PSI - Issue 22

Abdelkader Guillal et al. / Procedia Structural Integrity 22 (2019) 201–210 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

208

8

Fig 6.Variation of reliability index for two limit stat function

According to this figure, results obtained from LSF 2 using R6 curve show more conservatism than those given by LSF 1 . The factor C R6 reflects the potential interaction between plastic collapse and brittle fracture; therefore, this factor is a function of crack dimensions with value less than one. D - Sensitivity analysis The last part of this study is dedicated to investigate the importance of the deferent random variable included in both limit state functions that represent failure modes. The sensitivity analysis was conducted for case 4 ( = −0.997) . Figures 7 and 8 plot the sensitivity analysis of reliability index to mean values of random variables.

Fig 7. Evolution of sensitivity to mean value of random variables for LSF 1

From results mentioned in Fig 7, it is obvious in one hand that random variables K ic and t have the largest positive influence on reliability index results based LSF1 . Hence they are resistance variables and contribute to the resistance part of the limit state function. In the other hand, Paris law parameters (C. m) have the largest negative influence. Thus, they will contribute to load part of the limit state function. Sensitivity analysis with respect to LSF 2 is presented in Figure 8. In addition to toughness and pipe thickness, the yield strength of steel plays a secondary role in the resistance side of LSF2 . While in the stress side of LSF2 , the dominant variable is the pipeline diameter followed by crack length and Paris law parameters .