In the framework of the R6 development programme on leak before break NRG developped a probabilistic fracture-mechanics model for analyzing circumferential through-walled cracked pipe made of Type 304 stainless steel, subjected to bending loads. An elastic-plastic analysis has been carried out using ANSYS for estimating J-integral as a function of moment load. The validity of the J-integral based on the ANSYS calculations was evaluated by comparison with LBB.ENG2 method and Rahman’s calculations.
Probabilistic fracture analysis has been carried out using ANSYS Probabilistic Design System (PDS) module to find out the failure probability of a pipe as a function of applied moment. These results have been compared with LBB.ENG2 probabilistic calculations, which has been developed using MATLAB.
To achieve high efficiency, accuracy and robustness to design structural component with a low probability of failure, Advanced sampling methods (ADIS) have been used for probabilistic calculations. These ADIS results have been compared with Monte Carlo probabilistic results.
The probabilistic method is subsequently extended to a Leak Before Break case (LBB). It is demonstrated that the probability of failure reduces when more probabilistic data for input parameters is added instead of using deterministic safety factors.
