CAD-IGMS: Cyber attack detection in an intelligent grid management system via the deep learning enhancing grid reliability

Cyber security (CS) has emerged as a critical component in Intelligent Grid Management Systems (IGMS) due to the integration of Supervisory Control and Data Acquisition (SCADA) systems in modern power grids. However, this increased dependence on connectivity and technology leaving SG communication networks vulnerable to widespread attacks. To address these challenges, a novel Cyber Attack Detection system for IGMS (CAD-IGMS) using deep learning has been proposed. Renewable energy sources feed power through converters to a Load Source Inverter (LSI), with SCADA handling data acquisition and control. Features are extracted using a Cascaded Autoencoder and a Self-Attention based Convolutional Neural Network (SACNN) is utilized to detect and classify cyber-attacks in smart grids. Implemented in MATLAB, CAD-IGMS demonstrates superior performance, achieving accuracy improvements of 15.8 %, 17.2 % , and 20.6 % over Artificial Feed-forward Network (AFN), Genetically Seeded Flora Transformer Neural Network (GSFTNN), and Convolutional Neural Network based Long short term memory (CNN-LSTM), respectively.