1 INTRODUCTION
Research on damage process in the region of low confinement near openings, subjected to high in-situ stresses, suggests that it initiates in a moderately jointed brittle rockmass by the nucleation and growth of extension fractures within the intact rock material inside the rockmass, and that valid information can be extracted from laboratory tests on intact rock to assess the initiation of rockmass damage (Castro, 1996). This papers proposes that potential zones of damage initiation (DI) around deep excavations can be reliably predicted by performing elastic numerical analyses to identify regions where the deviatoric stress, (s1-s3), exceeds the threshold value at which stable crack growth commences for intact rock tested under uniaxial compression, s sc. Application of this criterion, successfully predicted the depth of the DI zones in the moderately jointed, brittle norite rockmass surrounding the Sudbury Neutrino Observatory (23 m diameter x 30.8 m high) excavated at a depth of 2070 m at INCO's Creighton Mine in Sudbury, Canada.
Field observations and an extensive literature review strongly suggests that, under compressive loading and conditions of low confinement, the damage process is dominated by the initiation and propagation of extension fractures within the intact rock. This means that the mode of damage initiation is by extension fracturing, which was observed to: a) nucleate, at a microscopic scale, from stress concentrators such as pores, pre-existing cracks and particularly, grain boundaries in most polycrystalline rocks; b) be a consequence of induced (or local) tensile stress, extension strain, or strain energy density around these stress concentrators; c) form at low angle (i.e.,<10º) or parallel to the direction of the major principal induced stress; and d) be sensitive to changes in the magnitude and direction of the major principal stress (Castro, 1996). Research on damage and failure processes in the region of low confinement near openings, subjected to high in-situ stresses, suggests that movements of blocks and along pre-existing, non-continuous discontinuities have only a minor effect on rockmass damage initiation, because the blocks do not have the kinematic freedom to allow translation or rotation. Thus, damage must initiate and progress within intact rock material. However, the progression of the damage process then creates new internal structures. The interaction of the newly formed internal fractures, together with the pre-existing discontinuities, increases the number and variety of kinematically feasible failure mechanisms that can be developed around the opening (Castro, 1996). This paper proposes that, even in a moderately jointed rockmass, the onset of damage is essentially controlled by the inherent properties of the intact rock material inside the rockmass and their relationship to the magnitude and orientation of the induced deviatoric stresses. If this hypothesis is correct, it means that valid information can be extracted from small-scale laboratory tests regarding the onset of damage, which occurs in a moderately jointed rockmass around an underground excavation at great depth, i.e., information can be transferred between different loading systems, the laboratory system and the field system, respectively.
