L51488 Effects of Stress Relief Due to Hydrostatic Testing on Girth Weld Failure

Fitness-for-purpose analyses are being increasingly used in the engineering assessment of the safety of gas pipelines. The potential need for this arises because of the likelihood of a defect being present in the girth weld. Many of these analyses make an allowance for the residual stress caused by the weld, whilst others do not. It has been argued that if a defect survives the field hydrostatic test that the residual stress distribution will be altered and this could subsequently affect the tolerable defect size analysis. Four full-scale tests were carried out, two at each wall thickness; one of which was as-welded and one hydrostatically tested. In order to simulate the field hydrostatic test a novel laboratory technique was developed that minimized the end load during pressurization. The results of the full-scale tests, that failed by ductile tearing, show that for similar defect sizes the strain to failure is the same whether in the as-welded or hydrostatically tested condition. The strain to initiate tearing, however, is quite different. In the case of the 24" (610 mm) thin wall tests the strain to initiate tearing increases from 0.09% (as welded) to 0.17% (hydra test). For the 36" (914 mm) thick wall tests the equivalent figures are 0.11% and 0.12%. These results are consistent with the radial deflection measurements which predicted a much higher axial residual strain for the thinner wall pipe material. The results, do point out the effect of the residual stress oncleavage failure. The overall results are consistent with the various Engineering Critical Assessment techniques, however, the approach of CSA Z184 may prove to be more appropriate. This approach makes no allowance for residual strain for either brittle or ductile failure.