Modeling and Analysis of Natural Convection Heat Transfer in Nanofluids

Nanofluids are engineered colloids made of a base fluid and nanoparticles (1–100 nm). The presence of nanoparticles causes a dramatic enhancement of thermal conductivity, an increase of convective heat transfer coefficient as well as of viscosity. These features make nanofluids suitable for the most common industrial cooling and heat transportation applications, for example in the heat exchanger whose performances can be dramatically improved. In the nanofluid literature it is not really evident the mechanism inside the unusual heat transport properties. Several studies concerning nanofluids were carried out to provide experimental data for different configurations and to find models suitable with these experiments. Unfortunately measurements available in literature seem to be affected by a significant dispersion so that some experimental data are not coherent with the others. The issue is that the properties of nanofluid are influenced by many factors such as the nature of the components, the nanoparticle size, shape and concentration, the temperature, the pH of the solution, the presence of surfactants (used to stabilize suspensions), and the charge state of the particle in suspension. Not all of these quantities are usually measured in an experimental campaign and then sometimes it is not possible to make a comparison between different experimental data available in literature. For this reason, several models proposed to validate experimental measurement work well only within a small range of validity, in terms of temperature or concentration interval or nanoparticle type. In this paper we consider always the nanofluid as a single phase and we compared different models presented in literature for the following properties: density, specific heat, viscosity and thermal conductivity. (All this properties depend, at least, on the nanoparticles concentration in the base fluid). The water-Al₂O₃ nanofluid is considered since several models and experimental data are available for this kind of fluid. The numerical simulations have been made by using the CFD code Fluent (release 6.3), where the models have been implemented by using external routines. The natural convection in a horizontal tube heat exchanger has been simulated in a wide region of conditions for which experimental data are available. Different models proposed in literature for viscosity and thermal conductivity have been considered, and compared to empirical models obtained by means a regression from experimental data. Aim of this work is to set suitable models which allows reproducing nanofluid behavior with a good accuracy in a wide region of different conditions.