Review of uncertainty methods for CFD application to nuclear reactor thermalhydraulics

Abstract : In the past ten years, the Working Group for the Analysis and Management of accidents (WGAMA) initiated activities to promote the use of CFD for Nuclear Reactor Safety (NRS). Best Practice Guidelines were written which are applicable to single phase CFD. Assessment requirements were also addressed. These activities provided more confidence in the application of CFD for safety by defining some conditions and requirements. However, no applicative methods were written about a possible quantitative evaluation of the uncertainty of predictions which is mandatory for complementing a Best-Estimate approach within a nuclear reactor licensing framework. A new activity reported here was then initiated to review the methodologies for determination of the uncertainty of CFD predictions applied to reactor thermalhydraulics. An OECD report was written which is summarized in this paper. A proposal for CFD UQ global approach shows the link between PIRT, Verification, Validation, and Uncertainty Quantification (UQ). The various sources of uncertainty are identified. Methods for uncertainty quantification are then reviewed, considering accuracy extrapolation and uncertainty propagation methods (with the possible use of meta-models). Then a few methods or elements of methods are summarized. The role of Separate Effect Tests and Integral Effect Tests in the UQ process is mentioned. Finally some conclusions are drawn, remaining needs are identified, and recommendations for further RandD and benchmarking of methods are given to progress on this topic.