This paper explores the basis for some of the current criteria for assessing the effect of metal-loss defects on the integrity of pipelines, with particular reference to important trends in the database of full-scale experiments on corroded pipe sections. Critical trends are identified as the foundation for an alternative to current integrity criteria for metal-loss defects. Finally, this paper outlines an approach to formulate an alternative criterion based on the limiting response of metal-loss defects observed in field and full-scale laboratory experiments. Part II of this paper develops this approach and presents numerically generated trends from parametric analysis that fix the functional form and the fitting-constants for this alternative criterion. From these trends a simple criteria for metal-loss defects is presented and demonstrated to be accurate based on comparisons to results from full-scale tests of end-capped pipe sections. Conclusions from Part I of this paper relate to the formulation of the alternative criterion and its practical application, and include the following points. The use of the actual yield stress in place of SMYS in B31G and other similar or related criteria offers little improvement in either their accuracy or consistency. In contrast, the uniaxial ultimate tensile stress in place of SMYS in B31G and other similar criteria consistently correlates data from burst tests of defect-free end-capped pipes with little or no scatter. Circumferentially aligned (coparallel) adjacent defects cause a local increase in the compliance of the pipe wall, which explains their failure at pressures greater than that for a geometrically similar isolated defect. Axially aligned (colinear) interacting defects have an equivalent defect length greater than that of the longest of the individual defects, leading to a reduction in failure pressure.

This content is only available via PDF.
You can access this article if you purchase or spend a download.