After uploading the historical photovoltaics generation data, an intelligent algorithm identifies the characteristics of the plant.
© iStockphoto
PVCAST is a project that develops a forecasting service that generates the seven day forecast for the production of a photovoltaic plant. The plant specifications, for example, component and inverter types or the orientation and incline of the system no longer matter.
The system is based on self-learning algorithms incorporating historic production data. Even temporary intrusions of shade or dirt responsible for losses in yield are automatically taken into account.
The use of high quality weather forecasts from a leading meteorological service supplier, in addition to satellite data, enables yield forecasts to be generated for numerous locations in hourly resolution and a forecast horizon of up to six days.
Fraunhofer Institute for Industrial Mathematics ITWM