Skip to content
Conference

Machine learning-assisted prediction of UV-induced degradation in perovskite solar cells with functional nano-coatings

Oct 2026 · International Conference on Photonics, Smart Grid and Energy Technology · Vol 14367, pp. 1436703 - 1436703-9 · 0 citations · 12 references
Engineering

Abstract

Although perovskite solar cells (PSCs) now deliver power conversion efficiencies (PCEs) above 25%, ultraviolet (UV)exposure continues to limit their operational lifetime. Functional surface layers can suppress UV-driven damage while preserving useful optical transmission, yet coating design must be considered together with absorber composition and aging conditions. This work develops a machine-learning workflow for estimating PCE retention after UV aging in nano-coated PSCs. The assembled data set contains 1,200 records described by 14 variables, including Br, Cs, and FA fractions, initial PCE, coating category and optical properties, and UV-aging conditions. Random Forest, XGBoost, Support Vector Regression (SVR), and a Multi-Layer Perceptron (MLP) were compared through five-fold cross-validation. On the held-out test set, XGBoost produced R² = 0.884, MAE = 4.45%, and RMSE = 5.74%. SHAP analysis indicated that exposure duration, UV intensity, and coating thickness contributed most strongly to the predicted retention values. The resulting model offers a rapid screening tool for narrowing the coating-design space before resource intensive laboratory validation.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.