Increasing resiliency against information vulnerability of renewable resources in the operation of smart multi-area microgrid

• Published in: Energy (Oxford) 2021-04, Vol.220, p.119776, Article 119776
• Main Authors: Tabar, Vahid Sohrabi, Ghassemzadeh, Saeid, Tohidi, Sajjad
• Format: Article
• Language: English
• Subjects: Adaptive systems; Artificial neural networks; Data points; Decision analysis; Distributed generation; Distribution functions; False data; Fuzzy logic; Information vulnerability; Injection; Learning algorithms; Machine learning; Multi-area microgrid; Probability distribution; Probability distribution functions; Real time; Renewable resources; Resilience; Resiliency; Sensitivity analysis; Smart grid; Sustainable yield
• ISSN: 0360-5442; 1873-6785
• DOI: 10.1016/j.energy.2021.119776

With the widespread use of information and communication technologies in smart grids, the vulnerability of these networks has increased significantly. In this paper, the operation of a smart electrical energy system is evaluated by considering the information vulnerability of renewable generators and their sensors. Hence, the false data injection process is modelled by the probability distribution function and different deviations to achieve real conditions. Since the attackers may have various information, an observation-action method is utilized to enhance their capability. Accordingly, an auxiliary variable is considered for real-time decisions and any modification which is required in the process. In return, to resilience the system and mitigate the impact of false data injection, a machine learning method, namely adaptive neuro fuzzy inference system, is used based on a threshold index. Implementing the method on a smart multi-area microgrid shows that if all data points are exposed to attack, the operation cost will be affected by about 8.52% and at least 70% of the false data into each sensor will be detectable. Moreover, sensitivity analysis validates that the wrong decision may be taken by attackers in real-time and, the percentage of detection will decrease if the threshold index increases. •Considering the information vulnerability of renewable generators as the critical components.•Modelling the behavior of attackers by a probability distribution and observation-action method.•Using the adaptive neuro fuzzy inference system based on a threshold index to detect and correct.•Validating the method by various scenarios and calculating the percentage of achievement.•Investigating the operation of a smart multi-area microgrid under false data injection.