Arc-Faults Detection in PV Systems by Measuring Pink Noise with Magnetic Sensors

Wenchao Miao, Xuyang Liu, K. H. Lam, Philip W.T. Pong

Research output: Contribution to journalArticlepeer-review

17 Scopus citations


The dc-arc detection is essential for dc systems to operate reliably and safely. Due to the randomness of dc-arc faults, it is difficult to define their characteristics for detection. The magnetic-sensing-based methodology using tunnel magnetoresistance (TMR) sensor was developed to distinguish the dc arc accurately. However, the noise from the power electronics such as maximum power pointing tracking (MPPT) controllers could affect the dc-arc detection in a photovoltaic (PV) system. In this paper, the arc faults which may occur in PV systems under various conditions were emulated in the dc system of 9 kW. It was verified that the proposed methodology of measuring pink noise can distinguish the dc arc accurately for PV systems from 48 to 300 V and current up to 30 A. An off-grid PV system was established to experimentally demonstrate the arc-detection method in the practical system. The experimental results confirmed that the technique could identify the dc arc correctly with an extended working region despite the noise of power electronics in MPPT. It was verified that the arc-detection system is capable of discriminating the normal operation and arc fault in the PV system. This technique is applicable and promising for arc-faults detection in PV systems.

Original languageEnglish (US)
Article number8693915
JournalIEEE Transactions on Magnetics
Issue number7
StatePublished - Jul 2019
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Electrical and Electronic Engineering


  • Direct current (dc) arc fault
  • fault detection
  • magnetoresistive sensor
  • photovoltaic (PV) system


Dive into the research topics of 'Arc-Faults Detection in PV Systems by Measuring Pink Noise with Magnetic Sensors'. Together they form a unique fingerprint.

Cite this