Predictive Direct Voltage Control of Induction Motor With Mechanical Model Consideration for Sensorless Applications

Finite control set model predictive control is a simple solution for control of power electronics instruments. One of the deterrent problems with this method is the weighting factor in the cost function. Adjusting the weighting factor has a significant impact on achieving the desired response. However, the optimum value of the weighting factor can be changed by varying the operation point. Predictive direct voltage control uses a simplified cost function without weighting factor. In this method, the only term of the cost function is the last control variable, i.e., voltage of the inverter. The reference voltage is calculated by the deadbeat method, and the feasible voltages are examined in the cost functions. However, the weak robustness is the main drawback of the method because the reference is calculated on the basis of the deadbeat method with similar problem, especially in sensorless applications. In this paper, a robust predictive direct voltage control is proposed. The prediction model is improved by an model reference adaptive system method taking the mechanical model of the motor into consideration. The method is evaluated by simulations and experiments in a wide range of the operating points.