Abstract
To investigate the equivalence between fracture toughness data obtained from MC(T) and larger specimens, more 600 fracture toughness tests were performed on six (four base and two welds) different RPV relevant steels in irradiated and unirradiated state. To this experimental round robin, 14 different laboratories
participated. In addition, numerical models were implemented in finite element models to support the experimental tests.
For all base materials agreement between data obtained from MC(T) and large specimens is satisfactory when accounting for macroscopic inhomogeneities in the steel (A508 Cl.3 and A533B JRQ). For weld materials, the analysis is more subtle: while agreement between MC(T) and large specimen data is satisfactory for ANP-5 when accounting for material heterogeneity, this is not the case for the 73W weldment, both in irradiated and unirradiated state.
For weldments, the scale of inhomogeneity in a weld was such that a 1T-C(T) specimen would sample across several weld beads, whereas a MC(T) specimen would sample less than a single weld bead. This suggests that metals and especially weld metals can exhibit microstructural inhomogeneity on a scale that significantly affects toughness testing. Both weld and base metals are likely to contain multiple constituent regions of substantial size such as recrystallized and as-deposited regions of a weld bead or islands of ferrite, bainite or martensite that reflect the dendrite spacing typical of castings. Each region contains a distribution of weak links which may or may not overlap with those in other regions. This underscores the necessity of testing a larger number of specimens to ensure an accurate assessment of inhomogeneity.
Finally, we note that all MC(T) complete data sets are inhomogeneous, regardless of the homogeneity or inhomogeneity in the reference data set. The small sampling volume combined with the weakest link theory interpretation means that local differences in the material may result in a large variation from one tested MC(T) specimen to another. Thus, rather than MC(T) specimens misinterpreting homogeneous materials as being inhomogeneous, MC(T) specimens are more adept at identifying inhomogeneity than larger specimens.
This last point should be taken as a strength of the use of MC(T), i.e., the capability to sample material properties very locally if desired. When a complete block or plate needs to be characterized, MC(T) specimens need to be collected from various locations within the block or plate in the same manner as with larger specimens.
Overall, fractography supports the use of toughness data from mini-CT specimens, if the tests are valid according to ASTM E1921-21. Ductility at edges of specimens does not dominate nucleation and the asymmetry of crack front does not dominate nucleation either. The fracture mode is likely to be unaffected by specimen size. A concentration of initiation sites in the central half of the specimen is observed due to the missing side grooves. However, since the relation between 𝐾jc and the distance to the initiation site is unaffected by specimen size, the stress and strain conditions at fracture initiation will be the same in the miniature and larger specimens.
participated. In addition, numerical models were implemented in finite element models to support the experimental tests.
For all base materials agreement between data obtained from MC(T) and large specimens is satisfactory when accounting for macroscopic inhomogeneities in the steel (A508 Cl.3 and A533B JRQ). For weld materials, the analysis is more subtle: while agreement between MC(T) and large specimen data is satisfactory for ANP-5 when accounting for material heterogeneity, this is not the case for the 73W weldment, both in irradiated and unirradiated state.
For weldments, the scale of inhomogeneity in a weld was such that a 1T-C(T) specimen would sample across several weld beads, whereas a MC(T) specimen would sample less than a single weld bead. This suggests that metals and especially weld metals can exhibit microstructural inhomogeneity on a scale that significantly affects toughness testing. Both weld and base metals are likely to contain multiple constituent regions of substantial size such as recrystallized and as-deposited regions of a weld bead or islands of ferrite, bainite or martensite that reflect the dendrite spacing typical of castings. Each region contains a distribution of weak links which may or may not overlap with those in other regions. This underscores the necessity of testing a larger number of specimens to ensure an accurate assessment of inhomogeneity.
Finally, we note that all MC(T) complete data sets are inhomogeneous, regardless of the homogeneity or inhomogeneity in the reference data set. The small sampling volume combined with the weakest link theory interpretation means that local differences in the material may result in a large variation from one tested MC(T) specimen to another. Thus, rather than MC(T) specimens misinterpreting homogeneous materials as being inhomogeneous, MC(T) specimens are more adept at identifying inhomogeneity than larger specimens.
This last point should be taken as a strength of the use of MC(T), i.e., the capability to sample material properties very locally if desired. When a complete block or plate needs to be characterized, MC(T) specimens need to be collected from various locations within the block or plate in the same manner as with larger specimens.
Overall, fractography supports the use of toughness data from mini-CT specimens, if the tests are valid according to ASTM E1921-21. Ductility at edges of specimens does not dominate nucleation and the asymmetry of crack front does not dominate nucleation either. The fracture mode is likely to be unaffected by specimen size. A concentration of initiation sites in the central half of the specimen is observed due to the missing side grooves. However, since the relation between 𝐾jc and the distance to the initiation site is unaffected by specimen size, the stress and strain conditions at fracture initiation will be the same in the miniature and larger specimens.
| Original language | English |
|---|---|
| Publisher | FRACTESUS |
| Number of pages | 36 |
| State | Published - Dec 2022 |
Publication series
| Name | FRACTESUS Reports |
|---|---|
| Publisher | FRACTESUS |
| No. | FRACTESUS D7.2 |