ANN-based analysis of heat transfer in blood-based hybrid nanofluid flow through a porous deforming channel.

This paper is devoted to the investigation of thermal transport features of an Ag-TiO2 hybrid nanofluid based on blood, a non-deforming porous channel which is being heated up in different ways, and with the aim to find out its possible application in the field of biomedical thermal management systems. The study deals with the behavior of flow and heat of a blood-based hybrid nanofluid consisting of silver (Ag) and titanium dioxide (TiO2) nanoparticles in a rectangular channel with porous walls taking into account generalized thermal flux model effects. The problem is then regarded as a 2D incompressible laminar isothermal flow. The resulting PDEs are transformed into nonlinear ODEs via similarity transformations and solved numerically in the Mathematica software using NDSolve in order to create reference data. The influence of Reynolds number (R), wall deformation rate (α), and porosity parameter (S) on velocity and temperature profiles is investigated. A Bayesian regularized artificial neural network (ANN-BRT) is built to predict both velocity and temperature profiles with the use of numerically generated reference data, where 70% of the data is used for training and 15% for testing and validation. The model demonstrates excellent performance, with absolute errors shrinking from 10-04 to 10-09 and MSE values reaching the extremely low figure of 1.25 × 10-12. The metrics for performance validate the presence of very good convergence and reliability. It was found that the velocity is positively affected by higher values of porosity and Reynolds number, particularly in the case of wall deformation, while the temperature distribution behaves the opposite way as it is affected by wall injection and suction under higher Reynolds number and porosity conditions. This study opens up a variety of medical applications including drug delivery, cancer therapy, wound healing, and thermal management systems.