Journal of Water Resources Planning and Management

The short-term optimal hydrogeneration planning is a complicated nonlinear constrained optimization problem with water delay time. To overcome the shortcomings of a standard genetic algorithm, this paper proposes a new real-value encoding self-adaptive chaotic genetic algorithm to solve this problem, which designs a new crossover operator in light of probability distribution function and a self-adaptive chaotic mutation operator combined chaotic dynamic character with artificial neural network theory. Constraints can be dealt with by using a simple direct comparison penalty function method without the need of any penalty coefficient. The feasibility of the proposed method is demonstrated for short-term generation scheduling of two test hydrosystems and the test results are compared with those obtained by the standard genetic algorithm in terms of solution quality and convergence characteristic. The simulation results show that the proposed method is capable of obtaining higher quality solutions.

It has often been cited that a water distribution system (WDS) hydraulic model needs to be highly accurate before it may be used in combination with a water quality model (WQM) to simulate the dispersion and decay of a residual disinfectant. However, even a well-calibrated WDS hydraulic model may not have data relating to the specific water demands during a given period, which may impede WQM calibration. This study examines using residual disinfectant data to calibrate a WQM under unknown or uncertain demands by calibrating a residential demand multiplier pattern (DMP) in tandem with the WQM parameters. Two artificial scenarios and one real case study are investigated. The artificial scenarios are used to (1) verify the proposed methodology under ideal conditions and (2) validate the proposed methodology when the hydraulic model and calibration data contain realistic errors. The real case study uses residual chlorine data and a WDS model for which a hydraulic and WQM calibration had been performed previously. The estimated demands from the real case study are validated using tracer test data. Results from the artificial case studies may be summarized as follows: (1) the proposed methodology can estimate the demands and calibrate WQM parameters correctly, although increasing model and calibration data errors adversely affect calibration results; (2) the calibrated WDS models reproduce the true residual chlorine concentrations with very little error. Results from the real case study indicate that the original WQM calibration was performed using underestimated WDS demands. Tracer test data confirm that the calibrated DMP provides good hydraulic velocities. The calibrated WDS model from the real case study is in good agreement with measured residual chlorine concentrations. The mean absolute error between the simulated chlorine concentrations from the calibrated network model and the observed values is $0.059\mathrm{mg}∕\mathrm{L}$ .

Reducing energy consumption of water distribution networks has never had more significance than today. The greatest energy savings can be obtained by careful scheduling of operation of pumps. Schedules can be defined either implicitly, in terms of other elements of the network such as tank levels, or explicitly by specifying the time during which each pump is on/off. The traditional representation of explicit schedules is a string of binary values with each bit representing pump on/off status during a particular time interval. In this paper a new explicit representation is presented. It is based on time controlled triggers, where the maximum number of pump switches is specified beforehand. In this representation a pump schedule is divided into a series of integers with each integer representing the number of hours for which a pump is active/inactive. This reduces the number of potential schedules (search space) compared to the binary representation. Ant colony optimization (ACO) is a stochastic meta-heuristic for combinatorial optimization problems that is inspired by the foraging behavior of some species of ants. In this paper, an application of the ACO framework was developed for the optimal scheduling of pumps. The proposed representation was adapted to an ant colony Optimization framework and solved for the optimal pump schedules. Minimization of electrical cost was considered as the objective, while satisfying system constraints. Instead of using a penalty function approach for constraint violations, constraint violations were ordered according to their importance and solutions were ranked based on this order. The proposed approach was tested on a small test network and on a large real-world network. Results are compared with those obtained using a simple genetic algorithm based on binary representation and a hybrid genetic algorithm that uses level-based triggers.

