PR-185-173600-R01 Reference Stress for Metal-loss Assessment of Pipelines

This project focused on quantifying the reference stress to be used in predictive models for assessing the effects of metal loss on pipeline integrity. The results of this project will work in concert with the outcomes of project EC-2-7 that examined sources of scatter in metal-loss predictions with respect to the metal-loss defect geometry. The methodology for developing a new reference stress included empirical and finite element analyses along with comparison of full-scale experimental results that indicate the failure behavior of defect-free pipe has dependence on the strain hardening rate, n, of the pipe steel. Since the strain hardening rate is often unreported in qualification test records and mill certification reports, the development of a new reference stress will seek to include the utilization of the ratio of yield-to-tensile strength (Y/T) as a surrogate for n. This approach ideally would be insensitive to pipe grade, and thus, allow broad application of the reference stress without increasing scatter or bias across grade levels. 

This work also compared the resulting metal-loss criterion with the new reference stress relative to the B31G and Modified B31G models using a dataset of approximately 75 full-scale burst test results for test vessels containing isolated defects. This comparison was performed by C-FER Technologies under sub-contract to EWI and quantified the prediction bias and prediction variability of the new criterion relative to those widely in use.