Journal of Water Resources Planning and Management

Although the Japanese water sector is economically and socially important, few empirical studies are available to help analysts and policy-makers understand the performance patterns in the industry. This study applies data envelopment analysis to 5,538 observations of 1,144 utilities that supplied drinking water between 2004 and 2007. With a comprehensive census of utilities, the present study controls for many factors affecting efficiency: region, prefecture, ownership/governance, water source, vertical integration (purchased or produced alone), water or integrated system, production, treatment, transport and distribution of water), peak factor, per capita consumption, customer density, water losses, monthly water charge, outsourcing, subsidies, gross prefecture product, and time. Thus, this study derives comprehensive conclusions regarding efficiency patterns in Japan. The analysis finds that the average level of inefficiency (weighted by volume) is 57% in the constant return to scale model, but only 24% for the (more flexible) variable return to scale model. Improving sector efficiency and transferring funds to more innovative sectors rather than using scarce funds to subsidize water distribution would benefit citizens. Thus, the application of advanced quantitative techniques to Japanese water utilities improves the understanding of efficiency patterns and underscores the importance of in-depth studies of the individual factors examined in this study.

A fully automated data-driven methodology for the detection of pipe bursts and other events that induce similar abnormal pressure/flow variations (e.g., unauthorized consumptions) at the district metered area (DMA) level has been recently developed by the authors. This methodology works by simultaneously analyzing the data coming on-line from all the pressure and/or flow sensors deployed in a DMA. It makes synergistic use of several self-learning artificial intelligence (AI) and statistical techniques. These include (1) wavelets for the de-noising of the recorded pressure/flow signals; (2) artificial neural networks (ANNs) for the short-term forecasting of pressure/flow signal values; (3) statistical process control (SPC) techniques for the short-term and long-term analysis of the burst/other event-induced pressure/flow variations; and (4) Bayesian inference systems (BISs) for inferring the probability that a pipe burst/other event has occurred in the DMA being studied, raising the corresponding detection alarms, and provide information useful for performing event diagnosis. This paper focuses on the (re)calibration of the above detection methodology with the aim of improving the forecasting performances of the ANN models and the classification performances of the BIS used to raise the detection alarms (i.e., DMA-level BIS). This is achieved by using (1) an Evolutionary Algorithm optimization strategy for selecting the best ANN input structures and related parameter values to be used for training the ANN models, and (2) an Expectation Maximization strategy for (re)calibrating the values in the conditional probability tables (CPTs) of the DMA-level BIS. The (re)calibration procedure is tested on a case study involving several DMAs in the U.K. with real-life pipe bursts/other events, engineered pipe burst events (i.e., simulated by opening fire hydrants), and synthetic pipe burst events (i.e., simulated by arbitrarily adding “burst flows” to an actual flow signal). The results obtained illustrate that the new (re)calibration procedure improves the performance of the event detection methodology in terms of increased detection speed and reliability.

A coupled binary linear programming–differential evolution (BLP-DE) approach is proposed in this paper to optimize the design of water distribution systems (WDS). Three stages are involved in the proposed BLP-DE optimization method. In the first stage, the WDS that is being optimized is decomposed into trees and the core using a graph algorithm. Binary linear programming is then used to optimize the design of the trees during the second stage. In the third stage, a differential evolution (DE) algorithm is utilized to deal with the core design while incorporating the optimal solutions for the trees obtained in the second stage, thereby yielding near-optimal solutions for the original whole WDS. The proposed method takes advantage of both the BLP and DE algorithms: BLP is capable of providing a global optimal solution for the trees (no loops involved) with great efficiency, and a DE is able to efficiently generate good quality solutions for the core (loops involved) with a reduced search space compared to the original full network. Two benchmark WDS case studies and one real-world case study (with multiple demand loading cases) with a number of decision variables ranging from 21–96 are used to verify the effectiveness of the proposed BLP-DE optimization approach. Results show that the proposed BLP-DE algorithm significantly outperforms other optimization algorithms in terms of both solution quality and efficiency.

The traditional optimization approach for water distribution mains is that of considering a single design scenario with prefixed nodal demands representing the peak values at the end of the life cycle of the construction. Instead, this paper presents a different approach for the design of water distribution mains aimed at considering the phasing of construction. It makes it possible to identify, on prefixed time steps or intervals (for instance 25 years), the upgrade of the construction rendering the network able to satisfy, during the expected life of the system, growing nodal demands related to the increment in the population served. To show the benefits of this approach in comparison to using a single design flow, an optimization methodology, aimed at introducing new pipes in the network as needed at each time step, was set up and applied to a simple case study, where two different scenarios were considered concerning the growth of the network. Results showed that this approach is able to yield better results when compared with the single flow design, because it enables short-term construction upgrades to be performed while keeping a vision of the expected long term network growth.

