<p>This study investigates the optimization and predictive accuracy of photovoltaic thermal systems’ thermal efficiency using advanced artificial intelligence algorithms, specifically the artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), and relevance vector machine (RVM). Experimental data was collected from a photovoltaic thermal system at the Research Institute of Petroleum Industry in Tehran, Iran, with critical variables including solar irradiance, inlet temperature, wind speed, and ambient temperature. The comparative analysis revealed that the artificial neural network model outperformed other algorithms, achieving the highest predictive accuracy with a root mean square error (RMSE) of 11.704 and an <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41660_2025_513_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="21" /> </InlineMediaObject> <EquationSource Format="TEX">$R^2$</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mi>R</mi> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation> value of 0.959, emphasizing its strength in capturing complex, non-linear data interactions. The adaptive neuro-fuzzy inference system and relevance vector machine models demonstrated moderate predictive capabilities, with root mean square error values of 14.704 and 19.606, and <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41660_2025_513_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="21" /> </InlineMediaObject> <EquationSource Format="TEX">$R^2$</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mi>R</mi> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation> scores of 0.936 and 0.887, respectively. These results highlight the transformative potential of artificial intelligence-driven models, particularly the artificial neural network, in enhancing photovoltaic thermal system efficiency, thereby supporting global renewable energy goals through improved system adaptability and energy yield. This study advances renewable energy optimization, illustrating that artificial intelligence algorithms can effectively manage intricate variable relationships in photovoltaic thermal systems. The demonstrated approach sets a foundation for further research into artificial intelligence-optimized renewable energy solutions, promoting more efficient and resilient infrastructures essential for sustainable development and climate action.</p>

Advanced Optimization Techniques Using Artificial Intelligence Algorithms for Thermal Efficiency Estimation of Photovoltaic Thermal Systems

期刊论文
发表日期：

2025年4月21日

查看全文

Hassan A. Hameed Al-Hamzawi,
Ali S. Abed Al Sailawi,
Qudama Al-Yasiri,
Mohammed Alktranee

摘要

This study investigates the optimization and predictive accuracy of photovoltaic thermal systems’ thermal efficiency using advanced artificial intelligence algorithms, specifically the artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), and relevance vector machine (RVM). Experimental data was collected from a photovoltaic thermal system at the Research Institute of Petroleum Industry in Tehran, Iran, with critical variables including solar irradiance, inlet temperature, wind speed, and ambient temperature. The comparative analysis revealed that the artificial neural network model outperformed other algorithms, achieving the highest predictive accuracy with a root mean square error (RMSE) of 11.704 and an $R^2$ $R^{2}$ value of 0.959, emphasizing its strength in capturing complex, non-linear data interactions. The adaptive neuro-fuzzy inference system and relevance vector machine models demonstrated moderate predictive capabilities, with root mean square error values of 14.704 and 19.606, and $R^2$ $R^{2}$ scores of 0.936 and 0.887, respectively. These results highlight the transformative potential of artificial intelligence-driven models, particularly the artificial neural network, in enhancing photovoltaic thermal system efficiency, thereby supporting global renewable energy goals through improved system adaptability and energy yield. This study advances renewable energy optimization, illustrating that artificial intelligence algorithms can effectively manage intricate variable relationships in photovoltaic thermal systems. The demonstrated approach sets a foundation for further research into artificial intelligence-optimized renewable energy solutions, promoting more efficient and resilient infrastructures essential for sustainable development and climate action.