Characterizing asphalt mixtures with random aggregate gradations based on the three-dimensional locally homogeneous model

Abstract

A three-dimensional (3-D) locally homogeneous model for asphalt mixtures is presented in this study, which can effectively balance the simulating accuracy and computational effort. Moreover, the locally homogeneous model accounts for various aggregate gradations to allow for the mesoscale analysis of the asphalt concrete (AC) under loading. In the first step of the locally homogeneous model development, weighted Voronoi tessellation technology was employed to divide the volume of asphalt mixture into several local parts in terms of the aggregate gradations and random particle locations with each part containing one or a few aggregates particles. Subsequently, the mesomechanical Mori–Tanaka method was used to calculate the homogeneous mechanical properties for each local space. Finally, 3-D models of asphalt mixtures were developed, which were further used for numerical simulations. The finite element method was used for the asphalt mixture mesoscale investigations. Two asphalt mixtures were studied consisting of dense- and gap-graded aggregates and labeled, respectively, AC-13 and stone mastic asphalt (SMA)-11. The efficacy of the numerical analysis was demonstrated by comparing model-generated responses and performance with compressive dynamic loadings test results. The models demonstrated the ability to characterize the linear viscoelastic performance for both asphalt mixtures. Furthermore, the locally homogeneous models of AC-13 and SMA-11 mixtures were coupled into global asphalt pavements to identify the influence of aggregate particles on their mechanical responses. The results indicated that the locally homogeneous models are able to consider both the pavement structure at the global scale and the internal feature of asphalt mixtures at the local scale.