In this study the validity of various analyses in the prediction of crack initiation, maximum load, and the corresponding displacements for 168-mm (6.625-in.)-diameter circumferentially complex-cracked pipe under pure bending was assessed. The results of six pipe fracture experiments on three materials (SA-376 TP304 stainless steel, Inconel 600, and A106 Grade B carbon steel) were used to verify the accuracy of these analytical predictions. All experiments were conducted at 288C (550F).

Two different sets of analyses were conducted using J-R curves and J-estimation schemes. J-resistance (J-R) curves were calculated from each pipe experiment using the -factor method. These results revealed that the J-R curves from the complex-cracked pipe experiments were significantly lower than J-R curves from 0.5T compact tension specimens. Furthermore, the pipe J-R curves decreased systematically with increasing ratios of surface crack depth to pipe wall thickness.

Predictions of loads and displacements in each experiment were made using four different J-estimation schemes. Good agreement was obtained between the predictions and the experimental data up to maximum load. Once past maximum load, however, the J-estimation schemes overpredicted loads and displacements.