A decentralized unknown input observer-based false data injection detection framework for DC microgrids

Dc microgrid cluster are widely adopted due to flexible utilization of renewable energy sources. However, their reliance on communication networks, make them vulnerable to stealthy cyber-attacks which can disrupt system dynamics and lead to unstable microgrid operation. To address this critical issue, a decentralized unknown input observer (DUIO) based false data injection (FDI) attack detection framework is proposed. First, a comprehensive modelling of DC microgrid cluster is developed considering line resistance, loads, grid forming converter, unknown disturbances. Subsequently, an unknown input observer (UIO) is carefully designed to monitor system dynamics. When malicious data is injected into measurement, UIO incorporates residual checking technique along with threshold comparing procedure to identify cyber-attack. To validate the framework, microgrid model tested in MATLAB environment against three distinct categories of FDI attacks and load changing scenario. The results show robust performance in terms of cyber-attack detection in DC microgrids which is essential for stable grid operations.