Adequacy analysis, validation and input uncertainty quantification for critical flow predictions using the CATHARE 3 code in the framework of the OECD-NEA ATRIUM project

• CEA’s results for the first exercise of the OECD/NEA ATRIUM project. • First application of the SAPIUM guideline for Inverse Uncertainty Quantification. • Adequacy analysis of the experimental database with an Analytical Hierarchy Process. • Simulation and validation of critical flow experiments with the system code CATHARE3. • Quantification of uncertainties associated to flashing and wall-to-liquid friction models. Defining the uncertainties of the physical models appearing in thermalhydraulic system codes is one of the most important challenges for the application of the Best Estimate Plus Uncertainty (BEPU) methodology. Several Inverse Uncertainty Quantification (IUQ) methods are available in literature and are used to assess the problem. To harmonize the use of such methods, a practical guideline was developed in the framework of the OECD-NEA SAPIUM project. The first attempt to use this guideline is the OECD-NEA ATRIUM project, which consists of two IUQ exercises and a forward-propagation on a LSTF test simulating an Intermediate Break Loss of Coolant Accident in hot leg. This work presents CEA’s results of the first exercise of the ATRIUM project, which is the first practical application of the SAPIUM guideline. The uncertainties associated to the critical flow predictions with the CATHARE3 code are quantified using three experimental set-ups (Sozzi-Sutherland, Super Moby-Dick and Marviken-CFT). The article details all the necessary steps to obtain reliable, robust and explicable uncertainties. The most representative tests are selected via an adequacy and experimental analysis. The validation of the code is demonstrated and the most influential physical models are determined via a dedicated sensitivity analysis. Using CEA’s methods for IUQ, such as standard CIRCE and CIRCE-2 Steps, the uncertainty of the flashing and the wall-to-liquid friction models are quantified and validated. The applicability and usefulness of the SAPIUM guideline is therefore demonstrated.