Multi-Objective Optimization of an Air-Cored Axial Flux Permanent Magnet Synchronous Machine with Segmented PMs based on Support Vector Machine and Genetic Algorithm

This paper focuses on the multi-objective optimization of an air-cored axial flux permanent magnet synchronous machine (AFPMSM) with segmented PMs. Support vector machine based on radial basis kernel function is combined with orthogonal design to create the surrogate model between output performance and design parameters of a coreless AFPMSM with segmented PMs efficiently, and genetic algorithm is used to obtain the global optimal solutions to decrease the torque ripple and increase the output torque. The effectiveness of the proposed method is verified by finite element analysis.