Journal of Water Resources Planning and Management

Electric power transmission lines in service in southwest China have fallen far behind the power sources. With rapid development of hydropower in this area, severe water spills for hydropower frequently occur during the flood season because of limited transmission capacity. Effective coordinated operation among large and small hydropower plants (LHPPs and SHPPs) in the area is a feasible approach for alleviating this problem. Using the Yunnan Power Grid (YNPG) of southwest China as an example, a long-term coordinated optimal operation model (COOM) for an integrated system of LHPPs and SHPPs is proposed in this paper. Total electricity that can be consumed or transmitted is defined as electricity absorption. Maximizing the total electricity absorption, which implies balancing high average water heads with reservoir spills, was selected as the objective. Satisfying multiple voltage transmission constraints is a critical issue. A progressive optimality algorithm (POA) was used to solve the model based on a multidimensional search in two-stage optimizations. Results from the YNPG case study show that the proposed method can achieve greater total electricity absorption while effectively reducing spillage.

Operational objectives and/or constraints of a reservoir system may need to be shifted at certain periods (i.e., transition periods) due to seasonal considerations of human interest and ecological benefits. Despite the fact that operational schemes in the transition periods are critical and of great interest to reservoir operation practice, the problem has received little attention in the literature. This paper presents a study on cluster analysis for identifying patterns of operational schemes during transition periods. The test case corresponds to ten major reservoirs of the Federal Columbia River Power System (FCRPS) in the United States. The operation horizon consists of two weeks during which the objectives of the reservoir system are shifted based on seasonal consideration for fish migration and survival. An optimization model based on an evolutionary algorithm is used to derive the optimal operational schemes under various inflow scenarios. A $K$-Spectral Centroid algorithm ($K\text{-}\mathrm{SC}$) is applied to the resulting operational schemes to find clusters of the schemes based on similarities of their temporal shapes. By investigating the relations between the clusters and the inflow scenarios, general patterns of operational schemes are identified. The analyses offer insights into the operational schemes during the transition period and broaden the understanding of short-term reservoir operation with shifting operational objectives.

Booster systems are used to ensure the water supply in multistory buildings, office buildings, and public institutions. The question is how to operate a booster system to achieve optimal power consumption. Usually, the individual pumps of a booster system are identical in construction. However, this analysis investigates the more general case of different pumps. For identical pumps, the answer is simple, easy to implement, and state of the art. Power consumption is minimal when all individual pumps provide the same partial flow rate and thus run at the same rotational speed. In the case of different pumps, the answer is more complex and is provided in this paper.

This study discerns optimal rules for coordinated operation of existing dams and eight newly constructed weirs in the Nakdong River through use of a system network while considering the water supply and operating facilities. To determine the operation rules, capacity was simulated by reservoir operations modeling and evaluated separately under the following three cases: (1) when dams were operated individually; (2) when they were operated in conjunction with other dams; and (3) when dams and weirs were operated in conjunction with each other. The water supply capacity of the Nakdong was highest in Case 3. When considering operational losses (9.1 and 14.8% in Cases 2 and 3, respectively), water supply decreased. Findings show that when water levels drop below the target, hydrological facilities should flexibly adjust water storage and flow rates at each dam in the system. Further studies are needed to investigate the influence of agricultural reservoirs and river channel discharged flows and to identify regions vulnerable to water shortages driven by future fluctuations in water resources due to climate change.

A number of pressure-driven analysis (PDA)–based hydraulic solvers have been proposed in the literature to address issues of negative pressures estimated by demand-driven analysis (DDA) solvers. However, the PDA methods reported so far attempt to achieve this by either developing a new PDA methodology, which requires modifying the source code of hydraulic solvers, or using iterative-type approaches in which artificial elements (like suitably chosen reservoirs) are added to network nodes until convergence is achieved. None of this is ideal, because the former is difficult to implement and the latter results in computationally inefficient PDA solvers that are difficult, and sometimes even impossible, to use in larger networks, especially under extended period simulation conditions. The PDA modeling approach proposed here does not require either of the aforementioned, because it is based on a single iteration-type algorithm, which involves connecting a set of artificial elements to each network node with demand and deficient pressure. This set consists of a check valve, a flow control valve, and a flow emitter. The new PDA method developed was validated on a number of benchmark and real-life networks under different flow conditions, clearly demonstrating its advantages when compared to existing methods. The key advantages include the simplicity of its implementation and the ability to predict network pressures and flows in a consistently accurate, numerically stable, and computationally efficient manner under pressure-deficient and normal-flow conditions in both steady-state and extended period simulations.

