Model Predictive MRAS Estimator for Sensorless Induction Motor Drives

This paper presents a novel predictive model reference adaptive system (MRAS) speed estimator for sensorless induction motor (IM) drives applications. The proposed estimator is based on the finite control set-model predictive control (FCS-MPC) principle. The rotor position is calculated using a search-based optimization algorithm which ensures a minimum speed tuning error signal at each sampling period. This eliminates the need for a proportional–integral (PI) controller which is conventionally employed in the adaption mechanism of MRAS estimators. Extensive experimental tests have been carried out to evaluate the performance of the proposed estimator using a 2.2-kW IM with a field-oriented control (FOC) scheme employed as the motor control strategy. Experimental results show improved performance of the MRAS scheme in both open- and closed-loop sensorless modes of operation at low speeds and with different loading conditions including regeneration. The proposed scheme also improves the system robustness against motor parameter variations and increases the maximum bandwidth of the speed loop controller.