Quantification of modeling uncertainties based on scaling laws in natural circulation decay heat removal

A Best Estimate Plus Uncertainty (BEPU) analysis is one of the good methods to estimate the uncertainty of phenomenon in a nuclear power plant dynamics. In BEPU analysis, a number of numerical analyses, in which input parameters are varied based on their probabilistic distributions, are carried out to obtain statistical characteristics of the output result. In general, the uncertainty of input parameters, such as a probabilistic distribution form and variance, are estimated based on experimental knowledge and/or engineering judgment. In the present research, we focus on a scaling law (dimensionless number) in constitutive equations from a view point of phenomenological theory. An influence of uncertainty in the dimensionless number and its dependency on BEPU analysis has been investigated. Plant dynamics analyses of Super Safe, Small and Simple (4S) reactor, being developed by Toshiba, are carried out under a natural circulation decay heat removal condition. In the analysis, uncertainties of the dimensionless numbers such as Nusselt, Reynolds, and Prandtl numbers are taken into consideration, as well as an uncertainty of decay heat power. The Latin Hypercube Sampling is applied to determine the input deck set. As a result, it is demonstrated that the parameter dependency on the output result can be revealed by using the dimensionless numbers. (author)