1. Introduction
In recent years, the rapid development of infrastructure construction in coastal cities and island areas, coupled with the acceleration of urbanization, has led to an increasing demand for concrete [
1,
2]. However, the production process of concrete requires the consumption of a significant amount of natural resources, such as fresh water and river sand [
3,
4,
5]. The excessive exploitation and use of these resources have resulted in a severe shortage and significant damage to the natural environment. Moreover, the transportation cost of traditional building materials is higher in construction projects located in coastal and island areas. In response to these challenges, researchers have proposed the use of seawater and sea sand as alternatives to freshwater and river sand in the production of concrete structures [
6,
7,
8]. Sea sand has the advantages of large resource reserves and wide distribution, and the physical properties are similar to river sand, but the grading is slightly different. Existing research indicates that the mechanical properties of seawater sea-sand concrete (SSSC) are comparable to those of ordinary concrete [
9,
10]. Consequently, SSSC presents not only substantial economic benefits but also satisfactory working performance, making it highly promising for widespread application. However, the high concentration of chloride ions in seawater and sea sand introduces a significant challenge, as it accelerates the corrosion of steel reinforcement, leading to potentially severe safety hazards. Therefore, there are difficulties in popularizing the application of SSSC.
In order to promote the application of SSSC and address the issue of steel bar corrosion, researchers have proposed the utilization of fiber-reinforced polymer bar (FRP bar) as a replacement for steel bars, in combination with SSSC in building structures [
11,
12,
13,
14,
15]. With advancements in technology and process refinement, the use of FRP bars has become more streamlined, cost-effective, and widespread in building and bridge structures. This has garnered increasing attention from scholars, who have begun to focus on the research regarding the application of FRP reinforcement [
16,
17,
18,
19,
20]. FRP bars are used in concrete as reinforcing steel bars, which have the advantages of chloride ion corrosion resistance, high tensile strength, lightweight properties, and fatigue resistance [
21,
22,
23,
24]. Therefore, the adoption of FRP bars instead of traditional steel bars plays an important role in resolving the issue of chloride corrosion in seawater sea sand. However, FRP bars tend to have low stiffness in concrete structures, large structural deflection in the use phase, easy to brittle damage, as well as high price and other shortcomings, making it difficult to be widely promoted in engineering applications. To address these challenges, steel–FRP composite bars (SFCBs) are a new type of reinforced composite material with steel reinforcement as the core and fiber layer as the outer layer. SFCB combines the favorable characteristics of both steel bars and FRP bars, including a high modulus of elasticity, stable secondary stiffness, excellent corrosion resistance, and a lower price point [
25,
26]. The application of SFCB in SSSC offers a theoretical solution to the problems associated with steel reinforcement erosion caused by chloride ions and the lack of ductility observed in FRP bars.
The substitution of SFCB for steel addresses two key problems, steel corrosion and the lack of ductility in FRP bars. A large number of previous studies have demonstrated the favorable bonding properties between FRP bars and concrete, resulting in a synergistic effect that enhances the integrity and safety of FRP-reinforced concrete [
27,
28,
29]. However, it is imperative to investigate whether SFCB exhibits similar bonding properties with concrete and creates a beneficial synergy. Moreover, building structures must not only perform well in comfortable environments but also withstand the impact of aggressive conditions such as freezing and thawing, high temperatures, seawater erosion, and wet–dry cycles. Scholars have already conducted relevant studies on the bonding performance between SFCB and concrete [
25,
30]. However, these studies have mostly focused on conventional environments, wet–dry cycles, and seawater immersion, with fewer investigations conducted on bond performance during freezing–thawing cycles, which present extreme conditions [
31,
32,
33]. Freezing–thawing environments have diverse and multifaceted effects on building structures. They can induce spalling, expansion cracking of concrete, interface cracking between different materials, accelerated carbonation of concrete, and increased sulfate corrosion, all of which contribute to a decline in structural performance [
34,
35,
36]. Therefore, ensuring the durability of the interfacial bond is crucial for maintaining the overall integrity of a building structure over a prolonged period under freezing–thawing conditions. Investigating the interfacial bonding properties between SFCB and SSSC under freezing–thawing cycles is imperative for the successful application of SFCB in such environments.