An inexact chance-constrained linear programming (ICCLP) model for optimal water pollution management at the watershed scale was developed. We selected the net expenditures of the alternative strategies, including initial capital investment and operating costs, as the objectives of water pollution management. The total environmental capacity of the water bodies at different probability levels $\left(q_i\right)$ was considered a key constraint; other constraints included in the model were government minimum requirements on farmland area, land cover, treatment rate of domestic wastewater and rural wastes, and certain technical constraints. The ICCLP model was applied to Lake Qionghai watershed in China for water quality improvement with the goal of achieving a minimum total cost. Alternative strategies were incorporated following discussions with shareholders and experts. A three-period optimization was conducted based on the alternative strategies; the model parameters were based on field investigations. Five probability levels were considered in the model: $q_i=0.01$ , 0.25, 0.50. 0.90, and 0.99. The model results showed that the total optimized costs were between US $\$[55,710.86,80,691.81]\times {10}^4$ and US $\$[72,151.39,101,338.6]\times {10}^4$ under different probability levels. The model results indicate that soil erosion treatment, nonpoint source control measures, and rural waste treatment have much higher costs than other strategies, and our findings indicate that the ICCLP model can effectively deal with optimal water pollution management under uncertainty at the watershed scale.

Many anthropic factors may contribute to the alteration of the natural evolution of the river environment. Among them, probably the most severe is the regulation of the stream flow by a dam. To minimize this impact normally a minimum flow release is guaranteed downstream from the dam. The minimum discharge constitutes an important parameter for the evaluation of the real impact of the reservoir on the catchments downstream. The assessment of this parameter is not easy because two conflicting objectives have to be taken into account: the maximum use of water and environmental protection. The objective of this study is to optimize, apply, and discuss different methods for evaluation of the minimum discharge found in the technical literature to the reservoir on the Menta stream in the province of Reggio Calabria, Italy. In this process particular emphasis is placed on the preservation not only of acceptable levels of discharge during dry months (summer) but also of the seasonal variability of the discharge that largely contributes to the safeguard of the biological potentialities of the stream. Among the methods tested, the one adopted by the Autorità di Bacino del Fiume Serchio (Serchio River Basin Authority, Italy), although modified, provides a good indication of the minimum discharge to be adopted.

A contaminant intentional intrusion into a water distribution system is one of the most difficult threats to address. This is because of the uncertainty of the type of the injected contaminant and its consequences, and the uncertainty of the location and intrusion time. An online contaminant sensor network is the main constituent to enhance the security of a water distribution system against such a threat. In this study a multiobjective model for water distribution system optimal sensor placement using the nondominated sorted genetic algorithm II is developed and demonstrated using two water distribution systems of increasing complexity. Tradeoffs between three objectives are explored: (1) sensor detection likelihood; (2) sensor detection redundancy; and (3) sensor expected detection time. Pareto fronts are plotted for pairs of conflicting objectives, and simultaneously for all three. A contamination event heuristic sampling methodology is developed for overcoming the problem of contamination event sampling.

The design of a sensor-placement scheme capable of detecting all possible contamination events for a water distribution system before consumers are put at risk is essentially impossible given current technologies and budgets. It is, however, possible to design sensor-placement schemes that optimize related objectives (e.g., minimize expected volume of contaminated water consumed prior to detection), but this requires the availability of hydraulic and water quality models for the distribution network and significant computational power, which are the main obstacles to the identification of optimal sensor locations. This paper describes a different approach that reduces the problem’s complexity by expressing a water distribution system as different graphs based on the information readily available from most, if not all, water utilities. The approach provides critical policy and decision support for utilities when hydraulic and water quality models are not available and/or when simulation-based techniques are computationally infeasible.

Following the identification of a contaminant in a water distribution network, a variety of response actions must be examined in order to implement the most beneficial consequence management strategy. Optimization techniques can be employed to determine the cost/benefit of reducing impacts to the network from contamination by isolating and/or flushing the system. In this current effort, we employ a genetic algorithm to minimize contaminant concentrations in a network while minimizing the cost of demand alteration. Application of this technique to two relatively simple networks demonstrates the usefulness of this optimization method as a consequence management strategy to reduce contaminant concentration. For the EPANET Example 1 network, the optimal response solution included closure of two pipes and alteration of the demand at one node, reducing the total network concentration by 95%, with a 73% increase in total network demand. For the Anytown network, the optimal response solution included altering the demand at four nodes, which resulted in a 12% increase in total network demand, while closing four pipes reduced the total network concentration by 54%.