Physics-informed neural network coupled with semi-analytical hybrid modeling of micropolar magnetohydrodynamic channel flow with heat and mass transport in a porous medium

The present study explores the steady magnetohydrodynamic (MHD) flow of a micropolar fluid within a porous channel, considering the effects of microrotation, spin-gradient viscosity, micro-inertia, and convective heat and mass transfer. A hybrid computational approach is developed by integrating a physics-informed neural network (PINN) with the semi-analytical Akbari–Ganji method (AGM), aiming to efficiently solve the highly nonlinear and coupled boundary value problem. The PINN model is trained directly on the governing equations and boundary conditions, while AGM serves as a benchmark for validation. A comprehensive parametric analysis is performed to investigate the influence of key nondimensional parameters, including magnetic field strength, micropolar coupling, and transport numbers. The findings reveal that magnetic forces tend to suppress microrotation and reduce thermal and solutal transport, whereas micropolar parameters enhance rotational motion and mass diffusion. The proposed PINN framework demonstrates high accuracy and strong agreement with the semi-analytical results, validating its effectiveness in modeling complex MHD flows. These results provide valuable insight into controlling microstructural effects for improved heat and mass transfer in porous media systems.