Parameter Analysis and Multilevel Design Optimization of a Permanent Magnet Claw Pole Machine with Hybrid Cores

The permanent magnet claw pole machine (PMCPM) with hybrid cores investigated in this paper is a new claw pole machine (CPM) with the modular design where its stator is composed of stator yoke made by silicon sheet and stator claw pole teeth made by soft magnetic composites (SMC). As the PMCPM is with very complex 3D structure and there is a great number of design parameter needs to be determined, optimizing the PMCPM is very time consuming and using the traditional multilevel optimization strategy (MOS) is an effective way. In this paper, the parametric analysis is proposed as a pre-design to the MOS, and a parameter-oriented grouping (POG) method based on parameter analysis is proposed to improve the optimization speed of MOS. Specifically, the objectives and design parameters are streamlined and grouped by using inter-objective correlation analysis (CA), parameter-to-objective sensitivity analysis (SA), and inter-parameter correlation analysis (RA) method, then the MOS is performed based on the parameter analysis results. The traditional MOS method are employed for the benchmark comparison, and the effectiveness of the proposed MOS based on POG is verified by the experiment.