PSI - Issue 54

Nikolai Kashaev et al. / Procedia Structural Integrity 54 (2024) 361–368 Kashaev et al. / Structural Integrity Procedia 00 (2023) 000 – 000

367

7

The exemplary determined quantities such as the area of the defect used for the calculation of the initial crack length and the distance to the surface a f of the two typical defects are shown in Fig. 7a-b. The results concerning the number of cycles to failure predicted using the fracture mechanics approach N predicted vs. the number of cycles to failure experimentally determined N measured are shown in Fig. 7c. As can be seen, most of the predicted results are within the five-fold error band, which verifies the appropriate prediction capability of the numerical method and the calibrated material parameters. However, for five specimens, the predicted fatigue lives are much lower than those determined experimentally. One reason for this could be in the sensitivity of the model to the initial crack length. For example, the initial crack length in the case of a spherical defect is more accurate to determine (Specimen 3, Fig. 7a) compared to a lack-of-fusion defect with a more complex shape (Specimen 4, Fig. 7b). Moreover, for a perfectly spherical defect from gas porosity, an accurate analytical solution for the stress intensity factor exists. That is not the case for irregularly shaped lack-of-fusion defects present in the additively manufactured material. For these kinds of irregular defects, numerical methods might yield better results, notwithstanding the influence of local microstructure. In conclusion, the results of the study show that the fracture mechanics-based model can be used to adequately predict the minimum fatigue life of the fatigue specimens extracted from the WAAM-fabricated structure. Prediction capability was particularly accurate in the case of porosity defects, while overestimated the effect of irregular lack-of fusion defects. Thus, a conservative prediction of the fatigue life is achieved.

Fig. 7. (a)-(b) Photos of fracture surface of failed specimen in the case of internal crack initiation from (a) a pore and (b) from a lack-of-fusion defect. (c) Predicted vs. experimentally obtained (measured) fatigue lives. 5. Conclusions Investigation of the fatigue strength of additively manufactured structures using specimens extracted from the deposited structure provides a unique opportunity to analyze the effects of both internal and surface defects on high cycle fatigue (HCF) behavior. Using this analysis, a fracture mechanics framework was adapted to achieve quantitative prediction of HCF life. The main conclusions of this work are summarized in the following bullet points: - The fatigue behavior of the investigated additive manufactured structures concerning the fatigue limit is on the level of a cast material, which can be attributed to the presence of pores and lack-of-fusion defects with submillimeter size. - The process-related defects are uniformly distributed in the structure so that in the case of specimens extracted for the fatigue analysis from the structure, both internal and surface crack initiation occurred. - Using the fracture mechanics-based model the minimum fatigue life of the fatigue specimens extracted from the WAAM-fabricated structure can be adequately predicted. Thus, a conservative prediction of the fatigue life is achieved. - The reason for the overestimation of the predicted fatigue life could be due to the overestimation of defect size used for the calculation of the initial crack length.