Cross-Weld Tensile Properties of Girth Welds for Strain-Based Designed Pipelines

For strain-based designed pipelines it is important to understand the relative tensile properties of both weld metal and pipe material in the pipe axial direction. The evaluation of weld metal tensile properties has typically involved extracting all-weld-metal tensile samples in the direction of the weld. In this study an evaluation of the application of “waisted” tensile samples to generate data has been conducted. Initial studies focused on finite element analysis to generate geometry factors for a range of specimen configurations to correct for the level of stress triaxiality. These factors were then applied to samples extracted from X70 and X100 pipe material to establish the validity of this approach. It was shown that, regardless of the radius of waisted specimens, very good agreement was obtained between the geometry-factor-corrected stress-strain curves and those generated from standard test specimens at true strains above 0.02. To achieve a better agreement between the corrected and standard tensile curves in and around yield it was necessary to use samples with a large radius (9 mm) where the stress concentration was low. Finite element analysis provided supporting evidence with respect to the effect of stress concentration associated with different specimen radii on the yielding pattern. These waisted samples were used to measure the tensile properties in all-weld-metal and cross-weld-metal directions for an X70 double joint (DJ) weld and an X100 mechanized pulsed gas metal arc (P-GMA) weld. Waisted samples taken from the double joint weld on X70 with radii of 3 mm and 9 mm showed no difference with respect to their orientation. Once stress-strain behaviour was corrected for geometry, the curves were in excellent agreement with the standard test specimens above 0.01 true strain in the case of the sample of radius 3 mm and for the whole curve for the sample with radius 9 mm. An assessment of the X100 weld identified a small difference between all-weld-metal and cross-weld-metal directions, with the latter displaying a lower stress between yielding and 0.03 true strain. The use of waisted samples of larger radius generated much better agreement with the standard specimens associated with their lower stress concentration. Because of the finite weld width, consideration needs to be given to the extent to which the reduced section may extend beyond the weld and the potential effect of mismatch in strength.