Optimization of the weekly operation of a multipurpose hydroelectric development, including a pumped storage plant

It is presented an optimization model based on genetic algorithms for the operation of a multipurpose hydroelectric power development consisting in a pumped storage plant (PSP) with weekly operation cycle. The lower reservoir of the PSP is supplied upstream from a peak hydropower plant (HPP) with a large reservoir and supplies the own HPP which provides the required discharges towards downstream. Under these conditions, the optimum operation of the assembly consisting in 3 reservoirs and hydropower plants becomes a difficult problem if there are considered the restrictions as regards: the gradients allowed for the reservoirs filling/emptying, compliance with of a long-term policy of the upper reservoir from the hydroelectric development and of the weekly cycle for the PSP upper reservoir, correspondence between the power output/consumption in the weekly load schedule, turning to account of the water resource at maximum overall efficiencies, etc. Maximization of the net energy value (generated minus consumed) was selected as performance function of the model, considering the differentiated price of the electric energy over the week (working or weekend days, peak, half-peak or base hours). The analysis time step was required to be of 3 hours, resulting a weekly horizon of 56 steps and 168 decision variables, respectively, for the 3 HPPs of the system. These were allowed to be the flows turbined at the HPP and the number of working hydrounits at PSP, on each time step. The numerical application has considered the guiding data of Fantanele-Tarnita-Lapustesti hydroelectric development. Results of various simulations carried out proved the qualities of the proposed optimization model, which will allow its use within a decisional support program for such a development.