Scholarlink

<p>This study analyzes unsteady heat transfer and boundary layer flow of nanofluids in a rotating double-disk system under nonlinear thermal radiation and magnetic effects. Two nanofluid cases are considered: Molybdenum Disulfide (MoS₂)–water and Multi-Walled Carbon Nanotube (MWCNT)–water. The governing nonlinear equations are transformed using similarity variables and solved numerically with the finite element method. Parametric effects of the Forchheimer number (Fr), magnetic parameter (M), and nanoparticle volume fraction (ϕ) on velocity and temperature distributions are investigated. Results show that increasing Fr enhances velocity but lowers temperature, while higher M suppresses velocity and increases temperature through Joule heating. MWCNT–water nanofluid demonstrates superior velocity performance, whereas MoS₂–water shows stronger thermal response. An artificial neural network (ANN) with three hidden layers was also developed to predict the skin-friction coefficient. The Levenberg–Marquardt optimizer achieved higher accuracy and faster convergence than LBFGS, enabling efficient prediction without repeated simulations. The combined FEM–ANN framework offers insights for designing nanofluid-based thermal management systems in applications such as electronic cooling, solar energy, and biomedical devices.</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-aided simulation of MHD nanofluid flow in double-disk systems with nonlinear thermal radiation

  • 期刊论文

  • 发表日期：

    2025年10月26日

查看全文
  • Nagaraju Vuppala,
  • Raja Shekar Mamidi

摘要

This study analyzes unsteady heat transfer and boundary layer flow of nanofluids in a rotating double-disk system under nonlinear thermal radiation and magnetic effects. Two nanofluid cases are considered: Molybdenum Disulfide (MoS₂)–water and Multi-Walled Carbon Nanotube (MWCNT)–water. The governing nonlinear equations are transformed using similarity variables and solved numerically with the finite element method. Parametric effects of the Forchheimer number (Fr), magnetic parameter (M), and nanoparticle volume fraction (ϕ) on velocity and temperature distributions are investigated. Results show that increasing Fr enhances velocity but lowers temperature, while higher M suppresses velocity and increases temperature through Joule heating. MWCNT–water nanofluid demonstrates superior velocity performance, whereas MoS₂–water shows stronger thermal response. An artificial neural network (ANN) with three hidden layers was also developed to predict the skin-friction coefficient. The Levenberg–Marquardt optimizer achieved higher accuracy and faster convergence than LBFGS, enabling efficient prediction without repeated simulations. The combined FEM–ANN framework offers insights for designing nanofluid-based thermal management systems in applications such as electronic cooling, solar energy, and biomedical devices.