It is a well-known fact that most of the developing countries have intermittent water supply and the quantity of water supplied from the source is also not distributed equitably among the consumers. Aged pipelines, pump failures, and improper management of water resources are some of the main reasons for it. This study presents the application of a nonlinear control technique to overcome this problem in different zones in the city of Bangalore. The water is pumped to the city from a large distance of approximately 100 km over a very high elevation of approximately 400 m. The city has large undulating terrain among different zones, which leads to unequal distribution of water. The Bangalore, inflow water-distribution system (WDS) has been modeled. A dynamic inversion (DI) nonlinear controller with proportional integral derivative (PID) features (DI-PID) is used for valve throttling to achieve the target flows to different zones of the city. This novel approach of equitable water distribution using DI-PID controllers that can be used as a decision support system is discussed in this paper.

This paper proposes a new methodology for the optimal design of water network sectorization, which is an essential technique for improving the management and security of multiple-source water supply systems. In particular, the network sectorization problem under consideration concerns the definition of isolated district meter areas, each of which is supplied by its own source (or sources) and is completely disconnected from the rest of the water system through boundary valves or permanent pipe sectioning. The proposed methodology uses graph theory principles and a heuristic procedure based on minimizing the amount of dissipated power in the water network. The procedure has been tested on two existing water distribution networks (WDNs) (in Parete, Italy and San Luis Rio Colorado, Mexico) using different performance indices. The simulation results, which confirmed the effectiveness of the proposed methodology, surpass empirical trial-and-error approaches and offer water utilities a tool for the design of multiple-source WDNs in isolated districts.

The establishment of an optimal pumping schedule is a cost-free, efficient measure that is easily adopted to reduce energy costs. This paper proposes an optimal pumping scheduling model for complex irrigation water distribution systems (WDS) for the case in which head and efficiency are independent of flow rate. The proposed model utilizes a simplified pumping scheme and a determinist linear programming (LP) optimization algorithm. A robust and versatile software application was also developed that is useful for WDS managers in adopting optimal pumping scheduling decisions for a specific electricity tariff structure and to satisfy all system performance constraints. The model also provides the option of testing and comparing different alternative electricity tariffs and allows managers to select the most cost-effective tariff for their system. The model is applied to a real WDS located in the province of Almería in southeastern Spain. The optimal pumping scheduling during an entire year is obtained for several alternative tariffs using real data on irrigation demands, pumping efficiencies, energy prices, and tariffs for the analyzed year. The most efficient electricity tariff is also obtained for that year. Finally, the actual management strategies in the studied year are compared with the optimal strategies provided by the model. Although the system managers chose an appropriate tariff for their system, the use of this optimization tool would have provided significant savings to their energy bills, exceeding 20% of the total energy costs.

Water-distribution operation is an essential part of an urban water-supply system to deliver high-quality water to consumers. Management of such an operation may involve deliberations of operation cost, system capacity, and environmental restriction, and is convoluted with many forms of uncertainties. In this paper, an integrated fuzzy programming and decision analysis (IFPDA) approach was proposed for a multilayer urban water-distribution system management under uncertainty. The system consisted of two water sources, four treatment plants, seven reservoirs, and seven consuming zones. Uncertain information associated with water demands, water-treatment capacities, water-transfer cost, and leakage rate was described by a trapezoidal-shaped fuzzy set and embedded into a fuzzy programming framework. The balance between the satisfaction degree of achieving the system objective and feasibility level of meeting the related constraints was analyzed by fuzzy inference procedures. The results indicate that the IFPDA approach was advantageous in (1) dealing with fuzzy uncertainties in the objective function, both sides of the model constraints, and the context of an urban water-distribution system management; (2) helping analyze tradeoffs between minimization of operation cost and reliability of running the system; and (3) linking optimization model outputs with decision analysis.

Adsorption to pipe wall materials significantly affects the fate and transport of certain contaminants in water distribution systems. For example, arsenate has a strong affinity for iron oxide, a substance common in water distribution pipes. In this paper a mathematical model for arsenate adsorption to iron oxide pipe wall materials is developed. The effects of two common assumptions on modeled arsenate transport are explored: a theoretical smooth pipe mass transfer coefficient and an assumption of rapid equilibrium of adsorption at the pipe wall surface. The effects of these assumptions are explored in a single pipe sensitivity analysis and found to yield significantly different results than parameters estimated from experimental data. In simulations of a hypothetical arsenate contamination event in a model water distribution system, the two assumptions result in different predictions of system contamination and contaminant exposure to consumers. These results indicate that even though water quality modeling plays an essential role in planning for distribution system decontamination, modeling assumptions must be carefully chosen.

