Search

Fluids injection for hydraulic stimulation and fracturing, typical in the development of enhanced geothermal systems (EGS) in granites, can reactivate deep faults and induce seismicity. Such faults typically contain chlorite coatings as an alteration product that may impact styles of deformation—aseismic through seismic. We performed low velocity shear experiments on simulated granite fault gouges under conditions typifying a geothermal reservoir at ∼4‐km depth with a confining pressure of 110 MPa, a temperature of 150°C, fluid pressures of 21–80 MPa, and chlorite contents of 0–100%, to investigate the influence of variation in effective stress and mineral composition on fault strength and stability. Our results show a transition from velocity‐strengthening to velocity‐weakening behavior in simulated granite gouge when the effective confining pressure was reduced from 89 to 30 MPa, characterized by a transition from fault compaction to dilation—as revealed by microstructural observations—with implications in enabling unstable failure. Conversely, increasing chlorite content stabilizes slip but reduces frictional strength. The microstructures of these mixed gouges exhibit shear localized on chlorite‐enriched planes and promoting fault sliding. These results suggest that earthquake ruptures occurring during fluid injection can be facilitated by effective stress variations and that both controlling fluid overpressures (effective stresses) and being aware of the presence of alteration minerals are both important controls in mitigating such injection‐induced seismic risks. Plain Language Summary: Enhanced geothermal systems (EGS) host an increasing number of induced earthquakes potentially linked to hydraulic stimulation. Rock cores recovered from geothermal reservoirs worldwide show an abundance of chlorite coatings on fault surfaces—present both natively and as a result of fluid circulation. To understand whether slip on deep fault will result in earthquakes, we measure the frictional properties of powdered granite fault rocks from the Pohang geothermal reservoir where an earthquake has occurred. We vary chlorite content and effective confining pressures and observe stable slip at higher effective confining pressures that transitions to unstable slip at lower effective confining pressures. Reducing the effective confining pressure destabilizes the fault behavior at in situ stress and temperature representative of the reservoir where an earthquake was observed. The addition of chlorite in the simulated gouge results in the opposite result, reducing the frictional strength but resulting in stable (aseismic) slip. Our results highlight the importance of effective stress variation and mineral composition in controlling fault frictional strength and stability, with their potential contributions to earthquake triggering. Key Points: Effective stress variation and chloritization exert competing controls on friction and stability of granite faultsLower effective stress accentuates velocity‐weakening behavior but retains high strengthConversely, chloritization reduces frictional strength but promotes stable failure [ABSTRACT FROM AUTHOR]

A shale fault reactivated during multistage hydraulic fracturing in the Changning block in the Sichuan Basin, southwest China, accompanied a cluster of small earthquakes with the largest reaching ML ~ 0.8. We illuminate the underlying mechanisms of fault reactivation through measurements of frictional properties on simulated fault gouge under hydrothermal conditions. Velocity‐stepping experiments were performed at a confining pressure of 60 MPa, temperatures from 30 to 300°C, pore fluid pressures from 10 to 55 MPa, and shear velocities between 0.122 and 1.22 μm/s. Results show that the gouge is frictionally strong with coefficient of friction of 0.6–0.7 across all experimental conditions. At observed in situ pore fluid pressure (30 MPa), the slip stability response is characterized by velocity strengthening at temperatures of 30–200°C and velocity weakening at temperatures of 250–300°C. Increasing the pore fluid pressure can increase values of (a − b) at temperatures ≥200°C, narrowing the temperature range where velocity weakening occurs. At the in situ temperature (90°C), the simulated gouge shows only velocity strengthening behavior and aseismic slip at elevated pore fluid pressures, contrary to the observed seismicity. We postulate that the aseismic slip at elevated pore fluid pressures may trigger seismicity by activating adjacent earthquake‐prone faults. Plain Language Summary: The Sichuan Basin of southwest China is the host to an increasing number of induced earthquakes potentially linked to the hydraulic fracturing for shale gas extraction. To understand whether the deep shale faults would slip unstably during hydraulic fracturing, we measure the frictional properties of powdered deep shale fault rocks (as simulated fault gouge) from a well in the Changning block in the Sichuan Basin which was identified with fault reactivation during hydraulic fracturing. We found that the simulated gouge slips stably at lower temperatures but unstably at higher temperatures. Elevating the pore fluid pressure stabilizes the fault slip at in situ and higher temperatures, contrary to the field observations. We postulate that the shale fault is prone to stable slip at higher pore fluid pressure, but this slip further can lead to the slip of adjacent unstable faults. Our results highlight the importance of combined temperature and pore fluid pressure effects on assessing the potential of induced seismicity from fluid injection activities. Key Points: Frictional properties of a shale fault reactivated by hydraulic fracturing are measured using simulated gouge under hydrothermal conditionsGouge favors aseismic creep at in situ temperature and remains stable at elevated pore fluid pressure, contrary to observed seismicityImportance of combined temperature and pore fluid pressure effects on assessing the potential of injection induced seismicity is highlighted [ABSTRACT FROM AUTHOR]

