Detection of false data injection attacks in smart grids using Recurrent Neural Networks

False Data Injection (FDI) attacks create serious security challenges to the operation of power systems, especially when they are carefully constructed to bypass conventional state estimation bad data detection techniques implemented in the power system control room. This paper investigates the utilization of Recurrent Neural Networks (RNN) as a machine learning technique to detect these FDI attacks. The proposed detection algorithm is validated throughout simulations of FDI in power flow data over the span of five years using IEEE-30 Bus system. The simulation results confirm that the proposed RNN-based algorithm achieves high accuracy in detecting anomalies in the data, by observing the temporal variation in the successive data sequence.