Multi-objective Design Optimization of the DPMSM Using RSM, Taguchi Method, and Improved Taguchi Method

Three methods i.e. Response surface, Taguchi, and improved Taguchi are used for multi-objective design optimization of the drilling permanent magnet synchronous motor (DPMSM) in this study. The optimization objectives are determined according to the performance requirements of the DPMSM aiming to maximize efficiency and minimize the torque ripple coefficient. 4 Of 8 electromagnetic parameters were screened as significant factors by factorial experiment. The response surface approach uses data from finite element analysis to fit a polynomial model between objectives and factors. The particle swarm optimization with constrict factor (PSOCF) algorithm is used for determining the Pareto frontier. In the Taguchi method, the effects of the noise factors on the optimization objective are considered. The combinations of significant factor levels for different objectives obtained from orthogonal experiments were fused into an optimal combination based on the influence proportion of factors. In the improved Taguchi method, the optimization objectives are fused into multiple performance characteristic indexes (MPCI) based on fuzzy inference. The MPCI is optimized by the Taguchi method. Then the sequential experiments are used to increase the range and effectiveness of the optimization. The characteristics and the optimization results of the three methods were compared and the improved Taguchi method was verified based on a test of the prototype motor.