Objectives of the service
The REGREEN solution supports the transition towards climate-neutrality in the energy sector by providing insightful data analytics to locate the highest-potential sites for renewable energy production. In particular, REGREEN focuses on co-generation from three renewable energy sources (RES): solar, wind and hydropower. The purpose is overcoming the fundamental challenges to RES penetration, namely availability, reliability, and transmission stability. If a site has a high generation potential from multiple RES, a system can be created where the fluctuations or unavailability of one source can be compensated by the energy generated from another source, thus guaranteeing availability and stability without resorting to batteries. REGREEN combines Earth Observations and actual power generation data from existing RES plants as training features of a Machine-Learning model that forecasts the RES power production at any desired location on the map.
Users and their needs
The REGREEN solution targets the following customer segments, initially in the Italian market:
-
Energy providers, who need to reduce their dependence on fossil fuels for energy production and increase their share of production from renewables. The REGREEN solution informs them about the most effective energy source to be exploited at existing fossil-fuel plants that need to be phased out and on new terrains to be used for renewable energy plants.
-
Energy and Satellite service providers, who lack innovative products to offer to clients interested in renewable energy generation and with the REGREEN solution have the opportunity to provide additional intelligence for customers willing to enter the energy markets.
-
Local authorities, that have the power to attribute land destination of use and issue energy production licenses by only relying on administrative constraints (distance from landmark, critical infrastructure, etc.). REGREEN can enhance their decision-making process with analytics about the energy generation potential of sites.
-
Local energy communities, that would benefit from the insights gained through the REGREEN solution to decide on the most convenient and effective sources to exploit, especially in rural areas.
The interactions with selected stakeholders in the various segments during the Feasibility Study have shown that Local authorities are the most motivated in acquiring the REGREEN solution in its current design and functionalities. It is in line with their objectives of decarbonisation, it can be easily combined with their other initiatives in this direction (e.g. the creation of charging stations for electric vehicles, the local generation of multi-purpose “green” hydrogen, etc.), and it supports decision making.
Energy providers gave additional ideas and requirements to enhance the REGREEN solution and make it fit for their purposes. Besides the identification of new sites for energy production, their needs include to know about the construction and operational constraints of the identified areas, and the monitoring of the existing infrastructure especially to protect it from the impact of meteorological events.
Service/ system concept
The REGREEN solution uses satellite data combined with regional renewables power output as input for a set of artificial-intelligence (AI) models. The input of the AI models are features such as daily radiation, temperature, wind speed, humidity, and terrain elevation, slope, and orientation. The target dataset on which the AI models are trained with supervised learning techniques is the local power output data of renewable energy sources (RES). As a result, the AI models learn to associate the input features at different geographical locations to the total power generated at that site from the different RES. This will be converted into a set of three energy generation potential indexes, one for each renewable energy source, the values of which the AI will be able to predict for any geographical location. The expected output of the project are high-resolution georeferenced maps of the energy generation potential indexes, that enable to visualise the sites with the highest potential for energy generation and co-generation from multiple RES.
Space Added Value
The project uses Earth Observations and ERA5 reanalysis data from the Copernicus Data Store, including weather-related variables, such as solar radiation, temperature, humidity, wind speed, and ground properties, such as elevation, slope, and orientation. While future versions of the REGREEN solution aim to incorporate direct Earth Observations from Sentinel missions, starting with ERA5 data offers advantages in terms of homogeneity, spatial and temporal resolution (approximately 2-10km and 1 hour, respectively, over Italy), and extensive coverage over several decades. Compared to ground monitoring, space observations enable a faster, more scalable and cost-effective solution to identify high-potential energy generation sites.
Current Status
COMPLETED.
The project activities focused on determining the technical and commercial feasibility of the REGREEN solution. On the technical side, we developed a system based on multiple machine-learning algorithms that associate the meteorological and geographical properties of an area with the amount of photovoltaic and wind energy that can be extracted from that area, as a function of the installed capacity. Because our initial target dataset for the training was at low spatial resolution, we developed a pre-processing pipeline to downscale the data and increase its resolution. This modified target dataset was subsequently used for the training of the machine-learning algorithms. While the pre-processed dataset does not allow to fully capture the differences of the geographical areas where energy is generated, it is sufficient to mimic the properties of an ideal dataset with geo-located information of the renewable energy generation and to demonstrate the feasibility of the machine-learning approach. The results of this analysis are encouraging and indicate that with the availability of more precise data for the target dataset, our solution is capable of forecasting the expected energy generation from multiple sources at any location on the map. The current spatial and temporal resolution of the predictions are the same as Copernicus ERA5 Land, which was chosen for this study for its high quality and homogeneous grid spacing. Future improvements will include the option to use satellite data at higher spatial resolution.
The technical development focused also in developing a user-friendly interface to retrieve this data. On the client side, the user access the REGREEN solution through a web application. The web application consists of a predictions page where the user can interact with a geographic map to retrieve the required data. When accessing the service, the user visualises the map of Italy and can overplot the a heatmap of the forecasted energy which can be obtained by co-generating from wind and solar.
By clicking on a particular point within the land borders of Italy, and choosing a date, the user can retrieve energy predictions for the selected date, and aggregated information about the month and the year. In addition, three Energy Generation Potential indices assess the low, medium, or high potential for energy generation from photovoltaic and wind at the selected location, and the potential for co-generation from the two sources. The assessment is conducted considering the amount of energy predicted for the selected day and comparing these with the annual distribution of generated energy as benchmark. The relevant weather variables used to generate the energy predictions for photovoltaic, wind, and co-generated energy are shown in a separated panel which can be accessed on demand.
Concurrently with the technical feasibility assessment, the activities to determine the commercial viability of the solution consisted in identifying the target users of the REGREEN solution to better define the value proposition. Interviews and a demonstration workshop were conducted with key stakeholders, namely energy providers in the Italian market, the representatives of a public national authority that issues the necessary permits to use such land for the construction of renewable power plants, a large research institution, and representatives of a satellite service provider and of a photovoltaic-energy service provider and consultant. These activities of stakeholder engagement identified the criteria influencing our respondents' decisions regarding the implementation of Renewable Energy Source projects and showed that, at present, these stakeholders lack information about the expected energy generation potential at the production sites. It also indicated that additional information might be necessary to make the REGREEN solution fulfil the needs of energy providers. These additional pieces of information can be included in the current REGREEN solution as additional data layers to be visualised on the map.
