SOME IMPLICATIONS OF THE USE OF THE PD6493 LEVEL-3 FAILURE ASSESSMENTPROCEDURE

Citation
Zh. Li et al., SOME IMPLICATIONS OF THE USE OF THE PD6493 LEVEL-3 FAILURE ASSESSMENTPROCEDURE, Fatigue & fracture of engineering materials & structures, 18(12), 1995, pp. 1499-1509
Citations number
16
Categorie Soggetti
Material Science","Engineering, Mechanical
ISSN journal
8756758X
Volume
18
Issue
12
Year of publication
1995
Pages
1499 - 1509
Database
ISI
SICI code
8756-758X(1995)18:12<1499:SIOTUO>2.0.ZU;2-1
Abstract
Under a natural mapping between the standard R-curve analysis diagram and the failure assessment diagrams of R6 and PD6493 Level 3 the R-cur ve becomes the RCI (R-curve image). It follows that whenever the asses sment point moves along the failure assessment line during ductile cra ck growth, the implication is that the failure assessment line is the RCI. This result is used to test the conservatism of a specific PD6493 Level 3 analysis by two methods. The first calculates the variation d uring crack growth of the applied elastic-plastic crack tip opening di splacement (CTOD) parameter (or 'driving force') which is implied by t he PD6493 analysis and then compares this variation with an independen t estimate of it. The second uses an assumed driving force to deduce t he CTOD resistance curve implied by the failure assessment line. It is shown by both methods and also by a direct R-curve analysis that this particular PD6493 analysis is conservative relative to an R-curve ana lysis which uses a crack driving force estimated by the EPRI (Electric Power Research Institute) procedures. However there is inconsistency between standard R-curve analysis and PD6493 Level 3 analysis in that the latter implies a material resistance which, for a given material, depends on the geometry of the structure. A similar inconsistency aris es in any failure assessment procedures like the Options 1 and 2 with Category 3 of R6, which require the assessment point to move on a geom etry independent failure assessment line during crack growth; indeed, even when the failure assessment line is geometry dependent there is f ull agreement with R-curve analysis only if the correct RCI is used as the failure assessment line. In a brief discussion it is noted that t he new failure assessment diagram studies involving multi-parameter fr acture mechanics may help to ameliorate these problems.