Design of power network fault diagnosis based on time series matching

Abstract

Common grid fault diagnosis does not fully utilize the alarm timing information generated by the fault. To solve this problem, this paper proposes a fault diagnosis method based on time series. The method analyzes the alarm hypothesis sequence generated by the grid fault and the time sequence actually received by the dispatch center, and utilizes the discrete characteristics of the edit distance and reflects the event discreteness and time continuity of the alarm information by adding the time distance. The calculated data of the similarity between the two sequences and the confidence of the alarm hypothesis sequence determine the faulty component.