The Film Boiling Look-Up Table: An Improvement in Predicting Post-CHF Temperatures

Abstract

During the past 50 years more than 60 film boiling prediction methods have been proposed (Groeneveld and Leung, 2000). These prediction methods generally are applicable over limited ranges of flow conditions and do not provide reasonable predictions when extrapolated well outside the range of their respective database. Leung et al. (1996, 1997) and Kirillov et al. (1996) have proposed the use of a film- boiling look-up table as an alternative to the many models, equations and correlations for the inverted annular film boiling (IAFB) and the dispersed flow film-boiling (DFFB) regime. The film-boiling look-up table is a logical follow-up to the development of the successful CHF look-up table (Groeneveld et al., 1996). It is basically a normalized data bank of heat-transfer coefficients for discrete values of pressure, mass flux, quality and heat flux or surface-temperature. The look-up table proposed by Leung et al. (1996, 1997), and referred to as PDO-LW-96, was based on 14,687 data and predicted the surface temperature with an average error of 1.2% and an rms error of 6.73%. The heat-transfer coefficient was predicted with an average error of -4.93% and an rms error of 16.87%. Leung et al. clearly showed that the look-up table approach, as a general predictive tool for film- boiling heat transfer, was superior to the correlation or model approach. Error statistics were not provided for the look-up table proposed by Kirillov et al. (1996). This paper reviews the look-up table approach and describes improvements to the derivation of the film- boiling look-up table. These improvements include: (i) a larger data base, (ii) a wider range of thermodynamic qualities, (iii) use of the wall temperature instead of the heat flux as an independent parameter, (iv) employment of fully-developed film-boiling data only for the derivation of the look-up table, (v) a finer subdivision and thus more table entries, (vi) smoother table, and (vii) use of the best of five prediction methods for areas where data are unavailable. The look-up table is based on 20,785 film-boiling data points, which were carefully selected from a data bank compiled by the University of Ottawa. These data were all believed to be obtained in the fully developed film-boiling region. A comparison of the fully developed film-boiling look-up table with the fully developed film-boiling database shows an overall rms error in heat-transfer coefficient of 10.58% and an average error of 1.71% (the corresponding errors of the previous heat-flux controlled look-up table with the updated data base are: 20.65% rms and 6.87 % average error). A comparison of the prediction accuracy of the look-up table with other leading film-boiling prediction methods clearly demonstrates the superiority of the present look-up table.