Response surface methodology (RSM) and artificial neural network (ANN) simulations for thermal flow hybrid nanofluid flow with Darcy-Forchheimer effects

Abstract

The goal of this study is to examine the radiative properties of a Cu−Al2O3/water hybrid nanofluid with respect to joule heating and suction effects over a stretched sheet. The copper and aluminium oxide nanoparticles are used to increase the thermal impact of water base fluid. The Darcy-Forchheimer theory is used to endorse the inertial and porous media effects and makes the model more interesting from the physical scenario. The partial differential equations (PDEs) are converted to nonlinear ordinary differential equations (ODEs) through similarity transformation and then numerically solved using the shooting technique with the assistance of a built-in function in Maple software. The impacts of relevant parameters on temperature and velocity profiles, as well as local Nusselt number and skin friction, are discussed. A derivative-based sensitivity analysis is performed to determine different values of physical parameters that affect response functions (skin friction coefficient, Nusselt number) under a given set of assumptions. The effect of the suction, magnetic, and permeability parameters on the solutions is observed to be significant. On the decreasing surface, an increase in copper nanoparticle volume fractions induces an improvement in local skin friction and a decrease in local Nusselt number. Moreover, in order to predict the skin friction and local Nusselt number, we employed an artificial neural network (ANN). This study revealed that the ANN method can predict skin friction and local Nusselt number (R2=0.98 and R2=0.99 respectively). As a result, this study is useful for engineers and scientists who want to learn about the features of the flow, its behavior, and how to anticipate it.