【摘要】 讨论了表面磁钢结构的永磁电机磁极形状的设计问题 .表面磁钢结构的永磁电机磁场波形由磁钢的形状确定 .如果磁场波形已知 ,通过求解电磁场反问题就可以获得理想的磁钢形状 .其求解过程一般是先转化为电磁场正问题 ,然后进行寻优叠代 .但在寻优过程中 ,利用传统寻优方法求解该问题难以获得满意结果 ,因此本文将遗传算法引入到电磁场反问题求解中来 ,并根据该问题的特点 ,提出了“基因子链最高位受限变异”以及“相邻基因子链相关变异”等概念 ,利用有限元方法与遗传算法相结合 ,对该电磁场反问题进行了寻优求解 ,最终获得了较理想的寻优结果更多还原

【Abstract】 In PM motors, the flux distribution is determined by pole shapes and if the flux distribution is already known, the pole shape desired can be obtained by solving an inverse electromagnetic field problem. In order to solve the problm, an ordinary electromagnetic field problem must be solved first, the results are then optimized. But it is hard to get satisfactory results using ordinary optimizing methods. A Genetic Algorithms (GA) is therefore used to solve this problem. Considering the peculiarity of this problem, the concepts of ′limited mutation of the genes on the highest bit′ and ′relative mutation of the neighboring genes′ are proposed. A pole shape satisfactory enough is obtained using Finite Element Method (FEM) and GA to solve the inverse electromagnetic field problem.更多还原