Research Article
Critical Feature Selection and Machine Learning-based Models for Monofacial and Bifacial Photovoltaics
@INPROCEEDINGS{10.1007/978-3-031-20398-5_12, author={Ansari Aadil Shahzad and Prajowal Manandhar and O. A. Qureshi and Ahmer A. B. Baloch and Edwin Rodriguez-Ubinas and Vivian Alberts and Sgouris Sgouridis}, title={Critical Feature Selection and Machine Learning-based Models for Monofacial and Bifacial Photovoltaics}, proceedings={Smart Grid and Internet of Things. 5th EAI International Conference, SGIoT 2021, Virtual Event, December 18-19, 2021, Proceedings}, proceedings_a={SGIOT}, year={2022}, month={11}, keywords={Generic Model Machine Learning Photovoltaics Monofacial Bifacial}, doi={10.1007/978-3-031-20398-5_12} }- Ansari Aadil Shahzad
Prajowal Manandhar
O. A. Qureshi
Ahmer A. B. Baloch
Edwin Rodriguez-Ubinas
Vivian Alberts
Sgouris Sgouridis
Year: 2022
Critical Feature Selection and Machine Learning-based Models for Monofacial and Bifacial Photovoltaics
SGIOT
Springer
DOI: 10.1007/978-3-031-20398-5_12
Abstract
Prediction accuracy has paramount importance in reliable PV solar plant performance. This helps with optimal plant design, economic assessment, smooth grid integration, and plant operations. Machine learning (ML) models help with faster, reliable, and accurate prediction of annual energy yield that is valid over a wide range of climatic conditions, module specifications, and site conditions. In this study, an ensemble-learning algorithm with regression trees is used to predict the performance of both monofacial and bifacial modules. Training data is prepared from parametric simulation results obtained using System Advisor Model (SAM) with an energy yield range of 120–584 kWh/m2for monofacial modules and 134–706 kWh/m2for bifacial modules. The results showed that ensemble-learning based ML algorithm can predict the energy yield of monofacial and bifacial modules with RMSE of 2.89 kWh/m2and 4.65kWh/m2.
