Aplicação de redes neurais artificiais na localização de faltas em linhas de transmissão

Abstract

The work presents the application of Artificial Neural Networks (ANNs) in the location of faults in transmission lines (TL). Two different methodologies are proposed, the first method, called generalist does not use the classification module, ie an ANN topology is used to estimate the location of all types of faults, in this method, were used as inputs the data recorded in both transmission line terminals and data from only one terminal of the TL, while the second has an ANN topology functioning as a fault classification module, and a topology for each type of fault to estimate the distance, this method uses as input the currents of the local bus. In order to perform the tests, the ATP software was used, considering the faults phase-earth, phase-phase, phase-phase-earth and phase-phase-phase to generate the database, already for implementation of the topologies ANN was used the NNTOOL toolbox of Matlab® software. After the tests were carried out, it was verified the efficiency of the methods that use ANN to locate faults, both with the generalist method and with the method of classification and location.