Annual Performance Estimation of a Multireservoir System Including a Pumped Storage Plant for the Mean Hydrological Year

Abstract

An existing Romanian hydro system (two reservoirs/peak hydropower plants), where the addition of a pumped storage plant with a weekly emptying/filling cycle was planned, was selected as a case study. For the main upstream reservoir/hydropower plant, an operational policy with a monthly time step was derived using a stochastic dynamic programming model. The optimal weekly scheduling problem of the whole system was solved by a genetic algorithm-based model. The monetary income maximization, by 2014 market energy prices for produced/consumed energy, was chosen as the objective function; any gain from ancillary services was ignored. An original methodology for coupling the monthly and weekly time-step results was used to obtain the annual performance estimation within the mean hydrological year conditions. At that level, the inclusion of the pumped storage plant would lead to a gain to nearly double the annual performance of the existing hydro system.