Optimization of Halbach magnetic roller for eddy current separation based on the response surface method and multi-objective genetic algorithm

摘要

Performance optimization of the magnetic roller of eddy current separator is an issue of great significance in the recycling industry. In this study, the Halbach array is proposed for the first time to be used in the eddy current separator. And the new concepts of area field intensity and magnet efficiency density are proposed to quantitatively evaluate the magnetic field intensity around the magnetic roller surface and the utilization efficiency of the magnet respectively. Then the effects of various structural parameters of the Halbach magnetic roller on the area field intensity and magnet efficiency density were systematically investigated by employing finite element analysis and response surface methodology. The results showed that the regression models of the area field intensity (R-2 = 0.9846, Pred.R-2 Y = 0.9420) and magnet efficiency density (R-2 = 0.9946, Pred.R-2 = 0.9782) are extremely significant and have good prediction capability. The magnetic roller can form a stronger magnetic field by selecting a larger magnetic block, and the utilization efficiency of the magnet is higher when the magnetic block's shape tends to be flat or has larger curvature. Furthermore, a multi-objective optimization model with the area field intensity, magnet efficiency density, and magnetic pole number is established, and a multi-objective genetic algorithm is applied to get the optimal combinations of various structural parameters. It was found that the points of the Pareto-optimal solution are concentrated on a spatial surface, which reflects the trade-off relationship among the three objectives. It also shows that the performance parameters of many Pareto-optimal solutions are superior to that of previous studies and the eddy current separators found in the vicinity. The proposed method framework and the results are useful for the design and optimization of the magnetic roller, which will improve the separation efficiency of eddy current separation. (C) 2020 Elsevier Ltd. All rights reserved.