PSI - Issue 28

Vinicius Carrillo Beber et al. / Procedia Structural Integrity 28 (2020) 1950–1962 V.C. Beber and M. Brede / Structural Integrity Procedia 00 (2019) 000–000

1959

10

(Beber et al., 2015).

20 30 40

5

Elastic Adhesive

Structural Adhesive

TAST Joint Scarf Joint Butt Joint

TAST Joint Scarf Joint Butt Joint

10 15

1

5

(b)

(a)

10 2 Nominal Shear (or Normal) Stress Amplitude 1 10 3 10 4 10 5

10 6

10 7

10 2

10 3

10 4

10 5

10 6

10 7

Nominal Shear (or Normal) Stress Amplitude

Number of Cycles to Failure [ - ]

Number of Cycles to Failure [ - ]

Figure 9 - Fatigue results: SN curves from butt, scarf and TAST joints for structural (left) and elastic (right) adhesives

One of the most suitable approaches to assess the multiaxiality effect on the fatigue behaviour is to use the p-q diagram (Beber et al., 2017). The p-q diagram is constructed as expressed in Figure 10: by taking the nominal stress amplitude ( σ a,i ) related to a certain fatigue lifetime ( N i ), the respective hydrostatic pressure ( p ) and Von Mises equivalent stress distribution ( q ) should be calculated. Then, a value of p and q should be selected from the stress distribution. In the present work, the maximum values of p and q were taken for the analysis. Then, the respective p and q values should be plotted. By connecting the p-q values of joints with different stress multiaxialities it is possible to construct iso-life lines. With this approach, the fatigue lifetime prediction of joints with several levels of stress multiaxiality can be carried out (Baumgartner, 2014). The p-q diagram of the structural adhesive is presented in Figure 11a based on the SN curves of the butt, scarf, and TAST joints. Here the values of p and q are related to the value of the stress amplitude. The plots were constructed from the SN curves in Figure 9a. The iso-life lines delimit the maximum values of p and q for a specific fatigue lifetime (e.g. red line for a lifetime of N = 1,000). This chart is very useful as it provides an approach for lifetime prediction of adhesively bonded joints regardless of their geometry, and consequently multiaxial state of stress (Beber et al., 2017). With regard to the structural adhesive, it is interesting to highlight that butt, scarf and TAST are spaced among the chart which can be correlated to their stress multiaxiality values as shown in Figure 8. Likewise the p-q diagram of the elastic adhesive (Figure 11b) was constructed from SN curves of the butt, scarf and TAST joints. However, since the multiaxiality of the scarf and butt joint is very similar (see Figure 8) some of the p-q points are very close to each other. Therefore, these results suggest that effect of stress multiaxiality is not only dependent on the geometry of the joint, but also on the mechanical properties of adhesives. 4.3. Fracture surface analysis Representative fracture surface images are presented in Figure 12 (for structural adhesive joints) and Figure 13 (for elastic adhesive joints). For both adhesives, no clear distinction was made between static and fatigue fracture surfaces. For structural adhesives joints, the same fracture pattern was observed for types of joints, namely near-surface cohesive failure. Similarly, elastic adhesive joints presented a pure cohesive failure. These cohesive fracture surface results suggest a proper surface preparation for both adhesive (da Silva et al., 2011) .