EUVIMEDEUVIMED entdecken
Crossref

Multi-Task Graph Learning for Resilience-Oriented Adaptive Microgrid Formation in Active Distribution Networks

Roshni Anna Jacob, Jie Zhang

IEEE Transactions on Power Delivery · 2026-08 · 41 (4) · 1541-1554

DOI: 10.1109/tpwrd.2026.3669817 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Power Delivery, 2026-08. Verfasst von Roshni Anna Jacob und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 47 Literaturangaben. Gefördert von: Department of the Navy, Office of Naval Research, National Science Foundation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

There is an alarming increase in extreme weather-related power outages, resulting in disruptions in the distribution network. With the availability of remote-controlled switches and distributed energy resources (DERs) in modern-day distribution systems, forming microgrids around grid-forming DERs can ensure continued power supply during outages. Such distributed and localized services can be used to enhance network resilience. Traditionally, microgrid formation has been treated as a proactive or restorative measure, with limited focus on its role in enabling immediate operational resilience following disruptions. Moreover, the adaptive potential of microgrids, accounting for spatial and temporal outage variability and dynamic network states, remains underexplored in the literature. This paper addresses this gap by integrating real-time automated control with microgrid formation using deep graph learning for outage mitigation. The partitioning of the distribution network into different clusters serviced by DERs is formulated as an edge classification problem, while load shedding to minimize load-generation imbalance is a node classification problem at the graph level. A multi-task learning framework based on graph convolution networks is developed to determine the microgrid boundaries and load pickup (or shedding). The proposed framework, implemented on the IEEE 13-bus, 37-bus, 123-bus, and a section of the 9500-bus distribution test networks, demonstrates real-time capability with generalized and high accuracy solutions.

Förderung

  • Department of the Navy, Office of Naval Research
  • National Science Foundation

Zitieren

APA

Jacob, R. A., & Zhang, J. (2026). Multi-Task Graph Learning for Resilience-Oriented Adaptive Microgrid Formation in Active Distribution Networks. IEEE Transactions on Power Delivery, 41(4), 1541-1554. https://doi.org/10.1109/tpwrd.2026.3669817

Vancouver

Jacob RA, Zhang J. Multi-Task Graph Learning for Resilience-Oriented Adaptive Microgrid Formation in Active Distribution Networks. IEEE Transactions on Power Delivery. 2026;41(4):1541-1554. doi:10.1109/tpwrd.2026.3669817

BibTeX
@article{Jacob2026,
  author = {Jacob, Roshni Anna and Zhang, Jie},
  title = {Multi-Task Graph Learning for Resilience-Oriented Adaptive Microgrid Formation in Active Distribution Networks},
  journal = {IEEE Transactions on Power Delivery},
  year = {2026},
  volume = {41},
  number = {4},
  pages = {1541--1554},
  doi = {10.1109/tpwrd.2026.3669817}
}
Kostenlos starten →