ABSTRACT

Nanofluids have shown higher heat transfer performance in comparison with conventional heat transfer fluids. Investigating the best model to predict/simulate the heat transfer coefficient of nanofluid is the main goal of this study, this is necessary as most previous studies done with different models do not agree.

Six models were studied using ANSYS fluent. The six models studied were Single phase, Discrete phase model, Volume of fluid, mixture model, Eulerian model and the new combined mixture and discrete phase model were studied in a 2D cylindrical pipe assuming axis-symmetry case with different thermal wall boundary conditions. The study was carried out assuming both constant and temperature dependent thermophysical properties of the basefluid in laminar, transitional and turbulent flow regime.

For the laminar flow regime, with particle size of 21nm and 3% volume fraction, Single phase showed the closest agreement with experiment with average deviation of 9% while Eulerian and DPM had an average deviation of 13% and 14% respectively, Mixture and combined model of mixture and DPM showed unrealistic increase in the heat transfer coefficient value. DPM predicted the heat transfer coefficient accurately in the case with particle size and volume fraction of 45nm & 4% and 150nm & 2% with a very small deviation. The single phase closely predicted the heat transfer value in case with particle size of 45nm and a volume fraction of 2% with an average deviation of 3%. From the cases carried out, the single phase and DPM are the closest model to experimental data. The Single phase is only as good as its correlation.