The problem of sensor placement for early-warning detection systems is a topical issue for industry and utilities who want to equip their networks with such technology. Solving the problem entails finding the best sensor locations that optimize a criterion such as detection rate or time to detection. Few methods exist concerning sensor placement that optimize the result of a source identification method. This paper fills the gap by coupling an adjoint source identification method and a Monte Carlo sensor placement algorithm. The first one is treated through the use of a backtracking algorithm. It uses binary responses at sensors to calculate the ranked list of potential contamination location nodes and contamination times. A criterion is then defined based on the source identification accuracy and specificity. Finally, two optimizing methods that maximize this criterion are proposed: a greedy algorithm and a local search algorithm, which are both coupled with a Monte Carlo method to give the locations of sensors that are the best suited for allocating the source of a contamination. These methods are tested on a 2,500-node network to evaluate their efficiency.

The identification of monitoring or sensor locations, or demand estimation, within drinking water distribution systems can be challenging given the size of realistic network models. Approaches such as skeletonization or aggregation can effectively reduce a network model and are generally appropriate for satisfying hydraulic objectives. However, a reduced hydraulic network model might not be appropriate for water quality objectives because of altered transport characteristics. This study proposes a clustering approach that groups nodes with similar water quality characteristics within the context of maintaining the original network structure. The proposed approach uses an input-output relationship to assess the hydraulic path between any two nodes. Using the hydraulic path information, a $k$-means clustering algorithm identified nodes with similar hydraulic paths. For two different case studies, as the number of clusters increased, the nodes within each cluster were shown to become more similar. The differences in water quality characteristics between the individual clusters also increased, demonstrating the ability to generate more distinct clusters of nodes. By identifying nodes with similar water quality characteristics, the resulting clusters can provide future opportunities, for example, to reduce the problem size for monitoring or sensor selection by assuming the nodes within a given cluster behave similarly.

A combined demand and roughness estimation is a critical step in order for the water distribution system model to represent the real system adequately. A novel two-level Markov chain Monte Carlo particle filter method for joint estimation of demand and roughness is proposed in this paper. First, an improved particle filter with ensemble Kalman filter modification to proposal density is adopted to track the non-Gaussian system dynamics and estimate demands. Then, the improved particle filter for demand estimation is nested into the Markov chain Monte Carlo simulation for roughness estimation. The method is very capable of quantifying the uncertainties associated with estimated or predicted values without requiring any assumptions of linearity and Gaussianity or any derivatives to be calculated. A strong nonlinear benchmark network with synthetically generated field data is utilized to validate the performance of this method. The results suggest that the proposed method is demonstrated to provide satisfactory demand and roughness values with reliable confidence limits. Some practical issues are also discussed to enhance the application potential of this method.

Water distribution systems (WDSs) are social infrastructures providing drinking water and must be capable of constant water supply while maintaining an appropriate water pressure. Hence, it is important to quantitatively evaluate the supply capacity of a WDS for design and operation purposes. As part of such efforts, several resilience indexes have been developed based on the energy flows within a network. Accurate estimation of a minimum required head at the demand node is critical for properly calculating the resilience indexes. This study proposes a novel approach of estimating the minimum required head, which considers the flow direction in pipes and the hydraulic gradient within a network. The proposed approach has been implemented and tested for six hypothetical grid-type networks that reflect various topographical characteristics. The application results revealed that the conventional resilience indexes are significantly affected by the nodal elevation and are generally overestimated due to incorrect calculation of the minimum required nodal head.

