INVESTIGATION ON SOLAR PV DEFAULTS BY USING ARTIFICIAL INTELLIGENCE AND DEEP LEARNING

Abstract

One of the main sources of renewable energy in the globe is solar electricity. Electricity may be produced anywhere the sun shines with solar devices. These systems should be evaluated on a regular basis to prevent efficiency losses in photovoltaic systems. This paper discusses the use of thermal pictures captured by solar panels to identify cell, module, and panel problems in PV systems. During the investigation, an Infrared Thermal Camera would be used to capture the thermal photographs of photovoltaic systems at Istanbul Aydin University. Using the obtained thermal pictures, a thermal data group would be formed with damages from the PV systems. These generated dataset would be used to train the convolutional neural network (CNN) powered by Visual Geometry Group 16 (VGG16) deep learning model. The embedded AI (Artificial Intelligence) computing system VGG16 would be used for this training. It would be determined during the VGG16 network's training that the defects listed in the training was adequately identified.