Empirical Analysis of Hydropower Scheduling

Abstract

A reservoir manager at a hydropower plant has to decide whether or not to release water in order to produce electricity, and the level at which to produce. These production levels have different related efficiencies as well as other related technical aspects. Often, the plant will produce at the most efficient, i.e. release water at a rate that produces the highest amount of electricity per unit of water. In this thesis, a structural estimation model was applied to an undisclosed hydropower plant in the Norwegian electricity price zone NO5, in order to discover the managers' preferences related to the different production levels. This model is based on time series models in order to replicate the managers' expectations of future conditions. The results show a greater willingness of the manager to produce at levels below than above the best efficiency point, which we argue is mainly due to the increased level of cavitation. They also imply that the reservoir managers' preferences have changed over time, showing an increased willingness to produce at production levels both above and below the most efficient level.