NEXT: Solar & power forecast

Solar resource is day/night intermittent and subject to fluctuating weather conditions. It is therefore essential to have accurate forecasts to increase solar penetration in the energy grid. A powerful and multi-horizon forecasting API helps developers optimize their operations by predicting the amount of solar energy that will be available at a specific location and time.

Solar forecasts rely on a combination of satellite images, NWP/ Weather forecasts, and ground measurements. Satellite images provides detailed information on current cloud cover and solar radiation, while meteorological models (NWP) predict the movement of clouds on a larger scale and their impact on solar irradiance. Ground measurements provide information on local weather conditions and instant PV plant production. These input data are constantly updated to ensure forecast relevance and accuracy. 

NEXT : how it works >

For optimal photovoltaic plant operation, there are 3 types of solar forecasts classified by temporality: nowcasting, short-term forecasts, and long-term forecasts.

Nowcasting, is a near real time forecast (few minutes before) and generally require a sky observation system with hemispherical cameras. They enhance measurements from usual on-site sensors and less precise satellite images. They are particularly useful for isolated hybrid sites in microgrids where the alternative to PV is not instantly accessible or in cases of significant energy constraint (industry).

Short-term forecasts (intra-day) predict solar irradiance and energy production for the next few hours of the day. They are essential to enable solar energy developers to optimize their operations in real-time for daily operation and for network operators to manage the integration of solar energy into the electrical grid.

Long-term forecasts (day ahead) predict solar irradiance and energy production from several days to several weeks. They  enable solar energy stakeholders to plan their operations and make decisions to balance energy supply and demand, ensuring electrical supply within the electrical grid.

Meteorological weather forecasts and satellite images are very good sources of information for irradiance forecasting and are available worldwide. However, their temporal and spatial resolution is insufficient to detect short-term cloud behaviors at the local scale.

The measurements from various monitoring sensors usually available on-site provide very valuable indications about the instantaneous state of the system with much better accuracy. CalibSun's forecasting algorithm is capable of taking them into account to constantly recalibrate forecasts based on real measurements in cases where other data sources may have incorrectly anticipated short-term variability.

NEXT Advanced includes real-time data retrieval by CalibSun is done through secure pipelines and, is strongly recommended to further improve forecast performance on short-term horizons.