Study on Screening Parameter Optimization of Wet Sand and Gravel Particles Using the GWO-SVR Algorithm

Abstract

The optimization of screening parameters will directly improve the screening performance of vibration screens, which has been a concern of the industry. In this work, the discrete element model of wet sand and gravel particles is established, and the vibration screening process is simulated using the discrete element method (DEM). The screening efficiency and time are used as evaluation indices, and the screening parameters including amplitude, vibration frequency, vibration direction angle, screen surface inclination, the long and short half-axis ratio of the track, feeding rate, and screen surface length are investigated. The results of an orthogonal experiment and range analysis show that the amplitude, screen surface inclination, and vibration frequency are significant factors affecting screening performance. Then, the support vector regression optimized with the grey wolf optimizer (GWO-SVR) algorithm is used to model the screening data. The screening model with excellent learning and prediction ability is obtained with the Gaussian kernel function setting. Moreover, the GWO-SVR algorithm is used to optimize the screening parameters, and the screening parameters with optimal screening efficiency and time are obtained. Furthermore, the effectiveness and reliability of the optimized model are verified using the discrete element calculation. The optimization strategy proposed in this work could provide guidance for the structural design of vibration screens and screening process optimization.