The Effect of Shallow Cracks on the Predicted Failure Pressure of Natural Gas Pipelines Containing Stress-Corrosion Cracks

Predicting the failure pressure from the dimensions of a defect in a pipe is a critical step in direct assessment and in-line inspection. There are reasonably accurate methods for regularly shaped defects, but predicting the failure pressure for stress-corrosion cracks is more difficult because of the irregular crack front and the existence of shallow cracks or shallow extensions of deeper cracks. This paper addresses the issue of shallow cracks. This is particularly important in the case of a very shallow crack that is long enough to be classified as Category 1, but is so shallow that is does not affect the strength of the pipe. It is common practice to assume that a Category 2 crack exists some place in that segment, and a hydrostatic test would be required, although such drastic action should not be necessary because of the negligible effects on the strength of the pipe. It has been suggested that grinding 10% of the wall thickness might be a useful way to define the effective length of cracks or cracks that have shallow extensions. The effect of removing shallow cracks by grinding was investigated by analyzing twenty-five cases for which the crack profile was known. The effect depends on how the crack depth profile is characterized and whether there is a neighboring crack nearby that might interact with the flaw of interest. It was found that grinding 10% off the wall thickness will affect a small percentage of the flaws (five of the twenty-five). For those that it does affect, if the crack depth is taken as the maximum depth, the predicted failure pressure (PFP) will be very conservative, and eliminating the shallow cracks will always make the PFP less conservative, but still more conservative than desired. If this method is used, decreasing the severity category by one should be considered.