PSI - Issue 7

Ana D. Brandão et al. / Procedia Structural Integrity 7 (2017) 58–66

61

4

Author name / Structural Integrity Procedia 00 (2017) 000–000

2.2. Specimen Characterization The characterization of the specimens comprised an analysis of the defect population present in a selected number of samples. This study was followed by the fatigue testing and the subsequent fracture surface analysis. Micro X-ray Computed Tomography (µCT) scans were performed on one sample per process parameters combination from Table 1. The acquisition was done in a Phoenix V|tome|x m 300 kV (General Electric) at a gun voltage of 230 kV, a current of 100 μA , with a timing per frame of 1000 ms and acquiring 1000 images over the complete rotation, using a filter of Cu with 0.5 mm thickness. The voxel size was around 18 μm. Reconstruction was performed using the Phoenix datos-x software from GE, using optimization functions for smoothing, beam hardening and post-alignment correction. The results obtained were analysed with the software VG Studio MAX 2.2.1 (Volume Graphics GmbH). The defect population, i.e. voids and inclusions were studied with the custom method VGDefX (v2.2). In order to perform a qualitative analysis of the overall porosity present in the different samples, a visualization aid from VG Studio MAX 2.2.1 was used, which allows the compression of different slices of a given volume in a single image. The result is an overlapped image of all the slices, portraying all the defects present in the volume analysed. Fig. 2 shows a representation of this tool and a typical overlap image of the voids present in the selected volume of an AM specimen. This approach permits a simplified qualitative inspection of the defect population in a sample, avoiding the screening of all the single slices. Fatigue testing was performed at room temperature on an Instron ElectroPuls™ E10000, with a stress ratio of R = -1, using a test frequency of 30 Hz, at stress amplitude of 120 MPa, according to ASTM E 466 (2015). The test was stopped at 2 × 10 7 cycles considering unbroken samples as runouts. Fracture surface analysis was performed via Scanning Electron Microscopy (SEM) examination, using a Zeiss EVO60 SEM at an accelerating voltage of 20 kV, Spot Size 402, I-probe current of 125 pA at a working distance of 10 mm and using Secondary Electron Detection.

Fig. 2. Representation of (a) the specimen and the volume selected for analysis (defined by the top and bottom blues planes) and (b) of a typical overlapping image of the voids present in a given volume of an AM specimen. This image represents a series of slices of a sample prepared with ID 4 parameters, and not the pores present on a single layer.

3. Results and Discussion 3.1. Fatigue Properties

The graph in Fig. 3 plots the number of cycles to failure at a stress level of 120 MPa for all samples tested for each combination of the manufacturing parameters, described in Table 1. From this graph, it can be observed that the samples produced with the parameter ID 2 showed the best performance in fatigue of this AlSi10Mg specimens. These were manufactured with a layer thickness of 30 µm and were all considered runouts (2 × 10 7 cycles). The impact of this parameter on the characteristics of this aluminium alloy produced via AM was also studied by Read et al. (2015) and Aboulkhair et al. (2017). Both authors reported that an increase in the layer thickness results in a growth of the number of defects (pores, lack of fusion) in the specimens, which is typical in an AM process. This can be explained by the fact that a larger layer thickness results in a reduced energy input, as shown in Eq. 1, (Read et al. (2015)), leading to lack of uniformity and smaller molten pool depth during deposition (Aboulkhair et al. (2017)). A similar tendency is observed in this work, where the samples produced with a layer thickness of 90 µm (energy input of 22 J/mm³) show the presence of a significant larger number of defects (see section 3.2.), when compared with the specimens