Concrete cracking is a critical and pervasive issue that not only compromises the structural integrity, load-bearing capacity, and aesthetic appearance of engineering projects but also provides direct pathways for harmful environmental ions, such as chlorides and sulfates, to penetrate. This process significantly accelerates the corrosion of internal steel reinforcement, ultimately reducing the overall service life and safety of concrete structures. Epoxy resin grouting materials have been widely adopted for repairing concrete cracks due to their exceptional interfacial bonding strength, superior mechanical properties, and favorable flow characteristics. However, in high-altitude and cold regions, such as Northwest China, repaired water conservancy and transportation infrastructure is frequently subjected to severe freeze-thaw (F-T) cycles. The repeated physical expansion and contraction of free water trapped within the structural micro-pores and crack interfaces generate immense internal frost heaving stresses. This cyclic stress may lead to accelerated material aging, adhesive debonding, or secondary cracking of the epoxy resin repair systems. Therefore, deeply understanding the degradation law and accurately predicting the service life of repaired concrete under long-term F-T environments are essential for ensuring engineering safety and optimizing maintenance strategies.

To systematically investigate the effect of different epoxy resin types on the freeze-thaw resistance of repaired concrete, this study selected three representative grouting materials: elastic epoxy resin (IIT), moisture-tolerant epoxy resin (IIC), and high-permeability epoxy grout (IVS). Additionally, an unrepaired intact concrete group (UNR) was prepared as a control reference to evaluate the relative repair effectiveness. Standard C30 concrete prismatic specimens, with dimensions of 100 mm × 100 mm × 400 mm, were cast and cured under standard conditions. The specimens were then artificially cracked to simulate realistic structural damage, followed by precise crack injection using the selected epoxy resins. Subsequently, the prepared specimens were subjected to a rapid F-T testing regime in accordance with standard testing protocols. Each F-T cycle lasted for 4 hours, with the core temperature precisely fluctuating between −17±2 ℃ and 8±2 ℃. Throughout the entire 300 cycles of testing, macroscopic failure morphology, mass loss rate, and relative dynamic elastic modulus (RDEM) were periodically measured and recorded every 50 cycles to comprehensively quantify the macroscopic and microscopic degradation processes.

The experimental results indicated distinct performance differences among the different repair materials under the combined action of low temperature and moisture. During the initial F-T stage (fewer than 100 cycles), the IIT group exhibited the most severe degradation trend, with a mass loss reaching 6.72% and the relative damage factor increasing sharply to 9.09%. Notably, these deterioration indicators were even higher than those observed in the unrepaired UNR group, suggesting a negative effect under specific initial conditions. In sharp contrast, the IIC and IVS groups demonstrated relatively stable performance with limited mass loss and damage accumulation. As the cycles progressed, macroscopic observations confirmed that while severe surface mortar spalling and coarse aggregate exposure occurred primarily in the concrete matrix, the epoxy-concrete bonding interfaces remained largely intact without obvious debonding, indicating strong interfacial adhesion for all selected resins.

To quantitatively predict the service life of the repaired structures, a continuous damage mechanics model was established based on the evolution law of the relative dynamic elastic modulus. Defining the structural failure threshold as the critical point at which the RDEM drops to 60% of its initial value, the predicted maximum F-T cycles for the UNR, IIT, IIC, and IVS groups were calculated as 413, 403, 429, and 456 cycles, respectively. Furthermore, a two-parameter Weibull probability distribution model was introduced to cross-validate the structural life prediction based on the experimental damage factors. The updated linear regression fitting results yielded shape parameters (k) for the UNR, IIT, IIC, and IVS groups of 3.03, 3.78, 1.82, and 1.28, respectively, reflecting the different acceleration rates of damage evolution. The scale parameters (λ), representing the characteristic service life at a 63.2% failure probability, were mathematically determined to be 580, 110, 710, and 1464 cycles, respectively. This rigorous statistical verification showed strong consistency with the deterministic damage mechanics model.

Comprehensive micro-mechanistic analysis revealed that the relatively poor F-T resistance of the IIT resin is primarily attributed to the addition of modified bisphenol A. Although this component imparts high static strength, it results in poor low-temperature fluidity at low temperatures, leading to significant air bubble entrapment and subsequent massive internal ice-expansion stresses. In contrast, the IVS resin, chemically modified with active diluents, exhibits superior capillary permeability, allowing it to deeply penetrate and effectively seal micro-defects at the crack tips, thereby blocking water ingress pathways. In conclusion, both the mechanics-based and probability-based models confirm that the ultimate order of freeze-thaw resistance from highest to lowest is as follows: IVS, IIC, UNR group, and finally the IIT group. These findings provide highly valuable theoretical guidance and practical reference for the targeted selection of optimal repair materials in cold-region engineering applications.