A new semi-empirical correlation for the evaluation of the dynamic viscosity of nanofluids

Thanks to their excellent heat transfer coefficient, nanofluids can be considered as ideal heat transfer fluids for a large number of relevant engineering and scientific applications. Precise assessments of their thermophysical properties are thus essential for reliable calculations. In this work, a new semi-empirical scaled correlation based on 8 parameters (volume fraction, temperature, base fluid critical temperature, base fluid density, base fluid critical density, nanoparticle diameter, base fluid molar mass, nanoparticle density) is introduced to evaluate the dynamic viscosity of nanofluids. The correlation is regressed and evaluated using a dynamic viscosity dataset for 32 nanofluids, including a total of 737 experimental points: 10 nanofluids have water as base fluid (Ag, Al 2 O 3 , Al 2 O 3 /CuO, C, CuO, diamond, Fe/Si, MWCNT, ND-Ni, TiO 2 ), 6 nanofluids have ethylene glycol (Ag, Al2O3, CeO 2 , Co 3 O 4 , SiC, TiO 2 /CuO), 11 nanofluids comprise different mixtures of water and ethylene glycol (Al 2 O 3 , MWCNT/WO 3 , CB, fGnP, G/Dp, G/Dr, nD87, nD97, TiO 2 ), 1 nanofluid has propylene glycol (SiC) and 4 nanofluids comprise different mixtures of water and propylene glycol (TiB 2 , TiB 2 /B 4 C, fGnP). The dynamic viscosity dataset was derived from experimental measurements documented in the scientific literature and conducted with samples that were prepared using consistent and reliable methods. The study evaluates the dynamic viscosity of nanofluids using 14 literature equations to verify their accuracy against the proposed correlation. Results show that the correlation has an average absolute relative deviation of 8.16%, which is significantly lower than that of the literature equations. A 4-fold cross-validation also shows that the correlation is resilient and accurate with different regression datasets. • A dataset of the dynamic viscosity of 32 nanofluids is considered. • A total of 737 experimental points is evaluated, selecting only stable nanofluids. • An 8-parameter, semi-empirical scaled correlation is proposed. • The correlation yields a lower deviation with respect to other existing equations. • A 4-fold cross-validation confirms the accuracy of the correlation.