MeteoFlow: Anticipating weather events to forecast and optimise renewable production
#R&D #Iberdrola projects #science
MeteoFlow is an integrated system for the prediction of medium and long-term meteorological variables in any type of renewable installation. Developed entirely by Iberdrola Renovables, the programme has been innovating for more than 15 years to adapt to the new challenges of the sector and has become a benchmark, with accuracy rates above those of other solutions on the market.
Pilar Orellana, MeteoFlow's performance monitoring manager, describes the project, its evolution and future challenges. Video voice transcription (Spanish version) [PDF]
MeteoFlow is Iberdrola's own computer system with the fundamental mission of estimating how much the Group's various renewable energy facilities will produce in the coming hours and days. This privileged information allows for optimal decision-making in the energy markets, as well as helping to find the time windows in which to schedule the Operation and Maintenance (O&M) of the facilities causing the least damage, among other functions.
Developed 100% by the Iberdrola Renovables team, and with more than 15 years of history, MeteoFlow has evolved to include the use of the most modern weather forecasting techniques, as well as machine learning, Artificial Intelligence and big data technologies. These advances have been made possible thanks to experts in different fields who are responsible for the maintenance, monitoring and development of the system.
Extensive technical expertise, detailed knowledge of renewable energy generation technologies, as well as seamless integration with other tools for facility operation has enabled MeteoFlow to maintain hit rates above those of other solutions on the market.
MeteoFlow The project in detail
Estimation of the electricity production of any renewable installation and weather forecasting at its sites.
- Integrated self-development system
- Projections for wind, photovoltaic and hydroelectric plants
- Serving more than 16GW in 12 countries
- Data processing and optimisation
- Real-time production data
- Prediction machine learning models
- METEOFLOW Optimal prediction of renewable production
- Historical production data
- Global weather information
- Sale of the energy produced
- Choice of windows maintenance
RENEWABLE ENERGY FORECASTING
Based on meteorological information - mainly wind forecasts for wind farms and solar radiation forecasts for photovoltaic plants - MeteoFlow produces production forecasts for all renewable facilities, including onshore and offshore wind farms, photovoltaic plants and, more recently, hydroelectric facilities. It also provides service to all countries where the Iberdrola group is present, tailoring forecasts to local requirements and end users.
From the data initially obtained, the programme refines the data with the latest machine learning and Artificial Intelligence techniques, currently achieving an hourly forecast of up to 96 hours and a daily forecast of up to 10 days for more than 450 wind and photovoltaic power plants.
All generation technologies, including renewables, must anticipate their output in the electricity markets by hours or even days, and there is a penalty if these forecasts are missed. The more accurate our energy production forecasts are, the higher the revenues of the installations, the greater their profitability and the greater the capacity for renewables to outperform other energy sources.
MeteoFlow has been providing the long- and short-term meteorological forecast maps on Iberdrola's intranet since 2016. In addition, for the specialised field of Renewables, it has a tool integrated within Domina - the set of tools for the operation and maintenance of facilities - that allows all users to have direct access not only to forecast maps, but also to meteorological and power forecasts adjusted to each of the facilities in Europe, the United States, Mexico, Brazil and offshore (offshore wind farms).
All of this is accompanied by a set of weather alerts for situations that can affect the facilities, such as extreme winds, storms or frost.
Another MeteoFlow functionality is hydraulic forecasting, which has a valuable application both in the calculation of the producible forecast in our hydropower facilities, as well as in supporting hydraulic operation and maintenance and energy management.
Through weather forecasting and observation, the tool indicates the probability of precipitation in 24 hours, the probability of snow, maximum and minimum temperatures and storms. In terms of flow prediction, it generates a daily forecast, up to a 10-day horizon, and is working on an hourly one, up to 48 hours.
Wave forecasting is essential for our many offshore installations: from wind turbines and the rest of the installations that make up an offshore wind farm, to the incipient floating photovoltaic energy installations. Wave forecasting, including wave height, supports the construction phase and helps optimise the operation and maintenance phase. MeteoFlow currently provides short-term wave forecasting services for up to 72 hours and long-term wave forecasting services for up to 10 days.
DISPLAY AND ANALYSIS TOOL
MeteoFlow offers various visualisation and analysis tools, on demand for Iberdrola's users, with which to try to get more and more out of the large amount of information stored in the system. The visualisation and study of this data provides support for operation and maintenance planning and energy management, as well as facilitating the analysis of results and deviations, re-evaluation and testing of new forecasting methods for the MeteoFlow team itself.
How MeteoFlow has evolved with big data
The statistical models used by MeteoFlow in its forecasts are fed with:
- Historical production data
- Weather forecast data from multiple sources
MeteoFlow with big data
More than 10,000 information points are selected
- Selection of information samples
- Data management and optimisation in a big data environment
- Selection of most influential points in prediction
The wind estimate was taken at 1 or 2 points on the vertical of a given installation.
- Much more accurate weather forecasting
- Increased use of renewable resources
- Maximising the supply of clean energy on the grid