Fault location in distribution networks using clustering techniques

Abstract

This paper studies an existing 13.8 kilovolt distribution network which serves an oil production field spread over an area of approximately sixty kilometers square, in order to locate any fault that may occur anywhere in the network using fuzzy c-mean classification techniques. In addition, the paper introduces several methods for normalizing data and selecting the optimum number of clusters in order to classify data. Results and conclusions are given to indicate the feasibility of the suggested fault location method.