Experimental and Mechanism Study on Axial Compressive Performance of Double Steel Tube Columns Filled with Recycled Concrete Containing Abandoned Brick Aggregate

Abstract

Recycled concrete is widely recognized as favorable for environmental protection and sustainable development. However, recycled concrete, especially containing abandoned brick aggregate, is rarely used in main structural members due to its inherent defects. Concrete-filled double steel tube columns (CFDSTCs), consisting of an outer and an inner steel tube with concrete filling the entire section, are effective in load bearing and deformation resistance. The structural application of abandoned brick aggregate, resulting from urbanization renewal, might be widened through CFDSTCs. This paper presents an experimental and analytical study aiming to investigate the axial compressive behavior of recycled-brick-aggregate-concrete-filled double steel tube columns (RBCDSTs). A total of six specimens were tested under concentric compression, including five RBCDSTs and one concrete-filled single steel tube column. The varied parameters included the replacement ratios (0% and 25%) of brick aggregate and the thickness ratio of the inner and outer steel tubes (0.75, 1, and 1.25). Theoretical analysis was also carried out. A new constitutive model of RBCDST was proposed and used in finite element analysis. The investigation indicated that, under the current conditions, the presence of the inner steel tube only increased the strength by 0.14%. When the inner and outer diameter ratio is 0.73, using a 25% replacement rate of bricks in the entire cross-section or only in the ring area of the cross-section will result in 21.1% and 10.1% strength decreases, respectively. For every 0.6% increase in the diameter-to-thickness ratio of the outer tube, the strength of RBCDST increases 16.3% on average.

1. Introduction

In recent years, the increased rate of industrialization and urbanization have led to the generation and release of larger amounts of construction and demolition (C&D) waste. These wastes account for a large part of solid waste in most countries around the world and are commonly and ineffectively disposed of in landfills at significant cost. On the other hand, the demand for building materials is continually increasing, which increases the consumption of natural resources and energy [1]. The mining, processing, and transport operations used for the acquisition and haulage of large amounts of aggregate consume considerable amounts of energy and produce carbon emissions. To conserve natural aggregate resources and minimize the environmental impact of C&D waste, recycled aggregate (RA), which is obtained from C&D waste, has been considered as an alternative to natural aggregates in structural concrete over the last two decades [2]. Many studies on the physical, mechanical, and durability-related properties of RA concrete (RAC) have been carried out over time [3,4,5].

The properties of concrete have close relative effects on the properties of aggregates used. Many properties of RAC have been widely recognized. Studies have shown that the mechanical strength of concrete made from recycled aggregate is lower than that of natural aggregates [6]. The compressive strength, elastic modulus, and splitting tensile strength of RAC decrease with an increase in the recycled aggregate replacement ratio. It can be attributed to the additional interfacial transition zone (ITZ), a weak zone between the mortar and the recycled aggregate surface. It may also be explained by the higher crushing values of recycled aggregate [7]. Research on RA has also focused on the water absorption capacity. Recycled aggregates have been shown to have more water absorption capacity than natural aggregates due to their greater porosity. The compressive strength of RAC increases with an increase in the effective water-to-cement ratio. RAC with a larger particle size or lower amount of attached mortar exhibit higher compressive strength. The elastic modulus and compressive strength are closely related to the properties of the parent concrete [8].

Given the well-known disadvantages of RAC—including lower strength, larger water absorption, higher shrinkage and creep, poorer durability, and unstable properties—its widespread structural application has been seriously limited [9,10,11]. Various attempts have been made to change the performance of RAC for practical engineering applications [12]. Many studies have demonstrated that the mechanical properties of RAC can be significantly improved when subjected to three-dimensional stress. In other words, under the lateral confinement provided by excellent tensile materials, like steel- or fiber-reinforced polymer, the mechanical properties of the core RAC could be improved significantly [13,14]. The concrete-filled steel tube (CFST) is an effective form for constraining concrete to enhance its performance under three-dimensional loading conditions. Many studies have outlined the advantages of CFST [15,16,17]. Jelena et al. [18] conducted a comprehensive review of the structural behavior of RAC-filled steel tubes, finding that while the influence of recycled aggregate replacement ratios on column axial performance is negligible, it significantly influences the initial axial stiffness and extent of long-term deformation. Abdikarim et al. [19] proposed a prediction model on the compressive strength of RAC-filled steel tube columns, which account for the diameter, thickness, and yield strength of the steel tube, length of the column, recycled aggregate replacement ratio, and concrete compressive strength. Yuan et al. [20] focused on the axial behaviors of an RAC-filled steel tube confined with GFRP. It was observed that as the substitution rate of recycled aggregate increased, the bulging deformation and ductility of the specimen also increased, while the bearing capacity and stiffness of the specimen decreased. Some studies have investigated the axial behaviors of RAC-filled double-skin hollow columns. Bai et al. [21] revealed that the influence of the replacement ratio on axial performance is negligible, which can be offset by improving the confinement stiffness. Xiong et al. [22] indicated that the rupture of the outer FRP tube and buckling of the inner steel tube is the representative failure mode of RAC-filled double skin columns. It is currently widely accepted that filling the hollow region of the double skin column with concrete, forming a double tube column, would be more effective for load bearing and deformation. The conclusion of Bai et al. [21] is consistent with the findings on double-tube solid columns reported by Xiong et al. [23]. The performance shortcomings of specimens with RAC can be remedied by strong confinement. The higher bearing capacity of double-tube solid RAC columns has been attributed to the core concrete delaying buckling of the inner steel tube [24]. Huang [25] conducted an experimental and analytical study on the progressive failure of GFRP-confined geopolymeric recycled aggregate. The effects of the confinement ratio, recycled aggregate content, and tube size were investigated.

In the past 10 years, crushed clay brick has been the main component in C&D waste in China [26]. The clay brick aggregates, with larger water absorption and higher compressive crush value, have many inherent defects compared to natural aggregate concrete. Numerous studies have attempted to improve the performance of recycled brick aggregate concrete (RBAC) in terms of various means. It was found that the structural effects of the crushed clay brick replacement ratio are much smaller than the corresponding effects in the material property tests. The deformation rate of RBAC is higher than that of concrete using natural aggregates. Ali Ejaz [27] focused on improving the performance of RBAC with steel clamps. Their research demonstrated that steel clamp confinement significantly modified the compressive failure of concrete, transitioning from brittle to ductile failure. Panumas [28,29] attempted to improve the performance of RBAC using hemp-fiber-reinforced polymer composites. Their research showed that the compressive strength and strain were enhanced by up to 181% and 564%. Analytical expressions predicting the stress–strain curves of HFRP-confined waste brick aggregate concrete were developed. Krisada Chaiyasarn [30] studied the effect of wrapping natural fiber rope on RBAC, concluding that the strain performance falls short compared to fully wrapped HFR confinement. Panuwat Joyklad [31] tried to improve the performance of RBAC with low-cost fiber chopped strand mat composites. Analytical expressions predicting the stress–strain curves were proposed. RBAC has a lower compressive strength than fresh concrete, and it can derive significantly enhanced strength and deformability under FRP confinement.