A substantial body of research has explored the durability of concrete in freezing–thawing environments, with scholars proposing various theories regarding the deterioration mechanism of concrete under such conditions [
37,
38,
39]. Freezing–thawing damage is commonly divided into two stages. The initial stage involves the solidification and expansion of internal liquid water, leading to crack formation and characterized by a loss of dynamic modulus and concrete strength. The subsequent stage involves the disintegration of the cement paste body, resulting in the shedding of surface mortar. This stage is characterized by mass loss and partial or complete destruction of the concrete structure. Considering this two-stage perspective, improving the cracking and permeability resistance of concrete can enhance its ability to withstand freezing–thawing losses, consequently improving the durability of the bond between reinforcement and concrete in freezing–thawing environments. To enhance the frost resistance of concrete, glass fiber plays a crucial role by limiting the extension of vertical cracks and promoting the bonding of parallel cracks. This optimization of the internal structure of concrete improves its crack resistance [
40]. The incorporation of an expansion agent in concrete contributes to the formation of expansion mineral crystals through a reaction with Ca(OH)
2, a product of cement hydration. This reaction slows down the cement hydration rate, generating expansion stress within the concrete. This stress helps resist part of the tensile stress resulting from drying contraction, inhibiting concrete deformation, and reducing cracks. Furthermore, an appropriate amount of expansion agent can enhance the seepage resistance of concrete. According to our previous studies [
41,
42], there is evidence of a synergistic effect when combining an appropriate amount of expansion agent with glass fiber. The concrete mixed with glass fiber and expansion agent can not only inhibit crack development but also improve the concrete compressive strength, flexural strength and toughness, and other aspects of performance [
8,
41,
42,
43]. Therefore, in this study, the incorporation of a suitable amount of glass fiber and expansion agent aims to enhance the crack resistance and seepage resistance of SSSC, ultimately improving its ability to resist freezing and thawing losses.
To ensure the frost resistance of building structures, it is crucial to examine the durability of the bond between reinforcement and concrete in freezing–thawing environments. Previous studies have focused on the bond durability between FRP bars and concrete, primarily through axial pullout tests [
25,
33]. However, these tests were conducted with a concrete cover thickness greater than 75 mm and a ratio of concrete cover thickness c to FRP reinforcement diameter d greater than 5, which does not align with the recommendation in ACI 440.1R-2015 stating that c/d should not exceed 3.5 [
44]. Consequently, the use of axial pullout tests may result in specimens with an inconsistent concrete protective layer thickness, leading to an overestimation of the bond durability between stressed reinforcement and concrete under freezing–thawing conditions. To address this issue, an eccentric pullout test is employed in this study to investigate the bond performance of the SFCB and SSSC interface under varying concrete protective layer thicknesses and freezing–thawing cycles. This approach allows for the consideration of the impact of concrete protective layer thickness on the frost durability of the bonded interface, while also being more representative of the actual engineering situation.
This study addresses the issues of chloride corrosion in traditional steel bars submerged in seawater sea-sand concrete and the limited ductility of FRP bars by introducing a novel composite steel bar known as SFCB. To enhance the crack resistance and seepage resistance of SSSC, an appropriate amount of glass fiber and expansion agent are incorporated. Consequently, the frost resistance of the concrete is improved. To assess the impact of the concrete protective layer thickness on the frost durability of the bonded interface, eccentric pullout experiments were conducted to replicate the actual engineering situation. Furthermore, the bonding performance of SFCB and GF-EA-SSSC interfaces was examined under varying concrete protective layer thicknesses and a number of freezing–thawing cycles. These investigations aim to promote the advancement and widespread application of seawater sea-sand concrete.
