Machine learning approach to detect arc faults based on regular coupling features

Run Jiang, Guanghai Bao, Qiteng Hong, Campbell Booth

Research output: Contribution to journalArticlepeer-review

9 Downloads (Pure)


During AC series arc faults (SAFs), arcing current features can change or vanish under different conditions. The phenomena make it challenging to detect SAFs. To address the issues, this paper presents a detection model based on regular coupling features (RCFs). After the model is only trained by the samples in single-load circuits, it can detect SAFs under unknown multi-load circuits. To extract RCFs, asymmetric magnetic flux is coupled by passing the live line and the neutral line through the current transformer. According to the unique signals, two time-domain features and one frequency-domain feature are extracted to represent RCFs, including impulse -factor analysis, covariance-matrix analysis and multiple frequency-band analysis. Then, the impulse factor and its threshold are used to preprocess the signals and decrease analysis complexity for the classifier. Finally, the experimental results show that the proposed method has significantly improved generalization ability and detection accuracy in SAF detection.

Original languageEnglish
Number of pages10
JournalIEEE Transactions on Industrial Informatics
Early online date23 Feb 2022
Publication statusE-pub ahead of print - 23 Feb 2022


  • arc faults
  • fault detection
  • machine learning

Cite this