A Data Mining Method for Constitutive Model of Hard Brittle Coal

Abstract

Data mining is an effective way to solve various types of complex nonlinear problems. And the anisotropy of the mechanical characteristics of coal is a typical nonlinear problem. Hard coal seam are usually characterized by brittleness, which makes them difficult to be crushed efficiently. It is important to improve the efficiency of hard coal crushing by the brittleness characterization method of hard coal is clarified. So the constitutive model and damage mechanism of hard brittle coal was studied in this paper. Damage model of hard brittle coal based on Weibull distribution was established. The damage transition point concept was proposed based on the deformation stage characteristics of hard brittle coal. And effective damage intervals for hard brittle coals can be determined based on the damage transition point. Uniaxial compression experiments of hard brittle coal were simulated using EDEM. The characterization of the effect of discrete elemental eigenstructural parameters on the brittleness of coal was analyzed. And quantitative relationships between shear modulus and brittleness, brittleness and damage transition point were fitted by Matlab. A discrete elemental constitutive model for hard brittle coals was obtained. Finally, the correctness of discrete elemental constitutive model was verified by comparing uniaxial compression experiments with other specimens. The results of this paper were expected to provide a theoretical basis for the crushing technology of hard brittle coal. And the numerical theoretical basis for data mining of the intrinsic model of hard and brittle coal was provided.