New wind energy predictions save millions

Risø National Laboratory, Technical University of Denmark, The Danish Meteorological Institute, Elsam and Energi E2 in Denmark have jointly developed a new method for predicting the energy produced by wind turbines. The method will save millions for electricity producers and consumers

In Denmark more than 5,000 wind turbines produce an average of more than 20 per cent of the Danish power consumption.

The electric utilities must supplement wind energy production with power from the liberalised electricity market. If they buy too much they have to sell the surplus power cheaply, and if they buy too little they have to buy extra at additional cost. It is thus extremely important to know precisely the volume of power, which the turbines are expected to generate, preferably a number of days ahead.

For more than ten years, scientists from Risø and from the Department of Informatics and Mathematical Modelling (IMM) at the Technical University of Denmark, DTU, have provided wind energy predictions to the electric utilities 48 hours ahead. These predictions have been based on meteorological data from the Danish Meteorological Institute, DMI, and local conditions. A new class of models is now also able to predict the uncertainty of the prediction up to a week ahead. The uncertainties are calculated by including predictions from many meteorological models or model runs, a so-called ensemble of meteorological models. The new models are thus able to predict the amount of the energy production in the coming week at a certain location, e.g., a wind farm or an entire region, and make an uncertainty estimate for that particular prediction.

The models for predicting wind energy are primarily supplied to electricity utilities which in this way know how much power they need to purchase on the liberalised electricity market in order to meet demand and ensure a stable power supply.

“The models allow the electricity utilities to save money”, says Gregor Giebel, Senior Scientist at Risø. “In Denmark, the electricity utilities must typically trade power for the next 24 hours by 12 noon the previous day. With a certain prognosis, the utilities are able to plan optimally and if it is uncertain, they need to have costly backup reserves.” However, Gregor Giebel points out that this is a financial game which will not result in the consumers being without power. Conversely, good prognoses help reduce the electricity price and the risk of power outages. Moreover, the new prognoses can be used to predict fuel consumption at power plants, or to decide when to disconnect a unit from the system for maintenance.