Uncertainty analysis for the predicted proportion of fuel rods failed during a hypothetical loss of coolant accident

摘要

One of the requirements in Germain licensing processes for pressurized water reactors (PWRs) is that the core damage extent must not exceed 10% in case of a hypothetical loss of coolant accident (LOCA). The core damage extent is the proportion P of fuel rods in the core bursting during the accident. Mainly two computer codes are applied to predict P: a thermal-hydraulics code and a fuel rod code evaluating the fuel rod behaviour under LOCA-thermal-hydraulics. This paper presents an approach to an uncertainty analysis for P based on the Monte Carlo technique. Prime intention of the analysis is to account for the numerous epistemic uncertainties involved in the calculation of P. To reduce the computational effort, the analysis is performed on the basis of a random sample of fuel rods. Thus, also the uncertainty in P due to the inference from only a sample from the core has to be taken into account. The Bayesian approach is used to describe the overall uncertainty in P. Uncertainty statements are derived from a random sample from the average posterior distribution of P. For comparison, also a frequentist approach based on an upper 95%-confidence limit is considered. Results from an illustrative uncertainty analysis are presented.