A two-stage optimal vector selection method for predictive torque control of a three-level VSI driven induction motor

Abstract

Conventional Finite State Predictive Torque control (FS-PTC) for three-level Neutral Point Clamped voltage source inverter (3L-NPC VSI) uses 27 voltage vectors for prediction and actuation. Using all voltage vectors for the prediction loop is not an ideal method as it increases computational burden. This paper proposes a less complex prediction loop method with selected number of voltage vectors for FS-PTC of a three level NPC driven induction motor. The number of voltage vectors is reduced based on a two-stage optimal vector selection algorithm. In the first stage, the algorithm considers the VSI as two-level and selects the most favourable long vector. In the second stage, among the short and the medium voltage vectors closest to the long vector which is selected in the first stage, the optimum vector is selected for prediction. Compared to 27 voltage vectors based prediction, this algorithm evaluates 15 selected vectors in total for prediction and actuation. The effectiveness of the proposed algorithm in terms of speed, torque and flux responses and capacitor voltage balancing is presented through simulation results from MATLAB/Simulink. Computational time is measured from a real-time simulation implemented on dSPACE DS1104 platform. The results show that the proposed method reduces the computation time significantly (by about 45%), while the dynamic and steady-state performances of the motor drive are retained similar to the conventional FS-PTC.