Innovative Design and Optimization of Flux Switching Machines for Transportation Electrification Using a Multi-Level Response Surface Approach

This paper presents a novel multi-level optimization methodology for flux switching wound field machines (FSWFMs), designed to meet the demands of transportation electrification. FSWFMs, with their high torque density and independence from rare-earth materials, hold significant potential for sustainable electric vehicle systems. However, challenges such as high torque ripple, suboptimal efficiency, and computational costs hinder their widespread adoption. The proposed methodology systematically integrates response surface methods (RSM), sensitivity analysis, and multi-objective genetic algorithms to overcome these limitations and improve key performance metrics such as torque, torque ripple, and efficiency. A structured sensitivity analysis is employed to classify design variables into non-sensitive, mild-sensitive, and strong-sensitive categories, enabling targeted optimization while ensuring computational efficiency. Using RSM as a surrogate model, the framework reduces the need for repeated finite element analysis simulations, achieving <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">${1 2, 5 0 0}$</tex> evaluations in just 6.5 hours compared to conventional methods requiring 10-15 hours for <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">${1, 0 0 0}$</tex> evaluations. Results demonstrate a 69.06 % reduction in torque ripple, an 11 % increase in torque, and enhanced energy efficiency. These advancements position FSWFMs as viable solutions for sustainable mobility systems, bridging the gap between theoretical optimization and practical implementation. The framework offers scalability and computational efficiency for broader industrial applications.