Randomised Sampling-Based False Data Injection Attacks Against Power System State Estimators

The rapid integration of renewable energy sources and electric vehicles into the traditional power grid introduces significant uncertainty and complexity. Accurately monitoring and understanding these uncertainties is crucial for efficient grid operation. The power utilities leverage supervisory control and data acquisition (SCADA) systems for automation, thus enhancing grid reliability by incorporating state estimation (SE) algorithms that play a critical role in control centres by providing real-time estimates of the grid's operating conditions. Integrating information and communication technologies further boosts grid efficiency. However, this integration also introduces new vulnerabilities to cyberattacks. Recently, with an efficient undermining of the critical vulnerabilities of the meters, remote terminal units (RTUs), communication networks, etc. adversaries have shown their potential to disrupt the normal operation of the grid, thus leading to grid outages and massive socio-economic impact. This work presents an efficient formulation of one such class of attacks namely the false data injection attack (FDIA) which can effectively inject inaccurate data within the system measurements whose errors are assumed to be uncorrelated zero mean Gaussian noises, without getting detected by the conventional bad data detectors (BDDs), thus leading to biased estimation of the operating states of the grid. This work focuses on investigating data-driven randomised sampling-based FDIAs as a deceptive cyberattack method on the power grid. The impact of these attacks on the test network (IEEE 14 bus) for the nonlinear SE algorithm depicts its performance under varying noise margins and the strength of attacks.