Automated real aggregate modelling approach in discrete element method based on X-ray computed tomography images

Abstract

Numerical representations of aggregate particles are critical to reconstruct microstructure of asphalt mixtures. The main objective of this study is to obtain an accurate and automated discrete element method (DEM) with a small number of sphere elements for any real aggregate particle. An automated DEM method was developed based on X-ray computed tomography (CT) images with aid of computer graphics and image processing techniques. The method consists of three main steps. First, or each image layer, two-dimensional (2D) outlines of aggregate particles in cropped image blocks are detected. Second, the mapping between an aggregate particle and its outlines on different layers is established. Based on the aforementioned mapping, the three-dimensional (3D) geometric model of each particle is constructed. Third, the 3D geometric model of the aggregate particle is automatically transformed into its discrete element representation. For each aggregate particle, the fitting accuracy between the 3D geometric model and discrete element representation can be conveniently adjusted. The obtained clumps of spheres can be virtually assembled into the discrete element microstructure model of asphalt specimen and used in numerical simulation. Based on four aggregate particle templates, a single-edge notched beam model was reconstructed and the numerical fracture simulation was conducted. The simulation results indicated that the proposed modelling method is applicable to reconstruct asphalt mixture models and conduct performance test simulation.