<p>The efficient thermal management of electronic devices remains barrier in the advancement of global digitalization. A major challenge to global digitalization is managing the cooling and heat control of electronic devices. Recently, scientific breakthroughs have been the result of creative and inventive research using machine learning technique in order to reach the needs of society. The integration of machine learning in research has catalyzed novel scientific discoveries pertinent to current challenges. The study investigates nanofluid flow within micro-channels, incorporating the Buongiorno model to assess the impacts of microorganisms, nanoparticle concentrations, and waste discharge. Employing similarity transformations, the study converts partial differential equations into dimensionless ordinary differential equations, subsequently addressed via the fourth-order Runge–Kutta technique. Furthermore, a novel application of artificial neural networks is presented to predict the thermal and solutal transport phenomena in micro-channel utilizing a Levenberg–Marquardt-based neural network. The main findings of this investigation demonstrate that the thermal transfer rate model achieves its validation performance at epochs 882, exhibiting an incredibly best MSE of <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41939_2025_911_Article_IEq1.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="98" /> </InlineMediaObject> <EquationSource Format="TEX">\(2.3929 \times 10^{ - 8}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>2.3929</mn> <mo>×</mo> <msup> <mn>10</mn> <mrow> <mo>-</mo> <mn>8</mn> </mrow> </msup> </mrow> </math></EquationSource> </InlineEquation>, while the mass transfer rate model reaches comparable MSE of <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41939_2025_911_Article_IEq2.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="98" /> </InlineMediaObject> <EquationSource Format="TEX">\(5.0043 \times 10^{ - 5}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>5.0043</mn> <mo>×</mo> <msup> <mn>10</mn> <mrow> <mo>-</mo> <mn>5</mn> </mrow> </msup> </mrow> </math></EquationSource> </InlineEquation> at epochs 140, demonstrating the model’s superior architectural design. Additionally, an inverse relationship is observed between solutal transfer rates and external pollutant concentrations.</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

Neural network-based analysis of thermal and mass transfer in nanofluid flow through a microchannel under pollutant concentration effect

  • Ram Prakash Sharma,
  • Sriram Praharaj,
  • Bimal Kumar Barik,
  • V. Vinay Kumar,
  • Abhishek Sharma,
  • Chandralekha Mahanta

摘要

The efficient thermal management of electronic devices remains barrier in the advancement of global digitalization. A major challenge to global digitalization is managing the cooling and heat control of electronic devices. Recently, scientific breakthroughs have been the result of creative and inventive research using machine learning technique in order to reach the needs of society. The integration of machine learning in research has catalyzed novel scientific discoveries pertinent to current challenges. The study investigates nanofluid flow within micro-channels, incorporating the Buongiorno model to assess the impacts of microorganisms, nanoparticle concentrations, and waste discharge. Employing similarity transformations, the study converts partial differential equations into dimensionless ordinary differential equations, subsequently addressed via the fourth-order Runge–Kutta technique. Furthermore, a novel application of artificial neural networks is presented to predict the thermal and solutal transport phenomena in micro-channel utilizing a Levenberg–Marquardt-based neural network. The main findings of this investigation demonstrate that the thermal transfer rate model achieves its validation performance at epochs 882, exhibiting an incredibly best MSE of \(2.3929 \times 10^{ - 8}\) 2.3929 × 10 - 8 , while the mass transfer rate model reaches comparable MSE of \(5.0043 \times 10^{ - 5}\) 5.0043 × 10 - 5 at epochs 140, demonstrating the model’s superior architectural design. Additionally, an inverse relationship is observed between solutal transfer rates and external pollutant concentrations.