Detection and mitigation against false data injection attacks using SHT and ANN in distributed control of DC microgrid

This article explores the susceptibility of network-connected DC Microgrids (MGs) to cyber threats, specifically focusing on combating false data injection attacks (FDIAs). It introduces a collaborative model that merges FDIA detection using Sequential Hypothesis Testing (SHT) with countermeasures derived from Artificial Neural Network (ANN) based control. Employing distributed control via adaptive droop, the proposed model integrates a detector that scrutinizes input currents for anomalies by comparing them with forecasted values provided by an ANN-based estimator. Upon detecting any deviation from the anticipated state, a countermeasure is activated, adjusting the adaptive droop control to utilize alternative measurements, thus minimizing the impact of an attack. To assess the efficacy of this approach, the study conducts dynamic simulations across diverse operational scenarios and tolerance thresholds. The findings demonstrate that the proposed method adeptly identifies falsely injected data and compensates for it even amidst severe attacks and prediction inaccuracies, thereby fortifying the security, reliability, and stability of DC MGs.