On malicious data attacks on power system state estimation

摘要: The problem of detecting and characterizing impacts malicious attacks against smart grid state estimation is considered. Different from the classical bad data detection for estimation, injected by an adversary must take into account carefully designed capable evading conventional detection. A Bayesian framework presented characterization fundamental tradeoffs at control center adversary. For center, a detector based on generalized likelihood ratio test (GRLT) introduced compared with schemes. adversary, tradeoff between increasing mean square error (MSE) vs. probability being detected characterized. heuristic design worst attack.