One of the difficulties in describing the rock mass behavior is assigning the appropriate constitutive model. This limitation may be overcome with the progress in discrete element software such as PFC, which does not need the user to prescribe a constitutive model for rock mass. In this paper, the model size of 30m × 30m was analyzed by using the fracture geometry from two tunnel sites. PFC simulations were carried out to examine the mechanical behavior of rock masses. From the numerical tests, it can be concluded that as the number of joint sets increased, the values of mechanical properties of rock masses were decreased to about 50% of those values of rock mass without joints. And the behavior of the rock mass changed from brittle to perfectly plastic with increase in the number of joints. Also the values of Young's modulus, Poisson's ratio and peak strength are almost similar from PFC model and empirical methods. As expected, the presence of joints had a pronounced effect on mechanical properties of the rock mass. More importantly, the mechanical response of the PFC model was not determined by a user specified constitutive model. So the discrete element model gives very contrasting results compared to the traditional model.
Although the evaluation of the mechanical properties and behavior of discontinuous rock masses is very important for the design of underground openings, it has always been considered the most difficult problem. The reason is that it is often impossible to carry out large-scale in situ tests and, although widely used, the correlations between strength parameters and quality indexes (for instance GSI or Q index) are still affected by considerable uncertainties (Ribacchi 2000).
The evaluation of rock mechanical properties such as deformation and strength properties can be achieved through the application of empirical relationships or by a theoretical approach based on numerical modeling. Both methodologies imply some assumptions and uncertainties that need to be considered.
Deformation properties and rock mass strength are not only dependent on the intact rock, but also on the fracture network (number and orientation of fracture sets, intensity, mineralization, and so on) and the presence of deformation zones. Therefore, characterization of both the intact rock and of the fractures is required to define the mechanical behavior of the rock mass.
Discontinuous rock masses are usually weaker and more deformable and are highly anisotropic when compared with intact rocks. So constitutive modeling of discontinuous rock masses has long been a subject of interest and numerous models have been developed in attempt to simulate their mechanical responses (Staub et al. 2002). Recent developments in numerical modeling that allow study of the overall response of a synthetic material containing discrete heterogeneities and discontinuities both at the micro (particle) scale and at the larger scale of jointed rock masses can greatly aid the interpretation and application of laboratory test results on these materials (Potyondy & Fairhurst 1999).
The methodology for the rock mechanical descriptive model was developed in Sweden.