Robust three-vector model predictive torque and stator flux control for PMSM drives with prediction error compensation

Abstract

The finite-set model predictive torque control (PTC) of permanent magnet synchronous motors has a strong dependence on motor parameters, where the three-vector PTC is more sensitive because it involves multiple voltage vectors, increasing the number of predictions. When the predictive torque and stator flux control is performed with mismatched parameters, the prediction errors of the torque and stator flux will be produced, resulting in the misjudgment of the optimal voltage vector and eventually leading to worse system control performance. In this work, a torque and stator flux prediction error reduction method (TVPTC) is proposed by online compensation in the three-vector PTC to improve parameter robustness. First, two loops of torque and stator flux are considered, and the prediction error models in three segments are established: delay compensation, torque and stator flux prediction, and vector action time calculation. Second, the main error sources and error transfer relations of each link are analyzed. Finally, the predicted values of the torque and stator flux at moment k + 1 are accurately compensated according to the online saved error values under the action of different voltage vectors. Simulation and experimental results confirm that the proposed TVPTC method can achieve reliable torque and stator flux prediction control when the motor parameters were mismatched.