Contributions to the multi-objective metaheuristic algorithm-based optimization of the analytical design of flux-switching generators

Flux-Switching Generators (FSGs) are stator-active electrical machines, wherein both the excitation mechanism and the armature windings are located on the stator. These machines possess significant advantages, including high power density, a compact and lightweight structure, efficient operation across a wide speed range, and low maintenance costs, high reliability, and robustness owing to their brushless design. These attributes to the position of FSGs as strong candidates for applications in electric vehicles, aerospace, small-scale wind turbines, and microgrid applications. However, this reluctance-based operational principle inherently causes FSGs to exhibit high torque ripple and significant harmonic components in the induced voltage. This thesis aims to develop an analytical design model for a low-power FSG and to perform a multi-objective optimization to enhance the performance of its output parameters. The first chapter of the thesis presents a comprehensive literature review on FSG architecture. In the second chapter, the effects of critical parameters such as skew angle, rotor pole tilt angle, and pole arc ratio on generator performance are analyzed using the Finite Element Method (FEM) on a 12-slot/10-pole FSG model. In the third chapter, as the core methodology of the thesis, an innovative MATLAB/Simulink-supported analytical design model based on the Magnetic Equivalent Circuit (MEC) was developed, which is more precise than existing approaches in literature. Through this model, analytical equations defining the generator's losses, volume, average torque, and torque ripple were derived. The results of the developed MEC model were validated against FEM analyses, confirming high accuracy. In the fourth chapter, these analytical equations were integrated with the Grey Wolf Optimization (GWO) algorithm to implement multi-objective optimization. Nine different design variables were utilized in the optimization. The multi-objective optimization aimed to minimize total losses, volume, and torque ripple while maximizing average torque. As a result of validating the GWO-based multi-objective optimization via FEM, significant improvements were obtained under different scenarios. In the scenario prioritizing average torque, the torque value was increased to 1,62 times that of reference design. In the scenario prioritizing torque ripple, the ripple was reduced to 0,33 times the reference value. Furthermore, in the scenario prioritizing total losses, the losses were reduced to 0,68 times. This study presents a methodology that successfully integrates MEC and meta-heuristic optimization, thereby accelerating the design process of FSGs and enabling the attainment of high-performance designs.