Transforming cyber security in modern power grids: the synergistic application of complex recurrent spectral network to address vulnerabilities and ensure resilience

This paper proposes transforming cyber security in modern power grids through the synergistic application of complex recurrent spectral network to address vulnerabilities and ensure resilience (TCS-MPG-CRSN) to address cyber threats in power grids. It utilises an adaptive two-stage unscented Kalman filter (ATSUKF) for data quality assessment, cleaning, and normalisation. The pre-processed data is fed into the complex recurrent spectral network (CRSN) for identifying and classifying cyber-attacks. The TCS-MPG-CRSN model was implemented in Python and achieved 20.77%, 34.55%, and 26.11% higher accuracy, 33.47%, 24.55%, and 22.18% higher precision, and 19.22%, 14.39%, and 24.66% higher recall compared to cyber-security in power systems using meta-heuristic and deep learning algorithms (CS-PS-ANN), classification of intrusion cyber-attacks in smart power grids using deep ensemble learning with metaheuristic-based optimisation (ICA-SPG-GWO), and attack graph model for cyber-physical power systems using hybrid deep learning (AGM-CPPS-DCNN).