Application of a hybrid genetic algorithm for optimal design of interior permanent magnet synchronous machines

摘要

Poor local search capability is a defect of genetic algorithms (GA), which make the GA easily trapped in a local extremum during the later evolution stages. However, as a local optimization method, the Taguchi method has strong local optimizing ability, which could overcome this shortcoming of GA. Therefore, in this paper, a hybrid genetic algorithm (HGA) which combines the GA with the Taguchi method is used to optimize the rotor shape of an IPMSM to obtain lower iron loss and torque ripple as well as higher average torque and efficiency. The optimization results of the HGA design is compared with the initial and GA designs. It is shown that the performance of the IPMSM is effectively improved by employing the GA and HGA, especially by adopting HGA. Moreover, better flux-weakening capability and less mass magnet is also obtained by these methods.