A Data-driven Multivariable Adaptive Cybersecurity Framework for Isolated Microgrids

Dynamic attack estimation in isolated microgrids (IMGs) is a significant concern due to their presence in various operating environments and the advancement of digital technologies. This paper introduces a dynamic online cybersecurity framework based on a data-driven approach for adaptive estimation and mitigation (EM) of IMG attacks with unknown inputs. The proposed framework utilizes a generative deep learning approach to process an unknown set of inputs, which is then used for EM of attacks in IMGs. The efficacy of the proposed framework is tested using a two-area connected IMG simulated in MATLAB. The performance is measured for simultaneous and separate occurrences of different forms of false data injection attacks. The results demonstrate that the proposed framework can provide adaptive EM of attacks with fast training speed, making it suitable for real-time applications and scenarios with limited computational resources.