Constellation, le dépôt institutionnel de l'Université du Québec à Chicoutimi

Résumé

Real-time hydropower operations planning requires many optimization models in order to efficiently manage the hydropower system. The short-term hydropower model is used on a daily basis to dispatch the volume of water available to the turbines. This paper considers a stochastic short-term model. Uncertainty of inflows is represented with scenario trees, which are a discrete representation of the continuous distribution of the inflows. In addition, the complexity needed in the scenario trees to maximize the energy production in a rollinghorizon framework is investigated. Methods to generate scenario trees require input parameters that may be laborious to determine, and defining the structure of the tree in a real-time context is time consuming. Three comparisons are conducted to assess the complexity required in the scenario trees to obtain a good operational solution for the short-term hydropower optimization model. The first involves generating a set of scenario trees built from inflow forecast data over a rolling horizon. The second replaces the entire set of scenario trees by the median scenario. The third replaces the set of trees by scenario fans. The method used to build scenario trees requires three parameters: number of stages, number of child nodes at each stage, and aggregation of the period covered by each stage. The question of finding the best values of these parameters is formulated as a black-box optimization problem that maximizes energy production over the rolling horizon. Numerical experiments with real data on three hydropower plants in series suggest that using a set of scenario trees is preferable to using the median scenario, but using a fan of scenarios yields a comparable solution with less computational effort.