Most research has focused on improving the performance of RBAC by single confinement using steel, fiber polymers, or other materials [32,33,34,35,36]. The wider application of BRAC requires stronger confinement for assistance. To date, few studies have focused on RBAC being confined using double confinement. Huang et al. [37] investigated using RBAC in a structural specimen through dual confinement with flax FRP and steel spirals. The research indicated that the dual-confined RBAC specimen presents better ductile properties compared to unconfined and single-confined RBAC specimens. The confined RBAC cylinders maintained better integrity and showed higher compressive strength due to the dual confinement. Gao et al. [38] investigated the improvement of RBAC compressive strength through dual confinement using PVC–FRP tubes. The study demonstrated that the FRP–PVC hybrid confining system notably enhanced the compressive strength, as well as the axial and lateral deformations, of the RAC–RBA. The comparison test results indicated that GFRP and CFRP tube confinement resulted in much greater enhancement in the ultimate compressive strength of RAC–RBA, due to the much higher tensile modulus and strength of these G/CFRP composites.

However, dual confinement using brittle FRP materials is often unstable and laborious. Additionally, the closely adhered double-layer confinement shows little difference compared to single-layer confinement. There is still a long way to go to effectively utilize RBA with dual confinement, as limited data and few studies are available in the existing literature.

This paper mainly focuses on improving RBA properties through dual confinement, which was provided by double steel tubes. A new composite column, the recycled-brick-aggregate-concrete-filled double steel tube column (RBCDST), was studied using six specimens. The axial compressive experiments, finite element analysis, and proposed formula aim to break new ground in addressing the troubling social problems of today.

2. Experimental Program

2.1. Test Specimens

In this study, a total of six short column specimens, including five recycled-brick-aggregate-concrete-filled double steel tube columns and one control concrete-filled steel tube, were tested at the structural laboratory of Shenyang Construction Engineering Quality Testing Center Co., LTD. (Shenyang, China). All specimens were 650 mm in height. For the double steel tubes, the inner and outer diameters of the steel tubes were 165 mm and 120 mm, respectively. The dimensions of specimens were mainly decided by considering the bearing capacity of the loading equipment and the traditional pattern in the existing literature. As shown in existing research, the height of short columns is nearly 2–3 times that of the diameter [39]. The distance between the inner and outer tubes is larger than the average particle size of the aggregate.

The considered parameters included the RBA replacement ratio of the inner core concrete and outer ring concrete. Recycled aggregate concrete with a 25% brick aggregate replacement ratio and common concrete without recycled aggregates were poured into six specimens, based on pre-design conditions. All specimens have a 4 mm thick inner tube and 4 mm, 5 mm, and 6 mm thick outer tubes, respectively. The inner steel tube was placed inside the outer tube and fixed at the center location of the outer tube with four short steel sticks around the inner tube, as shown in Figure 1. The steel sticks, measuring almost 22 mm in length and 6 mm in diameter, were pre-welded onto the surface of the inner steel tube before concrete casting. Detailed parameters of the specimens are shown in

Table 1. Detailed information on specimens.

No.	RBA Replacement Ratio of Inner Concrete /%	RBA Replacement Ratio of Ring Concrete /%	Thickness of Inner Steel Tube /mm	Thickness of Outer Steel Tube /mm
4-25-0-25	25	25	None	4
4-25-4-25	25	25	4	4
4-25-4-0	0	25	4	4
4-0-4-0	0	0	4	4
5-25-5-25	25	25	4	5
3-25-4-25	25	25	4	3

. The specimens in the table are named in the pattern of “A–B–C–D”: “A” and “C” denote the thickness (in mm) of the outer and inner steel tubes, respectively, and “B” and “D” refer to the RBA replacement ratio (expressed as a percentage) of coarse aggregates of the outer ring and inner core concrete.

2.2. Materials

2.2.1. Concrete

In this study, coarse aggregates were mixed with recycled brick aggregates instead of natural aggregates. A total of two mixtures were adopted, which corresponded to two replacement percentages—0% and 25%—for RAC. As shown in previous studies, a substitution rate of 25% is the point at which strength is significantly affected [40]. The concrete was manufactured by mixing the cement with ordinary water, fine aggregate, NA, and RBA. For cement, ordinary Portland cement with 42.5 MPa strength was adopted. River sand with a diameter of less than 2 mm was used as the fine aggregate. The particle size of the coarse aggregates, divided into natural aggregates and RBAs, ranged from 5 to 20 mm, as shown in Figure 2a,b and Figure 3b. The RBAs were meticulously screened manually from the mixing recycled aggregates, which were provided by Shenyang Ring City Ecological Environment Management Co., Ltd. (Shenyang, China). The RBAs, with special mechanical properties, were tested first, and the RBA properties were subsequently tested. Specimen preparation followed the Chinese code ‘Standard for Test Methods of Concrete physical and mechanical properties’ [41]. Details on the RBAs and concrete are outlined in

Table 2. Aggregate properties.

Aggregates	Apparent Density of Aggregate /kg/m3	Crush Value /%	Water Absorption /%
NA	2720	8.6	1.2
RBA	2340	33.2	15.2

. The mix ratio design and concrete properties are shown in

Table 3. Mix ratio design and properties of concrete.

Cement kg/m3	Water kg/m3	NA kg/m3	Fine Aggregates kg/m3	RBA kg/m3	RBA Replacement Ratio in Coarse Aggregates /%	Water-Reducing Admixture kg/m3	Cube Compressive Strength of Concrete /MPa	Elastic Module /MPa
540	165.0	790.0	900	0	0	16	51.7	30,446
540	195.0	592.5	900	197.5	25	16	39.8	20,807

. The admixture dosage recommended by the manufacturer was used for our specific concrete mix design, combined with our laboratory-optimized proportions to achieve the target workability and strength requirements. The concrete slump was tested during casting. The slump of recycled concrete is 158 mm on average. The recycled concrete meets the requirements for workability. After curing under the condition of 600 °C·day, compressive testing on a 150 mm × 150 mm × 300 mm specimen was carried out in the laboratory. The stress–strain curve of concrete with a 25% RBA replacement ratio is shown in Figure 3d. It is worth noting that, after the peak, the test block suddenly cracked, and the descending section could not be accurately measured.

2.2.2. Steel

The steel tubes were manufactured from Q235 steel. Three standardized tensile specimens were fabricated using steel with thicknesses of 3 mm, 4 mm, and 5 mm. The steel performance was tested based on tensile test methods for metal materials. Figure 3a shows the tensile specimens, and the mechanical properties of steel are presented in

Table 4. Properties of steel tubes.

