Probabilistic Minimum Night Flow Estimation in Water Distribution Networks and Comparison with the Water Balance Approach: Large-Scale Application to the City Center of Patras in Western Greece

Quantification of water losses (WL) in water distribution networks (WDNs) is a crucial task towards the development of proper strategies to reduce them. Currently, WL estimation methods rely on semi-empirical assumptions and different implementation strategies that increase the uncertainty of the obtained estimates. In this work, we compare the effectiveness and robustness of two widely applied WL estimation approaches found in the international literature: (a) the water balance, or top-down, approach introduced by the International Water Association (IWA), and (b) the bottom-up or minimum night flow (MNF) approach, based on a recently proposed probabilistic MNF estimation method. In doing so, we use users’ consumption and flow-pressure data from the 4 largest pressure management areas (PMAs) of the WDN of the city of Patras (the third largest city in Greece), which consist of more than 200 km of pipeline, cover the entire city center of Patras, and serve approximately 58,000 consumers. The obtained results show that: (a) when MNF estimation is done in a rigorous statistical setting from high resolution flow-pressure timeseries, and (b) there is sufficient understanding of the consumption types and patterns during day and night hours, the two approaches effectively converge, allowing for more reliable estimation of the individual WL components. In addition, when high resolution flow-pressure timeseries are available at the inlets of PMAs, the suggested version of the bottom-up approach with probabilistic estimation of MNF should be preferred as less sensitive, while allowing for confidence interval estimation of the individual components of water losses and development of proper strategies to reduce them.

Impact of active night population and leakage exponent on leakage estimation in developing countries

Methods for network leakage estimation include water balance, component analysis and minimum night flow (MNF) methods the latter of which involves subtracting the customer night use (QCNU) from night leakage and multiplying by the hour day factor (HDF). QCNU and HDF respectively depend on Active Night Population (ANP) and leakage exponent (N1). In most developing countries, these parameters are assumed in the MNF method thus introducing errors which makes setting realistic leakage reduction targets and key performance indicators (KPI) problematic. In this study, QCNU and HDF were evaluated by determining the relative error associated with ANP and N1 to establish localized rates for accurately estimating leakage in water networks. Between 7 and 11% relative error was associated with every 1% higher or lower ANP while up to 4% relative error was observed for every step considered. A linear relationship exists between the relative error associated with both and ANP although that of ANP is twice as high as This has technical implications on setting water loss reduction targets and investing in the water infrastructure. It is recommended that water utilities must establish localized ANP and values for accurate leakage estimation in water networks.

Water leakage detection and localization using hydraulic modeling and classification

A significant percentage of treated water is lost due to leakage in water distribution systems. The state-of-the-art leak detection and localization schemes use a hybrid approach of hydraulic modeling and data-driven techniques. Most of these works, however, focus on single leakage detection and localization. In this research, we propose to use combined pressure and flow residual data to detect and localize multiple leaks. The proposed approach has two phases: detection and localization. The detection phase uses the combination of pressure and flow residuals to build a hydraulic model and classification algorithm to identify leaks. The localization phase analyzes the pattern of isolated leak residuals to localize multiple leaks. To evaluate the performance of the proposed approach, we conducted experiments using Hanoi Water Network benchmark and a dataset produced based on LeakDB benchmark's dataset preparation procedure. The result for a well-calibrated hydraulic model shows that leak detection is 100% accurate while localization is 90% accurate, thereby outperforming minimum night flow and raw- and residual-based methods in localizing leaks. The proposed approach performed relatively well with the introduction of demand and noise uncertainty. The proposed localization approach is also able to locate two to four leaks that existed simultaneously.

Analysis of the effect of pressure control on leakages in distribution systems by FAVAD equation and field applications

Pressure has an important effect on the occurrence of failures/leaks in water distribution systems (WDSs) or the change of leakage in existing leakages. For this reason, monitoring the pressure is important especially for analyzing the changes in the day and night, determining the fluctuations and applying pressure management (PM) to ensure normal operating conditions. In this study, the effect of pressure on water losses and minimum night flow (MNF) was carried out according to the Fixed and Varied Area Discharge (FAVAD) approach which allows the amount of leakage to be calculated based on the change in pressure and field tests. The minimum flow rate and potential leakage were determined under the network operating conditions before the pressure control in region. Then, considering the features of the region, pressure was reduced with the pressure control system and MNFs and leaks were monitored. By reducing the pressure from 9.10 bar to 3.2 bar in the region, the MNF rate was reduced from 6.95 l/s to 3.29 l/s. The daily water savings in the system inlet volume is 78.44 m3/day and the annual saving is 28,624 m3 /year. The results obtained are very important for practitioners in terms of implementing PM in the field.

