \frac{\Delta u}{\Delta y}}$$

(8)

In this study, the energy equivalent method was used to determine the yield displacement of the member, as shown in Figure 11. The displacement corresponding to the point D is the yield displacement Δ_y. The ultimate displacement of the member usually refers to the displacement corresponding to a certain value of bearing capacity when the combined column enters the falling stage after reaching the peak load. There are some differences in the value of this value at home and abroad; in this study, we took the displacement corresponding to the point of falling to 85% of the peak load, and the displacement corresponding to point E is the ultimate displacement.

The column is symmetrical about the X axis and the Y axis, and it is subjected to force in the Z direction. A quarter of the column is selected and divided into five units to analyze its stress and strain. These five elements can well reflect the stress–strain curve of the composite column under axial load. In the cross-section at half the height of the composite column, concrete units at different positions were selected, and their stress–strain curves under axial load were extracted, as shown in Figure 12. The influence patterns of different parameters on the stress–strain curves of the concrete units were studied, and the mechanisms of this type of composite column under load were analyzed based on the results.

6.1. Confinement Effect Coefficient (ξ)

The N-Δ curves of composite columns with different ξ values after freeze–thaw cycles are shown in Figure 13, and the ultimate load (N_max) of composite columns with different ξ values under various numbers of freeze–thaw cycles (n_d) is depicted in Figure 14. It can be observed that when ξ ranges from 0.23 to 0.48, the rate of decrease in the specimens’ ultimate load capacity after the number of freeze–thaw cycles decreases with the increase in ξ, indicating that within this range, an increase in ξ helps to mitigate the damage caused by the freeze–thaw of concrete, which leads to a reduction in the load-bearing capacity. When ξ is between 0.48 and 0.6 and when the steel tube thickness exceeds 8 mm, the rate of decrease in the specimens’ load-bearing capacity after freeze–thaw cycles increases, adversely affecting the freeze–thaw resistance of the specimens.

The n-μ curves for composite columns with different ξ values are shown in Figure 15, from which it can be seen that as ξ increases, the ductility of the specimens under different n_d shows a trend of enhancement, and the effect is significant. As n_d increases, the overall μ of the specimens tends to decrease. When ξ is greater than 0.48, the specimens’ ductility decreases after the number of freeze–thaw cycles is reduced; therefore, in practical engineering, ξ should be greater than 0.48.

Figure 16 illustrates the stress–strain curves of units A, B, C, D, and E of composite columns with different ξ values after freeze–thaw cycles. From the curves, it can be seen that as ξ increases, the stress of each unit also increases, and the rate of decline in the stress–strain curves slows down accordingly. For specimens with an ξ of 0.48 and 0.60, the closer to the center of the specimen, the slower the rate of the stress–strain curve’s decline. After freeze–thaw cycles, the stress of each unit under different ξ values decreases in sync with the increase in n_d. As shown in the figure, when not subjected to freeze–thaw, after 100 and 200 freeze–thaw cycles, the peak stress of each unit under different ξ values does not differ significantly. However, after 300 freeze–thaw cycles, the peak stress of the specimens with an ξ of 0.48 (t = 8 mm) is significantly higher than those of other specimens, which is consistent with the previous load analysis results. It is evident that when ξ is 0.48 (t = 8 mm), the confining effect of the steel tube optimizes the cooperation between the steel tube and concrete. This not only improves the unit stress but also slows down the stress reduction in composite column units after freeze–thaw cycles, enabling the composite column to maintain a high overall load-bearing capacity even after 300 freeze–thaw cycles.

