Composite control based on FNTSMC and adaptive neural network for PMSM system

Area of Science:

• Control Systems Engineering
• Electrical Engineering
• Robotics

Background:

• Permanent magnet synchronous motors (PMSM) are crucial in various applications but susceptible to performance degradation due to parameter mismatch and external disturbances.
• Traditional sliding mode control methods often suffer from high-frequency chattering, limiting their practical implementation.

Purpose of the Study:

• To propose a novel fixed-time non-singular terminal sliding mode control (NFNTSMC) method enhanced with an adaptive neural network (ANN) for PMSM systems.
• To improve the dynamic performance, robustness, and reduce chattering in PMSM control.

Main Methods:

• Design of a fixed-time non-singular terminal sliding mode controller for nominal PMSM systems.
• Integration of an adaptive radial basis function (RBF) neural network to approximate and compensate for lumped disturbances online.
• Stability and fixed-time convergence analysis using the Lyapunov method.

Main Results:

• The proposed NFNTSMC with ANN effectively approximates unknown disturbances, reducing the required switching gain.
• Significant reduction in sliding mode chattering was observed, improving system smoothness.
• Validated satisfactory dynamic performance and strong robustness through simulations and experiments.

Conclusions:

• The developed control strategy offers enhanced dynamic performance and robustness for PMSM systems.
• The integration of ANN with NFNTSMC provides an effective approach to mitigate chattering and improve control accuracy.
• This work presents a systematic framework for designing advanced controllers for PMSM applications.