Root cause localization on power networks

Given information about cascading failures on a power network, localization of the root cause is of great importance in accelerating the process of recovering the power network from cascading failures. In order to locate the root cause, we need to answer two questions: (1) Which information can we use to locate the root cause of cascading failures? (2) What algorithm can locate the root cause effectively? In this paper, we present our recent results towards answering these two questions. In particular, we developed a root-cause localization algorithm by exploiting the concept of correlation network, which can describe the diffusion. The simulations based on different power networks are provided, in which we show our algorithm outperforms some other existing algorithms.