Continuous-Control-Set Model-Free Predictive Current Control With Voltage Correction of Five-Phase PMSM

Abstract

To lessen parameter dependence and suppress current harmonics, a continuous-control-set model-free predictive current control (MFPCC) strategy with voltage vector correction and duty cycle optimization is proposed for five-phase permanent magnet synchronous machine (PMSM) drives. Firstly, the MFPCC method with an ultralocal model is developed based on the mathematical model of a five-phase PMSM and the virtual voltage vector (VVV). The lumped parameters of the ultralocal model are estimated with current and voltage variations rather than observers. Thus, the stability of the five-phase PMSM drive is not affected. Secondly, the reason for the current distortion is investigated by analyzing the numerator and the denominator of the division operation in uncertain terms identification. A voltage vector correction mechanism is accordingly designed based on the power factor angle to decrease current distortions. Meanwhile, the duty cycle of the selected VVV is optimized to further reduce the tracking error. The proposed strategy not only theoretically eliminates the impact of machine parameters but also improves steady-state performance. Finally, the effectiveness of the MFPCC is validated by experiments.