A Separate-Effect-Based New Appraisal of Convective Boiling and its Suppression

Abstract

The development of convective boiling heat transfer correlations and analytical models has been based almost exclusively on the knowledge of global heat transfer coefficients, while the predictive capabilities of the correlation constituting components (typically additive convection and boiling) have remained usually elusive. This becomes important when, for example, developing a mechanistic subcooled void model based on wall heat flux partitioning, or when applying a correlation beyond its developmental range. In the latter case, the preponderance of the individual heat transfer mechanisms, through the phenomenon of boiling suppression, can become significantly different, thus leading to uncharted uncertainty extrapolations. An examination of existing experimental data, obtained under fixed hydrodynamic conditions, has allowed the isolation of the boiling heat transfer contribution over a broad range of thermodynamic qualities (0 to 0.8) and mass fluxes (1,100 to 3,900 kg/(m 2 Ás)) for water at 7.2 MPa. Boiling suppression has been quantified, thus providing valuable new insights on the basic functional relationships of boiling in convective flows. This work has allowed a new interpretation and representation of the standard flow ''boiling map'' (Collier's) to be developed. The convection enhancement and boiling suppression components (F and S) of the well-known Chen's correlation-an important constitutive relationship implemented in several best-estimate (realistic) thermal-hydraulics codes-have been individually determined, showing the pitfall of splitting the correlation for mechanistic boiling heat transfer modelling, and the important role of compensating errors in uncertainty extrapolation. An initial attempt to formulate a new correlation, based for the first time on segregated heat transfer components, is also included.

Development of a Wide-Range Pre-CHF Convective Boiling Correlation

“…The assessment of the Chen correlation against the CISE data 19) showed that the deterioration of the correlation predictive capability, when applied outside its intended domain, was due primarily to its boiling component, h boil ¼ Sh F:Z: , when previous attempts to improve this correlation led, for example, to the modification of its convective component h conv ¼ Fh D:B: instead 17,18) (as indicated above). Thus, while solving the following additive formulation,…”

“…19) Concerning the overall performance of the correlation, the measured-to-predicted (M=P) ratio for the heat transfer coefficient reached a value ðM=PÞ ¼ 2:9, and so, at relatively low wall superheats of about 12 K, with a sensitivity dðM=PÞ=dðÁT sat Þ ¼ 0:45/K. Furthermore, a new examination of the standard (Collier's) flow boiling map 24) indicated that the constant two-phase heat transfer coefficient over the ''low-quality'' region was not due to the predominance of boiling heat transfer, as generally (if not systematically) considered, but to a combination of increased convection and suppressed boiling as the quality increases, leading to a new interpretation and representation of the standard flow boiling map.…”

“…Furthermore, a new examination of the standard (Collier's) flow boiling map 24) indicated that the constant two-phase heat transfer coefficient over the ''low-quality'' region was not due to the predominance of boiling heat transfer, as generally (if not systematically) considered, but to a combination of increased convection and suppressed boiling as the quality increases, leading to a new interpretation and representation of the standard flow boiling map. 19) Stemming from the revealing insights gained in that study, the present work aimed at developing a new correlation whose objective is not high accuracy over a relatively ''narrow'' range of flow conditions, but a ''reasonable'' (or at least charted) accuracy when the correlation is used over the wide range of conditions encountered in thermal-hydraulics code simulations of light-water reactor (LWR) systems (covering pressures ranging from 0.1 to 15 MPa). Simplicity was sought in casting this new formulation, so as to lend itself to implementation in T-H codes without requiring important modifications of the code numerical solution scheme.…”

“…The importance of this quantity has been explicitly recognized in the development of the LAPUR boiling water reactor (BWR) stability code by implementing this quantity as a user-defined parameter. 21) In this plural context of BE calculations, an assessment of the correlation, 19) based on CISE experiments, 22,23) was undertaken at 7 MPa, a median pressure considering both boiling and pressurized water reactor systems, and a broad range of mass fluxes and qualities. The fundamental difference in that assessment was that the boiling heat transfer contributions were isolated purely on an experimental basis, while most, if not all, other studies relied essentially on correlations, and thus only the global heat transfer prediction could be assessed, as opposed to the individual constituting components.…”

“…In the prior study on the phenomenology of convective boiling and its suppression, 19) CISE experiments 22,23) were used to segregate the heat transfer mechanisms. In those experiments, the test section was a relatively short electrically heated tube, preceded by an adiabatic zone fed with saturated water at a selected enthalpy.…”