Genetic Algorithm and Fuzzy Based on The Taguchi Optimization to Improve The Torque Behavior of An Outer-Rotor Permanent-Magnet Machine

Abstract

The torque behavior of an outer-rotor surface-mounted permanent-magnet machine is improved by identifying seven pertinent design variables, including rotor height. The optimal design variables are revealed by analyzing 18 experiments determined by the Taguchi method for the minimum torque ripple, minimum total harmonic distortion of the induced voltage, and maximum average torque. In addition, the optimal design variables are obtained very quickly by using fuzzy inference mechanism and genetic algorithm (GA) based on the Taguchi method with the single response of the multi-response performance index instead of multiple responses. A considerable amount of multi-response improvement is achieved according to the results of the two optimizations.