Induced Microearthquakes and Seismicity-Permeability Relationships in Fractures
Open Access
- Author:
- Fang, Yi
- Graduate Program:
- Energy and Mineral Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 27, 2017
- Committee Members:
- Derek Elsworth, Dissertation Advisor/Co-Advisor
Derek Elsworth, Committee Chair/Co-Chair
Shimin Liu, Committee Member
Chris J Marone, Committee Member
Tong Qiu, Outside Member - Keywords:
- Induced microearthquakes
Frictional parameters
Permeability evolutions
Preexisting fractures
Fracture minerals
Fracture surface roughness
Enhanced Geothermal System
Geological sequestration of carbon dioxide - Abstract:
- In the past several years, induced microearthquakes (MEQs) related to energy development projects have garnered public attention. Large-scale fluid injection in geoengineering activities, such as enhanced geothermal reservoir stimulation, geological storage of CO2, shale reservoir stimulation, and deep disposal of wastewater can generate significant fluid overpressures and induce microearthquakes by reactivating preexisting faults and fractures that are widely distributed throughout the upper crust. The monitoring of fluid-injection-induced microearthquakes can provide significant information (e.g., timing, special distribution, and moment magnitude) in evaluating reservoir development (e.g., fracture distribution and reservoir hydraulic conductivity evolution), in informing production strategies (e.g., accommodation of production wells), and in assessing the risks of fluid injection activities (e.g., caprock integrity of CO2 reservoir). The following questions are addressed in this study: (1) what are the mechanisms and implications of in-situ feedbacks (e.g., monitored fluid-injection-induced MEQs distribution, seismic moment magnitudes, and fracture behaviors) in geothermal stimulation? (2) What information can be derived from MEQs to inform geothermal reservoir stimulation strategies? (3) Are there any potential relationships between induced seismic or aseismic slip and permeability evolution of fractures in such unconventional reservoirs and caprocks? (4) What factors play an important role in controlling these relationships? These questions are addressed in the five individual chapters of this thesis. Chapter 1 describes a case study of anomalous MEQs distribution: A bimodal depth distribution of fluid-injection-induced MEQs was observed in the 2012 stimulation phase of the Newberry Volcano EGS Demonstration project in Oregon, US. During 7 weeks of hydraulic stimulation of well NWG 55-29, 90% of MEQs occurred in the shallow reservoir (∼500 m to ∼1800 m), only a few occurred adjacent to the bottom of the open borehole (∼2500 m to ∼3000 m) while almost no seismicity was observed in the intervening interval (∼1800 m to ∼2500 m). Our analysis of frictional stability using spatial models for fluid pressure diffusion of injected fluids shows that the distribution of MEQs is consistent with observed casing damage, and a possible leak at ∼700 m, and is inconsistent with migration of fluids from the casing shoe. The role of fluid injection through the ruptured casing is further supported by the analyses of shear failure and pore-pressure diffusion. Finally, the absence of seismicity at intermediate depths is consistent with our laboratory determinations of frictional stability, showing velocity strengthening frictional behavior for samples from intermediate depths, bracketed by velocity neutral and weakening behavior for samples from shallower and greater depths. Chapter 2 introduces a method to constrain the evolution of fracture permeability at sufficiently fine resolution with observed in-situ MEQs data to define reservoir response. In this method, we propose a model that couples the moment magnitude to fracture aperture and then estimates the reservoir permeability at relatively high resolution. The critical parameters controlling fracture aperture and permeability evolution are stress-drop, the bulk modulus of the fracture embedded matrix, and the dilation angle of fractures. We employ Oda’s crack tensor theory and a cubic-law based analog to estimate the permeability of a synthetic fractured reservoir at various scales, demonstrating that the resolution of permeability is largely determined by the cellular grid size. Finally, we map the in-situ permeability of the Newberry EGS reservoir using observed MEQs during two rounds of reservoir stimulations in 2014. The equivalent mean permeability evaluated by each method is consistent and unlimited by representative elementary volume (REV) size. With identical parameters, Oda’s crack tensor theory produces a more accurate estimation of permeability than that of the cubic law method, but estimate differences are within one order of magnitude. The permeability maps show that the most permeable zone is located within the zone of most dense seismicity providing a reference for the siting of the production well. This model has the potential for mapping permeability evolution from MEQs data in conventional and unconventional reservoirs and at various scales. However, the impact of induced seismicity on fracture permeability evolution remains unclear due to the spectrum of modes of fault reactivation (e.g., stable vs. unstable). To better understand the hydro-mechanical behavior of reservoir due to stimulation, it becomes essential to understand the fundamental relationship between induced seismic and aseismic slip and permeability evolution of a fracture. As seismicity is controlled by the frictional response of fractures, Chapter 3 reports the results of experimental study of friction-stability-permeability relationships