What Is a Wind Power Forecast?

A wind power forecast predicts how much electricity a wind farm will feed into the grid over the coming minutes, hours or days. It is usually delivered as a time series in 15-minute steps, in kW or MW, for a single wind farm or an entire portfolio.

How a wind power forecast is created

The starting point is numerical weather prediction (NWP): weather models such as ICON from the German Weather Service or the IFS from ECMWF calculate the state of the atmosphere on a three-dimensional grid. For wind power, the decisive variable is the wind speed at hub height, typically 100 to 160 metres above ground. Because model levels rarely match the hub height exactly, the wind speed is interpolated vertically.

The wind speed is then converted into power using the power curve of the turbine type. Since output grows roughly with the cube of wind speed in the partial-load range, small errors in the wind forecast lead to large errors in the power forecast. That is why good forecasts don't stop at physics:

• Several weather models are combined, because each has its own strengths depending on region and weather situation.
• Statistical and machine learning models learn from the wind farm's measurements how it actually behaves, including wake effects, terrain and availability.
• Curtailments and unavailabilities are kept separate so the model doesn't learn from distorted data.
• Intraday updates align the forecast with the latest measurements.

What wind power forecasts are used for

Direct marketers use them to nominate schedules on the day-ahead and intraday markets. Utilities and balancing responsible parties need them to keep their balancing group even and avoid balancing energy. Grid operators use them for congestion management and redispatch.

Forecast quality is measured with metrics such as the normalised RMSE (nRMSE). In one of our day-ahead benchmarks for a 32 MW wind farm, the nRMSE was 9.3 %, about half the error of a simple persistence forecast. You can read more about how we calculate our forecasts on our wind power forecast page, and you're welcome to test it on your own wind farms.