Simulation study of convective and hydrodynamic turbulent nanofluids by turbulence models

Abstract

The numerical study of nanofluids as a two-phase flow (both as solid nanoparticles and in a liquid phase) has brought about a new approach to simulation in this area. Due to the lack of hybrid models to fully predict the flow characteristics of nanofluids under different conditions, a case can be made for developing homogenous models from numerical simulations. In this study, the convective heat transfer and hydrodynamic characteristics of nanofluids are investigated by simulation with ANSYS-FLUENT. Accordingly, four common types of nanofluids in horizontal turbulent pipe flows have been chosen from experimental data available in literature for modelling purposes. These nanofluids are Al2O3, ZrO2, TiO2 and SiO2. The simulations are done using the built-in models of ANSYS-FLUENT, namely the Mixture model and Discrete Phase Modelling (DPM). Comparing various appropriate turbulence models, the Realisable and Standard k-ɛ models have provided the same results in most of the simulations. The Reynolds stress model (RSM) overestimates pressure drops compared with the other k-ɛ models, while the re-normalisation group (RNG) model overestimates heat transfer coefficient. The anisotropy of instantaneous velocity in the RSM gives higher turbulent kinetic energy, dissipation rate and slip velocity between the particles and the main flow, which makes it an essential part of simulations. All the DPM results have shown the same trend, but with different percentages from measured data, which means that the number of particles plays a key role in the simulations. Any small weaknesses in DPM have a significant influence on the results due to the higher number of nanoparticles.