Scholarlink

<p>This study's main goal is to highlight the critical role that machine learning (ML) and artificial intelligence (AI) approaches play in the analysis of fluid flow and the solving of challenging engineering issues. The incorporation of AI- and ML-based approaches into computational frameworks has substantially enhanced the efficiency, accuracy, and robustness of numerical predictions. The goal of the current work is to investigate magnetised Newtonian nanofluid boundary layer flow over a cylinder with viscous dissipation effects. The governing nanofluid issues with nonlinear PDEs are developed using Buongiorno's model. An Implicit Finite Difference Method (IFDM) is used to compute numerical solutions of governing non-similar PDEs. However, the predicted solution is examined by MLP-ANN. The efficiency of the proposed model is examined with MSE, correlation index <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(R\)</EquationSource> </InlineEquation> and optimal curve fitness function. An optimal performance of MLP-ANN is examined with the addition of MSE [1.53 <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\times 10^{ - 9}\)</EquationSource> </InlineEquation>, 1.19 <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\times 10^{ - 9}\)</EquationSource> </InlineEquation>, and 9.58 <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(\times 10^{ - 10}\)</EquationSource> </InlineEquation>] against epoch [1000, 992, and 482] for scenario 1 with cases 1–3. The output response <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(C_{fx} \sqrt {Re_{x} }\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(Nu_{x} /\sqrt {Re_{x} }\)</EquationSource> </InlineEquation>, and <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(Sh_{x} /\sqrt {Re_{x} }\)</EquationSource> </InlineEquation> of Newtonian nanofluid flow to varying physical quantities is carefully examined and illustrated through graphical and tabular results. One significant outcome is that an upsurge in the thermophoresis and Brownian motion parameter leads to a reduction in <InlineEquation ID="IEq8"> <EquationSource Format="TEX">\(Nu_{x} /\)</EquationSource> </InlineEquation> and <InlineEquation ID="IEq9"> <EquationSource Format="TEX">\(Sh_{x} /\sqrt {Re_{x} }\)</EquationSource> </InlineEquation>, however, an increment in the drag force coefficient.</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

Multi-objective Artificial Neural Network-Based Modelling of Mixed Convection Non-similar Analysis of Friction Drag Force Along with Heat and Mass Transfer Effect

  • 期刊论文

  • 发表日期：

    2025年11月10日

查看全文
  • Muhammad Shoaib,
  • Muhammad Imran Khan

摘要

This study's main goal is to highlight the critical role that machine learning (ML) and artificial intelligence (AI) approaches play in the analysis of fluid flow and the solving of challenging engineering issues. The incorporation of AI- and ML-based approaches into computational frameworks has substantially enhanced the efficiency, accuracy, and robustness of numerical predictions. The goal of the current work is to investigate magnetised Newtonian nanofluid boundary layer flow over a cylinder with viscous dissipation effects. The governing nanofluid issues with nonlinear PDEs are developed using Buongiorno's model. An Implicit Finite Difference Method (IFDM) is used to compute numerical solutions of governing non-similar PDEs. However, the predicted solution is examined by MLP-ANN. The efficiency of the proposed model is examined with MSE, correlation index \(R\) and optimal curve fitness function. An optimal performance of MLP-ANN is examined with the addition of MSE [1.53 \(\times 10^{ - 9}\) , 1.19 \(\times 10^{ - 9}\) , and 9.58 \(\times 10^{ - 10}\) ] against epoch [1000, 992, and 482] for scenario 1 with cases 1–3. The output response \(C_{fx} \sqrt {Re_{x} }\) , \(Nu_{x} /\sqrt {Re_{x} }\) , and \(Sh_{x} /\sqrt {Re_{x} }\) of Newtonian nanofluid flow to varying physical quantities is carefully examined and illustrated through graphical and tabular results. One significant outcome is that an upsurge in the thermophoresis and Brownian motion parameter leads to a reduction in \(Nu_{x} /\) and \(Sh_{x} /\sqrt {Re_{x} }\) , however, an increment in the drag force coefficient.