Predicting thermal transport of blood-based penta-hybrid nanofluid in Fin geometries using deep neural networks and finite difference approach

Abstract

Nanofluids have garnered significant research interest due to their enhanced heat transfer and thermal characteristics. A novel hybrid nanofluid has exhibited exceptional thermal properties, combining five nanoparticles of uniform shapes with a base fluid, such as blood. This study investigates the influence of fin thickness, varying with length, considering the implications of internal heat production, convection, and thermal radiation processes in rectangular, convex, and triangular fin descriptions. Wet scenarios are interpreted to evaluate differences in thermal energy dynamics for fin shapes like Rectangular, Convex and Triangular. Darcy's model is employed to account for the material's porous nature. A finite difference scheme, implemented using Partial Differential Equation solver (PDSolve) in Maple (2024), provides graphical insights into fin effectiveness and thermal steady-state responses across various parameters. Incorporating Penta hybrid nanofluids enhances fin performance, with rectangular fins' Nusselt numbers (up to 1.936) proving more efficient, delivering faster thermal responses than triangular fins and convex fins. Further, using the Adam Optimisation algorithm, Convolutional Neural Networks were used to validate the current model. It was observed that these networks could accurately forecast the truth values, and the two findings matched, as indicated in Table 3 As a potential biological application, this research offers insight into optimising cooling systems for biomedical devices, such as heat exchangers in artificial organs.