Krishna S

This paper presents some results on the accuracy of normal form analysis for large deviations from the stable equilibrium point (SEP) in a power system. The results of studies on the region of convergence of Neumann series, which is required for normal form analysis, are presented. A polynomial function approximation is proposed to model the generator exciter limiter. Studies on the accuracy of higher order normal form analysis for large deviations from SEP are presented. Use of Prony analysis in addition to simulation is proposed for the verification of accuracy.

During a false data injection attack (FDIA), the attacker compromises the meter readings to corrupt the state estimator's output. One method of managing FDIA is by securing meters that are most vulnerable to attacks. The existing methods may not secure the most vulnerable meters, and may not work in practical scenarios where the utility has budget and time constraints. To solve these issues, a novel vulnerability index is proposed in this paper. The advantage of using the proposed index in finding meters to be secured is demonstrated using IEEE test systems.

In a false data injection attack to compromise the results of state estimation, the attacker intends to alter the meter readings communicated to the control center to his/her desired values to realize an objective. However, to avoid detection, the manipulated readings should not violate physical laws. Therefore, even if an attacker is able to attack a set of meters, he/she may not be able to inject an arbitrary magnitude of attack into all meters of this set. In this paper, we study the degrees of freedom in the attack vectors available to the attacker; higher degrees of freedom pose a higher risk to the utility. We propose an algorithm to determine the degrees of freedom in attack vectors and a method to find the set of meters to be secured in order to minimize the degrees of freedom in all attack vectors. Our results show that a significant reduction in degrees of freedom in all attack vectors is possible by securing a small set of meters.

Normal form analysis is used to assess the nonlinear behavior; it can also be used for siting controllers. Normal form analysis is straight-forward for systems governed by explicit ordinary differential equations. However, the analysis becomes complicated if the governing equations are differential–algebraic. The presence of SVC/STATCOM results in the governing equations for which the normal form analysis is not as straight-forward as that for explicit differential equations. In this paper, it is shown how semi-explicit differential–algebraic equations of index 1 are amenable to normal form analysis. In particular, methods are proposed for normal form analysis in the presence of SVC and STATCOM. Normal form analysis can be used to find the location of SVC and STATCOM that results in the best damping of rotor swings; the results are analyzed by case studies and also compared with the performance obtained by locating SVC and STATCOM using linearization methods.

A false data injection attack (FDIA) targeting measurements can bypass the distribution system state estimator and negatively affect the control and decision tasks of the distribution management system. A measurement becomes more vulnerable if the attacker has many sparse attack vectors to choose from to attack this measurement without getting detected. A fast graph-based algorithm is proposed in this paper to rank all measurements based on FDIA vulnerability in distribution networks. Though the problem of ranking meters based on vulnerability to FDIA is an offline problem, it is combinatorial and the brute force method for solving this problem is not feasible. The proposed algorithm assigns a higher rank to a measurement if it can be attacked by a large number of sparse attack vectors, by leveraging the radial structure of distribution networks and topological properties of power flows. The proposed algorithm has been shown to have applications in identifying meters that need to be secured in order to reduce vulnerability to FDIA, evaluating the impact of pseudo-meters and secured meters on FDIA vulnerability, and determining optimal measurement placement to mitigate FDIA vulnerability.