Democritus University of Thrace (DUTH)
Ανάπτυξη μεθοδολογίας εντοπισμού θέσης σφάλματος βραχυκύκλωσης σε σύγχρονα συστήματα διανομής ηλεκτρικής ενέργειας
Abstract
dc:descriptionElectrical Power Distribution Systems (EPDSs) are responsible for supplying electricity to end users in a reliable, high-quality, and cost-effective manner. Achieving this objective relies primarily on the optimal design of the distribution network. However, regardless of the quality of the system design, the occurrence of faults cannot be completely prevented, nor is it technically or economically justifiable to attempt to eliminate them entirely. Instead, the prompt clearance of faults, followed by the accurate identification of their location, is essential. These functions are performed by the protection system of the EPDS. To date, the systematic location of short-circuit faults without the need to dispatch field crews for visual inspection of the affected feeder or to perform field testing (e.g., using frequency generators) has not become common practice, even in conventional radial distribution systems without distributed generation. In modern EPDSs, however, the integration of distributed generation has enabled bidirectional power flow in distribution feeders, even when the network topology remains radial. This development necessitates not only the redesign of distribution system protection schemes but also the development of new fault-location methodologies capable of accounting for these operating conditions. At the same time, these methodologies can exploit the advanced protection, monitoring, control, and communication technologies that are progressively being deployed in modern distribution systems. This doctoral dissertation develops novel methodologies for the systematic and highly accurate location of short-circuit faults, taking into account the operational and topological characteristics of modern EPDSs. The proposed methodologies are based on the analysis of the faulted network using synchronized voltage and current phasors recorded at selected network buses before and during the fault. In addition, a methodology for the location of single-line-to-ground faults based on machine learning models is proposed. The models are trained using features extracted through Wavelet Scattering Networks (WSNs). The obtained results demonstrate that the features generated by Wavelet Scattering Networks provide a highly informative representation of electrical signals, enabling more effective analysis of distribution systems and significantly improving the accuracy of fault location.
Degree
thesis:*- Grantor dc:publisher
- Democritus University of Thrace (DUTH)
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Arsoniadis, Charalampos
- Αρσονιάδης, Χαράλαμπος
Subjects
dc:subject × 12- Τοποθεσία σφάλματος
- Κατανεμημένη παραγωγή
- Συστήματα διανομής ηλεκτρικής ενέργειας
- Fault location
- Distributed generation
- Electrical distribution systems
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Ηλεκτρική και Ηλεκτρονική μηχανική
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Electrical and Electronic Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/62564