摘要: Factors affecting the resistance of asphalt mixtures against low-temperature cracking are divided into three categories: pavement structure geometry, material properties, and environmental conditions. Most of the existing models for evaluating and predicting the resistance of asphalt mixtures against low-temperature cracking have focused more on environmental factors and pavement structure geometry. However, the impact of material properties has been neglected. The concepts of surface free energy (SFE) that focus on material properties have a high potential to assess different failures of asphalt mixtures. Accordingly, in this research, the effect of SFE parameters, along with mix design parameters and bitumen rheological properties, were studied on the resistance of asphalt mixture against low-temperature cracking using an artificial neural network (ANN) model. Two indicators of fracture energy and fracture toughness were considered as the criteria for the resistance of asphalt mixtures against low-temperature cracking which is measured by by the semi-circular bending (SCB) test. The sensitivity analysis results of the models indicated a positive correlation between the resistance of asphalt mixtures against low-temperature cracking and independent variables, including the apparent film thickness (AFT) on the surface of the aggregate, the index of aggregate (Ia), adhesion free energy (AFE), the ratio of the base to acid component of the aggregates (Base to Acid), cohesion free energy (CFE), the specific surface area (SSA) of the aggregates, and the stress relaxation rate (m-value). Among the variables studied, bitumen creep stiffness had a reverse relationship with the resistance of asphalt mixtures against low-temperature cracking. In addition, the sensitivity analysis results based on normalized data illustrated that the AFT and m-value were the most significant parameters affecting the asphalt mixtures' resistance against low-temperature cracking.