Improved ESO and Parallel Repetitive Control for Model-Free Predictive Current Control of PMSM

Traditional difference-free predictive current control (DPCC) relies on accurate models and cannot handle the unmodeled DC and AC disturbances present in surface-mounted permanent magnet synchronous motor (SPMSM) drive systems. Model-free predictive current control (MFPCC) based on an extended state observer (ESO) with a hyperlocal model reduces sensitivity to certain motor parameters and enhances system robustness. However, this approach is limited by the ESO’s bandwidth, resulting in poor observation of AC disturbances under various operating conditions and leading to non-negligible current harmonics in the system. This paper presents an enhanced MFPCC method based on an improved expanded state observer (IESO) coupled with a parallel switchable repetitive controller (SRC). This approach uses the IESO to decompensate the dominant sixth harmonic, while embedding a repetitive controller to suppress the twelfth harmonic amplified by the IESO during low-to-medium speed operation. Simulation experiments and comparisons demonstrate that this method improves dynamic response and suppresses current harmonics.