Machine learning-driven insights into efficiency optimization of Si solar cells assisted by CH3NH3PbBr3 perovskite and WS₂ nano-structures

Document Type : Original Article

Authors

1 Department of Physics, Imam Khomeini International University, P.O. Box: 34149-16818, Qazvin, Iran

2 Department of Physics, Imam Khomeini International University, P.O. Box: 34149-16818, Qazvin, Iran Department of Energy Engineering and Physics, Amirkabir University of Technology, P.O. Box 15875-4413, Tehran, Iran

Abstract

Machine learning techniques, by leveraging advanced algorithms can pave the way for the development of more efficient solar cells by accurately predicting their efficiency and identifying the most influential features that affect their performance. Identifying the most influential features facilitates the optimization of experiments while predicting efficiency reduces the number of experiments. This approach saves time and costs and ultimately, efficient solar cells will play a useful role in solving the energy crisis as renewable energy sources. For this purpose, in the first step of this study, machine learning techniques are used to predict the relative efficiency of silicon solar cells for 58 experimental data after drop-casting with certain concentrations of tungsten disulfide and CH3NH3PbBr3 perovskite nano-structures. It was found that the extreme gradient boosting model has the best performance. This model also showed promising results for 12 new data. In the second step, Shapley additive descriptions will investigate the most influential feature on cell efficiency. According to the SHAP results, deposition of tungsten disulfide nano-structures after perovskite on the silicon solar cell surface has the best performance to increase efficiency. In fact, the sequence of drop-casting of each kind of the nano-structures influences the efficiency based on the different interaction mechanisms.

Keywords

Advances in Energy Sciences and Technologies
Volume 1, Issue 1
June 2025
Pages 98-106

Files

History

  • Receive Date: 11 January 2025
  • Revise Date: 26 January 2025
  • Accept Date: 01 February 2025
  • Publish Date: 01 June 2025

Share

How to cite

Statistics