ABSTRACT
Rock materials are multiphase heterogeneous composites, consisting of mineral grains with preexisting defects in the forms of voids and cracks in their microstructures as shown in Figure 11.1. With the advance of experimental techniques, close observations on the specimen’s microscopic behaviour during rock failure processes become possible. For instance, Tapponnier and Brace (1976) investigated stress-induced microcrack developments within different mineral components in the Westerly granite. Wong (1982) further investigated the faulting mechanisms of different minerals in Westerly granite with different confining pressures and temperatures and concluded that the failure mechanism was related to both mineralogy and mineral grain orientation. Several numerical approaches have also been put forward to study the fracture processes of such heterogeneous materials, such as the lattice-based model (Blair and Cook, 1998), the RFPA (Realistic Failure Process Analysis) code (Tang et al., 2000), the local degradation model based on the FLAC (Fast Lagrangian Analysis of Continua) code (Fang and Harrison, 2002), the synthetic rock mass model (SRM) based on the PFC (Particle Flow Code) (Potyondy and Cundall, 2004). Although they succeeded in one aspect or another, most of these models were based on finite element or discrete element methods that in some respects have certain limitations. For instance, the continuumbased finite element models do not work well to capture the failure process featured by distortion and large deformation, fracture propagation and fragmentation. The discrete element-based methods, on the contrary, are not well defined in processing
the continuous deformations. Besides, few of these approaches can properly model the aggregates or grains in their microstructures and appropriately account for their effects in rock failure simulations.
