An Improved Model Free Predictive Current Control for PMSM with Current Prediction Error Variations

The conventional model predictive current control is a model-based control method, and the accuracy of the predicted currents is affected by motor parameters such as flux linkage, inductance, and resistance. To get rid of model parameters dependencies, a model-free predictive current control (MFCC) was proposed before, which can improve the parameter robustness without utilizing any knowledge of the initial motor parameters. However, the stagnant current update detection is one of the main problems that limit the current predictive performance. To solve this problem, a current prediction error model according to the contiguous instant current error variations is proposed to reconstruct the surface-permanent magnet synchronous motor (SPMSM) model in this paper. Afterwards, a novel MFCC method with the online parameter identification is developed. This method takes advantage of mathematical relationships in the current prediction error model, and the motor parameters can be updated within each period to improve prediction accuracy. Simulation and experimental results verify that this proposed MFCC method can significantly reduce the stagnation effect and improve MFCC performance under different parameter disturbances.