A Review on Modeling and Prediction of Soiling on Solar Photovoltaics and Thermal Collectors | ||
| Journal of Solar Energy Research | ||
| دوره 10، شماره 1، بهار 2025، صفحه 2135-2160 اصل مقاله (1.36 M) | ||
| نوع مقاله: Review Article | ||
| شناسه دیجیتال (DOI): 10.22059/jser.2025.392093.1544 | ||
| نویسندگان | ||
| Kudzanayi Chiteka1؛ Christopher Enweremadu* 2 | ||
| 1Department of Mechanical, Bioresources and Biomedical Engineering, University of South Africa, Science Campus, Florida South Africa | ||
| 2Department of Mechanical, Bioresources and Biomedical Engineering, University of South Africa, Florida 1710, South Africa | ||
| چکیده | ||
| Soiling of solar photovoltaic and thermal collectors can significantly reduce energy output, with losses reaching up to 78% in total yield and more than 1% per day in arid regions. Accurate soiling prediction is essential for optimizing system performance, minimizing downtime, and reducing operational expenses. This review critically examines empirical, analytical, and machine learning-based models used to forecast soiling effects. Empirical models, including transmittance loss and particulate matter-based approaches, report errors between <2% and 14%, while regression models show higher inaccuracies ranging from 40% to 93%. Analytical models such as the Bergin and Toth frameworks provide structured physical estimations but often require calibration and may overestimate under certain conditions. Machine learning and deep learning models demonstrate superior predictive performance, with image-based approaches achieving F1 score of 0.913 and models integrating environmental and image data reaching up to 97% accuracy. Despite these advancements, challenges remain, including limited availability of high-quality data, lack of generalizability across different climates, and insufficient real-time adaptability. This review also explores soiling mitigation strategies such as self-cleaning coatings, automated cleaning systems, and environmental monitoring tools. It emphasizes the need for hybrid, adaptive frameworks integrating artificial intelligence and Internet of Things technologies for improved accuracy and operational efficiency. | ||
| کلیدواژهها | ||
| Soiling effects؛ Machine learning؛ Soiling models؛ Energy yield loss؛ Soiling mitigation | ||
| مراجع | ||
|
| ||
|
آمار تعداد مشاهده مقاله: 590 تعداد دریافت فایل اصل مقاله: 374 |
||