Benchmark Analyses for Fracture Mechanics Methods for Assessing Sub-Clad Flaws - NESC-VI Final Report

摘要

The sixth project of the Network for Evaluating Structural Integrity (NESC-VI) deals with the fracture mechanics analysis of a set of 3 tests on beam specimens with simulated sub-surface flaws, which were performed by NRI Re¿ plc for the PHARE project ¿WWER Cladded Reactor Pressure Vessel Integrity Evaluation (with Respect to PTS Events)¿. The objectives were as follows:¿ to assess the capability to predict whether the cracks propagating into the cladding arrest or cause fullfracture, and¿ to assess the capability to predict the location of first initiation: near-surface or deep crack tip.The project was launched in December 2006 and completed in March 2009. It brought together a group of 10 organisations from NESC to perform comparative analyses of selected tests, based on a comprehensive datapack prepared by NRI. The investigations focussed almost exclusively on assessing the capability to predict the location of first initiation. The main results are as follows:¿ Comparison of analyses performed by individual partners showed that the FE simulations produced consistent predictions of the observed force vs. load-line displacement (or crack mouth opening displacement) behaviour. However the differences in predicted crack tip stress intensity, KJ, as a function of applied loading were greater than those found in similar intercomparisons made as part of previous NESC projects. This underlines the importance of periodically performing such exercises.¿ The influence of two modelling factors on KJ was clearly established: firstly for this type of specimen, for which the clad makes up almost 12% of the cross-section, the associated residual stresses have a significant effect in reducing KJ values and therefore need to be considered in "best-estimate" analysis. The second concerns the use of 2-D or 3-D models: in this case the 2D FE models underestimated KJ values and are considered non-conservative.¿ For this combination of test specimen geometry and flaw, constraint loss is expected at the near-surface tip. A range of constraint parameters were evaluated (elastic T-stress, elastic-plastic T-stress and Q) to confirm this. However only in two cases these were used in quantitative analyses: constraint-modified FAD and KIeff, both using elastic T-stress. These indicate that fracture is likely to initiate at lower (deep) tip, which is consistent with the limited high-speed video camera evidence. In general more systematic application of 2-parameter approaches is needed.¿ Both local approach models predicted initiation of cleavage fracture first from the lower crack front for medium and higher loads.Concerning the capability to predict whether the cracks propagating into the cladding arrest or cause full fracture, the two analyses performed indicate that when the load at first pop-in is low, crack arrest in the clad can be correctly predicted on the basis of the J-R curve, but that further work is needed to ensure the reliability of such approaches over the full load range.