Robust Multi-Objective Optimization for BEESM Based on Improved Climbing Algorithm

Robust optimization design of brushless electrically excited synchronous machines (BEESMs) is a problem that has received extensive attention. The increase in finite element calculation cost due to the increase in the number of motor parameters is one of the main problems faced by optimization. In this paper, a robust multi-objective optimization design method of BEESM based on an improved hill-climbing algorithm is proposed. All design parameters are divided into three subspaces according to the sensitivity by the sensitivity analysis method combined with Kendall's rank coefficient, thereby reducing the consumption required for finite element model (FEM) calculation. The screening problem of Pareto frontier solutions is solved by an improved hill-climbing algorithm. The candidate points to be optimized are screened through the improved climbing algorithm, and only the candidate points located on the Pareto frontier will be optimized, which ensures the high performance of the candidate points. Based on the noise problems that may occur in actual production and processing, the candidate points are robustly analyzed, and the optimal design is screened out. The robust optimization design method proposed in this paper can reduce the computational cost and improve the robustness of the motor based on improving the performance of the motor. © 2022, Progress In Electromagnetics Research B. All Rights Reserved.