Artificial intelligence modelling of cobalt chromite (CoCr 2 O 4 ) nanofluid flow over a curved surface with melting phenomena and nonlinear thermal radiation

This research looks at the incompressible flow of a cobalt chromite nanoparticle mixed with water and ethylene glycol in a 50:50 ratio, passing through a Darcy–Forchheimer porous medium with a curved stretching surface embedded in it. Included in the model are four typically handled separately physical effects: surface curvature, nonlinear thermal radiation, melting heat transfer at the wall and heat source/sink. By applying appropriate similarity transformations, the governing partial DEs are reduced to a coupled system of nonlinear ordinary DEs. These are then resolved using the MATLAB bvp5c solvers. The numerical data provided by the bvp5c solver is utilized to train the Levenberg–Marquardt Backpropagation (LMBP) artificial neural network. For a large variety of control settings, the local Nusselt number, skin friction coefficient, temperature field and velocity field are derived. It is found that the velocity increases with the curvature parameter but is reduced by porosity parameter. Furthermore, the heat generation parameter improves the temperature profile. Applying these findings to the design of porous-media energy devices, curved heat exchangers, polymer extrusion and ice/wax melting on heated surfaces are all viable thermal system options.