Secure Control of DC Microgrids under Cyber-Attacks based on Recurrent Neural Networks

DC microgrids have advantages when comparing to AC microgrids, i.e., more efficiency and less complexity to control. To control the DC microgrids, communication networks, and measurement devices such as voltage and current sensors are needed to be implemented to measure desired data and transmit them. Because of the existence of the communication network, the DC microgrids can be vulnerable to cyber-attacks. False data injection attacks (FDIAs) are a type of cyber-attacks that try to inject false data into the system. In DC microgrids, FDIAs can destruct the control application of the system by destroying the control objectives, i.e., current sharing and voltage regulation. This work introduces a method based on recurrent neural networks to detect and mitigate FDIAs in DC microgrids. The proposed strategy is based on a reference tracking application and it can also estimate the value of the false injected data. It is important to note that the proposed strategy is a decentralized approach and it needs only the local data and it does not need the information from other units. The performance of the proposed method is examined under different conditions. The obtained results prove the effectiveness of the proposed cyber-attack detection and mitigation strategy.