Series Arc Fault Detection Using Novel Signal Processing Technique

Abstract

A low current electric fault can be considered as a potentially hazardous phenomenon typically caused by loose connections, damaged wiring or degrading insulation in a power distribution system which can create high intensity heat capable of igniting the surrounding material such as curtains, wood framing or insulation. It can have detrimental effects on the electrical equipment also and can cause loss of property and life. To be able to develop an algorithm which can detect such low current electrical arc faults, it is imperative to investigate the behavior and characteristics of the current signal. Also, advancement in signal processing has enabled viable solutions for the reliable detection of low current faults. A detection algorithm for such faults has been proposed which is capable of differentiating fault condition from loads causing similar current signatures under normal operating conditions. This has become more common as we use a greater array of electronic appliances at our homes.The presence of series arcing causes imbalance in magnitude as well as frequency content of current in the power line. The paper is intended to explore signal processing approach by analysing power frequency and high frequency noise separately during series arc faults and also investigate the effectiveness of detection approach for a variety of loads. The occurrence of faults will be verified by using the experimental data for a variety of loads and compared with the technique present.