Book Description

During hydraulic fracturing in unconventional tight formations a high percentage of the injected fluid may remain in the formation and only a small portion of the fracturing fluid is typically recovered. Although spontaneous imbibition is mainly introduced as the main dominating mechanism, a clear understanding of the fundamental mechanisms through which the fracturing fluid would interact with the formation remains a challenge. The impact of these mechanisms on rock property changes is even more challenging but is important to account for post-fracturing reservoir characterization. In this study, an integrated analytical-experimental-numerical approach was adopted to study these issues using a case study within the Montney Formation in Farrell Creek field in northeast British Columbia. The results of experiments on Montney samples from different depths revealed that because of spontaneous water imbibition, the geomechanical properties of the samples were altered. Also, small scale heterogeneity in tight gas formations and shale results in these property changes occurring at various scales, such as beds. Property changes occurring along the beds and bedding planes, as a result of interaction with hydraulic fracturing fluid, can contribute to increased potential for shear failure along these planes. Therefore, a systematic micro-scale analysis (including micro-indentation and micro-scratch along the beds to capture micro-geomechanical responses) and macro-scale analysis (including ultrasonic measurements, uniaxial compressive loading in high and low capillary suctions and unloading-reloading cycles at varying capillary suction) have been developed and applied to capture the changes in rock behavior in different scales as a result of spontaneous water imbibition and how different behaviors in micro-scale would affect the responses in macro-scale. QEMSCAN analysis, nitrogen adsorption-desorption tests, thermogravimetric analysis (TGA), capillary condensation experiments, pressure-decay and pulse-decay permeability measurements and direct shear tests were also completed for quantitative analysis of minerals, pore shapes and porosity, initial water saturation, capillary suction as a function of water saturation, permeability and strength parameters in both macro-scale and micro-scale (bed-scale). QEMSCAN analysis indicated that mineral components were not the same in different beds and they could be categorized into quartz-rich and clay-rich. The results of the experimental phase indicated that the geomechanical and flow properties of Montney specimens were altered due to fluid imbibition. As the water saturation and capillary suction were changing in quartz-rich and clay-rich beds, they responded differently which would trigger some geomechanical behaviors in macro scale. In addition, it was observed that capillary suction would add extra stiffness and strength to the media and as it was diminishing, the media became weaker. A nonlinear response with hysteresis during unloading-reloading cycles at varying capillary suction implied that as a result of the water softening effect, the reduction in capillary suction and changing the local effective stress there is a high possibility of activation and propagation of pre-existing micro fractures. In the numerical modeling phase of this research, fully coupled poro-elastoplastic partially saturated models were developed that included transversely isotropic matrix properties and bed-scale geometry. Inclusion of bed-scale features in the numerical approach provided better analysis options since different properties of the adjacent beds (including different capillary suction change) that can trigger the failure in the planes of weakness (such as the interface between the beds) can be directly included in the model while it is not possible to have that in transversely isotropic numerical modeling. This implies that conventional numerical analysis of geomechanical responses originated from spontaneous imbibition needs to be revisited. Beds-included numerical analyses indicated that since the changes in local effective stress and rock mechanical properties were not the same in adjacent quartz-rich and clay-rich beds, differential volumetric strain along the interfaces between quartz-rich and clay-rich beds would take place which in turn generated induced shear stress components on the interface planes. For the interfaces where total shear stress along them exceeded the shear strength, failure occurred. Comparing the result of micro-geomechanical (bed scale) and macro-geomechanical analysis with the results of numerical modeling at reservoir in-situ conditions would suggest that as a result of post-fracturing spontaneous water imbibition in the studied Montney Formation, the failures/micro fractures would be generated along the interfaces. Then because of the propagation of activated pre-existing micro fractures in the adjacent beds followed by coalescence with the failed interfaces, a complex micro fracture network can be formed. Accordingly, rock mass geomechanical responses and flow properties would be affected which means that any numerical modeling or analytical approach to account for the production, refracturing and any other reservoir-related analysis without considering this fact is under question mark.