الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 167 |  |  |
| | | from | 167 | | |
Abstract
Microgrids are localized energy systems that integrate various distributed energy resources and loads, functioning either in connection with or independently from the main power grid. Most microgrids use hierarchical control architecture based on primary and secondary control levels. These control levels are communicating measurements to achieve the control objectives. Therefore, the wide use of communication layers in microgrids to transmit voltage and current measurements of each Distributed Generator Unit (DGU) increases the possibility of exposure to cyber-attacks. Cyber-attackers can manipulate the measured data to distort the control systems of microgrids, which may lead to a shutdown.
Based on the previous discussion, securing microgrids against cyber-attacks is mandatory. Hence, this thesis proposes distributed mitigation layers for the False Data Injection Attacks (FDIAs) on voltages and currents of DGUs in meshed DC microgrids. The proposed control strategy is based on integrating two layers for cyber-attack detection and mitigation to immune the primary and the secondary control loops of each DGU. The first layer is assigned to mitigate FDIAs on the voltage measurements needed for the voltage regulation task of the primary control loop. The second layer is devoted to the mitigation of FDIAs on the DGU current measurements, which are crucial for the secondary control level to guarantee the proper current sharing of each DGU. Artificial Neural Networks (ANNs), Kalman filters, and Lyapunov based state observers are employed to support these layers by estimating the authenticated states. Using the discrepancies between the estimated data and the measurements under cyber-attack, controllers are assigned to set proper corrective actions. These actions secure the primary and secondary loops to achieve accurate terminal voltage regulation and proper current sharing, respectively, for each DGU in the microgrid. Not only accurate state estimation is required for perfect cyber-attack mitigation, but fast control response is crucial. Therefore, different controllers: Adaptive PI (API), Fractional Adaptive PI (FAPI), Least Mean Fourth (LMF) adaptive PI, and Continuous Mixed P-Norm (CMPN) adaptive PI are proposed and compared while being used in the proposed two cyber-attack mitigation layers.
Different simulation and experimental case studies are provided to demonstrate the proposed mitigation layers’ effectiveness in detecting and mitigating cyber-attacks on voltage and current measurements. The dynamic performance of the proposed mitigation strategy based on the state estimators is evaluated using MATLAB software package to ensure the accurate operation of DC microgrids despite the existence of cyber-attacks on the transmitted measurements in the control strategy. Comparison between the proposed state estimators using different mitigation controllers is conducted to test their performance. Finally, the experimental case studies are carried out to support the simulation results.