Distribution Systems Real-Time Monitoring Cybersecurity: A Parameter Error Correction Model Against False Data Injection

As smart grid technologies are deployed, the many advantages of new measurement, control, and analysis come with added technical challenges. Specifically, the digitalization of the power grid and the increasing dependence on communication systems makes network real-time monitoring more vulnerable to cyber-attacks. Cyber-attacks, if not detected and corrected accurately, can lead to misinformation to the system operator. Current state-of-the-art models for real-time cybersecurity monitoring hypothesize that the measurement model is correct, without error. Although this assumption might be acceptable for systems and devices that are not dependent on communication networks, this can be considered a strong hypothesis for real-time monitoring of power grids. False data injection attacks on the measurement model are also possible. This work presents a parameter correction model against false data injection attacks. False data injection attacks on measurements and measurement models are simultaneously considered. A chi-squared hypothesis test is used for detection of cyber-attacks, while a normalized composed measurement error test is used for cyber-attack identification. The parameter correction model is then used if a modeling error is identified; otherwise, a measurement correction model is used if a measurement error is identified. The easy-to-implement model, built of the classical quasi-static state estimator, without hard-to-design parameters, suggests potential for real-life applications.