Issue 38

T. Morishita et alii, Frattura ed Integrità Strutturale, 38 (2016) 289-295; DOI: 10.3221/IGF-ESIS.38.39 291 non-proportional loading. In this stress level and the material of SS400; a relative good agreement of data correlations may be resulted from that the additional hardening is balanced with the reduction in failure life. Fatigue strength in the circle loading test σ w CI (150MPa) is lower than that in others σ w PP (175MPa). Fig. 3 shows observations of specimen surface in the push-pull and the circle loading tests at stress amplitude level around the fatigue strengths; Δσ eq /2=200; 175 and 150MPa. The observed location was set at the sufficient distances from a main crack which contributes directly to N f . In the push-pull loading test; the roughness caused by local plastic deformation can be observed clearly on the specimen surface only at Δσ eq /2=200MPa. In the circle loading test; on the other hand; the remarkable roughness can be observed at Δσ eq /2=200 and 175MPa in comparison with those in the push-pull loading test at each equivalent stress amplitude; which may be resulted from the increase in the number of activated slip systems due to the rotation of principal direction of stress under non-proportional loading. The roughness leads to more chance of initiation of microcracks and the earlier crack initiation. Consequently; the surface roughness causes reduction of the fatigue strength in the circle loading test. Fig. 4 is the failure life correlated by an equivalent total strain range based on von Mises Δε eq . The strain ranges used are those at the cycle of 0.5 N f in experiments. In this figure; the bold solid line is drawn by a universal slope curve [16] based on the experimental data in the push-pull loading test. The universal slope curve is given by 6.0 f 12.0     BN AN - f eq (1) where the coefficients A and B are equated as 3.5σ B / E and ε f 0.6 according to the definition of the universal slope method. E ; σ B and ε f are Yong’s modulus; a tensile strength and an elongation; respectively. In this study; A is put as the mechanical properties obtained from the tensile test but B is defined to fit the universal slope curve to the data of the push-pull loading test. In LCF region; failure life in the rev. torsion loading test is underestimated and conversely that in the circle loading test is overestimated out of the factor of 2 band. The same tendency of failure life was shown in the previous study of strain controlled multiaxial LCF test [11]; therefore it suggests that the failure life and the non- proportionality are not affected by the difference in the test control of strain or stress for the tested material. In the high cycle fatigue region; with decrease in strain range; failure life in the circle loading test approaches to that in the push-pull loading test. This trend indicates that the effect of non-proportional loading on failure life is decreased in the lower strain level; which will be mentioned in next. 10 3 10 4 10 5 10 6 10 7 100 150 200 250 300 350 Number of cycles to failure N f , cycles Equivalent stress amplitude , MPa  eq 2 MPa 175 σ PP w  MPa 145 σ CI w  Push-pull Rev.torsion Circle Figure 2: Correlation of N f with equivalent stress amplitude based on von Mises.