6.2. Cubic Strength of Concrete (f_cu)

To investigate the impact of freeze–thaw cycles on the axial compressive performance of composite columns with different f_cu grades, post-freeze–thaw axial compressive performance analyses were conducted on composite columns with f_cu values of C60, C70, and C80. The N-Δ curves for composite columns with different f_cu values after freeze–thaw cycles are shown in Figure 17a–c, and the variation curves of N_max for composite columns with different f_cu values after freeze–thaw cycles are depicted in Figure 18. As f_cu increases from 60 MPa to 80 MPa, the N_max of the composite columns gradually increases. After 300 freeze–thaw cycles, their ultimate load-bearing capacities decrease by 15.8%, 14.17%, and 13.96%, respectively, compared to their pre-freeze–thaw states. It is evident that the action of a freeze–thaw cycle causes a certain degree of damage to the core concrete inside the steel tube, and the extent of this impact varies between different n_d values, with higher n_d values leading to greater damage. Moreover, the higher the f_cu, the lesser the effect of the freeze–thaw cycles.

The n-μ curves for composite columns with different f_cu values are shown in Figure 19. When the f_cu of the composite column increases from 60 MPa to 70 MPa and 80 MPa, the μ values are 4.21, 4.27, and 4.30 respectively, showing an overall upward trend. After 300 freeze–thaw cycles, the μ values of the specimens decrease by 1.4%, 1.2%, and 0.1%, respectively. This indicates that as the f_cu increases, the overall ductility of the members does not change significantly. After 300 freeze–thaw cycles, the lower the f_cu, the greater the decrease in μ.

Figure 20 presents the stress–strain curves of Units A, B, C, D, and E of composite columns with different f_cu values after freeze–thaw cycles. From the curves, it can be observed that as f_cu increases, the peak stress of the units significantly rises. The more central the area near the specimen, the more pronounced the confinement effect, which gradually reduces the rate of the descending part of the curve, delaying the failure of the specimen. The trend in curve changes for specimens with different f_cu values after freeze–thaw cycles is essentially consistent. As n_d increases, the stress of units under different f_cu values gradually decreases. Even after 300 freeze–thaw cycles, the core region strength of C80 concrete remains around 60 MPa, enabling the specimen to maintain a high load-bearing capacity.

6.3. Slenderness Ratio (λ)

The N-Δ curves of composite columns with different λ values after freeze–thaw cycles are shown in Figure 21, while the variation curves of N_max for composite columns with different λ values after freeze–thaw cycles are presented in Figure 22. It can be observed from the figures that as λ increases, the N_max of the composite columns gradually decreases. After 300 freeze–thaw cycles, the N_max of the composite columns decreases by 14.17%, 15.04%, and 15.35% compared to their state of being unfrozen or unthawed. It is evident that N_max follows a pattern of gradual reduction with an increase in both λ and n_d, and the greater the λ, the more significant the decrease in N_max after the freeze–thaw cycles.

The n-μ curves for composite columns with different λ values are depicted in Figure 23. As the λ of the full-scale columns increases, the overall trend in their μ values decreases. After 300 freeze–thaw cycles, the decrease in μ for specimens with different λ values is 0.5%, 1.4%, and 1.2%, respectively. This indicates that as λ increases, the ductility of composite columns slightly improves. In contrast, the ductility generally tends to decrease with a rise in n_d, but the numerical change is small.

As represented in Figure 24, the stress–strain curves for Units A, B, C, D, and E of composite columns with different λ values after freeze–thaw cycles are illustrated. From the curves, it can be seen that the peak stress of specimens with different λ values generally shows a decreasing trend as n_d increases. Units A and B, located at the edges of the specimens, experience relatively small changes in peak stress values and a faster rate of stress decrease with an increase in n_d and λ. For the specimens in Unit C with λ = 11.08, the stress decreases more rapidly after freeze–thaw cycles compared to the other two groups. However, in Units D and E, the stress values for specimens with λ = 11.08 are significantly higher than the other two groups after freeze–thaw cycles, indicating a pronounced restraining effect of the steel tube on specimens with λ = 11.08.

6.4. Aspect Ratio (s/h_w)

The N-Δ curves for composite columns with different s/h_w values after freeze–thaw cycles are shown in Figure 25a–c, and the variation curves of N_max for composite columns with different s/h_w values after freeze–thaw cycles are presented in Figure 26. As observed from the figures, the N_max of the composite columns generally decreases with an increase in s/h_w. After 300 freeze–thaw cycles, the N_max of the composite columns decreases by 14.64%, 14.17%, and 13.77% compared to their pre-freeze–thaw states. It is evident that an increase in s/h_w can slow down the rate of N_max reduction caused by freeze–thaw cycles. Within the range of this study, the larger the s/h_w, the stronger the supporting effect of the steel web on the flanges, and the slower the rate of N_max reduction after freeze–thaw cycles.

