In France, nuclear reactor pressure vessels (RPV) integrity assessment follows a deterministic methodology and is performed according to the 1999 French regulation requirements with the use of safety coefficients. In the deterministic methodology, the input data are penalized in order to address the uncertainties.
EDF and AREVA developed some couplings between integrity assessment models and uncertainty treatment software in order to address explicitly the uncertainties on the input data.
In such analysis, the accumulation of conservatisms is replaced by a rational combination of parameters taking into account their uncertainty and variability.
This paper presents a benchmark campaign conducted jointly by EDF and AREVA to validate the couplings between the codes and to strengthen the confidence in the calculations.
The benchmark campaign was performed with several physical thermo-mechanical codes and uncertainty treatments software tools to evaluate the sensitivity of the analyses results as a function of code effects. The benchmark results are in good accordance between the several codes used. This conclusion gives confidence in the code couplings used for performing the uncertainty treatment analyses.
