چكيده به لاتين
In the present thesis, the flow and heat transfer of turbulent mixed convection of different nanofluids through a duct has been investigated. The nanofluid flow has been modeled via the single phase model with variable thermo-physical properties and two phase model with nanoparticles injection. However, the single phase model predicts the flow and heat transfer characteristics of nanofluid well, but the accuracy of two-phase model is higher than single phase model. The results of two-phase modeling indicates that at the fully developed region, the diffusion of thermophoresis is higher than Brownian motion and therefore the nanoparticles dispersion is higher at the core region of duct. Whilst at the entrance region the nanoparticles are dispersed uniformly. Different approaches are used to simulate the turbulent flow. The Scale-Adaptive Simulation (SAS) and Large Eddy Simulation (LES) approaches show higher accuracy than Reynolds Averaged Navier stokes (RANS) models in predicting heat transfer, pressure drop, capturing the coherent structures, velocity and temperature fluctuations of nanofluids. In parallel with velocity and temperature fluctuations, increasing the nanoparticles volume fraction, augments the Reynolds stresses and turbulent heat fluxes that enhance the energy transfer between nanofluid layers. Moreover, the artificial neural networks have been used to investigate the wide range of nanofluid flow considering the effect of Reynolds number, nanoparticles volume fraction and diameter on Nusselt number, friction factor and entropy generation to find the optimum range of nanofluid flow.
