Optimization of Predesign of Switched Reluctance Machines Cross Section Using Genetic Algorithms

Genetic algorithms (GA) have been applied in optimization of machine designs since the first publication in 1975 (J.H. Holland, 1992). In this paper, a practical implementation of this search technique in predesign of switched reluctance machines is presented. An optimized design was found by means of GA based on an objective function for maximizing an average torque of the machine, where dimensions and a thermal loading are specified. Moreover, an auxiliary objective for the most preferable geometries is utilized, well supporting a vector format of the model in GA. The simulation results verified and demonstrated the efficacy of the proposed strategy.