Physics informed neural network simulation of trihybrid nanofluid slip flow along a catalytic surface: Fuel cell applications

Abstract This study examines the influence of catalytic reactions on the flow of trihybrid nanofluids in a Darcy porous medium, motivated by improving fuel cell performance. Three nanofluids (mono‐(), binary hybrid () and tri‐hybrid ()) are considered. Exothermic catalytic reactions have been analyzed using Arrhenius kinetics. To analyze the thermo‐solutal transport phenomena, the primitive equations have been reduced to a set of similarity equations. The physics‐informed neural network (PINN) algorithm has been utilized to solve the transformed equations. Validation of the PINN results for certain fixed parameters has been included. Details of the coding, iterations and training losses are also provided based on L‐BFGS (i.e., the limited‐memory Broyden–Fletcher–Goldfarb–Shanno optimizer). The impacts of Darcy number, heat of reaction, reaction rate, and activation energy parameters on flow characteristics have been evaluated. Physical quantities and flow profiles have been analyzed through a novel deep‐learning approach using PINN. Velocity, temperature and solute concentration distributions, friction , Nusselt , and Sherwood number are computed. It is found that an elevation in the heat of reaction parameter significantly increases the temperature whereas it suppresses solutal concentration. A comparison of friction, heat, and mass transfer shows that for the hybrid nanofluid decreased by 12.27%, increased by 3.54% and increased by 3.54%. For the tri‐hybrid nanofluid, decreased by 12.92%, increased by 5.81%, and increased by An increase in the activation parameter enhances concentration. The present computations find applications in fuel cell plate design, bioreactor transport phenomena, and nano‐biochemical systems.