Short -term prediction of the generation of photovoltaic energy through the application of artificial intelligence to photographs of cloudiness on the generation system

Within the framework of sustainable development, the production of clean energy and energy sovereignty have been deployed in Ecuador several photovoltaic systems both at the experimental and production level. Although solar energy is practically inexhaustible, one of the main disadvantages of photovoltaic electricity generation is its fluctuation due to atmospheric effects; The clouds, depending on their height, thickness and speed, block the amount of light that impacts solar panels and consequently produce variation in the amount of electrical energy generated. The fluctuation in the generation of photovoltaic energy is as undesirable as inevitable; However, if a prediction of fluctuation can be established, this is very useful for generating companies since methods such as power soft -softening can be implemented to reduce the undesirable effects of fluctuation in the generation of electrical energy such as the condition of the quality of energy and the instability of electrical power systems. The implementation of a system is proposed that, based on the capture, storage and processing of photographs of the cloud position, can, through machine learning systems, predict within a prudent time the fluctuation in the generation of electricity, this information could be used to implement power softening systems. The results analysis will be carried out in the Micro-RED laboratory of the University of Cuenca.

20th UNIVERSITY RESEARCH PROJECT COMPETITION