Machine Learning Models for Photovoltaic Modules Operating Temperature Estimation in Semiarid Climate
Abstract
High photovoltaic (PV) modules operating temperatures impact energy conversion efficiency, particularly in semiarid regions. Our study investigates the application of supervised machine learning (ML) models to estimate PV modules operating temperature (Tpv) based on environmental. The dataset is used to train and evaluate different regression models: Linear Regression (LR), K-Nearest Neighbors (KNN), Random Forest (RF), and eXtreme Gradient Boosting (XGBoost). Among these, the RF model demonstrates superior performance and robustness. Global irradiance and ambient temperature consistently emerge as the most relevant predictors. The results highlight ML models potential to improve monitoring, thermal analysis, and operational optimization of PV systems in semiarid climates.
Keywords
Machine learning, Photovoltaic power, Operating temperature, Optimized physical model, Random forest