Alladini, Nahid, Wuthrich, Rolf and Hof, Lucas.
2025.
« Using random forest feature importance to rank environmental factors affecting PV degradation rates ».
In Proceedings of the CSME-CFDSC-CSR 2025 International Congress (Montreal, QC, Canada, May 25-28, 2025)
Coll. « Progress in Canadian Mechanical Engineering », vol. 8.
Preview |
PDF
285 - Using random forest feature import.pdf - Published Version Use licence: All rights reserved to copyright holder. Download (625kB) | Preview |
Abstract
This study uses machine learning techniques to explore the degradation of solar photovoltaic (PV) panels under varying environmental conditions. Using data from PV plants in Portugal alongside detailed weather datasets, the research aims to rank the influence of environmental factors on PV degradation. A Random Forest regression model was employed, demonstrating significant improvements in predictive performance through data augmentation and hyperparameter optimization. The analysis highlights wind speed, temperature, humidity, and cloud cover as critical degradation factors, providing actionable information to optimize solar panel installation and maintenance strategies. Despite promising findings, the study acknowledges limitations, including the size of the dataset, geographic scope, and computational constraints of the research process. This work contributes to the field by offering a systematic approach to understanding and mitigating environmental stressors on the performance of photovoltaic panels.
| Item Type: | Conference proceeding |
|---|---|
| Editors: | Editors ORCID Hof, Lucas A. UNSPECIFIED Di Labbio, Giuseppe UNSPECIFIED Tahan, Antoine UNSPECIFIED Sanjosé, Marlène UNSPECIFIED Lalonde, Sébastien UNSPECIFIED Demarquette, Nicole R. UNSPECIFIED |
| Professor: | Professor Hof, Lucas |
| Affiliation: | Génie mécanique |
| Date Deposited: | 18 Dec 2025 15:08 |
| Last Modified: | 18 Dec 2025 18:38 |
| URI: | https://espace2.etsmtl.ca/id/eprint/32355 |
Actions (login required)
 |
View Item |