Posters

Abstract

The growing penetration of a non-dispatchable resource of wind energy is rising concerns about future issues on power systems caused by power variability. This penetration could lead to system deviations, which can be dealt with Energy Storage Systems (ESS). The combination of a generation system such as wind turbines and an energy storage system is known as hybrid farm. This configuration allows new business opportunities, but also presents new technical problems. Until now, the most common business strategy for wind energy projects to reduce revenue risks is to establish Power Purchase Agreement (PPA) with an utility or an industrial customer. The advantages of a PPA include long-term price guarantees to cover the full lifetime of the project, which facilitates financing opportunities. However, PPAs are often set at prices with heavy discounts, compared with spot market prices, and under delivery penalties tend to appear on PPA contracts. This situation has carried owners to considerate the spot markets as an alternative. Spot markets requires short-term commitments which tend to be difficult for a non-dispatchable energy source such as wind. Nevertheless, ESS can provide flexibility to the overall system and help participation on such markets. This presentation is focused on the relevant aspects when defining a solution for the management of a ESS participating in spot markets by using optimization theory. It is common to prepare the energy bids for the market with an optimization problem, in which the decision variables are the power signals scheduled for each market period. However, neither price or generated power are known before presenting such bids to the market, Short-term forecasting techniques are mandatory to make a decision about how to participate in each market session. Such forecast tends to include uncertainties. Recent literature has dealt with this issue by using stochastic optimization techniques on this application. This presentation shows an analysis of the singularities of wind power forecasting, caused to the non-linear relationship between wind speed and generated power. Traditional stochastic optimization methods are not prepared for such non-linearities, since they assume the uncertainties have the shape of a normal probability density function. To correctly deal with this issues, non-conventional stochastic methods must be considered. On the other hand, electricity market rules strongly define the optimization problem and therefore must be studied and considered when defining the optimization model. Intuitively, a higher time horizon would generate larger benefits since allows to better planning, however, it is natural to forecasting techniques to increase its uncertainties as the time horizon grows. This higher unpredictability could cause greater deviation costs. This presentation will explain how the time window of power scheduling is restricted by market rules and how short-term decisions in market scheduling are nearly not affected by future events.