Modeling solar systems using artificial neural networks for performance prediction and investigation efficiency

One of the biggest problems in photovoltaic panel applications is that they have high costs in the initial installation stage. It is extremely important to know for the project and investments regarding photovoltaic power plants and the solar energy potential of the region where the power plant will be installed. The amount of electric energy produced by photovoltaic panels depends on air temperature, humidity rate, wind velocity and photovoltaic module temperature, and particularly solar radiation. Since knowing the output powers of the panels to be used in project and planning works for photovoltaic applications will provide a more accurate cost configuration, wrong investments will be avoided and the country budget will benefit from added value. Thus, it is very important to determine the effects of meteorological parameters involving non-standard test conditions of the region where panels will be operated on the panel power. In this thesis, by building measuring stations in three regions (Adıyaman-Malatya-Şanlıurfa) where environmental factors are different and the environmental factors (solar radiation, temperature, wind, humidity, PV module temperature) and output power of photovoltaic panels were measured and recorded for a year. By using Artificial Neural Network (ANN) algorithms on the large data set a model was obtained for estimating the power to be generated. In the trained ANN models, the estimation accuracy was 99.93%. By taking the data of the General Directorate of Meteorology as reference, models of ANN were trained by using Adıyaman province; and by using Malatya and Şanlıurfa data as test data, highly estimation accuracy was achieved. In the trained ANN models, the estimation accuracy was highly by using data of the Adıyaman Directorate of Meteorology. The artificial neural networks model was compared with the results of the Support Vector Regression model and it was found that the estimation was realized with higher accuracy with ANN. In addition, the effect of each meteorological data and the frequency of data collection on the ANN results were examined. With the artificial neural network models trained as a result of this study, the energy efficiency for the photovoltaic energy systems desired to be established by using meteorological parameters such as temperature, humidity, wind and solar radiation of various regions anywhere in the world can be estimated accurately. KEYWORDS: Artificial Neural Networks, efficiency, environmental factors, photovoltaic, renewable energy.