Stress Inversion and Damage Quantification in Tight Gas Shale with Application to Hydraulic Fracturing

This thesis aims to advance the quantitative analysis of stress and failure process in tight gas shale under hydraulic fracturing by integrating stress inversion, microseismic monitoring, acoustic emission and discrete element modeling techniques. With the introduction of a modified Bott hypothesis that allows for out-of-plane slip, a stress inversion algorithm is developed accounting for tensile components of the source mechanism. Synthetic test datasets are employed to quantify the error in stress determination that arises when the conventional stress inversion based on the original Bott hypothesis is applied in the presence of non-double-couple sources. The proposed method is evaluated using microseismic data collected from Barnett Shale in the Fort Worth Basin, Texas. Results show the modified method introduces a roughly 15 degree correction as compared with the inversion results from the conventional algorithm. With the same dataset, a case study is conducted to investigate the dynamic interactions between injected fluids and hydraulic fractures through the spatial-temporal analysis of microseismicity. Two types of triggering front expansion patterns are evident. With the presence of a dominant hydraulic fracture, the radius of the triggering front expands linearly with time, and the microseismic event cloud forms a planar shape with low tensile components. On the other hand, in the case of a complex fracture network with the absence of any major hydraulic fracture, the triggering front grows non-linearly with time, which can be treated as equivalent to a diffusion model. The microseismic events exhibit more tensile components and an equidimensional event cloud. Two stages of the microseismic dataset are analyzed and the derived fracture widths and fluid-loss coefficients fall into a realistic range of general observations. Two acoustic emission (AE) laboratory experiments are carried out to examine the failure behavior of Montney shale samples under conventional triaxial compression and fluid injection. Detailed analysis for the triaxial compression test includes deformation-induced velocity anisotropy, source hypocenter determinations, source mechanism analysis, and stress inversions. For the hydraulic fracturing test, the mechanical correlation with the AE activity is analyzed. The AE locations correlate reasonably well with the spatial distribution of shear fracture and hydraulic fracture imaged by X-ray computer tomography (CT) scanning. However, the signal-to-noise ratio of the AE waveform emitted from Montney shale sample is relatively low, especially for the hydraulic fracturing test, which makes the data processing quite challenging. Distinct characteristics of the AE activity in Montney shale are identified, which are different from those in granite, a type of rock that has been more extensively investigated in the past. These differences could arise from the different stress settings, the low brittleness and stiffness of shale. Additionally, the AE test under triaxial compression is simulated using dynamic micromechanical models based on the discrete element method, in which the transversely isotropic feature of Montney shale and moment tensor calculations are considered. Calibration and verification are conducted against the results of previous laboratory experiments in terms of the anisotropic behavior. The model results indicate some of the low energy AE signals might not be captured by the experimental recording system, and it further demonstrates that the shear fracture initiation contains high tensile failure components.