Number	Width /mm	Thickness /mm	Failure Load /kN	Yield Load /kN	Tensile Strength /MPa	Yield Strength /MPa
1	20.0	3.0	31.6	24.7	395	310
2	20.0	4.0	34.9	30.0	435	340
3	20.0	5.0	47.6	33.9	475	375

. The specimen preparation procedure followed the Chinese code ‘Metallic materials—Tensile testing—Part1: Method of test at room temperature’ [42].

2.3. Experimental Instrumentation

For the specimens tested in this study, four pairs of strain gauges were installed on the middle-height area equally installed along the circumference of the outer steel tube. Each pair included one longitudinal strain gauge to measure the axial strain during loading and one horizonal strain gauge to measure the lateral strain gauge, as shown in Figure 4a,b. The strain gauges attached to the outer steel tube surface are all adhesive-based strain gauges with resistance specifications of 3 × 20 mm, which ensures the accuracy of strain measuring of the steel tube. The resistance value of the strain gauge is usually measured before the test. An effective working resistance value of 120 ± 4 is considered. In addition, one linear variable differential transducer (LVDT) was used to monitor the longitudinal displacement of specimens for each specimen. The strain and displacement gauges on the specimen were all connected to the DH3825 dynamic signal acquisition instrument, which automatically collected the data. All the specimens were tested under axial compression provided by a 5000 kN pressure testing machine. Two ball hinges were placed at the top and bottom side of the specimen, respectively. Each ball hinge connected to a loading plate with high stiffness to ensure uniform distribution of compressive stress, as presented in Figure 4a,b.

2.4. Test Procedure

A pre-test procedure was conducted to ensure proper alignment, calibration of instruments, and system readiness prior to the actual loading test. The device, including the strain gauge and LVDT, were checked and calibrated. The test employed load-controlled loading, starting with a loading rate of 2 kN/s. Upon reaching half of the ultimate load, the loading rate was adjusted to 1 kN/s. As the calculated ultimate load approached, the loading rate was further reduced to 0.5 kN/s. The loading was stopped when the displacement reached 40 mm after the peak bearing capacity.

3. Test Results and Discussion

3.1. Test Phenomenon and Failure Modes

Figure 5 shows the failure modes of all specimens. Although each specimen has different characteristics, the test procedure and failure modes were basically similar. At the initial elastic stage, no obvious phenomenon was found for any of the specimens. When the specimen reached about 40–55% bearing capacity load, the crushing sound of concrete was heard. Outward buckling of the outer steel tube is the typical characteristic of specimens. Buckling in all specimens appeared after reaching the ultimate bearing load, typically at about mid-height or at one-third the height from the upper and lower ends.

In the process of the experiment, the bearing capacity, deformation, and lateral and axial strain were measured. The confinement factor of the inner and outer tubes and replacement ratio of RBA were studied through load–displacement curves and load–strain curves formed from the initial data.

The specimens were cut open along the longitudinal cross-section after the experiment, in order to observe the internal situation. Specimen 4-25-0-25 is a control specimen with a 4 mm thick steel tube and 25% RBA substitution ratio. For specimen 4-25-0-25, buckling occurred on the body of the steel tube, about 23 mm from the column foot, along the column height. After removing the steel surface, large oblique cracks, penetrating about 45° from the bottom to a 23 mm height along the column, could be seen on the core concrete. It is obvious that the sharp bulging of the external steel tube of the specimen is not caused by the expansion of the compression section of the internal concrete under axial pressure, but by shear failure of the internal RBAC. Further investigation is warranted, as steel-tube-confined RAC may exhibit more pronounced brittle failure characteristics compared to ordinary concrete columns confined by steel tubes. Specimens 4-25-4-25, 5-25-4-25, and 3-25-0-25 are all confined by double steel tubes. All three have a common RBAC substitution ratio (25%) and inner steel tube thickness (4 mm) but different outer steel tube thicknesses (3 mm, 4 mm, and 5 mm). Although the bulge appears on the steel tube of each specimen, as shown in Figure 5, the internal cutting surface shows that there are no obvious oblique shear cracks on internal concrete. It is demonstrated that the existence of the inner steel tube strengthens the shear resistance of the specimens, the ductility of which is obviously improved. Bulges were all located at a half to a third of the column height of specimens, resulting from the dilation of the inner concrete under axial load. Comparatively, specimen 5-25-4-25, with the highest thickness, exhibited the best deformation characteristics. Seen from the cross-section of the specimen, the ring sharp concrete between the inner and outer steel tubes was more extensively damaged, while damage to the core concrete confined by the inner tube was negligible. Specimen 4-25-4-0 is filled with different concretes in the inner tube and the gap area between inner and outer tubes. Specimen 4-0-4-0 is the specimen whose inner and gap area between inner and outer tubes are all filled with ordinary concrete. Specimens 4-25-4-0 and 4-0-4-0 also showed good symmetric deformation ability, but the concrete in the ring area of specimen 4-25-4-0 was more seriously deformed than that in specimen 4-0-4-0, and the concrete in specimen 4-0-4-0 was less damaged.

As shown in Figure 5, the fragmentation of aggregates always appeared on the specimens with RBAC, which can be attributed to the high crushing value of recycled concrete. However, the degree of fragmentation of the core recycled concrete aggregate is much less than that of ring concrete. It can be deduced that the load increase initially results in crushing of the ring concrete. The recycled brick aggregate, mortar, and concrete break together, resulting in a slight decrease in the bearing capacity of the component, while some areas of the outer steel tube begin to buckle. As the load continues to increase, the inner tubes begin to play roles, resulting in a secondary ascending phase.

3.2. Ultimate Load Analysis

The ultimate load of each specimen is shown in

Table 5. Ultimate load of specimens.

Name of Specimen	Ultimate Load /kN	Ultimate Strength /MPa
4-25-0-25	2048.6	93.6
4-25-4-25	2052.9	93.7
4-0-4-0	2485.8	113.5
5-25-4-25	2516.8	114.9
3-25-4-25	1862.2	85.1

. It can be seen from the data in the table that the bearing capacities of specimens with different conditions are different. It must be stated that the number of test specimens is indeed relatively small. Subsequently, the sample size needs to be expanded and the experiment repeated to enhance the reliability of the conclusion. The ultimate load of specimen 4-25-0-25 is 2048.6 kN, which is nearly the same as that of specimen 4-25-4-25. The ultimate load of specimen 4-25-0-25, with an inner steel tube, is just 0.14% higher than that of the counterpart without an inner steel tube. It is obvious that the inner steel tube under this condition had a limited effect on the ultimate load of specimens.

Based on the existing literature, a general conclusion can be drawn that the bearing capacity of specimens with internal pipes is generally higher than that of specimens without internal pipes, to varying degrees. As seen in

Table 6. Bearing capacity gain of double-tube confinement.

