Data-Driven Detection of Stealth Cyber-Attacks in DC Microgrids

Cyber-physical systems such as microgrids contain numerous attack surfaces in communication links, sensors, and actuators forms. Manipulating the links sensors is done to inject anomalous data that can be transmitted through cyber layer along with original stream. The presence of malicious, packets a dc microgrid create hindrances fulfilling control objectives, leading voltage instability affecting load dispatch patterns. Hence, detecting essential for restoration system stability. This article answers two important research questions: 1) Which data-driven detection scheme offers best performance against stealth cyber-attacks microgrids? 2) What improvement when fusing features (i.e., current data) training compared using single feature current)? Our investigations revealed adopting an unsupervised deep recurrent autoencoder anomaly superior other benchmarks. trained on benign generated from multisource model. Fusing 14.7% improvement. efficacy results verified experimental collected testbed subjected cyber-attacks.