Multi-Objective Optimization Strategy for Permanent Magnet Synchronous Motor Based on Combined Surrogate Model and Optimization Algorithm

When a permanent magnet synchronous motor (PMSM) is designed according to the traditional design theory, performance of often challenging achieve desired goal, and further optimization parameters usually required. However, strongly coupled, non-linear, multivariate complex system, it challenge optimize by methods. It needs rely on reliable surrogate models algorithms improve PMSM, which one problematic aspects optimization. Therefore, this paper proposes strategy based combination high-precision combined model method stator rotor structures interior PMSM (IPMSM). First, variables were classified into two layers with high low sensitivity comprehensive parameter analysis. Then, Latin hypercube sampling (LHS) used obtain sample points for highly sensitive variables, various methods are employed construct variables. Each target acquired points, from most accurate selected non-dominated ranking genetic algorithm-II (NSGA-II) find best. After optimizing high-sensitivity new finite element (FEM) built, Taguchi low-sensitivity Finally, analysis (FEA) was adopted compare initial optimized ones IPMSM. The results showed that improved prove effectiveness reliability proposed method.