Surrogate Model-Based Full Operating Condition Optimization Design for Permanent Magnet Synchronous Motors

To address the challenges of multi-objective coupling and high computational costs in the comprehensive operational optimization design of permanent magnet synchronous motor (PMSM), this paper proposes a multiobjective collaborative optimization method based on a fused surrogate model integrating response surface methodology and thermal network models. First, an optimal Latin hypercube sampling is employed to generate a sample set, constructing a second-order response surface model (RSM) for electromagnetic performance, with its accuracy validated using the multiple correlation coefficient$\mathbf{R}^{2}$. Subsequently, a reduced-order electromagnetic-thermal coupling model is established by incorporating a T-type equivalent thermal network, enabling rapid temperature rise prediction. On this basis, the NSGA-III algorithm is applied for multi-objective optimization of motor performance, volume, and temperature rise, where a reference point mechanism ensures uniformity of the solution set in highdimensional objective space. Finally, finite element simulations are conducted to validate the optimization results. The optimized motor exhibits a 4.17 % increase in peak torque, a 17.64 % improvement in torque-to-weight ratio, and a$\mathbf{9. 5 5 \%}$reduction in permanent magnet temperature rise, confirming the effectiveness of the proposed method. This study provides a novel approach for the efficient optimization design of PMSMs under complex operating conditions.