Scholarlink

<p>High operating temperatures are a major limitation for photovoltaic/thermal (PVT) systems, particularly in hot and arid regions, because they reduce electrical efficiency and limit practical energy output. Although several studies have investigated nanofluid and hybrid-nanofluid cooling for PVT systems, limited attention has been given to the experimental evaluation of TiO<sub>2</sub>–ZnO hybrid nanofluids at multiple weight fractions under real outdoor Iraqi climatic conditions, with direct comparison against conventional water cooling and machine-learning-based prediction of overall system efficiency. This study experimentally investigates the cooling performance of a PVT system using TiO<sub>2</sub>–ZnO hybrid nanofluids under outdoor Iraqi climatic conditions in May 2025. Hybrid nanofluids with weight fractions of 0.10, 0.15, and 0.25 wt% were circulated through the cooling system at a constant flow rate of 1&#xa0;L/min and compared with conventional water cooling. The system performance was evaluated in terms of photovoltaic cell temperature, electrical efficiency, thermal efficiency, power output, and overall efficiency. The results showed that the hybrid nanofluid improved the PVT performance compared with water cooling, with the best performance obtained at 0.25 wt%. At the peak operating condition, the 0.25 wt% hybrid nanofluid increased the electrical efficiency, thermal efficiency, power output, and overall efficiency by approximately 10%, 26.7%, 15.9%, and 20%, respectively, while reducing the photovoltaic cell temperature by up to 10.7%. In addition, machine-learning models were developed to predict the overall efficiency using the experimental dataset. The multilayer perceptron regressor (MLP) achieved the highest prediction accuracy, with R² = 0.9946, followed by the random forest regressor (RF) with R² = 0.9910, whereas the sequential neural network (SNN) showed lower accuracy with R² = 0.8825. These findings demonstrate that TiO<sub>2</sub>–ZnO hybrid nanofluids can effectively reduce photovoltaic module overheating and improve the thermal, electrical, and overall performance of PVT systems. The results also confirm that machine-learning models can serve as reliable tools for predicting PVT system behavior under real outdoor operating conditions.</p> Graphical abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine learning-assisted performance enhancement of pvt systems using TiO2–ZnO hybrid nanofluid cooling under Iraqi environmental conditions

  • 期刊论文

  • 发表日期：

    2026年7月16日

查看全文
  • Mustafa Hidayet Ibrahim,
  • Adnan M. Hussien

摘要

High operating temperatures are a major limitation for photovoltaic/thermal (PVT) systems, particularly in hot and arid regions, because they reduce electrical efficiency and limit practical energy output. Although several studies have investigated nanofluid and hybrid-nanofluid cooling for PVT systems, limited attention has been given to the experimental evaluation of TiO2–ZnO hybrid nanofluids at multiple weight fractions under real outdoor Iraqi climatic conditions, with direct comparison against conventional water cooling and machine-learning-based prediction of overall system efficiency. This study experimentally investigates the cooling performance of a PVT system using TiO2–ZnO hybrid nanofluids under outdoor Iraqi climatic conditions in May 2025. Hybrid nanofluids with weight fractions of 0.10, 0.15, and 0.25 wt% were circulated through the cooling system at a constant flow rate of 1 L/min and compared with conventional water cooling. The system performance was evaluated in terms of photovoltaic cell temperature, electrical efficiency, thermal efficiency, power output, and overall efficiency. The results showed that the hybrid nanofluid improved the PVT performance compared with water cooling, with the best performance obtained at 0.25 wt%. At the peak operating condition, the 0.25 wt% hybrid nanofluid increased the electrical efficiency, thermal efficiency, power output, and overall efficiency by approximately 10%, 26.7%, 15.9%, and 20%, respectively, while reducing the photovoltaic cell temperature by up to 10.7%. In addition, machine-learning models were developed to predict the overall efficiency using the experimental dataset. The multilayer perceptron regressor (MLP) achieved the highest prediction accuracy, with R² = 0.9946, followed by the random forest regressor (RF) with R² = 0.9910, whereas the sequential neural network (SNN) showed lower accuracy with R² = 0.8825. These findings demonstrate that TiO2–ZnO hybrid nanofluids can effectively reduce photovoltaic module overheating and improve the thermal, electrical, and overall performance of PVT systems. The results also confirm that machine-learning models can serve as reliable tools for predicting PVT system behavior under real outdoor operating conditions.

Graphical abstract