Arcing fault diagnosis using enhanced cross-correlation technique

Abstract

Abstract This study explored the potential use of cross-correlation as a technique for detecting arcing faults in a power system distribution network. The cross-correlation technique was employed to investigate the effect of each antenna placement as a detection device, time difference of arrival (TDOA), time delay, and correlation magnitude of arcing signals detected during on-line arcing fault measurement. The arcing fault was detected using four antennas that had been set up around the arc source point in a high voltage (HV) laboratory. The measurements were taken using a digital oscilloscope. For precise results, the Discrete Wavelet Transform (DWT) denoising technique combined with cross-correlation (CC) technique were applied using MATLAB software to identify the arcing signals detected in order to diagnose the differentiation between noisy and real arcing fault signals. Further assessment was carried out by performing a cross-correlation technique on the real arcing signals obtained to find the similarities and arrival time’s delay between single arcing signals’ placement. The outcome shows that all measurements including the time difference of arrival (TDOA), correlation magnitude, time delay, and antennas’ placement towards the arcing source point are valuable in determining the arcing signals detected precisely.