Stochastic uncertainty quantification for safety verification applications in nuclear power plants

There is an increasing interest in computational reactor safety analysis to systematically replace the conservative calculations by best estimate calculations augmented by quantitative uncertainty analysis methods. This has been necessitated by recent regulatory requirements that have permitted the use of such methods in reactor safety analysis. Stochastic uncertainty quantification methods have shown great promise as they are better suited to capture the complexities in real engineering problems. With advances in computational capabilities in recent times, these methods when utilized would provide distributions of safety important parameters computed by thermal hydraulic codes. In this study, a transient is simulated with a best estimate thermal hydraulic code, CATHENA. Stochastic uncertainty quantification and sensitivity analysis were performed using the OPENCOSSAN software which is based on the Monte Carlo method. The uncertainty and sensitivity analyses results were then utilized to update the dynamic Fault Semantic Network for safety verification. The effect of uncertainty in two input parameters (initial temperature and pressure) was investigated by analyzing the probability distribution of two output parameters. The first four moments of the output pressure and fuel pin temperature were computed and analyzed. The uncertainty in output pressure was 0.087% and 0.048% was found for the fuel pin temperature. These results are expected to provide insight for safety analyses by their utilization in updating the dynamic FSN.