Solution predictive Bayesian networks knacks for entropy optimized second-order slip velocity free convective nanofluid flow along Darcy-Forchheimer porous medium

Abstract

The aim of this study is to present incompressible, steady entropy optimized second-order slip velocity free convective nanofluid flow along Darcy–Forchheimer porous medium through neural network backpropagation with Bayesian Regularization Approach (NNBPBRA) under the influence of heat sink/source and viscous dissipation. The viscous dissipation effects and radiative heat flux are also employed in the presented fluid flow system. Nanomaterials are silicon dioxide and molybdenum disulfide, using water as the base fluid. To obtain the reference dataset for NNBPBRA, the ODEs which are obtained after simplifications of original fluid flow system in terms of PDEs with the appropriate application of transformation mechanism are solved numerically using state of the art numerical method. The outcome in terms of solution and error analyses plots for the modification of different parameters are interpreted using these reference datasets. The error histograms, regression analyses, and MSE indices are used to validate the performance of NNBP-BRA. In the presence of slip parameters and stratification variables, the skin friction coefficient, i.e. surface drag force, and heat transfer rate, i.e. Nusselt number, are also evaluated. The outcomes show that as the Brinkman number increases, the rate of entropy formation increases, but the temperature falls as the stratification parameter increases.