Experimental HIl implementation of RNN for detecting cyber physical attacks in AC microgrids

Abstract

In this paper, a real-time cyber intrusion detection mechanism based on recurrent neural networks is implemented for detecting cyber-physical attacks targeting AC microgrids (MG). An AutoRegressive eXogenous Neural Network (NARX) model is deployed as an Intelligent Detection System (IDS), to detect cyber-physical anomalies in the behavior of exchanged active power in a connected AC microgrid. Results are validated through a Hardware-in The loop simulation using the Opal RT real-time simulator and an external microcontroller board (Arduino) for Embedding the used Artificial Neural Network ANN.