Computation Efficiency and Robustness Improvement of Predictive Control for PMS Motors

High computational burden and model dependence of the conventional model predictive control of permanent magnet synchronous motors are regarded as the two shortcomings of this attractive control scheme. A new method is proposed in this paper to overcome the mentioned drawbacks. The method evaluates a cost function for only three admissible inverter voltage vectors in each sampling period for a two-level inverter without deteriorating motor performance. Furthermore, the prediction of the torque for the next sampling period(s) is achieved without using the motor parameters. The proposed finite control set predictive control is evaluated through the simulation and experimental tests for a permanent magnet synchronous motor. In addition, the motor performance under the proposed control is compared with those under two other predictive control methods. The results show that parameter mismatch does not affect the prediction of the torque under the proposed control method, thus resulting in improved robustness of the motor drive system against motor parameter variations.