Data-driven Earth-fault Localization via Neural Network Models for Classification and Regression

• Main Authors: Wormann, Julian, Amaya, Camilo, Eichelseder, Max, Duchon, Markus, Schwefel, Hans-Peter
• Format: Conference Proceeding
• Language: English
• Subjects: Accuracy; Analytical models; Data models; digital twin; earth fault localization; electricity distribution grids; Fault location; Impedance; Location awareness; Medium voltage; Neural networks; Numerical models; Predictive models
• ISSN: 2474-2902
• DOI: 10.1109/SmartGridComm60555.2024.10738067

Rapid detection and precise localization of earth faults are crucial for ensuring the safe operation of energy distribution grids. Although detection is generally accomplished using integrated power quality hardware, pinpointing the fault location frequently requires labor-intensive manual switching operations, which can lead to supply disruptions. This work explores two methods of automated, robust, and accurate earth fault localization in medium voltage power grids using machine learning techniques. A digital twin is utilized to simulate fault events thus enabling the generation of extensive data to facilitate the training of data-driven models for fault localization. We explore two shallow neural network models to tackle the fault localization problem. On the one hand, we predict potential fault locations via node classification. On the other hand, a regression model is realized, that maps input measurements to associated impedances.This research work considers real-world medium voltage grids and focuses on investigating the influence on accuracy, robustness and flexibility of the implemented fault localization methods. The aim is to characterize the application areas of each approach for distribution grid operators. Finally, we open a discussion on requirements and desirable improvements for future developments.