Three dimensional conjugate natural convection heat transfer of Mxene toluene nanofluid in a heat generating finned cubic cavity: finite element and ANN investigation

This study numerically investigates three dimensional natural convection heat transfer of an Mxene toluene nanofluid in a cubic cavity containing six internal heat generating fins using the Galerkin finite element method, motivated by the need for efficient passive thermal management solutions in high heat flux applications such as electronics cooling and solar thermal systems, where Mxene based nanofluids in three dimensional finned enclosures remain largely unexplored. The effect of the Rayleigh number, heat source, and nanoparticle volume fraction on the flow structure, temperature field, and average Nusselt number are systematically analyzed, and the results are validated through an energy balance check showing an imbalance below 2.2%. The results demonstrate that increasing the Rayleigh number significantly strengthens buoyancy driven circulation and improves convective heat transfer, with the average Nusselt number increasing approximately 57 folds as Ra rises from 10 3 to 10 6 , while the heat source parameter about 3.6%. Furthermore, an artificial neural network is developed to accurately predict the average Nusselt number, providing a fast and reliable alternative to computationally expensive CFD simulations with excellent predictive accuracy, achieving regression coefficients above 0.99. The results provide valuable insights for the development of high performance thermal management systems used in electronics cooling, energy storage, and solar thermal technologies.