Parameter calibration of discrete element simulation model for latosol particles in hot areas of Hainan Province

Latosol is the zonal soil of tropical rainforest or seasonal rainforest in the southernmost part of China. Compared with ordinary soil, latosol has stronger expansibility, flow plasticity and water holding capacity. In the cultivation of latosol, there are many problems, such as high resistance, low efficiency, and serious soil adhesion of parts touching the soil. In recent years, with the continuous improvement of agricultural mechanization level, discrete element method is gradually applied in the field of agricultural engineering. But the existing research is mainly aimed at the soil with weak viscosity, such as sandy soil, dry soil or rock soil. These research results are not suitable for the latosol, and there are few researches on the parameter calibration of the discrete element simulation model of latosol. In this paper, the simulation contrast test of soil repose angle and soil breaking resistance, and calibrate the contact parameters of the discrete element model of latosol had been carried out with "Hertz-Mindlin with JKR" contact model in the EDEM software. The actual repose angle of latosol with moisture content of 7.8% was measured to be 32.35 °. The recovery coefficient, static friction coefficient and rolling friction coefficient between latosol, latosol and contact parts were measured by design test. Seven coefficients were selected as test factors by Design-expert software and factor level was set. The repose angle was taken as test index and design parameter calibration test was carried out. Using Plackett-Burman Design test to determine the key factors affecting the repose angle of latosol include recovery coefficient of latosol-latosol, rolling friction coefficient of latosol-latosol, rolling friction coefficient of latosol-28MnB5 plate and surface energy. Using regressive analysis by Box-Behnken to fit the test results, and the repose angle regression model was obtained. The regression variance analysis of the model showed that the regression model was very significant, and the repose angle could be predicted according to the model. Taking the actual repose angle of latosol as the objective value, the regression model was optimized, and the optimal values of four significant factors were obtained: latosol-latosol recovery coefficient 0.40, latosol-latosol rolling friction coefficient 0.07, latosol-28MnB5 plate rolling friction coefficient 0.24, JKR surface energy 4.31 J/m2. The simulation experiment of the repose angle of EDEM was designed, and the image of the simulated repose angle was processed by python. The relative error between the simulated repose angle and the measured repose angle was 0.62%. The soil tank model was established to simulate the movement process of the breaking blade in the latosol, and the test was carried out to compare the fitting situation of the change of the breaking resistance. Comparion of soil breaking fitting resistance curve in numeral simulation and test, showed that the calibration of the parameters of the discrete element simulation model of latosol was accurate and reliable.