Detection and Classification of Fault Types in Distribution Lines by Applying Contrastive Learning to GAN Encoded Time-Series of Pulse Reflectometry Signals

Abstract

This study proposes a new method for detecting and classifying faults in distribution lines. The physical principle of classification is based on time-domain pulse reflectometry (TDR). These high-frequency pulses are injected into the line, propagate through all of its bifurcations, and are reflected back to the injection point. According to the impedances encountered along the way, these signals carry information regarding the state of the line. In the present work, an initial signal database was obtained using the TDR technique, simulating a real distribution line using (PSCAD™). By transforming these signals into images and reducing their dimensionality, these signals are processed using convolutional neural networks (CNN). In particular, in this study, contrastive learning in Siamese networks was used for the classification of different types of faults (ToF). In addition, to avoid the problem of overfitting owing to the scarcity of examples, generative adversarial neural networks (GAN) have been used to synthesise new examples, enlarging the initial database. The combination of Siamese neural networks and GAN allows the classification of this type of signal using only synthesised examples to train and validate and only the original examples to test the network. This solves the problem of the lack of original examples in this type of signal of natural phenomena which are difficult to obtain and simulate.

Modeling and Simulation of Time Domain Reflectometry Signals on a Real Network for Use in Fault Classification and Location

“…30 and Fig. 31) inside an existing installation, and on the other hand, the simulated data obtained have been processed for the classification of electrical faults [51] using the latest techniques of pattern recognition and deep learning.…”

“…During this period, it will record both the signals of the network in its fault-free operating state, as well as the faults that occur in the network. While these signals are being recorded, a fault location algorithm will be developed based on the database of 200 signals obtained in this work and the 10,000 synthesized signals obtained from these 200 that have been published [51] and whose database has been shared [52]. These results will be published and will also contain the error between the simulated, synthesized, and real signals obtained in the real network.…”

Multi‐view synergistic enhanced fault recording data for transmission line fault classification

“…Considering the temporal characteristics [28] of fault recording data and the advantages of Long Short-Term Memory (LSTM) [29] in processing long sequence data, LSTM has also been applied to fault classification tasks. In addition, some studies have explored the application of deep learning techniques such as Generative Adversarial Networks (GANs) [30] in transmission line fault classification.…”

TransSiamese: A Transformer-Based Siamese Network for Anomaly Detection in Time Series as Approach for Fault Location in Distribution Grids

“…According to the physics of transmission lines, these pulses travel through the network and are reflected at every bifurcation. Therefore, some of them are bounced back [13]. The magnitude of these reflections depended on the impedance of the line at each bifurcation.…”

“…In a previous study [13], the authors proposed a pulse injection solution based on the TDR technique, where a real network was modelled, and characteristic fault types were simulated. To avoid the lengthy simulation process, the authors generated a database of 200 signals examples [14] and synthesized the rest up to 10,000 using Generative Adversarial Networks (GANs).…”