Title:Anomaly Detection and Location in Distribution Network: A Data-Driven Approach

Abstract:In this paper, a data-driven approach is developed for the anomaly detection and location in distribution network. First, spatio-temporal matrices are formulated by using massive online monitoring data collected through multiple-time-instant monitoring devices in the distribution network. Based on the random matrix theory for the spectrum analysis of those data matrices, a real-time anomaly detection and location algorithm is developed. During the above process, the linear eigenvalue statistics are defined to indicate data behavior, and the anomaly latencies are analyzed and located. As for those low-dimensional data matrices formulated in the distribution network, by using tensor product, an increasing data dimension method is developed for them to be analyzed by using the random matrix theory. Theoretical justifications for the developed approach are provided through Marchenko-Pastur Law and Ring Law. Case studies both on the synthetic data from IEEE standard bus system and the real-world online monitoring data in a power grid corroborate the feasibility and effectiveness of the proposed approach.