ICF13B

13th International Conference on Fracture June 16–21, 2013, Beijing, China -8- a) d3 extrapolated from d8 estimates b) d22 extrapolated from d8 estimates c) d22 extrapolated from d3 estimates d) d3 extrapolated from d22 estimates Figure 5. Extrapolation to different specimen geometries (1, 5 and 99 % percentile curves) improve the model. Different surface quality can also be excluded, as all specimens were machined in the same workshop, therefore supposedly having undergone the same surface treatment. In particular, both the d3 and d8 specimens were fabricated from 25 mm diameter rods. Although, this implies that the surface of the d3 specimens is closer to the centre of the rods than the surface of the d8 specimens and possibly having experienced both different cooling rates, the difference is negligible bearing in mind the original rod diameter. One could also question the existence of a size effect. Though this assumption might be true for the d3 and d8 specimens, it is obvious from Fig. 6 that the d22 specimens have lower lifetimes than the d3 and d8 specimens for the same stress ranges. Nevertheless, in the region of low stress ranges, the d3 specimens tend to present higher fatigue lifetimes than the d8 specimens. On the other hand, the prediction for the d22 specimens based on the d8 estimates (Fig. 5b) is quite good lying almost all failure data for the d22 specimens between the 1 and 99 % - percentiles. However, a tendency to underestimate the fatigue strength is noticeable since by the extrapolation the data with highest lifetime are assigned to failure probabilities higher than 99 %. 10 4 10 5 10 6 10 7 10 8 180 200 220 240 260 280 300 320 340 N [Cycles] Δ σ [MPa] 1% 50% 99% 10 4 10 5 10 6 10 7 10 8 180 200 220 240 260 280 300 320 340 N [Cycles] Δ σ [MPa] 1 % 50 % 99 % 10 4 10 5 10 6 10 7 10 8 180 200 220 240 260 280 300 320 340 N [Cycles] Δ σ [MPa] 50 % 1 % 99 % 10 4 10 5 10 6 10 7 10 8 180 200 220 240 260 280 300 320 340 N [Cycles] Δ σ [MPa] 1 % 50 % 99 %

RkJQdWJsaXNoZXIy MjM0NDE=