A comparison of analysis methodologies for predicting cleavage arrest of a deep crack in a reactor pressure vessel subjected to pressurized-thermal-shock loading conditions

Several calculational procedures are compared for predicting cleavage arrest of a deep crack in the wall of a prototypical reactor pressure vessel (RPV) subjected to pressurized-thermal-shock (PTS) types of loading conditions. Three procedures examined in this study used the following models: (1) a static finite-element model (full bending); (2) a radially constrained static model; and (3) a thermoelastic dynamic finite-element model. A PTS transient loading condition was selected that produced a deep arrest of an axially oriented, initially shallow crack according to calculational results obtained from the static (full-bending) model. Results from the two static models were compared with those generated from the detailed thermoelastic dynamic finite-element analysis. The dynamic analyses modeled cleavage-crack propagation using a node-release technique and application- and generation-mode methodologies. Comparisons presented here indicate that the degree to which dynamic solutions can be approximated by static models is highly dependent on several factors, including the material dynamic fracture curves and the propensity for cleavage reinitiation of the arrested crack under PTS loading conditions. Additional work is required to develop and validate a satisfactory dynamic fracture toughness model applicable to postcleavage arrest conditions in an RPV