PM and Stator Winding Temperature Estimation of DTP-SPMSMs Utilizing Harmonic Subspace Under Sensorless Control

Permanent magnet (PM) flux linkage and stator resistance are critical thermal indicators for online monitoring PM and winding temperatures of the dual three phase surface-mounted permanent magnet synchronous machines (DTP-SPMSMs). However, both estimations are compromised by position error under sensorless control. This paper proposes a simple online estimation method of PM flux and resistance for the DTP-SPMSMs utilizing harmonic subspace to eliminate the influence of position error under vector space decomposition model. Firstly, when a pair of positive and negative position-offsets are simultaneously injected into two sets of DTP windings, the voltage responses in harmonic subspace can be utilized to construct the PM flux estimation model that is independent of position error, resistance, stator inductance, and inverter nonlinearity. Secondly, by controlling the ratio of dq-axis currents in harmonic subspace to be opposite to that in fundamental subspace, the resistance can be independently estimated without influence of position error, PM flux, inductance, and inverter nonlinearity in harmonic subspace. Consequently, the estimated PM flux and resistance can be utilized to monitor PM and winding temperatures. Moreover, both estimations have no influence on torque output due to the utilization of the characteristics of DTP windings. The proposed method has been validated experimentally.