A Novel False Data Injection Method Targeting on Time-series analysis in Smart Grid

With the development of smart grid, it has been facing aggravating cyber threats. The cyber security of smart grid relies on the identification of latent threats, including false data injection (FDI). Existing researches on FDI mostly target on the network structure of smart grid, while ignoring its time-series features, which are effectless under the detection strategies based on time series analysis. This paper proposes a novel FDI method based on Time-series Generative Adversarial Networks, which generates false data by the adversarial process of generating against time-series analysis. The simulation experiments on SimBench dataset validates the proposed method's destructiveness and stealth against the bad data detector and time-series detection. The proposed method demonstrates a new threat situation under the detection strategy based on time series analysis. Through the proposed FDI method, we will have effective data support for the development of the time-series detection strategy and the defense ability in smart grid.