Deep learning investigation of water-based tetra hybrid nanofluid across a shrinking cylinder for variable electrical conductivity with thermal radiation

Aims and objective: This study investigates the impact of varying electrical conductivity and viscosity on flow using a unique design of intelligent Bayesian regularization neural networks (IBRNN). The study focuses on several significant physical aspects of a tetra hybrid nanomaterial model consisting of silver (Ag), magnesium oxide (MgO), titanium dioxide (TiO2) and zirconium oxide (ZrO2) with water as the base fluid. Investigating the effects of mass suction, variable electrical conductivity, variable viscosity, Magnetohydrodynamic, Joule heating, thermal radiation, Smoluchowski temperature, and Maxwell Velocity slip conditions on a combination of tetra hybrid nanoparticles is the primary goal of this work. Design/methodology: The study approach entails converting fundamental equations into a dimensionless form by the use of a similarity transformation technique and implementing the bvp4c numerical computing method in MATLAB. We train and test the ANN-IBR approach exploitation reference datasets derived from numerical calculations to estimate flow solutions under diverse physical parameter situations. The ANN-IBR model effectively moves water and tetra hybrid nanoparticles around a cylinder, according to the research. Mean square error analysis, transition state analysis, histogram analysis, and regression analysis show its accuracy compared to reference data. Findings: The temperature profile exhibits dual behavior in response to the magnetic parameter and the changing electrical conductivity parameter. As the radiation parameter rises, the temperature profile and the Nusselt number both grow in magnitude. Mass suction enhances the velocity profile, whereas the viscosity parameter diminishes the velocity profile of tetra hybrid nanofluid. There is a 0.6% improvement in heat transfer rate for tetra hybrid nanofluid with Rd = 0.8 and a 0.82% improvement for S = 2.0 when compared to nanofluid. This work influences the development of thermal protection systems for aerospace applications, where materials reach high temperatures and thermal load management depends on their conductive properties.