PSI - Issue 57

Cristian Bagni et al. / Procedia Structural Integrity 57 (2024) 598–610 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

608

11

process described above, some preliminary test data from the ongoing testing programme is presented. It is worth highlighting, however, that this data represents just a small part of the overall testing programme and is not sufficient to derive SN parameters. Therefore, it must be considered for illustration purposes only in this paper. The data presented below was obtained by testing lap shear specimens under sinusoidal axial loading. The test set-up is shown in Fig. 8. The displacements were measured using an additional deflectometer calibrated to an accuracy of 0.005mm (higher compared to the accuracy of the position transducer of the test rig). Dimensions of the specimens and materials cannot be disclosed for confidentiality reasons.

Fig. 8. Test set-up.

The tests confirmed that the failure mode was predominantly cohesive, but close to the interface between the adherends and adhesive. The normalised stiffness vs cycles plots (see for example Fig. 9) showed that the stiffness remained reasonably constant for at least 50% of the total life, supporting the findings from previous research works (Crocombe et al. (2002), Quaresimin and Ricotta (2006b) and Graner Solana et al. (2007)), and that crack initiation represents a significant portion of the total life of the joint. Finally, Fig. 10a summarises the LN datapoints obtained from the analysis of the test data for two possible failure criteria: test stopped and 2.5% stiffness drop. Fig. 10b instead shows the SN datapoints, obtained by reverse-engineering the LN datapoints, necessary to derive the SN parameters. It is not uncommon that during the manual process of specimens manufacturing and adhesive application, the bond quality might be compromised. In that case, the fracture surfaces must be carefully analysed to determine if the observed failure mode is still representative to what is expected from production joints. This will determine how the corresponding datapoints are treated during data post-processing (i.e. whether to include or exclude them from the analysis to generate the SN curve). In Fig. 10, the two datapoints that lie out of family might fall into this category . They are designated as ‘Scatter only’, and t heir suitability to be included into the data post processing for the SN curve generation must be carefully evaluated. It is recognised that the number of datapoints currently available and presented in Fig. 10 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 adhesive joints.