This paper proposes five methods for improving the quality of Pareto optimal solutions of multiobjective optimal water distribution network (WDN) design problems: (1) three warm initial solution methods, (2) the postoptimization method, and (3) the guided-search method. The five methods were demonstrated through resilience-based design of the Hanoi network. The guided-search method, considering the reasonable range of decision variables, was identified as the best method with respect to the nondomination and diversity of the obtained Pareto solutions. Then, the effect of considering known initial solutions (e.g., least-cost solutions obtained from single-objective optimal design) on the final Pareto solution quality was investigated using the guided-search method. Finally, the guided-search method was compared with five multiobjective optimization algorithms widely used in the WDN research community through the resilience-based design of well-known benchmark WDNs (i.e., two-loop, Hanoi, Balerma, and P-city).

Increasing demand for water coupled with reduced water availability in many regions of the world is leading to growing water scarcity and calls for implementation of a range of technological, institutional, and economic solutions. Water-economy models (WEMs), which integrate the complex interrelationships between hydrologic and economic systems, are effective tools for analyzing these issues and for providing appropriate solutions across varied spatial and temporal scales. These models can be powerful tools for examining potential future changes in water resources systems, including the effects of climate change, socioeconomic changes, and infrastructural and policy responses to water resource management challenges. In this paper, the WEM models used to answer water economy questions are classified into two general categories on the basis of their structure: network-based (simulation or optimization) hydroeconomic models, and economywide (input-output or computable general equilibrium) models. This paper highlights the primary differences in the applications and interpretations obtained using these approaches, analyzes the distribution of questions that different WEMs have been used to answer, and discusses previous work and efforts to integrate across model types. Findings suggest that additional efforts are needed to more realistically account for the range and complexity of linking water systems and society, particularly regarding ecology and water quality, and the food and energy sectors. Additionally, the broader economic impacts of water-related processes, for example those related to interregional trade dynamics, the distribution of income, and migration, should be investigated further. In effect, because of the inherent complexity in the economic dynamics underlying many water systems, such tools can challenge intuition and provide critical insights that are relevant to more effective management of transboundary water resources and related sectors.

Water exists in both a physical and virtual sense. This paper presents a noncompetitive water input-output model for investigating water allocation. The model achieves a robust first-pass diagnostic of water flows in Urumqi, China. Generalized water resources amounted to $4.973\mathrm{billion}{\mathrm{m}}^3$ in 2012, of which fresh, recycled, indirect virtual, and net external virtual water accounted for 22, 4, 12, and 62%, respectively. The water deficit reached $3.678\mathrm{billion}{\mathrm{m}}^3$, approximately 2.84 times the amount of physical water on hand. External virtual water levels amounted to $3.45\mathrm{billion}{\mathrm{m}}^3$, 3.14 times the volume of freshwater available. Agricultural activities generated 1.5% of the gross domestic product, but such activities consumed 60% of all freshwater available. In addition, $0.37\mathrm{billion}{\mathrm{m}}^3$ of local water flowed out of Urumqi in a virtual sense. As a result, generalized water resources management is dependent on the following two basic measures: improving the water infrastructure construction mechanisms to increase physical water supplies, and improving trade with water-rich regions to increase virtual water supplies.

Climate variability and expansion of communities and their infrastructure have increased vulnerability of coastal regions to flood. Flooding can cause failure of infrastructure, which could result in seriously interrupting or even paralyzing the daily life in those communities. Manhattan in New York City is a coastal metropolitan area surrounded by several waterbodies. Interactions among these waterbodies at the time of flooding significantly affect the water level fluctuation. Devastation caused by Superstorm Sandy in October 2012 is evidence of how interconnected NYC waterbodies amplified the adverse effects on the shoreline and the coastal system. Proposing models to study the coastal areas where there are interactions among waterbodies that affect water level variations, and results in infrastructure vulnerability to flooding, is a challenge. In response to this challenge, the main objectives of this study are to simulate water-level variation in the New York Bay (lower Manhattan) during storm events, and to investigate the effect of storms on coastal infrastructure. For this purpose, by using the concept of system dynamics (SD), a dynamic simulation model (DSM) has been formulated to model the interactions among the water entities of the system for normal water level (harmonic tide) simulation. In the DSM framework, the governing equations among different waterbodies are used based on flood (forward) and ebb (backward) effects in each tidal cycle. The simulated water level by DSM is then fine-tuned using an artificial neural network (ANN) model, providing a better match between simulation results and the observed values. An envelope graph is then developed to simulate surge elevation at the time of storm in a representative station (Battery Park south of lower Manhattan). A load-resistance concept has been developed to investigate the effect of flood (storm) as load on the infrastructure resistance. Then, the probability of coastal infrastructure failure to flood load is determined. Results show the significant value of using a coupled SD-ANN platform to simulate water-level fluctuations at the time of storm. The developed load-resistance concept is a framework to link the simulated storm (load) to the resistance of lifelines (infrastructure) of coastal communities. This study targeted two important growing concerns in the coastal communities that can be further integrated, replicated, and extended in this and other similar settings.

