Optimization design of a doubly salient 8/6 SRM based on three computational intelligence methods

Abstract

The aim of this paper is to optimize an 8/6 doubly salient switched reluctance machine using three computational intelligence methods, which include particle swarm optimization, a genetic algorithm, and differential evolution. Three cases are investigated where different parameters are considered like the stator pole arc, rotor pole arc, and ratios, which define the stator yoke and rotor thickness. The objective functions considered are the average torque and the torque-to-weight functions. The simulations are carried out using MATLAB and FEMM software. The optimal results found are compared with the initial design, and it is shown that high improvements are achieved.