Journal of Water Resources Planning and Management

Studies have shown that real-time control (RTC) of pressure in water distribution networks (WDNs) can be very effective for leakage reduction. The aim of RTC is to regulate the pressure level over the WDN by guaranteeing a pressure that is nearly constant and as low as possible at the critical node. To this end, a pressure reducing valve (PRV) at the network inlet can be remotely controlled to ensure the optimal service level, regardless of upstream pressure and inflow. Field experiments in a district of Benevento, Italy are carried out to assess the suitability of the proposed approach. Results show that the system guarantees the required pressure at the critical node at all times. The controller is effective in optimizing pressure levels over the entire WDN by also reducing excessive pressure in the other nodes. Experiments show that the pressure control is effective in reducing leakage, as confirmed by the minimum night flow (MNF) measurements.

Evolutionary optimization of reservoir operation is subject to complex physical and operational constraints. Constraint-handling techniques (CHTs) in this field are predominantly problem-specific or based on certain evolutionary algorithms; generally applicable CHTs are seldom tested against reservoir scheduling problems. This study proposes an independent CHT to accommodate the reservoir operation constraints, called the nondomination rank-based adaptive method (NRAM). The NRAM is straightforward to use and free of parameter tuning. The process emphasizes exploiting information from infeasible individuals and preserving them to promote convergence to global optima on a feasible space boundary. Moreover, the method adjusts the population composition dynamically to facilitate exploration or local search. The NRAM was applied to the hydropower scheduling of the Three Gorges Reservoir and Gezhouba Reservoir in China. Results show that the NRAM performs slightly better than three other well-regarded CHTs but requires mildly longer computational time. In addition, the genetic algorithm with the NRAM outperforms dynamic programming that is commonly used for hydropower scheduling. The operation schedules the NRAM provides are well suited for maximizing hydropower generation with all constraints satisfied.

The conventional method of delineating wellhead protection areas using travel times is not well suited for application in wellfields containing seepage lakes. The concentration factor is identified as an intrinsic vulnerability measure that does not have the shortcomings of travel time, and the theory of concentration factors is presented. Idealized scenarios are used to demonstrate the features of using concentration factors. It is shown that using travel times in the presence of seepage lakes can indicate an increased well vulnerability, when in fact there is a reduced vulnerability. It is further demonstrated that in wellfields without seepage lakes, travel times do not give a consistent measure of vulnerability, and concentration factors are a preferred measure.

Two implementations of the Pareto archived dynamically dimensioned search (PADDS) algorithm using different selection metrics are applied to 12 water distribution network (WDN) design benchmark problems from the literature. Convex hull contribution (CHC) and hypervolume contribution (HVC) are used as selection metrics for PADDS making this study the first to assess their relative performance on WDN design problems. Past research applied five state-of-the-art multiobjective evolutionary algorithms (MOEAs) to these 12 benchmark problems to generate the best-known Pareto fronts (PFs). The PADDS-CHC and PADDS-HVC both find all solutions on the known true PFs of the first three problems. Together, both PADDS results augment the previously best-known PFs in the nine other benchmark problems with new PF solutions, some of which dominate previous best-known PF solutions, to define updated best-known PFs. Comparative results against five state-of-the-art MOEAs show PADDS derived best-known PFs are equal or better than all other algorithms in 11 of 12 WDN design problems. A comprehensive comparison between PADDS-CHC and PADDS-HVC performance on the largely convex benchmark problem Pareto fronts reveals the different responses of PADDS algorithm to increment of computational budget. An innovative measure called effective archive size (EAS) is introduced to quantify the portion of PADDS archived solutions that play the dominant role in directing PADDS toward the final PF. Tracking the EAS value throughout the search revealed that compared with PADDS-HVC, the EAS of PADDS-CHC is typically close to an order of magnitude smaller. In fact, the PADDS-CHC algorithm generates candidate solutions from a surprisingly small effective archive size that ranges from only 16 to 73 solutions across the 12 benchmark WDN problems while being only 24 for the largest problem.

The progressive optimality algorithm (POA) is commonly used to identify optimal hydropower operation schedules in China. However, POA may not converge within a reasonable time for large and complex problems because its computational burden grows exponentially with the expansion of system scale. In order to effectively alleviate the dimensionality problem of POA, an improved POA variant called orthogonal progressive optimality algorithm (OPOA) is introduced in this paper. In the OPOA, an orthogonal experimental design is used to replace the exhaustive combinatorial evaluation at each POA two-stage subproblem. The theoretical analysis shows that POA and OPOA have exponential and approximately polynomial growth in computational complexity, respectively. The proposed method is applied to a large-scale multireservoir system located on the Wu River in China. The results indicate that, compared with POA, OPOA can remarkably enhance the computing efficiency in different cases, showing its practicability and feasibility for multireservoir system operation.

