Neural network-based adaptive dynamic surface control for permanent magnet synchronous motors.

Area of Science:

• Control Systems Engineering
• Artificial Intelligence in Engineering
• Electrical Machines

Background:

• Permanent magnet synchronous motors (PMSMs) are crucial in modern industrial applications.
• Controlling PMSMs is challenging due to parameter uncertainties and load torque disturbances.
• Traditional control methods can suffer from complexity explosion in backstepping designs.

Purpose of the Study:

• To develop a novel adaptive dynamic surface control (DSC) strategy for PMSMs using neural networks (NNs).
• To address parameter uncertainties and load torque disturbances in PMSM drive systems.
• To simplify the controller structure and reduce the number of adaptive parameters.

Main Methods:

• Utilizing neural networks (NNs) to approximate unknown nonlinear functions within the PMSM drive system.
• Constructing a novel adaptive DSC to mitigate complexity issues inherent in backstepping.
• Implementing a simplified adaptive neural controller with a single adaptive parameter.

Main Results:

• The proposed adaptive neural DSC effectively handles parameter uncertainties and load torque disturbances.
• The controller structure is significantly simpler compared to existing methods.
• Guaranteed convergence of tracking errors to a small neighborhood of the origin was demonstrated through simulations.

Conclusions:

• The developed adaptive neural DSC offers an effective and simplified control solution for PMSMs.
• The technique shows potential for practical implementation in PMSM drive systems.
• Simulation results validate the performance and robustness of the proposed control strategy.