The n-μ curves for composite columns with different s/h_w values are shown in Figure 27. As the s/h_w of the composite columns increases from 0.15 to 0.35, μ shows a trend of first decreasing and then increasing. With the increase in n_d, the μ of specimens with different s/h_w values follows different patterns of change, but generally exhibits a downward trend, with μ decreasing by 0.9%, 0.9%, and 0.7% after 300 freeze–thaw cycles, respectively. It can be seen that the variation in s/h_w has a certain impact on the ductile performance of specimens after freeze–thaw action, but the impact is relatively small.

Figure 28 illustrates the stress–strain curves for Units A, B, C, D, and E of composite columns with different s/h_w values after freeze–thaw cycles. From the curves, it can be observed that with the increase in n_d, the peak stress of specimens under different s/h_w values shows a downward trend, but the degree of decrease is slight. The stress–strain curves for Units A, B, D, and E are relatively unaffected by n_d and s/h_w, with no significant overall changes. Notably, for Unit C, when s/h_w = 0.25, the stress–strain curve for this unit is significantly influenced by n_d, showing a faster rate of decrease compared to the other two curves. Overall, the specimens with s/h_w = 0.35 exhibit superior unit stress levels compared to the other two groups of specimens, suggesting that an s/h_w setting of 0.35 is recommended.

6.5. Aspect Ratio of Holes (d/h_w)

The N-Δ curves for composite columns with different d/h_w values after freeze–thaw cycles are shown in Figure 29, and the variation curves of N_max for composite columns with different d/h_w values after freeze–thaw cycles are presented in Figure 30. It can be seen that as d/h_w increases, the N_max of the composite columns generally tends to decrease. After 300 freeze–thaw cycles, the N_max of the composite columns decreases by 14.22%, 14.17%, and 14.76% compared to their pre-freeze–thaw states. This indicates that the larger the d/h_w, the less support it provides to the flanges, resulting in a greater decrease in N_max. However, compared to other parameters, d/h_w has a relatively small impact on the N_max of specimens after freeze–thaw cycles.

The n-μ curves for composite columns with different d/h_w values are shown in Figure 31. As the d/h_w of the composite columns increases from 0.5 to 0.7, μ first decreases and then increases. With the increase in n_d, the μ of specimens with different d/h_w values generally declines gradually, with μ decreasing by 0.9%, 1.4%, and 1.0% after 300 freeze–thaw cycles, respectively.

Figure 32 illustrates the stress–strain curves for Units A, B, C, D, and E of composite columns with different d/h_w values after freeze–thaw cycles. From the figure, it can be observed that as d/h_w increases, the peak stress of the units initially increases and then decreases. An excessively large d/h_w weakens the supporting effect of the steel web on the flanges, leading to a reduction in the confining effect of the steel tube. With the increase in n_d, the peak stress of units for specimens with different d/h_w values shows a downward trend. After freeze–thaw cycles, Units A and B show little change in peak stress and stress–strain curves due to n_d and d/h_w. In Unit C, specimens with a d/h_w of 0.6 and 0.7 exhibit a noticeable decrease in their stress–strain curves before and after freeze–thaw. In Units D and E, except for specimens with a d/h_w of 0.7, which experience a sudden drop in their stress–strain curves after 100 freeze–thaw cycles, the stress–strain curves of other units show no significant changes. Overall, specimens with an aspect ratio of 0.6 exhibit superior unit stress levels after freeze–thaw cycles compared to the other two groups of specimens. Therefore, it is recommended that d/h_w be set at 0.6.

