Detection of False Data Injection Attack in Smart Grids via Interval Observer

This paper focuses on the detection of false data inject attacks (FDIAs) in smart grids. An interval observer-based detection scheme against the FDIAs is proposed, by considering the stealthy characteristics of FDIAs. Based on the constructed physical dynamics grid model, we design interval observer to estimate the interval state of physical dynamics accurately. To address the limitation of precomputed threshold, an interval residual-based detection standard is proposed. The residual evaluation functions and detection threshold in traditional attack detection methodologies are replaced by the interval residuals which can be regarded as the time-varying detection threshold. Finally, the effectiveness of the developed detection method is demonstrated by using a detailed study on the IEEE 37-bus smart grid system.