DC Serial Arc Fault Detection based on Visibility Graphs

Sarmed Hussain, Dominik Waffler, Ahmed Alnaggar, Hans Georg Herzog

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Unlike parallel arcs, serial arcs cannot be detected with conventional fuses for short circuit and overcurrent protection, which are currently used in vehicles. This is due to the current drop below the rated current in the case of a serial arc. This applies to both the low-voltage and the high-voltage vehicle system. Undetected serial arcs for an extended period of time can severely damage vehicle components and may even cause a fire due to the high temperatures involved. This paper presents a novel method for detecting standing serial arcs based on converting the time series of current measurements into visibility graphs. A neural network is further used to classify serial arcs based on differences in the degree distributions of the visibility graphs pre, during, and post the arc.

Original languageEnglish
Title of host publication2021 IEEE Vehicle Power and Propulsion Conference, VPPC 2021 - ProceedingS
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665405287
DOIs
StatePublished - 2021
Event18th IEEE Vehicle Power and Propulsion Conference, VPPC 2021 - Virtual, Gijon, Spain
Duration: 25 Oct 202128 Oct 2021

Publication series

Name2021 IEEE Vehicle Power and Propulsion Conference, VPPC 2021 - ProceedingS

Conference

Conference18th IEEE Vehicle Power and Propulsion Conference, VPPC 2021
Country/TerritorySpain
CityVirtual, Gijon
Period25/10/2128/10/21

Keywords

  • 48 V vehicle power system
  • Arc fault detection
  • DC serial arc
  • Degree distribution
  • Neural network
  • Visibility graph

Fingerprint

Dive into the research topics of 'DC Serial Arc Fault Detection based on Visibility Graphs'. Together they form a unique fingerprint.

Cite this