Journal of Water Resources Planning and Management

Forum papers are thought-provoking opinion pieces or essays founded in fact, sometimes containing speculation, on a civil engineering topic of general interest and relevance to the readership of the journal. The views expressed in this Forum article do not necessarily reflect the views of ASCE or the Editorial Board of the journal.

In this paper a new method is proposed that solves for the steady state of pressure-dependent models (PDMs) with flow control valves. Rather than model flow devices individually, the method solves the more general problem in which a water distribution system (WDS) has some link flows constrained to lie between upper and lower, or possibly equal, set limits. No heuristics are used to determine device states. The method is shown to be fast, and its effectiveness is demonstrated on PDM WDSs with up to about 20,000 links and 18,000 nodes and 60 link flow constraints, some of which prescribe a fixed flow. The method has applications in network management, network design, and flow control to deal with water distribution where there is insufficient supply.

In the water systems analysis literature and practice, planning (i.e., dam sizing) and management (i.e., operation design) have been for long time addressed as two weakly interconnected problems, and this often resulted in oversized, poorly performing infrastructures. Recently, several authors started exploring the interdependent nature of these two problems, introducing new integrated approaches to simultaneously design water infrastructures and their operations. Yet, the high computational burden is a likely downside of these methods, a large share of which require solving one optimal operation design problem for every candidate dam size, making it unfeasible to explore the entire planning and associated operation decision space. This paper contributes a novel reinforcement learning (RL)-based approach to integrate dam sizing and operation design while significantly containing computational costs with respect to alternative state-of-the-art methods. The approach first optimizes a single operating policy parametric in the dam size and then searches for the best reservoir size operated using this policy. The parametric policy is computed through a novel batch-mode RL algorithm, called Planning Fitted Q-Iteration (pFQI). The proposed RL approach is tested on a numerical case study, where the water infrastructure must be sized and operated to meet downstream users’ water demand while minimizing construction costs. Results show that the proposed RL approach is able to identify more efficient system configurations with respect to traditional sizing approaches that neglect the optimal operation design phase. Furthermore, when compared with other integrated approaches, the pFQI algorithm is proven to be computationally more efficient.

This study presents an analysis of possible policy measures to face water scarcity for agriculture in a long-term climate projection (2070–2100). Water availability was computed with the Water Availability and Adaptation Policy Analysis (WAAPA) model. WAAPA simulates water management in a basin at a monthly time scale, accounting for streamflow, reservoir storage, evaporation, and environmental flows. Focus was placed on urban and irrigation demands. Water was first allocated to urban demands (assumed as a priority), and the remaining water resources were allocated to irrigation. First, this availability was compared to an estimated projected evolution of water withdrawals to identify potential conflicts over water supply. If insufficient water was available to satisfy all irrigation demands, policy measures, such as increasing reservoir storage, improving the efficiency of urban water use, and changing the environmental flow allocation, were applied and their effects on irrigation water availability analyzed. The analysis of the Duero-Douro (Spain-Portugal) Basin showed that, although the expected impacts of climate change on water availability for agriculture are high, several policy options can partially compensate for those impacts. Comparatively, a change in environmental flow requirements was the policy measure that most affected Duero-Douro Basin management.

Water-supply systems (WSSs) and electrical power systems (EPSs) are highly interdependent critical infrastructures. The electrical energy required for pumping in WSSs and cooling water required for power plants in EPSs are major interdependencies. Failure of either of the two independently operated infrastructures can lead to a cascading failure of both the systems. A combined operations control methodology for WSSs and EPSs taking into consideration the inherent interdependencies is required to ensure reliable operations. An optimization-simulation model is presented for the real-time operation of water-supply canal systems (WSCSs) under critical conditions during short-term and long-term emergency events such as limited electrical energy and/or limited water availability, electrical grid failures, extreme droughts, or other severe conditions related to natural and manmade disasters. WSCSs are used for the conveyance of raw water from sources such as lakes, reservoirs, or rivers to water treatment plants that supply treated water to consumers through water distribution systems (WDSs). The approach interfaces the optimization-simulation model for WSCSs with an optimization-simulation model for WDSs to provide for a comprehensive decision-making tool for the control of WSCSs and WDSs. Two WSCSs optimization methodologies are presented including a nonlinear programming approach and an optimization-simulation approach that interfaces a genetic algorithm (MATLAB) with the US Army Corps of Engineers Hydraulic Engineering Center’s (HEC) River Analysis System (HEC-RAS) simulation model. A steady-state analysis of the WSCSs is performed for each time period of operation. The new methodologies for determining pump and gate operations under limited power and/or water availability are illustrated using two example canal systems.

