Study On The Identification Method Of Series Arc Fault

In this paper, the research of series arc fault has been launched. This paper studies the characteristics of series arc fault and looks for effective methods to detect, which has great significance to ensure the safe operation of power system stability and improve the quality of power supply.A series arc fault generator was built according to UL1699. Experiments were carried out under different load conditions. Loop current waveforms with and without series arc fault were obtained. First, the current signal was decomposed and reconstructed by wavelet transform. On the basis of current signal decomposition and reconstruction, the irregular degrees of signals in each frequency band were quantified with approximate entropy algorithm and the feature vectors of current signal were acquired. Then, all the feature vectors were used as input variables of Support Vector Machine (SVM). The series arc fault can be recognized by classifying those feature vectors with SVM. It shows that the feature vectors obtained by wavelet approximate entropy algorithm can be used to diagnose series arc fault.