Scholarlink

<p>The lower thermal behavior of solar-based thermal systems limits the contribution of solar systems to meet current energy demand of industries. The Flat Plate Solar Air Heater (FPSAH) is extensively utilized in many applications requiring reasonable heat but struggles from inherent limitation in convective heat release and mediocre efficiency. In this study, the challenges are addressed with a novel means of dual mode augmented technique. This mode integrated absorber surface of the FPSAH system by introducing two radiation reflectors on either edge of the rectangular channel. Primarily these reflectors forward back the solar irradiance over the absorber plate, thus increasing the actual solar flux. Simultaneously, a W-shaped artificial rib roughness pattern is merged on the underneath (air-side) of the absorber plate. This coarseness is intended to persuade measured flow disorder inside the channel that disrupt the boundary layer development and may consequently augment convective heat transfer. Experimental testing is conducted with different combinations of roughened absorber surface and radiation reflectors. The performance enhancement is evaluated in terms of Nusselt number (<i>Nu</i>) and thermal efficiency of the FPSAH system. The maximum <i>Nu</i> achieved is 1.63 times higher using a set of radiation reflectors along with W-shaped roughness on the absorber surface compared to the plain configuration without radiation reflector. Finally, artificial neural network (ANN) and machine learning (ML) algorithms were used to predict Reynolds number in each set of experiments. A very good curve fitting was achieved by the Robust Regression algorithm with <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(R^2 = 0.99\)</EquationSource> </InlineEquation> for the testing dataset and <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(R^2 = 0.94\)</EquationSource> </InlineEquation> by the Random Forest Regression algorithm for ML.</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

Thermal analysis of flat plate solar air heater system with radiation reflectors and W-shaped roughness: artificial neural network & machine learning approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月2日

查看全文
  • Piyush Kumar Jain,
  • Kawal Lal Kurrey,
  • Vikas Pandey,
  • Jitesh R. Shinde,
  • Abhishek Narayan Tripathi,
  • Niraj Kumar Dewangan

摘要

The lower thermal behavior of solar-based thermal systems limits the contribution of solar systems to meet current energy demand of industries. The Flat Plate Solar Air Heater (FPSAH) is extensively utilized in many applications requiring reasonable heat but struggles from inherent limitation in convective heat release and mediocre efficiency. In this study, the challenges are addressed with a novel means of dual mode augmented technique. This mode integrated absorber surface of the FPSAH system by introducing two radiation reflectors on either edge of the rectangular channel. Primarily these reflectors forward back the solar irradiance over the absorber plate, thus increasing the actual solar flux. Simultaneously, a W-shaped artificial rib roughness pattern is merged on the underneath (air-side) of the absorber plate. This coarseness is intended to persuade measured flow disorder inside the channel that disrupt the boundary layer development and may consequently augment convective heat transfer. Experimental testing is conducted with different combinations of roughened absorber surface and radiation reflectors. The performance enhancement is evaluated in terms of Nusselt number (Nu) and thermal efficiency of the FPSAH system. The maximum Nu achieved is 1.63 times higher using a set of radiation reflectors along with W-shaped roughness on the absorber surface compared to the plain configuration without radiation reflector. Finally, artificial neural network (ANN) and machine learning (ML) algorithms were used to predict Reynolds number in each set of experiments. A very good curve fitting was achieved by the Robust Regression algorithm with \(R^2 = 0.99\) for the testing dataset and \(R^2 = 0.94\) by the Random Forest Regression algorithm for ML.