Study On Mining Slope Stability Based On Improved Material Point Method

Supported by the national key research and development program(No.2018YFC1504802),this paper focuses on the numerical method of material point method,which is suitable for modeling and analysis of large deformation problems.The specific research is carried out from four aspects: improving the material point method to improve the stability and simplify the algorithm,developing GPU parallel program to solve the problem of large amount of calculation of the material point method,verifying the improved algorithm and program,and applying it to the stability simulation of mining slope.The research contents and results are as follows:(1)The momentum mapping scheme of affine particle in cell method is applied to the classical material point method,and MPM-APIC is proposed to greatly improve the stability;based on MPM-APIC,the affine velocity matrix is deduced as the first order approximation of velocity gradient,and an improved MPM-APIC is proposed.It saves the steps of solving the velocity gradient of material points by interpolation of grid nodes,which greatly simplifies the rate of rotation,the rate of deformation and the solving process in a single time step.The algorithm has better stability while reducing the complexity and computation amount by more than half.Under the improved MPM-APIC framework,the higher order shape functions are studied and integrated into them,and a relatively complete computing framework is developed.(2)In view of the time-consuming calculation of material point method,a GPU parallel material point method is designed and developed.One is a set of functional program with concise code developed based on the idea of functional programming;the other is based on the idea of Objected-Oriented,making the material point method program more advanced encapsulation,and developing a set of easy to use and subsequent development of material point method class.According to the characteristics of the two sets of programs,the corresponding pre-processing and post-processing modules are studied and developed.(3)Verify the improved algorithm and program.The numerical stability advantage of the algorithm is verified through sand column collapse.By simulating 1-D quasi-static compression and uniaxial compression,the accuracy of the algorithm is proved in two dimensions of one dimension and two dimensions.After the simulation of the aluminum bar experiment proves that the improved algorithm still has high accuracy in simulating the dynamic problem,the efficiency of the algorithm and the developed program is verified.By simulating the landslide of a typical shape slope,the effectiveness of the algorithm and program on the actual scale problem is verified.(4)Based on the improved MPM-APIC algorithm,the Jianshanying landslides were simulated by using self developed material points method class.This paper reveals the deformation and evolution process of Jianshanying landslide under the action of mining;explores the deformation and failure characteristics of Jianshanying landslide under different mining sequences;simulates the landslide process and makes a preliminary inquiry into the application of material point method in mining problems.