Optimizing Photovoltaic Performance: A Comparative Analysis of XGBoost and Gradient Boosting for MPPT

Abstract

This work compares Gradient Boosting Regression (GBR) and Extreme Gradient Boosting (XGBoost) for MPPT in PV systems. Both models were implemented and trained with three years of temperature and irradiance data available at NASA to predict the optimal maximum power point voltage (Vmp). The models developed were put to the test in MATLAB/Simulink. Experimental results showed that GBR achieved a power conversion efficiency of 95.7% under AM 1.5 irradiance (1000 W/m<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>), which was much higher than XGBoost of 73.5%.In addition, the accuracy of GBR was better than that of XGBoost (MAE = 0.18888) and presented higher robustness and more rapid convergence speed in the rapid fluctuation of irradiance. These results demonstrate the performance of GBR in real-time MPPT and the benefits of using ensemble learning strategies in intelligent solar energy harvesting.