Design and Implementation of Arcing Detection on Low Voltage with Evaluation of Discrete Wavelet Transform

In this study, parallel arcing detection is designed and implemented. Arcing is one of the disturbances in electrical power systems. The fault current generated by arcing typically happens very quickly with a short period. This study focuses on designing a hardware device for parallel arcing fault detection. The algorithm in this research is using wavelet transform as a signal processing technique by considering the selection of the mother wavelet and decomposition for optimizing parameters of the DWT method. The system of this DWT is to transform the maximum current values read by the hardware into an output signal called the detail coefficient of the current maximum. Based on the DWT output data in normal and switching conditions, the maximum detail coefficient value obtained is 150, with 3 points having this value. Therefore, to detect parallel arcing, two thresholds are used: the detail coefficient current maximum value should be ≥150, and the number of fault points should be ≥3 points. In this study, DWT parameters were evaluated by using 3 different variations of mother wavelet and 2 variations of decomposition level to determine the best arcing detection results. From the results obtained, it shows that the use of mother wavelet parameter Coiflet 4 and decomposition level 5 is the best DWT performance in detecting parallel arcing faults with a maximum current coefficient detail output value of 3108 and number of fault points at 5 points.