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

517

5

Fig. 3. Fracture surface of the specimen 30-2 (CS 2) under uniaxial tensile load at R=0.1, σ amp = 40 MPa, N = 5.52 ∙ 10 5 cycles, (b) cavity shrinkage at the origin of the failure with a defect size of AREA 1/2 = 1619 µm, (c) fracture surface of the specimen 53-1 (SS 2) under uniaxial tensile load σ amp = 40 MPa, N = 1.29 ∙ 10 5 cycles, the sponge shrinkage at the origin of the failure has an estimated defect size of AREA 1/2 = 6766 µm (d) zone F that shows a cluster of porosity that form the final sponge shrinkage. σ amp is the stress amplitude that caused the failure of the specimen. Figures 3c and 3d show the fracture surface of the specimen 53-1 classed as sponge shrinkage grade 2 (SS 2). It can be observed that a single initiation site is not detectable and that the gage section of the specimen is affected by a distributed porosity. At a local scale, the micro-shrinkages that form the sponge shrinkage have a morphology related to the dendritic microstructure of the material (Figure 3d). A clear defect contour is difficult to be identified and the defect size has been estimated by considering the whole zone that is affected by the distributed porosity. The most probable crack initiation and propagation scenario is that the micro-shrinkages, which are close to the specimen surface, start to initiate micro-cracks. During the test the micro-cracks coalesce in order to form a stable propagation zone until the crack is long enough to activate the final instable crack propagation process. The remaining part of the specimen is characterized by a ductile failure. The specimens classed as cavity shrinkage with a grade ranging from 3 to 4 are normally affected by the presence of a distributed porosity that surrounds the cavity shrinkage. Also in this case the defect size has been estimated by considering the whole zone that is affected by the porosity. From a point of view that is strictly related to the shrinkage family it can be observed that the effect of the global defect morphology (isolated cavity or distributed porosity) does not seem to have a significant impact on the fatigue limit. Any difference have been detected between cavity and sponge shrinkages of the same grade (grade 2 and 3) on the fatigue test results. 3.2. Fatigue tests on specimens with surface artificial defects with a local morphology modification In order to better understand the effect of the defect morphology, fatigue tests have also been conducted on specimens with artificial surface defects.

/ σ a-max

σ a (MPa)

0.0 0.2 0.4 0.6 0.8 1.0

C

1E+ 05

1E+ 06

1E+ 07

500 µm

C

50 µm

Number of cycles

(a)

(b)

(c)

Fig. 4. (a) S-N diagram obtained for the specimens with artificial surface defects (EDM) at R=0.1 (frequency of 106 Hz), σ amp is the stress amplitude, σ amp-max is the value of the maximal stress amplitude used for the stress normalization. (b) Fracture surface of the specimen 25-2 classed as reference (grade < 1) at R = 0.1, σ amp = 60 MPa, N = 3.99 ∙ 10 5 cycles, failed on an artificial spherical defect with a local morphology variation (EDM), defect size AREA 1/2 = 715 µm, (c) zone C of the fracture surface showing the local morphology variation (EDM), defect size AREA 1/2 = 92 µm. The artificial defects are machined using the Electron Discharge Machining (EDM) technique. Two defect morphologies have been tested, a classical spherical defect and a spherical defect with a local perturbation (that does not affect the global defect size). The machining of the spherical defect is performed using a pure Copper electrode (99% of Cu) with a diameter of 1 mm. With this kind of machining, the final defect size expressed as AREA 1/2 (calculated on the fracture surfaces) is of about 700 µm. In this condition the fatigue limit of the reference material