Scholarlink

<p>A machine-learning-assisted numerical study is conducted on inclined magnetohydrodynamic (MHD) micropolar bioconvective flow over a curved permeable stretching surface, incorporating non-Darcy porous resistance. The mathematical model combines thermal radiation, chemical reactions, mass diffusion, and motile microorganisms to represent coupled bio-magneto-thermal transport phenomena. Similarity transformations are used to reduce the governing nonlinear boundary-layer equations to a system of ordinary differential equations, which are solved numerically using MATLAB’s bvp4c solver. The resulting numerical dataset is used to train a feedforward Artificial Neural Network (ANN) with the Levenberg-Marquardt algorithm to predict key engineering parameters. Results show that increasing the magnetic parameter reduces velocity by approximately 9%, whereas non-Darcy porous resistance reduces fluid motion by 18%. Thermal radiation increases the temperature field by approximately 14%, whereas a higher Schmidt number reduces the concentration by nearly 73%. Similarly, higher Peclet numbers reduce the number of microorganisms by approximately 71%, illustrating the importance of advective transport. The ANN model shows high predictive accuracy, evidenced by a coefficient of determination close to 1 and low RMSE and MAE values. This combined numerical and machine-learning method provides an accurate and computationally efficient approach to studying complex inclined MHD micropolar bioconvective systems, which are important for advanced thermal and bioengineering applications.</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

Machine-learning-based predictive assessment of inclined MHD micropolar bioconvection over curved porous stretching sheet

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月13日

查看全文
  • Bandaru Gangadri,
  • P. Durgaprasad

摘要

A machine-learning-assisted numerical study is conducted on inclined magnetohydrodynamic (MHD) micropolar bioconvective flow over a curved permeable stretching surface, incorporating non-Darcy porous resistance. The mathematical model combines thermal radiation, chemical reactions, mass diffusion, and motile microorganisms to represent coupled bio-magneto-thermal transport phenomena. Similarity transformations are used to reduce the governing nonlinear boundary-layer equations to a system of ordinary differential equations, which are solved numerically using MATLAB’s bvp4c solver. The resulting numerical dataset is used to train a feedforward Artificial Neural Network (ANN) with the Levenberg-Marquardt algorithm to predict key engineering parameters. Results show that increasing the magnetic parameter reduces velocity by approximately 9%, whereas non-Darcy porous resistance reduces fluid motion by 18%. Thermal radiation increases the temperature field by approximately 14%, whereas a higher Schmidt number reduces the concentration by nearly 73%. Similarly, higher Peclet numbers reduce the number of microorganisms by approximately 71%, illustrating the importance of advective transport. The ANN model shows high predictive accuracy, evidenced by a coefficient of determination close to 1 and low RMSE and MAE values. This combined numerical and machine-learning method provides an accurate and computationally efficient approach to studying complex inclined MHD micropolar bioconvective systems, which are important for advanced thermal and bioengineering applications.