Defensing Against Measurement Attacks on Sub-Transmission Level

This paper develops a correlation index for Anomaly Detection System (ADS) to detect and identify wide area measurement attacks at sub-transmission level. Inspired by forensic analysis, we study the pattern of measurement attacks at sub-transmission power systems by developing an index of correlated substation attacks. The attacking strategy is reconstructed as a one stage two-level optimization problem that models the interaction between falsified measurements and power system operation decisions. The proposed correlation index is computed as the "attacker's decision under limited resources" by solving the mixed integer problem transformed from the attacking model. We show that this correlation index may greatly improve ADS detection rate without increasing false alarms. With the proposed correlation index, the existing protection scheme is reinforced against correlated attacks at sub-transmission power systems. The effectiveness of the developed correlation index is demonstrated through numerical simulations of the IEEE 39-bus system.