Experimental and Machine Learning-Based Electro-Thermal Performance Enhancement of PV/T Renewable Energy Systems with Advanced Cooling and Power Optimization Techniques for Sustainable Buildings

Ahmad F. Tazay (aElectrical Engineering Department, Faculty of Engineering, Al-Baha University, Al Baha, Saudi Arabia)

International Journal of Renewable Energy Development
Accepted Articles

The electrical power output and conversion efficiency of photovoltaic (PV) systems decrease significantly due to increased thermal losses at elevated module temperatures. Unlike conventional PV/T enhancement approaches, this study integrates recycled porous media, hybrid nanofluid PCM cooling, and data-driven optimization within a unified electro-thermal framework. This work experimentally investigates and optimizes the electro-thermal performance of a PV/T renewable energy system under real climat-ic conditions of Madinah, Saudi Arabia. An extreme gradient boosting (XGBoost) machine learning model was employed to pre-dict PV/T electrical and thermal performance and identify the dominant parameters affecting power generation and system effi-ciency. Four modified PV/T configurations were developed and compared with a conventional PV unit: Al₂O₃/water nanofluid cooling, recycled copper porous media, PCM containing recycled copper fillings, and Cu/water nanofluid integrated with PCM. These configurations significantly enhance heat transfer coefficients and thermal dissipation, leading to improved electrical per-formance. The results revealed that the mean surface temperatures of PV/T@Al2O3, PV/T@Al2O3+CP, PV/T@Al2O3+CPPCM, PV/T@Cu+CPPCM were reduced by 6.05 (3.16 oC), 10.30 (5.51 oC), 15.31 (8.17 oC), 17.38% (11.98 oC) compared to standard PV, respectively. Consequently, electrical energy production increased by 7.57%, 18.85%, 27.85%, and 31.49%, respectively, compared with the standard PV unit of 610 W. The PV/T@Cu+CPPCM configuration achieved the highest electrical and thermal efficiencies of 15.90% and 24.85%, respectively, with an electrical efficiency improvement of up to 22.39%. Techno-economic analysis indi-cates significant improvements in LCOE and payback period by 136 160% and 160 181%, respectively. The XGBoost model demonstrated high predictive accuracy and robustness, confirming its capability for real-time performance prediction and sys-tem-level optimization. The results demonstrate a strong coupling between thermal regulation and electrical output enhancement, confirming the effectiveness of the proposed electro-thermal optimization strategy. The improved thermal regulation enhances PV output stability and facilitates efficient integration with power electronic interfaces, including DC-DC converters, MPPT, and in-verter-based renewable systems. This work presents a novel integrated electro-thermal optimization approach that combines hy-brid cooling techniques with data-driven modeling to maximize PV/T system performance under real operating conditions. This work presents a novel integrated electro-thermal optimization approach that combines hybrid cooling techniques with da-ta-driven modeling to maximize PV/T system performance under real operating conditions for sustainable buildings in Saudi Arabia.