Statistical core design methodology using the vipre thermal-hydraulics code

Abstract

An improved statistical core design methodology for developing a computational departure from nucleate boiling ratio (DNBR) correlation has been developed and applied in order to analyze the nominal 1.3 DNBR limit on Westinghouse Pressurized Water Reactor (PWR) cores. This analysis, although limited in scope, found that the DNBR limit can be reduced from 1.3 to some lower value and be accurate within an adequate confidence level of 95%, for three particular FSAR operational transients: turbine trip, complete loss of flow, and inadvertent opening of a pressurizer relief valve. The VIPRE-01 thermal-hydraulics code, the SAS/STAT statistical package, and the EPRI/Columbia University DNBR experimental data base were used in this research to develop the Pennsylvania State Statistical Core Design Methodology (PSSCDM). The VIPRE code was used to perform the necessary sensitivity studies and generate the EPRI correlation-calculated DNBR predictions. The SAS package used these EPRI DNBR correlation predictions from VIPRE as a data set to determine the best fit for the empirical model and to perform the statistical analysis. The PSSCDM not only includes the EPRI correlation/test data standard deviation but also the computational uncertainty for the particular VIPRE code model used and the new PSSCDM composite box design correlation. The resultant PSSCDM equation adequately mimics the EPRI DNBR correlation results well with an uncertainty of 3.89%. The combined uncertainty yields a reduced new DNBR limit of 1.18, for the specific lumped channel and subchannel VIPRE model, correlation and coefficients. Although the PSSCDM is based on a typical PWR core VIPRE model, this PSSCDM approach can be easily applied to other PWR plant-specific VIPRE models.