Pressure management is an effective method to reduce water loss. A common way to reduce pressure and water loss in water distribution networks (WDNs) is to use pressure-reducing valves (PRVs). In recent years, pumps used in turbine mode [pump as turbine (PAT)] started to appear as a viable option for reducing pressure, water loss, and pipe failure in addition to energy production. In this study, a recently installed full-scale PAT system is presented for the Antalya City WDN in Turkey. The system was installed on a bypass line and operated in parallel with a PRV. The system’s performance was continuously monitored online for flow rate, generated power, and inlet and outlet pressures. The installed PAT system works efficiently for a wide range of inflows ($130–300{\mathrm{m}}^3/\mathrm{h}$), and the produced energy varies between 0.7 and 7.0 kWh for a reduction of approximately 100 kPa (1 bar) pressure head in general. To the best of the authors’ knowledge, the installed PAT system is the first full-scale application in a WDN to reduce water loss and to recover energy with a hydraulic model application.

The implementation of nodal demands in current transient modeling and analysis in the urban water distribution system (WDS) is usually based on steady-state conditions or empirical-based approximations, lacking for appropriate evaluation and scientific guidance of the nodal demand effect on transient modeling and analysis for accurately reproducing transient responses. This paper develops a probabilistic analysis and evaluation framework based on Monte-Carlo simulation (MCS) and global sensitivity analysis (GSA) methods for prior-known evaluation of transient nodal demand effects in WDS. An expression of the demand impact factor (DIF) is first proposed for indicating the influence of nodal demand effect based on a Lagrangian nodal demand model and implemented in the developed evaluation framework. Parameter sensitivity analysis is then conducted to evaluate the relevance and importance of the DIF to transient modeling and analysis in WDS. The results indicate that the DIF is highly relevant to the inherent attributes and initial steady state of the system, which can thus be used to facilitate the prior-known evaluation of the importance and influence of nodal demand effect in WDS. To demonstrate the application procedure and the applicability of the proposed probabilistic method, a realistic WDS is adopted for numerical investigation. The application results and analysis confirm the effectiveness and improvement of the developed method in this study on the guidance of accurate transient modeling and analysis under the influence of nodal demands in WDS.

In many modern water networks, an emerging trend is to measure pressure at various points in the network for operational reasons. Because leaks typically induce a signature on pressure, these routine measurements can be used to develop nonintrusive leak detection approaches. This research employs successive linear approximation methods, based on linear programming and mixed integer linear programming, in a simulation-optimization framework to explore an alternative leak detection methodology for urban water distribution networks based on pressure measurements. The methods attempt to minimize the absolute differences between observed and simulated pressure values at the sensors to determine a linear combination of leaks that most closely approximates the observed pressure pattern. Steady-state and time-varying models of differing complexity (from small published networks to a 27,000-node network for a U.S. utility) were used to test the method. Results are presented to illustrate the method’s effectiveness under different conditions. The methods are shown to work well when pervasive pressure data and hydraulic models representing true operational conditions are available. The methods developed in this work are not intended to replace traditional leak detection methods; rather, they are meant to work in concert with available methods to more accurately and efficiently isolate leak locations and reduce water loss.