7. Mechanism of Composite Columns under Axial Compression after Freeze–Thaw Cycles

7.1. Failure Modes of Specimens

Under axial loading, all specimens exhibit similar failure patterns. This paper presents the deformation and failure modes of two typical specimens, as shown in Figure 33. The failure modes before and after freeze–thaw are similar. Due to the compression of concrete, the central part of the specimens bulges outward, and the steel tube has a confining effect on the concrete, causing both the steel and concrete in the specimens to deform simultaneously. The central part of the specimens begins to protrude outward and presents an outward convex damaged shape. The honeycomb steel web undergoes plastic deformation, and the circular holes are gradually flattened. The honeycomb steel web serves not only to connect the two steel–concrete flanges but also to support the core concrete. This structural design can effectively delay the failure of the specimens, making them act as a single entity and preventing the flanges from bending inward. After 300 freeze–thaw cycles, the core concrete shows a larger area of internal damage compared to its pre-freeze–thaw state, the degree of internal damage in the specimens increases, the specimens fail more quickly, and the supporting effect of the honeycomb steel web on the flanges at both ends weakens.

7.2. Analysis of the Load-Bearing Process of the Specimen

The load–displacement curve trend of the full-scale composite column after undergoing freeze–thaw cycles is essentially the same as when it has not experienced freeze–thaw cycles. It can be roughly divided into four stages: the elastic stage (OA section), the elastoplastic stage (AB section), the load reduction stage (BC section), and the residual deformation stage (CD section). The stress of STHCC filling throughout the entire process is provided, along with cloud maps of different characteristic points in the four stages of non-freeze–thaw and post-freeze–thaw cycles, as seen in Figure 34.

The results indicate that the stress in the STHCC near the steel web is relatively high and distributed within the range of the web hole. In the elastic stage, the steel tube and concrete work independently. The stress at the edges of the web and the core concrete at the corners exhibit stress concentration. After 300 freeze–thaw cycles, the bearing capacity of the specimen decreases due to the significant damage to the core concrete. Compared to the non-frozen specimen at the same time, the longitudinal displacement increases, and the steel tube exerts restraint on the concrete earlier, resulting in stress concentration around the web, and an increase in the stress of the steel tube, as shown in Figure 35a. In the elastoplastic stage, due to the axial force, the middle part of the concrete undergoes circumferential deformation and is squeezed against the outer steel tube. At this point, the steel tube begins to exert restraint on the concrete, and the larger the axial force, the more obvious the restraint effect. As the circumferential deformation of the concrete increases, the concrete enters the elastoplastic stage and undergoes plastic deformation, and the steel tube exhibits a bulging phenomenon. At this time, the stress distribution of the steel tube is uniform, and the stress value is relatively high. Freeze–thaw cycles have a minimal impact on the stress process in this stage, and the stress processes before and after freezing–thawing are similar, as shown in Figure 35b. In the load reduction stage, due to the maximum circumferential deformation at the mid-span of the specimen and the support effect of the web, significant stress appears at the center of the concrete near the web. In contrast, local buckling deformation occurs at the mid-height of the steel web. After freeze–thaw cycles, the circumferential deformation increases, the steel tube yields earlier, the damage to the edges of the concrete at the quarter-height and half-height is more significant, and the specimen tends to be damaged, as shown in Figure 35c. In the residual deformation stage, the specimens that underwent 300 freeze–thaw cycles have a more extensive core concrete damage range and more severe damage, as shown in Figure 35d. Therefore, it is essential to pay attention to the earlier deformation and damage of the composite column with freeze–thaw damage and consider the impact of parameters on the bearing capacity of the composite column after freeze–thaw cycles.

