An Improved Deadbeat Predictive Current Control Based on Exact Compensation of Inverter Nonlinearities

The deadbeat current predictive control (DPCC) of PMSMs relies on an accurate motor drive system model and is susceptible to unmodeled factors in the drive system. As the most significant unmodeled disturbances, inverter nonlinearity cause problems such as steady-state current static error, current harmonics, and tracking lag in DPCC. Therefore, in this paper, the inverter nonlinearity is precisely modeled by comprehensively considering the dead time, conduction delay, conduction voltage drop, and parasitic capacitance. Then, the proposed method is experimentally verified on the Zynq platform. The results show that, compared with the traditional DPCC, the proposed improved method can effectively eliminate current deviation and achieve ideal current predictive control.