PSI - Issue 7

A. Rotella et al. / Procedia Structural Integrity 7 (2017) 513–520 Antonio Rotella et al. / Structural Integrity Procedia 00 (2017) 000–000

518

6

(grade < 1) is reduced of about 25%. Figure 4a shows the S-N diagram obtained with the two artificial defect configurations. As reported in section 3.1, the local morphology of a natural defect is driven by the dendritic structure, in order to reproduce this effect, a small local defect has been introduced at the tip of the principal spherical defect. The defect has been machined using the EDM technique. The electrode has a diameter of 50 µm and is made of pure Copper (99 % Cu). The final local defect size (estimated on the fracture surface) is of 92 µm (Figure 4c). The additional defect has not a detrimental effect with respect to a simple spherical defect and the estimated fatigue limit is the same for both configurations. It is important to notice that the addition of an artificial defect will not always initiate a critical crack that leads the specimen to failure. In two cases the specimen failed because of the presence of an oxide or a natural shrinkage. Nevertheless an analysis of the artificial defect surface showed that the artificial defect is critical enough to initiate a crack that not leads the specimen to failure. This result shows that globally the local defect morphology is not a first order parameter influencing the fatigue limit. A local defect modification (that does not affect the defect size) is not influent towards the fatigue limit and the result is the same obtained for a simple spherical defect. Therefore, the defect type is a parameter that should be taken into account. The different defect nature can substantially influence the critical crack initiation site. It seems that the more complex geometry of an oxide film or natural defect (shrinkage) is more critical, as crack initiation site, with respect to a spherical artificial defect. Nevertheless this effect is not visible when the local artificial defect morphology is modified. 3.3. Kitagawa-Takahashi diagram (influence of defect size on the fatigue limit) A Kitagawa-Takahashi diagram [(Kitagawa et al. (1976)] has been plotted in order to investigate the effect of the defect size on the fatigue limit of the A357-T6 cast Aluminum alloy (Figure 5).

/

Fatigue limit (MPa)

1

R= 0.1

0

1E-02

1E-01

1E+ 00

/

Fig. 5. Normalized Kitagawa-Takahashi diagram showing the influence of the defect size on the fatigue limit, all the specimens failed because of a surface defect, the fatigue limit is estimated at R=0.1 and is expressed as a stress amplitude ( σ D0.1 ta ), σ D0.1-max ta is the value of the maximal fatigue limit (calculated as stress amplitude) used for the stress normalization. The parameter AREA 1/2 is the defect size, AREA max 1/2 is the value of the maximal defect size used for the size normalization. A Kitagawa-Takahashi diagram is characterised by a critical defect size that indicates a reduction of the fatigue limit of the material. It is important to notice that in this work, a deterministic approach has been used since the experimental database is not sufficiently large to allow the application of a probabilistic approach. Nevertheless it is well known that the critical defect size is not a unique value but defines a border between the reference material (low probability to reduce the fatigue limit for a given defect size) and a degraded material (higher probability to reduce the fatigue limit for a given defect size). Referring to the normalized Kitagawa-Takahashi diagram of Figure 5 it can be observed that the fatigue limit of the material is reduced starting from a given defect size. This effect is more pronounced between a normalized defect size of 0.05 and 0.2. After this value the diagram starts to saturate. The only difference can be observed for the specimens classed as CS 4 that show a lower fatigue limit for a defect size that is similar to the one estimated for the specimens classed as SS 2 and SS 3. Nevertheless as already reported in section 3.1, the defect size estimation for the specimens that are characterized by a distributed sponge shrinkage (CS 3/4 and SS 2/3) is affected by the uncertainty to define a proper defect contour. From a global point of view, the diagram shows that the effect of the oxides is more pronounced in absence of other natural defects. In presence of natural defects the effect of the oxides is not highlighted and the failure always occurs on shrinkages. The defect size has a