Source	Specimen	Type	Concrete	Bearing Capacity Gain%
Rong et al. [11]	GRCC(5)-1,2	GFRP	RAC	1.2%
	DRCC102(5)-1,2	GFRP/steel
	GRCC(8)-1,2	GFRP	RAC	25%
	DRCC140(8)-1,2	GFRP/steel
YL Long et al. [39]	1S0FL	Steel	Concrete	12%
	1S1FL-100-2	Steel/FRP
	1S1FL-150-2			22%
	1S1FL-200-2			30%
	1S0FH			12%
	1S1FH-150-3
	1S1FH-200-3			19%
Talha et al. [43]	CC1-OT1	Steel/steel	Concrete	-
	CC1-SC1-OT1			66%
	CC2-OT1			-
	CC2-SC1-OT1			32%

, the gain of the bearing capacity is related to the type of confined object, confining material, and diameter ratio of the outer and inner tubes. The larger the constraint diameter, the higher the bearing capacity. For RAC, the effect of double-tube restraint on enhancing the bearing capacity is not as significant as that of ordinary concrete, especially for those with lower strength.

Specimens 4-25-4-25, 5-25-4-25, and 3-25-4-25 share the same concrete arrangement and inner steel tube thickness but differ in outer steel tube thickness. As shown in Figure 6a, the ultimate loads of specimens 3-25-4-25, 4-25-4-25, and 5-25-4-25, with outer steel tube thicknesses of 3 mm, 4 mm, and 5 mm, are 1862.2, 2052.9, and 2516.8 kN, respectively. It can be seen that the ultimate load of specimen 4-25-4-25 with a 4 mm thick outer tube is 10.2% higher than with a 3 mm thick outer tube, and the ultimate load of specimen 5-25-4-25 with a 5 mm thick outer tube is 35.2% higher than with a 3 mm thick outer tube. It can be concluded that the thickness of the outer tube has a significant effect on the ultimate load of RBCDST. The bearing capacity of RBCDST increases with the gradual increase in outer steel thickness.

As shown in Figure 6b, specimens 4-25-4-25, 4-25-4-0, and 4-0-4-0 have the same steel tube configuration but differ in the concrete fill, with different RA replacement ratios. It can be seen from the data in Table 5 that the specimens in this group (4-25-4-25, 4-25-4-0, and 4-0-4-0) have similar values, indicating a subtle effect of the recycled aggregate replacement rate on the concrete strength under double confinement. The bearing capacity of specimen 4-25-4-25, filled with 25% brick aggregate mixed recycled concrete, is 2052.9. The bearing capacity of specimen 4-0-4-0, filled with 0% brick aggregate mixed recycled concrete, is 2485.8. The ultimate load of 4-0-4-0 seemed to be 21.1% higher than that of 4-25-4-25. The bearing capacity of specimen 4-25-4-0, with a natural aggregate concrete filling inner tube and 25% brick aggregate mixed recycled concrete filling ring region between the outer and inner steel tube, is 2256.6 kN. The ultimate load of specimen 4-25-4-0 is 9.9% higher than that of specimen 4-25-4-25. Specimen 4-0-4-0 was found to have the highest ultimate load, indicating that stronger inner core concrete and ring concrete has great significance to the bearing capacity of double steel tube columns.

3.3. Axial Stress–Strain Curve Analysis

Figure 7a–c present the axial stress–strain curves. In Figure 7a, 4-25-0-25 and 4-25-4-25 have nearly equal ultimate stress values. The initial stiffness of specimen 4-25-4-25 with an inner tube is slightly smaller than that of specimen 4-25-0-25. However, the ultimate strength of specimen 4-25-4-25 occurs during the second upward stage of the curve. The specimen with an inner tube maintained a long-term stable bearing capacity after the peak stage, as well as slight recovery after the peak. This may be attributed to the function of the inner steel tube. The stress–strain curve of recycled brick concrete constrained by most double steel tubes shows an inferior fovea curve and secondary ascending section, which is mainly related to the constraint mechanism of the double steel tubes and the force characteristics of recycled brick aggregate concrete. After the outer steel tube yields, the inner steel tube continues to provide constraints: at the initial stage of loading, both the inner and outer steel tubes and the concrete bear the force together, and the load increases with the increase in displacement. When the outer steel tube reaches the yield strength first, the growth of its restraint capacity slows down. At this time, the upward trend in the load may slow down, or a platform section may appear. However, as the inner steel tube has not yet yielded, it can still provide effective restraint to the concrete. As the displacement increases, the restraint effect of the inner steel pipe gradually becomes prominent, further enhancing the compressive strength of the concrete and thus forming a secondary ascending section (R3-8).

Figure 7b presents a comparison of the axial stress–strain curves of specimens 4-25-4-25, 5-25-4-25, and 3-25-4-25, which have various outer steel tube thicknesses. In Figure 7b, the shapes of the axial stress–strain curve of specimens with 5 mm thick outer tubes are different from those with 3 mm and 4 mm thick outer tubes. The lack of axial strain after the peak was due to damage to the strain gauge. The curves of specimens 4-25-4-25 and 3-25-4-25 also exhibit a large difference. The curves of specimens 4-25-4-25 and 5-25-4-25 exhibit a distinct concave stage, whereas the curve of specimen 3-25-4-25 shows an opposite, convex curve. This difference in stress–strain curves is often attributed to “strong confinement” and “weak confinement”. For a common concrete-filled tube column, regardless of whether a steel tube or FRP tube is used, the fullness of the stress–strain curve is guaranteed by sufficient restraint. The mutational point and concave stage exist on the curves of specimens with weaker confinement. Nevertheless, for the double-steel-tube-confined recycled concrete mixed with brick aggregates, the larger confinement may result in the premature fragmentation of brick aggregate. The specimens with stronger confinement present concave curves. Stronger confinement increases the bearing capacity but may be detrimental to maintaining it over time. Therefore, it is critical to match the constraint level to the core concrete properties to ensure optimal performance. In this study, 3-25-4-25 is the appropriate confinement for recycled brick aggregate concrete.

Figure 7c presents a comparison of axial stress–strain curves of specimens 4-25-4-25, 4-25-4-0, and 4-0-4-0, which indicates the replacement ratio effect on the behaviors of specimens. As presented in Figure 7c, the shapes of two curves (4-25-4-25 and 4-0-4-0) were similar, including the ascending stage, first peak point, concave stage, and recovery stage. Specimen 4-25-4-25, with a 25% brick recycled aggregate replacement ratio, has the lowest first peak stress but more pronounced behavior in the recovery stage compared to the other two specimens. Specimen 4-0-4-0, filled with common concrete without brick recycled aggregates, presents similar first peak stress and recovery performance. However, it can be seen that specimen 4-25-4-0, with different replacement ratios of concrete in the inner and outer tubes, presents the best behaviors, demonstrating higher first peak load and the best recovery performance. Specimen 4-25-4-0 has the best confinement combination, whose stress–strain curve is full and smooth without a concave stage. It is concluded that a combination of a strong core and weak ring region is the best choice for double steel tube columns.

