DHMGAT: A Dynamic and Hierarchical Multi-Head Graph Attention Network for Fault Location in Distribution Networks

Wang, Linfeng; Liu, Hang; Dong, Yu; Feng, Shengtao; Li, Xuefei; Liu, Ziqian; Li, Guohao; Zhou, Jiajun

doi:10.3390/en19133100

This is an early access version, the complete PDF, HTML, and XML versions will be available soon.

Open AccessArticle

DHMGAT: A Dynamic and Hierarchical Multi-Head Graph Attention Network for Fault Location in Distribution Networks

by

Linfeng Wang

¹,

Hang Liu

¹,

Yu Dong

¹,

Shengtao Feng

¹,

Xuefei Li

^2,*

,

Ziqian Liu

²

,

Guohao Li

²

and

Jiajun Zhou

²

¹

State Grid Handan Electric Power Supply Company, Handan 056000, China

²

School of Computer Science, Wuhan University, Wuhan 430072, China

^*

Author to whom correspondence should be addressed.

Energies 2026, 19(13), 3100; https://doi.org/10.3390/en19133100

Submission received: 27 May 2026 / Revised: 24 June 2026 / Accepted: 25 June 2026 / Published: 30 June 2026

(This article belongs to the Special Issue Transforming Power Systems and Smart Grids with Deep Learning)

Download

Browse Figure

Versions Notes

Abstract

Fault location in distribution networks is challenged by dynamic topology changes and heterogeneous equipment. This paper proposes a Dynamic and Hierarchical Multi-Head Graph Attention Network (DHMGAT) that overcomes the limitations of static graph assumptions. Unlike methods that treat network structure as fixed or neglect line parameters, DHMGAT employs a hierarchical multi-head attention mechanism to encode topology dynamically. An Edge Feature Encoding Module fuses physical line attributes—impedance and switch states—directly into node embeddings. A Topology-Gated Pooling mechanism adapts to radial structural variations, and a Physics-Constrained Data Augmentation strategy ensures robustness under limited-sample and anomalous-data conditions. Evaluated on the IEEE 33-node and IEEE 123-node systems under comprehensive fault scenarios, DHMGAT achieves localization accuracies of 96.70% and 94.31%, respectively, with near-perfect calibration (ECE = 0.066). It maintains accuracy above 92% under high-noise conditions and N-1 topological reconfiguration, and above 88% under severe feature loss (up to 30% missing data), substantially outperforming conventional graph neural networks.

Keywords: distribution network; graph attention networks; fault location; dynamic topology; deep learning

Graphical Abstract

Share and Cite

MDPI and ACS Style

Wang, L.; Liu, H.; Dong, Y.; Feng, S.; Li, X.; Liu, Z.; Li, G.; Zhou, J. DHMGAT: A Dynamic and Hierarchical Multi-Head Graph Attention Network for Fault Location in Distribution Networks. Energies 2026, 19, 3100. https://doi.org/10.3390/en19133100

AMA Style

Wang L, Liu H, Dong Y, Feng S, Li X, Liu Z, Li G, Zhou J. DHMGAT: A Dynamic and Hierarchical Multi-Head Graph Attention Network for Fault Location in Distribution Networks. Energies. 2026; 19(13):3100. https://doi.org/10.3390/en19133100

Chicago/Turabian Style

Wang, Linfeng, Hang Liu, Yu Dong, Shengtao Feng, Xuefei Li, Ziqian Liu, Guohao Li, and Jiajun Zhou. 2026. "DHMGAT: A Dynamic and Hierarchical Multi-Head Graph Attention Network for Fault Location in Distribution Networks" Energies 19, no. 13: 3100. https://doi.org/10.3390/en19133100

APA Style

Wang, L., Liu, H., Dong, Y., Feng, S., Li, X., Liu, Z., Li, G., & Zhou, J. (2026). DHMGAT: A Dynamic and Hierarchical Multi-Head Graph Attention Network for Fault Location in Distribution Networks. Energies, 19(13), 3100. https://doi.org/10.3390/en19133100

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

DHMGAT: A Dynamic and Hierarchical Multi-Head Graph Attention Network for Fault Location in Distribution Networks

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI