A Speed-Sensorless FS-PTC of Induction Motors Using Extended Kalman Filters

A sensorless finite-state predictive torque control (FS-PTC) strategy uses stator current, estimated stator and rotor flux, and estimated rotor speed to predict stator flux and torque. Direct application of measured stator currents and using a noisy estimated speed in the prediction model degrade the steady-state performance in terms of higher current total harmonic distortion (THD), torque ripple, and flux ripple, particularly at low speeds. This paper proposes an extended Kalman filter (EKF)-based, which is a promising state observer, improved prediction model of sensorless FS-PTC for induction motor drives. The EKF has been used to estimate rotor speed, rotor/stator flux, and stator currents accurately. The estimated stator currents, instead of measured currents, are fed back to the prediction model, and thus, small stator current THD is confirmed. Depending on the commanded speed, either the rotor current model or the open-loop stator voltage model is proposed for the EKF to achieve better performance in a wide speed range, including the field-weakening region. The proposed control system has been verified experimentally, and excellent torque and flux responses, robustness, and stable operation at lower and higher speeds have been achieved.