基于响应面与机器学习的曲料颗粒离散元参数标定OA

To improve the accuracy of discrete element method simulations for qu,the key contact parameters of qu particles were calibrated using the particle angle of repose as the evaluation index,combining physical experiments with numerical simulations.The physical parameters and angle of repose of qu were determined through physical experiments.The Plackett-Burman design was employed to identify the key contact parameters significantly affecting the angle of repose of qu,and the steepest ascent method was used to optimize the range of these parameters.Based on the response surface results,machine learning was applied to determine the optimal parameter set.The results showed that the decision tree regression model outperformed random forest,SVR,KNN,and XGBoost models in terms of prediction accuracy and stability for the angle of repose.The optimal parameter combination for qu was determined as follows:coefficient of static friction between qu particles of 0.774,coefficient of rolling friction of 0.513,and JKR surface energy of 0.228.Using the calibrated parameters,angle of repose and die compression tests were conducted.The relative error between the experimental and simulated angle of repose was 0.64%,while the relative errors for compression displacement and compression ratio were 1.14%and 1.17%,respectively.This research provides a reference for the compression process of qu and related discrete element analyses.