PSI - Issue 42

Aleksander Omholt Myhre et al. / Procedia Structural Integrity 42 (2022) 935–942 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

941

7

further increase in hydrostatic stresses and leading to higher local hydrogen concentrations, can explain the difference in the measured EI values compared to the vintage counterparts. Another factor pivotal for the mechanical response both in air and in hydrogen is the microstructural anisotropy presented in the vintage steels due to the hot rolling manufacturing. Evidence of the effect of the elongated grain morphology in the rolling direction and the harder bands consisting of pearlite and bainite, which can lead to higher constraints in the deformation, particularly during necking can be found in the plot in Figure 5. In fact, it seems evident that the anisotropy in the fracture surfaces is reduced in hydrogen compared to in air. However, in Material B, parallel ridges can be seen on the fracture surface, where mixed ductile and brittle fracture features are present. These could also be related to the microstructural bands, where high diffusivity ferrite meets low diffusivity pearlite and bainite (Ogawa et al. 2021). By comparing the post-mortem fracture surfaces morphologies in air, the differences between the Materials relate to the asymmetric shape present in the vintage steels due to the previously discussed microstructural anisotropy. The surfaces consist of dimples, where the size and shape are somewhat different, most probably being related to the grain size. Some differences are visible at the macroscale for the fracture surfaces of the specimen tested in hydrogen: due to the ridged appearance in the banded, vintage pipes, and a wedge-shaped region with flatter topography perpendicular to the bands. At the higher magnification, the same types of mixed dimpled and quasi-cleavage features are present in all materials near the outer edges of the specimen, while the centre remains fully ductile. Since the in situ charging conditions allowed for a continuous supply of hydrogen, the ductile centre likely corresponds to the final stage of the rupture, whose speed is too rapid for hydrogen supply to keep up and exert any influence. The larger nominal strain seen in the curves of Material C versus Material B during necking may indicate that the growth and propagation of hydrogen-induced surface micro-cracks occurred more rapidly in Material B, resulting in a greater loss of the net-section than predicted from the nominal strains obtained from the displacement data. Such a hypothesis is supported by the presence of the surface side cracks, which are more "open" in Material C, thus indicating a greater degree of crack tip blunting. The cause of the proposed increase in micro-crack propagation in the vintage materials could be related to a combination of increased hydrogen accumulation at the boundaries between ferrite matrix and the pearlite bands, and the increased constraints in the direction of deformation due to the anisotropic properties, together providing more favourable crack propagation paths. Based on the quantitative EI values obtained, Material B and Material C perform the overall worst and best, respectively, in terms of the resistance to hydrogen degradation. However, care should be taken in using these results for the evaluation of in-service suitability. Environment (Henthorne 2016), strain rate (Margot-Marette, Bardou, and Charbonnier 1987) and the specimen design and surface finish (Le Hong 2001) are among the factors which can impact the quantitative results of SSRT. Assessment based on the quantitative hydrogen degradation values obtained from SSRT tests to design guidelines in other loading conditions is not straightforwardly valid, as exemplified in the amendment of the ASME B31.12 code due to the similar fatigue crack growth rates of X52 and X70 steels, despite the latter having a higher hydrogen susceptibility based in part on SSRT testing (Amaro et al. 2018). 5. Conclusions • All the materials tested revealed considerable loss in ductility when tested in hydrogen, while the negligible hydrogen effect was recorded on the strength. • Material B, position 2, showed the highest hydrogen susceptibility as assessed through the embrittlement index, while the modern Material C was the least affected by the exposure to hydrogen. • Fracture surface examinations in hydrogen consistently unveiled lateral surface cracks both in the necked and uniaxially stresses material zones regardless of the materials and positions tested. The quasi-cleavage facets and secondary cracking represent the most typical features on the fracture surfaces; however, the centre of the cross section remains dimpled. • The main difference in the magnitude of hydrogen degradation effects is likely related to the larger amount of impurities and the presence of pearlitic-bainitic banding in the vintage pipelines compared to the homogenous modern material. • The present work focused on the slow strain rate testing, the results of which were used as part of a screening process aimed to identify the most suitable materials for the comprehensive characterization of their response,