Solar irradiation estimation with meteorological data using multi layer neural network approach

Abstract

The depletion of fossil fuels and the release of carbon dioxide into the atmosphere have increased the importance of alternative energy sources. Therefore, electricity generation is increasing using renewable energy sources. Solar energy has an important place among renewable energy sources. The reach of solar irradiation to the earth, which is an important parameter for solar power plants, depends on different climatic conditions. The efficiency of the solar power plant depends on the predictive accuracy of the solar irradiation. Accurate irradiation estimation improves the efficiency of the Photovoltaic (PV) plant, enabling accurate and efficient programming of the grid and improving power quality. In this study, simultaneous solar radiation values were predicted through a Multilayer Perceptron (MLP) model utilizing atmospheric pressure, relative humidity, ambient temperature, and wind speed parameters obtained from a station established for the measurement of meteorological data. Furthermore, the relationships between the input parameters employed in the prediction model and the output parameter, which is the solar radiation value, were investigated, along with their impact on the prediction accuracy. In the study using the error test method, solar irradiation values were estimated with high accuracy.