关键词: 永磁同步电机, 齿槽转矩, 径向基神经网络法, Isight, 多岛遗传算法, 有限元仿真

Abstract: Aiming at a low-speed permanent magnet motor with 32 poles and 36 slots,on the Isight optimization platform,approximate models between cogging torque,magnetic flux leakage coefficient and design variables  were set up separately by optimal latin hypercube sampling and the radial basis function (RBF) method,and multi island genetic algorithm (MIGA) was used to the optimal design.The relative error between the finite element simulation results and the predicted value of the RBF approximation model was 1.6%,and the optimized cogging torque was 59% lower than the initial value,which proves the effectiveness of the optimization method.The use of Isight command stream improved the efficiency of optimization.