Secure MPC/ANN-Based False Data Injection Cyber-Attack Detection and Mitigation in DC Microgrids

Abstract

Direct current (DC) microgrids can be considered as cyber-physical systems due to implementation of measurement devices, communication network, and control layers. Consequently, dc microgrids are also vulnerable to cyber-attacks. False-data injection attacks (FDIAs) are a common type of cyber-attacks, which try to inject false data into the system in order to cause the defective behavior. This article proposes a method based on model predictive control (MPC) and artificial neural networks (ANNs) to detect and mitigate the FDIA in dc microgrids that are formed by parallel dc-dc converters. The proposed MPC/ANNbased strategy shows how MPC and ANNs can be coordinated to provide a secure control layer to detect and remove the FDIAs in the dc microgrid. In the proposed strategy, an ANN plays the role of the estimator to implement in the cyber-attack detection and mitigation strategy. The proposed method is examined under different conditions, physical events and cyber disturbances (i.e. load changing and communication delay, and time-varying attack), and the results of the MPC-based scheme is compared with conventional proportional-integral controllers. The obtained results show the effectiveness of the proposed strategy to detect and mitigate the attack in dc microgrids. Index Terms-Artificial neural network (ANN), cyber-physical dc microgrid, false-data injection attack (FDIA), model predictive control (MPC). I. INTRODUCTIOND IRECT current (DC) microgrids offer more advantages when compared to ac microgrids, e.g., higher reliability, more efficiency, less complex control, and also easier interface with electronic loads, energy storage systems, and renewable energy resources because of their natural dc behavior [1], [2]. Distributed generators are connected to the microgrids commonly by controllable converters [3]. To control the dc microgrids, three control layers are implemented in a hierarchical control Manuscript

Attack and Protection Technology of Intelligent Terminals for New Energy Internet Transmission, Transformation, and Distribution

“…Meanwhile, FDIA with cryptic features exhibits diverse features [12]. ①Incomplete attack capabilities lead to different false data with different degrees of sparsity.…”

A Comprehensive Review of Cybersecurity in Inverter-Based Smart Power System Amid the Boom of Renewable Energy

“…Those very networks can have more than one hidden layer and feedback, and their output often VOLUME 4, 2016 varies depending on either the weighted sum of all the inputs and the activation function. The most recent yet interesting usage of ANN to detect cyberattacks can be viewed in [124], in which the model predictive control/artificial neural network (MPC/ANN) defense strategy was proposed. The role of MPC is to inject a certain amount of data into the system to quickly move the effect of FDIA and help the system heal.…”

“…In contrast to non-real-time simulations, these devices not only act as virtual systems whose communication data and characteristics match that of the real power network but also achieve real-time simulation speed. While they come in the form of HIL, they may be more comprehensive with the inclusion of a dedicated software tool-chain [178] and support for interfacing with simulation softwares to leverage the existing features in these software [124]. API is included as part of the solution to facilitate the integration of other hardware and software, such as data acquisition and actuator.…”

An Explainable Intelligent Framework for Anomaly Mitigation in Cyber-Physical Inverter-based Systems

“…ANNs are utilized to improve power sharing, distributed generation management, and resilient control design in multi-DG microgrids. Specifically, ANNs are designed using historical voltage and current measurements to act as an estimator and an observer layer for FDI detection and mitigation in cooperative controlled DC microgrids [26]- [29]. The application of ANNbased control for AC microgrids is not common as manifested from the usage of ANN as an observer layer in [25] and in reference tracking applications for DC microgrids in [28].…”