Application of Neural Network Analysis to Study Thermally Stratified  Hybrid Nanofluid Over a Variable Thickness Sheet

Enhancements of industrial thermal system energy efficiency will involve high quality working fluids with high heat transfer properties. This work examines inclined magnetohydrodynamic (MHD) flow of hybrid nanofluid particles (TiO$_2$ and Fe$_3$O$_4$/H$_2$O) on a surface whose thickness is varying. The mathematical formulation uses the effects of internal heat generation, thermal radiation under a magnetic field together with convective boundary conditions. We transformed the equations to a coupled ordinary differential equations(ODEs) on using adequate similarity transformations. The characteristics of solutions are analyzed using a machine learning model based on neural network (NN) approaches that is trained using the Levenberg-Marquardt (LM) optimization scheme . Validation of the correctness of the proposed approach is ensured by comparing the NN predictions and the computational results. In parametric analysis, the magnitude of the Prandtl number ($Pr$) and thermal stratification parameter ($St$) increases and results in a fall in temperature of the hybrid nanofluid system. Velocity profiles decrease when the magnetic fields are stronger and velocity slip conditions are increased. In addition, the analysis quantified the effect of physical parameters on the surface heat transfer rate and the skin friction coefficient.