through the concurrent measurement of frictional and hydraulic properties of artificial fractures in Green River shale (GRS) and Opalinus shale (OPS). We observe that carbonate-rich GRS shows higher frictional strength but weak neutral frictional stability. The GRS fracture permeability declines during shearing while an increased sliding velocity reduces the rate of permeability decline. By comparison, the phyllosilicate-rich OPS has lower friction and strong stability while the fracture permeability is reduced due to the swelling behavior that dominates over the shearing induced permeability reduction. Hence, we conclude that the friction-stability-permeability relationship of a fracture is largely controlled by mineral composition, and that shale mineral compositions with strong frictional stability may be particularly subject to permanent permeability reduction during fluid infiltration. Chapter 4 extends previous studies to explore frictional stability-permeability relationships of fractures and identify the role of mineralogy (i.e., tectosilicate, carbonate, and phyllosilicate content). In this study, we perform a series of direct-shear experiments on saw-cut fractures of natural rocks and sintered fractures with distinct mineralogical compositions. Our experimental results indicate that the friction-permeability relationship is controlled by mineralogy. Frictional strength and change in permeability both decrease with an increase in either phyllosilicate or carbonate content as frictional instability (a-b) increases. With this relationship, we speculate that planar fractures with low frictional stability exhibit permeability enhancement after seismic slip in the frame of rate-state friction theory. This relationship implies a new mechanical-hydro-chemical (MHC) coupling loop via a linkage of frictional properties, mineralogy, and permeability. Previous experiments suggest that frictional strength and stability are primarily controlled by the mineralogical content of fracture material. The permeability of smooth fractures declines monotonically with displacement due to the generation of wear products. Chapter 5 investigates the effect of roughness on permeability evolution and frictional behavior using artificially fabricated fractures with specified roughness features. The experimental results show that (1) both smooth and rough fracture surfaces exhibit velocity strengthening frictional behavior for small net displacement and evolves to velocity neutral and velocity weakening with greater displacement. (2) Rougher surfaces exhibit higher velocity strengthening frictional behavior and higher frictional strength due to the presence of cohesive interlocking asperities during shearing. Seismicity may not be induced on rough fracture surfaces. (3) The roughness pattern exerts a dominant control on permeability evolution over the entire shearing history. Permeability evolves monotonically for smooth fractures but in a fluctuating pattern for highly roughened fractures. A higher roughness is likely to result in alternating compaction and dilation during shearing. Significant permeability damage may occur for rough samples when asperities are highly worn with wear products blocking fluid pathways. (4) There is no obvious correlation between permeability evolution and frictional behavior for rough fracture samples when fractures are subject to sudden sliding velocity change. Implications of our lab-scale experimental results suggest that characterization of fracture geometry would be beneficial for better understanding and managing induced seismicity and permeability development. In shale reservoir stimulation, fractures are propped to increase the permeability of the formation. On the other hand, the proppants may also influence the frictional strength of fractures. Thus, in the appendix, we explore the evolution of friction and permeability of a propped fracture using shearing-concurrent measurements of permeability during constant velocity shearing experiments. We observe that (1) the frictional response is mainly controlled by the normal stress and proppant thickness. High normal stress results in the crushing of proppant particles although this change in particle size distribution has almost no impact on the frictional response of the proppant-fracture system. The depth of shearing-concurrent striations on fracture surfaces suggests that the magnitude of proppant embedment is controlled by the applied normal stress. Moreover, under high normal stress, the reduced friction implies that shear slip is more likely to occur on propped fractures in deeper reservoirs. The increase in the number of proppant layers, from mono-layer to triple-layers, significantly increases the friction of the propped fracture due to the interlocking of the particles and jamming, suggesting that high proppant density during emplacement would help stabilize the fractures during injection. (2) Permeability of the propped fracture is mainly controlled by the magnitude of the normal stress, the proppant thickness, and the proppant size. Permeability of the propped fracture decreases during shearing due to proppant particle crushing and related clogging. Compared to the multi-layered specimen, the mono-layer case which has fewer displacement degrees-of-freedom exhibits the smallest initial permeability due to proppant embedment. Proppants are prone to crushing if the shear loading evolves concurrently with the normal loading. These combined conclusions suggest that the use of high-density proppants not only provides high hydraulic conductivity for hydrocarbon production but may also help to mitigate the risk of induced seismicity.