Downscaling algorithms and assimilation of proprietary data
Training of neural networks on historical data
Constant updating of the forecast flow
Forecast correction and tuning with real-time data from the plant
Weather variables designed for plant maintenance and management
Alerts and critical-condition notifications for intense or extreme events
The combined use of the power curves of the individual wind turbines and the training of neural networks on the plant’s historical energy-yield data guarantee an accurate energy-yield forecast with limited errors in the short to medium term.
A dedicated post-processing algorithm provides, when real-time energy-yield data are available, a continuously updated and calibrated very-short-term forecast data stream.
You can also have a set of weather variables designed specifically for plant maintenance and management: wind speed and direction, air temperature, relative humidity, air density, pressure, forecast precipitation.
In the event of intense or extreme events, weather alerts and critical-condition notifications are available, with the aim of reducing economic losses and the risk of damage to personnel or plants.
hourly or sub-hourly time step
Forecast horizon from 24-48-72 hours to 2
weeks
2 or 4 daily updates
weather parameters: wind speed and direction,
temperature, relative humidity, precipitation,
air pressure and density.
wind power forecast (single turbine /
entire plant)
weather alerts in the event of events
potentially harmful and dangerous to
the plant and personnel
different output formats (csv, xml, xls) and different
transmission systems (ftp server, web server,
e-mail, API )
one month free trial
IDEAM Srl runs its modelling entirely in-house.
Grid resolution, physical schemes and parameterizations have been selected so as to reduce bias and increase the accuracy of wind-speed forecasts.
Using the power curves supplied by the turbine manufacturer gives the best wind-energy forecast estimate for a given wind speed and direction.
If time series of wind speed and energy produced at the specific site of interest are available, post-processing techniques and neural networks can be implemented to improve the final forecast of the plant’s energy yield.
If the plant’s production data are available in real time, a constantly updated forecast, corrected with a dedicated nowcasting algorithm, can be provided.
Share this page on social media