Multiphysics Hierarchical Optimization of Switched Reluctance Machines Based on an Improved Hybrid Method

A multiphysics hierarchical optimization for switched reluctance machines (SRMs) is presented, considering both electromagnetic and vibration performances. The basic topology and control strategy of SRMs are briefly introduced, based on which the geometric and control variables are identified. Given the huge computational cost of the multiphysics coupled-field finite element method (FEM), a hybrid method (HM) with improved efficiency is developed for SRM vibration predictions. It integrates the radial electromagnetic forces originating from the 2-D FEM into the analytical vibration syntheses based on a three-layer simply supported cylindrical shell model. To further enhance the optimization efficiency, the optimization variables are classified into three levels based on the sensitivity analysis and physical domains, and each level is optimized sequentially with distinct focuses and strategies. The effectiveness and accuracy of the proposed method are validated by the electromagnetic-mechanical coupled-field FEM and experimental results.