PSI - Issue 7

S. Romano et al. / Procedia Structural Integrity 7 (2017) 275–282

281 7

S. Romano et al. / Structural Integrity Procedia 00 (2017) 000–000

a

b

c d Fig. 6. Statistics of extremes estimates for di ff erent material volumes with 95% scatter band (green) and characteristic maximum (red) for: (a) material 1; (b) material 2; (c) material 3; (d) material 4.

Table 3. Maximum defect size estimated compared to the maximum measurement in V c . Material Max. experimental (mm) Maximum characteristic estimate (mm)

2.5% percentile (mm)

97.5% percentile (mm)

1 2 3 4

7.8 4.1 7.5 7.2

8.0 4.3 7.9 6.9

7 1 3.7 6.6 6.0

8.9 4.9

11.4

7.8

4. Discussion

In all the materials analysed, the presence of voids could always be considered irrelevant with respect to degenerate graphite, except in the cases in which both the defect types were detected in the same regions. In general, di ff erent behaviours could be expected considering materials showing a better overall quality, reason why considering both the defect types in the same analysis gives a more robust result. Peaks-over threshold maxima sampling was seen to ensure the best performance. The choice of the threshold can slightly influence the results, but a unique value can be used for all the materials, equal to 1.25 mm. Note that the application of POT does not allow to make estimates for a defect size lower than u . Some anisotropy and heterogeneity of the material and of defect elongation have been detected in di ff erent samples and even inside the same one. The three dimensional measurements obtained by computed tomography allow obtaining a more precise characterization