Physics-Informed Neural Networks for Dissipative Micropolar Nanofluid Flow with Microrotation Dynamics and Zero Nanoparticle Mass Flux

This research presents a physics-informed deep learning framework for investigating the magnetohydrodynamic flow of a dissipative non-Newtonian micropolar nanofluid induced by a stretching sheet, incorporating Stefan blowing, internal heat generation, and the zero nanoparticle mass flux condition. The physical model consists of the interplay between the microrotation dynamics, resistance of porosity on the microrotation, Brownian diffusion, and thermophoretic transport phenomenon. The numerical solutions for the nonlinear yielded equations that result from the above interaction are obtained by employing a PINN that considers the laws of physics and boundary conditions. With this technique, the flow behavior, temperature, concentration, and microrotation fields can be predicted accurately without requiring huge datasets. This shows the ability of PINNs to numerically treat highly-coupled nonlinear transport equations in a very efficient manner compared to other traditional methods. The important discoveries from this study include that the porous and magnetic factors increased the skin friction coefficient, but the magnetic effect and viscous dissipation decreased the rate of heat transfer, and the thermophoresis effect decreased the rate of mass transfer while the Brownian effect increased it. The precision of the PINN algorithm is confirmed by comparison of the results with the earlier findings, which proves very high accuracy and hence the robustness of the current computing framework. Results of this research are useful for the development of some thermal management systems, energy converters, cooling methods, chemical reaction processes, fuel cell technology, porous media reactors, and ocean engineering involving the transport of complicated non-Newtonian nanofluids.