3.4. Transverse Strain Analysis

Figure 8a–c present the axial stress–transverse strain curves of specimens. The axial stress–lateral strain values in Figure 8 are all average values measured by the axial and lateral strain gauges on the surface of the outer steel tubes. It can be seen from Figure 8a that the deformability of specimens without an inner tube is much worse than those with double tubes. The lateral strain of double-tube specimens also shows a second stage of increase.

Figure 8b presents the transverse strain curves of specimens 4-25-4-25, 4-25-4-0, and 4-0-4-0. It can be seen from Figure 8b that the curve shape of lateral strain is similar to that of axial strain. The axial strain is higher than the lateral strain under unique load. The strain curve of specimen 4-25-4-25 presents a more obverse secondary ascending stage, which is different from the mutant shape observed for specimens 4-25-4-0 and 4-0-4-0. Figure 8c is the axial and lateral strain curves of specimens with different outer tube thicknesses. It is obvious that the transverse strain of the specimens under the three working conditions varies little with the load, at the initial stage before 1200 με. With the increase in load, the curve starts to deviate. With the increase in the wall thickness of the outer tube, the deformability of the specimen also increases gradually.

At present, there are many studies focused on recycled concrete. The corresponding stress–strain curves of the specimens show different shapes. As shown in Figure 9a, the stress–strain curve of the single-steel-tube-restrained brick recycled concrete column shows a relatively smooth downward curve. Figure 9b is also a study of steel-tube-confined recycled concrete. The curves of the aforementioned specimens all exhibit the characteristics of a weakly constrained concave curve.

4. Constitutive Model of Double-Confined RBAC

The constitutive model reflects the relationship between the strain and stress of the specimen, which is vital for the research and application of the specimen. Many studies focus on the constitutive model of the relative material, steel or RBAC, and some also focus on the constitutive model of single- or double-confined specimens.

4.1. Constitutive Model of Steel Tube

Many constitutive models of steel tube could be derived from the existing literature. The model used by Mizan Ahmed [45] performs well in describing reinforced concrete members confined with double steel tubes as follows:

$$\left\{\begin{array}{ll}{\sigma }_s=E_s{\epsilon }_s& {\epsilon }_s\le 0.9{\epsilon }_{sy}\\ {\sigma }_s=f_{sy}{\left({\displaystyle \frac{{\epsilon }_s-0.9{\epsilon }_{sy}}{{\epsilon }_{st}-0.9{\epsilon }_{sy}}}\right)}^{{\displaystyle \frac{1}{45}}}& 0.9{\epsilon }_{sy}<{\epsilon }_s\le {\epsilon }_{st}\\ {\sigma }_s=f_{su}-{\left({\displaystyle \frac{{\epsilon }_{su}-{\epsilon }_s}{{\epsilon }_{su}-{\epsilon }_{st}}}\right)}^n(f_{su}-f_{sy})& {\epsilon }_{st}<{\epsilon }_s\le {\epsilon }_{su}\end{array}\right.$$

(1)

where n = E_st (ε_su − ε_st)/(f_su − f_sy); σ_s and ε_s represent the axial steel stress and strain; f_sy and ε_sy stand for the steel yield strength and yield strain; and ε_st denotes the strain at the onset of strain hardening, equal to 0.005. ε_su is equal to 0.2, and E_st can be calculated as 0.02E_s. E_st is the elastic modulus of the steel tube. The second formula was derived from the research of Liang [46], and the third was derived from a study carried out by Mander [47].

4.2. Constitutive Model of Confined RBAC

Many models of confined RBAC are available in the recent literature [32,33,34,35,36,37,38,46,47,48,49]. However, most of them were derived from research on specimens with a square section, or studies of RBAC confined by kinds of polymer material as FRP. Some models were derived from RBAC confined by circle steel tube, such as Chen [44], who noticed that the phenomenon of the constitutive model of confined RBAC has a close relationship with the RBA content. The unified stress–strain model for confined concrete proposed by Mander et al. [48] has been employed to simulate the vertical stress response of the infilled concrete for RACDSTs [49].

$${\sigma }_{c,v}=f_{cc}^{\prime }x\beta /(\beta -1+x^{\beta })$$

(2)

$$f_{cc}^{\prime }=1.08f_c(1+c_{\alpha }{c}_{f_y}{c}_{f_{cu}})$$

(3)

$$c_{\alpha }=6.625{\alpha }^2+5.96\alpha$$

(4)

$$c_{f_y}=1.034f_y/345-0.016$$

(5)

$$c_{f_{cu}}=-0.393f_{cu}/60+1.42$$

(6)

$$x={\epsilon }_{c,v}/{\epsilon }_{cc,b}$$

(7)

$${\epsilon }_{cc,b}={\epsilon }_{cc,r}(1+0.7r_{CCB})$$

(8)

$${\epsilon }_{cc,r}=(1+0.21r)[{\epsilon }_{cl}+10\alpha (26.9f_y-1885)(0.0063f_{cu}+0.622)]\times {10}^{-6}$$

(9)

$${\epsilon }_{c1}=(700+150\sqrt{f_{cu}})\times {10}^{-6}$$

(10)

$$\beta =E_{c,b}/(E_{c,b}-E_{\sec })$$

(11)

$$E_{c,b}=E_{c,r}(1-0.5r_{CCB})$$

(12)

$$E_{c,r}=(1-0.2r)\cdot {10}^5/(2.2+34.7/f_{cu})$$

(13)

$$E_{\sec }=f_{cc}^{\prime }/{\epsilon }_{cc,b}$$

(14)

In these formulas, the stress of specimen σ_c,v is related to the peak strength of confined concrete f_cc’, index x, and β. f_c is equal to 0.76; f_cu, c_α, c_fy, and c_fcu are coefficients that account for the influence of the steel-to-concrete area ratio, steel yielding strength, and concrete compressive resistance on the strength enhancement of core concrete. ε_c,v is the peak strain of confined concrete. ε_ccb denotes the peak strain of RBAC. r_CCB is the contained ratio of brick. E_cb and E_cr are the elastic moduli of RBAC and RAC. In this study, the content ratio of RAC was regarded as zero.

