L51580 Safety Factors in the Assessment of Realistic Defects in Pipeline Welds

The British Standards Institute document PD6493:1980 uses the crack tip opening displacement (CTOD) design curve as a method for determining defect acceptance levels in welded joints. The document is under revision and current proposals for the fracture section of the revised document employ three levels of assessment procedures, arranged as a progression from simple stress treatments producing conservative results, to critical analyses which incorporate more rigorous stress analyses and contain no inherent factors of safety. The proposed Level 1 assessment procedure is generally consistent with the CTOD design curve methodology of the current document. Changes include a more simplified treatment of stresses and a plastic collapse analysis. Level 1 is intended as a preliminary screening method before moving to the higher-level procedures as appropriate. Level 2 is seen as the most appropriate assessment level for the majority of industrial applications. It employs a collapse modified strip yield model, and allows for more complex stress analysis treatments than Level 1. These treatments take account of the plasticity interaction between primary and secondary stresses. Level 3 provides the most advanced assessment route. It is based on a reference stress approach and in general requires development of accurate stress-strain data for a particular material. Its main areas of application are for materials exhibiting a high work hardening characteristic or for tearing instability analyses.Result: Experimental work was carried out to generate data against which to validate the numerical models used in the subsequent theoretical investigation of safety factors inherent in defect assessment procedures. It is considered appropriate to examine the proposed Levels 1, 2 and 3 assessment procedures in preference to the CTOD design curve of the current document since this three-tier structure will soon supercede the PD6493:1980 approach. The similarities between the Level 1 methodology and the design curve, and also between the Level 2 approach and the previously investigated collapse modified strip yield formulations allow a systematic investigation of any relative change in safety factors between the old and the new methods.Benefit: The validation studies demonstrated that numerical models using true stress-logarithmic strain data gave reasonable predictions of the experimental global displacement and local crack tip response. Plane stress and plane strain succeeded in providing lower and upper bounds respectively to the experimental load-displacement behavior for plain and welded configurations. Plane stress conditions were considered the most suitable for the theoretical safety factor investigation. Level 1 assessment was found to be generally conservative but safety factors less than unity could occur for defects located in severe residual stress fields under high levels of load. Level 2 and Level 3 assessments produced consistent levels of safety when realistic defects at welds were considered which included over-matching weld metal and weld cap reinforcement.