Parameter calibration of soil particles contact model of farmland soil in northwest arid region based on discrete element method

Abstract: In order to solve simulation distortion problems of soil resistance and dynamic movement with discrete element method, the hysteretic spring contact model (HSCM) and the liner cohesion model (LCM) were used to simulate farmland soils in order to investigate their soil plasticity behavior and dynamics, which could be used for predicting the resistance of farm tools. To achieve better simulation results, soil basic physical parameters, contact mechanical parameters and contact model parameters of HSCM and LCM were needed. Soil basic physical parameters included particle size distribution, soil particle density, elasticity modulus, shear modulus, poisson's ratio and bulk density. Among them, the elasticity modulus was obtained by the related literature. Soil contact mechanical parameters included cohesion between soil particles formed due to the action of the water liquid bridge, cohesive strength and internal friction angle obtained from soil shear test with 6 kinds of different soil moisture. Contact model parameters were composed of coefficient of restitution, static friction coefficient and dynamic friction coefficient. The soil particles were 5 mm so that the simulation time could be shortened. Because soil simulation particles were larger than actual ones, additionally in order to achieve better simulation results, a simulation model of soil contact parameter prediction model was established according to central experimental design method (Box-Behnken) used for a comprehensive analysis of fewer trial times. Taking static friction coefficient, dynamic friction coefficient and shear strength as test factors, and soil accumulation angle as response value, 17 sets of simulation tests were carried out according to response surface method test arrangement. The results showed that the influence of static friction coefficient, dynamic friction coefficient and the interaction between dynamic friction coefficient and cohesion strength on repose angle were significant. Interaction between static friction coefficient and dynamic friction coefficient, quadratic term of static friction coefficient had a significant effect on repose angle. Taking repose angle of soil with moisture of 1%, 4%, 8%, 12%, 16% and 20% as the response value, fixing the cohesion strength, we predicted the static friction coefficient and dynamic friction coefficient using the established regression model. Then, the results were validated by soil shear test. The results showed that the relative error of shear strength was 1.18%-9.31%, and the relative error of internal friction angle was 0.55%-4.07%. In order to verify the feasibility of the proposed soil contact model, a test by using duckbill opening hole test device in soil was carried out and the results showed that when the displacement was 0-50 mm the deviation of simulation and test was big. But when the displacemtn was 50 mm, the relative error was 0.928%, which could be used for design of duckbill structure in the future.