Challenges exist in managing groundwater resources because of spatiotemporally variable pumping activities as well as complex subsurface hydrogeology. In addition, excessive water exploitation induces an imbalance among multistakeholder benefits. In this study, a nonlinear high-order multiobjective optimization model was constructed to derive optimal freshwater pumping strategies and explore the optimality through regulation of pumping locations. Three objectives concerning water supply, energy cost, and environmental problems were formulated into a groundwater management framework that maximizes the total groundwater withdrawal from potential wells and minimizes the total energy cost for well pumping, and groundwater level variations at monitoring locations. Binary variables were incorporated into the groundwater management model to control the operative status of the pumping wells. An improved Nondominated Sorting Genetic Algorithm II (NSGA-II) was developed to increase solution convergence and linked with a high-fidelity groundwater model (MODFLOW-2005) to solve the optimization problem. The improved NSGA-II was expedited on a parallel computing platform to alleviate the computational burden. The effectiveness of the proposed methodology was demonstrated by an application to the Baton Rouge multiaquifer system in southeastern Louisiana. Nondominated trade-off solutions were successfully achieved through the proposed approach and were an optimum with regard to the goals and corresponding consequences. Operative status of the pumping wells, pumping rates, and distances from observation wells to the pumping wells produced distinctive optimization responses. In conclusion, the proposed approach is an appealing method for determining the optimal extent to which the three objectives concerning water supply, energy cost, and environmental problems can be achieved.

In this paper, we propose a novel methodology for altering the area monitored by water quality sensors in water distribution systems (WDS) when there is suspicion of a contamination event. The proposed active contamination detection (ACD) scheme manipulates WDS actuators, i.e., by closing and opening valves or by changing the set-points at pressure controlled locations to drive flows from specific parts of the network in predetermined paths and enable the sensors to monitor the quality of water from previously unobserved locations. As a consequence, the monitoring coverage of the sensors is increased and some contamination events occurring within those areas can be detected. The objective is to minimize the contamination impact by detecting the contaminant as soon as possible, while also maintaining the hydraulic requirements of the system. Moreover, the methodology facilitates the isolation of the contamination propagation path and its possible source. We demonstrate the ACD scheme on two networks analyze the results and open the discussion for further work in this area.

The California State Water Project (SWP) along with the Central Valley Project (CVP), under various environmental regulations, manage California’s complex water storage and delivery system. An important part of the water system regulations strictly limit salinity intrusion into the Sacramento–San Joaquin Delta, which is a complex tidal estuary with many factors influencing its salinity. The nonlinear relationship of these factors on salinity make the system operations challenging. Operational models [e.g., California Water Resources Simulation Model (CalSim) and CalSim Lite (CalLite)] are used to provide guidelines to decision makers for efficient planning and management of the system. But these operational models are not designed to directly simulate the salinity. The hydrodynamic and water quality model, California Department of Water Resources (DWR) Delta Simulation Model II (DSM2), is needed to simulate the salinity. Because of a linking problem and longer simulation time of DMS2, it is impractical to use DSM2 directly in operational models. This paper presents the development, improvement, and successful application of an artificial neural network (ANN). The ANN, when fully integrated into CalSim and CalLite, emulates the Delta salinity so that the operational models, when coupled with the ANN, can simulate the salinity management in the Delta. The newly developed and improved ANN reported in this research, when used in the CalSim model, provides more accurate insights on the salinity regime in the Delta, which is conducive to more efficient use of the freshwater in the Delta resulting in the more efficient overall operation of the SWP and CVP.

