Thermal and solutal transport in a magnetized binary micropolar flow over Riga plates with Soret and Dufour effects using artificial neural networks

Abstract

The hybrid nanoparticles dispersed in an ethylene glycol–water mixture are treated as a micropolar nanofluid. Analyzing the behavior of this mixture in squeezing channel configurations is essential for applications such as heat exchangers and cooling systems employed in the production of thin sheets from molten materials. This study aims to utilize copper (Cu) and aluminum (Al) nanoparticles to enhance the efficiency of heat transfer processes, either for heating or cooling purposes. The fluid motion and heat exchange between the plates are analyzed using the Navier–Stokes equations along with the first law of thermodynamics and Fick's law of diffusion, under the assumption that the base fluid exhibits micropolar characteristics. In addition to conventional viscous dissipation, the thermal dissipation arising from the micropolar effects is thoroughly examined. The governing equations are solved numerically using a fourth-order Runge–Kutta method combined with a shooting technique. The outcomes of the present analysis are validated by comparing them with previously published results, showing excellent agreement and confirming the accuracy of the current model. A feed forward neural network is developed and trained using the Levenberg–Marquardt back propagation algorithm to predict the outcomes corresponding to various parameter variations. The training process utilizes a reference dataset generated from numerical simulations obtained through the solver. The temperature profile increases with higher Hartmann number (Z), nanoparticle volume fraction (ϕ), Soret number (Sr), and Dufour number (Du), indicating that magnetic field strength and thermodiffusive effects enhance heat transfer performance.