Prediction on dynamic modulus of asphalt concrete with random aggregate modeling methods and virtual physics engine

Asphalt mixture usually needs compaction prior to servicing in field. Traditionally, finite element methods cannot simulate the compaction of asphalt mixture in field. Although discrete element methods (DEM) are able to describe such a process, the intrusion among aggregates always exists and influences the calculation accuracy. Additionally, the central difference method used in DEM requires very fine elements, which significantly jeopardizes the calculation efficiency. This study proposed an innovative systematic method to generate virtual asphalt mixture based on random aggregate modeling method and virtual physics engine theory. The aggregates were virtually generated with random aggregate modeling methods and imported into the virtual physics engine, where the movements of aggregates in asphalt concrete during compaction was simulated. The coordinates of aggregates after compaction were extracted, which were incorporated into the Convex Hull Algorithm to determine the new locations of aggregates after compaction. The original contribution of this study is that a pre-processing method of generating virtual asphalt concrete was proposed, which can successfully introduce compaction of asphalt mixture into finite element models and solve the intrusion problem of aggregates. The dynamic modulus of asphalt concrete simulated in finite element method (FEM) based on this modeling method was found much closer to the experimental results, comparing to the traditional analytic methods. It indicates the proposed method valid in generating virtual asphalt concrete in FEM. (C) 2016 Elsevier Ltd. All rights reserved.