Longmaxi formation shales are the major target reservoir for shale gas extraction in Sichuan Basin, southwest China. Swarms of earthquakes accompanying hydraulic fracturing are observed at depths typifying the Longmaxi formation. Mineral composition varies broadly through the stratigraphic section due to different depositional environments. The section is generally tectosilicate‐poor and phyllosilicate‐rich with a minor portion the converse. We measure the frictional and stability properties of shale gouges taken from the full stratigraphic section at conditions typifying the reservoir depth. Velocity‐stepping experiments were performed on representative shale gouges at a confining pressure of 60 MPa, pore fluid pressure of 30 MPa, and temperature of 150°C. Results show the gouges are generally frictionally strong with friction coefficients ranging between 0.50 and 0.75. Two phyllosilicate + TOC (total organic carbon)‐poor gouges exhibited higher frictional strength and velocity weakening, capable of potentially unstable fault slip, while only velocity strengthening was observed for the remaining phyllosilicate + TOC‐rich gouges. These results confirm that the frictional and stability properties are mainly controlled by phyllosilicate + TOC content. Elevating the temperature further weakens the gouges and drives it toward velocity weakening. The presence of observed seismicity in a majority of velocity‐strengthening materials suggests the importance of the velocity‐weakening materials. We suggest a model where seismicity is triggered when pore fluid pressures drive aseismic slip and triggers seismic slip on adjacent faults in the same formation and distant faults in the formations above/below. The effect of pore pressure transients within low‐permeability shale gouges is incorporated. Our results highlight the importance of understanding mechanisms of induced earthquakes and characterizing fault properties prior to hydraulic fracturing. Key Points: Longmaxi formation shales show a strong variation in mineral composition due to the different depositional environmentsA small proportion of phyllosilicate‐poor gouges exhibit velocity weakening at hydrothermal conditions, indicative of potentially unstable slipAseismic fault slip during hydraulic fracturing may reactivate adjacent and distal unstable faults and trigger the seismicity [ABSTRACT FROM AUTHOR]

The injection‐induced slip of a fault containing along‐the‐fault geometrical complexities such as permeable cracks, dilational jogs and branches, was studied numerically using a plane‐strain hydraulic fracture model. The fault is modeled by placing complexities evenly spaced on either side of the inlet where fluid coming from a steady source is injected at a constant rate. The fault slip obeys the Coulomb friction law with slip weakening of the coefficient of friction. The applied shear stress is less than the residual fault strength, corresponding to a situation where a fault undergoes stable slip behind a slowly advancing rupture front. The numerical results show that the segments of complexity may not only delay slip zone extension and fluid flow, but temporarily increase slip rates. Short‐term faster slip rates occur after cracks and jogs are pressurized. The slip rate can reach values typical of microearthquake events, accompanied by rising and then dropping pressure and producing a stress release. Especially, the discontinuous slip sources are separated by the right‐angle jogs that do not slip. The slip on branches can be activated by fluid invasion and an associated normal effective stress reduction, and the slow slip sources eventually move from the fault to the branches. Even if a hydraulic fracturing treatment presents a low risk of generating dynamic slip, when fractures intersect faults containing geometrical complexities, fast slip events may be induced. The spatiotemporally varying slip rates and patterns presented provide an alternative interpretation for recorded seismic slip signals. Key Points: Along‐the‐fault geometrical complexities affect the slip location and rate.Short‐term fast slip rates can be generated by along‐the‐fault dilational segments.Fast slip duration and spatial pattern from seismicity may provide an estimate of rupture geometries. [ABSTRACT FROM AUTHOR]

