In this study a probabilistic Leak-Before-Break (LBB) analysis was carried out based on the R6 FAD Option 1 assessment method. The method uses the material fracture toughness and yield stress in order to determine, deterministically, a Critical Crack Length (CCL) and a Leakage Rate (LR) through a crack. In order to define the fracture toughness of the material, the Master Curve approach was used accordingly to BS7910:2013 Annex J. Initially, deterministic analyses were carried out and the fracture toughness and yield stress were set to 190 MPa√m and 158 MPa, respectively. In order to implement a probabilistic approach, the yield stress and fracture toughness were introduced as stochastic parameter. The Fracture toughness was generated using a Weibull distribution to match the Master Curve. The distribution was built such that 190 MPa√m represents the 5% probability fracture toughness. The Yield stress (0.2% proof strength) was generated using a normal distribution with standard deviation 10.35 MPa such that the average value was 175 MPa and the lower bound (5% of probability of occurrence) was 158 MPa. The choice of building the distribution as above mentioned was justified by the fact that in structural integrity assessment the lower 5% is generally used for material parameters. Thus, once a Detectable Leakage Rate (DLR) was determined, it was possible to assign an implicit probability of failure to the deterministic case. The calculations were then extended by using several LR formulas. The calculations were carried out making use of the probabilistic software RAP++ coupled to MATLAB. The probabilities of failure were calculated with regard to a postulated DLR and a DLRSF corrected with a safety factor of 10. The probabilities of failure for the DLRSF were proved to be 9 to 15 times higher than for the postulated DLR case, which leads to the opportunity of conservatism reduction.

This content is only available via PDF.
You do not currently have access to this content.