PREDICTION OF DEM PARAMETERS OF COATED FERTILIZER PARTICLES BASED ON GA-BP NEURAL NETWORK

To provide an efficient and reliable calibration method with reduced time cost increased accuracy, the angle of repose (AoR) in simulation is batch-processed based on Python GA-BP neural network used to improve prediction accuracy DEM parameters coated fertilizer particles. The single-factor test data were firstly interpolated obtain sufficient training samples, thus avoiding drawback that BP tends fall into local minimum during process. Then was trained combination orthogonal test, fitted correlation coefficients all greater than 0.975, indicating algorithm has strong generalization performance good stability. predicted values matched expected output values, can accurately predict nonlinear function output, be approximated as actual function. With AoR value, value obtained 24.457° when coefficient restitution (CoR), static friction (CoSF), rolling (CoRF) 0.509, 0.176, 0.0332, respectively, relative error 0.068%, well-fitted could