Parallel Arc Fault Detection Using S-Transform and Convolutional Neural Network Under Simultaneous Power Quality Disturbances

Parallel Arc fault is one of the causes of electrical fires, which potentially poses a major threat to people\u2019s safety. In a low-voltage distribution system, a parallel arc fault may occur simultaneously with power quality disturbances. Arc fault detection may be increasingly difficult because the characteristics of power quality (PQ) disturbances have non-stationary signals such as voltage sags and transients. Therefore, Stockwell Transform (ST) was proposed for analyzing parallel arc fault under simultaneous power quality (PQ) disturbances, as it was superior for analyzing non-stationary waveform. The study aims to examine the parallel arc fault signal model through mathematical models and experiments. The time-frequency features of parallel arc fault were obtained from ST analysis. Then, the dataset of time-frequency was provided to vary the number of tests. The obtained time-frequency features were used as input into a convolutional neural network (CNN) for classification recognition of parallel arc fault under simultaneous PQ disturbances. Eleven types of PQ disturbances were selected to evaluate the performance of the proposed methods. The simulation results showed that the proposed methods could detect parallel arc fault under simultaneous PQ disturbances with satisfactory and accurate results