New strategies for quantifying and propagating nuclear data uncertainty in CUSA

The uncertainties of nuclear cross sections are propagated to the key parameters of nuclear reactor core through transport calculation. The statistical sampling method can be used to quantify and propagate nuclear data uncertainty in nuclear reactor physics calculations. In order to use statistical sampling method two key technical problems, method of generating multi-group covariance matrices and sampling method, should be considered reasonably and efficiently. In this paper, a method of transforming nuclear cross section covariance matrix in multi-group form into users' group structures based on the flat-flux approximation has been studied in depth. And most notably, an efficient sampling method has been proposed, which is based on Latin Hypercube Sampling (LHS) combined with Cholesky decomposition conversion. Based on those method, two modules named T-COCCO and GUIDE have been developed and have been successfully added into the code for uncertainty and sensitivity analysis (CUSA). The new modules have been verified respectively. Numerical results for the TMI-1 pin-cell case are presented and compared to TSUNAMI-1D. The comparison of the results further support that the methods and the computational tool developed in this work can be used to conduct sensitivity and uncertainty analysis for nuclear cross sections.