Identifying network properties that influence the vulnerability of power systems to extreme contingencies

Research into transmission system vulnerability has shown that different power networks respond differently to large multiple contingencies, in terms of how much load shedding is required to restore a power system to within its operating limits after the contingency. Specifically, the mean and variance in blackout sizes differ substantially among different cases. This paper aims to understand why these differences exist by studying relationships between the properties of several test cases and the vulnerability of those networks to severe multiple contingencies. The results suggest that the graph density (the relative number of transmission lines) and the geographic dispersion of loads significantly influence the vulnerability of networks to multiple contingencies.