Novel Predictive Current Control of Permanent Magnet Synchronous Machine: DBC and FOC Algorithms.

Abstract:

Introduction: This work presents a novel predictive current control methodology for Permanent Magnet Synchronous Motors (PMSM). The study develops a dq-model of the PMSM in the rotor reference frame and establishes discretized motor models to facilitate the implementation of advanced control algorithms. The primary focus was on designing and evaluating a Dead-Beat Predictive Control (DBC) algorithm and comparing its performance with the conventional Field-Oriented Control (FOC) algorithm, which was developed due to its technological maturity and similarities with predictive methods. Methods: A complete closed-loop control system for PMSM was constructed using MATLAB/Simulink, incorporating the DBC and FOC algorithms. The FOC relied on PI controllers, whose tuning was optimized through a first-principles tuning algorithm. Both control schemes aimed to ensure that the dq-currents accurately tracked their references, thus controlling speed, torque, and rotor position effectively. Additionally, Space Vector Pulse Width Modulation (SVPWM) was implemented with an optimized switching sequence to minimize losses and harmonic distortion. The inverter switching strategy involved dividing the switching period into segments, applying zero and active vectors symmetrically, and analyzing the rectifier operation across different angle regions. Results: Simulation results demonstrated that both the DBC and FOC models effectively tracked current and speed references with minimal transients. The DBC exhibited a faster response, settling within approximately 0.12 seconds, 40% quicker than the FOC, which settled in 0.2 seconds without overshoot. During speed reversals, the DBC maintained sinusoidal phase currents with minor drops, and the d-axis current remained effectively at zero with deviations near 10^(-17). The system achieved a rapid speed response, reaching 500 RPM within 0.2 seconds, with the phase currents remaining balanced and sinusoidal. The FOC showed about 16% overshoot but still provided a very fast response, making it suitable for applications where rapid response is critical. Conclusion: The developed DBC and FOC algorithms successfully controlled PMSM, with the DBC demonstrating superior speed response and quicker settling times. Both methods maintained stable, sinusoidal phase currents and precise current tracking during transient conditions such as speed reversals. The simulation results validate that the proposed predictive control approach (DBC) offers improved performance over the conventional FOC, making it a promising candidate for high-performance motor control applications requiring fast dynamic response and minimal transient disturbance.