Sensorless Estimation for Stator Winding Temperature of Automotive Electric Motors Based on Sequential Current Pulse Injection

For ensuring the electric motor safety of electric vehicles (EVs), the temperature sensors are embedded in the stator windings for diagnosis and monitoring, but constrained by their lifetime and vulnerability. To solve this issue, a sensorless initial temperature estimation method for winding is proposed. The innovation lies in that the initial temperature estimation is based on a combination of thermal models of the electric motor and traction inverter. Through the sequential current pulse injection, the electrical characteristics of the traction system circuit were extracted and the winding initial thermal state was identified according to the designed algorithm. This method is highly portable and can provide an initial value for various temperature estimation models to improve the model accuracy and its rate of convergence. It is suitable for all types of electric motors with low phase resistance and driven by a three-phase full bridge, especially oil cooling. Finally, the effectiveness of the presented model is verified by the simulations and experimental results.