One of the large scale mock-up experiments (Mock-Up 2) of the recent STYLE European project was a ductile fracture test carried out on a pipe containing a circumferential through-wall crack located in a repair weld and loaded under four-point bending. This experiment has been used as a basis for undertaking a comparative probabilistic fracture mechanics benchmark study by a number of the organisations participating in STYLE.
In launching the benchmark study, the pipe and crack dimensions were specified as fixed values, applied bending stress (moment) was considered as a deterministic variable and three values were taken for weld residual stress (assuming constant membrane values). Tensile and fracture toughness material properties were considered to be random variables and these were specified by log-normal distribution parameter values. In undertaking their evaluations, the participants were asked to evaluate J-integral versus applied bending moment, probability of crack growth initiation versus applied bending moment, probability of net-section collapse versus applied bending moment and carry out various sensitivity analyses.
This paper focuses mainly on the contribution to the benchmark study primarily based on the R6 methodology and includes examples of contributions provided using other methods.
