PSI - Issue 37

Koji Uenishi et al. / Procedia Structural Integrity 37 (2022) 397–403 Uenishi and Nagasawa / Structural Integrity Procedia 00 (2022) 000 – 000

399 3

acccc

Tensile loading

Normalized maximum in-plane shear stress 4.5

b

c

c ccccccc c

10 mm

1

Regions of stress amplification

2

Rupture front wave

2

Tensile loading

3

0

Fig. 1. (a) Archetypal specimen having multiple pre-existing small-scale parallel cracks that dip vertically and model a geological fault plane near a horizontal free surface [unit: mm]. (b) Fracture development, all dynamic, in the specimen under uniaxial tension with a constant strain rate (1.3  10 − 2 /s); and (c) Finite difference simulations indicate crucial roles of the main fracture-induced waves in the generation of subsidiary fractures. Here, the upward main fracture travels at an experimentally observed sub-Rayleigh speed, 563 m/s (modified after Uenishi and Nagasawa (2021)). specimen by the main fracture. As before, the dynamic secondary fracture is arrested, for some 300  s, but resumes its downward propagation. Fractures at all three stages may radiate dynamic waves into the far-field. On the other hand, pre-existing parallel cracks with smaller dip angles tend to induce rather different fracture behavior, i.e. quasi-static evolution of the main fracture. In the third case, the pre-existing cracks have a dip angle of as small as 15 degrees (Fig. 3). First, the main fracture extends upward relatively slowly and connects the pre-existing cracks in a step-by-step fashion. This quasi-statically extending main fracture is arrested well below the top free surface, but almost simultaneously downward secondary fractures can be initiated and they propagate dynamically. From the seismology point of view, the secondary fractures radiate much more kinetic (seismic) energy into the far-field in a shorter time span and may play more crucial physical roles than the main one. However, note also that it may be hard to “detect” the local fracture evolution like in Fig. 3(b) via global stress-strain curves obtained by a testing machine. In the figure, branching of the secondary fractures, which results in the merge of the main and secondary fractures and split of the entire specimen is also identifiable. In the fourth case with a dip angle of 30 degrees, the crack distribution pattern is totally different from the other cases. In specific positions the specimen has zones of damage consisting of small-scale cracks (Fig. 4). By intuition, such initially damaged zones are expected to enhance fracture development in the specimen, but the experimentally obtained snapshots do indicate that both upward quasi-static main and downward dynamic secondary fractures are “captured” in the damage zones . Thus, pre-existing damage zones can absorb fracture development, and a local geometrical change can really govern the global and local fracture behavior. 4. Conclusions This study has shown several examples of the global and local fracture behavior in brittle specimens with different initial inclination angles and distribution patterns of sets of pre-existing small-scale parallel cracks. In general, if the pre-existing cracks are more steeply inclined with respect to the externally applied tensile load, both the main and the oppositely moving secondary fractures travel dynamically. On the contrary, when the inclination angles are smaller, the main fracture is prone to extend quasi-statically but the secondary ones can propagate dynamically. The experimental observations also indicate that jump, arrest and resumption of propagation of fractures can quite easily occur in our two-dimensional specimens with small-scale cracks. These drastic changes of fracture behavior may imply the mechanical importance of small-scale cracks in understanding the generation of a cluster of earthquakes.