Water distribution systems (WDSs) face increasing cyber-physical risks, which make reliable anomaly detection essential. Many data-driven models ignore network topology and are hard to interpret, while model-based ones depend strongly on parameter accuracy. This work proposes a hydraulic-aware graph attention network using normalized conservation law violations as features. It combines mass and energy balance residuals with graph attention and bidirectional LSTM to learn spatio-temporal patterns. A multi-scale module aggregates detection scores from node to network level. On the BATADAL dataset, it reaches F1=0.979, showing 3.3pp gain and high robustness under 15% parameter noise.
@article{arxiv.2601.12426,
title = {Graph Attention Networks with Physical Constraints for Anomaly Detection},
author = {Mohammadhossein Homaei and Iman Khazrak and Ruben Molano and Andres Caro and Mar Avila},
journal= {arXiv preprint arXiv:2601.12426},
year = {2026}
}