Accurate prediction of mixing at cross junctions is important in water quality models because most of the intersecting pipes use cross junctions in water distribution systems. To describe the characteristics of complex mixing phenomena, an empirical mixing model at the cross joints is proposed based on a large number of experimental data. The orthogonal experiments by numerical simulations show that the diameter ratio of inlets and the Reynolds number ratios of inlets and outlets have a dominant effect on the solute mixing at cross junctions. These dominant parameters are then chosen to investigate the cross-joint mixing in the experimental setup. The results indicated that the cross joint connected by pipes with a large size difference has a more complete mixing than the one with a small pipe size difference, if other conditions are kept the same. The mixing formulas for solute concentrations at outlets are fitted by an improved particle swarm optimization (PSO) algorithm based on the experimental data. The interpolation method is used to obtain the mixing information for other diameter ratios of the inlets. The established mixing model can be incorporated into standard EPANET to analyze the water quality in water distribution systems.

Emerging applications in water distribution systems place new challenges on water quality modeling tools, including the need for an explicit mathematical representation and simulation of large ensembles of contamination scenarios. We present a computational framework, referred to as Merlion, that creates an explicit mathematical model for water quality in drinking water distribution systems that is appropriate for embedding within other numerical applications (e.g., optimization). This model is efficiently generated for large water distribution systems and represents an all-to-all mapping (inputs include injections at all possible nodes and time steps, and outputs include concentrations at all possible nodes and time steps), which is necessary for many security applications. The Merlion framework is compared to water quality simulations using EPANET on a set of network models ranging in size from 10 to approximately 13,000 nodes. The simulation results show excellent agreement. Furthermore, the explicit linear model can be used to evaluate a large number of tracer scenarios very quickly, speeding up current security tools by approximately an order of magnitude.

Surge vessels provide efficient protection against low and high transient pressures in water distribution systems. However, they can be quite expensive, and any reduction in surge vessel size can significantly reduce surge protection costs. Graphical and other heuristic methods reported in literature are limited to sizing surge vessels for simple rising mains. Attempts to use more structured optimization techniques have been largely unsuccessful because of their impractical computational requirements. This article proposes a robust framework for developing surge protection design tools and demonstrates the usefulness of the framework through an example surge vessel sizing tool. The essence of the proposed framework is in the identification of key transient response parameters that influence surge vessel characteristics from seemingly unmanageable transient response data. This parameterization helps the sizing tool to exploit the similarity between transient responses of small pipe networks and subsections of large pipe networks. The framework employs a knowledge-base of transient pressures and flows derived from several small network models and corresponding optimal surge vessel sizes obtained from genetic algorithm (GA) optimizers. Key transient response parameters were identified from this knowledge-base and used as input variables for a neural network model along with the associated surge vessel characteristics as output variables. The trained neural network model was successfully applied for complex pipe networks to obtain optimal surge vessel sizes for transient protection. Neural network model predictions were compared with optimal surge vessel characteristics, and their performance was evaluated by transient simulation models to assess the efficiency of the proposed framework and sizing tool. Application of this framework has the advantage of developing surge protection sizing tools independent of network system schematics and boundary conditions.

As a common approach to reservoir operating policies, water levels at the end of each time interval should be kept at or above the rule curve. In this study, the policy is captured using rationing of the target yield to reduce the intensity of severe water shortages. For this purpose, a hybrid model is developed to optimize simultaneously both the conventional rule curve and the hedging rule. In the compound model, a simple genetic algorithm is coupled with a simulation program, including an inner linear programming algorithm. In this way, operational policies are imposed by priority concepts to achieve the optimal water allocation and the target storage levels for reservoirs. As a case study, a multipurpose, multireservoir system in southern Iran is selected. The results show that the model has good performance in extracting the optimum policy for reservoir operation under both normal and drought conditions.

