A Frequency-Domain Fault Location Method for Underground Cables in MMC-HVdc Systems

Accurate fault location with a short data window is crucial and challenging for cable faults in the modular multilevel converter high voltage direct current (MMC-HVdc) system. Aiming at this issue, this article proposes a frequency-domain fault location method, which accurately and comprehensively a...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on instrumentation and measurement 2024, Vol.73, p.1-12
Hauptverfasser: Song, Guobing, Yan, Jifei, Chang, Zhongxue, Zhang, Chenhao, Gao, Xiaoping
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Accurate fault location with a short data window is crucial and challenging for cable faults in the modular multilevel converter high voltage direct current (MMC-HVdc) system. Aiming at this issue, this article proposes a frequency-domain fault location method, which accurately and comprehensively addresses the frequency dependency and unbalance of the parameter matrices without any approximation while considering the distributed parameters. First, the frequency-domain decoupling is analyzed for dc cables with the metallic sheath and armor grounded at two terminals. Second, a method is proposed to calculate the voltage distribution along the cable using voltages and currents at one side of the cable based on the numerical Laplace transform (NLT). Subsequently, the cable parameter calculation method suitable for the NLT is presented, and a multiround iterative algorithm for voltage distribution calculation is proposed. Ultimately, the complete fault location scheme is presented. Extensive simulations validate that the proposed method can achieve fault location with minor errors using a 20 kHz sampling rate and a 4 ms data window. Besides, high accuracy is maintained even under the influence of large fault impedance and noise. Compared to existing methods, the approach exhibits lower fault location error while requiring less computational complexity.
ISSN:0018-9456
1557-9662
DOI:10.1109/TIM.2024.3413194