Non-invasive health status diagnosis of solar PV panel using ensemble classifier

In this article, a non-invasive health monitoring of solar photovoltaic (PV) panels using Artificial Intelligence (AI) is investigated. Proper maintenance of solar PV panels is crucial for ensuring their safe, reliable and efficient operation. An AI based non- invasive condition monitoring technique is adopted for diagnosing the health status of solar PV panels from the thermal images of the panels. The proposed approach involves acquiring thermal images, preprocessing, superpixel image segmentation, feature extraction and ensemble classification. The aim is to identify the faulty solar panels and monitor their health status with high accuracy. Superpixel image segmentation is introduced to enhance feature extraction reliability, reduce outliers and prevent overfitting in fault diagnosis. The proposed approach is tested on the collected dataset of solar PV panel thermal images and the results obtained are compared with other ensemble algorithms such as subspace, robust boosting, linear boosting and adaptive boosting. The results show that the proposed approach achieves high accuracy of about 99.99% in identifying faulty solar panels.