A novel dynamic prescribed performance fuzzy-neural backstepping control for PMSM under step load

Area of Science:

• Electrical Engineering
• Control Systems
• Robotics

Background:

• Permanent magnet synchronous motor (PMSM) systems often face challenges with time-varying parameters and input constraints, especially under step load disturbances.
• Traditional prescribed performance functions can lead to issues like exceeding predefined errors, control singularity, and system instability during load changes.

Purpose of the Study:

• To propose a dynamic prescribed performance fuzzy-neural backstepping control approach for PMSM systems.
• To address performance degradation caused by nonlinear time-varying parameters and input constraints.
• To improve the transient and steady-state performance of PMSM systems under step load.

Main Methods:

• A novel finite-time asymmetric dynamic prescribed performance function (FADPPF) was developed to overcome limitations of traditional methods.
• A backstepping controller was designed integrating a speed function (SF) and a fuzzy neural network (FNN).
• The FNN approximates uncertain nonlinear system functions, while the SF and FADPPF ensure system performance.

Main Results:

• The proposed FADPPF demonstrated dynamic self-adjusting capabilities and effectiveness under step load conditions.
• Lyapunov analysis confirmed the stability of the developed control strategy.
• Simulation results validated the feasibility and superiority of the proposed control scheme compared to existing methods.

Conclusions:

• The dynamic prescribed performance fuzzy-neural backstepping control approach effectively manages PMSM systems with time-varying parameters and input constraints.
• The novel FADPPF is crucial for maintaining system stability and performance during load variations.
• The integrated control strategy offers a robust and superior solution for PMSM control applications.