LAUSR

Reconciling simultaneous water quantity and quality aspects in drawing optimal reservoir operational strategy involves extensive computational burdens. Surrogate based optimization techniques (SBOTs) are common approaches to overcome computational bottlenecks of numerical hydrodynamic simulation models coupling evolutionary algorithm in simulation-optimization approaches. In this study, the reservoir high resolution CE-QUAL-W2 model is replaced by the lower resolution CE-QUAL-W2 and/or static ANN to speed up the optimization process. These surrogate models could consider the complex relationships to emulate the main dynamics of HR CE-QUAL-W2 model due to various reservoir operational strategies. The performance of various SBOTs, based on adaptive and sequential surrogate models coupled with particle swarm optimization algorithm, are evaluated in deriving optimal reservoir operational strategies. Then adaptive surrogate model, as the more efficient and accurate one, is applied to derive long term optimal reservoir operational strategy in the selective withdrawal scheme. The results show application of the proposed approach could enhance downstream water temperature, water demand satisfactions, and hydropower peak energy generation compared with the standard operation policy (SOP) in Karkheh reservoir during 15-year-time horizon.