4.3. Constitutive Model of Double-Confined RAC

Recent research shows that columns made of concrete, confined by double tubes, have drawn increasing attention [45,50,51,52]. In the study by Rong [11], the confined pressure model of core concrete confined with an inner tube and outer tube were proposed as Equations (15)–(23). Equations (22) and (23) are obtained from Lam and Teng’s model [53].

$${\displaystyle \frac{f_c}{f_{co}}}=1+[1.065+{\displaystyle \frac{1}{2.7w^{2.7}}}]{\displaystyle \frac{f_{l1}}{f_{co}}}+i(1+3.19{\rho }_k^{0.5}{\rho }_{\epsilon }^{0.5})$$

(15)

$${\displaystyle \frac{{\epsilon }_c}{{\epsilon }_{co}}}=1+2.4{\rho }_k^{0.27}{\rho }_{\epsilon }^{1.03}$$

(16)

$$w=f_l/f_{cc}$$

(17)

$$f_{l1}=2f_y{t}_s/D_s$$

(18)

$$f_{cc}=f_{co}+(1.065+{\displaystyle \frac{1}{2.7w^{7.2}}})f_{l1}$$

(19)

$$i=9j$$

(20)

$$j=t_f/D_f+t_s/D_s$$

(21)

$${\rho }_k={\displaystyle \frac{2E_f{t}_f}{(f_{co}/{\epsilon }_{co})D}}$$

(22)

$${\rho }_{\epsilon }={\displaystyle \frac{{\epsilon }_l}{{\epsilon }_{co}}}$$

(23)

In this group of formulas (Equations (15)–(23)), f_l refers to the lateral confining pressure, w is the confinement ratio, t_s is the thickness of steel tube, D_s is the cylinder diameter, ρ_k and ρ_ε are the confining stiffness ratio and strain ratio, respectively, and ε_l is equal to 0.002. Equations (15)–(23) can be used in the double-steel-tube-confined core RBAC.

4.4. Proposed Constitutive Model of Double-Steel-Tube-Confined Core RBAC

On the basis of the research mentioned above, a new constitutive model of double-steel-tube-confined RBAC was proposed. The constitutive model of steel tube proposed by Mizan Ahmed [45] can be used in the new model. In addition, according to the conclusion of the research, the longitudinal compressive stress in the ring concrete in the CFDSTC was similar to that in the corresponding concrete in the CFST column. The ring concrete was confined mainly by the outer steel tube, and confinement provided by the inner steel tube can be ignored. Thus, the constitutive model of confined RBAC with a single steel tube, given by Equations (2)–(14), can be adopted in predicting the mechanical behavior of ring concrete in double-steel-tube-confined RBAC.

The constitutive model of core concrete is the most important part to DSTC. The core concrete is confined by both the outer and inner steel tubes. Based on the research of Chen et al. [44] and Rong [11], the constitutive model of RBAC confined by a double steel tube can be conducted as follows:

$$\left\{\begin{array}{ll}{\sigma }_{cc}=f_{cc}x\beta (\beta -1+x^{\beta })& 0<{\epsilon }_c\le {\epsilon }_{ccr1}\\ {\sigma }_{cc}=f_{cc1}+{\displaystyle \frac{f_{cc2}-f_{cc1}}{{\epsilon }_{ccr2}-{\epsilon }_{ccr1}}}({\epsilon }_c-{\epsilon }_{ccr1})& {\epsilon }_{ccr1}<{\epsilon }_c\le {\epsilon }_{ccr2}\end{array}\right.$$

(24)

$$f_{cc1}=k{f}_{co}(1+C_{\alpha }{C}_{f_y}{C}_{f_{cu}})$$

(25)

$$f_{cc2}=f_{cc1}+i{f}_{cc1}(1+C_{\alpha 2}{C}_{f_{y2}}{C}_{f_{cu2}})$$

(26)

$$i=\lambda (t_{si}/d_{si}+t_{so}/D_{so})$$

(27)

$${\epsilon }_{ccr1}=(1+0.21r)[{\epsilon }_{cl}+10\alpha (26.9f_y-1885)(0.0063f_{cu}+0.622)]\times {10}^{-6}$$

(28)

$${\epsilon }_{ccr2}=[1+2.4{\left({\displaystyle \frac{2E_s{t}_s{\epsilon }_{co}}{D{f}_{co}}}\right)}^{0.27}{\left({\displaystyle \frac{{\epsilon }_l}{{\epsilon }_{co}}}\right)}^{1.03}]{\epsilon }_{co}$$

(29)

$$\lambda =3.3138(t_0/D_0{)}^{0.9326}$$

(30)

In this group of formulas (Equations (24)–(30)), f_cc1 and f_cc2 denote the ultimate strength of single-confined concrete and double-confined concrete, respectively. ε_cc1 and ε_cc2 are the corresponding strain values. Parameters k and i stand for the strength reduction factor. This parameter can be determined through regression of the test results. First, Equation (25) is suitable for RBAC-filled steel tubes. According to the test result, parameter k can be modified as 0.957. As shown in Equation (21), parameter i is related to the confinement of the outer and inner tubes. Therefore, parameter i can be considered from the test results. In this study, i is different for the double-tube-confined RBAC, with a difference in the thickness-to-diameter ratio of the outer steel tube, as shown in Figure 10.

A comparison of the dual-confined core RBAC proposed model curve and the experimental curve of unconfined RBAC is shown in Figure 11. As can be seen from the figure, the elastic sections of the two curves match well. From this, it can also be concluded that the two have the same elastic modulus.

5. Model Validation of FEM

According to the theoretical research, the finite element study of double-steel-tube-confined RBAC was conducted using ABAQUS (V6.14-1) commercial finite element analysis software, aiming to validate the effectiveness of the constitutive model proposed above.

5.1. Model Establishment

The circular RBCDST column is composed of three parts: the outer and inner steel tubes, sandwiched concrete between the inner and outer steel tubes, and core concrete confined by a double tube, as shown in Figure 12. The core and sandwiched concretes were modeled using 3D solid continuum elements of brick, all with linear interpolation and normal Gauss integration. Eight-node reducing integration solid element (C3D8R) was adopted to simulate the concrete and end loading plate to avoid the locking effect during the calculation. The inner and outer steel tubes were simulated using the shell model of a four-node quadrilateral membrane with reduced integration (M3D4R).

