Multi-objective heat transfer optimization of hybrid nanofluids in porous medium

An artificial intelligence (AI) driven computational study has been conducted to investigate non-linear heat and mass transfer of hybrid nanofluid across a stretching surface placed in a porous media under magnetohydrodynamic (MHD) effects. This study will serve as a platform to evolve sustainable thermal management system through balance of heat transfer enhancement and entropy generation. A hybrid model incorporating Brownian motion, thermophoresis, mixed convection and Darcy–Forchheimer resistance is developed to examine two-dimensional, steady, laminar boundary-layer characteristics. High-quality numerical results obtained using a finite-volume based solver are used to construct a robust meta-model for fast evaluation. A multi-objective approach based on NSGA-II is utilized to optimize the parameters of the hybrid model for maximizing Nusselt number and minimizing entropy generation. An increase in the thermophoretic and Brownian diffusion parameters enhances the entropy generation by more than 40%. The increase in heat transfer (optimum for moderate values of Prandtl number P r = 3 ) leads to a corresponding enhancement of entropy generation. An increase in the magnetic field (Hartmann number) suppresses convection and creates a thicker thermal boundary layer, which results in a higher Nusselt number and increased entropy production. A robust surrogate model is developed with R 2 values higher than 0.996 to facilitate efficient design space exploration. A novel CFD–AI paradigm is developed and used to uncover a nuanced trade-off between thermal performance and irreversibility, and thus provides useful design and optimization guidance for next-generation hybrid nanofluid thermal systems.

Hali tarjima qilinmagan