Short term power forecasting

As wind-generated electricity increasingly impacts on many countries’ grids, its fluctuating nature makes a uniform approach to managing energy from all sources crucial. Accurate short-term forecasting of wind power can help to achieve this goal. Independent system operators who have to manage supply and demand on a grid system can avoid many of the balancing issues if accurate forecasts of wind output are available. Anybody trading in wind energy can negotiate a better price if they have precise information about the volume they have to sell at any particular time. A study published in 2006 revealed the potential benefit of using GL Garrad Hassan’s forecasting service to be a €10/MWh increase in the traded value of wind energy.

With global wind energy forecasting expertise and experience going back nearly three decades, GL Garrad Hassan now operates services for a range of clients, with over 33 GW of wind, in  countries all around the world. Utilities, energy traders, wind farm operators and ISOs rely on this highly accurate and reliable forecasting service with its operational availability of 99.9%.

The forecasting service combines GL Garrad Hassan’s wind energy expertise and experience with the best weather model inputs to provide state-of-the-art forecasts of wind farm output. The service is operated from a secure server farm in close proximity to the Internet backbone, to which data are sent and from which the power forecasts are issued in client-specific formats.

The forecasting process involves four steps: first, forecasts of meteorological parameters for the wind farm area are obtained from numerical weather prediction (NWP) providers; second, the weather input is refined to suit the wind farm site and sophisticated wind farm models run; third, detailed data from the wind farm are used to run auto-adaptive algorithms, which adjust the forecast to site-specific conditions; fourth, the client is provided with regularly updated site-specific forecasts of wind farm output delivered by e-mail, FTP or website. Typically, forecasts are updated hourly and each one contains 7-day predictions, but alternative frequencies and horizons can be easily accommodated.