Surrogate-model-based optimization design of Wound-Field Flux Switching Machine for industrial applications

Abstract

The Wound-Field Flux Switching Machine (WFFSM) is a promising solution for high-performance industrial applications, offering high torque density, brushless operation, and controllable excitation flux. However, optimizing its design is challenging due to nonlinear flux interactions and multiple competing objectives. This study presents a surrogate-model-based optimization framework using response surface methodology (RSM) and multi-objective genetic algorithm (MOGA) to enhance the electromagnetic performance of a 1.5 kW WFFSM. Finite element analysis (FEA)-driven sampling and RSM surrogate modeling enable efficient exploration of the design space. The optimized WFFSM achieves a 77.7% reduction in torque ripple compared to the initial model, along with significant improvements in torque output and enhanced back-EMF characteristics. The proposed approach contributes a novel, systematic optimization strategy specifically tailored for industrial WFFSM applications, ensuring improved efficiency, reliability, and adaptability in next-generation electric machines.