Publication Date: June 5, 2025
Access counts:0
A New Method for Realistic Seismic Probabilistic Risk Assessment and Its Application to Next-Generation Reactors
-Seismic Fatigue Failure Probability Assessment Considering Post-Buckling Behavior-
Fig. 1 Structural analyses of the entire reactor vessel using the finite element method and zooming analyses
Fig. 2 Failure probabilities of the reactor vessel
Conventional seismic probabilistic risk assessments (PRAs) for fast reactors are conducted assuming that the reactor vessel (RV) which is an important component for safety remains within the elastic deformation range and ignoring margin from RV buckling to actual RV failure; hence, RV buckling is considered a failure. Thus, the assessed failure probability is higher than the actual value, and RV failure that results in the reactor level loss is a dominant factor in determining the core damage frequency. However, RV buckling does not necessarily lead to rapid RV collapse or rupture as it depends on changes in the natural RV period caused by buckling distortion. Instead, oscillatory deformation induced by seismic activity can lead to RV fatigue failure. Recognizing this, we have developed a seismic failure probability assessment method that considers the post-buckling deformation behavior.
In this method, the finite element mesh model of the RV is first created, and the meshed structure is analyzed using the finite element method to determine its deformation and strain responses under multiple ground motion levels that are several times the input design-basis ground motion level (Fig. 1a). Second, regions that may experience severe localized strain, which can cause fatigue failure hotspots, are identified (Fig. 1b). Subsequently, a zooming analysis is conducted on these identified regions using their fine mesh models (Fig. 1c). Based on the obtained strain history, the cumulative fatigue damage fraction is calculated. By repeating this process for different input ground motion levels, the threshold ground motion level that will cause RV fatigue failure is determined. Finally, this study estimated the post-buckling fatigue failure probability as a function of the ground motion level by incorporating uncertainties.
As a case study, we have applied this method to the RV of a next-generation loop-type fast reactor. The result indicates that the post-buckling fatigue failure probability is substantially lower than the failure probability evaluated by using conventional PRA method (Fig. 2).
There might be a possibility that RV fails in failure modes other than the post-buckling fatigue failure. Evaluations of failure probabilities for these failure modes are future works.
If you have any comments or feedback about this page, please click the button to share your thoughts with us.