PSI - Issue 7

Gianni Nicoletto et al. / Procedia Structural Integrity 7 (2017) 133–140 Gianni Nicoletto/ Structural Integrity Procedia 00 (2017) 000–000

137

5

structure” and α’ martensite. The lamellar spacing and the alpha grain boundary thickness decrease with increase cooling rate.

3.2. Tensile behavior Tensile test results on the specimens subjected to the different post fabrication heat treatments are presented in Fig. 4. Ultimate tensile strength UTS and elongation at break E% are considered. Several observations can be made: i) mechanical properties for all heat treated samples are above the specific minimum ASTM F2924-14 requirements for additive manufactured Ti6Al4V, namely UTS= 895 MPa and E%= 10%; ii) the post fabrication heat treatments influence mainly the elongation at break, which is an index of material ductility; iii) HT1 and HT2, which differ for the cooling rate, show a similar ultimate strength but an improved elongation at break of the HT1 treatment; iv) HT3 data show the worst mechanical properties of the four heat treatments in terms of both strength and elongation; v) HT4 specimens show high ultimate tensile strength and the maximum elongation at break.

a) b) Fig. 4 Influence of post fabrication heat treatment on tensile properties of SLM Ti6Al4V a) ultimate strength b) elongation at break

3.3. Fatigue behavior The rotating bending fatigue tests performed to determine the influence of the post fabrication heat treatment are presented in Fig. 5 and show that this influence is more significant than on static properties. In detail, the HT1 specimens show good fatigue properties with a low data scatter. As regard the heat treatment HT2, while the tensile results were similar to HT1, the fatigue behavior is definitely worse than that of HT1 with remarkably shorter lives. The HT3 treatment, which was characterized by the worst tensile results, shows a good high cycle fatigue behavior, close to that obtained with the HT1 treatment. A good performance, both tensile and in fatigue, is determined with HT4 specimens. It is pointed out that the run-outs at the highest stress amplitudes were associated to this HT4 heat treatment but with a somewhat high scatter. The influence of the as-built surface on fatigue was investigated on a set of specimens heat treated according to HT1 and then machined on an automatic lathe. The comparison of the as-built and of the machined data is presented in the fatigue plot of Fig. 6. An experimental as-built surface fatigue knock-down factor C as-built = 0.3 is estimated using the respective fatigue strengths at 10 7 cycles. Although the roughness of the as-built SLM surface is the result of several mechanisms that are not discussed here, it is reminded that the surfaces of a SLM part of complex geometry may be too expensive or even impossible to machine. As an additional information, although the surface roughness of the present as-built rotating bending specimens was not measured directly, as-built flat specimens produced under the same conditions and oriented in the Z direction were characterized by an average roughness R a = 13.4 µ m and R z = 80.7 µ m, see Konečná et al (2017).