Study on the shear strength of asphalt mixture by discrete element modeling with coarse aggregate morphology

High shear strength is crucial for asphalt pavements to resist permanent deformation at high temperatures, which is largely determined by the skeletal structure of coarse aggregates. In this study, four size ranges of polyhedral coarse aggregate clumps were built to simulate the morphology and dimensional characteristics of coarse aggregate. Subsequently, virtual triaxial shear tests were simulated using discrete element models. Three gradation types were analyzed to study the effects of aggregate structure, granularity composition, and coarse aggregate modulus on the shear strength of asphalt mixtures. The results show that it is feasible to obtain accurate morphological features of coarse aggregates by controlling the minimum diameter of the filled spheres in a polyhedral coarse aggregate box when constructing a discrete element model. The virtual triaxial shear test model developed in this study can be effectively utilized to simulate the shear mechanical properties of asphalt mixtures under various confining pressures. The numerical simulation results indicate that the cohesion of asphalt mixture AC-13 is the best, while the internal friction angle and shear strength of SMA-13 are the highest. The ranking of factors that affect the cohesion and internal friction can be summarized as: aggregate structure ≫ coarse aggregate modulus > granularity composition, while the order of factors that affect the shear strength is: aggregate structure ≫ granularity composition > coarse aggregate modulus. This indicates that the aggregate structure is crucial for the high-temperature performance of asphalt mixtures. This study addressed the issue of high variability in laboratory test results for macroscopic properties of asphalt mixtures and analyzed the skeleton role of coarse aggregate in asphalt mixtures from a mesoscopic perspective. The findings parameters for explaining the strength mechanism. The discrete element models also facilitate the optimization of asphalt mix design with reduced labor and time consumption on laboratory testing.