Optimal positioning of geo-referenced short circuit sensors for faster fault finding using genetic algorithm

Abstract

The performance of electric energy distribution system is strongly affected by faults. System restoration is critical due to difficulties in searching process by the utilities maintenance crew. Usually fault searching methods employed by utilities may lead to different locations to the same fault, mainly if the feeder is dense connected. There have been many approaches in order to overcome this problem. One popular method applies short-circuits indicators to be allocated along the feeders. Nevertheless, their cost can be a limitation for the method. In this context, this work issues the problem of optimal short-circuit indicator allocation along electrical network. The allocation is optimally accomplished in order to minimize the fault searching area of the utility maintenance crew. As the the problem formulation leads to a multimodal discrete nonlinear optimization, a genetic algorithm has been the optimizer tool selected to accomplished it. The technique has been tested on a model of a feeder of an Brazilpsilas Utility. The results showed a significant improvement over ad-hoc methodologies usually adopted by utilities.