The growing hazard of induced seismicity driven by the boom in unconventional resources exploitation is strongly linked to fault activation. We perform laboratory measurements on simulated fault gouges comprising powdered reservoir rocks from major oil and gas production sites in China, to probe the control of mineral composition on fault friction and stability responses during reservoir stimulation. Double direct shear experiments were conducted on gouges with phyllosilicate content ranging from 0 to 30 wt.% and grain sizes <150 μm, at constant normal stresses of 10–40 MPa and conditions of room temperature and water saturation. The velocity step and slide‐hold‐slide sequences were employed to evaluate frictional stability and static healing, respectively. Results indicate that the mineralogy of the gouges exhibit a strong control on the frictional strength, stability, and healing. Phyllosilicate‐rich samples show lower frictional strength μ and higher values of (a − b), promoting stable sliding. For the gouges studied, the frictional strength decreases monotonically with increasing phyllosilicate content, and a transition from velocity weakening to velocity strengthening behavior is evident at 15 wt.% phyllosilicates. Intermediate healing rates are common in gouges with higher content of phyllosilicates, with high healing rates predominantly in phyllosilicate‐poor gouges. As an indispensable component in reservoir rocks, the carbonates are shown to affect both the frictional stability and healing response. These findings can have important implications for understanding the effects of mineralogy on fault behavior and induced seismic potential in geoengineering activities, particularly in reservoirs in China. Key Points: Frictional strength and stability of powdered gouges are affected by the relative amount of tectosilicates, phyllosilicates, and carbonatesGouges with less phyllosilicates exhibit higher healing ratesThe presence of carbonates can affect gouge frictional stability and healing [ABSTRACT FROM AUTHOR]

• The Gonghe geothermal reservoir comprises granites dominated by quartz and feldspar but with high chlorite content. • Granitic gouges tend to show velocity-neutral to slight velocity-weakening behaviors at in-situ hydrothermal conditions. • Instability on granite faults has important implications for understanding seismicity observed during hydraulic fracturing. The Gonghe Basin in northwest China has significant potential for the recovery of deep geothermal fluids. However, a large number of earthquakes were observed during stimulation by hydraulic fracturing with the maximum magnitude reaching ∼ M L 2. To understand the mechanisms of deep fault stability and the clusters of earthquakes, we recovered seven granite cores from the Gonghe geothermal reservoir and powdered them to simulate fault gouges. XRD results show that the granites are dominated by quartz and feldspar and show a high chlorite content. Triaxial shear experiments were conducted on the seven gouges to explore the frictional and stability properties at conditions typifying the depths of 2450–3600 m in the Gonghe geothermal site. Results show that the friction coefficients of the tested gouges are high and close to 0.70. At hydrothermal conditions, all gouges show slight velocity-weakening to velocity-neutral behavior, which is indicative of potentially unstable fault slip. Our results have important implications for understanding the fault stability behavior and induced seismicity during hydraulic fracturing in granite geothermal reservoirs. [ABSTRACT FROM AUTHOR]

