Artificial neural network analysis of electroosmotic triple diffusion Ree-Eyring nanofluid transport in a complex...

Complex cilia waves have significant applications in biomedical engineering and physiological fluid transport, where they accurately model the coordinated motion of cilia in respiratory tracts, reproductive systems, and microfluidic devices. They also enhance mixing, pumping efficiency, and targeted transport in drug delivery systems, bio-inspired micropumps, and lab-on-chip technologies. An electroosmotic MHD Ree-Eyring nanofluid transport through a symmetric transverse porous channel is investigated by incorporating triple diffusive convection, thermal radiation, viscous dissipation, and Joule heating under the influence of a complex cilia wave. A mathematical model is presented in which the nonlinear governing equations are transformed using lubrication approximations and solved numerically using the robust BVP5C collocation technique, and an Artificial Neural network (ANN) framework is employed to authenticate and predict numerical solutions. The temperature profile exhibits an approximate 35-55% increase with increasing Brinkman number, indicating strong enhancement in thermal energy due to viscous dissipation effects. A slight rise of nearly 5-8% is observed with increasing Brownian motion parameter, while the radiation parameter causes an approximate 8-12% reduction in temperature because of enhanced thermal energy dissipation. Furthermore, increasing the heat source parameter Q elevates the temperature by nearly 10-18%. The ANN model demonstrated excellent agreement with the BVP5C numerical solutions for both temperature and velocity profiles under varying different parameters. Extremely low prediction errors were observed with MSE of order 10^-8, MAE of order 10^-5, and R²values approaching unity, confirming the high precision and robustness of the ANN framework. Furthermore, the error distributions, heatmaps, and cumulative analyses verified that the ANN could serve as a fast and reliable surrogate model for complex nonlinear fluid flow problems. It is observed that increasing radiations parameter reduces the Nusselt number, whereas larger Brickman number, Brownian motion parameter and heat generation parameter, enhance heat transfer due to intensified viscous dissipation, nanoparticle diffusion, and internal heat generation.