Application of Artificial Intelligence in Detecting and Classifying Faults of Solar Panels

Abstract

Solar energy has always been an important field, which has received a lot of attention and research in the world. One of those problems is the methods of diagnosing, detecting, and classifying faults in the solar panel system. Indeed, such methods are being widely studied with the aim of improving power quality, reliability and as well as ensuring safety when operating solar PV systems. Solar panels are one of the most important elements of a solar power system and by itself there are always problems that can be mentioned such as short circuit fault, open circuit fault, aging condition, discolor, cracks on the surface, … This paper will focus on researching a new method combining I-V curve analysis of solar cells and artificial intelligence algorithms in solar cell fault detection and classification. The failures that can be detected and classified in this study include short circuit, partial shading and hybrid fault. In order to recognize certain faults, the data are firstly simulated and analyzed on the change of I-V characteristics. Then, Principal Components Analysis (PCA) algorithm is introduced to reduce the dimensionality of the data sets and by comparing the changes of parameters with the SVM model, the system will predict the number of solar panels with which type of fault in a solar cell branch. Overall, the research has gathered 680 samples in total and the result shows positive outcomes since the accuracy of recognizing exceeds 90%.