PVSCOPE™ combines and offsets meteorogical models, satellite images and real-time measurements to minimize prediction errors for each geographical scale and for each time horizon.

 

Our software merges multiple forecasts to generate the most accurate prediction :

1-Use of data from local observation : PV production measurement are collected by the information system installed on the production site (10 min timestep)

2- Use of satellite data : satellite images are provided by the Centre for Space Meteorology (10 to 30 min timestep).

3- Use of meteorological modelisation : day weather forecasts are recovered from agencies that perform the weather forecast and reactualize all 12h.

Then, this data are used to generate forecasts using several innovative algorithmic methods:

  • Corrected persistence models from measured production data (Very short term, less than 1h)
  • Forecast from geostationary satellite observations  (Short term. Between 1h and 6h.)
  • Climatology-based forecast (Between  6h and 24h)
  • Forecast from meteorological numerical prediction  (Between 6h and 24h)
  • Merging multiple forecasts to minimize prediction error

 

 

pvscopE2

PVSCOPE™ was developed in partnership with the Dynamical Meteorology Lab (Ecole Polytechnique, Paris), for generating day-ahead and intraday PV power production forecasts.