Detection Localization and Recovery of False Data Injection Attacks on Power Grids Based on SA-DCNN and AE-LSTM

The power system has gradually evolved into a power information physical system with deep coupling of power and information systems as advanced information and communication technology has advanced. False data injection attack (FDIA), an emerging form of information attack, can quickly affect the physical system via a strongly coupled system, tampering with system operation data without being detected by the bad data detection module, posing a serious threat to the safe and reliable operation of the power system. As a result, this paper employs FDIA of power grid as the research object, and first constructs FDIA base data to meet the actual conditions based on the AC power system where part of the network information is known. Then, a self attention-deep convolutional neural network based FDIA detection and localization algorithm is proposed, and the attacked data are recovered using an autoencoder-long and short-term memory neural network based on the correlation of the measurement data itself. Finally, a thorough experimental validation of the FDIA construction, detection, localization, and recovery algorithm is carried out using the IEEE 14 and 118 systems. The results show that the proposed comprehensive algorithm has excellent performance and can be used to effectively defend against FDIA against power grids.