8. Equation for Ultimate Bearing Capacity of Composite Columns under Axial Compression after Freeze–Thaw Cycles

To clarify their axial mechanical performance after experiencing freeze–thaw cycles, an extended parametric analysis was conducted for this class of composite columns. The confinement effect coefficient, concrete strength, slenderness ratio, web hole opening, and the number of freeze–thaw cycles were taken as the main parameters. On this basis, this study used superposition theory to fit and regress the axial load-bearing capacity equation for such composite columns. Due to the confinement effect coefficient and concrete strength, the enhancement of the load-bearing capacity is considered. And due to web hole openings and the number of freeze–thaw cycles, the weakening of the load-bearing capacity is considered. The basic form of the axial load-bearing capacity is proposed, as shown in Equation (9), and the specific expressions for three of the coefficients are provided, as shown in Equations (10)–(13). The expanded expression for the axial load-bearing capacity is presented in Equation (14).

$$N_u^c=A\cdot N_D(f_s{A}_s+\varphi \cdot h\cdot f_c{A}_c+\beta \cdot f_s{A}_s)$$

(9)

$$A=a{\lambda }^2+b\lambda +c$$

(10)

$$N_D=\left(d\cdot n_d^{\alpha }+e{n}_d+f\right)$$

(11)

$$\varphi =1+g\xi$$

(12)

$$\beta ={\left({\displaystyle \frac{d}{h_w}}\right)}^i+{\left({\displaystyle \frac{s}{h_w}}\right)}^k$$

(13)

$$N_u^c=A\cdot N_D\cdot \left\{f_s{A}_s+h\cdot \left(1+g\xi \right)\cdot f_c{A}_c+\left[{\left({\displaystyle \frac{d}{h_w}}\right)}^i+{\left({\displaystyle \frac{s}{h_w}}\right)}^k\right]\cdot f_s{A}_s\right\}$$

(14)

In the equations, f_s represents the strength of the steel; f_c represents the strength of the concrete; A_s, A_c, and A_w represent the cross-sectional areas of the flange steel tube, concrete, and steel web, respectively; α is the slenderness reduction factor; φ represents the steel tube confinement enhancement coefficient; β is the steel web reduction factor; N_D is the freeze–thaw influence factor; n_d represents the number of freeze–thaw cycles; λ represents the slenderness ratio of the specimen; ξ is the confinement effect coefficient; d/h_w is the aspect ratio of holes; and s/h_w is the aspect ratio. Parameters a, b, c, d, e, f, g, h, i, k, and α are all coefficients.

Through statistical regression using the 1st Opt software, the regression results are as follows: a = −0.017; b = 0.36; c = −0.899; d = −1.07 × 10⁻¹⁹; e = −0.0004; f = 0.99; g = 0.031; h = 0.8; i = 92; k = 39; and α = 6.74. Through substituting these coefficients into Equation (13), the final expression for the axial load-bearing capacity is obtained, as shown in Equations (15)–(17).

$$N_u^c=A\cdot N_D\left\{f_s{A}_s+0.8\cdot (1+0.031\xi )\cdot f_c{A}_c+\left[{\left({\displaystyle \frac{d}{h_w}}\right)}^{92}+{\left({\displaystyle \frac{s}{h_w}}\right)}^{39}\right]\cdot f_s{A}_W\right\}$$

(15)

$$A=-0.017{\lambda }^2+0.36\lambda -0.899$$

(16)

$$N_D=\left(-1.07\times {10}^{-19}\cdot n_d^{6.74}-0.0004n_d+0.99\right)$$

(17)

According to Equation (14), the ultimate axial load-bearing capacity (ULCC) of the 48 specimens after experiencing freeze–thaw cycles was calculated, denoted as ($N_u^c$). The results obtained from the equation calculations were compared with the ULCC derived from finite element simulations, denoted as ($N_u^s$). The comparison results are presented in

Table 4. Comparison of axial load-bearing capacity of specimens.

