关键词: 初级分裂式永磁直线电机, BP 神经网络, NSGA-Ⅱ智能算法, 电磁场有限元

Abstract: A partitioned-primary permanent magnet linear direct-drive motion unit structure was proposed. The motor employed the principle of magnetic modulation to increase motor thrust density. The permanent magnets and armature were integrated and arranged in segments, which reduced the amount of permanent magnets used and provided better heat dissipation conditions, making it suitable for long-stroke applications. The structural design method of segmented
electromagnetic structures was studied to enhance electromagnetic thrust. The reluctance thrust was calculated using the finite element method for electromagnetic fields, and the factors influencing the electromagnetic structure that improve reluctance thrust and suppress fluctuations were analyzed. The results indicate that the segmentation method can effectively enhance the motor's thrust level. Given the complexity of electromagnetic structural parameters and multiple performance targets, a BP neural network surrogate model of the motor was established using the finite element method, and the NSGA-Ⅱ intelligent optimization algorithm was applied for multi-objective optimization. After optimization, the motor’ s thrust was increased by 13. 93%, and thrust fluctuations were reduced by 13. 89%. This partitioned-primary motor holds significant application potential in the field of linear direct-drive motion units.

Key words: partitioned-primary permanent magnet linear motor, BP neural network, NSGA-Ⅱ intelligent algorithm, electromagnetic field finite element