Swarms of earthquakes during shale gas exploitation in the Changning area of Sichuan Basin indicate that hydraulic fracturing induces seismicity both within the target reservoir but also to depths of several kilometers below the horizontal well. These remote earthquakes are possibly triggered by total stress perturbations resulting from the hydraulic fracturing. We use a dislocation-based analytical model to simulate multistage hydraulic fracturing of three horizontal wells at a single well pad to explore the spatiotemporal evolution of total stress perturbations. Results show that the number and distribution of fracturing stages affect both the distribution and magnitude of stress changes and that the stress change diminishes with distance. The undrained injection-induced stress change is below 10-3 MPa at distances ≥1 km for first-stage fracturing but reach 10-1 MPa for multistage fracturing of 30 stages in three wells. Undrained stress changes scale linearly with the magnitude of fluid leakoff into the formation – halving the effective fracture width halves the induced stress magnitudes and with an identical distribution – limiting the potential for fault reactivation. Scaling analysis for pressure diffusion distal from the reservoir indicate that the short-term impact is indeed essentially undrained. Estimates for long-term depletion identify a similar induced stress signal of opposite sign but with similar Coulomb potential for reactivation in the long-term. Such magnitudes of Coulomb stress changes suggest the possibility of fault reactivation on critically-stressed faults at kilometer separation from the injection both in the short-term due to stimulation and in the long-term resulting from depletion. • A dislocation theory based model defines the spatiotemporal evolution of stress perturbations after hydraulic fracturing. • Multistage fracturing accumulates a stress perturbation with a maximum amplitude of ~0.1 MPa at a distance of 1 km. • Fracturing as a single well pad can result in a Coulomb failure stress capable of reactivating critically-stressed faults. [ABSTRACT FROM AUTHOR]

A multi-scale numerical investigation for fault slip and casing deformation is carried out by combining a large-scale distinct element model to simulate the hydraulic fracturing and fault slip and a small-scale finite difference model to analyze casing deformation. The result of large-scale model shows that the pore pressure elevation causes the slippage of fault and thus the deformation of an intersected wellbore and casing. Both fluid viscosity and injection rate have significant impacts on the shear displacement while the effect of injection volume is determined by the number of natural fractures in the reservoir. The result of small-scale model shows that the fault sliding causes the stress concentration near the fault plane and thus gives rise to the casing yielding. Casing grade, casing sheath thickness and cement sheath thickness have limited impacts on casing deformation while the countermeasure of no cementing is effective on mitigating casing deformation. Moreover, optimizing the well trajectory can help to decrease casing deformation. The field data in Sichuan Basin demonstrates that decreasing the injection rate could effectively mitigate the casing deformation. But the effect of total injection volume is less significant probably due to the fact that the shale reservoir is intensively fractured. • A large-scale distinct element model is utilized for simulating hydraulic fracturing and fault slip. • A small-scale finite difference model is applied for analyzing casing deformation. • Casing grade, casing sheath thickness and cement sheath thickness have limited impacts on casing deformation. • Casing deformation can be mitigated by optimizing well trajectory or decreasing injection rate. [ABSTRACT FROM AUTHOR]

The number of background earthquakes (M L ≥ 0) in the southern Sichuan basin, southwest China, has increased thirtyfold as a result of hydraulic fracturing. Background events are originally deep (4–6 km) within the sedimentary section but build into a multimodal distribution both at depth and in the shallow stimulated reservoir (2–4 km) - representing a counterpoint to the usual triggering of seismicity on deep sub-reservoir basement faults. Surprisingly, the largest events (M L ≥ 3) evolve in the deep sedimentary strata (4–6 km) that are hydraulically isolated from the injection zone (2–4 km) by low permeability layers. We evaluate the friction-stability rheology of the strata within the full stratigraphic section to define the feasibility of nucleation within these shallow and deep strata. These show velocity-neutral to velocity-weakening behavior in the shallow reservoir transitioning to more strongly velocity-weakening with increase in both depth and temperature. Poroelastic stress calculations confirms that stress transfer, rather than transmitted fluid pressures, are capable of directly reactivating critically-stressed faults at depth, with fluid pressures the triggering source within the shallow reservoir. • HF results in significant increase in seismic event rates at two distinct depths – in and below the reservoir zone. • Direct fluid diffusion and effective stress reduction nucleate earthquakes near the fracturing zone. • Earthquake-prone sedimentary layers host earthquakes in deep & hydraulically isolated zone by poroelastic stress transfer. [ABSTRACT FROM AUTHOR]

