A MECHANISTIC MODEL FOR CRITICAL HEAT FLUX OF SUBCOOLED FLOW BOILING

摘要

This paper presents a mechanistic model for the prediction of CHF of subcooled flow boiling based on liquid sublayer dryout model. In the model, By writing critical wavelength of Helmholtz Instability to both left and right sides of vapor blanket and by assuming these two wavelengths are equal to each other, the vapor blanket velocity U{sub}B can be written as a simple function of average velocity of liquid bulk V{sub}l which can be obtained by the knowledge we have known. Then, on the base of U{sub}B, other important parameters such as vapor blanket length L{sub}B and liquid sublayer thickness δ can be calculated easily. The model is simple with explicit physics nature, and is characterized by the absence of empirical constants. To verify the present model, two databases (include about 2400 points) are used. One gathered by Celata used to verify his model is characterized by high mass velocity and low-medium system pressure. The other gathered by Pei is characterized by high pressure and low-medium mass velocity. The verification showed that present model could keep its validity in a wide ranges of operating conditions (mass velocity up to 70 Mg/m{sup}2s, system pressure up to 17.5 MPa). Figure A-1 shows a comparison of calculated versus experimental CHF. About 89% of data points are predicted within ±30%. Comparison between Celata model and the present model shows that although present model shows a little worse prediction than Celata model with the data base collected by Eclat, it shows a much better prediction with the data base collected by PEI. A general better prediction than Celata model is obtained with both the databases collected by Celata and PEI. Predictions of some important parameters in liquid sublayer dryout model, such as vapor blanket velocity U{sub}B, initial liquid sublayer thickness δ, vapor blanket diameter D{sub}B and vapor blanket length L{sub}B, are also compared with Celata model during a wide range of operating conditions. The result shows that although the basic thought of two models diverse greatly, the predictions of these parameters differ not so much. The model gives a higher prediction at low L/D and shows this low L/D effect on CHF turns much more obvious with the increase of system pressure. That means, with the pressure increasing, on one side, L/D effect acts in much wider area (L/D effect area begins at higher L/D value). On the other side, in the L/D effect area, CHF jumping amplitude also increases greatly. The reason for this is still not clear. For the CHF at low L/D, as thought by authors, more investigation is needed.