Multiresponse optimization to improve the torque behavior of an outer-rotor permanent-magnet machine using gray relational analysis based on the Taguchi method

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 gray relational analysis based on the Taguchi method with the single response of the gray relational grade instead of multiple responses. A considerable amount of multiresponse improvement is achieved according to the results of the two optimizations. Performance improvements of 20.6%, 32.0%, and 24.5% are obtained for the average torque, the torque ripple, and the total harmonic distortion of the back-EMF, respectively.