When large openings are created in rock masses, their stability is typically evaluated by comparing the induced stress state to a properly formulated failure criterion. The parameters included in the criterion are usually determined from both laboratory properties and in situ rock mass characteristics, combined to define the large scale material behavior. In this paper, the authors use a recently developed multiaxial (3D) failure criterion with a new scale effect function to analyze the stability of two underground excavations in hard rocks. To do so, a simple numerical code based on an elastic-brittle (plasticity) type model is used in combination with the proposed criterion and scale effect function. Comparison of the calculated results and field observations shows a fairly good agreement between predicted and actual responses.
In many situations in rock engineering, there is a concern for the possible failure of rock mass around underground openings due to large induced stresses. To treat this type of problem, it is now a common approach to use numerical codes to calculate the stress field around the excavation. One can then evaluate the potential for instability of the rock mass by comparing the induced stress components to material strength as given by an approximate failure criterion. When only elastic behavior is considered, the extent of the failure zone can be underestimated because the analysis allows the material to be loaded beyond peak strength, thus neglecting the softening behavior of failed zones. This softening of failed rock beyond its available strength induces a stress redistribution in adjacent zones which may, in turn, be brought to failure because of the added demand. To deal with this aspect of rock behavior, non linear inelastic models have to be used. A simplified, yet very convenient approach is to consider that the portions of rock mass that reach failure can only bear stresses equal to the residual strength. The loss of strength (known as post-peak behavior) can be gradual but it is typically treated as sudden. Such type of constitutive behavior is sometimes referred to as elastic-brittle. With this approach, it has been shown that a reasonable evaluation of the failure zone geometry can be achieved (e.g. Hoek et al. 1995; Read & Martin 1996; Li et aI. 2000). When using this type of approach, the failed rock (and rock mass) characteristics are however difficult to define. In some cases, the material strength is simply reduced to zero, as if the failed portion was removed. In other cases, a residual cohesion and/or frictional component remain so that the failed material stays in place but with a reduced strength.. In this paper, the authors address this question by performing stability analyses of openings, taking into account the "removability" of the material in the failed zones. For that Purpose, local gravity induced stresses are compared to the material residual strength. The analyses are performed with a recently proposed failure criterion that is expressed from the full components of the stress tensor.
