Predefined-Time Synchronization of Chaotic Systems of Permanent-Magnet Synchronous Generators via Neural Network Control

Chaotic behavior in power systems that are integrated with permanent-magnet synchronous generators (PMSGs) poses a significant threat to stability and security. Existing control methods often suffer from slow convergence, reliance on precise system models, or the inability to guarantee convergence within a predefined time. To address these issues, this paper develops a predefined-time synchronization control scheme for chaotic PMSG systems under unknown nonlinearities and external disturbances. First, an adaptive neural network with variable exponent coefficients is constructed to approximate unknown system dynamics online. Second, a predefined-time stability criterion is established, ensuring global convergence of synchronization errors within a user-specified time, independently of initial conditions. Third, the proposed controller achieves superior disturbance rejection without requiring prior knowledge of disturbance bounds. Numerical simulations demonstrate that the proposed method outperforms conventional finite-time control in convergence speed, control smoothness, and robustness to parameter variations—offering a practical and theoretically guaranteed solution for enhancing the stability of PMSG-based power systems.