Leakage detection in water pipe networks using machine learning

<p><span>Leakage in the water supply system is a world problem that happens everywhere, not only in China but also in Japan, the US, and Europe. It not only results in the waste of water resources but also raises safety issues in drinking water. The traditional solution is the Minimum Night Flow method with manual leak detectors. This solution could only find leakage at night. The engineers have to search the leaking point randomly using leak detectors. It not only highly relies on domain knowledge and expertise but is also labor-consuming. The response time is quite long, might be a couple of days to several days. Here, time series analysis based on a dynamic time warping algorithm is used to detect anomalies in time series of pressure stations and flow stations, and the risk coefficient of each pipe network is determined by using a neural network combined with existing data. The water treatment plants don't even have to install new sensors if the budget is limited.</span></p>

Practical Estimation of Minimum Night Flow in Water Distribution Networks: Large-scale Application to the City of Patras in Western Greece

<p><strong>Abstract</strong></p><p>In the present work we develop and test a non-parametric statistical methodology to obtain point estimates of Minimum Night Flow (MNF) in Water Distribution Networks (WDNs). The methodology constitutes a simplified version of the approach of Serafeim et al. (2021) for confidence interval estimation of background losses in WDNs, that simultaneously analyzes all night flow measurements, producing robust estimates independent of the nominal resolution of the available data.</p><p>In addition to being simpler to apply and computationally more efficient, the developed method can be applied to any WDN independent of its size, age and overall condition, its  specific geometric characteristics (intensity of altimetry, average diameter etc.), inlet/operating pressures, and the nominal resolution of the flow data.</p><p>The effectiveness of the method is tested via a large-scale application to the WDN of the City of Patras in western Greece, which consists of 79 Pressure Management Areas (PMAs) with more than 700 km of pipeline grid. To do so, we use flow data at 1 min temporal resolution, provided by the Municipal Enterprise of Water Supply and Sewerage of the City of Patras, for the 4-month winter period from 01 November 2018 – 28 February 2019, which are progressively averaged to coarser temporal resolutions, in an effort to test the sensitivity of the developed method to the nominal resolution of the data.  </p><p>The obtained point estimates of MNF are assessed on the basis of the confidence intervals obtained by the approach of Serafeim et al. (2021), highlighting the accuracy and robustness of a simple non-parametric approach in providing MNF point estimates at a minimum of effort.</p><p><strong>Acknowledgements</strong></p><p>The research work was supported by the Hellenic Foundation for Research and Innovation (H.F.R.I.) under the “First Call for H.F.R.I. Research Projects to support Faculty members and Researchers and the procurement of high-cost research equipment grant” (Project Number: 1162).</p><p><strong>References</strong></p><p>Serafeim, A.V., G. Kokosalakis, R. Deidda, I. Karathanasi and A. Langousis, (2021) Probabilistic Estimation of Minimum Night Flow in Water Distribution Networks: Large-scale Application to the City of Patras in Western Greece (submitted).</p>

Analysis of MNF and FAVAD Models for Leakage Characterization by Exploiting Smart-Metered Data: The Case of the Gorino Ferrarese (FE-Italy) District

Leakages in water distribution systems have great economic and environmental impacts and are a major issue for water utilities. In this work, the water balance and the Minimum Night Flow (MNF) method for evaluating the amount of water loss, as well as the power and Fixed and Variable Area Discharge (FAVAD) equations for analyzing the relationship between leakage and pressure, were applied to a fully monitored District Metered Area (DMA) located in Gorino Ferrarese (FE, Italy). Time series of (a) the water consumption of each user, (b) the DMA inflow, and (c) the pressure at the DMA inlet point were monitored with a 5 min time step. The results of an analysis carried out by exploiting the collected time series highlighted that: (a) The application of the MNF method based on literature values can lead to significant inaccuracies in the presence of users with irregular consumption, and (b) the estimation of the parameters of the power and FAVAD equations is highly affected by the amounts and types of observed data used.