Learning physics and temporal dependencies: real-time modeling of water distribution systems via Kolmogorov–Arnold attention networks

Abstract Real-time modeling is vital for the intelligent management of urban water distribution systems (WDSs), enabling proactive decision-making, rapid anomaly detection, and efficient operational control. In comparison with traditional mechanistic simulators, data-driven models offer faster compu...

Full description

Saved in:
Bibliographic Details
Main Authors: Zekun Zou, Zhihong Long, Gang Xu, Raziyeh Farmani, Tingchao Yu, Shipeng Chu
Format: Article
Language:English
Published: Nature Portfolio 2025-08-01
Series:npj Clean Water
Online Access:https://doi.org/10.1038/s41545-025-00505-y
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849767333084528640
author Zekun Zou
Zhihong Long
Gang Xu
Raziyeh Farmani
Tingchao Yu
Shipeng Chu
author_facet Zekun Zou
Zhihong Long
Gang Xu
Raziyeh Farmani
Tingchao Yu
Shipeng Chu
author_sort Zekun Zou
collection DOAJ
description Abstract Real-time modeling is vital for the intelligent management of urban water distribution systems (WDSs), enabling proactive decision-making, rapid anomaly detection, and efficient operational control. In comparison with traditional mechanistic simulators, data-driven models offer faster computation and reduced calibration demands, making them more suitable for real-time applications. However, existing models often accumulate long-term prediction errors and fail to capture the strong temporal dependencies in measured time series. To address these challenges, this study proposes the Kolmogorov–Arnold Attention Network for the real-time modeling of WDSs (KANSA), which combines Kolmogorov–Arnold Networks with attention mechanisms to extract temporal dependency features through bidirectional spatiotemporal processing. Additionally, a multi-equation soft-constraint formulation embeds mass and energy conservation laws into the loss function, mitigating cumulative errors and enhancing physical consistency. Evaluations on a benchmark network and a real-world system demonstrate that KANSA achieves high-accuracy real-time estimation and pattern fidelity while maintaining engineering-grade hydraulic balance.
format Article
id doaj-art-eda8324d24ff4bc7beab9e1e57d90cfa
institution DOAJ
issn 2059-7037
language English
publishDate 2025-08-01
publisher Nature Portfolio
record_format Article
series npj Clean Water
spelling doaj-art-eda8324d24ff4bc7beab9e1e57d90cfa2025-08-20T03:04:16ZengNature Portfolionpj Clean Water2059-70372025-08-018111210.1038/s41545-025-00505-yLearning physics and temporal dependencies: real-time modeling of water distribution systems via Kolmogorov–Arnold attention networksZekun Zou0Zhihong Long1Gang Xu2Raziyeh Farmani3Tingchao Yu4Shipeng Chu5College of Civil Engineering and Architecture, Zhejiang UniversityGuangzhou Water Supply Co. LtdCollege of Civil Engineering and Architecture, Zhejiang UniversityCentre for Water Systems, Department of Engineering, University of Exeter, ExeterCollege of Civil Engineering and Architecture, Zhejiang UniversityCollege of Civil Engineering and Architecture, Zhejiang UniversityAbstract Real-time modeling is vital for the intelligent management of urban water distribution systems (WDSs), enabling proactive decision-making, rapid anomaly detection, and efficient operational control. In comparison with traditional mechanistic simulators, data-driven models offer faster computation and reduced calibration demands, making them more suitable for real-time applications. However, existing models often accumulate long-term prediction errors and fail to capture the strong temporal dependencies in measured time series. To address these challenges, this study proposes the Kolmogorov–Arnold Attention Network for the real-time modeling of WDSs (KANSA), which combines Kolmogorov–Arnold Networks with attention mechanisms to extract temporal dependency features through bidirectional spatiotemporal processing. Additionally, a multi-equation soft-constraint formulation embeds mass and energy conservation laws into the loss function, mitigating cumulative errors and enhancing physical consistency. Evaluations on a benchmark network and a real-world system demonstrate that KANSA achieves high-accuracy real-time estimation and pattern fidelity while maintaining engineering-grade hydraulic balance.https://doi.org/10.1038/s41545-025-00505-y
spellingShingle Zekun Zou
Zhihong Long
Gang Xu
Raziyeh Farmani
Tingchao Yu
Shipeng Chu
Learning physics and temporal dependencies: real-time modeling of water distribution systems via Kolmogorov–Arnold attention networks
npj Clean Water
title Learning physics and temporal dependencies: real-time modeling of water distribution systems via Kolmogorov–Arnold attention networks
title_full Learning physics and temporal dependencies: real-time modeling of water distribution systems via Kolmogorov–Arnold attention networks
title_fullStr Learning physics and temporal dependencies: real-time modeling of water distribution systems via Kolmogorov–Arnold attention networks
title_full_unstemmed Learning physics and temporal dependencies: real-time modeling of water distribution systems via Kolmogorov–Arnold attention networks
title_short Learning physics and temporal dependencies: real-time modeling of water distribution systems via Kolmogorov–Arnold attention networks
title_sort learning physics and temporal dependencies real time modeling of water distribution systems via kolmogorov arnold attention networks
url https://doi.org/10.1038/s41545-025-00505-y
work_keys_str_mv AT zekunzou learningphysicsandtemporaldependenciesrealtimemodelingofwaterdistributionsystemsviakolmogorovarnoldattentionnetworks
AT zhihonglong learningphysicsandtemporaldependenciesrealtimemodelingofwaterdistributionsystemsviakolmogorovarnoldattentionnetworks
AT gangxu learningphysicsandtemporaldependenciesrealtimemodelingofwaterdistributionsystemsviakolmogorovarnoldattentionnetworks
AT raziyehfarmani learningphysicsandtemporaldependenciesrealtimemodelingofwaterdistributionsystemsviakolmogorovarnoldattentionnetworks
AT tingchaoyu learningphysicsandtemporaldependenciesrealtimemodelingofwaterdistributionsystemsviakolmogorovarnoldattentionnetworks
AT shipengchu learningphysicsandtemporaldependenciesrealtimemodelingofwaterdistributionsystemsviakolmogorovarnoldattentionnetworks