A hydropower resource assessment (HRA) can be considered a reconnaissance survey in which the total hydropower potential in a region is quantified. Hydropower potential for purposes of this paper is defined as the potential power that can be extracted from a reservoir. A product of mean annual streamflow flowing in a stream reach and the hydraulic head in a potential reservoir to be constructed in the stream reach with turbine efficiency and unit weight of water determines the power potential. Traditionally, an HRA considered only hydrologic and hydraulic information to quantify the total hydropower potential in a region. With the advancements in computational technology, the HRA can be enriched with geospatial environmental information. This additional information can help to identify the potential stream reaches and their rank for future hydropower development in a region. This paper considers three different types of information, and hence, three criteria, including (1) hydrologic and hydraulic information; (2) environmental information; and (3) combined hydrologic and hydraulic and environmental information to search for potential stream reaches. A methodology was proposed to identify the high-power-density sites for potential development using these three criteria. The methodology was tested to a USGS subbasin in the state of California. The impacts of the criteria on potential site identification are clearly observed in the results.

Over the past decade, many performance indicators have been developed for water utilities to track their system performance. This study proposes a set of normalized and time-integrated benchmarking performance indicators for sustainable long-term management of water distribution and wastewater collection networks. The benchmarking performance indicators are aggregated into three categories: (1) infrastructure, (2) sociopolitical, and (3) financial. To demonstrate the use and value of the benchmarking performance indicators, a system dynamics model is used to present a case study for three water utilities in southern Ontario, Canada. This study shows that the benchmarking performance indicators will allow water utilities with different attributes (such as number of customers, network pipe age profile, pipe material type, network size, and location) to benchmark the long-term variation in their performance with other utilities regionally and nationally. Furthermore, the benchmarking performance indicators can be used to forecast the future behavior of the system to ensure decision-making policies that will drive improvements and best practices.

Energy efficiency plays a large role in the sustainability effort of water utilities. In gravity-fed pipe networks that present excessive pressures, the installation of micro turbines or pumps-as-turbines for energy recovery can provide significant benefits. The main obstacles associated with this kind of solution include the high cost of implementation (equipment and installation) and the time-consuming task of searching for potential and feasible sites for the power plant implementation. In this work, a numerical methodology for the analysis of any water supply and distribution network in order to optimally locate sites for maximum energy recovery and to preliminarily select or design turbines is proposed and validated with a case study. The idea of this decision support tool is to facilitate the assessment of the potential for energy production in networks and encourage the implementation of this type of efficiency measure. The proposed approach consists of two main stages: (1) identification of potential sites for energy recovery through an optimization procedure and (2) selection of adequate types of turbine and preliminary feasibility analysis.

The existing risk-based approaches for quantifying expected flood damage loss may not be suitable in the changing environment when streamflow becomes nonstationary. Conditional value-at-risk ($CVa{R}_{\alpha }$), a modified form of value-at-risk ($Va{R}_{\alpha }$), that takes account of both the magnitude and probability of damage loss has been widely applied to water resources issues. However, quantification of $CVa{R}_{\alpha }$ under nonstationary conditions has not been addressed in the literature. This study proposes an approach that incorporates $CVa{R}_{\alpha }$ in quantifying the flood damage loss under nonstationary conditions, and then $CVa{R}_{\alpha }^n$ associated with flood risk over a specified time horizon is deduced. To illustrate the concept of $CVa{R}_{\alpha }$, this study applies $CVa{R}_{\alpha }$ into the adaptive flood limited water level (FLWL) optimization. With China’s Three Gorges Reservoir (TGR) as a case study, the results indicate that (1) the $CVa{R}_{\alpha }$ not only can represent the possible flood damage loss over a time horizon in the future, but also can reflect the flood risk probability by choosing a suitable confidence level; and (2) it is reliable to take account of the $CVa{R}_{\alpha }$ value as a constraint when studying the adaptive FLWL optimization problem.

This paper proposes an agent-based modeling approach to assess water users’ behavior for water demand management (WDM) in a river basin. In this procedure, each agent controls its own strategy regarding whether to conserve or consume more water in order to achieve a better economic return based on an initial allocation scheme. The effects of agents’ behaviors on their own economic returns and the aggregated impacts of individual behavior on the system are investigated. A positive incentive given to water conservers encourages agents to implement WDM strategies, which in turn improve water use efficiency. A case study using this new agent-based approach reveals that agricultural users are the main contributors to water conservation. Compensation given to water conservers more than covers the benefit loss from lower water consumption, whereas other users gain benefits from the use of the conserved water. The results also indicate that the implementation of WDM strategies is beneficial for the overall system from both economic and ecological perspectives.

