.png){kind=icon}
.png){kind=icon}
Return
Application of machine learning for magnetohydrodynamic peristaltic motion of ternary hybrid nanofluid through a curved channel
.png){kind=icon}
PRE
AI
J
F
M
DOI：10.1080/02286203.2026.2685214.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This study explores the thermal analysis of magnetohydrodynamic (MHD) peristaltic flow of ternary hybrid nanofluids (MHD-PFTHNFs) through an asymmetric curved channel. The primary objective is to analyze the heat transfer characteristics of ternary hybrid nanofluid CuO+MgO+GO/H2O,THNFs while considering the effects of Joule heating, porous media, heat sink/source, Hall current, an applied magnetic field, and thermal radiation. Furthermore, velocity and thermal slip boundary conditions are taken into account. The governing equations are addressed numerically by employing the NDSolve scheme through Mathematica software after applying lubrication theory. Additionally, a novel intelligent computing approach is implemented, employing a Multi-Layer Perception Feed-Forward Backpropagation Artificial Neural Network (MLP-FFBANN) combined with the Levenberg-Marquardt Algorithm (LMA) to enhance the solution efficiency. The numerical technique is used to create a highly accurate dataset for MLP-FFBANN, facilitating the examination of the temperature profile, velocity distribution, and heat transfer at the boundary. The proposed MLP-FFBANN algorithm, developed using Artificial Intelligence (AI), is designed to solve the MHD-PFTHNFs model. The algorithm achieves optimal performance, with Mean Squared Error (MSE) values of 1.4629E−08, 9.9701E−08, 1.0589E−08, 4.1987E−08, 8.2668E−08, 5.4892E−04, 5.549E−08, and 2.4661E−09, corresponding to 811, 363, 100, 504, 100, 38, 646, and 1000 epochs, respectively. The smallest magnitudes of the error distributions on histograms, near-optimal regression metrics, and low MSE values suggest that the technique is accurate and precise. The outcome reveals that heat transfer improves for greater values of Hartmann number, curvature parameter, and heat generation parameter. Furthermore, the maximum velocity of THNFs shifts from the upper wall to the lower wall of the channel, revealing that the upper wall exhibits a higher fluid velocity compared to the lower wall as the curvature parameter increases.
Keywords:
Peristalsis
multi-layer perception
machine learning
Levenberg-Marquardt backpropagation
artificial neural network
Journal
I
IF:
3.9
Papers:
596
Citations:
1.5K
Organization
Cited Papers