Decision Trees Applied to Fault Locations in Distribution Systems with Smart Meters

The outages of power supply caused by faults in electric power systems are a serious issue because of the potential damage they can cause to the elements of the distribution systems, to the concessionaires and to the consumers, compromising the energy quality available. In this context, this work aims to locate faults in distribution systems employing voltage measurements made by smart meters allocated in different points of the electric network. The proposed approach is based on Decision Trees intelligent systems. It is also used the sensing of the passage of the fault current through the smart meters to indicate the location of the fault compared to the positions of the smart meters. The results obtained demonstrate that the use of redundant information provided by several smart meters minimizes the location errors, which reduces the search area for the network maintenance teams, allowing faster energy supply restorations.