Diagnosis of External Faults in Photovoltaic Systems based on a Deep Learning approach

Due to the growing global demand for electricity energy, photovoltaic systems are becoming increasingly important as a continuous and environmentally friendly alternative. They ensure the continuity of electrical production in a healthy and sustainable manner. To ensure the efficiency and optimal performance of these systems, an effective diagnostic model is urgently needed to classify faulty and working solar cells. In recent years, deep learning methods have been used to analyse and process images, providing new insights and guidance in the field of fault diagnosis in PV systems. This research proposes a comparative study of the deep learning models ResNet50, VGG-19, and AlexNet to test their effectiveness in analysing and classifying defective solar cells from non-defective cells using EL images.