The effects of rock inherent heterogeneity and grain size on hydraulic fracture initiation and propagation for different propagation regimes are investigated through two dimensional discrete element modeling. Random particle assembly is used to mimic rock inherent heterogeneity in the numerical model while regular particle assembly is used as the reference. The rock inherent heterogeneity mainly affects the hydraulic fracture net pressure in the viscosity dominated regime and the effect is more profound in the toughness dominated regime. In the toughness dominated regime, in addition to the increase of net pressure relative to the regular particle sample, the hydraulic fracture profiles in the random particle sample also show larger tortuosity and asymmetry caused by the local heterogeneity, and the fracture growth of one of the wings can be temporarily arrested. Numerical simulations show that the effective toughness of the random particle sample is larger than that of the regular particle sample. This is caused by tortuosity, in which case the net pressure in the random particle sample is also affected by the local geometrical arrangements of the particles. Also, the apparent toughness is influenced by the magnitude of initial stress, which comes in addition to the tensile strength of the contact bond and the particle radius. The effect of stress anisotropy has limited effect on the hydraulic fracture propagation for both the viscosity and toughness dominated regimes. [ABSTRACT FROM AUTHOR]

Abstract This paper introduces a new method to predict conductivity in channel fracturing by implementing analytical solution in Discrete Element Method (DEM). First, an analytical model for channel fracturing conductivity is proposed. Then, a DEM model calibrated by using experimental results is set up to investigate the deformation of proppant pillar. Finally, the analytical model of fracture conductivity is implemented in the DEM model to predict conductivity during fracture closing. Parametric analyses are carried out to understand the effects of four factors: proppant size combination, concentration, time ratio τ and elastic modulus to stress ratio λ. Conductivity generally decreases during fracture closing and increases with the increasing proppant size and proppant concentration. The ratio of the pulsing time of proppant laden fluid to the pulsing time of the clean fluid is the key parameter for the field operation. A large time ratio could enhance the pillar stability though it may lead to the damage of fracture conductivity. For given rock modulus, closing stress and proppant size, the optimal range of time ratio can be given as a guidance for field operation. The field application of channel fracturing in the Shengli Oilfield proves that the optimized range of time ratio τ based on the proposed theoretical model is valid. The porosity of proppant pillar first decreases due to compression and then increases during fracture closing period because of the breakdown of the proppant pillar and the resulted particle movement outward. This paper gives insights for understanding the channel fracture conductivity and provides a practical tool for the optimization of channel fracturing design in the field. Highlights • During closing, the pillar porosity first decreases and then increases while the fracture conductivity generally decreases. • The ratio of the pulsing time of proppant laden fluid to the pulsing time of the clean fluid is a key parameter. • For given rock modulus, closing stress and proppant size, the range of time ratio can be optimized to guide field operation. [ABSTRACT FROM AUTHOR]

Injecting proppant to prop open fluid-driven fractures in subsurface reservoirs is one of the key missions of hydraulic fracturing. However, quantitative evaluation of the distribution of successfully propped fractures is limited due to the infeasibility of direct measurement. This work defines an indexing parameter for field practice to estimate the proportion of proppant-filled fractures in the reservoir – the proppant filling index (PFI). A new data-driven workflow, combining numerical models and an ensemble learning algorithm, is proposed and trained on field records of both screen-out and near screen-out cases and is then applied to predict PFIs for regular cases. The algorithm performance is promoted via variable importance measure (VIM) analyses and a backward elimination strategy. Four screen-out and twelve regular cases are presented to demonstrate the predicted PFI and highlight its potential utilizations. The new PFI and workflow evaluate the proppant injection quantitatively and reveal any mismatch between proppant injection and underground fractures, which may be essential for post-fracturing analyses and reservoir characterization to improve both oil & gas recovery, the sequestration of CO 2 , storage then recovery of H 2 and the recovery of deep geothermal fluids as important components in enabling the energy transition. • An indexing parameter (PFI, the proportion of proppant-filled fractures) is defined. • PFI reveals the mismatch between proppant injection and fractures in reservoirs. • An ensemble learning workflow is built to predict PFI for field practice. • Proppant injection is quantitatively evaluated for post-fracturing analysis. • Fracture evolution is qualitatively described for characterizing fractured reservoirs. [ABSTRACT FROM AUTHOR]

