A research team from the Department of Physics at the University of Jaén (UJA), in collaboration with the Department of Computer Science at Carlos III University in Madrid (UC3M), designed a database of freely available estimated photovoltaic solar energy that could be generated in Spain. Analyzed using Shirenda_pv, it has been developed using artificial intelligence and techniques, and when combined with other tools, it allows estimating the number of solar installations needed for nationwide implementation and the ideal location to create a national electrical system based on renewable energy.
The study, conducted under the Met4lowCar Research Project of the Ministry of Science, Innovation, and Universities, also received funding from the University and the Ministry of Research and Innovation, addressing a critical point in the transition to the mentioned system, such as the need for reliable resource databases (solar, wind, and hydro energy), high-resolution, temporal, and long-term coverage. «Only with this type of project database can the study be classified.»
Until now, physical models have been used to generate this type of database, simulating the behavior of actual solar plants in meteorological data. «The problem is that to simulate solar generation in Spain, parameters like tilt or panel technology are unknown. This leads to significant errors, making existing databases not reliable enough,» explains Pozo.
Experts applied a different methodology, combining climate generation data and real energy data from 2015-2020 provided by Spain’s Electricity (REE) to train machine learning models. They used three years for training and two for validation, and with the algorithms obtained, they estimated the production of photovoltaic solariums from 1990 to 2020 for all of Spain.
As a result, they created Shirenda_PV, a database of photovoltaic solar generation covering three decades with a time resolution. The tool is presented in the article titled «A new method for modeling renewable energy production using ERA5: solar photovoltaic energy,» published in the Renewable Energy Journal.
Where and when does energy occur most?
This database allows analyzing changes in solar generation over several years, the most stable and productive regions. «There are areas in Spain with very stable solar generation over 30 years, and others that sync especially with demand, even though their production is lower. This type of area can be of great interest for a renewable-based electrical system, helping reduce generation variability,» says the UJA expert.
The study identifies regions in the south and southwest, especially the extremity, as having the highest solar production. It also reveals that winter is the most challenging season in Spain, with a very pronounced fluctuation. This is due to the North Atlantic Oscillation (NAO), the most important phenomenon of climatic variability in Europe, which can vary by up to 20% from one winter to another. Specifically, in negative phases, generation can decrease by 16%, especially in the south, the country’s breadbasket, while in positive phases, it increases by 10%.
Solar, wind, and hydro energy
Shirenda_pv not only allows optimizing the location of solar plants but also anticipating climatic variability and its impact on energy production. Moreover, it is open access, facilitating its use by the research community, businesses, and administrations responsible for energy transition.
The researchers have also developed similar databases for wind and hydro energy generation, all freely accessible, to complement the record of renewable potential in Spain. In the upcoming project, already underway, they use all three to analyze the optimal electrical system based on these energies and study their behavior and reliability in climate change, among other objectives.