Computational modelling and thermophysical characterization of Kelvin–Voigt fluid flow: A neural network approach

Abstract This research provides insight into the unsteady, fully developed natural convective flow and heat transfer of conducting viscoelastic Kelvin–Voigt zero order fluid in a vertical annulus with variable viscosity, induced and radial magnetic field. Quadratically and exponentially varying viscosity impacts on the flow field, magnetic and thermal fields are also analyzed. The nonlinear coupled governing PDEs are solved using an implicit Crank–Nicolson finite difference method and the Thomas algorithm to turn up the nonsimilar solutions. The effect of pertinent physical parameters on flow, magnetic, and thermal fields are discussed through graphs. In the course of numerical analysis, it is noted that enhancing the variations in viscosity parameters enhances skin friction and flow rate. The fluid velocity and local heat transport increase with buoyancy force. The artificial neural network model with the backpropagating Levenberg–Marquardt algorithm is adopted to analyze the heat transfer characteristics in quadratically and exponentially varying viscosity cases.