For more than 20 years, NOVELTIS has been working towards improving and calibrating earth observation instruments embedded in satellites. The development of models, for example to forecast the evolution of clouds, was a natural evolution for NOVELTIS. The addition of A.I. allowed to get an excellent reliability of the forecast.

NOVELTIS got involved very early in the development of renewable energies, because observing the earth from space raises awareness on the necessity to reduce dramatically the carbon impact of human activity on Earth. Availability of clean energy at a reasonable cost is paramount for our current way of life and it is one of the most motivating subjects for NOVELTIS teams.

Globally, the name of the game is to manage as well as possible the available solar resource and to optimize the balance of the grid that receives the electricity. We are talking about profitability for the operators and anticipating the load for the grid.

Enables to know what will be deliverable on a 7 day horizon (1 to 7 days in advance).

The forecast is obtained from a WRF high resolution model with meso scale, recalculated once a day on all points of the globe, using the weather data of the GFS weather model. This digital deterministic multi-scale model of the latest generation transforms all these weather data into a forecast of irradiation, as a function of the reflectance and the luminance of cloud masses. These data are finally adapted to the local context, taking into account the horizon line (mountains, shadow zones, …) in order to narrow down as close as possible to the actual conditions of the plant. These fine scale post-treatments ensure the quality of the final forecast.

Calculations of the solar irradiation (Global Horizontal Irradiation) coming out of the models are converted into deliverable electricity (Kwh) by physical modelisation of the plant (slope, orientation, number and type of panels, number and type of inverters). If the MOS option is activated, the production of electricity is computed in a statistical manner.

Provides a forecast of the deliverable energy for the same day and a monitoring of solar power plants

The short-term forecast uses the satellite pictures of 5 geostationary satellites, so as to cover the entire planet. For example, Europe is covered by a Meteosat MSG geostationary satellite managed by EUMETSAT. Subsequently the Helioclim model is used to convert satellite pictures into cloud cover indexes, with a 3km resolution.

NOVELTIS has developed a model that gathers all available information and takes 3D clouds into account, to allow for the fact that high altitude clouds behave differently from low altitude ones. These clouds can be different in terms of direction, speed, luminance, cover… The movement of the various types of clouds is then modelled to obtain a reliable short-term irradiation forecast.

This solution enables the monitoring of photovoltaic farms, which advantageously replaces pyranometers, both in terms of reliability and operating costs.