This work extends the hydro-mechanical phase-field fracture model to non-isothermal conditions with micromechanics based poroelasticity, which degrades Biot's coefficient not only with the phase-field variable (damage) but also with the energy decomposition scheme. Furthermore, we propose a new approach to update porosity solely determined by the strain change rather than damage evolution as in the existing models. As such, these poroelastic behaviors of Biot's coefficient and the porosity dictate Biot's modulus and the thermal expansion coefficient. For numerical implementation, we employ an isotropic diffusion method to stabilize the advection-dominated heat flux and adapt the fixed stress split method to account for the thermal stress. We verify our model against a series of analytical solutions such as Terzaghi's consolidation, thermal consolidation, and the plane strain hydraulic fracture propagation, known as the KGD fracture. Finally, numerical experiments demonstrate the effectiveness of the stabilization method and intricate thermo-hydro-mechanical interactions during hydraulic fracturing with and without a pre-existing weak interface. [ABSTRACT FROM AUTHOR]

This work introduces 3D lattice modeling of hydraulic fracturing initiation and near-wellbore propagation for different perforation models, including spiral perforation, oriented perforation and Tristim perforation. For each perforation model, a total of six perforation tunnels are explicitly modeled and the representative intermediate states are chosen to analyze the results. The numerical simulation results show that the perforation tunnels guide the generation of initial microcracks at the roots of the perforation tunnels once the injection starts, however, the subsequent fracture propagation is controlled by the relative locations of perforation tunnels and the stress interference among different perforation tunnels. The spiral perforation gives the highest breakdown pressure while the Tristim perforation gives the lowest. Trans-wellbore fracture surfaces caused by the generation of large amount of microannulus cracks is the main reason of the pressure breakdown. The magnitude of the breakdown pressure is associated with the level of fracture complexity generated before the breakdown. Despite the variation of breakdown pressure for different perforation models, both the initial minimum pressure after the breakdown and the propagation pressure are nearly identical. The study provides a theoretical basis for understanding fracture initiation and near-wellbore tortuosity. • Peform 3D lattice modeling of fracturing initiation and near-wellbore propagation for different perforation models. • Reveal the micro-scale mechanism of near-wellbore complexity through the explicit modeling of hydraulic fracturing. • Provide a theoretical basis for understanding fracture initiation and near-wellbore tortuosity. [ABSTRACT FROM AUTHOR]

An accurate understanding on hydraulic behaviors of fluid flow through fractured rocks is of central importance for both underground resources exploitation and CO 2 geological sequestration. Despite its importance, fluid flow through fractures surrounded by permeable matrix is rarely investigated. Here, this study aims to explore fluid flow through fractures surrounded by permeable host rocks using a combined experimental and numerical approach. In particular, core-flooding experiments were sequentially conducted on both intact and fractured cores subjected to different rock types, under varying confining stress and fracture roughness. According to the experimental results, fracture permeability decreases rapidly with increasing confining stress and fracture roughness, while matrix permeability is less affected by confining pressure, resulting in an increment in contributions of matrix flow to the overall flow with increased confining pressure. In addition, numerical simulations were performed on digital fractured rocks with a broadened range of roughness from 2.58 to 18.2, and under confining pressure ranging from 2 MPa to 53 MPa. The fluid flows within matrix and fracture were assumed to follow linear Darcy's law and nonlinear Forchheimer's law, respectively. The simulated results also show that K m / K f increases with the increase of confining pressure, reflecting the enhancement in contributions of matrix permeability to the overall transportation. If K m / K f = 0.1 is defined as a threshold over which matrix permeability should not be ignored, then the critical confining pressure of 18 MPa, 20 MPa, and 46 MPa is determined for mudstone, sandstone and limestone, respectively. In addition, the fluid flow through rough-walled fractures is observed to be susceptible to normal loading and fracture roughness, which directly alters flow paths within the fractures. With an increment in confining pressure and fracture roughness, the contact area between rough fracture surfaces grows, promoting flow tortuosity and generating eddies inside fracture, which consequently increases flow resistance and hence weakens fracture permeability. Ultimately, by fitting the simulated data, an empirical equation linking fracture permeability with confining stress, fracture roughness and hydraulic gradient is provided. This study provides quantitative characterizations on fluid flow through fractures accounting for permeable matrix, and thus can potentially improve the accuracy in predicting hydraulic properties of fractured rocks. • Fluid flow in fractures surrounded by permeable rocks are systematically explored. • Contributions of matrix permeability to the fracture flow is quantified. • An empirical equation is proposed for quantitative characterizations of fracture flow. [ABSTRACT FROM AUTHOR]

