Fuzzy Neural Network‐Assisted Super‐Twisting Fractional‐Order Terminal Sliding Mode Frequency Control of Cyber‐Physical Power System

ABSTRACT This work addresses an adaptive super‐twisting fractional‐order terminal sliding mode control (AST‐FOTSMC) framework for a cyber‐physical power system under unknown malicious distributed denial‐of‐service (DDoS) and false‐data injection attacks. The proposed frequency control architecture effectively integrates fractional‐order calculus with a super‐twisting switching law to balance chattering suppression, fast finite‐time state convergence, design flexibility, and cyber‐resilience. A dual‐loop fuzzy neural network (DLFNN) is additionally formulated to approximate mismatched plant uncertainties, including unmodelled dynamics, exogenous inputs, and parametric perturbation. Adaptive tuning of the DLFNN parameters enhances its approximation ability, enabling accurate estimation of unknown dynamics. An in‐depth stability analysis is conducted using the Lyapunov approach in conjunction with Barbalat's corollary. The superiority of the proposed frequency control method is qualitatively demonstrated under malicious cyber‐attacks and mismatched uncertainties, outperforming other reported control methodologies in terms of resilience and cyber‐tolerance. Overall, the results confirm that the proposed DLFNN‐assisted AST‐FOTSMC framework achieves stable, cyber‐resilient power‐frequency regulation with smoother control effort. Lastly, the practicality of the proposed control algorithm in the cyber‐physical environment has been verified through real‐time implementation and testing on the OPAL‐RT 4610XG platform.