Mesostructural fracture modeling of asphalt mixtures: Stochastic heterogeneity coupled with bilinear cohesive elements

Under the action of traffic loads, asphalt mixtures will crack to different degrees, which directly affects the serviceability and durability of pavements. While cohesive zone models (CZM) have been widely used to simulate fracture in asphalt materials, existing approaches often treat the mixture as homogeneous or employ simplified aggregate representations, limiting their ability to capture realistic crack paths. This study presents a novel mesoscale fracture modeling approach that explicitly couples a stochastic heterogeneous aggregate structure with a bilinear Cohesive Zone Model (CZM) to simulate crack initiation and propagation. The model captures the random shape, size, and spatial distribution of aggregates by converting 3D volumetric gradation to 2D probabilistic distributions. The embedded bilinear CZM elements at both the aggregate-mortar interface and within the mortar matrix enable simultaneous simulation of interfacial debonding and matrix cracking. The key novelty lies in the systematic integration of statistically representative aggregate heterogeneity with CZM, which produces aggregate-deflected crack paths that closely match experimental observations—an improvement over homogeneous models that predict unrealistic straight-line cracking. The proposed framework provides a more realistic and efficient tool for analyzing mesoscale fracture mechanisms in asphalt mixtures, overcoming limitations of homogeneous models and purely 2D simulations. Limitations include the current 2D simplification and validation against a single experimental dataset, which will be addressed in future work through 3D model extension and multi-condition experimental validation.