Techn. Fakultät Website deprecated and outdated. Click here for the new site. FAU-Logo

Dipl.-Phys. David Bernecker

Alumnus of the Pattern Recognition Lab of the Friedrich-Alexander-Universität Erlangen-Nürnberg

Improving the reliability of renewable energies by using methods from computer vision and pattern recognition.
Solar irradiance forecasting

Solar power plants rely on the solar radiation of the sun for power production, therefore, their output is strongly influenced by the current weather conditions. Especially clouds occluding the sun can lead to huge drop-offs in the power production.

Irradiance forecasting methods have until recently relied on numerical weather models and satellite imagery. Both of these methods have in common, that their spatial and temporal resolution is not high enough to make accurate forecasts for single solar power plants.

Our approach to complement the existing forecasts consists of the following steps:

  1. Monitor the sky with an on-site camera
  2. Register the cloud movement with methods from computer vision (ie. tracking, non-rigid registration)
  3. Forecast the cloud movement to predict occlusions of the sun
  4. Establish a forecast for the irradiance based on possible occlusions in the future