跳到正文
arXiv cs.AI· Ojas Sanghi, Anvita Bhagavathula, Alvaro Carbonero, Priya L. Donti·· 1 小时前AI 评分25

神经网络交流潮流计算的分布外泛化研究

Toward Out-of-Distribution Generalization in Neural AC Power Flow

AI 导读

一项 NeurIPS 2026 AI4PowerGrids 研讨会口头报告的系统评估显示,物理信息损失、多电网训练和扰动数据增强可提升基于图神经网络的潮流模型从较小系统向未见的 500 母线系统的零样本迁移能力。其中基于边属性扰动的数据增强最多可将迁移性能提升 54.6%,物理信息损失优于纯监督目标;多电网训练仅匹配最佳单一电网,随机线路/发电机 dropout 则无稳定收益。

正文 · 原文

View PDF HTML (experimental)

Abstract:Growing uncertainty in power system operations from renewables integration, extreme weather, and data center loads challenges the reliability of traditional power flow solvers at scale. Machine learning-based proxies are a potentially promising alternative, but the prohibitive cost of generating training data for large grids makes out-of-distribution generalization a key capability for deployment; currently, models trained on small grids lack that capability. In this work, we systematically evaluate how physics-informed losses, multi-grid training, and perturbation-based data augmentation improve the zero-shot transfer of three graph neural network-based power flow models from smaller systems to a larger, unseen 500-bus system. We find that (1) physics-informed losses improve transfer relative to purely supervised objectives; (2) multi-grid training on smaller systems matches the best single grid but does not close the gap to target-grid training; (3) perturbation-based augmentation on edge attributes improves transfer by up to 54.6%; and (4) random line/generator dropout does not consistently improve performance.
Comments: 11 pages, 5 figures. Accepted as an oral presentation at the NeurIPS 2026 Workshop on AI Foundations for Power Grids: From Models to Deployment at Scale (AI4PowerGrids), Sydney, Australia
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.23729 [cs.LG]
  (or arXiv:2611.23729v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2611.23729

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Ojas Sanghi [view email]
[v1] Wed, 18 Nov 2026 02:41:59 UTC (144 KB)

来源:arXiv cs.AI · arxiv.org