Uniaxial compression tests were conducted on prismatic Barre Granite specimens with two pre-cut, straight, open flaws. Using a high-speed video system, crack initiation, propagation, and coalescence were observed. Coalescence patterns for the granite specimens fit into a previous framework (developed for Molded Gypsum and Carrara Marble) except for one new coalescence pattern. The crack initiation stress and the maximum stress were measured for each specimen, and these results are interpreted, also.
Crack coalescence, which is the linkage of pre-existing flaws, is a common phenomenon in nature. The current study examined coalescence in Barre Granite as an extension of previous work presented by the MIT rock mechanics group . A high-speed video system was used to record the coalescence of two artificial flaws in Barre Granite (the term flaw will be used in this paper to refer to an artificially made, pre-existing crack). This observation method made it possible to distinguish between shear and tensile cracks during their formation and propagation as well as record the sequence of cracks during coalescence. The observed coalescence behavior was complex, but also fit into the framework proposed earlier for Molded Gypsum and Carrara Marble .
2.1. Specimen Preparation
While this study used Barre Granite, several other materials have been used in the past being part of the continuing research of the MIT rock mechanics group. For nearly forty years, the group has studied the behavior of discontinuous (jointed) geo-materials . In this context, a specific study began 17 years ago to experimentally investigate the effects of material type, geometric parameters, and loading conditions (Table 1) on fracture initiation, propagation, and coalescence [1, 4, 5, 6]. Bobet continued his work at Purdue University . Several other groups have also researched crack coalescence in geo-materials [e.g. - 8, 9, 10]. To enable one to make comparisons between this study and previous work, similar procedures for specimen preparation were followed. Prismatic specimens of Barre Granite specimens with dimensions ~152 mm x~76 mm x~25 mm (6" x 3" x 1") were prepared. North Barre Granite, Inc. cut slabs ~25 mm (1") thick with a diamond saw. The other two dimensions (6" and 3") were cut with an OMAX waterjet, which cuts with a high-pressure mixture of water and a garnet abrasive. Two open, straight flaws 12.7 mm (0.5") long were cut in each specimen with the waterjet. The flaws were cut with different geometric relationships. Ligament length (L) was always equal to flaw length. The flaw inclination angle was varied (ß= 0°, 30°, 45°, 60°, and 75°) for two values of bridging angle (a= 0°?and 60°). As a consequence, flaw pairs were either coplanar (a= 0°) or left stepping (a= 60°) and could be nonoverlapping, partially overlapping, or completely overlapping. Three specimens were prepared and tested for each flaw-pair geometry. The two geometries summarized in Table 2 were geometries identical to those tested by Wong for Molded Gypsum and Carrara Marble .
Table 1. Parameters tested by the MIT rock mechanics group (both specimen parameters and loading conditions). Refer to Figure 1 for explanation of terms marked with an asterisk.(available in full paper)