Online thermal parameter identification for permanent magnet synchronous machines
Shuai Xiao, Antonio Griffo
Abstract
Temperature monitoring of permanent magnet synchronous machines (PMSMs) is of great importance because high temperatures can significantly shorten the lifetimes of motor components. Accurate temperature predictions can be achieved using reduced‐order lumped parameter thermal networks (LPTNs) with accurate thermal parameters. In this study, an online estimation method based on the recursive Kalman filter algorithm is introduced for online identification of the thermal resistances in a three‐node LPTN representing motor stator iron, stator winding and permanent magnet. The identification procedure requires a rotor temperature measurement, which is provided by an accurate pulse‐width modulation‐based estimation method. The proposed methodology is experimentally validated and applied to real‐time fault detection of the motor cooling system.