基于多物理场近似模型的高速永磁电机多目标优化设计

High-speed permanent magnet machine (HSPMM) is attracting more attention due to its advantages of high power density and efficiency, small size, light weight and fast dynamic response. The design of HSPMM is a nonlinear, multi-physics coupled process that makes it difficult to build an accurate mathematical model to optimize design parameters. This paper presents a multi-objective optimization method based on multi-physics approximate model (MPAM). This method uses a MPAM to replace the multi-physics serial design process, and directly calls the MPAM for calculations in the optimization process, which can effectively solve the time-consuming problem and avoid the problem of non-convergence in the process of finite element model call. Finally, a 1.1MW, 18 000r/min HSPMM is produced and related experiments are carried out, the feasibility of the method proposed in this paper for HSPMM optimization is verified.