Non-linear multiobjective optimization for control of hydropower plants network

In this paper, a non-linear multiobjective optimization model is developed to obtain optimal annual scheduling for control of power generation in serial or parallel hydropower plants. Weighted-sum method is used to convert multiobjective optimization to single objective problem. Then, Sequential Quadratic Programming (SQP) is used, based on using Newtonpsilas method and Kune-Tucker conditions, to solve optimization problem. Maximization of power generation benefits and minimization of specific water consumption are assumed as objective functions. In this study, stochastic flows from external sources that enter to each reservoir (like rivers and rain) are assumed based on some scenarios for dry, median, and wet years. These various dry, median and wet water inflow scenarios are generated from the gathered fifty-year historical inflow measurements. Besides, turbine power generation is obtained from hill diagrams. The case study refers to hydropower plants network in Karoon River Basin in southern Iran. The corresponding optimization model has 240 variables and 660 constraints and MATLAB software is used to develop the model. Solving such non-linear multiobjective optimization model, annual rule curves that show optimal operation for reservoirs and total yearly energy production are presented for control of the cascaded units.