Optimal Design of IPMSM for Electric Bus Using a Sub-Domain Algorithm with Dynamic Area Sampling

Abstract

The optimal design of an electromagnetic machine requires enormous computation time and includes multi-optima in a design area. Multimodal optimization is a useful method for multi-optima problems. In this paper, a novel multimodal algorithm, specifically a sub-domain algorithm with dynamic area sampling (SADAS), is proposed. A conventional kriging has a memory problem that occurs when increasing the grid density for resolution of interpolation. The SADAS solves this problem by using a sub-domain method. Furthermore, a novel sampling method, dynamic area sampling, is proposed to improve the capability of peak search and convergence speed. The validity of the SADAS is verified by test functions and it is applied to the optimal design of an interior permanent magnet synchronous motor for an electric bus.