Towards Real -Time Distinction of Power System Faults and Cyber Attacks

This paper presents a methodolop, to distinguish between three-phase faults and GOOSE cyber attacks, aimed at opening the circuit breakers in the power grid. We propose a scheme that utilizes Phasor Measurement Unit (PMU)-enabled monitoring of power grid states, and communication network packet logs in the substation. In this scheme, by leveraging both cyber and physical data correlations and applying a Seasonal Autoregressive Moving Average (SARNIA) model, we successfully distinguish between 3 -phase faults and cyber attacks. The proposed scheme is tested using the benchmark IEEE 9 -bus system, and can distinguish cyber attacks from faults in less than 0.2s. This demonstrates the usefulness of the proposed scheme for power system cyber security analytics.