Modelo para otimização da programação diária da operação de usinas hidrelétricas em cascata com a consideração das incertezas hidrológicas

Abstract

The daily schedule of hydroelectric plants aims to determine which units will be operating and their respective generation levels for the following day. The representation of this process is subject to imperfections in the representation of reality due to the multiple uncertainties that condition it. The present work focused on exploring considerations about the resolution method, on the assumptions of the problem and on its uncertainties in the optimization of the problem. The work methodology is divided into three stages. The work methodology is divided into three stages, applied in a modeling/formulation of the unit commitment problem present in the literature. The first step consists of developing an alternative solution methodology, based on Constructive Heuristics. In turn, in the second stage, changes in the modeling are analyzed/proposed of the problem in order to verify possible gains for those involved. Finally, the third step is to consider uncertainties in the modeling, in this case the forecasts of daily flows natural tributaries to each plant in the cascade are considered uncertain. The use of the alternative strategy to solve the problem proved to be an approach capable of providing a solution for the case studied. In regarding the second stage, two modifications of the original problem were carried out. In the first change, it was verified that the flexibility of the hourly restriction of attendance of the generation target increases the efficiency in the studied operation points. In the same sense, making it more flexible to meet the aggregate generation target for the entire cascade also improves the cascade’s operating point. The results of the third stage concern the insertion of hydrological uncertainties in the original problem. For the model adopted, there was no difference in the number of units dispatched per plant per time step, considering the variation of inflows. Finally, as a conclusion, the methodology proved to be able to explore different alterations of an original problem from the literature and evaluate its impact on results.