Detection and analysis of series arc using non-conventional methods in low-voltage switchboards

Abstract

Abstract Detection and analysis of series arc in low-voltage switchboards have significant meaning for preventing the electrical fires. However, the conventional current and voltage methods have low a sensitivity to sense the minute arc discharge, leading to the fail operation of arc fault circuit interrupter. Therefore, this paper dealt with the application of non-conventional methods, including the ultra-violet (UV), acoustic emission (AE), and transient earth voltage (TEV) sensor in arc detection, for the purpose of improving the detection sensitivity and reducing the potential electric fires. Three types of typical arc faults in low-voltage switchboards were simulated and the actual detection environment was configured. From the results, the wavelength of UV light emitted from arc was 200–400 nm and the arc-induced AE signal had a frequency range of 40–600 kHz. The TEV signals generated from three types of arc faults presented different frequency spectrums, based on which the time-frequency map was used to classify the fault type.