A combination of strengths of physical and statistical forecasting methods

EW2 - The new solar power forecast

  • Direct marketing
  • TSO / DSO
  • Utilities
  • Individual systems for optimising self-consumption

Wind power forecast with feed-in management

  • Direct marketing portfolios
  • Intraday trading
  • automated trading strategies


The combination of both methods achieves a strikingly improved forecasting quality and represents a serious alternative to the market leader  .

Physical methodology

The calculation of the generation potential for every day and every hour of the year is a unique feature for solar power forecasts and is necessary because, for example, the highest position of the sun at the beginning of summer does not correspond to the highest extraterrestrial radiation and this in turn does not correspond to the highest irradiation on the ground in the annual cycle. Only after this maximum power has been determined - assuming that the whole of Germany is under a cloudless sky at normal temperature - is the data from the weather models taken into account and put into relation to it.

Statistical model mix with learning effect

We use a mix of the best weather models for wind and solar power forecasts and weight each forecast horizon according to the respective model specialities. Components include both, the best high-resolution individual model for weather forecasting and the leading models with global calculation. For longer forecast horizons, ensemble calculations find their way into our forecasts, whose bandwidth also allows statements about the confidence.

Maximum flexibility

Our database includes every single wind and PV plant in Germany. This enables us to deliver individual forecasts for direct marketing, for distribution grids or even nationwide forecasts of a comparably high quality.

Probabilistic labelling of uncertainty ranges

Weather forecasts over a longer period of time are naturally associated with uncertainties, which also applies in particular to the generation forecasts of solar and wind plants. We meet this challenge by explicitly naming the uncertainty range that enables sufficiently accurate forecasts, but at the same time indicates their probability of occurrence.

Independence from observational data

The physically based calculation methodology combined with the complete consideration of the individual plants enables us to be independent of feed-in values and other observation data.

System security

All data and programmes are backed up and mirrored several times so that the highest reliability in availability and punctuality is guaranteed. A sophisticated technical infrastructure also supports the provision of data in any desired format and thus without any integration effort on the part of the customer.

You want to know more?

Contact us