Shale mechanical properties are evaluated from laboratory tests after a complex workflow that covers tasks from sampling to testing. Due to the heterogeneous nature of shale, it is common to obtain inconsistent test results when evaluating the mechanical properties. In practice, this variation creates errors in numerical modeling when test results differ significantly, even when samples are from a similar core specimen. This is because the fundamental models are based on the supplied test data and a gap is, therefore, always observed during calibration. Thus, the overall goal of this study was to provide additional insight regarding the organization of the non-linear model input parameters in borehole simulations and to assist other researchers involved in the rock physics-related research fields. To achieve this goal, the following parallel activities were carried out: (1) perform triaxial testing with different sample orientations, i.e., 0°, 45°, 60°, and 90°, including the Brazilian test and CT scans, to obtain a reasonably accurate description of the anisotropic properties of shale; (2) apply an accurate interpretative method to evaluate the elastic moduli of shale; (3) evaluate and quantify the mechanical properties of shale by accounting for the beddings plane, variable confinement pressures, drained and undrained test mechanisms, and cyclic versus monotonic test effects. The experimental results indicate that shale has a significant level of heterogeneity. Postfailure analysis confirmed that the failure plane coincides nicely with the weak bedding plane. The drained Poisson’s ratios were, on average, 40 % or lower than the undrained rates. The drained Young’s modulus was approximately 48 % that of the undrained value. These mechanical properties were significantly impacted by the bedding plane orientation. Based on the Brazilian test, the predicted tensile strength perpendicular to the bedding plane was 12 % lower than the value obtained using the standard isotropic correlation test. The cyclic tests provided approximately 6 % higher rock strength than those predicted by the monotonic tests.
Read full abstract