MACHINE LEARNING COMPUTATIONS FOR MHD BIOCONVECTION PERISTALTIC TRANSPORT OF NON-NEWTONIAN NANOFLUID FLOW CONTAINING GYROTACTIC MICROORGANISMS WITH POROUS MEDIUM

This study introduces a new approach to machine learning-based numerical computing using a Levenberg-Marquardt algorithm-based multi-layer perceptron (MLP) feed-forward back-propagation with artificial neural networks (LMA-MLPFFBP) to model the electrically conducting bioconvection peristaltic movement of Reiner-Philippoff nanofluid (ECBPM-RPN) through a symmetric channel. This investigation considered the Reiner-Philippoff nanofluid and Buongiorno's nanoliquid models. The modeled flow situation includes factors such as heat generation, Brownian diffusion, magnetic fields, mixed convection, porous media, heat dissipation, thermophoresis diffusion, and gyrotactic microorganisms. The fluid-saturated porous medium is represented using a modified Darcy's law. Furthermore, slip-boundary conditions are applied to the channel walls. The governing equations of this problem are simplified using negligible Reynolds number and long wavelength approximations, and the resulting system is numerically solved employing the BVP4c algorithm based on the finite difference scheme in MATLAB. Furthermore, a dataset is generated through the numerical computation for the proposed LMA-MLPFFBP, considering fourteen scenarios for different profiles such as axial velocity, concentration of nanoparticles, density of motile microorganisms, and nanofluid's temperature to study the peristaltic motion of Reiner-Philippoff (R-Ph) nanofluid model by varying the pertinent flow parameters. The dataset is divided into three parts: 10% for training, 10% for testing, and 80% for validation. The reliability and efficacy of LMA-MLPFFBP are verified through the error histogram, performance, regression analysis, and fitness curves based on mean squared error (MSE), which vary from 10<sup>(-10)</sup> to 10<sup>(-8)</sup> ANNs-predicted results are further validated through tables and graphs for heat, density of motile microorganisms, mass transfer rates, axial velocity, concentration profile, and nanofluid’s temperature.

Перевод пока недоступен