Intelligent MADS with clustering and elastic net and its application to optimal design of interior PM synchronous machines

Optimal design of the electric machines is interlinked with the optimization algorithm for obtaining the quantitative optimum and numerical analysis of electric machines for the objective function. The optimization algorithm is required to evaluate the objective function minimal times since numerical analysis of the electric machines consumes a lot of time. In this paper, intelligent MADS with a clustering and elastic net is newly adopted for the optimal design of electric machines to raise the reliability of the global optimization and fast convergence time. In particular, the proposed algorithm is employed to the optimal design of the interior permanent-magnet PM synchronous machine minimizing torque ripples, which has been regarded as one of the complicated design problems.