We study the resilience of real world power grids to targeted adversarial attacks. Prior blackouts have shown that failures in the power grid can cascade, starting from a single failure, leading to a large number of failed nodes. In this paper, we study the problem of identifying a set of k critical nodes, whose failure/attack leads to the maximum number of tripped nodes. There has been a lot of work on this problem, but it has been mainly restricted to simple networks and failure models with either steady state analysis or DC power flow. In this paper, we perform AC power flow based transient analysis on a detailed power grid model. We find that a simple greedy approach yields node sets with higher crit-icality than a degree based approach, which has been suggested in many prior works. Furthermore, we observe that the cascades exhibit a nonmonotonic behavior as a function of k.
展开▼
