For rapid detection of arc fault and immediate break of fault lines

this paper presents a diagnostic method using wavelet transform to analyze the energy variation of the arc fault current in characteristic frequency bands. Adopting db5 wavelet function to decompose both the current in normal operation and the current of series arc fault into 6 layers

the energy variation of the arc fault current and the normal current in characteristic frequency bands is accordingly obtained. The frequency bands where detailed wavelet signals d4 and d5 exist can represent the specific character of the arc fault current. With this specific character

the arc fault can be accurately diagnosed. This present arc fault diagnosis technology is especially applicable to the arc fault with linear loads.