Issue34

T. Itoh et alii, Frattura ed Integrità Strutturale, 34 (2015) 487-497; DOI: 10.3221/IGF-ESIS.34.54 494 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 (%) Uniaxial fatigue λ=0 Biaxial fatigue λ=0.4 Biaxial fatigue λ=0.5 Biaxial fatigue λ=1 Non-polish Polish Figure 10 : Comparison of failure lives at each principal stress ratio’s test. Dependence of defamation behavior on multiaxiality Fig. 11 shows variation of mean axial strain (  z mean ) as a function of cycle ( N ). In the figure, the vertical axis is normalized by mean axial strain at 10 cycles (  z mean10 ) and the horizontal normalized by ( N / N f ). In all the tests, mean strains increase gradually with increase of cycles, i.e. , cyclic ratcheting occurs because the stress amplitude employed was relative high comparing with the yield stress 113 GPa. The degree of the ratcheting becomes larger with increasing  and the larger ratcheting may leads to the reduction in failure life. In structural components using Ti-6Al-4V, on the other hand, the stress level under service loadings is usually much lower than that employed in this study. In this case, no or small ratcheting will occur resulting in smaller effect of  on failure life. However, the property of failure life under multiaxial loading at lower stress levels is still open question and additional experimental tests at the stress revel are required to confirm them. Figure 11 : Variation of mean axial strains at each principal stress ratio’s test. Surface crack and effects polishing on failure life Fig. 12 shows cracks observed on specimen surface fatigued at each  ’s test. In the figure, the upper side shows surface cracks observed by optical microscope and the lower side schematic showings of the main crack. Using the polished specimen, the crack in the λ=0 test shows a zigzag shape crack. Each zigzag crack is approximately 100  m in length and has propagation directions of  45 degrees from the specimen axial direction and then the macro-crack formed by linking of each zigzag crack propagated normal to the axial direction. In the tests at λ=0.4 and 0.5, observed cracks are similar to that at λ=0 although length of each zigzag crack is slightly smaller. In the λ=1.0 test, a main crack is observed in the direction of specimen axis and small cracks are also observed around the main crack in various direction. Using the non-polished specimen, on the other hand, the crack at λ=0 test is slightly different from that at λ=0 test with the polished specimen. The crack shows the zigzag one but it may be formed by linking cracks propagating along cutting 0 0.2 0.4 0.6 0.8 1 0 1 2 3 4 λ ＝ 0 λ ＝ 0.4 λ ＝ 0.5 λ ＝ 1.0 Mean axial strain normalized by ε z mean10 ε z mean /ε z mean10 Number of cycle normalized by N f N / N f