A Data-Driven Detection strategy of False Data in Cooperative DC Microgrids

Distributed cooperative control strategies of DC microgrids (DCMG) reduce the detriment of communication delays, packet loss and link failure compared with centralized control. However, they are vulnerary to cyber-attacks. The operating objectives can be deviated by false data. Firstly, the adverse effects of false data are explained and modeled. A data-driven strategy based on linear regression is then proposed to remove the false data by offline learning and online judging of the transient process in DCMG without affecting the dynamic response. It successfully solves the problem of parameter selection of resilient control. Finally, the detection strategy is verified by detailed time-domain simulation.