Validation studies and interpretation of the Oskarshamn-2 1999 stability event with SIMULATE-3K

The OECD/NEA recently launched an international benchmark on a combined feedwater transient and stability event that occurred at the Swedish Oskarshamn-2 (O2) nuclear power plant (NPP). The primary benchmark objective is to assess advances in coupled neutronic/thermal-hydraulic codes for simulations of challenging transients including the appearance of unstable power oscillations. The Paul Scherrer Institut (PSI) is participating in this benchmark in order to enlarge the validation basis of its advanced stability analysis methodology currently under development for Swiss B WRs and based on the state-of-the-art SIMULATE-3K (S3K) code. This paper presents the development, optimization and validation of a S3K model for the O2 benchmark, including the analysis of the entire February 1999 event as well as additional stability tests considered as blind cases. With the optimized model, the S3K solution is compared to available benchmark event data both for at steady-state and transient conditions. For the latter, the qualitative as well as quantitative behavior of S3K results is compared and discussed in relation to the experimental observations. The modeling aspects found in this context to mostly affect the S3K ability to reproduce the event are also presented. As a result, the S3K model indicates that if the reactor scram had not been introduced, the observed diverging oscillations would reach a maximum amplitude before decaying back into a stable state. However, it is also found that the core could instead have evolved into limit-cycle oscillations if a stabilization of the feedwater flow and temperature had occurred just before the scram signal. Finally, the S3K model optimized for the event analysis is subsequently validated in a blind mode approach against several stability tests conducted before and after the event. Considering that the uncertainties in measured decay ratio values are not known and could be relatively large, the obtained results can on a qualitative basis be considered as satisfactory.