Deep Reinforcement Learning-Based Controller for Field-Oriented Control of SynRM

Synchronous reluctance motors offer several advantages that make them suitable for use in electric vehicle traction systems. Motor-drive systems constitute the most significant share of the energy consumption of electric vehicles. Controller performance is essential for achieving accurate, stable, efficient, and safe motor control. Conventional controllers, such as PID controllers, remain popular owing to their simplicity, ease of implementation, and effectiveness in many control applications. However, advanced control techniques may offer better performance and robustness for more complex or challenging control tasks. Successful studies employing deep reinforcement learning have been conducted across various control applications, including motor control. Because deep reinforcement learning is a relatively recent approach to control, its utilization in motor control remains a subject of ongoing research and development. This paper presents the application of deep reinforcement learning as a speed-control strategy for synchronous reluctance motors. The simulation results are presented to demonstrate the effectiveness of the proposed deep reinforcement learning control strategy. These results highlight the improved robustness of the system to speed changes and load disturbances and demonstrate the superior performance achieved in synchronous reluctance motor speed control compared with the conventional PI control method.