Surface Mesh Generation of Large-scale Digital Rock Images in 3D

Abstract

A surface mesh generation method for describing and applying the large-scale 3D digital rock images is presented in this paper. The accurate description of rock structure needs high resolution digital images involving a huge number of voxels, which make it difficult and even impossible to directly use in conventional numerical analysis (e.g. FEM and FVM). In addition, as regards to processing such large-scale original data itself, computer capability and efficiency are also quite challenging. We proposed a meshing method based on a simplified centroidal Voronoi tessellation (CVT) to efficiently describe the rock structures with acceptable element quantity and quality and apply for further simulation and processing. The method is implemented in the following steps: (1) extracting material interfaces in rock images; (2) creating a background mesh with size field on the surfaces; (3) generating nodes with a simplified CVT method; (4) constructing and optimising triangular meshes with the above nodes. A one-billion-voxel rock image is utilized to demonstrate the effectiveness of the proposed method.