4. Calibration of Constitutive Models
The bond stress–slip principal model, as the key to bond research, offers a comprehensive depiction of the bond interaction between reinforcement and concrete. Currently, a large number of models reflecting the bond–slip behavior of reinforced concrete have been developed, leading scholars to establish FRP bar-concrete stress–slip constitutive models, primarily the BPE model and CMR model [
60,
61,
62]. However, the ontological model for the stress–slip relationship between SFCB and concrete remains imperfect. Notably, existing models do not account for the effects of the freezing–thawing environment and the thickness of the concrete protective layer on bond performance. Therefore, it is necessary to extend the bond stress–slip ontological model of SFCB with concrete to more accurately depict the bonding performance between SFCB and SSSC. Given the enhanced frost durability of SSSC with the incorporation of GF and EA, the effects of the number of freezing–thawing cycles and the thickness of the concrete protective layer are also considered. In this study, the commonly used modified Bertero–Eligehausen–Popov (mBPE) model and Cosenza–Manfredi–Realfonzo (CMR) model are used as references to establish the bond stress–slip ontological model about SFCB and GF-EA-SSSC.
Among them, the bond stress–slip relationship of the mBPE model can be expressed as follows:
where
a and
p are empirical parameters, and
τr and
sr are the bond stress and slip at the intersection of the descending and residual stages, respectively.
However, the mBPE model employs a horizontal straight line to approximate the residual segment curve, resulting in an inadequate prediction of the residual stage. In this study, only the ascending and descending segments of the mBPE model were fitted, disregarding the residual stage.
On the other hand, the CMR model primarily focuses on describing the ascending phase of the curve and does not fully capture the entire bond stress–slip curve. The bond stress–slip relationship in the CMR model is expressed as follows:
where
α and
β are empirical parameters.
In comparison, the CMR model exhibits a better fit for the ascending phase than the mBPE model. The fitting results of the improved intrinsic model match well with the experimental data, as shown in
Figure 23. The specific empirical parameter values are given in
Table 6.
5. Conclusions
In this study, 13 concrete mix proportions were devised, and 39 specimens of SFCB embedded in SSSC underwent eccentric pullout tests. The investigation focused on assessing the bond durability of SFCB and GF-EA-SSSC under freezing–thawing conditions. The study considered five sets of freezing–thawing cycles (T) and four varying thicknesses of the concrete protective layer (c) as the independent variables. The primary conclusions drawn from this research are as follows:
(1) The SSSC undergoes a loosening and brittleness transformation under freezing–thawing cycles. Incorporating glass fiber and expansion agents can enhance the ability of concrete to resist freezing–thawing damage. Initially, cracks emerge from the weakest point of the concrete. As the number of freezing–thawing cycles increases, damage accumulates from the exterior to the interior, leading to a gradual escalation in both crack size and quantity. Upon reaching 150–200 freezing–thawing cycles, significant spalling occurs on the surface of the specimen, with an expanding spalling area. The damage pattern shifts from tensile damage to diagonal cut damage as freezing–thawing cycles progress.
(2) As the number of freezing–thawing cycles increases, there is a concurrent decrease in compressive strength and an increase in mass loss. In the initial stages of the freezing–thawing cycle, the decrease in compressive strength is gradual, and mass loss is minimal. Microcracks have a negligible impact on compressive strength, and the specimen retains its structural integrity. After surpassing 100 freezing–thawing cycles, concrete damage accumulates significantly. The internal structure transitions from dense to loose, leading to extensive spalling from the exterior to the interior. Compressive strength exhibits a noticeable decline, accompanied by an increase in quality loss. By the time 200 freezing–thawing cycles are reached, the loss of compressive strength amounts to 55.05%, and the quality loss rate reaches 6.36%.
(3) Analyzing the stress–strain curve reveals that freezing–thawing action not only impacts the compressive strength of concrete but also influences the characteristics of the concrete damage stage. With an escalation in the number of freezing–thawing cycles, compressive strength diminishes, the peak strain initially rises and then decreases, and the stress–strain curve sharply descends after reaching the peak point. This phenomenon indicates an increase in brittleness and a decrease in ductility for SSSC. The freezing–thawing cycles induce heterogeneous damage within the concrete interior. Stress concentration during the damage stage results in the manifestation of brittle damage.
(4) The dynamic elastic modulus experiences a decline as the number of freezing–thawing cycles increases. In the initial phase of the freezing–thawing cycle, the decrease in dynamic elastic modulus is gradual. However, as the number of freezing–thawing cycles advances, the impact of freezing–thawing damage intensifies, progressing from the outer layer to the inner layers of concrete. Consequently, the development of cracks and pores within the concrete becomes increasingly profound, leading to an accelerated rate of decrease in dynamic elastic modulus.
