Fractional-Order Finite-Time Control of Cyber-Physical Power System With Time-Delayed Switching and Distributed DoS Attacks Using RBF-Neural Network

This paper presents the design and application of an adaptive fractional-order (FO) cyber-resilient frequency protocol to counteract aggregated plant uncertainties and maintain grid stability and security in the presence of malicious hybrid cyberattacks. A FO-integral terminal sliding mode control (FO-ITSMC) framework is implemented to achieve effective transient shaping and to guarantee power-frequency equilibrium in large-scale cyber-physical power systems. To further strengthen the controller functionality, a radial basis function neural network (RBF-NN) compensator, employing a FO steepest-descent algorithm with momentum learning strategy, is designed and integrated with FO-ITSMC to enable faster cyber-mitigation and improved resilience. Through qualitative analysis, the practicality and effectiveness of the proposed controller are demonstrated in comparison with backstepping control, TSMC, and ITSMC, under identical operating scenarios. The simulation results confirm the superiority of the proposed controller in terms of chattering suppression, cyberattacks mitigation, improved transient response, and better steady-state accuracy over the above-mentioned methodologies in the face of hybrid cyberattacks. Furthermore, the boundedness of the estimation error and finite-time closed-loop stability have been rigorously analyzed and guaranteed using Lyapunov theory.