Real-time prediction of rotor temperature with PMSMs for electric vehicles

The problem of wide variation and challenge in direct measurement of the rotor temperature of permanent magnet synchronous motors can lead to decreased excitation performance of the permanent magnets or even permanent demagnetization when the temperature is too high, resulting in reduced thermal safety and reliability of the motor. The paper proposes a real-time rotor temperature prediction method that does not require additional hardware and has a low computational load, considering factors such as power-off duration, rotor temperature at power-up, and rotor temperature rise. Firstly, the initial temperature of the rotor is estimated based on the ambient temperature and downtime. Then, the change of the rotor temperature is estimated based on the motor loss and heat dissipation model, and finally, the rotor temperature at the next moment is predicted. The proposed rotor temperature real-time prediction method has been validated through bench and on-road tests, demonstrating its real-time and accurate performance. This has enhanced the thermal safety and extreme condition output capability of the motor, laying the foundation for achieving efficient and reliable operation of the motor.