Search

Abstract: Modern power systems are vulnerable to false data injection attacks (FDIAs) due to the widespread applications of two-way communication networks for system operation and control. Diagnosing such malicious attacks are of great significance for the resilient operations of power systems. Firstly, to detect the presence of FDIAs, a novel FDIA detection scheme is proposed in this thesis for three-phase distribution systems based on zero-sequence voltage (ZSV). From the voltage and power measurements, the bus voltages are estimated, and then the estimated ZSV is calculated as the sum of the estimated bus voltages on the three phases to represent the degree of unbalance of the distribution system. Via mathematical analysis of the linear distribution system state estimation (DSSE) model, the distribution of the estimated ZSV under the normal condition is derived, based on which a whitening process is adopted on the estimated ZSV to weaken the effect of measurement noises. The L2-norm of the whitened ZSV vector is then compared with a predefined threshold for the FDIA detection. Moreover, the probability of false alarm of the proposed scheme is derived, which can be utilized to determine the detection threshold for a desired tolerance of false alarm rate. The proposed FDIA detection scheme is validated on several IEEE Test Feeders and simulation results show the effectiveness of the proposed scheme in detecting FDIAs in three-phase distribution systems. To further specify the locations of FDIAs, a ZSV-based FDIA localization scheme is also proposed in this thesis for three-phase distribution systems. Based on the estimated ZSV of a bus relative to others, the bus under attack is localized. By eliminating the effect of estimation error, an estimation of the injected attack is obtained. With an optimal test to minimize the error rate, the phase under attack can be diagnosed. By analyzing the estimated ZSV values in three steps, the bus localization and phase diagnosis can be achieved. Simulation results verify the effectiveness of the proposed FDIA localization method with high accuracy.