PSI - Issue 19

Yukio Miyashita et al. / Procedia Structural Integrity 19 (2019) 604–609 Author name / Structural Integrity Procedia 00 (2019) 000–000

607

4

Figure 3 shows result of fatigue strength test for SA and SB. In the figure, S - N curves for the base material with loading direction parallel (LD) and perpendicular (TD) to the extruded direction are also shown. According to Fig.3, weld specimens show lower strength compared to base material. However, difference in fatigue strength was not significant between welds produced in Institute A and B. The both weld specimens showed large scatter in result of fatigue strength test regardless of the institutes. Figure 4 shows examples of fracture origin observed in SA and SB. According to fracture surface observation, fatigue crack initiated from a weld defect in all weld specimens tested. A weld defect was also found as fracture origin in the previous paper concerning on fatigue characteristic of magnesium alloy weld (S. Hamada et al.,(2010)). Results of fatigue crack growth tests in SA and SB are shown in Fig.5(a). The weld specimens showed higher fatigue crack growth resistance compared to the base material. Fatigue crack growth curves arranged by effective stress intensity factor range, Δ K eff is shown in Fig.5(b). Crack closure could not be measured in TD specimen due to large deformation induced by extrusion process, therefore crack growth curve arranged by Δ K eff in TD specimen is not shown in Fig.5(b) and later Fig.7(b). It was speculated that difference of fatigue crack growth resistance between weld and the base material was mainly due to difference in crack closure behavior. Relationship between fatigue life and stress intensity factor range, Δ K for SA and SB is shown in Fig.6. Stress intensity factor indicated in the figure was calculated from size obtained as area for a weld defect observed at fracture origin and applied stress amplitude by using equations (2) for surface crack and (3) for internal crack (Y. Murakami, 1993 ). According to the figure, fatigue life was well arranged by stress intensity factor and fatigue strength at 10 7 cycles was possibly evaluated by threshold stress intensity factor range, Δ K th .

(b) SB, σ a =60 MPa, N f =19,842 cycle

(a) SA, σ a =60 MPa, N f =33,743 cycle

Fig.4 Observation of fracture origin.

10 -6

(a)

(b)

LD SA SB

10 -7

10 -11 10 -10 10 -9 10 -8 da/dN, m/cycles

10 -12

0.1

0.5 1

5

Δ K eff , MPam 1/2 Δ

Δ

Fig.5 Relationship between fatigue crack growth rate, da / dN and (a) stress intensity factor range, Δ K and (b) effective stress intensity factor range, Δ K eff in SA and SB.