Planning teams need confidence that the numerical tools they use to estimate potential for mine instability are sufficient to quantify the hazards. The tools must be quantitative, field validated, and the hazardous phenomena must emerge in these models as a function of the governing physics.

In seismically active mines, this essential task is more difficult as high rates of deformation and dynamic loading are still comparatively infrequent events. Reliable simulation of support and reinforcement for dynamic loading is even more complex.

The goal of this paper is to summarize a framework that has been used for reliable numerical rock mechanics simulation at several mines for forecasting of potential for induced seismicity.


From an engineering perspective, the mining engineer designing ground support and excavations in a mine must undertake two broad tasks:

  • Estimation of the potential for dynamic deformation and seismic events, including appreciating the mechanisms by which damage may occur to an excavation, including both rapidly (dynamically) increasing dilation of the rock mass due to increasing stress as well as the effects of seismic waves.

  • Simulation of the response of installed support and the discontinuous rock mass around the excavation.

The only way to simulate significant excavation loading, deformation and energy change is to properly capture the physics of rock damage. If the extent and magnitude of the damage in the model is wrong, the displacements and energy changes, including estimates of seismic potential will also be wrong. The complete stress-strain behaviour of rock must be correctly captured.

A tool that can correctly forecast the rock mass behaviour on different length scales should include:

  • The ability to incorporate explicit defects down to the length scale of interest.

  • The strain softening, dilatant behaviour of the rock mass.

  • The simulation of post-failure response of rock and support.

The modelling framework discussed in this paper attempts to satisfy these requirements. It consists of a constitutive model that describe the stress-strain behaviour of rock masses and structures and the application of an explicit Finite Element (FE) algorithm to represent the rockmass as a three-dimensional discretized volume with discontinuities and heterogeneous material domains. The main ingredients are:

  • The continuum regions of the rockmass are modelled as strain-softening dilatant materials. This means that as strain increases the material softens, weakens, and dilates. All parameters can vary at different rates with respect to strain changes, and this allows approximation of complex stress-strain behaviour of real rock masses. A generalization of the Hoek-Brown yield criterion (Hoek et al., 1992) is used for the continuous regions of the rockmass derived from a more generic and versatile formulation of (Menetrey&Willam, 1995).

  • The behaviour of explicit discontinuities is approximated using cohesive elements. The constitutive behaviour of the cohesive elements can be defined using the presented constitutive model, or a constitutive model specified directly in terms of traction versus separation.

  • The seismic potential of the modelled rock mass can be assessed by considering the modelled Rate of Energy Release (RER), which is the maximum instantaneous rate of energy release within a unit volume during a period of the simulation.

This content is only available via PDF.
You can access this article if you purchase or spend a download.