Uma nova proposta para detecção de faltas de alta impedância usando características no domínio do tempo e da frequência

Abstract

High impedance faults (HIFs) present a great risk to people and animals that are in the vicinity of its occurrence. This occurs the due to the fact that this type of fault is not detected by conventional protection devices, thus allowing this problem to persist until someone reports its occurrence, which cannot always happen in time to avoid major losses. This work aims, through a new methodology, to present a solution to this type of problem, using a method to detect transients together with a method responsible for classifying if the detected event is in fact a high impedance fault or not. After detecting the transient, the method uses characteristics both in time and frequency domains to distinguish a HIF among another events commom to the electrical distribution system. In case of the time domain, the method searches for the buildup and shoulder stages and, once these characteristics are detected, the method then analyzes the behavior of the third harmonic magnitude, which results from the distortions caused by the electric arc present in the HIF. In order to validate the effectiveness of the technique, it was assembled a test system in ATPdraw®, in which, in addition to the transients that normally occur in the distribution systems, three different electric arc models were also implemented and used to simulate HIFs. For this system, the methodology presented promising results acting correctly for all analyzed cases. Key-words: High Impedance Fault, Power Systems Protection, Distribution System, Eletric Arcing, Signal processing.