{"title":"ST-GPINN:一个时空图物理信息神经网络,用于增强配水系统的水质预测","authors":"Tianwei Mu, Feiyu Duan, Baokuan Ning, Bo Zhou, Junyu Liu, Manhong Huang","doi":"10.1038/s41545-025-00499-7","DOIUrl":null,"url":null,"abstract":"<p>Data-driven models often neglect the underlying physical principles, limiting generalization capabilities in water distribution systems (WDSs). This study presents a novel spatio-temporal graph physics-informed neural network (ST-GPINN) for water quality prediction in WDSs, integrating hydraulic simulations, physics-informed neural networks (PINNs), and graph neural networks (GNNs) to capture dynamics and graph-based network connectivity while approximating partial differential equations (PDEs). ST-GPINN discretizes WDSs using virtual nodes to enhance spatial granularity, employs an Encoder-Processor-Decoder architecture for predictions. Validated on Network A (a small-scale network with 9 junctions and 11 pipes) and Network B (a real large-scale WDS with 920 junctions and 1032 pipes), ST-GPINN outperforms others, achieving a MAE of 0.0073 mg/L, RMSE of 0.0121 mg/L, and <i>R</i><sup>2</sup> of 88.91% in Network A, and a MAE of 0.008 mg/L, RMSE of 0.0098 mg/L, and <i>R</i>² of 98.91% in Network B. Its scalability and accuracy highlight ST-GPINN’s potential for water quality predictions.</p><figure></figure>","PeriodicalId":19375,"journal":{"name":"npj Clean Water","volume":"14 1","pages":""},"PeriodicalIF":11.4000,"publicationDate":"2025-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"ST-GPINN: a spatio-temporal graph physics-informed neural network for enhanced water quality prediction in water distribution systems\",\"authors\":\"Tianwei Mu, Feiyu Duan, Baokuan Ning, Bo Zhou, Junyu Liu, Manhong Huang\",\"doi\":\"10.1038/s41545-025-00499-7\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>Data-driven models often neglect the underlying physical principles, limiting generalization capabilities in water distribution systems (WDSs). This study presents a novel spatio-temporal graph physics-informed neural network (ST-GPINN) for water quality prediction in WDSs, integrating hydraulic simulations, physics-informed neural networks (PINNs), and graph neural networks (GNNs) to capture dynamics and graph-based network connectivity while approximating partial differential equations (PDEs). ST-GPINN discretizes WDSs using virtual nodes to enhance spatial granularity, employs an Encoder-Processor-Decoder architecture for predictions. Validated on Network A (a small-scale network with 9 junctions and 11 pipes) and Network B (a real large-scale WDS with 920 junctions and 1032 pipes), ST-GPINN outperforms others, achieving a MAE of 0.0073 mg/L, RMSE of 0.0121 mg/L, and <i>R</i><sup>2</sup> of 88.91% in Network A, and a MAE of 0.008 mg/L, RMSE of 0.0098 mg/L, and <i>R</i>² of 98.91% in Network B. Its scalability and accuracy highlight ST-GPINN’s potential for water quality predictions.</p><figure></figure>\",\"PeriodicalId\":19375,\"journal\":{\"name\":\"npj Clean Water\",\"volume\":\"14 1\",\"pages\":\"\"},\"PeriodicalIF\":11.4000,\"publicationDate\":\"2025-08-07\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"npj Clean Water\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://doi.org/10.1038/s41545-025-00499-7\",\"RegionNum\":1,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, CHEMICAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"npj Clean Water","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1038/s41545-025-00499-7","RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, CHEMICAL","Score":null,"Total":0}
ST-GPINN: a spatio-temporal graph physics-informed neural network for enhanced water quality prediction in water distribution systems
Data-driven models often neglect the underlying physical principles, limiting generalization capabilities in water distribution systems (WDSs). This study presents a novel spatio-temporal graph physics-informed neural network (ST-GPINN) for water quality prediction in WDSs, integrating hydraulic simulations, physics-informed neural networks (PINNs), and graph neural networks (GNNs) to capture dynamics and graph-based network connectivity while approximating partial differential equations (PDEs). ST-GPINN discretizes WDSs using virtual nodes to enhance spatial granularity, employs an Encoder-Processor-Decoder architecture for predictions. Validated on Network A (a small-scale network with 9 junctions and 11 pipes) and Network B (a real large-scale WDS with 920 junctions and 1032 pipes), ST-GPINN outperforms others, achieving a MAE of 0.0073 mg/L, RMSE of 0.0121 mg/L, and R2 of 88.91% in Network A, and a MAE of 0.008 mg/L, RMSE of 0.0098 mg/L, and R² of 98.91% in Network B. Its scalability and accuracy highlight ST-GPINN’s potential for water quality predictions.
npj Clean WaterEnvironmental Science-Water Science and Technology
CiteScore
15.30
自引率
2.60%
发文量
61
审稿时长
5 weeks
期刊介绍:
npj Clean Water publishes high-quality papers that report cutting-edge science, technology, applications, policies, and societal issues contributing to a more sustainable supply of clean water. The journal's publications may also support and accelerate the achievement of Sustainable Development Goal 6, which focuses on clean water and sanitation.