Instability of hard rockmasses, the role of tensile damage and relaxation

Abstract

In low stress environments, stability of rockmasses around underground excavations is typically controlled by rockmass structure. At depth and in highly stressed areas the rockmass strength controls stability. Tensile strength and tensile damage processes as well as abundant relaxation or reduced confinement often play dominant roles as agents of rockmass conditioning and failure control. This study focuses primarily on stability issues related to underground excavations in "hard" rockmasses, rockmasses that are sparsely to moderately jointed with rock of high strength, high stiffness, and brittle response.

This thesis is divided into two parts. In Part I, the author examines structurally dominated failure modes, investigating the role of residual tensile rockmass strength and relaxation on the stability of tunnels and mining stopes. Because of the demonstrated significance of these mechanisms, the results of this study are used to modify empirical design tools for stability assessment.

In Part II, the author investigates the importance of microscopic and mesoscopic tensile crack growth on compression-induced damage and near-excavation fracturing of heterogeneous rockmasses at depth. The sensitivity of insitu rockmass compressive strength to low near-excavation confining stress is examined. It is demonstrated that tensile failure processes, even under moderately compressive stress fields, control rockmass failure and fracture development near excavations in hard rock. The findings of this work explain the success of empirical strength criteria and damage prediction tools, improving confidence for future design applications.

Two simple numerical analogues for rock and rockmass behaviour are used in this study. The voussoir beam model is applied to the study of structurally controlled instability while an elastic-brittle bonded-contact discrete element model is used to study damage processes of stressed rock.

For excavations in sparsely to moderately jointed hard rockmasses, the extent of gravity driven failures can be prevented or limited if the joints, which bound the potentially unstable blocks, are even slightly discontinuous. That is, even a small proportion of intact area (rock bridge) over the plane of a joint can generate significant self-supporting, load bearing capacity. The support pressure achieved for most practical support systems can be matched by a very small proportion of rock bridges as demonstrated using example analyses of wedge and jointed beam (voussoir) stability. This observation has implications for short-term or first-pass support design as well as for accurate back analysis of failures and stability assessment of areas that are affected by mining.

Limit equilibrium analyses of rock blocks and wedges are highly conservative if the effects of clamping (compressive stresses) are ignored. Conversely, abutment relaxation or unclamping of excavation backs and walls may also be a significant factor controlling rockmass stability and failure. This relaxation can be the result of geometry change, mining induced local stress change, e.g., undercutting or inappropriate wall design, or abutment deterioration. Empirical stope and tunnel design guidelines are modified in this work to account for the effects of relaxation.

An empirical insitu damage criteria for use with elastic models, based on a reduced uniaxial compressive strength intercept and a limited dependency on confining stress, has proven useful and dependable for the prediction of rockmass damage around excavations. This criterion closely resembles the experimentally determined threshold for crack initiation in laboratory and field scale tests. The success of this approach is related to the nature of the damage process in hard rock.

Two important aspects of this process, investigated in this thesis, are the tensile nature of damage mechanisms and the consequent sensitivity to low confinement conditions. Mechanisms of extension crack initiation, accumulation and interaction and the impact of low effective confining stress are explored through a review of the nature of solid bonding and damage, a study of granite behaviour in laboratory strength tests, and through the use of a bonded contact analogue. In particular, the numerical model, Particle Flow Code (discrete element) is used to examine, as a baseline case, a damage accumulation process in which individual crack extension is restricted.

While such accumulation in a heterogeneous solid shows a first order sensitivity to confinement, it is then demonstrated that increased sensitivity, to low confinement, of actual rockmass yield strength is related to the mechanics of crack extension. Rock yield strength is shown to be coincident with the onset of crack interaction. Mechanisms which cause individual nucleating cracks to extend further, do significantly increase the interaction potential and reduce the yield strength. A number of mechanisms affecting this process are addressed in this work.

Finally, it is shown that many of these mechanisms act together insitu, reducing the yield strength of rockmasses to the crack initiation threshold. These findings improve our understanding of the rockmass failure process and explain the success, in near-excavation conditions, of the empirical damage threshold described above. Now, verified mechanistically, this threshold, used with robust three dimensional elastic boundary element analyses, can be applied with confidence to determine the potential extent of excavation-induced rockmass damage and associated support requirements for complex mine openings or underground civil works.