Probabilistic analysis of fatigue crack initiation and growth in reactor pressure boundary components

F. A. Simonen, D. O. Harris, M. A. Khaleel, D. Dedhia

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Procedures for the probabilistic analysis of the initiation of fatigue cracks in pressure boundary components of commercial light water reactors (LWRs) are described. This analysis is then combined with a probabilistic fracture mechanics analysis to predict the probability of such initiated cracks growing to become through-wall cracks, thereby causing leaks or catastrophic breaks. The crack initiation procedures are based on data analyses performed by Argonne National Laboratory (ANL) using data from fatigue tests of pressure boundary materials in LWR environments. The ANL work provided relations giving the probability of crack initiation as a function of the number of cycles for a given cyclic stress. The ANL correlations included the influence of the strain rate, sulfur content, oxygen content of the reactor water and temperature. The probabilistic fracture mechanics model of the pcPRAISE code was modified to account for crack initiation and the possible linking of fatigue cracks at multiple initiation sites. Example calculations are presented to show the effects of crack initiation and crack linking on the probabilities of through-wall cracks and the relative probabilities of large versus small leaks. The results obtained are of use in assessments of risk contributions of components with high cyclic usage and the risks posed by extending the lifetime of the components beyond their initial design life.

Original languageEnglish
Pages (from-to)239-247
Number of pages9
JournalAmerican Society of Mechanical Engineers, Pressure Vessels and Piping Division (Publication) PVP
Volume410
StatePublished - 2000
Externally publishedYes

Fingerprint

Dive into the research topics of 'Probabilistic analysis of fatigue crack initiation and growth in reactor pressure boundary components'. Together they form a unique fingerprint.

Cite this