화학공학소재연구정보센터
Chemical Engineering Science, Vol.131, 172-186, 2015
A correlation for interfacial area concentration in high void fraction flows in large diameter channels
Two phase flows exist as a part of many industrial processes, including chemical processes, nuclear reactor systems, and heat exchangers. In all of these applications the interfacial area concentration is an important parameter for evaluating the interactions between the phases, including drag forces, heat transfer or chemical reaction rates. Many models for interfacial area concentration exist for dispersed bubbly flows; however this type of flow only exists at relatively low void fractions. Very few correlations exist for the prediction of cap-turbulent, slug, or churn-turbulent flows. In this paper a new correlation for predicting the interfacial area concentration beyond bubbly flows in large diameter pipes is derived using a two-bubble-group method (spherical and distorted bubbles as Group-1 bubbles and cap and churn-turbulent bubbles as Group-2 bubbles) and the two-group interfacial area transport equation. The derivation assumes steady state and fully developed flow, and is based on interfacial area transport source and sink terms for large diameter pipes developed by Smith et al., 2012a. Int. J. Heat Fluid Flow 33, 156-167. The resulting equations can be used to predict the void fraction for each group of bubbles and the Sauter mean diameter for each group of bubbles in addition to the total interfacial area concentration. The model is then benchmarked based on the data collected by Schlegel et al., 2012. Exp. Therm. Fluid Sci, 41, 12-22; Schlegel et al., 2014. Int, J. Heat Fluid Flow 47, 42-56. It is found that the correlation predicts the data for Sauter mean diameter of Group 1 bubbles with RMS error of 23.3% and bias of +1.83%. For Group 2 bubbles the RMS error is 24.0% and the bias is +5.35%. This indicates that the correlation somewhat over-predicts the bubble sizes. In spite of this the prediction error remains reasonable compared to the accuracy of previous correlations, and given that the experimental uncertainty can be as high as 15% for some flow conditions. The RMS error and bias in the total interfacial area concentration are 22.6% and -4.29%, respectively. This is consistent with the over-prediction of the Sauter mean diameters, but again is reasonable considering the experimental uncertainty and the prediction error of previous correlations. The model is also able to predict the trends found in the experimental data with varied liquid and gas velocities, representing a large improvement over previous modeling efforts. An expanded database of accurate interfacial area concentration measurements at higher pressures would allow further improvement of the model benchmark and expansion of the range of applicability of the model. (C) 2015 Elsevier Ltd. All rights reserved.