Publication

Abstract : In fluid flow applications, endothermic and exothermic chemical reactions are important, especially in the scientific and engineering fields. They make advanced modeling and optimization of complex systems possible when combined with artificial neural networks (ANNs). As a result, the present investigation uses ANNs to study the influences of heat source/sink, endothermic/exothermic chemical reactions, and Darcy-Forchheimer porous media on the two-dimensional, stable, incompressible flow of bio-convection nanofluids through a stenosed artery (cylinder). Using appropriate similarity equations, non-linear partial differential equations are transformed into ordinary differential equations, which are then resolved with RKF-45 and the shooting technique. Important engineering coefficients were also investigated. Outcomes show that in an endothermic chemical reaction, the temperature profile increases as the chemical reaction parameter rises, whereas in an exothermic chemical reaction, the reverse behaviour is observed. The C f % shows negligible variations with the addition of nanoparticles at about 8.2 % across distinct parameter values. The N u % is strongly influenced by nanoparticles, increasing significantly for λ 1 > 0 than λ 1 < 0 . Model-Agnostic-Meta-Learning relative studies show high convergence; it generalizes effectively on unknown data. Error histogram studies validate performance analysis, training is stable, and the predicted values almost equal the actual values, proving its effectiveness.