PSI - Issue 34

Harry O. Psihoyos et al. / Procedia Structural Integrity 34 (2021) 253–258 Harry O.Psihoyos et al. / Structural Integrity Procedia 00 (2021) 000 – 000

257

5

Consequently, the most critica l defect for fa ilure crack initiation and propagation lies at the surface of this layer. The melt pools which did not fulfill the LOF criterion a t the layer of the thinner cross-sectionarepresented in Figure 1.

Fig. 1. Melt pools susceptible for lack-of-fusion defect formation in the layer of the smallest cross-section (indicated with pink dots).

4.2. Fatigue life estimation Due to the location of defect on the surface of the thinner cross section of fatigue specimen, the most suitable model (K-solution) for fatigue life is the one proposed by Forman and Shivakumar (1986) for surface crack growth in solid cylinders. The initia l size of defect was considered 150 μ m based on the experimenta l observations of specimen fractured surface of Du et a l. (2021). The results of fatigue life prediction model are presented in the Fig.2. As it can be seen, the correla tion between prediction results and experimenta l data is very good. It should be mentioned that experimenta l data concern the cases where cracks initiated from the surface. Regarding the ana lytica l LEFM simula tion, crack growth did not take place for stress amplitudes less than 200 MPa , mainly because the defect size was not adequatenot causeany crack growth. This issue will be further investiga ted in future studies.

150 170 190 210 230 250 270 290 310

Experiment - SLM Ti-6Al-4V LEFM simulation

Stress Amplitude, σ α [MPa]

1.E+04

1.E+05

1.E+06

1.E+07

Number of Cycles to Failure, Ν f

Fig. 2. Comparison between experimental S-N fatigue data and predicted results