Scholarlink

<p>The present study investigates random forest regression analysis and entropy generation in an unsteady Williamson tri-hybrid nanofluid with Darcy–Forchheimer, chemical reaction and bioconvection over a stretchable rotating disk. Convective boundary conditions are imposed on the thermal, concentration and microorganism field. Gold (Au), titanium dioxide (TiO<sub>2</sub>) and silicon dioxide (SiO<sub>2</sub>) nanoparticles are incorporated into the base fluid blood, as this combination offers a promising multifunctional platform for cancer treatment, controlled bio-heat transfer and diagnostic imaging. The governing partial differential equations are transformed into ordinary differential equations by similarity transformations and solved by the fifth-order Runge–Kutta–Fehlberg shooting technique. The increments of the Williamson parameter reduce the velocity profile in both the radial and azimuthal directions for steady and unsteady flow cases. An increment in the porous parameter enhances the Bejan number. To evaluate the reliability of numerical solutions, a random forest machine learning model is used with 300 numerical datasets which accurately predicts physical quantities in close agreement with the numerical values. For the radial skin-friction coefficient, the random forest yields (<i>R</i><sup>2</sup>) coefficients of determination of 0.990206 and 0.988812 for the training and testing datasets, respectively, with an overall mean absolute percentage error of 0.62%. The training and testing root mean square rrror (RMSE) of azimuthal direction skin friction is 0.020583 and 0.019681 and overall mean absolute percentage error is 0.52%.</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

Random forest machine learning model and entropy generation in unsteady Williamson tri-hybrid (Au + TiO2 + SiO2)/blood nanofluid with Darcy–Forchheimer porous rotating disk

  • 期刊论文

  • 发表日期：

    2026年8月1日

查看全文
  • P. Soniya,
  • Poulomi De

摘要

The present study investigates random forest regression analysis and entropy generation in an unsteady Williamson tri-hybrid nanofluid with Darcy–Forchheimer, chemical reaction and bioconvection over a stretchable rotating disk. Convective boundary conditions are imposed on the thermal, concentration and microorganism field. Gold (Au), titanium dioxide (TiO2) and silicon dioxide (SiO2) nanoparticles are incorporated into the base fluid blood, as this combination offers a promising multifunctional platform for cancer treatment, controlled bio-heat transfer and diagnostic imaging. The governing partial differential equations are transformed into ordinary differential equations by similarity transformations and solved by the fifth-order Runge–Kutta–Fehlberg shooting technique. The increments of the Williamson parameter reduce the velocity profile in both the radial and azimuthal directions for steady and unsteady flow cases. An increment in the porous parameter enhances the Bejan number. To evaluate the reliability of numerical solutions, a random forest machine learning model is used with 300 numerical datasets which accurately predicts physical quantities in close agreement with the numerical values. For the radial skin-friction coefficient, the random forest yields (R2) coefficients of determination of 0.990206 and 0.988812 for the training and testing datasets, respectively, with an overall mean absolute percentage error of 0.62%. The training and testing root mean square rrror (RMSE) of azimuthal direction skin friction is 0.020583 and 0.019681 and overall mean absolute percentage error is 0.52%.