PSI - Issue 57

Cristian Bagni et al. / Procedia Structural Integrity 57 (2024) 859–871 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

869 11

Fig. 6. Test set-up.

The normalised stiffness vs cycles plots (example shown in Fig. 7) showed that the stiffness remained reasonably constant for a portion of the tests, before rapidly decreasing to a new stiffness value that stayed reasonably constant until the final sharp stiffness decrease at specimen failure. This is in agreement with the findings of other research works (Moroni (2019) and Wu et al. (2021)). In particular, the first rapid stiffness decrease can be associated to the failure of the adhesive, after which only the mechanical joints carry the load and contribute to the fatigue life of the joint. This can be observed in Fig. 8, where it can be seen that the adhesive failed cohesively (but close to the interface between the adherends and the adhesive) until the crack reached the rivets and then a crack initiated and propagated through one of the adherends. Fig. 7 also shows that the portion of the total joint life provided by the adhesive increases as the load decreases: for high load/short life tests the life represents a relatively small portion of the overall life of the joint (Fig. 7a), while for low load/long life tests the life of the adhesive represents about 50% of the total life of the specimens. The 10% stiffness drop can be considered representative of the failure of the adhesive, as shown in Fig. 7, and therefore, Fig. 9 provides a rough idea of the level of conservativism introduced when the life of the mechanical joints is neglected. This is important information when deciding whether to adopt the first or the second methodology. Adopting the first methodology and assuming 10% stiffness drop as the failure criterion (representative of the complete failure of the adhesive), the LN datapoints obtained from the tests (Fig. 9a) can be reverse-engineered, as described above using Eq. 5, to obtain the SN datapoints (Fig. 9b) necessary to derive the SN parameters. It is recognised that the number of datapoints currently available and presented in Fig. 9 is limited. However, further tests belonging to the same test programme are currently underway. Once the test programme is complete, the full set of test results will be used to further validate the proposed methodology and may be used to derive bespoke fatigue parameters of the tested hybrid joints.

Fig. 7. Representative examples of normalised stiffness vs cycles plots for (a) short life test and (b) long life test.