Arc Fault Detection Method Based on Signal Energy Distribution in Frequency Band

Abstract

Each year, fires caused by arc fault bring great loss and damage. But due to its similarity to working waveforms of many household appliances, arc waveforms are very difficult to detect in domestic circuits. This paper focuses mainly on how to distinguish it from a variety of electrical waveforms, and thus lays the foundation for the development of measuring equipment. Arc fault waveforms of circuits with different power factor and some typical working household appliances' waveforms are collected with a set of experimental devices, and analyzed in terms of time and frequency, after which the low frequency component of the signal is filtered through wavelet decomposition and the high-frequency part of the signal is decomposed with the improved wavelet packet decomposition. Then histograms are made based on the signal energy of different frequency bands. A new arc fault detection process can be deduced by the position and the magnitude of maximum energy in frequency bands.