A multi-stage methodology for fault location in radial distribution systems

In the context of smart grids and microgrids, fault location is an important point that can be explored with all the possibilities of this new paradigm. This study presents a multi-stage methodology that involves decision trees as well as artificial neural networks for fault location purposes. Another aspect to be considered is the use of an allocation of smart meters in the distribution system for the tests. The IEEE 34-bus distribution system was used to validate the methodology. This system was chosen due to its considerable number of monophasic branches and its extension. The focus of this paper is the multiple estimation problem, in which, for a same distance of the fault, there is more than one possibility to locate it. Confusion matrices were used to explore the results and show the performance of the proposed algorithm.