PSI - Issue 17

Jürgen Bär et al. / Procedia Structural Integrity 17 (2019) 300–307 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

306

7

clear separation of the sheets by steps and secondary cracks is clearly visible. The surface of the secondary cracks is very smooth, with nearly no visible structure (figure 7).

b

a

Fig. 7. SEM images of the final fracture of a specimen fatigued in liquid hydrogen (LH2).

4. Discussion

The experiments have shown that the fatigue behavior of AA2198 T851 at cryogenic temperature is clearly improved compared to room temperature. The adaption of the Gecks-Och-function to the measured cyclic lifetimes shows a very good match. The adaption results in values for the tensile strength and the fatigue limit and allows a direct comparison of these values. Kohout and V ě chet (2001) proposed another function describing the complete S N-curve using parameters having a clear geometrical meaning. Unfortunately it was impossible to adapt these function on the measured data points of the fatigue experiments. The ultimate tensile strength  UTS and the fatigue limit  FL show a different increase in the absolute as well as the relative values, therefore a prediction of the S-N-curve for LH2-conditions by a simple shift based on the UTS-values is not possible. Additionally the form of the S-N-curve is changed between RT and LH2-conditions. The transition from the low cycle fatigue to the finite fatigue life region is smoother in LH2 and shifted to higher cyclic lifetimes, the transition to the fatigue limit is shifted to a higher cyclic lifetime, too. The shift of the first transition point is considered in a model proposed by Kohout (2000). However, the model does not include the shift of the fatigue limit and the transition of the finite fatigue life region to the fatigue limit. Therefore an extrapolation of the S-N-curve with this model is only possible within a limited range. This clearly indicates that a prediction of the S-N-curve for cryogenic temperatures with the currently available models is not possible. Therefore, an optimized construction using the full capacity of the material at low temperature is only possible when the fatigue behavior is determined experimentally. The fractographic investigations have shown, that under LH2 conditions the fracture surface show distinct differences compared to specimen tested under RT conditions. All specimens tested in LH2 show a sheet-like structure leading to a pronounced roughness of the crack surface. This phenomenon is well known for Al-Li-alloys tested under cryogenic conditions. Venkateswara et al (1989) found in their fracture toughness tests on several Al-Li alloys a strong dependence of the measured values of the orientation of the specimens. The increase of the fracture toughness with decreasing temperature is ascribed to a mechanism of crack-divider delamination toughening. The splitting of the material into several individual sheets hinders the propagation of cracks through the material. In the fracture toughness experiments this splitting takes place in front of the crack tip. It is unknown, when the splitting process in the fatigue experiments presented in this work takes place. However, the results of the fractographic investigations show that in most cases no clear identifiable fatigue crack area is found in case of the specimens tested in LH2. It can be assumed that the splitting takes place in an early stage during fatigue