Algorithm for Fault Localization of Distribution Feeders by Weighted Probability Method

This study aims to develop an algorithm for fault localization in distribution feeders to improve the reliability of the medium voltage network by fast and accurate localization during emergency conditions. Matrix row/column operations are performed by the algorithm on a connection matrix of a graphical network model based on the incidence matrix to identify the faulty feeder. Exact and complete fault localization is ensured by the weighted probability method, proposed in this study. The algorithm is designed to coordinate with remote terminal units (RTU) of switching devices. Testing was performed using prototype software and a demonstration network panel. 11kV overhead distribution network of Sri Lanka was taken as the case study. The speed of the process was compared with data collected from utility and effective results were observed.