Assessment of compensating error in the WCOBRA/TRAC-TF2 thermal-hydraulic code

US NRC Regulatory Guide 1.157 outlines the acceptable models and methods by which the performance of an Emergency Core Cooling System (ECCS) can be calculated in a best-estimate loss-of-coolant accident (BE LOCA) analysis. Empiricism in complex, two-phase thermal modeling requires accurate uncertainty assessment on many fronts, including compensating errors, or those model errors which, when applied in combination, could result in good but misleading prediction of the peak cladding temperature (PCT) during the transient. For example, a model that over-predicts vapor temperature and heat transfer coefficient might predict heat flux, and therefore PCT, in accordance with experimental data. The potential existence of this type of error must be investigated to assure confidence in the code's assessment of full plant safety analysis. An extensive evaluation of potential compensating error in the predicted results of integral and single-effects tests was performed in the qualification of the WCOBRA/TRAC BE methodology. Experimental results such as FLECHT forced reflood, ORNL high pressure film boiling, and single tube film boiling tests provided the basis for evaluation. This paper focuses on the studies associated with the reflood heat transfer. The intent is to provide the details of the compensating error analysis in modeling post-critical heat flux (post-CHF) heat transfer phenomena by looking at a subset of the FLECHT tests with very high and very low forced reflood rates. (authors)