Robust Model Predictive Control With Simplified Repetitive Control for Electrical Machine Drives

Conventional model predictive control (MPC) for electrical drives relies on the accuracy of system model, thus, its performance will deteriorate due to disturbances such as parameter mismatch and current distortion. To improve the robustness of MPC, a compensated scheme with simplified repetitive control (SRC) is proposed in this paper. Since the prediction of MPC is corrected with the assistance of compensation, a much more reasonable output is provided to promote system performance. An open-loop structure of repetitive control (RC) with two resonant units is applied for the proposed SRC to guarantee stability and maintain high accuracy for tracking disturbances, and moreover, SRC responds fast and just needs half of the data memory used by an ideal RC. Additionally, an output filter is designed to correct the quantizing error of frequency, thus, the frequency adaption is enhanced. To validate the feasibility and the effectiveness of the proposed method, a compensated MPC for permanent-magnet synchronous machine drive system is utilized. Both simulation and experiment results prove significant performance enhancement of the SRC-based robust MPC, and compared with the conventional MPC, it provides nearly 30% reduction of total harmonic distortion (THD) in stator phase current.