An ICT System to Assist Earth Fault Localization

• Published in: Dependable Computing - EDCC 2021 Workshops pp. 51 - 64
• Main Authors: Schäler, Christine, Fotso, Serge Foko, Drenjanac, Domagoj, Chaves, Juan Felipe, Duchon, Markus
• Format: Book Chapter
• Language: English
• Published: Cham Springer International Publishing
• Series: Communications in Computer and Information Science
• Subjects: Electricity distribution grids; Fault localization; ICT architecture; Simulation
• ISBN: 3030865061; 9783030865061
• ISSN: 1865-0929; 1865-0937
• DOI: 10.1007/978-3-030-86507-8_5

Earth faults are hard to locate, but prominent failures in medium voltage electrical grids. However, to localize them in short time is essential to prevent total grid failures. There are different approaches to detect these faults, and also to determine or limit their geographical positions, in order to enable a quick elimination of the cause and thus to restore a stable network condition. To this end, an ICT system assisting fast localization is needed. Developing such an ICT system is challenging, as one needs to connect heterogeneous data and algorithmic sub-systems, scalability and low communication latency. In this paper, we propose a scalable service-oriented software architecture featuring these challenges. The architecture consists of (1) a so-called data fusion hub serving as connecting point of all sub-systems, (2) the earth fault application that localizes the faults based on signature comparison, and (3) a Message Queuing Telemetry Transport message bus for low-latency communication. Simulations are used to determine the electrical characteristics in the form of signatures to possible fault locations. Measurements of real earth faults are then compared with the simulated data to draw conclusions about the position of the fault. Once an earth fault is detected by the distribution system operator, the system calculates the probable fault locations based on the measured values and visualizes the results in the operator's network topology. In the evaluation, we show that the proposed system provides a scalable and robust solution to address the challenges of fast localization and visualization.