Artificial-Based Sensorless Deadbeat Control for PMSM in Electric Vehicle Applications

This research work designs an artificial-based sensorless deadbeat control (ABS-DBC) method for permanent magnet synchronous motors (PMSMs). The proposed strategy encompasses the merits of deadbeat control (DBC) with artificial neural network (ANN), to ensure adequate control performance of PMSM drive systems. Combining these two methods improves robustness, enhanced steady-state and dynamic response by the DBC and learning capabilities of artificial intelligence, respectively. The DBC contributes to fast dynamic response, while the ANN technique augments the system's efficacy to tolerate nonlinear behavior inherent in PMSM drives. The ANN is trained to effectively estimate the rotor speed and rotor angle under some severe conditions, for instance, low speed region. The proposed ABS-DBC method is then verified through PSIM simulation tool, exhibiting superior performance in terms of precise tracking capability covering low, medium, and high speed, and reduced estimated speed ripples compared to classical control methods.