PSI - Issue 18

G. Quino et al. / Procedia Structural Integrity 18 (2019) 507–515 G. Quino et. al/ Structural Integrity Procedia 00 (2019) 000–000

513

7

Fig. 6. Time history of the cumulative number of breaks and load.

Fig. 7 pictures the results of the experiment described in Section 3, and the numerical experiments using the Weibull strength distributions and the SM apparent strengths. There are various irregularities that can be observed in the experimental load-strain behaviour, while the Weibull model curve looks very smooth and shows a nicely stable fracture. As already mentioned, this is due to various factors such as difference in diameters of the fibres within the bundle, non-uniform load sharing, misalignments, manufacture induced defects, etc., that are ignored by the Weibull approach. Qualitatively, it can be observed that the apparent distribution of strengths provides a numerical test with more “realistic” features (ups and downs along the entire Strain vs. Load curve) that better agrees with the experimental curve. To quantify how close to the experimental and the numerical curves are, a residual e is defined: �� � � � ����� � � � ��� � �� � , (9) where ��� � � is the load at the strain measured in the experiment, and ����� � � is the load at the same given strain for either the model based on the Weibull distribution or the one that considers the apparent distribution of strengths found with the SM technique.

Fig. 7. Results of numerical experiments compared with the real one.