Using a technique of statistics and artificial intelligence, known as clustering, scientists from the Solar Energy Institute of Polytechnic University of Madrid (IES – UPM) and the Institute of Micro and Nanotechnology of the Higher Council of Scientific Research (IMN-CSIC) have found a practical way to include in your calculations all the changes that occur in the solar spectrum to predict the production of photovoltaic solar energy. The study published in Nature Communication, allows finding in a few hours of calculation, the optimal design of multi-junction solar panel for each location.
Throughout the day and through the seasons of the year, the position of the sun and the atmospheric conditions change and this makes the light that reaches the photovoltaic panels has different characteristics. The most important change occurs in the spectral content of light, which consist of the distribution of colors of light. Thus, for example, at midday the light is more “blue”, while in the afternoon it is more “red”. As the photograph initially used as single pigment (black and white), to later go to use three colors, the solar panels of the future will be multi-junction type, combining several materials to better exploit the spectrum of sunlight.
Artificial Intelligence at photovoltaic energy
But the energy production of multi-junction panels depends to some extent on the color changes that occur in sunlight. For this reason, these panels are manufactured to produce the maximum energy for certain color of the light and, therefore, the changes produced by the position of the sun and atmospheric conditions give rise to losses in production. In order to reduce these losses, we try to design the panels looking for the optimum of global energy production and not for a specific color. But, due to the infinite variety of atmospheric conditions combined with different positions of the sun, this optimization is very complicated.
The work carried out by researchers shows that data sets with thousands of solar spectra can be reduced to a few characteristic spectra using artificial intelligence techniques, and used successfully to predict annual average efficiency based on cell design solar.
Renovables, E. (2019). La Inteligencia Artificial al servicio de la energía fotovoltaica. [online] ECOticias.com. Available at: https://www.ecoticias.com/energias-renovables/192922/Inteligencia-Artificial-servicio-energia-fotovoltaica [Accessed 29 Mar. 2019].
J. M. Ripalda, J. Buencuerpo & I. García. "Solar cell designs by maximizing energy production based on machine learning clustering of spectral variations". Nature Communications volumen 9, Article number: 5126 (2018)
Iván Garcia, William E. McMahon , Aron Habte , John F. Geisz , Myles A. Steiner , Manajit Sengupta , Daniel J. Friedman. "Spectral Binning for Energy Production Calculations and Multijunction Solar Cell Design". Progress in Photovoltaics: Research and Applications: 1-7 (2017)
Test related to solar panels can be carried out in specialized climatic chamber to check their characteristics.