Quickest line outage localization under unknown model

Abstract

Line outage detection and localization play pivotal roles in enhancing the overall reliability of the electricity grid. The existing line outage detection and localization techniques often rely on the assumption that the information about the grid, such as the topology and system parameters, is known perfectly. In practice, however, such information bears uncertainties due to inaccuracies in lines parameters and the historical data. This paper studies line outage detection and localization under the assumption that the line reactance values are known only partially, and aims to find the minimum number of measurements to perform outage detection and localization with target reliability. Specifically, based on the nominal values of line reactance it proposes a stochastic graphical framework that capitalizes on the correlation among the measurements generated across the grid, and designs data-adaptive data-acquisition and decision-making processes for the quickest localization of the lines in outage. The paper also analyzes the sensitivity of the proposed algorithm to the changes in the line reactance values and shows its robustness to line reactance uncertainties.