A Steady-State-Based Predictive Current Control Method for CSI-PMSM Systems

To suppress AC-side oscillation and improve steady-state current quality in current-source-inverter-fed permanent magnet synchronous motor (CSI-PMSM) systems, this paper proposes a predictive current control method based on steady-state characteristics. An equivalent model of the CSI-PMSM system is developed in the synchronous rotating dq reference frame, and the steady-state characteristics of the filter capacitor current are analyzed. The analysis shows that the capacitor current is generally nonzero under steady-state operation, whereas its deviation from the steady-state reference component should converge to zero. Based on this property, a discrete predictive model is constructed, and the stator current tracking error and the capacitor current deviation are incorporated into the cost function to achieve coordinated current tracking and LC oscillation suppression. In addition, a deadbeat preselection and a local finite-candidate optimization scheme are adopted to reduce the online computational burden. Experimental results obtained from a 3.7 kW CSI-PMSM platform demonstrate that, compared with conventional multi-loop PI control, the proposed method significantly reduces dq-axis current ripples and DC-link current fluctuations, while decreasing the stator current total harmonic distortion from 9.87% to 2.25%. These results verify the effectiveness and engineering feasibility of the proposed steady-state-consistent predictive current control method.