PSI - Issue 19

Lloyd Hackel et al. / Procedia Structural Integrity 19 (2019) 452–462 Valentin LOURY--MALHERBE/ Structural Integrity Procedia 00 (2019) 000–000

457

6

In general, there is concern for AM samples about voids and inclusions in the materials that might be randomly distributed and thereby initiate failure over a large region [16]–[19]. The 4-point bend testing configuration is especially effective because it provides a very uniform test stress over a large fraction of the gauge section and thus samples relatively large volumes of the material. Figure 4 shows our finite element analysis prediction of the stress and strain developed during loading of the test samples. As can be seen, the test setup develops very uniform strain over the entire region between the load rollers. to begin the tests, strain gauges were applied to a first sample, the sample loaded on the machine and a 7000 MPa (1000 lb) load applied. For this load the strain gauge showed 615 microstrain which was within 3% of the 595 microstrain predicted indicating that the applied stress was accurately calibrated to the loading and was consistent with our modeling. In operation stress loadings were reported to this measured/predicted load-to-stress calibration result. The analysis shown in Figure 4 gives a maximum stress along the gauge section of 134 MPa (19.170 ksi) for an applied load of 4.448 kN (1000 lbs). For the fatigue test results shown in this report a maximum load of 22.65 kN was applied(5093 lbs) which generated a maximum stress of 683 MPa (97.6 ksi). Literature values for the yield stress of AM In718 [20] give a value of 791 MPa (113 ksi) at 0.2% yield. Literature values from Special Metals Corporation [21] give a yield stress of 1050 MPa (150 ksi) at 0.2% elongation rating the AM material at 24% lower yield strength.

Figure 4. FEA model of AM In718 fatigue test sample shows uniform stress generated across the gauge region. Strain values were measured and confirmed to be within 3% of the predicted value.

With the basic samples fabricated, specific surface treatments were applied including single shot peening (0.3 mm size shot with medium A-scale intensity), dual shot peening with a light and then heavy peening (0.3 mm size shot with medium A-scale intensity followed by peening with 0.1 mm shot size with high C-scale intensity), and standard laser peening. The laser peen was applied to the top curved sections of the samples only with each layer of peening applied at 8 GW/cm2 irradiance and 18 ns of laser pulse duration with no ablative layer on the rough AM surfaces. Some samples were also hand polished over the gauge radius using 600 grit paper followed by 1000 grit and then 3000 grit. Also because the LP+TME process indicated meaningful stress retention in wrought In718, we explored variants of this processing by adding in more of the peening plus annealing steps. More specifically our notation Process A represents laser peening + 8 hour anneal at 600°C + laser peening + 8 hour anneal at 600oC + laser peeing. Process A ′ adds another iteration of anneal + laser peening and Process A ″ yet another. This A ″ process represents a total of 5 peening steps and 4 annealing steps.