Methodology for Forecasting the Power Production of PV Plants During Solar Eclipse
Keywords: multiple linear regression, artificial neural network, solar energy, PV, energy estimation
On July 2, 2019, a total solar eclipse occurred in the southern part of South America. This affected mainly the electrical systems of; Chile, Argentina, Peru, Bolivia, Paraguay, and Brazil. Considering this type of events, this research proposes a procedure that allows predicting the energy production of an industrial size photovoltaic plant under conditions of high variation of solar irradiance. The methodology is broken down into two stages: in the first, a heuristic methodology is developed to forecast irradiance, while in the second, supervised learning machines are used to estimate the production of photovoltaic plants. The results show that the overall error is of the order of 21%, which were contrasted with the measurements obtained during the eclipse of 2 July.
|IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
|Fecha de publicación:
|Año de Inicio/Término: