Scholarlink

<p>This study investigates the two-dimensional viscous, steady, incompressible magnetohydrodynamic (MHD) flow of a nanofluid over a linearly stretching sheet, taking into account buoyancy forces, viscous dissipation, internal heat generation, chemical reactions, activation energy, suction and convective boundary conditions. Such a comprehensive model is crucial for applications in biomedical heat transfer, nuclear reactors, solar collectors, polymer processing and aerospace engineering. The governing nonlinear partial differential equations are reduced using similarity transformations and solved numerically by the spectral quasi-linearisation method (SQLM). A feed-forward backpropagation multilayer perceptron (MLP) artificial neural network (ANN), trained using the Levenberg–Marquardt algorithm, is developed for efficient prediction of the Nusselt number. The results reveal a significant sensitivity of flow, temperature and concentration profiles to thermophysical parameters, including thermophoresis, Brownian motion, Biot numbers and activation energy. We observed that velocity augments with rising thermophoretic parameter and Eckert number, while it declines with increasing Brownian motion parameter and suction parameter. The temperature of the fluid increases with rising Brownian motion parameter, Eckert number and thermal Biot number, whereas it declines with increasing suction parameter. Concentration increases with rising thermal and solute Biot numbers, whereas it decreases with increasing activation energy. Due to the enhancement of the thermophoretic parameter from 0.2 to 1, the Sherwood number and skin friction coefficient increase by 345.47 and 7.22%, respectively, while the Nusselt number decreases by 4.75%. For the rise in thermal Biot number from 100 to 120, the Nusselt number and Sherwood number increase by 22.29 and 25.67%, respectively, whereas the skin friction coefficient decreases by 0.10%. Skin friction coefficient, Nusselt number and Sherwood number increase by 0.10, 0.23 and 0.44%, respectively with the rise in activation energy from 0.1 to 2. This combined modelling and ANN-based prediction approach is novel and offers accurate, computation-free thermal analysis.</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

Prediction analysis of heat transfer with convective boundary layer through an artificial neural network

  • 期刊论文

  • 发表日期：

    2026年4月7日

查看全文
  • Shovan Sarkar,
  • Hiranmoy Mondal,
  • Prabir Kumar Kundu

摘要

This study investigates the two-dimensional viscous, steady, incompressible magnetohydrodynamic (MHD) flow of a nanofluid over a linearly stretching sheet, taking into account buoyancy forces, viscous dissipation, internal heat generation, chemical reactions, activation energy, suction and convective boundary conditions. Such a comprehensive model is crucial for applications in biomedical heat transfer, nuclear reactors, solar collectors, polymer processing and aerospace engineering. The governing nonlinear partial differential equations are reduced using similarity transformations and solved numerically by the spectral quasi-linearisation method (SQLM). A feed-forward backpropagation multilayer perceptron (MLP) artificial neural network (ANN), trained using the Levenberg–Marquardt algorithm, is developed for efficient prediction of the Nusselt number. The results reveal a significant sensitivity of flow, temperature and concentration profiles to thermophysical parameters, including thermophoresis, Brownian motion, Biot numbers and activation energy. We observed that velocity augments with rising thermophoretic parameter and Eckert number, while it declines with increasing Brownian motion parameter and suction parameter. The temperature of the fluid increases with rising Brownian motion parameter, Eckert number and thermal Biot number, whereas it declines with increasing suction parameter. Concentration increases with rising thermal and solute Biot numbers, whereas it decreases with increasing activation energy. Due to the enhancement of the thermophoretic parameter from 0.2 to 1, the Sherwood number and skin friction coefficient increase by 345.47 and 7.22%, respectively, while the Nusselt number decreases by 4.75%. For the rise in thermal Biot number from 100 to 120, the Nusselt number and Sherwood number increase by 22.29 and 25.67%, respectively, whereas the skin friction coefficient decreases by 0.10%. Skin friction coefficient, Nusselt number and Sherwood number increase by 0.10, 0.23 and 0.44%, respectively with the rise in activation energy from 0.1 to 2. This combined modelling and ANN-based prediction approach is novel and offers accurate, computation-free thermal analysis.