Recent J-R curve testing of A302-B plate has shown significant specimen size effects. If the J-R curves of a material are dependent on size and the larger the size, the lower the J-R curve, serious questions as to the usability of J-R curve information from nuclear reactor pressure vessel surveillance programs could arise. It is important, then, to be able to identify the existence of this phenomenon from J-R curve test records and either correct the data for size effects or account for this size effect in failure assessments.
Other studies, such as the HSST irradiation work by ORNL and the EPRI-sponsored work by Westinghouse, have demonstrated mixed results. From a review of all these test results, it appears that there are two causes of the observed specimen size effects, one due to loss of constraint in the small specimen and the second due to metallurgical inhomogeneities. One of the objectives of this investigation was to develop procedures to identify if either of these conditions exists.
Several evaluation models have been applied to determine their usefulness in identifying these size effects. The models investigated were: 1. The key curve approach. 2. DPFAD (deformation plasticity failure assessment diagram) approach. 3. Load-displacement prediction based on the EPRI handbook.
The following data were analyzed: 1. HSST Irradiation Tasks II and III weld metal compacts. 2. Westinghouse/EPRI RP 1238-2 A508 0.5Tto 10Tcompacts. 3. A302-B plate tests by MEA.
In order for an evaluation method to be useful, the methodology must be able to identify test results in which the specimen size influences the J-R curve. By using a number of different evaluation schemes on test data where a range of sizes and where size effects have been observed, the effectiveness of the different methodologies can be assessed.
The results of the key curve, DPFAD, and load-displacement methodology predictions were compared to the test data. When good agreement is seen in one procedure, good agreement is seen in all procedures. Likewise, when one prediction/comparison looked bad, all the predictions were bad. However, the DPFAD approach seemed to better highlight the extent of disagreement than was shown in the other two procedures.
Any one of the procedures investigated could be used to demonstrate that a size effect is present in a data set. The DPFAD approach seems to best identify the test results from compact specimens which are not J-controlled and can be used when only one specimen size has been tested. The lack of J-control may be due to incorrect crack measurements during the test, metallurgical anomalies, or the loss of specimen constraint. The key curve can be used to identify J-R curve test results within a set which differs from others in the set. It, however, does not appear useful in identifying specimen size effects if only one specimen has been tested. The load-displacement procedure can be useful if a specimen exhibits plane strain behavior, but it is not as sensitive as a measure of loss of J-control growth as DPFAD.
The second objective of the investigation was the development of remedial actions to correct J-R curve data in which specimen size effects are evident. The DPFAD approach was used to demonstrate that the various size effects seen with the A302-B plate material could be corrected. The assessment points from each compact specimen which deviated from the DPFAD curve were recalculated by increasing the amount of ductile tearing and adjusting the calculated J-R point via ASTM E 1152-87 test standard such that these assessment points fall on the DPFAD curve. This has been shown to be equivalent to /-controlled crack growth up to the limit load of the specimen for displacement-controlled loading.
Conclusions are that for the A302-B plate material, the l/2Tand 1 T specimens may not be large enough to produce usable J-R curves for this material.