In the single-winding consequent-pole bearingless permanent magnet synchronous motor (SW-CP-BPMSM), the torque and suspension systems utilize a shared winding configuration. This structure significantly intensifies inter-system coupling. Furthermore, the presence of non-linear and strongly coupled relationships among structural parameters, combined with inherent coupling and conflicts between optimization objectives, makes the unified optimization of key performance indicators for both the torque and suspension systems a substantial challenge. To address these issues, this paper proposes a dual-system interactive optimization strategy based on the classification of sensitive variables. First, the strategy employs the Sobol method to conduct a global sensitivity analysis. By defining dual-system coupled sensitive parameters and single-system sensitive parameters, the method achieves dimensionality reduction through parameter classification. Subsequently, Response Surface Methodology (RSM) and Back Propagation (BP) neural network surrogate models are constructed for the suspension and torque systems, respectively. A progressive optimization process—comprising single-system optimization followed by dual-system interactive optimization—is then performed on the single-system and dual-system sensitive variables to determine the final optimal parameters. Finally, a comparative simulation analysis of the key performance indicators for both the torque and suspension systems before and after optimization is conducted. The results validate the feasibility and effectiveness of the proposed optimization strategy.
Yuan et al. (Tue,) studied this question.