Towards a Consolidated Approach for the Assessment of Evaluation Models of Nuclear Power Reactors

Abstract

The STARS project at the Paul Scherrer Institut (PSI) has adopted the TRACE thermal-hydraulic (T-H) code. For analyses involving interactions between system and core, a coupling of TRACE with the SIMULATE-3K (S3K) LWR core simulator has been developed. In this configuration, the codes and associated simulation models play a central role to achieve a comprehensive safety analysis capability. Therefore, efforts have now been undertaken to consolidate the validation strategy by implementing a more rigorous and structured assessment approach for TRACE applications. The principle is to systematically track the evolution of a given set of predicted physical Quantities of Interest (QoIs) over a multidimensional parametric space. If properly set up, such environment should provide code developers and code users with persistent (less affected by user effect) and quantified information (sensitivity of QoIs) on the applicability of a simulation scheme (codes, methodology and input models) for steady state and transient analysis of full LWR systems. Through this, for each given transient/accident, critical paths of the validation process can be identified that could then translate into defining reference schemes to be applied for downstream predictive simulations. To illustrate this approach, this validation strategy is applied to an inadvertent blowdown event that occurred in a Swiss BWR/6. The transient was initiated by the spurious actuation of the Automatic Depressurization System (ADS). Here, the validation approach progresses through a number of dimensions: a) different versions of the TRACE code, b) the "methodology" dimension. In this case imposed power and an updated TRACE core models are investigated, and c) the "nodalisation" dimension, where changes to the input model are assessed. For each step in each validation dimension, a common set of QoIs are investigated. For the steady-state results, these include fuel temperatures distributions. For the transient part of the present study, the evaluated QoIs include the system pressure evolution and water carry-over into the steam line. It has been seen that the improvements to the model predictions resulted in a small impact on the system pressure gradient, thus confirming a persistency of the downstream mechanical stress estimate, whereas the water carry-over could vary by up to 150% as function of the adopted simulation methodology.