PSI - Issue 42

J.R. Antunes et al. / Procedia Structural Integrity 42 (2022) 588–593 Author name / Structural Integrity Procedia 00 (2019) 000–000

592

5

initiation and propagation of any defects. It was also noted that there was a substantial difference in magnitude between the two stress components. This is typically due to the small variation in temperature in the transverse direction, along with the lack of restraint. Regarding the observed fatigue test results for unpeened samples, it was clear that the as-welded pristine FSW samples showed lower fatigue life than the base material samples. This was expected given the effect of softening in the mechanical properties of the welded alloy. Moreover, these as-welded samples failed consistently in lip region of the weld, on the shoulder side of the sample, more specifically on the advancing side. This behaviour can be attributed to the heat input being higher on the advancing side of the weld compared to the retreating side, due to the overlapping directions of both the tool rotation and pin travel speed. 6.2. Effect of peening on the residual stresses and fatigue life of FSW joints. In order to understand the variation of stresses after peening along the transverse direction to the weld, the results were re arranged for a fixed depth of 100 µm, as can be observed in Figure 7.

Figure 7 - Residual stress results after peening for a fixed depth of 100 µm.

As can be seen, the magnitude of the residual stresses decreases with distance to the weld centreline. Moreover, the stresses in the nugget region (weld centreline) are 1.5 times higher compared to the edge of the weld, and 4 to 6 times higher than the ones found outside the peened region. The results also highlight the high level of biaxiality between the two stress directions (approx. 50MPa difference between stress components). These results reveal the robustness of the peening treatment. FSW is prone to a particular type of defect called “lack of penetration” (LoP) defect, where there is incomplete bonding of the two sheets of metal on the root side of the weld due to insufficient heat input through the thickness. Given that the peening treatment achieved a penetration depth of about 500 µm in the transverse direction (coinciding with the loading direction, and therefore being the crack opening direction), this means that a defect up to that dimension will be completely surrounded by a compressive residual stress field. This could potentially decrease crack growth or even prevent initiation. In the event that the defect is bigger than 500 µm, a significant portion of its extension will still be under the effect of compressive residual stresses. Another benefit of this particular peening treatment is that significant compressive residual stresses were achieved at the surface, which is particularly important to avoid crack initiation from the surface texture and surface defects. As for the balancing tensile residual stresses, these were found at the centre through the thickness direction, which is a non-critical region of these samples, presumably with no defects. In terms of fatigue performance, life was increased by a factor of 2.4 for the second lowest applied stress. This effect was lessened as the stress increased. This is expected as the peak compressive stress achieved by the treatment was only 175 MPa, while the maximum tensile stress applied during testing was 350 MPa. This confirms once again the robustness of the treatment, given that for 240 MPa, peening restored 65% of the as-welded sample’s life when compared to the pristine material with no weld. 7. Conclusions In conclusion, this work helped to verify that fatigue is a complex mechanism and that failure in the FSW joint will be a competition between two main factors: