Predictive model of sunlight based on Artificial Intelligence for the optimization of electric power generation

Abstract

Due to the continuous depletion of resources and the increased social awareness of environmental problems, there is currently a constant need for energy, which has led researchers to seek new technologies to optimize the production of electrical energy. The present investigation has the type of applied investigation, it has a predictive level, its design is experimental and with a quantitative approach. The CRISP-DM data mining methodology was used, which is considered the de facto methodology for projects dedicated to extracting data values. In this sense, the data was collected, establishing the following variables: date, time, elevation angle and azimuth, later, the data was ordered in the Excel program, to be later processed in the Orange software. Which was used to predict the coordinates of the movement of the sun, so that it can optimize the electrical energy through the solar light tracker system. In addition, linear regression precision metrics were used and the design of the proposed prototype was carried out.