Journal of Water Resources Planning and Management

Drinking water systems are critical to human health and economic development. System design considers treatment infrastructure to remove contaminants from water supplies and distribution system infrastructure to transport and maintain the quality of treated water to households and businesses. Therefore, treatment and distribution components should be evaluated together when considering system designs and the associated effects on water quality. Planning models are often used to aid decision makers in evaluating drinking water system changes over time. This work proposes a planning modeling approach to determine the optimal treatment and alternative supply infrastructure location and capacity within an existing distribution system. The model consists of four stages: (1) simulating system hydraulic and water quality behavior to set optimization model parameters; (2) selecting potential system configurations through optimization; (3) validating the system hydraulic and water quality behavior through simulation; and (4) selecting network treatment configuration. Results indicate an optimal number of treatment locations that provide the best delivered water quality. For the demonstration network, more than three plants did not provide any improvement in water quality.

Drinking water systems provide consistent and safe supplies of water across urban areas through the use of treatment and distribution infrastructure. Design of drinking water infrastructure requires incorporation of multiple conflicting stakeholder goals, such as cost and water quality. This work expands on a previously developed integrated modeling framework that uses simulation and optimization models for long-term planning of treatment infrastructure in the context of existing distribution systems. The optimization model component is a multi-objective formulation. The expanded model framework is demonstrated on the USEPA Net3 treatment network. The multiobjective optimization includes three objectives—minimizing cost, maximizing systemwide chlorine residual, and minimizing the demand-weighted water age (a proxy for water quality)—to identify the optimal location and capacity of treatment plants. Results show the trade-offs between each pair of objectives. Of a maximum of five possible treatment locations, three treatment plants were optimal for the water quality objectives, whereas one treatment plant was optimal for the cost objective. The capacity allocation and sites selected for treatment differed between the two water quality objectives.

This study presents a novel approach to real-time water quality state (chlorine concentration) and reaction parameter estimation in water distribution systems (WDSs) using ensemble Kalman filter (EnKF)–based data assimilation techniques. Two different types of EnKF-based methods are used in this study: noniterative restart-EnKF (NIR-EnKF) and iterative restart-EnKF (IR-EnKF). The use of these data assimilation frameworks for addressing key uncertainties in water quality models, such as uncertainty in the source or initial concentration of chlorine and uncertainty in the wall reaction parameter, is studied. The effect of ensemble size, number and location of measurement nodes, measurement error, and noise are also studied extensively in this work. The performance of the proposed methodology is tested on two different water networks: a brushy plains network and a large, citywide WDS, the Bangalore inflow network. The results of the simulation study show that both the NIR-EnKF and IR-EnKF methods are appropriate for dealing with uncertainty in source chlorine concentration, but the IR-EnKF method performs better than the NIR-EnKF method in the case of reaction parameter uncertainty.

Maintaining acceptable drinking water quality in First Nations communities has been an ongoing issue in Canada, resulting in frequent drinking water advisories (DWAs). Information about drinking water systems on reserves was combined with historical advisory data to investigate the relationships between system attributes and DWAs using data mining. Decision trees were developed to identify key factors linked to the occurrence, frequency, duration, and causes of DWAs. The results showed that trends varied considerably by province, despite federal management. Operator treatment certification was shown to be associated with the duration of advisories, where systems with fully trained operators were less likely to have long-lasting DWAs. The findings from data mining can be useful for informing future regulatory and funding decisions in efforts to reduce DWAs.

A new method to identify pipes with insufficient hydraulic capacity is proposed. This method can be applied to assess the future evolution of network performance under climate change (CC). It is based on hydrologic/hydraulic simulations using the Storm Water Management Model (SWMM) and single observed rainfall events. The evolution of the hydraulic performance with time is simulated by increasing the intensity of these rainfall events by a factor depending on the CC predictions for the study area. The proposed method is applied to two Canadian separated and combined sewer networks. The method identified the constraining pipe sections that could cause hydraulic dysfunctions in the networks in both current and future climates. For the two networks, the number of constraining pipes depends on rain events and is anticipated to increase in the future climate. The proposed method can be applied to various types of networks to assess the network performance and project the evolution of the hydraulic performance of individual pipes over time, making it a useful tool for the planning of drainage network renewal under CC.

This case study implements long-term projections of domestic water demand for a UK water company, Thames Water. Projections of per household consumption (PHC) and households were combined to yield future demand. Regression models predicted PHC using the determinants of occupancy, property type, ethnicity and rateable value, drawing on 2006–2015 domestic water-use data as a baseline. A model was developed for diffusing savings in per capita consumption (PCC), drawn from published studies of interventions. PCC declines were converted to PHC reductions using baseline ratios. Interventions were grouped into Business as Usual, Light Green (limited intervention), and Dark Green (extreme intervention) scenarios. Projected households were generated by property type, occupancy, and ethnicity for Thames Water’s resource zones for 2011 to 2101 and multiplied by projected PHCs to yield water-demand projections. By 2101, the 2011 water demand of 1,225 million liters a day grew 90% under Business as Usual, 69% under Light Green, and 46% under Dark Green.

Reliability of water distribution systems is typically divided into mechanical and hydraulic reliability. This paper introduces two reliability surrogate measures: the mechanical reliability estimator and the hydraulic reliability estimator. These estimators were developed based on simplifications of a broader definition of stochastic reliability and were tested in the multiobjective design optimization of the Hanoi and Fossolo benchmark networks. The numerical values of the estimators were compared with existing reliability surrogate measures (the residence index, the network resilience index, and flow entropy) through correlation analysis. Results showed that the proposed estimators concur with existing optimization algorithms. They also show that previous indexes correlate with at most one of the types of reliability, while a trade-off between them seems to appear, especially in one of the study cases.