Topology-Aware Graph Neural Network-based State Estimation for PMU-Unobservable Power Systems

Shiva Moshtagh, Behrouz Azimian, Mohammad Golgol, Anamitra Pal · 2025

Read Topology-Aware Graph Neural Network-based State Estimation for PMU-Unobservable Power Systems by Shiva Moshtagh, Behrouz Azimian, Mohammad Golgol, Anamitra Pal (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-03493.txt.

Description

Moshtagh, Azimian, Golgol, and Pal estimate the state of power grids that lack enough precision sensors to watch directly. Classic estimators are slow online and assume tidy noise, while earlier learning models break when the grid's wiring changes or data drops out. Their 2025 model is a graph neural network built around the grid's own shape.

Layers that pass messages along lines, plus attention heads that weigh key links, let the model cope with wiring changes and missing data. The authors also derive a bound tying estimation error to how much the wiring changed. On several test grids, the model beat classic and learning rivals under wiring changes, sensor failures, bad data, and messy noise. They pitch it for real-time grid control rooms.

Catalog record

CREATORShiva Moshtagh, Behrouz Azimian, Mohammad Golgol, Anamitra Pal
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE65 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-03493

Browse the full library · About the Tower of Babel