In the present paper, the characteristics of the granite burst under loading/unloading conditions was studied by means of both laboratory experiment and finite-discrete element method (FDEM). After gaining the mechanical properties of the granite samples under uniaxial compression conditions, the parameters taken as input arguments in the FDEM modelling were then carefully calibrated so as to provide some basic data including Young's modulus, Poisson's ratio, uniaxial strength, Lame constants, cohesive strength, and frictional angle for further being used in simulating granite bursts. All the simulating results by the FDEM were basically in agreement with that obtained from the uniaxial compression test and triaxial granite bursts in laboratory, respectively. This study shows that that such computational mechanics of discontinua can be employed to gain powerful insight into the failure mechanism of granite burst, which could also be a useful tool to clarify the collapse mechanism of rock structures in rock engineering design and rock test scheme optimisation.
Based on discontinuum state a priori, discontinuous modeling techniques that are commonly referred to as discrete element methods (DEM) treat the material directly as an assembly of separate blocks or particles, which was originally proposed by Cundall (1971) and Cundall and Strack (1979) from the viewpoint of analogous molecular dynamics simulation to better account for and understand the interaction between the blocks. The distinct feature of these formulations is that a constantly changing problem of geometry and contact conditions can be addressed by solving the equations of motion using an explicit time marching scheme (most often in a ‘leap-frog’ format), while updating contact force histories as a consequence of contacts between different discrete elements and/or resulting from contacts with model boundaries (Bicanic 1998). The most notable implementations of DEMs are arguably represented by the universal distinct element code (UDEC) (Itasca Consulting Group Inc., 2013) and the particle flow code (PFC) (Itasca Consulting Group Inc., 2012). Another formulation is the discrete crack framework that is based upon some criteria for an advancement of crack front, which is coupled with extensive local and global re-meshing while simulating by numerical methods such as finite element technique, commonly using fracture theories. Certainly, it is relatively easy to trace the propagation of single crack front governed by some principles of energy release rates (Ingraffea 1997) but tracing of multiple cracks leads to topological and algorithmic complexities. Otherwise, there is no consideration of interacting cracks in these formulations of finite element or enhanced finite element techniques (e.g., in XFEM and ABAQUS). Although PFC is able to simulate the behaviour of rock cracking under laboratory experiment conditions and in situ observations (Feng et al 2006, Hsieh et al 2008, Sagong et al 2011, Singh et al 2012), how the computational parameters such as the stiffness at the directions along normal and tangent at the contacts are linked with both the macroscopically experimental results in laboratory and in situ macro-scale responses to need extensive validation. PFC and DEM are used to predict the cracking and fragmentations after damage, which usually require both a large body of blocks or particles and a very small time step to model the continuous/discontinuous problems of interest. Consequently, large scale computation to the formulation further requires much more computer time to complete an analysis for study. Recently Lisjak and Grasselli (2014) presented state of the art review on discrete modeling techniques for fracturing processes in rock masses to better help understanding for the current trend emerged in the field of rock mechanics as simulation tools. Among all the formulations in simulating the continuous and discontinuous behaviours of rock cracking, the combined finite-discrete element method that belongs to the discipline of computational mechanics of discontinua holds a specific advantage because it inherits all the merits of ordinary DEM and enables to yield a more accurate estimate of the contact force and deformation as finite elements are embedded in (Munjiza 2004, Munjiza et al 2011, Munjiza et al 1995).
