Advanced intelligent fault detection for solar panels: incorporation of dust coverage ratio calculation

Abstract

Fault detection is pivotal in ensuring optimum performance of photovoltaic systems. Faults can practically affect energy production and compromise overall system reliability. Responding to this challenge, convolution image processing automates visual inspection through digital imaging, allowing for the identification of defects from a sequence of images. This study introduces a comprehensive approach for smart detection of fault in solar panels. Therefore artificial intelligence techniques are applied, utilizing YOLO_NAS for defect identification and employing OpenCV for dust coverage rate calculation. The achieved results by using YOLO_NAS model for fault detection demonstrate significant precision and recall values, reaching 0.96 and 0.89, respectively. These findings confirm the effectiveness of the proposed methodology in smartly detecting faults, including dust, and accurately quantifying the dust coverage. This approach is promising for the management and maintenance of solar photovoltaic systems by enhancing their reliability and efficiency.