Advanced fault location scheme for superconducting cables based on deep learning algorithms

Tsotsopoulou, Eleni and Karagiannis, Xenofon and Papadopoulos, Theofilos and Chrysochos, Andreas and Dyśko, Adam and Hong, Qiteng and Tzelepis, Dimitrios (2023) Advanced fault location scheme for superconducting cables based on deep learning algorithms. International Journal of Electrical Power & Energy Systems, 147. 108860. ISSN 0142-0615 (https://doi.org/10.1016/j.ijepes.2022.108860)

[thumbnail of Tsotsopoulou-etal-IJEPES-2022-Advanced-fault-location-scheme-for-superconducting-cables]
Preview
Text. Filename: Tsotsopoulou_etal_IJEPES_2022_Advanced_fault_location_scheme_for_superconducting_cables.pdf
Final Published Version
License: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 logo

Download (2MB)| Preview

Abstract

This paper addresses the challenge of fault localization in power grids which incorporate Superconducting Cables (SCs) and presence of inverter-connected generation, by proposing a novel data-driven fault location scheme. The developed fault location algorithm utilizes the transformation of time domain fault current and voltage signatures, to time-frequency domain and exploits the advantages of a Convolutional Neural Network (CNN) to estimate the fault position along SCs. The proposed algorithm has been tested using a verified model of SC, and the results revealed that it can provide precise fault localization for a wide range of fault scenarios, including different fault types, fault resistance values and fault inception angles. Furthermore, the proposed scheme robustness has been verified against different influencing factors accounting for very small increments of fault location, additive noise and different value of sampling frequency. For validation purposes, the effectiveness of the CNN-based algorithm has been compared with other data-driven algorithms and the relevant advantages have been highlighted.