Auto Encoder and DWT for Arc Fault Prediction in Electrical Systems

Abstract

The most frequent reasons for arc faults are electrical problems, such as outdated wiring and bad connections. Electrical fires are caused by arc faults because they release molten metal and produce high temperatures. These kinds of flames result in a great quantity of loss and disaster each year. A novel approach to identifying residential series and parallel arc faults is presented in this research. To simulate arc failures in series and parallel circuits, an arc simulation model is used. The fault detection algorithm is then used to develop the Discrete Wavelet Transform (DWT) signal processing technique in MATLAB/Simulink, which is used to obtain the fault features. Next, it was discovered that db2 and one level of the wavelet transform were the proper mothers and levels for extracting arc-fault features. MATLAB Simulink was utilized to construct and model