Estudo da implementação de um método para detecção de faltas de alta impedância baseado nas medições dos fasores harmônicos

Abstract

This works presents the implementation of a method for high impedance fault detection in power distribution systems. The method uses the harmonics phasors measurement to detect the arc current, typically, founded in high impedance faults. For the estimation of amplitude and phase of the harmonics, it is used the Sliding Window Recursive DFT, and to distinguish the high impedance fault events from another events in the electrical system, it was designed an euclidean distance classifier. Moreover, there are two additional algorithms to improve the method performance, which are: an overcurrent detector and a broken conductor detector based on the ANSI 46BC standard. To analyse the method accuracy, it was generated a signal database containing some events that are present in the electrical system such as capacitor bank switching, nonlinear load switching, inductor motor start, low and high impedance faults and distributed generation connection. The results shows both a step by step method performance analysis, as well as a quantitative analysis using a considerable testing set of signals. Lastly, it is presented a comparison between the proposed methodology and some other methods founded in the high impedance fault detection literature.