Comments on probabilistic fracture mechanics analysis of cracked components

Nam Su Huh, Do Jun Shim, Yun Jae Kim, Kyu Ho Lee, Young Jin Kim

Research output: Contribution to journalConference articlepeer-review


This paper investigates the effect of elastic-plastic J estimate on calculated failure probabilities for cracked components. To investigate effect of elastic-plastic J estimate on failure probabilities, firstly circumferential through-wall cracked pipes under pure bending is considered, which is important in the leak-before-break analysis. For the analysis, scatters in the applied load, mechanical and fracture properties are considered using normal and log-normal distributions. More importantly, the applied elastic-plastic J-integral is estimated according to four different estimation methods. Resulting failure probabilities are calculated using the efficient first-order reliability method (FORM). The results show that failure probabilities can be significantly different, depending on the elastic-plastic J estimate, suggesting that absolute failure probabilities might have less significance. Based on these investigations, the accuracy of elastic-plastic defect assessment method embedded in API RP579, popularly used in petrochemical industry for integrity assessment of defective structures, is systematically investigated based on the 3-D finite element analyses. The results show that the J estimate in API RP579 can be highly inaccurate for some cases, which in turn would affect failure probabilities.

Original languageEnglish
Pages (from-to)1755-1760
Number of pages6
JournalKey Engineering Materials
Issue numberIII
Publication statusPublished - 2004
EventProceedings of the 11th Asian Pacific Conference on Nondestructive Testing - Jeju Island, Korea, Republic of
Duration: 2003 Nov 32003 Nov 7


  • First Order Reliability Method
  • Leak-Before-Break
  • Monte Carlo
  • Probabilistic Fracture Mechanics
  • Reference Stress

ASJC Scopus subject areas

  • General Materials Science
  • Mechanics of Materials
  • Mechanical Engineering


Dive into the research topics of 'Comments on probabilistic fracture mechanics analysis of cracked components'. Together they form a unique fingerprint.

Cite this