PSI - Issue 19

Yukitaka Murakami et al. / Procedia Structural Integrity 19 (2019) 113–122 Yukitaka Murakami et al./ Structural Integrity Procedia 00 (2019) 000–000

116

4

Based on this grade, only the specimens of Ti-6Al-4V with HIP + Surface polish is evaluated to be Fatigue-Grade 5. On the other hand, the specimens without HIP and without Surface polish stays in Fatigue-Grade 1. 4. Influential factors to be considered We can pick up the following influential factors which may influence the fatigue performance of AM materials.  Surface roughness , Defects , Powder size , Microstructure (phase, grain size, etc.) , Build direction , etc . Figure 3 shows a typical defect observed at fracture origin s . The size of defect is estimated by the √area which is defined by a smooth contour of defect (Masuo et al (2018), Yamashita et al (2018), Murakami (2019)). It is clear from Figs. 1 and 2 that the influence of as-built surface roughness is most detrimental and defects are second most detrimental. Figure 4 shows the transverse section and top view of an as-built specimen manufactured by EBM and DMLS. The electron beam melting (EBM) and direct metal laser sintering (DMLS) are two most common 3D printing methods for additive manufacturing. In EBM, a focused electron beam scans the bed that metal powder layer is deposited, resulting in localized melting and solidification of powder. In DMLS, a focused layer is used instead of electron beam. It is evident that powder size influences the morphology of surface roughness. It should be noted from Fig. 1 and Fig. 4 that microstructure (Masuo et al (2018)) and build direction (Yamashita et al (2018)) are not the first priority from the viewpoint of fatigue strength in the presence of surface roughness and defects. In other words, only minor improvement of fatigue performance by changing microstructures may be expected after removing the influence of surface roughness and defects. Therefore, we must focus our attention on the influence of surface roughness and defects before we discuss the problem of microstructures and build direction. It is difficult to define the equivalent defect size of the surface roughness of as-built specimens, because the equivalent defect size √ area eq of surface roughness inversely calculated by the √ area parameter model is larger than 1000  m which definitely exceeds the category of small crack. Therefore, currently we cannot avoid a very large safety factor if we apply AM as-built materials to power components. In other words, AM defects remaining after removing as-built surface roughness by surface polish can be treated by small crack models while the as-built surface roughness should be categorized between small cracks and long cracks.

(a) EBM with HIP

(b) DMLS with HIP

(c) EBM with HIP

(d) DMLS with HIP

Fig. 4 Surface morphology of as-built specimens. (a) and (b): Transverse section, (c) and (d): Top view.

5. AM defects as the small crack problem and the presence of nonpropagating crack at fatigue limit The equivalent crack size of surface roughness of AM materials mostly exceeds the category of small crack (Murakami (2019)). The size of AM defects is mostly smaller than 1000  m and the fatigue of AM defects must be treated by the mechanics of small crack. The √area parameter model (Murakami and Endo, M (1994)), Eq. (2), is a convenient model which has been applied successfully to various fatigue problems of small defects and cracks.  w = C ( HV +120)/(√ area ) 1/6 (2) Where,  w : fatigue limit (MPa), C : 1.43 for surface defects, 1.56 for internal defects, HV : Vickers hardness (kgf/mm 2 ) and √area : equivalent defect size (  m). The fatigue limit in presence of defects is not directly proportional to HV and the constant 120 reflects the easiness of development of crack closure mechanism for soft metals (Murakami (2019)).