Multilevel and Multiobjective Optimization of a Six-Phase SRM with Bezier Curve

This paper presents parameter optimization based on multilevel optimization strategy (MOS) and shape optimization based on the Bezier curve for a 12/10 switched reluctance motor (SRM). First, the multiobjective optimization problem, including objectives and constraints, are defined. Then, the finite-element model (FEM) is validated by the experiment and the Kriging model is validated by FEMs to prove the accuracy of this optimization. Second, parameter optimization is conducted based on MOS. The parameters are divided into three subspaces according to the result of comprehensive sensitivity analysis, and each subspace’s optimization is performed sequentially. Finally, the shape optimization based on the Bezier curve is conducted when the performance of the investigated motor cannot be improved by parameter optimization. The initial and optimal designs are compared and discussed. From the discussion, it can be found that the performance of the motor can be effectively improved through param-eter optimization and shape optimization.