Analysis for 3D thermal conducting micropolar nanofluid via artificial neural network

Abstract

This paper considers the Darcy–Forchheimer flow over a micropolar nanofluid by using an intelligent backpropagated neural network with Levenberg–Marquardt scheme. The PDEs governing the DFF-MNFM are reduced into ODEs through some appropriate transformations. A reference dataset is prepared from HAM by changing several key parameters, such as the porosity parameter (γ), Reynolds number (Re), coupling parameter (N₁), rotation parameter (Kr), coefficient of inertia (Fr), viscosity gradient parameter (N₂), and Brownian motion parameter (Nb), for all proposed IBNN-LMS scenarios. The estimated solutions of the IBNN-LMS are analyzed and compared with reference results. The results suggest that for high values of the Reynolds number, Re, the fluid velocity is increased at the surface, and with Kr, increasing velocity on the surface of the fluid increases but decreases beyond the surface. A rise in the value of γ enhances velocity closer to the surface while diminishing the velocity beyond the surface distance. The rise of N₁ enhances the speed of the microrotation of fluid closer to the surface. In addition, the higher temperature and concentration profiles enhance the value of Nb. For the validation of IBNN-LMS approach, its efficiency is justified through convergence analysis of MSE, regression indices, and error spectrum evaluations that represent its robustness in solving complicated fluid flow problems.