摘要
This article emphasises finding solutions for fluid flow and heat transfer-related problems through the Levenberg-Marquardt back-propagation technique.The solutions are developed for a three-layered channel with the porous medium in the middle layer.The main motive of the numerical experiment is to investigate the parametric effects on the Cu-Al_(2)O_(3) hybrid nanofluid in the central layer,Cu nanofluid in the left layer and Al_(2)O_(3) nanofluid in the right layer.The training and testing data for generating the solution are sought through shooting technique.Levenberg-Marquardt back-propagation solutions show that the error for the training data is very close to zero.The computational domain is extended using a machine learning approach for various parametric values with zero Jacobian error.Results show that the slippery nature of the left wall has a noticeable effect in the hybrid nanofluid channel compared to the other layers.Also observed that the porosity decreases the velocity as the solid space dominates the fluid space and thus has a strong opposing force,reducing its velocity.
