Tracking of Winding and Magnet Hotspots in SPMSMs Based on Synergized Lumped-parameter and Sub-domain Thermal Models

Lumped-parameter thermal models (LPTMs) have been widely used to estimate critical temperatures in permanent magnet synchronous machines (PMSMs). However, in some cases, they may be incapable of predicting the hotspots and even cause temperature misestimation. In this paper, the synergized lumped-parameter and sub-domain thermal models are proposed for surface-mounted PMSMs to track the hotspots in the most vulnerable components, i.e. active/end-windings and rotor magnet. Firstly, the T-type elementary network based LPTM is developed to predict the average temperatures in main components and the boundary conditions. Secondly, the boundary conditions are applied to three sub-domain thermal models for the active/end-windings and the rotor magnet to predict the respective hotspots and to further correct the misestimated average temperatures. Finally, the proposed synergized thermal model is verified by the experiments on a prototype surface-mounted PMSM under various load conditions.