En
期刊文章
文章精选
当期目录
过刊浏览
高被引
高下载
特邀文章
专刊合集
待刊文章
Volume 118
您当前的位置：首页 > 期刊文章 > 当期目录 > Volume 118
A constructive sphere packing method for DEM using centroidal Voronoi tessellation and maximum inscribed spheres
Hui Yuan, Yongzhi Zhao *
Institute of Process Equipment, College of Energy Engineering, Zhejiang University, Hangzhou, 310027, China
10.1016/j.partic.2026.09.007
Volume 118, November 2026, Pages 364-376
Institute of Process Equipment, College of Energy Engineering, Zhejiang University, Hangzhou, 310027, China
E-mail: yzzhao@zju.edu.cn
Highlights

• A constructive sphere packing method combines CVT and maximum inscribed spheres.

• Nonoverlapping packings are generated in complex and concave 3D domains.

• CVT iterations govern particle size distribution and packing fraction.

• Computational time increases approximately linearly with particle number.


Abstract

Discrete element method (DEM) simulations generally require an initial particle packing before calculation. This preprocessing step can be time-consuming, particularly for large particle systems and domains with complex geometries. In this study, a constructive method for sphere packing generation is presented based on centroidal Voronoi tessellation (CVT) and maximum inscribed spheres. Initial seeds are distributed in the target domain using a regular background grid, and the corresponding Voronoi cells are then constructed. Lloyd iterations are subsequently performed to regularize the spatial partition, after which the maximum inscribed sphere is generated within each final cell. The method is applicable to complex and concave 3D domains and produces sphere packings without initial particle overlap. The normalized particle radius distributions are approximately normal, with limited sensitivity to the domain geometry and particle number. Their statistical characteristics are governed mainly by the number of CVT iterations. As the iteration number increases, the mean normalized particle radius and packing fraction increase, whereas the particle size dispersion decreases. Both computational time and memory consumption increase approximately linearly with particle number over the investigated range. These results show that the proposed method provides a practical geometric approach for generating large-scale sphere packings for DEM simulations.

Graphical abstract
Keywords
Discrete element method; Particle packing; Centroidal Voronoi tessellation; Maximum inscribed sphere
文章导读