A Neural Network and NSGA-II Based Multi-objective Optimization Design Method for Permanent Magnet Synchronous Machine

This paper presents a multi-objective optimization design method for permanent magnet synchronous machine (PMSM) by using back propagation neural network combined with non-dominated sorting genetic algorithm II (NSGA-II). First, neural network is used to map out the nonlinear relationship between the structural parameters of the machine and the output performances, such as average torque and torque ripple. Then NSGA II is implemented to find out the Pareto-optimal front. According to the Pareto-optimal front, optimal structure parameters of the PMSM for the best output optimization results are selected.