Resilient operation of DC microgrid against FDI attack: A GRU based framework

Abstract

DC microgrid is the most susceptible to cyber-attacks as the communication channel is involved for the implementation of the secondary controller. Accordingly, the false data are injected into the transmitted data (i.e., DC bus voltage) and it may lead to deteriorating the system performance. To address these issues, the gated recurrent unit (GRU) based mechanism is presented to eliminate the false data injection (FDI) attack for the resilient operation of the DC microgrid. The presented GRU-based framework is divided into two parts: 1) estimation strategy: an offline-trained GRU based network is employed herein for online evaluation of the actual DC bus voltage, and 2) mitigation strategy: GRU based trained network is exploited herein with an amalgamation of the proportional-integral (PI) controller to counteract the malicious cyber-attack. The presented GRU-based framework has several advantages such as ease of implementation and computationally efficient, unlike state-of-art methods. The sensitivity analysis is investigated herein to validate the effectiveness of the presented GRU-based framework over state-of-art techniques. Simulation results show satisfactory performance under manifold operating scenarios such as bias injection attack and time-varying attack. In addition, the quantitative and qualitative comparative performances are performed herein to demonstrate the efficacy of the presented framework.