PSI - Issue 39

A.L. Pinto et al. / Procedia Structural Integrity 39 (2022) 409–418 Author name / Structural Integrity Procedia 00 (2019) 000–000

417

9

Table 3 presents the predicted crack initiation angles for all the tests and methods investigated in this work. It also shows the measured experimental angles reported in (Almeida et al., 2020). It is worth noticing that the Method 2 does not provide straight crack estimates. In this case, results displayed in Table 3 consider the straight line connecting the crack initiation site with its tip. A similar approach is considered for the experimental angles. Through the data presented in Table 3, it is noted that 93.75% of the results obtained with Method 1 are within the experimental error margin, however, only 18.75% of the results obtained by Method 2 are within the experimental scatter. Concerning the worst estimates provided by the Method 2, it seems that microstructural features and localized plasticity near the surface might not be fully neglected when simulating the propagation of very small cracks. However, the average stress/strain evaluations considered in the CDM approach seemed to properly capture the macro behavior of initiation cracks.

Table 3 – Crack angle prediction for all tests using Methods 1 and 2. Load config. Test Exp. (°) Average (°)

Method 1 (°)

Method 2 (°)

21.4 ± 12.2 29.5 ± 6.4 31.4 ± 8.0 25.9 ± 5.4 31.5 ± 0.3 30.4 ± 8.4 32.4 ± 13.6 37.2 ± 13.2

T1 T2 T1 T2 T3 T1 T2 T1 T2 T1 T2 T1 T2 T1 T2 T3 T1 T2

12.8 30.0 28.0 36.6 24.0 25.7 37.0 29.7 22.0 31.3 31.7 36.3 24.4 40.7 16.7 39.7 46.5 27.9

C1

27

10

C2

25

10

C3

23

9

C4

35

20

C5

35

16

C6

35

15

C7

25

9

C8

25

10

6. Conclusions This paper investigates two different approaches, namely Method 1 and Method 2, to predict crack initiation direction in fretting problems subjected to partial slip conditions. Experimental data from fretting tests conducted on an AA7050- T7451 alloy (Almeida et al., 2020) were used for validation. In this setting, the following conclusions could be drawn: i. Method 1, which is based on the CDM in conjunction with SWT’s fatigue model, provided crack initiation angle estimates in good accordance with the experimental observations. ii. On the other hand, Method 2 did not prove itself accurate in predicting crack initiation angle. Almost all the predicted angles by this method were out of the experimental margin of error.