This study presents hollow-cylinder testing, a representative experimental approach for investigating borehole stability problems, using three-dimensional bonded-particle discrete element model (DEM) to identify failure mechanisms along the open hole by considering the effect of rock anisotropy. A transversely isotropic model was reproduced as an intact rock model embedding a smooth-joint contact model of which microparameters were calibrated against the laboratory results acquired from triaxial tests on overburden shale. A series of hollow-cylinder tests were performed on this transversely isotropic models that have five different inclined bedding planes, which were then compared to the laboratory observations. The result obtained from three-dimensional bonded-particle DEM model was able to capture the instability of the inner hole, i.e., a considerable reduction of effective confining pressure when the hole axis is sub-parallel to bedding planes. Furthermore, the DEM model manifested the overall failure patterns in the vicinity of the hole such as shear failure or spalling, which highly depend on the inclined angle of bedding planes.
Rock anisotropy is an important consideration in various engineering rock mechanics, e.g., borehole instability in shale formation especially when drilling sub-parallel to bedding planes (Økland et al., 1998). Thus, care must be taken when performing wellbore stability analyses on highly anisotropic rock formations. Of many forms of anisotropic models, a transversely isotropic model, which is pertinent to representing laminated rock formations (Jaeger et al., 2007), was adopted in this study.
Recent studies have showed that the two-dimensional bonded-particle discrete element model (DEM) can effectively simulate the elastic and strength behavior of transversely isotropic rocks (Chiu et al., 2013; Duan and Kwok, 2015; Park and Min, 2015; Wang et al., 2016). However, two-dimensional modeling is not sufficient to capture stress distributions and associated failures observed in near borehole, which is relied upon heavily to the degree of bedding inclination. Therefore, three-dimensional modeling is needed for a true representation of anisotropic rock.
The objective of the study is the three-dimensional bonded-particle DEM modeling of transversely isotropic rock to investigate the validity of hollow-cylinder test results by comparing the laboratory observations with the numerical results.