The requirements for an emergency core cooling system (ECCS) evaluation model that is acceptable for a pressurized water reactor licensing analysis are detailed in Appendix K to 10CFR50. The purpose of these requirements is to ensure that such an analysis will yield a conservative upper bound to the maximum cladding temperature and cladding oxidation that can result from a postulated loss-of-coolant accident (LOCA). By its nature, therefore, this model is inappropriate to indicate the actual anticipated results of a LOCA. Furthermore, a quantitative assessment of the conservatism inherent in the licensing model is unavailable. To produce realistic LOCA results, a calculation was performed at Combustion Engineering (C-E) for the reactor in its System 80™ nuclear steam supply system, using a best-judgment ECCS evaluation model. The best-judgment model is a C-E first-generation best-estimate model that uses the basic Appendix K licensing computer programs, but in which the bounding conservatisms required by Appendix K are relaxed for selected parameters and models of primary concern in a LOCA analysis. The important differences between the best-judgment model and the Appendix K licensing model are as follows: 1. In the best-judgment calculation, nominal values of certain reactor system parameters were used in place of the bounding, conservative values assumed in the licensing calculation. Of primary importance are the relaxation of the U.S. Nuclear Regulatory Commission (NRC)-imposed double-ended guillotine break, and 20% contingency on the American National Standards standard decay heat generation curve. Nominal values were also assumed for the containment building physical parameters and wall condensing heat transfer coefficients, which influence the calculation of transient containment pressure. 2. It was assumed that offsite power was lost upon pipe rupture, but that auxiliary power from the diesel generators was available to active ECCS and other safeguard components following the normal startup and loading sequence. All active safeguard systems were assumed to be operating at nominal capacity in their most likely condition throughout the accident. Power, from the coasting-down turbine generator, was maintained to the reactor coolant system pumps during the blowdown, and the pump rotor was assumed to coast down during reflood. 3. A critical flow model deemed by C-E to be appropriate for break flow rate calculations was used. In the licensing LOCA analysis, the maximum local power density was adjusted such that the Appendix K model yielded a peak clad temperature approximately equal to the criteria limit of 2200°F (1204°C), thus establishing a corresponding operating limit. The best-judgment calculation, performed at the same indicated peak local power density, yielded a maximum clad temperature that was 980°F (544°C) lower than that predicted by the Appendix K model. At such low temperatures, clad oxidation and rupture will not occur. An additional calculation was performed in which the peak local power density was decreased to a value that permits full-power operation, but limited operating flexibility; the maximum cladding temperature decreased an additional 100°F (56°C). Although no attempt has been made to specify a statistical confidence level for either the assumptions or the results of this analysis, it is evident that predictions of the consequences of a LOCA that are obtained from an ECCS evaluation model conforming to 10CFR50, Appendix K, are extremely conservative.