Pump scheduling is a key element in water distribution systems operation. Modeling this problem requires a mixed integer nonlinear program (MINLP) formulation. Even linearization schemes of mixed integer linear programs (MILPs) are typically beyond the capability of real-time optimization frameworks. In this study, we explore different levels of MILP approximations by reducing the number of binary decision variables (i.e., different binarization levels). In addition, we present a simple demand forecast model and evaluate the performance and approximation accuracy of the suggested approach in a real-time optimization framework under a receding horizon operation mode. The results show that the balance between approximation accuracy and solution efficiency is biased. That is, a simple low-accuracy approximation may yield an efficient and practical solution algorithm that results in a near-optimal solution.

This study explores the ways to introduce physical interpretability into the process of optimizing operating rules for multireservoir systems with multiple objectives. Prior studies applied the concept of direct policy search (DPS), in which the release policy is expressed as a set of parameterized functions (e.g., neural networks) that are optimized by simulating the performance of different parameter value combinations over a testing period. The problem with this approach is that the operators generally avoid adopting such artificial black-box functions for the direct real-time control of their systems, preferring simpler tools with a clear connection to the system’s physics. This study addresses this mismatch by replacing the black-box functions in DPS with physically based parameterized operating rules, for example by directly using target levels in dams as decision variables. This leads to results that are physically interpretable and may be more acceptable to operators. The methodology proposed in this work is applied to a network of five reservoirs and four power plants in the Nechí catchment in Colombia, with four interests involved: average energy generation, firm energy generation, flood hazard, and flow regime alteration. The release policy is expressed depending on only 12 parameters, which significantly reduces the computational complexity compared to existing approaches of multiobjective DPS. The resulting four-dimensional Pareto-approximate set offers a variety of operational strategies from which operators may choose one that corresponds best to their preferences. For demonstration purposes, one particular optimized policy is selected and its parameter values are analyzed to illustrate how the physically based operating rules can be directly interpreted by the operators.

Manantali is a dam located on the Senegal River and is mainly used for hydropower production. Before the dam’s construction, the annual river flood alimented the flood recession agriculture, a practice based on natural irrigation and fertilization of the flood plain, used traditionally by the local populations downstream. Analysis of the actual reservoir operation shows that annual floods have been largely reduced for the benefit of hydropower production. Moreover, the Senegal River Basin authority is evaluating the construction of different new dams, which could reduce even further the water available for flood support, given that the current operational focus is on satisfying hydropower demand. This study investigates the effects of an optimal reservoir operation strategy that maximizes hydropower production only, analyzing the results of this strategy in terms of effects on the two main objectives, i.e., hydropower production and flood support. The problem of finding optimal reservoir operation strategy is solved by applying the stochastic dual dynamic programming method. Results show the existence of a release strategy in which both objectives improve ($+9\%$ for hydropower and $+7\%$ for flood production) with respect to the historically observed operation. This solution, however, may require the electric system to compensate for the variability in energy supply along the year.

Over the last decade, utilities, governments, and businesses have increasingly come to realize that financial considerations are not the only factors driving consumer behavior; rather, social and psychological factors play a significant role as well. For example, water demand management strategies rely on customers understanding how to reduce their individual water consumption and apply this understanding to their everyday activities. The proliferation of smart metering technology has enabled water utilities to better quantify end-user demand and provide near-real-time feedback to their customers. However, relatively few studies have examined the effectiveness of online feedback in promoting water-saving behavior. This study engaged water customers from three different cities in Israel in focus groups to analyze their behavioral incentives to conserve water and their preferences for online feedback applications. Focus group participants revealed a greater interest in environmental conservation than economic concerns, perhaps due to a low price of water. Participants were generally not interested in learning about their daily water consumption but rather preferred to receive alerts only about urgent matters such as a leak in their home or abnormal consumption.