Radio-Frequency Electrical Arcing Detection Using Low-Cost Mini Spectrum Analyzer

Abstract

Radio-frequency (RF) electrical arcing detection by using a low-cost mini spectrum analyzer called TinySA is proposed. The electromagnetic (EM) frequency spectrum emitted by a local discharge (electrical arcing) will be monitored by TinySA instead of an expensive conventional spectrum analyzer. A suitable monitored frequency of 347 MHz and bandwidth of 1 MHz is employed in this study. The detection algorithms including a mean normalization, standard deviation, and variance of the real-time electrical arcing signal are studied. The DC electrical arcing generator is employed as an electrical arcing source in this paper. It is found that TinySA can be efficiently employed to detect the transient signal of the electrical arcing. The frequency spectrum of the electrical arcing has randomly occurred versus time. In an electrical arcing event, the mean normalization values of some frequencies are greater than the mean of the whole frequency spectrum. Moreover, variance and standard deviation of the frequency spectrum can be employed as detection parameters. Finally, the mean normalization performed better than the variance and standard deviation parameters in the case of very weak signal amplitude.