Simultaneous Stator Winding and Permanent Magnet Temperature Estimation for Permanent Magnet Synchronous Machines

This paper proposes a noninvasive Kalman filter based simultaneous stator winding and permanent magnet (PM) temperature estimation approach for permanent magnet synchronous machines (PMSMs). An estimation model involving the winding and PM temperatures is firstly derived from the machine model with machine inductances being cancelled. Thus, the winding and PM temperatures can be directly estimated from the measurements available in the drive system by using the derived model without the need of inductances. To improve the estimation performance, a state-space model is developed from the derived model and Kalman filter is employed for simultaneous winding and PM temperature estimation. Compared with existing approaches, the proposed approach is noninvasive and independent from machine inductance variation and it is able to provide accurate estimation of winding and PM temperature simultaneously. The proposed approach is validated on a laboratory PMSM drive system.