Local detection of replay attacks and data anomalies on PMU measurements of smart power grids via tracking critical dynamic modes

This work proposes an approach based on critical dynamic modes to detect four types of phasor measurement unit (PMU) data anomalies in smart power grids. Smart power grids as modern power systems involve new information and communication technologies (ICTs), relying on real-time data, like data of PMUs. Due to the dependence of PMUs on communication technology, PMUs are prone to different data anomalies and false data injection attacks, such as replay attack, which has been rarely considered in previous works. Since detecting abnormal data is necessary for the proper function of the power system, most detection methods have been data-based approaches that may need a large amount of data and lead to more complexity. This paper proposes an online detection approach based on the minimum number of dynamic modes for the local detection of a replay attack and three types of bad data injection on PMU measurements. For this purpose, a distributed modeling of the power system is considered. Then, a replay attack and bad data injections are detected locally by tracking critical dynamic modes. The proposed approach can detect simultaneous data anomalies on more than one PMU. The effectiveness and accuracy of this approach are evaluated through simulations on the 10-machine New England 39-bus power system.