Anisotropic dynamic and static mechanical properties of organic-rich shale: The influence of stress

Abstract

Understanding the relationship between dynamic and static mechanical properties of organic-rich shales is crucial for successful in situ stress profile prediction and hydraulic fracturing stimulation in unconventional reservoirs. However, the relationship between dynamic and static properties remains ambiguous, considering the complex rock microstructure and subsurface stress environment. We have reported pseudotriaxial tests on a pair of outcrop Eagle Ford shale plugs, with the axial load applied perpendicular and parallel to bedding planes, to investigate the effects of intrinsic anisotropy and anisotropic stress on dynamic-static relationships. The bedding-parallel Young’s modulus is larger than the bedding-normal one dynamically and statically, whereas there exist complex relations among three static Poisson’s ratios, which are attributed to the intrinsic anisotropy induced by the lenticular texture and finely laminated alignment of kerogen. Along with a stress increment, static tests respond to superpositions of the elastic, viscoelastic, and nonelastic properties, whereas dynamic tests, with more than two orders of magnitude smaller strain amplitude, only reflect the elastic properties of rocks. As a result, the static properties characteristically exhibit more stress dependence than the dynamic properties. Moreover, the evolutions of static properties, especially two static Poisson’s ratios in the horizontal plug, are significantly influenced by the applied stress orientation with respect to the bedding plane. Lastly, we calculate four independent stiffnesses using the five static mechanical parameters with the assumption of transverse isotropy to compare with those calculated from ultrasonic velocities at different stress levels. Finally, when the deviatoric stress is approximately 20 MPa, static parameters derived from stress loading, unloading, and reloading almost intersect together. At this stress level, dynamic and static stiffnesses demonstrate a reasonable correlation with the fitting coefficient of approximately 1.4.