While renewable energy is already widely used in some countries, one of the biggest challenges in others is proving that it can contribute to European electricity supply at a level comparable to conventional power sources. Wind currently meets 8 % of European electricity needs, but is capable of meeting more than 12 % by 2020.

To integrate increasing amounts of wind power into existing electricity systems, renewable power plant operators must know in advance how much energy a plant can deliver. This is not straightforward, because unlike conventional power plants, the power output of a wind power plant is as variable as the weather.

Grid operators cannot rely on wind power plants to deliver the scheduled level of energy production in real-time and therefore, maintain an adequate level of power reserve for dispatch. If wind power production forecasts are not accurate, they risk under- or overcommitting their power resources.

Further down the market chain, energy traders also need to know how much energy a renewable plant can deliver over the next few hours and days. Wind power volume not only has an influence on the daily operation and management of the electricity grid, but also on price.

The wind power prediction system, developed within the EU-funded SAFEWIND project, is able to provide accurate and reliable forecasts. It is the commercial spin-off of a series of research and development activities that started with the ANEMOS project in 2002.

"The idea behind the first ANEMOS project was to bring together the expertise available in Europe and advance the state of the art in wind power forecasting," explains project coordinator George Kariniotakis from the Centre of Processes, Renewable Energies and Energy Systems (PERSEE) at MINES ParisTech and ARMINES, France.

Focus on 'extremes'

Following ANEMOS, research continued in the EU-funded ANEMOS.PLUS, which improved wind power forecasts for 'usual' operating conditions further still. SAFEWIND then focused on the prediction of extreme wind events such as 'ramps'.

A 'wind ramp' is a drastic change in wind speed at a specific location over a very short timeframe that causes a significant increase or decrease in wind power generation over a short time period. Such situations pose a challenge to power system operators seeking to balance generation with demand and to operate the grid in a secure way.

The SAFEWIND partners tackled the challenge of forecasting ramps by combining multiple weather models and measurements from synoptic stations with different statistical wind power prediction models. This approach provided optimal accuracy.

In addition, they developed an alarm system for 'extreme' situations that informs interested users of expected power surges and shut-down events as early as six hours in advance.

"Sceptics may doubt that it is possible to predict energy production, given the variability of wind," notes Kariniotakis. "But today, wind power forecasting is widely used and is a common business practice. Grid operators worldwide rely on accurate predictions for expected wind energy input."

Predictions worldwide

The SAFEWIND project drew on models from several of its partners, which rank among the top wind power prediction providers in Europe and are together responsible for predicting over 50 gigawatts of installed wind power capacity worldwide. Between them, project partners have commercially provided predictions to utilities and grid operators in Australia, Denmark, France, Canada, the US and many other countries.

As a result of this collaboration, the SAFEWIND team was able to improve on what was currently available – the energy industry has recognised this accomplishment and wind power predictions are now being used much more widely than before. The commercial adoption of the project’s research products has allowed for the incorporation of user feedback into the models in order to refine the forecasts' accuracy.

The SAFEWIND project concluded in 2012. In the years following, the prediction tools developed have been intensively tested by customers, both for forecast quality as well as software reliability and user-friendliness. Challenges remain before the wind power forecasts can fully satisfy the needs of their various end-users – the team is hoping to secure funding so that it may solve them.

SAFEWIND website