Specimens	nd	fcu/MPa	λ	fyk /MPa	ξ	${\mathit{N}}_{\mathbf{u}}^{\mathbf{s}}$/kN	${\mathit{N}}_{\mathbf{u}}^{\mathbf{c}}$/kN	$\frac{\left|{\mathit{N}}_{\mathbf{u}}^{\mathbf{c}}-{\mathit{N}}_{\mathbf{u}}^{\mathit{s}}\right|}{{\mathit{N}}_{\mathbf{u}}^{\mathbf{c}}}$
S1	0	70	11.08	345	0.23	22,031.7	21,639.1	1.78%
S2	100	70	11.08	345	0.23	21,127.6	20,764.7	1.72%
S3	200	70	11.08	345	0.23	20,042.4	19,882.9	0.80%
S4	300	70	11.08	345	0.23	17,710.2	18,900.1	6.72%
S5	0	70	11.08	345	0.35	23,737.4	23,534.9	0.85%
S6	100	70	11.08	345	0.35	22,685.6	22,583.9	0.45%
S7	200	70	11.08	345	0.35	21,646.8	21,624.8	0.10%
S8	300	70	11.08	345	0.35	20,624.8	20,555.9	0.33%
S9	0	70	11.08	345	0.48	25,641.7	25,414.8	0.88%
S10	100	70	11.08	345	0.48	24,605.8	24,387.9	0.89%
S11	200	70	11.08	345	0.48	23,670.1	23,352.2	1.34%
S12	300	70	11.08	345	0.48	23,524.9	22,197.9	5.64%
S13	0	70	11.08	345	0.6	27,481.6	27,268.7	0.77%
S14	100	70	11.08	345	0.6	26,518.8	26,166.8	1.33%
S15	200	70	11.08	345	0.6	25,520.2	25,055.6	1.82%
S16	300	70	11.08	345	0.6	24,511.1	23,817.1	2.83%
S17	0	60	11.08	345	0.27	19,766.7	19,201.6	2.86%
S18	100	60	11.08	345	0.27	18,832.9	18,425.7	2.16%
S19	200	60	11.08	345	0.27	17,912.1	17,643.3	1.50%
S20	300	60	11.08	345	0.27	16,642.0	16,771.1	0.78%
S21	0	80	11.08	345	0.2	24,511.7	24,076.6	1.78%
S22	100	80	11.08	345	0.2	23,233.5	23,103.7	0.56%
S23	200	80	11.08	345	0.2	22,065.3	22,122.6	0.26%
S24	300	80	11.08	345	0.2	20,489.6	21,029.0	2.63%
S25	0	70	11.78	345	0.23	21,603.4	21,206.8	1.84%
S26	100	70	11.78	345	0.23	20,500.0	20,349.9	0.73%
S27	200	70	11.78	345	0.23	19,500.0	19,485.7	0.07%
S28	300	70	11.78	345	0.23	18,355.2	18,522.5	0.91%
S29	0	70	12.47	345	0.23	20,920.9	20,428.8	2.35%
S30	100	70	12.47	345	0.23	19,839.5	19,603.3	1.19%
S31	200	70	12.47	345	0.23	18,500.0	18,770.9	1.46%
S32	300	70	12.47	345	0.23	17,710.2	17,843.0	0.75%

and Figure 36. Upon comparison, it is found that the difference between the results calculated by the equation and those from the simulations is 6.72%. This difference meets the requirements for engineering error, indicating that the equation can accurately calculate the ULCC of such composite columns after experiencing freeze–thaw cycles.

9. Conclusions

An analysis of the axial mechanical properties of 48 full-scale STHHC composite columns was conducted, with the confinement effect coefficient (ξ), concrete strength (f_cu), freeze–thaw cycles count (n_d), slenderness ratio (λ), aspect ratio (s/h_w), and aspect ratio of holes (d/h_w) as the main control parameters. Within this parameter range, the axial mechanical properties of this class of composite columns after freeze–thaw cycles were studied, leading to the following conclusions:

(1)

The load-bearing capacity of STHHC composite columns increases with an increase in ξ and f_cu and decreases with a rise in λ. The influence of s/h_w and d/h_w on the ultimate load-bearing capacity of the composite columns is relatively small. The ductility of STHHC composite columns increases with the increase in ξ and decreases with a rise in λ and f_cu. The impact of s/h_w and d/h_w on the ductility of the specimens is minimal.

