Document Type
Presentation
Publication Date
8-1-2014
Abstract
Hydroelectric generation is the major source of electric energy in the Pacific Northwestern region of the United States, and efficient operation of that system while meeting environmental constraints and reserve capacity demands is an important economic, environmental, and social issue. This paper describes efforts to develop a new stochastic short-term scheduling model (with perhaps a 3-week planning horizon) for the ten major reservoirs operated by the federal Bonneville Power Administration (BPA) on the Columbia and Snake River systems. The analysis incorporates time-delays (up to 24 hours in a model with time steps increasing from 6 hours initially perhaps to 24 hours); non-economic turbine dispatch with operational constraints; and inflow and load uncertainty (reflecting wind generation) through use of Ensemble Streamflow Predictions (ESP) augmented to include load uncertainties (ESLP). Synthetic ESLPs will be generated for the model testing effort. The intent is to evaluate the benefits of alternative representations of uncertainty subject to all of the operational constraints, both physical and those that result from environmental concerns. Large BPA storage projects can include many turbines of different types; for example, Grand Coulee has 27 turbines of 4 different types. To make system optimization faster and more reliable, concave “powerhouse” functions are pre-computed which are as economically efficient as possible given estimated turbine performance characteristics, and operational dispatch and release constraints. Powerhouse generation functions are forced to be concave if such constraints are consistent with the data; in other cases mandated fish-passage constraints result in non-economic turbine dispatch sequences and often limit allowable discharge ranges, both of which complicate the computation of the loading of individual turbines and the optimization of the hydropower system. Pre-computation of powerhouse functions is an effective decomposition technique for this large stochastic nonlinear optimization problem.
Comments
Session S7-03, Special Session: Optimizing Short Term Reservoir Operations III