Pre-processing And Micro Parameters Selection Of The Bonded Discrete Element Model

Anisotropy and heterogeneity are the remarkable properties of rock materials.It is of great significance to study the fracture process of rock for the stability and safety rock engineering.In recent years,the bonded-particle model has become an effective numerical tool for studying the dynamic fracture process of rock due to its advantages in characterizing the discrete features of rock esalily.There are two main problems of solving the rock fracture problem by using the bonded-particle model are the costly pre-processing and artificial randomness of the selection of mesoscopic parameters.The purpose of this paper is to study the preprocessing modeling and the selection of the mesoscopic parameters based on the newly-developed open-source discrete element program DICE2 D developed by Tianjin University.(1)A fast generation method of the bonded-particle model based on the level set method is established.Firstly,four bonded-particle model algorithms based on the triangular finite element mesh are proposed.For the boundary processing problem,the quadratic model processing technique is proposed by the characteristics of the level set.In addition,a Gaussian random distribution function method is proposed to solve the problem of particle configuration crystallization which is the defect of the particle model generation technology based on finite element mesh.The recommended values for the selection of specific parameters are given by the detiled study of the selection and effects of the random parameters.Finally,The results of rock dynamic cutting simulation using the newly developed bonded-particle model are compared with the results of the conventional discrete element model.It is proved that the fast generation method of bonded discrete element is effectiveness and efficiency.(2)The ability of the bonded-particle model to solve the elastic boundary value problem is systematically studied,and the quantitative relationship between the mesoscopic parameters and the macroscopic parameters is established,where the effects of particle arrangement,particle influence area,shear stiffiness ratio on the macroscopic Poisson's ratio and elastic modulus of the bonded-particle model are mainly taken into account.It's shown that the geometrical arrangement and the shear interaction ratio between particles are the most important factors for reproducing the Poisson's ratio of the bonded-particle model.Furthermore,the ability of the bonded-particle model to solve the elastic boundary value problem is also studied,which reveals mismatch of the boundary value solutions between bonded-particle model and traditional continuum mechanics when considering long-range interaction and near-incompressible problems.And the coupling interaction between elastic parameter selection and fracture morphology calculation is also revealed.(3)A fast generation method of bonded particle discrete element model based on image vectorization partition is developed.Firstly,the partition of the target model region are obtained by fast image segmentation algorithm,and the outer contour of each partition are characterized by constructing a level set function.On this basis,the triangle mesh of the target region is realized by the finite element meshing algorithm based on the level set method,and the bonded-particle model of each region is generated by the fast generation method of the bonded-particle model developed in this paper.Finally,the whole bonded-particle model is formed by combining and assembling all the partitions.In the end,several common examples have shown that the method can effectively characterize complex rock structures.