LGPMI - Detecting Solar Panel Hotspots and Diode Failures with Thermal

Conference paper

Younes Benazzouz, Djilalia Guendouz
The first national conference on energy and its applications (CNEA2024), 2024

Read

Cite

APA Click to copy
Benazzouz, Y., & Guendouz, D. (2024). Detecting Solar Panel Hotspots and Diode Failures with Thermal Cameras and Deep Learning. In The first national conference on energy and its applications (CNEA2024).

Chicago/Turabian Click to copy
Benazzouz, Younes, and Djilalia Guendouz. “Detecting Solar Panel Hotspots and Diode Failures with Thermal Cameras and Deep Learning.” In The First National Conference on Energy and Its Applications (CNEA2024), 2024.

MLA Click to copy
Benazzouz, Younes, and Djilalia Guendouz. “Detecting Solar Panel Hotspots and Diode Failures with Thermal Cameras and Deep Learning.” The First National Conference on Energy and Its Applications (CNEA2024), 2024.

BibTeX Click to copy

@inproceedings{benazzouz2024a,
  title = {Detecting Solar Panel Hotspots and Diode Failures with Thermal Cameras and Deep Learning},
  year = {2024},
  author = {Benazzouz, Younes and Guendouz, Djilalia},
  booktitle = {The first national conference on energy and its applications (CNEA2024)}
}

Abstract

This research paper explores the use of deep learning, specifically the YOLOv11 model, in detecting defects in solar panels using thermal imaging. The focus is on two common types of faults: Hotspot Faults and Bypass Diode Faults. Hotspot faults, caused by malfunctioning cells, result in localized overheating, while bypass diode faults lead to thermal anomalies affecting entire lines of cells. The YOLOv11 model demonstrated strong performance in identifying hotspot faults, accurately detecting areas of overheating. However, its detection of bypass diode faults was less effective, revealing the need for further improvement in distinguishing and localizing these larger, more widespread thermal anomalies.