Pantograph arc location estimation using resonant frequencies in DC railway power systems

Research output: Contribution to journalArticlepeer-review

10 Citations (Scopus)
41 Downloads (Pure)


Pantograph arcing in electrified railway systems not only reduces the power collection quality of a locomotive but can also damage pantograph strips and overhead lines (OHLs). Most research detects pantograph-to-OHL arcs based on onboard voltage/current measurements, pantograph cameras, and so on. The use of onboard voltage/current data, though being cost-effective, rarely reflects arc locations along OHLs. This article develops an arc positioning method, which matches the position-dependent resonant frequency (RF) of an OHL with the RF extracted from voltage measurements in a pantograph arc event. A particular 20-km DC railway line supplied by two substations is first modelled in MATLAB/Simulink, with the model effectiveness being assessed based on voltage measurements in an arc event. Then, the OHL-related RFs estimated by the model are validated by the Tableau formula and discussed alongside impacts on RFs based on line models, locomotive locations, and line lengths. These evaluations permit the generation of an RF curve that links OHL-related RFs with arc locations. The arc positioning method is tested based on the pantograph arc events presumed at various positions along the 20-km line, showing errors within 0.2 km at certain locations. The ability to determine arc locations will permit periodic inspections to be performed on the determined line sections.

Original languageEnglish
Pages (from-to)3083-3095
Number of pages13
JournalIEEE Transactions on Transportation Electrification
Issue number4
Early online date24 Feb 2021
Publication statusPublished - 1 Dec 2021


  • DC traction power supply system
  • overhead line
  • pantograph arc location estimation
  • resonant frequency


Dive into the research topics of 'Pantograph arc location estimation using resonant frequencies in DC railway power systems'. Together they form a unique fingerprint.

Cite this