In this study, a techno-economic analysis of 16 decentralized wastewater treatment plants (WWTPs) based on various technologies was performed in northern India. Six elements including treatment performance, land use, capital investment, operation and maintenance, cost of treatment, and electricity consumption are discussed in this study. The technologies assessed are extended aeration (EA), moving bed bioreactor (MBBR), sequential bioreactor (SBR), rotating biological contactor (RBC), on-site package (aerobic and/or anaerobic), and membrane bioreactor (MBR). The results indicate that the treatment efficiencies of all cluster-type WWTPs differed significantly from on-site anaerobic package (AnP) types. On the other hand, the treatment efficiencies of on-site aerobic package (AP) systems were reasonably comparable to those of cluster-type plants. The land use for on-site package treatment systems was estimated between 0.125 and $0.8{\mathrm{m}}^2$ per population equivalent (pe) and higher than for cluster-type WWTPs, which require between 0.039 and $0.159{\mathrm{m}}^2/\mathrm{pe}$. The data collected from this study show that package plant treatment costs are high compared with those of cluster types, ranging from US$0.0676 to 0.1045 ($\pm 10\%$) and from US$0.0353 to $0.1891/{\mathrm{m}}^3$ ($\pm 15-20\%$), respectively. Cost-benefit analysis (CBA) was undertaken for all the WWTPs and among the considered environmental benefits nitrogen removal contributed to the most. Moreover, the operation of WWTPs was found to be economically viable even without the sale of treated water, except for one AnP plant. For all plants evaluated, specific power consumption (SPC) was found to vary between 0 and $1\mathrm{kW}/{\mathrm{m}}^3$.

One consistent projection for the western United States has been for increasing peak streamflow as the global climate warms. Although past studies have characterized some aspects of future streamflow projections, this effort exploits new data sources to estimate changing peak flow frequency based on output from many climate projections, which drive a physically-based hydrology model. Using historic and projected future streamflow simulations at 421 sites across the western United States, changes in peak flows for common recurrence intervals are estimated. A parameter relating the recurrence interval needed for design in the present to produce a future required recurrence interval is derived and mapped across the United States, illustrating a method for adapting design to a changing hydrologic setting. For this demonstration, using a higher business-as-usual greenhouse gas emissions pathway, peak flow increases were proportionately larger for the more rare 100-year (or 1% exceedance) event than the 10-year event. Compared to 1971–2000, the domain-wide peak flow magnitude is projected to increase by 14–19% for early 21st century and 31–43% by the end of the 21st century, depending on recurrence interval. Impacts under lower emissions pathways will be more modest (or occur further in the future). In terms of return period, by the end of the 21st century, the 100-year event of the late twentieth century is projected to be approximately a 40-year event, representing a 2.5-fold increase in occurrence probability. This approach offers a strategy for regional planners to incorporate these projected changes into design based on flood flow frequency.

Eutrophication is an environmental imbalance arising from excess nutrients and causing a decrease of water quality. This study assessed the water quality of the Vacacaí Mirim reservoir in the city of Santa Maria, Brazil, according to Brazilian Environmental Standard from 2010–2011, using a numerical model to simulate temperature, dissolved oxygen, 5-day biochemical oxygen demand, ammonium, nitrite-nitrate, ortho-phosphate, total phosphorus, and chlorophyll-$a$. Calibration was performed manually and was found to be satisfactory when the accuracy was checked using mean absolute error and root-mean square error. Extreme case scenarios were proposed to monitor the response of the reservoir to droughts, heavy rainfall, population increases, and wastewater treatment. Simulations show that a 30-day period of dry weather causes the reservoir to become hypereutrophic. Increased precipitation simulation could not translate the interactions that occurred in the environment. Increased pollution load scenarios show that high concentrations of total phosphorus and chlorophyll-$a$ change the trophic state of the reservoir to hypertrophic. Simulations of reduced wastewater dumping show that a 66% decrease in dumping volumes considerably improve water quality and the reservoir becomes mesotrophic.

This paper presents a comparison between an inflow model-based and an inflow model-free optimization method applied to a hydropower system. Widely used stochastic dynamic programming (SDP) and the evolutionary multiobjective direct policy search (EMODPS) methods are used, respectively, as model-based and model-free methods. Main results show that the model-free approach provides a better representation of the complex inflow correlations. Stochastic dynamic programming suffers from the temporal decomposition of the problem that allows only autoregressive exogenous (ARX) models to be used. However, because inflow uncertainty is implicitly represented through simulation with EMODPS, long time series are required to accurately characterize the probability of all possible events. To tackle this drawback and to avoid only learning the data set, a new regularization framework is introduced to improve the policy robustness on unseen data sets. Moreover, this study highlights how the preselected family of functions defining the policy can reduce the performance of the EMODPS method. This study is based on the real-world case of Kemano, located in British Columbia, Canada. This system is challenging because of the long-term inflow correlation due to long snow accumulation periods and its multiple objectives.