Optimization-simulation for short-term reservoir operation under flooding conditions

The problem of optimal operation of river-reservoir systems under flooding conditions is formulated in this study as a combination of particle swarm optimization (PSO) algorithm and a simulation model of river flood routing. Both hydrologic (Muskingum and Muskingum-Cunge) and hydraulic (solving Saint-Venant equations) flow routing methods are linked to the PSO algorithm with the objective function of minimizing flood damage in downstream areas. The Preissmann method is used to numerically solve the continuity and momentum equations. The developed optimization-simulation models were first tested in the case study of flood control operation of Bishop Dam, a benchmark problem in HEC-5 software. The models' performance was verified through their ability to reduce peak discharge at downstream control points compared to that resulting from applying gate regulation curve (GRC) operation policies available in HEC-5. Then the PSO-Venant model was used in the river-reservoir system of Upper Gotvand dam, a real case study in south-west Iran. Results show the significance of the model in determining optimal reservoir releases and minimizing flood damages.