Leveraging graph-based leak localization in water distribution networks

Registro completo de metadados
MetadadosDescriçãoIdioma
Autor(es): dc.contributorUniversidade Estadual Paulista (UNESP)-
Autor(es): dc.creatorRolle, Rodrigo P.-
Autor(es): dc.creatorRodrigues, Weliton C.-
Autor(es): dc.creatorTomazini, Lucas R.-
Autor(es): dc.creatorMonteiro, Lucas N.-
Autor(es): dc.creatorGodoy, Eduardo P.-
Data de aceite: dc.date.accessioned2025-08-21T21:46:09Z-
Data de disponibilização: dc.date.available2025-08-21T21:46:09Z-
Data de envio: dc.date.issued2025-04-29-
Data de envio: dc.date.issued2023-12-31-
Fonte completa do material: dc.identifierhttp://dx.doi.org/10.1109/MetroInd4.0IoT61288.2024.10584129-
Fonte completa do material: dc.identifierhttps://hdl.handle.net/11449/308101-
Fonte: dc.identifier.urihttp://educapes.capes.gov.br/handle/11449/308101-
Descrição: dc.descriptionIn recent years, the technological resources of the Digital Era have been employed to improve sustainability, especially in the cities. As the global population increases and moves to urban areas, there is a growing need for utilities such as water, electricity, gas, and others. The current Smart City concept is strongly tied to Information and Communication Technology (ICT) towards improving sustainability and ensuring efficient usage of scarce resources in the urban area. This paper intends to evaluate Graph Neural Networks (GNN), a class of graph-based Deep Learning algorithms, to detect and locate water leakage under the data availability restrictions of Low Power Wide-Area Networks (LPWAN), a popular class of wireless sensor networks in IoT/Smart Cities applications. A case study Water Distribution Network (WDN) was developed to obtain data for training and validation. Also, linear regression was employed to minimize the number of sensor nodes, aiming to reduce implementation costs. The results indicate that the graph-based approach tied with linear regression in intermediate nodes can provide up to 80% accuracy, even under the data restrictions of LPWAN. Also, the usage of linear regression improved the mean accuracy of the GNN algorithm by approximately 18% in all three simulated cases in comparison to the situation without data from intermediate (junction) nodes, even with 37% fewer sensor nodes available.-
Descrição: dc.descriptionSão Paulo State University (UNESP)-
Descrição: dc.descriptionSão Paulo State University (UNESP)-
Formato: dc.format192-197-
Idioma: dc.languageen-
Relação: dc.relation2024 IEEE International Workshop on Metrology for Industry 4.0 and IoT, MetroInd4.0 and IoT 2024 - Proceedings-
???dc.source???: dc.sourceScopus-
Palavras-chave: dc.subjectGraph Learning-
Palavras-chave: dc.subjectLeakage Localization-
Palavras-chave: dc.subjectLP-WAN-
Palavras-chave: dc.subjectSmart Cities-
Título: dc.titleLeveraging graph-based leak localization in water distribution networks-
Tipo de arquivo: dc.typeaula digital-
Aparece nas coleções:Repositório Institucional - Unesp

Não existem arquivos associados a este item.