基于改进萤火虫算法的永磁同步电机多模态优化设计  

Multi-modal optimal design of permanent magnet synchronous motor based on improved firefly algorithm

作　　者：夏斌[1] 王超[1] 孙鑫 宋春丽 XIA Bin;WANG Chao;SUN Xin;SONG Chunli(School of Electrical Engineering,Shenyang University of Technology,Shenyang 110870,China;Transformer Business Department,Tebian Electric Apparatus Stock Co.,Ltd.,Shenyang 110027,China;Anshan Information Engineering School,Anshan 114001,China)

机构地区：[1]沈阳工业大学电气工程学院,辽宁沈阳110870 [2]特变电工沈阳变压器集团有限公司变压器事业部,辽宁沈阳110027 [3]鞍山市信息工程学校,辽宁鞍山114001

出　　处：《电机与控制学报》2024年第4期131-138,共8页Electric Machines and Control

基　　金：辽宁省教育厅项目(LJKQZ2021051)。

摘　　要：永磁同步电机的结构参数选择往往是多模态、非线性的优化问题,同时其结构优化设计又存在计算分析时间过长、计算效率过低、收敛慢等问题。文中针对此类问题提出了一种基于改进萤火虫算法的新型多模态优化算法。改进萤火虫算法主要引用了等值线概念并通过克里金(Kriging)代理模型自适应确定子计算区域,从而获得多个全局最优解和局部解,并且引入单纯形法和改进停止条件以提高算法效率。此外,研究了分层余一法用于检验代理模型精度。最后,将其应用到永磁同步电机的结构优化中,使齿槽转矩成功降低了38.33%。The structural parameter selection of permanent magnet synchronous motor(PMSM)is often a multi-modal and nonlinear optimization problem.At the same time,the structural optimization design of PMSM has some problems,such as long computational analysis time,low computational efficiency and slow convergence.In order to solve these problems,a novel multimodal optimization algorithm based on improved firefly algorithm was proposed.In the improved firefly algorithm,the concept of contour lines was introduced and the sub-computation region was determined adaptively through Kriging surrogate model,so as to obtain multiple global optimal solutions and local solutions,and simplex method was introduced and stop condition was improved to improve efficiency of the algorithm.In addition,a stratified leave-one-out method was studied to verify accuracy of the surrogate model.Finally,it is applied to the structure optimization of PMSM,and the cogging torque is successfully reduced by 38.33%.

关 键 词：永磁同步电机 多模态优化 萤火虫算法 克里金代理模型 等值线 留一法 

分 类 号：TM351[电气工程—电机]