Towards Detecting Long-term False Data Injection Attacks: An Entropy-based Approach

Longxing Jin,Yunqi Yan,Ying Chen,Wei Fu,Danlong Zhu,Zhisheng Wang
DOI: https://doi.org/10.1109/PandaFPE57779.2023.10140733
2023-01-01
Abstract:False data injection (FDI) attacks have been regarded as serious cybersecurity threats to power systems. This paper proposed an attack detection method, which involves measuring uncertainties in both physical and cyber systems caused by FDI attacks. Using a new concept called Cyber-Physical System Entropy (CPSE), two indicators are developed: the grid CPSE (G-CPSE) and the bus CPSE (B-CPSE). An online detection method is coined for discovering long-term FDI attacks according to spatial and temporal increments of G-CPSE and B-CPSE. Test results on the IEEE 39-bus system show that both model-based and learning-based FDI attacks can be detected effectively and sensitively using G-CPSE and B-CPSE.
What problem does this paper attempt to address?