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

117

5

Figure 5 shows the application of the √ area parameter model to the surface-polished specimens. The fatigue limit of the fractured specimens were estimated from the size of defect observed at fracture origin and HV and Eq. (2). The applied stress   a was normalized by the estimated fatigue limit  w . The specimens which ran out up to 10 7 were re tested at higher stress until the specimens fail, and then the killer defects were identified. It must be noted that individual specimens have individually different fatigue limit due to presence of defects with individually different sizes. This is the cause of scatter and we cannot define one definite fatigue limit for an AM material. It is very difficult to find nonpropagating cracks at fatigue limit for AM specimens, because we don’t know the size of killer defect prior to fatigue test.

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4

EBM Surface polish without HIP EBM Surface polish without HIP №A EBM Surface polish without HIP №B DMLS Surface polish without HIP DMLS Surface polish without HIP №C DMLS Surface polish without HIP №D New machine Surface polish without HIP New machine TP No.3-1-6-9

σ a / σ w

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

Cycles to failure N f (cycles)

Fig. 5 Modified S - N data.  a : applied stress,  w : estimated fatigue limit.

(a) σ a =400 MPa, 1.0×10

7 cycles

(b) σ

a =400 MPa, 2.0×10

7 cycles

(c) σ

a =500 MPa, 9.7×10

4 cycles

Fig. 6 Nonpropagating cracks emanating from an AM defect. The cracks did not grow from N =10 7 to N =2x10 7 cycles at 

a =400MPa. These

cracks grew by additional stress cycles N =9.7x10 4 cycles when the specimen failed at 

a =500MPa from another subsurface defect hidden below

surface.

Figure 6 shows a nonpropagating crack which was luckily observed at the test stress. Thus, the presence of nonpropagating crack at fatigue limit confirms that the fatigue limit of AMmaterials were determined by the threshold condition for small cracks. Thus, it can be confirmed in Fig. 5 that the √area parameter model works very well. The evidence of nonpropagting crack is the basis of the applicability of the √area parameter model. So far, we have always observed the presence of nonpropagating cracks at small defects at the fatigue limit (Murakami and Endo, T (1980), Murakami (2002)). At the same time, we confirmed that there exists a critical nondamaging size of defect dependent on hardness of materials also on AM materials. The reason why the threshold for small cracks is different from long cracks is explained in details in Murakami (2019).