Entropy generation and artificial neural network prediction for EMHD dusty tangent hyperbolic nanofluid flow over a rotating porous disk

This study investigates the electromagnetohydrodynamic (EMHD) transport of a two-phase dusty tangent hyperbolic nanofluid over a rotating porous disk in a Darcy-Forchheimer medium, incorporating the Buongiorno nanofluid model, thermal radiation, viscous dissipation, Joule heating, non-uniform heat generation/absorption, Arrhenius activation energy, and interphase momentum/heat exchange with the dust phase. The coupled consideration of these mechanisms is important for understanding and controlling momentum, heat, mass-transfer, and entropy-generation characteristics in complex rotating nanofluid systems. The governing equations are reduced via similarity transformations and solved numerically using the MATLAB bvp4c solver, validated against published limiting-case results. The results show that the Hartmann and electric field parameters accelerate both phases, radiation and heat source raise temperatures, and Schmidt number and chemical reaction reduce concentration. Skin friction strengthens with Hartmann number but weakens with porosity; entropy generation increases with Brinkman number, Hartmann number, and radiation, while the Bejan number reflects the shifting balance between fluid-friction and heat-transfer irreversibility. An artificial neural network (ANN) is subsequently developed to predict the radial and azimuthal skin-friction coefficients, Nusselt number, Sherwood number, and Bejan number, achieving and 𝑅 2 values of approximately 0.999 for all outputs. Overall, the numerical results characterize the coupled transport and thermodynamic behavior of the proposed model, while the ANN provides accurate prediction of the engineering response parameters within the investigated parameter domain