Potential of physics-informed neural networks for analysing chemically reactive fluid flow past a deformable rotating cone due to generated and absorbed heat propagated by Arrhenius kinetics

The investigation of thermofluid transport in deformable and rotating geometries is crucial in recent engineering, since these flows occur in heat management devices, catalytic reactors, energy storage systems, and innovative manufacturing processes. Motivated by these applications, the current study inspects the flow of a chemically reactive fluid induced by an extending or contracting rotating cone subjected to wall-deforming boundary conditions. Additionally, this model integrates the combined effects of endothermic and exothermic chemical reactions, magnetic field, porous medium, non-linear thermal radiation, and cross-diffusion phenomena to assess the fluid flow, heat, and mass transport features. Using the similarity variables, the partial differential equations are converted to ordinary differential equations. Furthermore, the reduced equations are solved numerically using the finite difference method. Moreover, a physics-informed neural network architecture is presented to enhance prediction accuracy by integrating fluid flow physics with artificial intelligence. The major findings of the present study show that the velocity profiles get more intense as the ratio of deformation to rotation parameter values increases. The thermal profile increases for the exothermic case and decreases for the endothermic case as the values of the chemical reaction parameter increase. As the radiation parameter values increase, the Nusselt number decreases by approximately 15.7%. The Nusselt number decreases by around 42.3% in the exothermic case and increases by around 24.1% in the endothermic case when the chemical reaction parameter increases. These results offer both methodological innovation and valuable insights with applications in energy-efficient flow systems, porous media transport, and chemical processing.