PSI - Issue 68

Robert Sundström et al. / Procedia Structural Integrity 68 (2025) 1081–1090 Robert Sundström / Structural Integrity Procedia 00 (2025) 000–000

1087

7

investigators that did not report differences used large wall thickness, 2.5 to 3.5 mm. Where differences in fatigue life were found, they were 1.4 to 1.6 times lower for the hollow specimen than the solid. A finite element analysis (FEA) was done comparing solid and hollow specimens in tensile testing, where it was seen that hollow specimens had elongations that were smaller as the inner diameter increased (Ogata and Ono 2019). An outer diameter of 6.25 mm and inner diameter of 0 (solid specimen), 1, 2 and 3.5 mm (hollow specimens) was used, giving diameter ratios of 0.16, 0.32 and 0.56 for the hollow specimens. At an internal pressure of 100 MPa, the specimen with the largest inner diameter (3.5 mm) exhibited approximately half the elongation of the specimen with the smallest inner diameter (1 mm), which was nearly equal to the elongation of the solid specimen. Elongations were larger when no internal pressure was applied. In summary, these studies point towards wall thickness effects on low cycle fatigue life when hollow specimens are tested in argon, air or reactor water environments. An effect of wall thickness is also seen on elongation in tensile testing, according to finite element analysis. Some additional work comparing solid and hollow specimens tested in hydrogen gas has been done by researchers at Fraunhofer IWM in Germany. They compared the hollow and solid specimen geometries for austenitic stainless steels (Michler et al. 2023), X60 pipeline steel (Michler et al. 2022) and ferritic steels (Michler et al. 2024). Both geometries were tested with in-situ hydrogen, i.e., the solid specimen was tested in an autoclave and the hollow specimen was filled with hydrogen gas. For X60 pipeline steels (Michler et al. 2022), austenitic stainless steels (Michler et al. 2023) and ferritic steels(Michler et al. 2024), they found that compared to a solid specimen, the hollow specimen had a lower reduction of area in air but a higher in hydrogen. When relative reduction of area is calculated, it is thus lower for solid specimens than for hollow specimens. This can have the effect that when the relative reductions of area are used to classify metals into different embrittlement categories in order that they can be ranked, the results for the hollow and solid specimens will fall into different categories (Michler et al. 2024). Yield and ultimate tensile strength were broadly comparable between the two geometries, except for some hollow specimens of ultra high strength steels tested in hydrogen which had a much lower UTS than the solid specimen in hydrogen. This was due to a higher notch sensitivity which resulted in failure at a lower strain, caused by drilling which gave an inner surface from which cracks could more readily initiate compared to the polished gauge length of a solid specimen. All three studies pointed out that necking behavior is different between the two specimen geometries, as the necking proceeds from the outer surface while cracking proceeds from the inner surface in a hollow specimen. In a solid specimen, necking and cracking occurs on the same outer surface, giving a lower reduction of area compared to the hollow specimen in gaseous hydrogen. Hollow and solid specimens in light water reactor (LWR) environment have also been compared (Twite et al. 2016), using a specimen with 6 mm inner diameter, 12 mm outer diameter, 3 mm wall thickness, and a polished inner surface. Fatigue testing results pointed to differences in fatigue lives between the two specimen geometries. Material tested was 304L austenitic stainless steel. Under testing in high-temperature pressurized water reactor (PWR) environment, solid specimens had fatigue lives 1.6 to 5.5 times higher than hollow specimens, with the largest difference found at long fatigue lives. For the most part, the difference was between 1.6 to 2.1 times greater. The authors argued that there can be less constraint on plasticity in a hollow specimen as the crack grows from the inner surface to the outer. This would mainly affect the growth of long cracks, resulting in differences in fatigue lives at high or intermediate strain amplitudes. At lower strain amplitudes the lives are mainly determined by crack nucleation and short crack growth where this effect would not be present to the same extent. This is not what was observed in the fatigue testing results. Where multiple cracks had formed in hollow specimens, they formed at the same distance along the gauge length as the main crack, likely caused by local elevation of strain and stress caused by the first crack. A similar investigation was performed by other researchers using a specimen with same dimensions on the gauge (Asada et al. 2017). Material tested was low-alloy pressure vessel steel and carbon steel. Outer and inner surfaces of the hollow specimen were polished to carefully remove the machined surface layer. They found no differences in fatigue life above 0.4 % strain amplitude but some differences below 0.3 %, though a smaller than what the previous study (Twite et al. 2016) found. The authors reviewed other articles comparing solid and hollow specimens, finding differences smaller than a factor of two. The authors recommended a wall thickness of 3 mm and grips with flanges to avoid bending moments.