Statistical study and a complete overview of nanofluid viscosity correlations: a new look

Nanofluids are considered the top candidates to replace surface cooling systems, making it essential to study the effect of nanoparticles on thermophysical properties of the base fluid when it is added. Viscosity is a crucial factor in heat transfer, especially convection heat transfer. In most of the studies published, the correlations obtained from experiments were performed without examining statistical tests, and the effect of different parameters, including temperature, volume (mass) fraction, etc., on the viscosity of nanofluid in the proposed correlations was not specified. Moreover, some correlations it was shown that the elimination of one of the parameters had no effect on the response of that correlation. For statistical analysis, analysis of variance and sensitivity analysis were used to determine the relationship of the correlation with its variable parameters. The results showed that approximately 27.2% of the correlations presented for the ethylene glycol-based nanofluid and 27.7% of the correlations presented for the water-based nanofluid are reliable. Finally, as until now, no accurate correlation has been provided for the viscosity in a wide temperature and volume fraction range. According to the R-square statistical index, viscosity models were obtained in this study with an accuracy of 97.01% and 96.08% for water- and ethylene glycol-based nanofluids, regardless of the nanoparticle type. Also, the RMSE value was improved by 35.82% and 49.84% compared to the best correlation presented by the researchers for estimating the viscosity of water-based nanofluid and ethylene glycol-based nanofluid, respectively.