Reservoir operations play an important role in water resources planning and management. During periods of drought, when inflows to the reservoir are low, it may not be possible to satisfy the target delivery (planned demand), and shortage occurs. Reservoir hedging rules, which include a set of rule curves and rationing factors, are used to guide the operation under such a situation. In particular, hedging rules are employed to balance the current shortage in supply with future storage to avoid severe water shortages in the future. In general, hedging rules are determined at the planning stage, or at the early stage of a reservoir’s operating life. However, after years of operation, the reservoir system, as well as the demand characteristics, may change. Hence, it becomes necessary to reevaluate and update the existing hedging rules to incorporate changes that have taken place. The purpose of this paper is to devise new hedging rules for an existing multireservoir system. An original mixed-integer nonlinea...

The study introduces a supervised committee machine with artificial intelligence (SCMAI) method to predict fluoride in ground water of Maku, Iran. Ground water is the main source of drinking water for the area. Management of fluoride anomaly needs better prediction of fluoride concentration. However, the complex hydrogeological characteristics cause difficulties to accurately predict fluoride concentration in basaltic formation, non-basaltic formation, and mixing zone. SCMAI predicts fluoride by a nonlinear combination of individual AI models through an artificial intelligent system. Factor analysis is used to identify effective fluoride-correlated hydrochemical parameters as input to AI models. Four AI models, Sugeno fuzzy logic, Mamdani fuzzy logic, artificial neural network (ANN), and neuro-fuzzy are employed to predict fluoride concentration. The results show that all of these models have similar fitting to the fluoride data in the Maku area, and do not predict well for samples in the mixing zone. The SCMAI employs an ANN model to re-predict the fluoride concentration based on the four AI model predictions. The result shows improvement to the CMAI method, a committee machine with the linear combination of AI model predictions. The results also show significant fitting improvement to individual AI models, especially for fluoride prediction in the mixing zone.

This paper develops a multicommodity flow model to optimize water distribution and water quality in a regional water supply system. Waters from different sources with different qualities are considered as distinct commodities that concurrently share a single water distribution system. The model can accommodate two-way flow pipes, represented by undirected arcs, and the perfect mixing condition. Additionally, blending requirements are specified at certain control nodes within the system to ensure that downstream users receive the desired water quality. The optimization model is highly nonlinear and solved by a hybrid genetic algorithm (GA). The GA is first used to globally search for the directions of all undirected arcs. Then a generalized reduced gradient (GRG) algorithm embedded in the GA is used to optimize the objective function for fitness evaluation. The proposed methodology was first tested and verified on a hypothetical system and then applied to the regional water distribution system of the Metro...