Multi-level optimisation design of a yoke-less axial-field flux-switching permanent magnet motor considering the fault-tolerant capability

Abstract

This paper proposes a multi-level optimisation design strategy aimed at enhancing the fault-tolerant capability of a Yoke-less Axial-Field Flux-Switching Permanent Magnet (YASA-AFFSPM) motor. The research addresses the crucial requirement for robustness and reliability in electric motor systems utilised in safety-critical applications. By employing a comprehensive multi-objective optimisation approach, the motor's design is refined at three levels based on the main effect stratification of design variables for the YASA-AFFSPM motor. The design variables are divided into three levels and optimised, resulting in the attainment of a comprehensive optimal solution that considers multiple design goals, including average torque, cogging torque, and fault-tolerant capability. The optimal design places a primary emphasis on attaining a substantial level of fault tolerance for both open- and short-circuit faults. This is accomplished by reducing the ratio of mutual-inductance to self-inductance. To streamline optimisation, the Kriging model approximates finite element analysis at each level. Furthermore, the effectiveness of the design is showcased through the utilisation of a dual-channel winding configuration. A prototype is constructed, and experimental validation demonstrates a significant improvement in fault-tolerance performance, with results indicating enhancements of 14% in average torque and 41% in cogging torque.