• News
  • Press Releases
  • Renewable News
  • Wind

Machine Learning in Support of Complex Wind Data Management


Karlsruhe, 21/02/2014. EWC Weather Consult take part in the European largest and most important wind event, the EWEA 2014 Annual Event, 10-13 March 2014 in Barcelona. In Hall 6, Stand 6B42, the weather service providers and wind experts from Karlsruhe will be presenting their solutions in complex weather data management. The service portfolio includes MCP (Measure Correlate Predict) wind measurement procedures, analyses for assessing wind farm locations, simulation calculations (based on historical data) and forecasts for power feed-in, together with solutions for optimum combination of (third party) wind forecasts.

With more than 60 nations participating, the event will provide the international wind community with a platform incorporating exhibition, conference and network facilities. EWC, weather service provider in the energy sector since 1999 with a leading position in the forefront of technology relative to the long-term correction of yield calculations and weather measuring service data processing, will actively approach traders, planners, project developers, investors and network as well as wind energy plant operators. To meet the ever-increasing demands of the wind market in terms of highly accurate forecasts, EWC compile complex physical forecasts, calibrate measuring data and combine a very wide diversity of forecast models and model parameters, MERRA reanalyses, satellite images and other wind data sources. Thus, energy traders, for example, are supported by a Meta forecasts based on machine learning techniques: an overall portfolio forecast improved through combining a range of various types as well as independent third-party forecasts. The advantage achieved is the highly improved forecast quality, which is evident, for example, as regards the day-ahead forecast error (15% lower MAE = Mean Absolute Error).

A new product launch at the E-world energy & water 2014 in Essen in February was the renewable energy index, a partial solution to the energy meteorological analysis and a long term index depicting current power production. This comprises meteorological time series and fields based on MERRA reanalysis and simulation calculations. It provides an opportunity for comparing current energy production data with a 35-year production review taking into account build-up data.

Further focal points for the South German wind experts at the fair are high-resolution, up-to-date wind (and solar) power forecasts for enhanced intraday- and day-ahead forecasts. These forecasts, over the past year, have found themselves similarly catapulted into a new and significantly higher quality league by using machine learning techniques. The short-term forecasts suitable for trading are calibrated with current feed-in data from existing wind farms and any errors eradicated by means of probabilistic fault characterisation.

 

To read the full content,
please download the PDF below.