Stator Temperature Estimation Strategy Under Three-Phase Winding DC Current Injection

The stator temperature of a Permanent Magnet Synchronous Motor (PMSM) can be derived directly from its stator resistance. However, conventional online stator resistance identification methods based on voltage equations are susceptible to non-ideal factors such as inverter nonlinearity, rotor position estimation errors, and parameter coupling. To address this, this paper proposes a stator temperature estimation method based on three-phase winding DC current injection. This method first establishes an independent stator resistance identification model by injecting DC signals into specific synchronous rotating axes, effectively circumventing the influence of rotational speed, position errors, and parameter coupling inherent in conventional models. Furthermore, the underlying mechanism of inverter nonlinearity on resistance and temperature identification accuracy is analyzed. Critically, to avert potential destabilizing effects from online inverter nonlinearity compensation during high-speed operation, a compensation strategy utilizing moving-average filtering of the inverter nonlinear error voltage is implemented, significantly enhancing the accuracy of online stator resistance and temperature identification.