PSI - Issue 75

D. Tousse Tchamassi et al. / Procedia Structural Integrity 75 (2025) 450–456 Tousse / Structural Integrity Procedia (2025)

452

3

and from the expanded part of a connection, along the axis of the parent tube, with a 5-mm gauge diameter and a 8 mm gauge length. The gauge region of all specimens was mirror- polished with 15 µm down to 0.25 µm diamond pastes (R a < 0.48 µm). LCF tests were conducted using a triangular wave form, a strain rate of 10 -3 s -1 , with or without prior tensile strain of either 2% or 4%; the specimen elongation was controlled using an extensometer attached to the specimen gauge over a length of 5 mm. For the tensile prestrained specimens, prior deformation was first applied in tension and the LCF strain was superimposed right after, without any unloading, also starting in tension. Experimental conditions are gathered in Table 1. Most tests were interrupted just before final fracture, as soon as the maximum tensile stress dropped by 300 MPa with respect to its initial value. The shape of the last hysteresis loop indicated that a long crack was indeed propagating across the LCF specimen (Fig. 2a). This procedure ensured that the side surface of the specimens was not significantly modified by final ductile fracture and by the associated high amounts of strain and damage development. Some specimens were then broken by tensile loading at low temperature (80°C) to expose fatigue cracks for further observation, again without significant additional deformation.

Table 1. Experimental conditions of low cycle fatigue tests. ‘*’ symbols indicate specimens that were observed using X -ray tomography.

Initial deformation state No prior deformation

strain amplitudes, ±ε total (%)

0.5 – 0.65 – 0.7 – 0.8 – 0.9* – 1.0* – 1.2*

2% tensile strain 4% tensile strain

0.6 – 0.7 -0.8 -1.0 0.5 – 0.6 – 0.8- 1.0 0.5 – 0.6* – 0.8*- 1.0

Expanded region of the connection

2.3. Investigation of damage development and fracture

Fracture surfaces as well as some specimen side surfaces (namely, the undeformed material after fatigue tests at ±0.5%, ±0.8%, and ±1.0%) were observed after the tests using a Zeiss Sigma 300 scanning electron microscope (SEM) using secondary electron imaging, a high voltage of 5 kV, an aperture of 30 µm, and a working distance between 5 mm and 30 mm according to the required magnification. For some specimens (indicated with an asterisk in Table 1), side surface observations indicated that the main (longest) crack initiated from the extensometer attachment regions. This could suggest that the slight indents left by the extensometer on the side surface of the specimen might have triggered fatigue crack initiation and governed the fatigue lifetime of those specimens. If so, no other deep crack was expected to be observed in their gauge region. To determine whether other deep cracks were also present, X-ray absorption contrast tomography was conducted for those specimens at Navier Laboratory, using a 180 kV high voltage, a voxel size of 5 µm, and a beam current of 29 µA. The imager captured the images at a frame rate of 0.5 Hz, with frame averaging set to 8. A total of 928 projections of each scan were acquired to enable the 3D reconstruction. 3. Results and discussion 3.1. Low cycle fatigue behavior 3.1.1. Cyclic behavior without prior deformation Fig. 2 illustrates the cyclic behavior of the initially undeformed material. As already observed by Guguloth et al. (2014) for high-strength martensitic steels, cyclic softening was observed whatever the prescribed strain amplitude. The lower apparent modulus observed during unloading for the last cycle suggests the presence of a deep crack (Fig. 2a). By superimposing stress-strain curves at mid-life, a Masing effect was evidenced, i.e., the loading part of the curves were very similar, especially for strain amplitudes larger than ±0.5%. The stress amplitude became rather independent of the strain amplitude, for a given number of cycles, as shown in Fig. 2b. This means that the elastic strain amplitude was the same, and that the plastic strain amplitude strongly increased with the total strain amplitude. Thus, as expected, higher total strain amplitudes led to lower fatigue lifetimes.