Impact of Uncertainty on Prediction of Supercritical CO2 Properties and Nusselt Numbers

Abstract

Abstract The impact of measurement uncertainty on Nusselt number correlations for supercritical CO2 is investigated. Selection of appropriate reference quantities for these correlations, such as thermal conductivity, appears to have a large effect on accurately predicting heat transfer rates. Supercritical CO2 work heavily depends on real fluid property data, which show that fluid properties have very large gradients with respect to temperature and pressure near the critical point. The sharp gradients imply heat transfer predictions are highly sensitive to the accuracy of temperature and pressure experimental measurements in this region. Root sum of squares uncertainties at various property values indicate predictably large (on the order of 1000%) uncertainties in calculated Reynolds, Prandtl, and Nusselt numbers near the critical point. Interestingly, uncertainties remain several times the calculated value for operating pressures between 7.5 and 8.5 MPa, which are common in the experimental literature—highlighting a need for careful application of correlations near the pseudocritical line and the benefits of presenting dimensional data in the literature.