Dedicated local-scale weather models
Tuning of the solar radiation forecast
Statistical post-processing algorithms and neural networks
Data analysis and forecasts also for plants with solar tracking
Algorithms trained on satellite data and time series of ground measurements
Direct and diffuse components of radiation
Local-scale weather models, specially tuned for solar radiation forecasting, are used in combination with statistical post-processing algorithms and neural networks to offer the best forecast of solar-radiation variables for each particular site.
The algorithms are trained using solar radiation data from satellites and time series of ground measurements, of both irradiance and electricity production.
A proprietary algorithm effectively separates direct and diffuse radiation, the two parameters needed to forecast the energy yield of plants with tilted panels or with solar tracking.
Specific data and products are available for operators in energy production and dispatching: aggregated and regional forecasts, the possibility of continuous short-term updates, etc.
For individual photovoltaic plants we provide forecasts for:
for example temperature, humidity, pressure, precipitation, wind speed
also split into the components:
• direct and diffuse radiation
• near-infrared, visible and ultraviolet
• radiation incident on a tilted plane
• radiation incident on a plant equipped with a solar tracker
forecasts with hourly or sub-hourly granularity
forecast horizon from 24-48-72 hours up to 2 weeks
forecasts updated 2-4 times a day
information on the reliability of forecasts through standard deviation and percentiles
In the field of forecasts dedicated to photovoltaics, IDEAM Srl has developed a proprietary algorithm (MOSRH) able to take into account the presence of humidity and cloud cover along the entire atmospheric column, and which outputs a solar radiation forecast with minimal errors, especially in the most complicated situations with partly cloudy skies.
For electricity transmission and distribution operators, IDEAM Srl can integrate the solar radiation forecast with the electricity production forecast for specific zonal aggregates, in order to reduce risks and improve grid stability.
In particular, the following parameters are available:
• estimate of instantaneous electricity production
• hourly forecast of electricity produced
• daily forecasts and forecasts aggregated by area
• actual (measured) irradiance values
• ramp rate
Post-processing algorithms and neural networks are used to improve forecast accuracy.
If the plant’s radiation and energy production data are available, they can be used to train the neural networks and offer the best possible product. Alternatively, the algorithms can be trained using satellite data and weather data from other station networks.
Share this page on social media