三维颗粒生成算法(英文)

Three-dimensional particles generation algorithm

颗粒生成是进行散粒体细观数值模拟的前提.为快速生成指定级配的颗粒并减少自由堆积过程,针对三维球形及多面体颗粒生成,改进了颗粒生成的前进面算法.其基本思想是:随机选取3种粒径,生成3个相互接触的球,3个球的球心构成初始的2个前进面;随机选取前进面和一种粒径,在前进面的外侧生成新的球,使新球与构成前进面的3个球相切;更新前进面,进而生成新的球,如此往复直到生成全部颗粒或填满容器.在球的表面随机生成点,这些点构成一个凸锥-多面体,进而生成多面体颗粒.编制了前进面法颗粒生成程序,5个算例证明了该算法的高效性.采用此法生成的颗粒不一定是全局最优的排列,但可以保证局部最密分布.该算法的最大优点是可以大大减少CPU运算时间.在主频为2.8GHz的PC机上生成5万个颗粒只需86s.

Particles generation is the pre-step of discrete element simulation of practical problems.One advancing front face（AFF） algorithm for three-dimensional sphere/polyhedron particles packing is advanced.It can be used to constructively generate a random initial packing of spheres / polyhedrons with different radii within a 3D domain.Firstly,select three radii randomly and generate three spheres which contact with each other.The centroids of the initial three spheres form two advancing front faces.Then,select an AFF as the active face,randomly select a radius and generate a new sphere which contacts with the three old spheres forming the current active AFF.Update AFF list,select a new active AFF and generate a new sphere until all needed spheres have been generated or the domain has been filled in.If polyhedron particles are required,vertices are generated randomly on the sphere surface and these points are connected to form a convex hull.Go through all spheres and polyhedron particles are formed and filled in the 3D domain.The generated packing is not a globally optimal arrangement,but a locally highest density is achieved from the algorithmic point of view.The performance of the algorithm is illustrated by several examples.The major benefit of this algorithm is the significant reduction of CPU time required for the preparation of an initial discrete object configuration in discrete element method （DEM）/discontinuous deformation analysis（DDA） simulations.It is demonstrated that it takes only 86 s for the proposed algorithm to generate 50 thousand particles on a PC with one 2.8 GHz processor.