Real-time detection of malicious PMU data

The growing installation of phasor measurement units (PMUs) provide grid operators wide-area situational awareness while introducing additional vulnerabilities to power systems from the cyber security point of view. This paper presents an online method to detect ongoing contingencies in the system and bad data injection on its PMU network. To do so, the principal component analysis is applied to leverage the spatial and temporal correlations among the synchrophasor data. Pattern match and data reconstruction are proposed to identify incident types and find their most possible locations. Case studies are carried out on a 150-bus system to demonstrate the effectiveness of the proposed scheme.