Water, wastewater, and stormwater infrastructure systems are intrinsically interdependent, impacting the performance and operations of each individual infrastructure. The demands placed on and revenues generated for these infrastructures are impacted by human interactions, such as price elasticity and population dynamics. The consequences of these human–infrastructure interactions may be seen in shrinking cities, where chronic population decline has led to underfunded and underutilized infrastructures. This paper evaluates human-water and wastewater systems sector interdependencies in shrinking cities to explore water demand, utility revenues, public support, and payoff periods for retooling alternatives. To assess the dynamic behavior of the water sector systems, a hybrid framework is used that incorporates agent-based and system dynamics modeling. Driving the approach is data gathered from published and publicly available literature, two case studies, and a survey deployed to residents of 21 U.S. shrinking cities. This study quantifies the endogenous, physical interdependencies and the exogenous, complex human interactions. The epistemic uncertainty associated with human–infrastructure interactions is explored by incorporating stochastic parameters rather than the traditional deterministic parameters. Furthermore, the framework enables the assessment of interdependencies with parameters tailored to a city’s unique characteristics.

Reliability, resilience, and vulnerability are the most commonly used performance criteria for water supply planning and management. However, there is lack of understanding of the relationships among these criteria. This paper aims to reveal the relationships among them by using emerging many-objective visual analytics. To measure different aspects of water supply systems in terms of reliability, resilience, and vulnerability, a suite of five metrics are considered: water supply reliability, mean and maximum deficits of water supply, and mean and maximum durations of water shortage. Results obtained in this study reveal that both conflicting and synergetic relationships exist between reliability and resilience (mean deficit of water supply), and between vulnerability (mean duration of water shortage) and resilience (mean deficit of water supply) in different regions of the objective space. A more complete picture of the relationships among reliability, resilience, and vulnerability than reported in the prior literature is provided in this paper thanks to the use of many-objective analysis. This study provides an in-depth understanding of the relationships and can help decision makers make an informed decision in the management of water resources systems.

Low and constant pressure can be maintained throughout a water distribution system by setting the pressure at remote consumer locations and using the pressure to control the speed of a variable speed pump (VSP). The prospect of incorporating hydraulics theory into a controller is investigated, with the goal of improving on generic controllers. Five new controllers are proposed here, four of which depend on hydraulics theory. These controllers, which set the speed of the VSP, are investigated numerically. A parameter-dependent controller that does not require the flow in the pump to be known is developed and shown to significantly improve on the performance of conventional (parameter-dependent) proportional control (PC). Next, a parameter-free controller that requires the flow in the pump to be known is proposed and shown to outperform PC, even though PC has a tunable parameter, and perform comparably to the best new parameter-dependent controllers. The parameter-dependent controllers (when optimally tuned) perform best overall. The efficacy of many of the new controllers shows that hydraulics theory can lead to improved controllers.

An adaptive hybrid genetic algorithm (GA) and cellular automata (CA) method is proposed for solving implicit stochastic optimization of reservoir operation problems. The method is based on a decomposition approach in which the reliability constraints are handled with GA, whereas the resulting deterministic problem is solved with a CA model. Two versions, binary and integer GA, were employed for handling the reliability constraints of the problem. In the first one, GA was used to determine the success/failure pattern of the operation, whereas in the latter, only failure periods were determined with GA. The proposed method was used for monthly water supply and hydropower operation of an existing reservoir and the results are presented and compared with those of a GA model. To demonstrate the efficiency and scale independency of the model, short-term, medium-term, and long-term operations are considered assuming different target reliabilities. Comparison of the results with those of a GA model shows the superiority of the proposed method.

Security issues have become increasingly important within drinking water distribution systems, leading to the development of event detection algorithms (EDAs) to provide timely detection of intrusion events. The current study develops a systemwide event detection algorithm that integrates the alarm information from localized model-based EDAs at individual monitoring stations with the probabilistic contamination source identification algorithm to estimate the probability that a node-time pair may have been contaminated. An alarm threshold can be set on the resulting contamination probabilities, which results in a systemwide EDA that takes advantage of integrating the distributed detection at all sensors rather than relying on single sensors. The systemwide EDA was evaluated with two injection scenarios simulated at two different injection locations to assess performance when all sensors observed the event versus when only one sensor observed the event. The results showed that for both injection locations and both injection scenarios, the integrated systemwide EDA outperformed the use of only the localized EDA for a series of detection thresholds based upon a range of false positive rates. These results support the assumption that including information from multiple sensors would improve event detection performance.