(2)

As ξ increases, the rate of decrease in the ultimate load-carrying capacity of the specimens after freeze–thaw cycles (n_d) gradually slows down. When ξ is between 0.23 and 0.48, the increase in ξ helps to mitigate the damage caused by the n_d to the concrete, with more noticeable confinement effects in areas closer to the center of the specimen. As the concrete strength grade (f_cu) increases, the degree of influence of n_d lessens. The n_d causes a certain degree of damage to the core concrete inside the steel tube. The higher the n_d, the greater the degree of damage. As λ increases and the number of n_d increases, the ultimate load-carrying capacity of the specimens will gradually decrease.

(3)

Moreover, the larger the λ, the greater the decrease in the ultimate load-carrying capacity of the specimens after n_d, and the steel tube has a weaker confining effect on specimens, and the stress decrease in the central unit position is more apparent. With the increase in n_d, the ductility of the composite column gradually decreases, but the numerical change is small, and the impact is minimal. Within the research range of this study, the larger the s/h_w, the slower the rate of decrease in the ultimate load-carrying capacity. The change in d/h_w has a minimal impact on the ultimate load-carrying capacity of the specimens after n_d.

(4)

All specimens exhibited excellent load-bearing capacity and freeze–thaw resistance. During loading, the high-strength concrete and steel undergo synchronous deformation, causing the middle part of the specimen to bulge. At this time, the honeycomb steel web will gradually be compressed, and the honeycomb holes will become progressively elliptical. This indicates that the steel web also participates in the loading process. The web provides certain support to the concrete flanges on both sides of the steel tube, making the confining effect of the inner side of the flange more apparent.

(5)

STHHC composite columns that have undergone freeze–thaw cycles can be divided into four main stages when subjected to axial pressure: the elastic stage, the elastoplastic stage, the load decrease stage, and the residual deformation stage. Based on the consideration of the enhancement of the load-bearing capacity of the composite column by ξ and f_cu, and the weakening of the load-bearing capacity by the web opening and n_d, the 1stOpt software was used for fitting and regression to obtain the calculation equation for the axial load-bearing capacity of STHHC composite columns under freeze–thaw cycles. In comparing the ultimate load-bearing capacity calculated by the equation with that obtained from finite element analysis, the maximum error is 6.72%, which meets the engineering accuracy requirements.

(6)

In order to reduce the damage caused by the decrease in the bearing capacity caused by the freeze–thaw of concrete, it is suggested that the ξ of the composite column after freeze–thaw cycles should be controlled at about 0.48. From the perspectives of giving full play to material properties and the economy, it is recommended to use 70 MPa for f_cu after freeze–thaw cycles. After a certain number of freeze–thaw cycles, the ultimate bearing capacity of the specimen will gradually decrease. At the same time, with the increase in the λ of the specimen, the decrease in the bearing capacity after the freeze–thaw cycles is greater. When the λ is between 11.08 and 11.78, the ultimate bearing capacity of the specimen decreases slightly under different freeze–thaw times. It is recommended that the λ is about 11.08. If the s/h_w is too large, the initial stiffness of the composite column will be weakened. Therefore, it is recommended to use 0.6 for the s/h_w of this type of composite column under freeze–thaw cycles. If the d/h_w is too small, the initial stiffness of the composite column will be reduced, and if it is too large, the ductility of the specimen after freeze–thaw cycles will be reduced. Therefore, this paper suggests that the d/h_w should be 0.25, and the thickness of the web should be 10 mm.

(7)

In this paper, the axial compression mechanical properties of STHHC composite columns under freeze–thaw cycles are analyzed. However, in practical engineering, members completely subjected to axial compression are rare. Therefore, in subsequent research, the eccentric compression mechanical properties of such composite columns after freeze–thaw cycles can be determined. The research on the axial compression performance in this paper is based on finite element simulation, and then the axial compression test of the STHHC composite columns under freeze–thaw cycles was carried out to compensate for the lack of experimental research in this paper. In addition, the concrete material used in this study was high-strength concrete. In the subsequent research, combined with the needs of modern buildings, new materials such as fiber-reinforced concrete, reactive powder concrete, and recycled concrete can be added to the steel pipe to study its mechanical properties under a freeze–thaw environment. The constitutive model provides technical support for the application of new materials.