A Novel Three-Stage Optimization Design Method of Asymmetric-PM Variable Flux Memory Machine Considering Magnet-Axis-Shifting Effect

Abstract

Asymmetric-permanent-magnet variable flux memory machine (APM-VFMM) commonly suffers from the optimization issues, i.e., variable magnet-axis-shifting (MAS) effects result in distinctly differentiated electromagnetic characteristics under multiple magnetization states (MSs). Thus, this article proposes a novel three-stage optimization method for APM-VFMM considering the MAS effects to achieve efficient optimization. In the first stage, an analytical model based on the magnetic equivalent circuit (MEC) and torque models considering the MAS effects is developed to determine the key optimization objectives of the investigated machine while theoretically revealing the internal mechanisms for the improvement of the flux regulation range and torque capability. In the second stage, the comprehensive sensitivity analyses of the main geometric parameters are implemented based on a self-organizing map (SOM) method. Subsequently, the geometric parameters with higher sensitivity are selected for the multiobjective optimization process, thus effectively reducing the computational burden. In the third stage, a multimode circuit control model is developed and utilized in a single finite element analysis (FEA) case to consider multiple MSs and different MAS effects simultaneously for multiobjective optimization. The FEA predictions and experimental tests carried out on an APM-VFMM prototype confirm the effectiveness of the proposed three-stage optimization method.