Research on Response Surface Modeling Method of Permanent Magnet Synchronous Motor Based on Sequential Experimental Design

The optimization process of traditional permanent magnet synchronous motor (PMSM) based on finite element calculation and control variable method has problems such as time-consuming and difficulty in achieving optimal design. It affects the solution efficiency and target response expectation of the PMSM multi-objective optimization problem. In order to solve the problem of high coupling degree of PMSM design factors and difficulty in visualizing the graph, the response surface nonlinear models with torque ripple and average torque per unit current as optimization objectives are established respectively in this paper based on factor significance analysis and central composite sequential experimental design method to realize the nonlinear mathematics modeling of PMSM under multi-variable interactive response. Subsequently, an improved response surface modeling method based on experimental space optimization is proposed to solve the problem of prediction deviation in response surface nonlinear modeling. This method constitutes a second-order response surface test scheme by adding appropriate test data points on the basis of the factorial experiment design, which can effectively improve the nonlinear modeling efficiency of the PMSM response surface.