Utilizing deep learning towards real-time snow cover detection and energy loss estimation for solar modules

Conversion of solar energy using photovoltaic (PV) panels faces challenges due to snow accumulation on PV surface in cold regions. Despite existing methods to assess this impact, there remains a gap in real-time detection and accurate quantification of the energy loss. This study introduces a novel deep learning-based method for detecting snow coverage on PV panels for maximizing solar energy conversion. The model achieved a Dice score of 0.81 when trained on a diverse dataset of PV images, achieving a 44% improvement over conventional computer vision methods. Energy losses from snow on solar panels showed the model's predictions align closely with ground-truth data, achieving an error under 5% in snow coverage prediction. Six PV arrays were analyzed for energy loss estimation under varying snow coverage conditions. The analysis showed that the model could reliably predict energy losses due to snow accumulation with a mean error of 0.05 kWh/m 2 /month. The maximum energy loss was 0.23 kWh from a large PV array system covering 117 m 2 area. Analysis of the impact of snow coverage duration on energy loss showed a saving potential of 0.13 kWh/m 2 using timely clearing of snow coverage. This study highlights the effectiveness of CNN-based models in the early detection and measurement of snow coverage for improving the management and maintenance of PV systems. • Developed a method to detect snow on solar panels in real-time. • Model identifies snow with high accuracy of 81% on PV surface. • Precisely estimates energy loss due to snow under 5% error. • Demonstrates a 44% improvement over conventional computer vision methods. • Study enhances PV system efficiency, management, and maintenance.