An optimal cyber–physical attack strategy on DC microgrids

Abstract

This paper proposes an optimal attack strategy against Direct Current MGs (DCMGs). In the proposed strategy, attackers first eavesdrop on the data communicated between a target converter and the MG Control Center (MGCC) to model (i) the converter by its state-space representation using system identification techniques, and (ii) the rest of the system by its Thevenin equivalent circuit using the least-square method. Afterwards, the obtained model is used in an optimal control framework to design a False Data Injection Attack (FDIA) that controls the converter’s output to achieve the attackers’ objective, e.g., triggering the overcurrent or under-voltage protection of the converter. The proposed framework fulfills the objective while minimizing the rate of change of the converter’s voltage and the required control effort during the FDIA (to keep the attack stealthy) in the presence of the attack constraints. Such an attack strategy provides cyber-security experts with insight into potential attacks and their impacts on the operation and stability of DCMGs, and helps them to design and test their cyber-security measures more effectively. Simulation results in MATLAB/Simulink corroborates the effectiveness of the proposed attack strategy.