Computationally efficient hybrid PM loss prediction method for surface‐mounted permanent magnet machine over entire working conditions

Abstract

This study proposes a computationally efficient hybrid approach to calculating permanent magnet (PM) loss in surface-mounted PM machine (SPMM) over the whole working points. By analysing the impact of current amplitude, current angle, rotor speed, and PM temperature on the corresponding subdomain PM loss results, a simplified model that considers harmonic current and explicitly expresses the relationship between PM loss and working point variables is derived. Unlike the corresponding subdomain method, the proposed model does not involve a Fourier expansion or summation over a large number of terms, which simplifies the calculation and makes the PM loss under any operating condition can be acquired instantly. Thus, the needs for PM loss table and massive experiments or calculations are eliminated. The accuracy of this hybrid approach is validated against the finite element method results of a six-phase SPMM over a wide range of working conditions. However, the cross-verification studies demonstrate that the proposed method has good efficiency and high precision, making it a promising tool for online PM loss prediction.