Modified Red Fox Optimizer With Deep Learning Enabled False Data Injection Attack Detection

Recently, power systems are drastically developed and shifted towards cyber-physical power systems (CPPS). The CPPS involve numerous sensor devices which generates enormous quantities of information. The data gathered from each sensing component needs to accomplish to reliability which are highly prone to attacks. Amongst various kinds of attacks, false data injection attack (FDIA) can seriously affects energy efficiency of CPPS. Current data driven approach utilized for designing FDIA frequently depends on distinct environmental and assumption conditions making them unrealistic and ineffective. In this paper, we present a modified Red Fox Optimizer with Deep Learning enabled FDIA detection (MRFODL-FDIAD) in the CPPS environment. The presented MRFODL-FDIAD technique mainly detects and classifies FDIAs in the CPPS environment. It encompasses a three stage process namely pre-processing, detection, and parameter tuning. For FDIA detection, the MRFODL-FDIAD technique uses multihead attention-based long short term memory (MBALSTM) technique. To improve the detection performance of the MBALSTM model, the MRFO technique can be exploited in this study. The experimental evaluation of the MRFODL-FDIAD approach was performed on standard IEEE bus system. Extensive set of experimentations highlighted the supremacy of the MRFODL-FDIAD technique.