EUVIMEDEUVIMED entdecken
Crossref

Spatiotemporal PMU and SCADA Measurement-Based Physics-Embedded Deep Learning Method for Fault Detection and Localization in Active Distribution Networks

Donglei Cao, Junjun Xu, Tengfei Zhang, Qinran Hu, Zaijun Wu

IEEE Transactions on Instrumentation and Measurement · 2026 · 75 · 9009213-9009213

DOI: 10.1109/tim.2026.3744011 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Instrumentation and Measurement, 2026. Verfasst von Donglei Cao und 4 weiteren Autor:innen. Die Arbeit nennt 25 Literaturangaben. Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Active distribution networks with high-penetration distributed generation (DG) suffer from poor fault diagnosis performance: traditional physical models depend on precise grid parameters, while pure data-driven deep learning ignores electrical propagation laws, leading to weak robustness under measurement noise and DG fluctuations. This article proposes a physics-embedded spatiotemporal graph convolutional network (STGCN) for joint fault detection and localization using hybrid phasor measurement unit (PMU)/supervisory control and data acquisition (SCADA) measurements. Normalized node admittance and symmetrical sequence components are embedded into graph convolution to capture spatial electrical coupling and temporal fault transients, and a multilabel classifier identifies seven typical faults, including short circuits, DG failures, load redistribution attacks (LRAs), and communication faults. A mutual information (MI) optimization module further extracts key measurement nodes and features to realize interpretable high-precision fault localization under severe interference. Simulations on a modified IEEE 123-node system with synthetic and noisy PMU/SCADA datasets show the framework outperforms benchmarks in efficiency, accuracy, robustness, and adaptability, balancing physical interpretability, anti-interference ability, and real-time inference for practical distribution network monitoring, with consistent superiority verified under both mild and severe interference datasets. This demonstrates great potential for real-time measurement, monitoring, and fault diagnosis in modern active distribution networks.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Cao, D., Xu, J., Zhang, T., Hu, Q., & Wu, Z. (2026). Spatiotemporal PMU and SCADA Measurement-Based Physics-Embedded Deep Learning Method for Fault Detection and Localization in Active Distribution Networks. IEEE Transactions on Instrumentation and Measurement, 75, 9009213-9009213. https://doi.org/10.1109/tim.2026.3744011

Vancouver

Cao D, Xu J, Zhang T, Hu Q, Wu Z. Spatiotemporal PMU and SCADA Measurement-Based Physics-Embedded Deep Learning Method for Fault Detection and Localization in Active Distribution Networks. IEEE Transactions on Instrumentation and Measurement. 2026;75:9009213-9009213. doi:10.1109/tim.2026.3744011

BibTeX
@article{Cao2026,
  author = {Cao, Donglei and Xu, Junjun and Zhang, Tengfei and Hu, Qinran and Wu, Zaijun},
  title = {Spatiotemporal PMU and SCADA Measurement-Based Physics-Embedded Deep Learning Method for Fault Detection and Localization in Active Distribution Networks},
  journal = {IEEE Transactions on Instrumentation and Measurement},
  year = {2026},
  volume = {75},
  pages = {9009213--9009213},
  doi = {10.1109/tim.2026.3744011}
}
Kostenlos starten →