Design and development of the photovoltaic station of Donetsk National Technical University

Abstract

This article considers the principle of creation and operation of a stand called "Photovoltaic Station", which is located on the basis of "Donetsk National Technical University" in Pokrovsk. The mathematical model of this stand is also given. The purpose of this study is to analyze the operation of a real solar power plant and develop a mathematical model to predict the results that will be obtained when working with a physical model. The article uses both empirical research methods: description, measurements, experiment and comparison, and theoretical - formalization. To achieve the main task in the work the following stages are performed: Development of constructive decisions concerning change of an angle of inclination of solar panels; Improving models of graphical programming of photovoltaic systems using artificial intelligence; Forecasting the results of the solar panel using neural networks; Identifying the value of the efficiency of photovoltaic installations that use optimized parameters. The results of such a system will allow additions to the topology of the artificial neural network, which will help determine the correct direction of the neural network. The correct direction of the artificial neural network means the case when the difference between the objective function and the output of the neural network tends to converge. The paper presents the method of selection and parameters of the main elements of the installation, block diagrams, electrical diagrams and models of the developed structures. A mathematical model of the stand was developed and modeled in the Matlab software package. The obtained results allowed us to conclude that the mathematical model has the same properties as a real photovoltaic station and can be used in the future to model accidents and pre-accidents in the electrical network. It was also concluded that it is important to pay attention to the architecture and method of neural network training in this power system.