Sensitivity of Stress-Strain Parameters in Leak-Before-Break Evaluations

Leak-before-break (LBB) evaluations involve the use of deterministic fracture mechanics to establish the margin between critical and leakage flaw sizes in order to assure that leaks can be detected by the plant leak detection system before a through-wall flaw reaches critical flaw size. When the material is semi-ductile, the fracture mechanics evaluations involve the use of elastic-plastic fracture mechanics (EPFM) consisting of the J-integral and tearing modulus (J/T) analyses. An important input into the J/T analyses is the Ramberg-Osgood (R-O) material stress strain parameters which describe the stress-strain curve of the material of interest. These are also key inputs in the determination of leakage associated with through-wall flaws. If the stress-strain curve of the material of interest is available, the R-O parameters can be determined from power law curve fit. However, in most cases, archival material of existing plant piping is not readily available to determine the actual stress-strain curve. In the absence of the actual stress-strain curve, several approximate methods for determining the R-O parameters from basic mechanical properties have been proposed in the literature. These approximate methods however produce different sets of R-O parameters. In this paper, the effect of using different sets of R-O parameters from three R-O formulations on LBB analyses results is investigated. EPFM analyses are performed to determine the critical through-wall flaw lengths with the various sets of the R-O parameters for various materials and various pipe sizes. The same sets of parameters are then used to determine the leakage associated with through-wall flaws. The results of the evaluation indicate that different sets of R-O parameters can yield different critical flaw sizes as well as leakage flaw sizes, thus resulting in different margins in LBB evaluations. Considering the margins involved in LBB evaluations (factor of two on critical flaw size and factor of 10 on leakage), it is believed that these differences are small enough that any of the three correlations presented in this paper for determination of the R-O parameters can be adequately used in LBB evaluations employing EPFM analyses.Copyright © 2007 by ASME