A Novel Series Arc Fault Detection Method Based on Modes-Based Wigner-Ville Distribution and Feature Enhancement Network
IEEE Transactions on Instrumentation and Measurement · 2026 · 75 · 3517213-3517213
DOI: 10.1109/tim.2026.3704190 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Instrumentation and Measurement, 2026. Verfasst von Dejie Sheng und 5 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Gefördert von: Science and Technology Project of Hebei Education Department, National Natural Science Foundation of China, Tianjin Science and Technology Plan Project.
Abstract (Originalsprache)
With the increase in load type and power consumption complexity in low-voltage distribution systems, the features of alternating current (AC) arc faults are seriously obscured, which can easily lead to missed detection. Therefore, it is crucial to deeply explore and enhance arc fault features to achieve reliable detection of arc faults in low-voltage power consumption scenarios. In this article, an arc fault test platform is built for the real electricity scene, and we collected arc fault current signals under 56 actual working conditions. Aiming to address the challenges of time–frequency feature aliasing and the weak features of low-current arc faults (which result from regular current masking), a method utilizing modes-based Wigner-Ville distribution (MBWVD) for extracting key frequency band information and constructing time–frequency spectra is proposed, thereby effectively improving feature separability in the time–frequency domain. Subsequently, a feature enhancement module (FEM) based on the high-frequency features and randomness of arc current is designed and integrated with a deep-learning detection model to realize the adaptive enhancement of key features. Experiments show that the model achieves a detection accuracy of 99.65% under masking conditions and low-current arc fault conditions. Finally, the arc fault detection prototype is developed, and the function verification experiment is carried out. The test results of the prototype demonstrate that the detection time meets the requirements of the IEC 62606 standard ( $\le 120$ ms) under various load types, complex combined working conditions, and severe masking effects, validating its feasibility for practical applications.
Förderung
- Science and Technology Project of Hebei Education Department
- National Natural Science Foundation of China
- Tianjin Science and Technology Plan Project
Zitieren
APA
Sheng, D., Xu, J., Chang, G., Zhang, Z., Guo, Z., & Wang, Y. (2026). A Novel Series Arc Fault Detection Method Based on Modes-Based Wigner-Ville Distribution and Feature Enhancement Network. IEEE Transactions on Instrumentation and Measurement, 75, 3517213-3517213. https://doi.org/10.1109/tim.2026.3704190
Vancouver
Sheng D, Xu J, Chang G, Zhang Z, Guo Z, Wang Y. A Novel Series Arc Fault Detection Method Based on Modes-Based Wigner-Ville Distribution and Feature Enhancement Network. IEEE Transactions on Instrumentation and Measurement. 2026;75:3517213-3517213. doi:10.1109/tim.2026.3704190
BibTeX
@article{Sheng2026,
author = {Sheng, Dejie and Xu, Jiahao and Chang, Guanghui and Zhang, Zizhe and Guo, Zhitao and Wang, Yao},
title = {A Novel Series Arc Fault Detection Method Based on Modes-Based Wigner-Ville Distribution and Feature Enhancement Network},
journal = {IEEE Transactions on Instrumentation and Measurement},
year = {2026},
volume = {75},
pages = {3517213--3517213},
doi = {10.1109/tim.2026.3704190}
}
EUVIMED