5.2. Material Model

The finite element analysis was conducted by using the constitutive model of the corresponding material. In this study, five kinds of materials were simulated: core RBAC, ring section RBAC, core common concrete, ring common concrete, and steel. The behavior of the steel tube was simulated with the constitutive model described in Section 4.1. The relationship between strain and stress can be calculated with Equation (1). The simulation of ring RBA concrete behaviors was conducted using the constitutive model of confined RBAC outlined in Section 4.2. The behaviors of core RBA can be expressed as Equations (24)–(30) in Section 4.4. The constitutive relationship model of single- and double-confined common concrete can be found in previous research [45], as Equations (31)–(39). Equation (31) is derived from studies conducted by Mander et al. [48] and Lim and Ozbakkaloglu [54]. Equations (33) and (36) are derived from Lim and Ozbakkaloglu’s research [54] and modified by Ahmed et al. [55]. Equation (34) is derived from the literature [47]. Equations (38) and (39) are provided by Mander et al. [48], and Equation (39) is provided by De Nicolo et al. [56].

$$\left\{\begin{array}{l}{\sigma }_c={\displaystyle \frac{f_{cc}^{\prime }({\epsilon }_c/{\epsilon }_{cc}^{\prime })\lambda }{{({\epsilon }_c/{\epsilon }_{cc}^{\prime })}^{\lambda }+\lambda -1}}\\ {\sigma }_c=f_{cc}^{\prime }-{\displaystyle \frac{f_{cc}^{\prime }-f_{cr}}{[1+{\left(\frac{{\epsilon }_c-{\epsilon }_{cc}^{\prime }}{{\epsilon }_{ci}-{\epsilon }_{cc}^{\prime }}\right)}^{-2}]}}\end{array}\right. \begin{array}{l}0\le {\epsilon }_c<{\epsilon }_{cc}^{\prime }\\ {\epsilon }_c\ge {\epsilon }_{cc}^{\prime }\end{array}$$

(31)

$$\lambda ={\displaystyle \frac{{\epsilon }_{cc}^{\prime }{E}_c}{{\epsilon }_{cc}^{\prime }{E}_c-f_{cc}^{\prime }}}$$

(32)

$$E_c=4400\sqrt{{\gamma }_c{f}_c^{\prime }}$$

(33)

$${\gamma }_c=1.85D_c^{-0.135}$$

(34)

$${\displaystyle \frac{f_{cr}}{f_{cc}^{\prime }}}=1.242-0.0029\left({\displaystyle \frac{D}{t}}\right)-0.0044{\gamma }_c{f}_c^{\prime }$$

(35)

$${\epsilon }_{ci}=2.8{\epsilon }_{cc}^{\prime }{({\gamma }_c{f}_c^{\prime })}^{-0.12}\left({\displaystyle \frac{f_{cr}}{f_{cc}^{\prime }}}\right)+10{\epsilon }_{cc}^{\prime }{({\gamma }_c{f}_c^{\prime })}^{-0.47}\left(1-{\displaystyle \frac{f_{cr}}{f_{cc}^{\prime }}}\right)$$

(36)

$$f_{cc}^{\prime }=\left(1+{\displaystyle \frac{k_1{f}_{rp}}{{\gamma }_c{f}_c^{\prime }}}\right){\gamma }_c{f}_c^{\prime }$$

(37)

$${\epsilon }_{cc}^{\prime }={\epsilon }_c^{\prime }+{\displaystyle \frac{k_2{f}_{rp}{\epsilon }_c^{\prime }}{{\gamma }_c{f}_c^{\prime }}}$$

(38)

$${\epsilon }_c^{\prime }=0.00076+\sqrt{(0.626{\gamma }_c{f}_c^{\prime }-4.33)\times {10}^{-7}}$$

(39)

For common concrete serving as ring section concrete, the confined stress can be expressed as Equation (40), provided by the literature [57].

$${\displaystyle \frac{f_{rp}}{f_{sy}}}=\left\{\begin{array}{l}0.043646-0.000832\left({\displaystyle \frac{D_0}{t_0}}\right)\\ 0.006241-0.0000357\left({\displaystyle \frac{D_0}{t_0}}\right)\end{array}\right. \begin{array}{l}21.7\le {\displaystyle \frac{D_0}{t_0}}\le 47\\ 47<{\displaystyle \frac{D_0}{t_0}}\le 150\end{array}$$

(40)

For common concrete serving as the core concrete, the confined stress can be expressed as Equations (41) and (42), provided by [45].

$$f_{rpi}=2.2897+0.0066\left({\displaystyle \frac{D_0}{t_0}}\right)-0.1918\left({\displaystyle \frac{D_i}{t_i}}\right)-\left[0.0585\left({\displaystyle \frac{D_0}{t_0}}\right)-0.3801\left({\displaystyle \frac{D_i}{t_i}}\right)\right]{\zeta }^{-1}$$

(41)

$$\zeta ={\displaystyle \frac{A_{so}{f}_{sy,o}+A_{si}{f}_{sy,i}}{A_{sc}{\gamma }_c{f}_{sc}^{\prime }+A_{cc}{\gamma }_c{f}_{cc}^{\prime }}}$$

(42)

In this group of formulas (Equations (31)–(42)), λ is the index control, the initial slope, and the curvature of the ascending branch. f_cr represents the concrete residual strength, and ε_ci denotes the concrete strain corresponding to the inflection point. D_c is the concrete core diameter of the circular CFST column. D₀ and D_i represent the diameters of the outer and inner steel tubes. A_s0, A_si, A_sc, and A_cc are the cross-sectional areas of the outer tube, inner tube, ring concrete, and core concrete, respectively. f_cc′ and f_sc′ are the strengths of the core concrete and ring concrete. f_syi and f_sy0 are the yield stresses of the inner and outer tubes, respectively. The material used in finite element analysis is presented in

Table 7. Materials examined in FEM analysis.

Materials	Formulas
Steel tube	Equation (1)
Ring region RBAC	Equations (2)–(14)
Ring region normal concrete	Equations (40)–(42)
Core region RBAC	Equations (24)–(30)
Core region normal concrete	Equations (31)–(39)

. The stress–strain curves are shown in Figure 13a–e.

The steel model was implemented in ABAQUS using an elastic-plastic software interface. The elastic modulus and Poisson’s ratio were necessary to input. Poisson’s ratio was 0.3, and the elastic modulus was determined using the calculated value of code. In the plastic phase, plastic strain and plastic stress were inputted. The yield stress was derived from the measurements obtained in the test.

Various confined concretes were expressed by the plastic damage model of ABAQUS. In the elastic phase, the calculated elastic modulus was used, and the Poisson’s ratio was 0.2. In the plastic damage phase, plastic compressive strain and stress were input. The elastic stage was usually defined as 0.6 of the ultimate strength.

5.3. Loading and Boundary Condition

The loading conditions were applied using the boundary conditions of each cross-section. Two high-stiffness endplates were laid out at the end of the specimens. The bottom endplate was fixed against all degrees of freedom. On the other hand, the node in the center of the top endplate was fixed against all types of rotation and against lateral displacements. The top plate for all specimens was able to deform along the longitudinal (z) axis, upon which the compressive loads were to be applied. The vertical displacement was applied by specifying a reference point, as shown in Figure 14.

5.4. Interaction and Surface

The surface interaction of specimens, with six components, was complex. Considering the actual situation from the experiment, the core concrete and sandwiched concrete were tied to the top and bottom endplates, respectively. A steel endplate was chosen as the master surface. The concrete was also the slave surface during its interaction with both endplates. The core concrete suffered from double confinement, having little sliding displacement between the concrete and inner steel tube. In this study, a little slip was ignored during simulation, which is helpful for convergence analysis. However, the slippage between the concrete and the outer steel tube is significant, which cannot be ignored. In the simulation of a concrete outer steel tube surface, the term “normal behavior” refers to the pressure developed between the surfaces of each pair, whereas the term “tangential behavior” describes the extent of friction and the occurrence of slippage between the two surfaces, caused by high shear stresses. Herein, “hard contact” was selected to represent the normal behavior, while “rough” friction was used for the plate–concrete and plate–steel interactions and the “penalty” friction for the core–tube interaction with a coefficient of friction equal to 0.3.

