ADVANCED NEURAL COMPUTING FRAMEWORK FOR PREDICTING FLOW ATTRIBUTES OF NONSTANDARD NANOFLUIDS INCORPORATING MULTIPLE NOVEL PHYSICAL PARAMETERS

Premier consent to commence present artifact is to investigate the physical characteristics of magnetized nanofluid flow past a stretchable surface, along with multiple physical aspects. For this purpose, viscous fluid (water) is considered as a base fluid and copper nanoparticles are induced in it to form a nanofluid. For comprehensive examination of the current study, viscous dissipation, Joule heating, and slip conditions are also included. Mathematical configuration of the problem is attained in the form of dimensional partial differential equations after utilization of associated conservation laws. Furthermore, the modeled coupled differential system is converted into dimensionless ordinary differential equations by operating the appropriate non-similarity procedure. The solution to this problem is attained through numerical and machine-learning-based approaches. Numerically, the RK-4 method in conjunction with the shooting method is utilized to resolve the governing system of equations, whereas the Levenberg-Marquardt strategy is capitalized to predict behavior of data obtained from the numerical scheme. The results are portrayed in graphical and tabular manner for the velocity and temperature distributions by accounting for a wide range of sundry parameters in a comparative manner for the situation when nanoparticles are either present or absent. Quantities of physical interest, such as the skin friction coefficient and average Nusselt number, are also enumerated against the flow controlling parameters. The Nusselt number is found to be increased up to 3% when nanoparticles are present, in contrast to the condition when nanoparticles are absent. An increase up to 4.3% is observed in the heat transportation rate due to consideration of the magnetic field factor and 2.4% elevation due to permeability aspects. Conversely, the slip parameters tend to cause a reduction in the Nusselt number up to 5% instead of the situation when no-slip boundary constraint is applied.