Consideration of anisotropy is important in various rock types and applications, e.g., borehole stability in shale formation especially when drilling sub-parallel to bedding planes. In this study, three-dimensional bonded-particle discrete element modeling was performed in a transversely isotropic rock using PFC3D as an extension to the previous work conducted in two dimensions (Park and Min, 2015). Rock anisotropy was taken into account by means of inserting a smooth joint model and verification conducted in terms of both elastic and strength variation with respect to anisotropy angle compared well with the analytical solutions. This study includes validation of discrete element model where the modeled strength and elastic behavior of shale closely matches with those of the laboratory observations. The DEM model was able to capture the key anisotropic characteristics observed in the laboratory test such as drastic reduction of elastic modulus and strength affected by transversely isotropic plane which is often the weakest and most deformable. Finally, the three-dimensional discrete element model with embedded smooth joints was successfully applied to simulate the laboratory hollow-cylinder testing, which provides an improved insight into the wellbore instability observed in situ.
Anisotropy of rock is generally observed in various rock types, e.g., shale that contains a set of parallel bedding planes. These weak cohesive planes contribute to large deformation and low shear strength, thereby resulting in unpredictable borehole damage especially when drilling sub-parallel to bedding planes (Økland et al., 1998; Zhang, 2013; Meier et al., 2015). Thus, the mechanical behavior of bedding planes with respect to their orientation should be properly considered in borehole stability analyses.
Discrete element method (DEM) particulate system is a powerful approach for simulating the material dynamic behavior under a given boundary condition without the need to establish the constitutive law (Potyondy and Cundall, 2004). The main advantage of bonded-particle DEM is its outstanding ability to monitor the evolving micromechanisms that can account for rock failure process. Furthermore, applications of bonded-particle DEM have been expanded into a complex rock mass modeling in larger scale in which natural fractures in rock mass have been taken into account using the discrete fracture network (DFN) model by means of smooth joint model (Mas Ivars et al., 2011). While previous studies on the fractured rock mass behaviors focused more on the natural fractures that are normally separated from each other with very low mechanical properties, there is a need for investigation regarding the behavior of weak cohesive planes within the intact rock, such as bedding planes that manifest transverse isotropy of rocks.