Quantification of condensation model uncertainty based on large scale PCS test SCOPE

Abstract

Quantification of model input uncertainties is very important to development of Best-estimate plus uncertainty (BEPU) accident analysis method. In particular, condensation model uncertainty on the Sherwood number should be quantified accurately in the development of containment response BEPU analysis method. In this paper, Sobol’s total indices were used to assess the parameter sensitivity, nonparametric statistics method with Wilks correlation was used to quantify the Sherwood number uncertainties, and calibration intervals of the condensation factor were obtained with an optimization method. The objective of the optimization was to ensure that the Sherwood number uncertainties quantified from test data were enveloped by those calculated from model prediction. The optimization process was automatically implemented with specified objective function and constraint condition. The results show that the developed method is able to reasonably determine the condensation model calibration interval. The method presents high accurateness, robustness and flexibility and could be conveniently applied to model uncertainty quantification based on separate effects tests and simple integral effect tests.