A coupled simulation-optimization with reanalysis-based short-term weather forecasts from a regional climate model (RCM) is proposed to optimize an irrigation scheduling problem. Using different physical configurations of the climate extension of a weather research and forecasting model (CWRF) that is driven by national atmospheric model project reanalysis data, five ensemble outlooks of 15 consecutive daily forecasts have been generated during five different crop-growing seasons. Six daily climatic variables are forecasted, namely, rainfall, minimum temperature, maximum temperature, humidity, wind speed, and solar radiation. To correct the forecasts for any inherent bias, the quantile mapping method is applied to all six daily climatic variables. After bias correction, a skill assessment of the reanalysis-based RCM forecasts indicate that only the first three climatic variables are predicted with reliable accuracy; thus, average climatic means are used to replace the remaining three variables (humidity, wind speed, and solar radiation). The framework is applied to the Havana Lowlands region, Illinois, as a case study, and the value of forecasts is assessed against two baseline scenarios: no-rain forecast (a pessimistic case) and average climatology (a normal case). Using reanalysis-based RCM forecasts to guide farmers’ irrigation decisions could yield about 1–3% in expected profit gain and 4–6% in water reduction when compared to the no-rain forecast scenario, and 1–6% in expected profit gain when compared to the average climatology scenario. This study is a first preliminary attempt to use an ensemble of weather simulations in the optimization of irrigation scheduling, and the developed framework can be used to incorporate operational forecasting once the reanalysis boundary is replaced by global weather forecasts.

Water systems in snowmelt-dominated hydroregions such as California’s Sierra Nevada mountains are sensitive to regional climate change, hydropower systems in particular. In this study, a water resources management model was developed for the upper west slope Sierra Nevada to understand the potential effects of regional climate warming on hydropower at the watershed scale, a scale that has been largely neglected but is important for hydroregional planning. The model is developed with the Water Evaluation and Planning system (WEAP) and includes most water management infrastructure in the study region. Hydropower is simulated assuming historical long-term electricity demand and a spill minimization rule. The method is suitable for simulating generation for most of the main watersheds in the region. To assess the potential effect of climate warming, uniform air temperature increases of 0°C, 2°C, 4°C, and 6°C were considered, with no change in precipitation, to approximate regional warming through 2100. The highly productive northern Sierra Nevada sees large reductions in hydropower generation with decreases in annual runoff. The central watersheds see less reduction in annual runoff and can adapt better to changes in runoff timing. Generation in southern watersheds, which are less productive, decreases. Results from this study can help identify which watersheds might easily adapt to climate change, where hydropower is likely to conflict with other uses, and where more detailed operational studies are needed.

The common-law Riparian Doctrine, the legal basis for water allocation in the eastern United States, restricts qualified users along a shared waterway to reasonable withdrawals. Recent water shortages and droughts have highlighted the need for more specific withdrawal restrictions, and laws have been passed. What is still missing, however, is a water allocation mechanism that is consistent with the Riparian Doctrine and one that can accommodate downstream demands in quantifying the withdrawal limits and reservoir release rules. The present study proposes one such mechanism, called proportionally fair (ProFair) allocation, which fulfills this need. It is realized by a so-called bottleneck algorithm (BA), which allots water proportionally in a large stream network through a series of successive steps. It is discovered that the ProFair allocation can be simulated by a log-linear programming (LLP) model. The equivalence of the LLP model and the BA is proved. The Upper Neuse River watershed in North Carolina serves as an example to demonstrate the effectiveness of the ProFair mechanism.

The objectives of this project were to analyze available Sacramento–San Joaquin Delta, California, levee risk information in a Composite Risk Management matrix and examine the results for management decision support. U.S. Army Corps of Engineers (USACE) guidance documents define risk as the “probability and severity of loss linked to hazards” and prescribe a composite risk assessment method. The Delta Risk Management Strategy performed for a group of state and federal partners provided analyses of the relative probability of hazards and severity of risks in the Delta and provide the information needed for a risk analysis compliant with USACE requirements. Composite Risk assessment provides rank-ordered lists of the highest risk zones–those with the greatest probability of failure combined with the most severe consequences–for several hundred protected areas in the Delta. Although uncertainties in the absolute magnitude of the results make them most useful for comparisons, the actual values of the probabilities and consequences are alarming. For example, Sargent Barnhart Tract, northwest of Stockton, has a mean annual failure rate of 0.07, or an expected levee failure every 14 years, with a probable 96 fatalities for a nighttime seismic-induced failure. Adjacent tracts with only slightly lower failure probabilities put another 500 lives at risk. An area of the Suisun Marsh has a projected failure rate of 0.5, or once every two years, with maximum possible damages exceeding $250 million. The Sacramento Pocket Area, with a mean annual failure rate of 0.006, has over $9 billion at risk. Although refinements to these risk estimates are possible, this paper and the Delta Risk Management Strategy analyses provide more than sufficient evidence that flooding in the Sacramento–San Joaquin Delta presents significant risks to California and the nation. Hundreds of lives and billions of dollar damages are at risk. Urgent action is necessary to manage those risks.