Estimation of the remuneration of hydro plants in a market environment using an iterative under-relaxation approach

Abstract

One of the challenges that generation companies having hydro stations are facing corresponds to build the most adequate bids to send to day-ahead markets, maximizing their profit and taking into account the expect inflows, market prices and the interdependency between hydro plants in cascades. As a contribution to address this problem, this paper describes a short-term optimization model to build one-week schedules for a set of hydro power plants so that the outputs can be used to bid in the day-ahead market. This model is coordinated with a medium-scale (one year) model that inputs the value of using the water in the short-term problem. The developed model considers the non-linear relationship between the electric power, the net head and the turbine discharge and it takes in consideration pumping as well. The solution approach is based on an under-relaxed iterative procedure based on the algorithm described in [1] and the players are considered as price-takers, so that market prices are input variables. To solve the medium-term problem we propose a similar model that sends information about the final week volumes of each power plant to the short-term problem. On the other hand, we conducted a scenario analysis regarding market prices and inflows to internalize uncertainty. Finally, the model was successfully applied to a real size Portuguese cascade and the paper includes several results to illustrate the application of the developed models.