Flow forecasting for leakage burst prediction in water distribution systems using long short-term memory neural networks and Kalman filtering DOI Creative Commons
Lauren McMillan, Jawad Fayaz, Liz Varga

и другие.

Sustainable Cities and Society, Год журнала: 2023, Номер 99, С. 104934 - 104934

Опубликована: Сен. 12, 2023

Reducing pipe leakage is one of the top priorities for water companies, with many investing in higher quality sensor coverage to improve flow forecasting and detection leaks. Most research on this topic focused through analysis data from district metered areas (DMAs), aiming identify bursts after their occurrence. This study a step towards development 'self-healing' infrastructure systems. In particular, machine learning deep learning-based algorithms are applied anomalous experienced during (new leakage) DMAs at various temporal scales, thereby aiding health monitoring distribution uses dataset over 2,000 Yorkshire, containing time series recorded 15-minute intervals year. Firstly, method isolation forests used anomalies dataset, which verified as corresponding entries mains repair log, indicating occurrence bursts. Going beyond detection, proposes hybrid framework named FLUIDS (Forecasting Leakage Usual Intelligently Distribution Systems). A recurrent neural network (RNN) mean forecasting, then combined forecasted residuals obtained real-time Kalman filter. While providing expected day-to-day demands, also aims issue sufficient early warning any upcoming or possible leakages. For given forecast period, can be compute probability exceeding pre-defined threshold, thus allowing decisions made regarding necessary interventions. inform targeted strategies that best utilize resources minimize disruptions by addressing detected predicted burst events. The proposed statistically assessed compared against state-of-practice minimum night (MNF) methodology. Finally, it concluded performs well unobserved test both regular flows.

Язык: Английский

Socio-technical challenges towards data-driven and integrated urban water management: A socio-technical network approach DOI Creative Commons
Liliane Manny

Sustainable Cities and Society, Год журнала: 2022, Номер 90, С. 104360 - 104360

Опубликована: Дек. 24, 2022

Data-driven and integrated urban water management have been proposed to reduce surface pollution in light of climate change urbanization impacts. Besides technological innovation, data-driven require information exchange among many actors, e.g., operators, engineers, or authorities. With the aim achieving a more profound understanding socio-technical infrastructures, such as systems, I draw on approach networks study actors infrastructure elements well multiple relations in-between. In this article, investigate whether underlying dependencies influence social interactions exchange. More specifically related management, analyze potential challenges, organizational fragmentation, data access, diverging perceptions. Based empirical from three case studies Switzerland, provide inferential results obtained fitting exponential random graph models. Findings showed that actors’ relatedness affects their Among cases, presence challenges varied is potentially contingent upon system size, form, progress terms management. Thus, incorporating perspective could help improve policy design implementation aiming achieve sustainable cities.

Язык: Английский

Процитировано

26
Advanced Techniques for Monitoring and Management of Urban Water Infrastructures—An Overview DOI Open Access
Anca Hângan, Costin-Gabriel Chiru, Diana Arsene

и другие.

Water, Год журнала: 2022, Номер 14(14), С. 2174 - 2174

Опубликована: Июль 9, 2022

Water supply systems are essential for a modern society. This article presents an overview of the latest research related to information and communication technology water resource monitoring, control management. The main objective our review is show how emerging technologies offer support smart administration infrastructures. paper covers results cities, big data, data analysis decision support. Our evaluation reveals that there many possible solutions generated through combinations advanced methods. Emerging open new possibilities including functionalities such as social involvement in offers researchers area monitoring management identify useful models designing better solutions.

Язык: Английский

Процитировано

24
Micro hydro power generation in water distribution networks through the optimal pumps-as-turbines sizing and control DOI
Michael K. Kostner, Ariele Zanfei, Jacopo C. Alberizzi

и другие.

Applied Energy, Год журнала: 2023, Номер 351, С. 121802 - 121802

Опубликована: Сен. 7, 2023

Язык: Английский

Процитировано

15
An unsupervised method to exploit low-resolution water meter data for detecting end-users with abnormal consumption: Employing the DBSCAN and time series complexity DOI
Hani Ghamkhar, Mohammadreza Jalili Ghazizadeh, Seyed Hossein Mohajeri

и другие.

Sustainable Cities and Society, Год журнала: 2023, Номер 94, С. 104516 - 104516

Опубликована: Март 17, 2023

Язык: Английский

Процитировано

14
Flow forecasting for leakage burst prediction in water distribution systems using long short-term memory neural networks and Kalman filtering DOI Creative Commons
Lauren McMillan, Jawad Fayaz, Liz Varga

и другие.

Sustainable Cities and Society, Год журнала: 2023, Номер 99, С. 104934 - 104934

Опубликована: Сен. 12, 2023

Reducing pipe leakage is one of the top priorities for water companies, with many investing in higher quality sensor coverage to improve flow forecasting and detection leaks. Most research on this topic focused through analysis data from district metered areas (DMAs), aiming identify bursts after their occurrence. This study a step towards development 'self-healing' infrastructure systems. In particular, machine learning deep learning-based algorithms are applied anomalous experienced during (new leakage) DMAs at various temporal scales, thereby aiding health monitoring distribution uses dataset over 2,000 Yorkshire, containing time series recorded 15-minute intervals year. Firstly, method isolation forests used anomalies dataset, which verified as corresponding entries mains repair log, indicating occurrence bursts. Going beyond detection, proposes hybrid framework named FLUIDS (Forecasting Leakage Usual Intelligently Distribution Systems). A recurrent neural network (RNN) mean forecasting, then combined forecasted residuals obtained real-time Kalman filter. While providing expected day-to-day demands, also aims issue sufficient early warning any upcoming or possible leakages. For given forecast period, can be compute probability exceeding pre-defined threshold, thus allowing decisions made regarding necessary interventions. inform targeted strategies that best utilize resources minimize disruptions by addressing detected predicted burst events. The proposed statistically assessed compared against state-of-practice minimum night (MNF) methodology. Finally, it concluded performs well unobserved test both regular flows.

Язык: Английский

Процитировано

14