The importance of understanding the phenomena associated with rock fracture has long been fully appreciated in rock mechanics. This is clearly apparent from the special attention paid to rock fracture in the comprehensive papers published within the past four years.1-5 These papers present a picture of how the knowledge of brittle fracture of rock was gradually enriched and also show that what was often taken for granted three or four years ago is no longer considered as valid today.
The Griffith hypothesis,6thought in the past to predict the strength of rock, is now known to be applicable to fracture initiation only. The knowledge of a failure criterion for fracture initiation is not sufficient, however, as after fracture initiation the rock is still a useful material even possessing increasing load-bearing ability. The importance of the fracture propagation process, subsequent to fracture initiation, is thus obvious. Yet, relatively little attention has so far been paid to brittle fracture propagation in rock. As it is now realized that understanding of this process is a prerequisite for practical rock mechanics applications, studies of propagation of brittle fracture in rock constitute an important aspect of rock mechanics research.
It has been pointed out 4 that it is essential that the failure mechanism is fully understood before a failure criterion is postulated. Many failure criteria have been propounded in the past on the basis of theoretical reasoning alone and could not be verified by experimental evidence.
Recently, a hypothesis for the mechanism of brittle fracture of rock was propounded4, describing the events taking place in the rock from the initial application of gradually increased load to complete disintegration of the material.
In Fig. 1, the axial stress vs. the lateral, volumetric, and axial strain is plotted for Witwatersrand quartzite specimens subjected to gradually increased uniaxial compressive load. The characteristic events taking place in rock during this process are marked on the graph.
The process starts with crack closure when the cracks, inherent in any rock, close due to the application of the compressive load. Only after completion of crack closure do the stress-strain curves become linear, the strain being elastic. Further deformation leads to fracture initiation when the closed cracks start to propagate (onset of fracture propagation). This stage is associated with departure from linearity for the lateral and volumetric strain-stress curves but not for the axial strain-stress relationship. Now stable fracture propagation prevails; it is called stable because it can be halted by stopping the gradual load increase. During this process, a part of the elastic energy is released to extend the crack surfaces. When critical energy release is attained, unstable fracture propagation starts; it is called unstable because the propagation cannot be controlled by load control alone any more. This process is associated with departure from linearity for the axial strain-stress curve and change in curvature sign for the volumetric strain-stress curve. Next, strength failure occurs, i. e., the loss of the maximum load-bearing capacity.