A novel model combining monthly inflow forecasting and multiobjective ecological reservoir operations was developed to improve the efficiency of water resource allocation. To forecast monthly inflow values, an improved support vector machine (SVM) model was constructed using meteorological and hydrological data as inputs. To improve forecasting accuracy, this improved model, along with a basic SVM, was coupled to two data preprocessing methods: empirical mode decomposition and singular spectrum analysis (SSA). Comparisons between all the models revealed that the improved SSA-SVM model had superior predictive accuracy. China’s Three Gorges Reservoir (TGR) was taken as a case study for combining the improved SSA-SVM forecasting model with an operation model. A multiobjective ecological reservoir operation model was created for the TGR and optimized using the nondominated sorting genetic algorithm II. The simulation results indicate that the proposed coupled model for the TGR would outperform actual TGR operation; multiobjective ecological operations using inflow forecasting could alleviate water shortages. The societal, economic, and ecological implications of different TGR operation modes are discussed.

Water supply systems are diversifying and expanding in response to climate pressures and population growth. However, these water grids present challenges for the water supply manager in identifying optimal operating options for the short term. This study demonstrates the final step in a framework to address these challenges, multicriteria analysis, using a case study based on the South East Queensland water grid. A shortlist of nine water grid operating options have been identified, which are optimal in terms of minimizing total operational cost, maximizing water security, and minimizing spills from reservoirs, over a 5-year period. This study assesses the performance of each of these nine operating options against a wider set of 18 criteria reflecting cost, supply reliability, environmental flow, water quality, reservoir spill, and water security concerns. The weighted summation multicriteria analysis technique is used to combine and rank performance of the nine operating options against the 18 criteria. An operating option is selected that performs best on average across the 18 criteria and four scenarios of preference weights. This operating option comprises a set of operating rules that can form the basis of a short-term optimal annual operating plan.

The construction of water transfer projects can have a considerable impact on the operation of the receiving reservoir. This study investigates the change of the objective trade-offs in multiobjective reservoir operation problems due to the introduction of water transfer using a case study of the east-to-west water transfer project in northeastern China. Two optimization cases are constructed to analyze the trade-off changes: a base case with no water transfer that considers four objectives, i.e., minimizing industry water shortage, minimizing agriculture water shortage, minimizing water spillage, and maximizing ecological satisfaction; and a future postconstruction case that considers an additional objective to minimize the amount of water transferred. Results obtained from the case study show increasing water transfer substantially reduces the intensity of the competition between industrial and agricultural water shortages, and the objective trade-offs among water spillage, ecological satisfaction, and agricultural shortage index are substantially changed because of water transfer. In addition, the amount of water transferred with high efficiency regarding each objective is identified, and three solutions of different orders of magnitude in diverted water have been recommended for informed decision making considering efficiency and benefit. This study implies that many-objective visual analytics can be used to determine the optimal amount of water transferred in terms of water efficiency revealed in different objective trade-off spaces.

CANARY, developed by the U.S. Environmental Protection Agency (USEPA), is a freeware designed for contamination events detection in water distribution systems. CANARY has several imbedded statistical methods for analyzing water quality data to detect contamination events. The imbedded methods require calibration of their parameters to achieve good performance. However, the multiobjective nature of the problem creates a conflict between high sensitivity which results in good detection but with many false alerts, and low sensitivity which results in a small number of false alarms but with poor detection. In this work, a MOGA-CANARY add-in for CANARY autocalibration is introduced. This tool could be used by CANARY users to find the optimal parameter configuration that fits their system. MOGA-CANARY gives the users the whole set of Pareto optimal system configurations based on their defined objectives. The results of MOGA-CANARY are compared with existing manual calibration methods proposed in CANARY documentation.