Neural Network and Fractional-Order Control of Permanent Magnet Synchronous Machine (PMSM)

The Permanent Magnet Synchronous Machine (PMSM) is known for its high efficiency and ability to maintain consistent power across a wide speed range in motor drive systems. However, achieving this performance requires robust and reliable control systems. This paper presents a robust control approach proposing fractional-order control and neural network techniques as a replacement for traditional Proportional-Integral (PI) controllers in PMSM control systems. Using the Simulink Design Optimization Toolbox, an optimized fractional-order Proportional-Integral-Derivative (FOPID) controller was installed on the PMSM model. Simultaneously, the neural network controller employed an error correction learning mechanism to minimize the deviation between the system's output and the desired reference. Simulation results in MATLAB/Simulink demonstrate superior performance of the proposed control system compared to the conventional PI controller, achieving enhanced precision and stability in PMSM operations.