Deep Learning for High-Impedance Fault Detection: Convolutional Autoencoders

High-impedance faults (HIF) are difficult to detect because of their low current amplitude and highly diverse characteristics. In recent years, machine learning (ML) has been gaining popularity in HIF detection ML techniques learn patterns from data successfully HIFs. However, as these methods based on supervised learning, they fail reliably any scenario, fault or non-fault, not present the training data. Consequently, this paper takes advantage unsupervised proposes a convolutional autoencoder framework for (CAE-HIFD). Contrary conventional autoencoders that normal behavior, (CAE) CAE-HIFD learns only signals eliminating need presence non-HIF scenarios CAE training. distinguishes HIFs operating conditions by employing cross-correlation. To discriminate transient disturbances such capacitor load switching, uses kurtosis, statistical measure probability distribution shape. The performance evaluation studies conducted using IEEE 13-node test feeder indicate detects HIFs, outperforms state-of-the-art techniques, is robust against noise.