• A grain-based DEM coupled with a pore network model is established to investigate the hydraulic fracture growth in polymineralic rock. • A new algorithm for the toughness-dominated regime is proposed to simulate SC-CO 2 fracturing. • The effects of fracturing fluids and inherent microstructures of rock on hydraulic fracture growth at grain-scale are investigated. Rock is a typical heterogeneous material composed of inherent microstructures at the grain scale. In this paper, a grain-based discrete element model (DEM) coupled with a pore network model is developed to study the interaction behavior between hydraulic fractures and the inherent microstructures of rocks. The numerical model parameters are calibrated using the experimental results on Pocheon granite, and then the model is validated by the plane strain Khristianovic-Geertsma-de Klerk (KGD) analytical solution. The fracture propagation in polymineralic rock involves many unique phenomena at the grain scale, such as intragranular fractures splitting grains, intergranular fractures along grain boundaries, fluid lag, and rock fragments. Compared with high viscosity fluid, supercritical carbon dioxide (SC-CO 2) driven fractures tend to separate grain boundaries with low local resistance and propagate less smoothly and continuously, more asymmetrically and tortuously. Furthermore, the uniqueness of microstructures controlled by mineral distribution can lead to large variability in fracture paths, but the fracture properties at the macro-scale have not much difference. As the intergranular bonding strength or average grain size increases, a transition from intergranular fracture-dominated to intragranular fracture-dominated is reported. In addition, it is found that the existence of weak grains can result in more intragranular fractures within weak grains and fewer intergranular fractures associated with weak grains. Rock fragments are likely created as a result of the interaction between hydraulic fractures and weak grains and grain boundaries. Our results show that the low-viscosity fracturing fluid and the strong micro heterogeneity of microstructures are prone to result in complex fracture propagation. [ABSTRACT FROM AUTHOR]

A propagating fluid driven fracture in a rock mass is expected to interact with geological interfaces on a wide variety of length scales. The vertical growth of hydraulic fractures in layered rocks is of pivotal importance for the successful stimulation in reservoir development. In this study, 2D discrete element modeling is used to investigate the influence of the stiffness and toughness ratio, as well as stress contrast between layers on the hydraulic fracture height growth. In particular, the ultimate goal is to better understand mechanisms of the fracture height containment by contrasts of different rock properties and to quantitatively determine which parameters provide a stronger influence. In addition, the analysis is performed in the context of hydraulic fracture regimes, whereby the dominant dissipation mechanism in the system can either be associated with fracture toughness or viscous fluid flow. As a starting point, we investigated the propagation of a plane strain hydraulic fracture from a low stiffness layer to a high stiffness layer and vice versa, while keeping the stress constant. The influence of stress on hydraulic fracture propagation in layered rocks is investigated afterwards. The numerical results demonstrate that the hydraulic fracture can either directly pass through the geological interface, be arrested at the interface, or stop before reaching the interface. The interface itself is assumed to be perfectly bonded, therefore no slippage is considered. Ability of the hydraulic fracture to approach the interface is first determined by the elastic modulus ratio of the two adjacent layers. Once reached the interface, the further growth is then affected by the toughness ratio between the layers. After that, if the fracture crosses the interface, then it is affected by the stress contrast. The propagation regime has an important influence on the fracture propagation in layered rocks. If the propagation regime is closer to the viscosity dominated, the hydraulic fracture is likely to cross the interface. In contrast, it is more difficult for a fracture to cross the barrier if the propagation regime is near the toughness dominated. A map of fracture crossing behavior versus propagation regime and contrast in properties has been constructed, that can be used to quantify strength of mechanical barriers and to deduce hydraulic fracture height growth behavior for various scenarios. [ABSTRACT FROM AUTHOR]