Maximum power point tracking of the photovoltaic system using artificial neural networks

Abstract

In this study, a photovoltaic system is simulated for maximum power point tracking using an artificial neural network. The interest in photovoltaic systems is increasing with inadequacy of fossil resources and rise in demand for energy. The variation of photovoltaic system performance depending on environmental conditions reduces efficiency. It is possible to prevent this by reaching maximum power point. Many techniques have been developed to operate system at maximum power point. With the spread of artificial intelligence, smart systems have started to be used in determining maximum power point. Artificial neural networks are intelligent systems that have the ability to learn, generalize and make decisions. In this study, maximum power point is determined with an artificial neural network whose inputs are temperature and radiation. The system was simulated in MATLAB/Simulink environment. A total of 1000 data were used in training of network in which Levenberg-Marquardt algorithm was used, 70% in training, 15% in validation and 15% in testing stages. It was observed that power of photovoltaic panel decreased from 225.1W to 46.9W, and power of boost converter from 220.9W to 45.75W when radiation was reduced from 1000W/m2 to 200W/m2 at regular intervals. When temperature was increased from 25°C to 45°C at regular intervals, it was determined that power of photovoltaic panel decreases from 225.1W to 194.6W, and power of boost converter from 220.9W to 190.7W. At constant radiation (1000W/m2) and temperature (25°C), system has the highest efficiency value of 98%. At constant temperature and different radiation conditions, it was seen that efficiency decreased as radiation value decreased. Similarly, under constant radiation and different temperature conditions, temperature increase caused a decrease in efficiency. The results show that photovoltaic system with artificial neural network based maximum power point tracking reaches maximum power point under changing environmental conditions. © 2023, Gumushane University. All rights reserved.