(5) Following freezing–thawing cycles, the majority of specimens subjected to the eccentric pullout test exhibited rebar pullout damage. The bond performance at the bonded interface of the pulled specimens diminished, and there was a continuous decrease in the quantity of residual fibers on the concrete interface, along with wear on the SFCB interface. Maintaining a constant number of freezing–thawing cycles, an increase in the thickness of the concrete protective layer altered the damage mode of entangled fibers. The shift occurred from fracture to spalling and subsequently from spalling to abrasion. Notably, residual fibers remained at the interface, and the SSSC has a better restraining effect on the SFCB. With a fixed concrete protective layer thickness, an escalation in the number of freezing–thawing cycles resulted in a gradual reduction in bond interface damage. The damage at the concrete interface transitioned from bond–slip damage to freezing–thawing damage.
(6) The bond stress–slip curves for specimens exhibiting pullout damage vary with different thicknesses of the concrete protective layer and the number of freezing–thawing cycles. In the initial loading stage, the bond stress between SFCB and SSSC demonstrates elastic characteristics. Once the curve reaches its peak stress value, it transitions into the descending and residual sections. The bond stress–slip curve exhibits multiple peaks before and after freezing and thawing. As the thickness of the concrete protective layer increases, the peak value of the curve gradually rises, indicating enhanced bond stress between SFCB and SSSC. Conversely, with an increase in the number of freezing–thawing cycles, the peak of the curve decreases, signifying a reduction in the bond stress between SFCB and SSSC. The second peak is generally much smaller than the first peak, indicating that the pullout damage is ductile in nature.
(7) Under various thicknesses of the concrete protective layer and differing numbers of freezing–thawing cycles, the trends observed in bond strength, relative slip, and bond stiffness exhibited similarities. With a constant concrete protective layer thickness, an increase in the number of freezing–thawing cycles resulted in a decrease in both bond strength and bond stiffness, accompanied by an increase in relative slip. After 200 freezing–thawing cycles, the bond strength decreased by 42.31% relative to the unfrozen-thawed condition, and the loss of bond stiffness could reach 45.39%. The restraining effect on SFCB pullout was found to improve with a thicker concrete protective layer. Additionally, the rate of loss in bond strength and bond stiffness reduced as the protective layer thickness increased. Formulas for bond strength, relative slip, and bond stiffness were established, taking into account the effects of the concrete protective layer thickness and the number of freezing–thawing cycles in the eccentric pullout test. The relative error between theoretical and experimental values was within 5%.
(8) In this study, GF and EA were introduced to enhance the frost durability of SSSC, and the mBPE model and CMR model were refined to account for the influence of the number of freezing and thawing cycles as well as the thickness of the concrete protective layer. These modifications aimed to establish a bond stress–slip constitutive model for the interaction between SFCB and the composite material of GF-EA-SSSC. The fitting results of the improved constitutive model closely resembled the experimental values, demonstrating a high level of accuracy in the predictive capabilities of the model.
(9) This paper investigates the bonding properties between steel–FRP composite bar (SFCB) and glass fiber with expansion-agent-reinforced seawater sea-sand concrete (GF-EA-SSSC) interface using eccentric pullout experiments under different thicknesses of concrete protective cover and a number of freezing–thawing cycles. The eccentric pullout test in this study utilized a single SFCB, without considering the impact of spacing multiple SFCBs on the bond strength of GF-EA-SSSC after freezing–thawing cycles. This divergence from practical engineering standards suggests the potential for refining the test methodology. Following 200 freezing–thawing cycles, the specimens underwent nearly complete freezing, an occurrence rarely observed in practical engineering scenarios. Future research could delve into variations in dynamic elastic modulus and ultrasonic testing post freezing–thawing damage to specific cross-section areas, and the influence of the bonding performance of the interface between steel and concrete, so as to propose a more reliable evaluation method for reinforced concrete members after freezing–thawing damage in practical engineering.