Advanced fault location in MTDC networks utilising optically-multiplexed current measurements and machine learning approach

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

This paper presents a method for accurate fault localisation of DC-side faults in Voltage Source Converter (VSC) based Multi-Terminal Direct Current (MTDC) networks utilising optically-multiplexed DC current measurements sampled at 5 kHz, off-line continuous wavelet transform and machine learning approach. The technical feasibility of optically-based DC current measurements is evaluated through laboratory experiments using commercially available equipment. Simulation-based analysis through Matlab/Simulink® has been adopted to test the proposed fault location algorithm under different fault types and locations along a DC grid. Results revealed that the proposed fault location scheme can accurately calculate the location of a fault and successfully identify its type. The scheme has been also found to be effective for highly resistive fault with resistances of up to 500 Ω. Further sensitivity analysis revealed that the proposed scheme is relatively robust to additive noise and synchronisation errors.
LanguageEnglish
Pages319-333
Number of pages15
JournalInternational Journal of Electrical Power and Energy Systems
Volume97
Early online date24 Nov 2017
DOIs
Publication statusPublished - 30 Apr 2018

Fingerprint

Electric fault location
Electric current measurement
Learning systems
Additive noise
Wavelet transforms
Sensitivity analysis
Synchronization
Electric potential
Experiments

Keywords

  • fault location
  • multi-terminal direct current
  • travelling waves
  • optical sensors
  • machine learning
  • pattern recognition

Cite this

@article{02433824242a4ffb9d30e75ae3a15091,
title = "Advanced fault location in MTDC networks utilising optically-multiplexed current measurements and machine learning approach",
abstract = "This paper presents a method for accurate fault localisation of DC-side faults in Voltage Source Converter (VSC) based Multi-Terminal Direct Current (MTDC) networks utilising optically-multiplexed DC current measurements sampled at 5 kHz, off-line continuous wavelet transform and machine learning approach. The technical feasibility of optically-based DC current measurements is evaluated through laboratory experiments using commercially available equipment. Simulation-based analysis through Matlab/Simulink{\circledR} has been adopted to test the proposed fault location algorithm under different fault types and locations along a DC grid. Results revealed that the proposed fault location scheme can accurately calculate the location of a fault and successfully identify its type. The scheme has been also found to be effective for highly resistive fault with resistances of up to 500 Ω. Further sensitivity analysis revealed that the proposed scheme is relatively robust to additive noise and synchronisation errors.",
keywords = "fault location, multi-terminal direct current, travelling waves, optical sensors, machine learning, pattern recognition",
author = "D. Tzelepis and A. Dyśko and G. Fusiek and P. Niewczas and S. Mirsaeidi and C. Booth and X. Dong",
year = "2018",
month = "4",
day = "30",
doi = "10.1016/j.ijepes.2017.10.040",
language = "English",
volume = "97",
pages = "319--333",
journal = "Electrical Power and Energy Systems",
issn = "0142-0615",

}

TY - JOUR

T1 - Advanced fault location in MTDC networks utilising optically-multiplexed current measurements and machine learning approach

AU - Tzelepis, D.

AU - Dyśko, A.

AU - Fusiek, G.

AU - Niewczas, P.

AU - Mirsaeidi, S.

AU - Booth, C.

AU - Dong, X.

PY - 2018/4/30

Y1 - 2018/4/30

N2 - This paper presents a method for accurate fault localisation of DC-side faults in Voltage Source Converter (VSC) based Multi-Terminal Direct Current (MTDC) networks utilising optically-multiplexed DC current measurements sampled at 5 kHz, off-line continuous wavelet transform and machine learning approach. The technical feasibility of optically-based DC current measurements is evaluated through laboratory experiments using commercially available equipment. Simulation-based analysis through Matlab/Simulink® has been adopted to test the proposed fault location algorithm under different fault types and locations along a DC grid. Results revealed that the proposed fault location scheme can accurately calculate the location of a fault and successfully identify its type. The scheme has been also found to be effective for highly resistive fault with resistances of up to 500 Ω. Further sensitivity analysis revealed that the proposed scheme is relatively robust to additive noise and synchronisation errors.

AB - This paper presents a method for accurate fault localisation of DC-side faults in Voltage Source Converter (VSC) based Multi-Terminal Direct Current (MTDC) networks utilising optically-multiplexed DC current measurements sampled at 5 kHz, off-line continuous wavelet transform and machine learning approach. The technical feasibility of optically-based DC current measurements is evaluated through laboratory experiments using commercially available equipment. Simulation-based analysis through Matlab/Simulink® has been adopted to test the proposed fault location algorithm under different fault types and locations along a DC grid. Results revealed that the proposed fault location scheme can accurately calculate the location of a fault and successfully identify its type. The scheme has been also found to be effective for highly resistive fault with resistances of up to 500 Ω. Further sensitivity analysis revealed that the proposed scheme is relatively robust to additive noise and synchronisation errors.

KW - fault location

KW - multi-terminal direct current

KW - travelling waves

KW - optical sensors

KW - machine learning

KW - pattern recognition

UR - http://www.sciencedirect.com/science/journal/01420615

U2 - 10.1016/j.ijepes.2017.10.040

DO - 10.1016/j.ijepes.2017.10.040

M3 - Article

VL - 97

SP - 319

EP - 333

JO - Electrical Power and Energy Systems

T2 - Electrical Power and Energy Systems

JF - Electrical Power and Energy Systems

SN - 0142-0615

ER -