Enhancing photovoltaic system efficiency through a digital twin framework: A comprehensive modeling approach

Photovoltaic (PV) systems contribute significantly to renewable energy generation, but their efficiency and reliability are often hindered by environmental conditions, thermal inefficiencies, and a lack of predictive operational insights. Existing solutions, such as advanced material design, cooling systems, artificial intelligence-based modeling, Internet of Things, address these limitations to some extent but often focus on isolated and limited aspects. This study introduces a new Digital Twin framework that integrates physical modeling based on MATLAB Simulink environment, analytical formulations, and artificial intelligence-based Gradient Boosting Regression Trees. Real-time data from an established on-grid 2.88 kW PV system is utilized to validate the framework, ensuring practical applicability and accuracy. Unlike traditional methods, this comprehensive approach enables real-time monitoring, predictive maintenance, and operational optimization under varying environmental conditions. The findings demonstrate significant improvements in system performance, showcasing enhanced predictive accuracy of 99.77 % and dynamic adaptability. Through the utilization of real-time data extracted from the established PV system, the framework provides a cost-effective solution for modeling large PV systems, ensuring practical and sustainable energy management with optimal operation.