5.5. Verification of Model

The numerical analyzed load–displacement curves are shown in Figure 15a–f. It is obvious that the ascent curves of most specimens have better agreement with the test curves; the specimen 4-0-4-0 is an exception. The elastic stiffness of the numerical curve has a larger difference to the tested one. This anomaly can be attributed to the insufficient consideration of the slippage between the concrete and the steel tube during the simulation, which led to the degradation of stiffness. However, for the other components containing RBA, the constitutive relationship has already taken into account the influence of the replacement rate of recycled aggregates on the material’s stiffness. Therefore, the ascent curves are relatively well matched. For all the specimens, the bearing capacity can be accurately predicted. On the other hand, most of the specimens have a poor match between FE results and test results in the after-elastic phase. The elastic constitutive adopted in the current finite element model is limited in its ability to accurately describe the nonlinear and inelastic behaviors of brick aggregates during brittle failure. Complex mechanical responses such as crack initiation and propagation cannot be effectively captured by the existing constitutive. In terms of material parameter characterization, parameters such as the microscopic damage evolution involved in the brittle failure of brick aggregates are difficult to accurately obtain and input into the model through conventional tests, resulting in limited reflection of such failures by the model.

However, for CFST, the descending branch dropped more quickly than those of the experimental curves, which means that the actual CFST specimen exhibited better ductility in the experiment. As shown in Figure 7, the curves for specimens 4-25-4-25 to 3-25-4-25 exhibit an early descent stage and secondary ascending section. The finite analysis struggles to simulate this effect because it ignores the crushing of brick aggregates during loading.

Figure 16a–d presents the failure mode of the numerical model, which is similar to those actual specimens in Figure 5. With the increase in axial compression, the specimens gradually exhibit eccentrical displacement, and the buckling of steel tube happened at the half to third column height location. The stress on the outer steel tube is larger than on the inner steel tube. Regardless of outer tube or inner tube, the strain distribution of the bottom is larger than on the top. The strain distribution of ring and core concrete has similar strain distributions. Generally speaking, it can be seen from the figures that the proposed model of the constitutive relation model combined with the finite element model can basically accurately predict the properties of retrofitted CFST columns.

5.6. Parametric Study

On the basis of the finite element analysis mentioned above, a parametric study of the layout of the original and additional steel tube was conducted. The influence of the diameter-to-thickness ratio of the inner and outer tubes and the influence of the diameter ratio of the inner and outer tubes were studied.

5.6.1. Influence of Diameter-to-Thickness Ratio of Outer Tube (D₀/t₀)

Figure 17 presents the load–displacement curve differences between specimens with different diameter-to-thickness ratios of outer tubes, when other conditions changing were not considered. It can be seen from the figure that the bearing capacity increased obviously with the decreasing D₀/t₀. The average decreasing ratio is 13.2%, and the ductility of specimens is nearly the same. For the initial stiffness, the specimen with the lower diameter-to-thickness ratio has a larger elastic modulus.

5.6.2. Influence of Diameter-to-Thickness Ratio of Inner Tube (D_i/t_i)

The load–displacement curves of specimens with different diameter-to-thickness ratios of inner tubes are shown in Figure 18. It is obvious that, different from the regular results presented in Figure 17, the influence of D_i/t_i on the bearing capacity is much lower than for the outer steel tube. With the parameter of D_i/t_i changing from 40 to 24, the bearing capacity also increases gradually. The average decreasing ratio of the bearing capacity is 4.1%. It can be concluded that the diameter-to-thickness ratio of the outer tube plays a more important role in the bearing capacity of double-confined recycled brick aggregated concrete.

5.6.3. Influence of Diameter Ratio of Inner and Outer Tube (D_i/D₀)

Finally, the influence of the diameter ratio of the inner and outer tubes on the bearing capacity of specimens was studied. As shown in Figure 19, the load–displacement curves all present an ascending trend. The bearing capacity increased with an increase in D_i/D₀ value. In other words, although a small diameter is good for improving the confined effect, a larger confined area of inner steel tube is also beneficial for bearing load.

In this study, the number of test specimens and situations was relatively small. The applicable situation is limited. First, the replacement ratio is relatively simple. Second, the diameter ratio of the outer and inner steel tubes is fixed. More situations will be researched in future work.

6. Conclusions

Based on the research presented above, the conclusions that can be drawn are as follows:

• The recycled-brick-aggregate-concrete-filled single-steel-tube column exhibits a clear shear failure mode under axial load. Double-steel-tube confinement can effectively change the failure mode. The buckling is only concentrated on the outer steel tube, approximately located within the range of a half to a third of the specimen height, whereas the inner steel tube has a relatively low rate of deformation.
• The ultimate strength of specimens is related to the replaced ratio of brick aggregate content, diameter-to-thickness ratio of the outer and inner steel tube, and outer and inner tube diameter ratio. Under the same double steel tube condition, if 25% brick aggregate recycled concrete is used only in the ring area of the section, the ultimate strength will be reduced by 9.1%. If 25% brick aggregate recycled concrete is adopted in the whole section, the ultimate strength will be reduced by 17.4%. Under the same concrete condition, the ultimate strength of the specimen improved with increasing of outer and inner steel tube diameters. For each 1 mm increase in the thickness of the outer tube, the maximum increase in the ultimate strength of the component can reach 22.6%. For each 1 mm increase in the thickness of the inner tube, the maximum increase in the ultimate strength of the component is just 4.1%.
• The double-steel-tube-confined recycled brick aggregate concrete column has two kinds of stress–strain curves. One is “full and convex”, while the other is “concave” with a secondary ascending phase. The shape of these curves reflects the suitability of the constraint combination. Excessive constraints may result in premature cracking of brick aggregates. Specimens 4-25-4-0 and 3-25-4-25 are relatively suitable combinations.
• A constitutive relationship of double-confined brick recycled aggregate concrete was proposed based on the existing literature, which can be used to express the mechanical behavior of inner core concrete in a double steel tube column.
• Finite element analysis shows that, as the diameter-to-thickness ratio of the inner and outer steel tubes decreases, the specimens’ ultimate strength shows an increasing trend. However, relatively speaking, the influence of the inner steel tube is not as obvious as that of the outer steel tube. The closer the diameters of the inner and outer steel tubes, the higher the ultimate strength obtained. Therefore, more attention should be paid to the restraint effect of the outer steel pipe during reinforcement.