Improving leakage management in urban water distribution networks through data analytics and hydraulic simulation

Abstract

Worldwide, water utilities are finding it increasingly difficult to meet the growing water demand. The problem, already acute in view of the urbanization trends, is compounded by the age of the infrastructure: one third of water utilities have 20% or more of their pipelines nearing the end of their useful life. This paper outlines an innovative approach for improving leakage management processes through the adoption of data analytics techniques and hydraulic simulation. More in detail, the aim is to provide analytical leaks localization and correspondent severity estimation in order to reduce time and costs for interventions and rehabilitation while improving asset management. The technical solution has been developed as a set of web services able to interoperate with other technological systems usually adopted by urban water distribution utilities, such as Supervisory Control And Data Acquisition (SCADA) system, Customer Information System (CIS), Geographical Information System (GIS) and Hydraulic Simulation tools. This paper presents some results obtained by the application of the approach to a real case study and also proves how its effectiveness may be improved in a sectorized water distribution network.

Overview of Energy Management and Leakage Control Systems for Smart Water Grids and Digital Water

“…Izquierdo et al [74] proposed a procedure for estimating anomalous pipe states based on neuro-fuzzy theory. Years later, Candelieri et al [75] used a combination of SCADA, GIS, and customer information systems, to better address leakage control. This work has multiple extensions by approaching leakage using big data [76] or by optimal procedures for sensor placement [77].…”

Results from a water-saving maintenance campaign at Cape Town schools in the run-up to Day Zero

“…Numerous studies have used it for this purpose (Oliver, 2005;Cheung et al, 2010;Peters and Ben-Ephraim, 2012;Candelieri et al, 2013;Loureiro et al, 2016;García et al, 2008;Alkasseh et al, 2013), we cite only a few examples here. Although the period for MNF was defined differently for some of those studies, all were in the range of 00:00 to 05:00 daily.…”

Identifying Typical Urban Water Demand Patterns for a Reliable Short-term Forecasting – The Icewater Project Approach

“…As an example, this data has been widely used, even coupled with simulation software and Machine Learning, to offer innovative leakage management functionalities [10][11][12][13][14][15]. More recently, this data have been largely used to perform short-term forecast to dynamically optimize pump scheduling under diurnal and seasonal variations of the water demand, in order to maintain a satisfactory level of the service while reducing costs for caption, treatment, storage and distribution.…”