Developing graphical detection techniques for maintaining state estimation integrity against false data injection attack in integrated electric cyber-physical system

The merging of power grid, information, and communication technology promotes the intelligent development of smart grid, which is also more prone to cyber attack threats. Especially, the intelligently designed False Data Injection (FDI) attacks severely disturb the normal management and state estimation operations in power system. In this paper, the graphical detection technology which uses Graph Network (GN) is developed for detecting tampered measurements without external knowledge and manual preprocess of historical data. To solve the detection of FDI attacks location issue from the diversified dimensionality in power systems, the Capsule Network combined with GN is developed, which can extract preserve the detailed properties around each note such as location, direction, connection, etc. To evaluate the superior performance of proposed method, the proposed detection technology is carried out through standard IEEE 30-bus and IEEE 118-bus systems. The simulation results demonstrate that the proposed method can detect FDI attacks accurately with different attack sparsity and magnitude of disturbances.