Multi-Objective Optimization Framework of a Radial-Axial Hybrid Excitation Machine for Electric Vehicles

Abstract

This paper proposes a multi-objective optimization framework for a radial-axial hybrid excitation machine (RAHEM) to provide higher average torque, better flux regulation ability and smaller torque ripple, which are applied to electric vehicles (EVs). The design variables related to multiple-objective are analyzed by sensitivity stratification. Non-dominated sorting genetic algorithm II (NSGA-II) based on response surface model (RSM) is adopted for the high sensitivity layer variable. The advantages are selected with the pareto optimal solutions (POS), while the low sensitivity layer variables are optimized by sensitivity ranking for single parameter scanning. The optimization function compares the two sensitive layers results to obtain the optimal design. Three-dimensional (3-D) finite element analysis (FEA) is used to compare the electromagnetic performance of initial and optimal designs. Finally, a prototype is manufactured to verify the effectiveness of the proposed framework.