Optimization and Performance Prediction of an Air‐Based PVT System Using Machine Learning
Nguyen Van Minh, Reda A. Haggam, Mohamed Bechir Ben Hamida, Amr Sayed Hassan Abdallah, S. Purushothaman, B. Srimanickam, Senthil SampathABSTRACT
Photovoltaic‐thermal (PVT) systems are promising renewable energy technologies capable of generating both electrical and thermal energy simultaneously. However, the performance of air‐based PVT systems is often limited by inadequate thermal management and difficulties in predicting thermo–electrical behavior under varying operating conditions. In this study, an experimental and machine learning‐based framework was developed to evaluate and optimize the performance of an air‐based PVT system with different internal obstacle configurations. Experiments were conducted under real climatic conditions in Chennai, India, at two mass flow rates. Support vector regression (SVR) and decision tree (DT) models were used to predict electrical efficiency, whereas response surface methodology (RSM) optimized the operating parameters. The SVR model showed better prediction accuracy than DT. The optimal operating condition achieved 13.5% electrical efficiency and 39.4% electrical–thermal efficiency.