In petroleum engineering and geophysics, it is often assumed that the rocks are completely rigid bodies with a totally interconnected pore space and that the fluid within the pores does not affect and are independent of the strains in the porous media. These assumptions are often not accurate and also unrealistic because the pore pressure effects are of great importance in all of the geomechanical processes occurring in the subsurface. The hydraulic and mechanical processes are coupled so that the rock deformation causes pore pressure changes and fluid flow (displacement relative to the solid). The time- dependent coupling of the hydraulic and mechanical processes can be described by the theory of poroelasticity. Application of this theory requires the availability of material parameters through experiments. In this work, the main poroelastic parameters are determined for some rock types of interest. The focus of this work is concentrated in low porosity rocks that are commonly encountered. Experimental procedures under drained, undrained and unjacketed conditions were initially completed in Berea Sandstone. Then, Indiana Limestone, Westerly Granite and Welded Tuff specimens were tested in order to obtain Skempton’s pore pressure parameter B, Biot’s coefficient of effective stress α, Bulk Modulus and Grain compressibility. Preliminary results suggest that the parameters B, K and α will change in accordance to the permeability and the porosity of the rock, while K_(S) would depend more on the mineralogy and deposition characteristics of the rock.

Hydraulic fracturing of a naturally-fractured reservoir is a challenge for industry, as fractures can have complex growth patterns when propagating in systems of natural fractures in the reservoir. Fracture propagation near a natural fracture (NF) considering interaction between a hydraulic fracture (HF) and a pre-existing NF, has been investigated comprehensively using a two dimensional Displacement Discontinuity Method (DDM) Model in this thesis. The rock is first considered as an elastic impermeable medium (with no leakoff), and then the effects of pore pressure change as a result of leakoff of fracturing fluid are considered. A uniform pressure fluid model and a Newtonian fluid flow model are used to calculate the fluid flow, fluid pressure and width distribution along the fracture. Joint elements are implemented to describe different NF contact modes (stick, slip, and open mode). The structural criterion is used for predicting the direction and mode of fracture propagation. The numerical model was used to first examine the mechanical response of the NF to predict potential reactivation of the NF and the resultant probable location for fracture re-initiation. Results demonstrate that: 1) Before the HF reaches a NF, the possibility of fracture re-initiation across the NF and with an offset is enhanced when the NF has weaker interfaces; 2) During the stage of fluid infiltration along the NF, a maximum tensile stress peak can be generated at the end of the opening zone along the NF ahead of the fluid front; 3) Poroelastic effects, arising from fluid diffusion into the rock deformation can induce closure and compressive stress at the center of the NF ahead of the HF tip before HF arrival. Upon coalescence when fluid flows along the NF, the poroelastic effects tend to reduce the value of the HF aperture and this decreases the tension peak and the possibility of fracture re-initiation with time. Next, HF trajectories near a NF were examined prior to coalesce with the NF using different joint, rock and fluid properties. Our analysis shows that: 1) Hydraulic fracture trajectories near a NF may bend and deviate from the direction of the maximum horizontal stress when using a joint model that includes initial joint deformation; 2) Hydraulic fractures propagating with higher injection rate or fracturing fluid of higher viscosity propagate longer distance when turning to the direction of maximum horizontal stress; 3) Fracture trajectories are less dependent on injection rate or fluid viscosity when using a joint model that includes initial joint deformation; whereas, they are more dominated by injection rate and fluid viscosity when using a joint model that excludes initial joint deformation.

Natural hydraulic fractures (NHFs) are fractures whose growths are driven by fluid loading. The fluid flow properties of the host rock have a primary, but hitherto little appreciated control on the NHF propagation rates. This study focuses on investigating the impacts of host rock fluid flow on the propagation and pattern development of multiple NHF in a poroelastic media. A realistic geomechanical model is developed to combine both the fluid flow and mechanical interactions between multiple fractures. The natural hydraulic fracture propagation is observed to consist of a series of crack-seal processes indicating incremental stop-start growth. Growth timing is on the scale of millions of years based on recent natural fracture growth reconstructions. These time scales are compatible with some model scenarios. My newly developed numerical model captures the crack-seal process for multiple NHF propagation. A sensitivity study conducted to investigate the impacts of different fluid flow properties on NHF propagation shows that permeability is a predominate influence on the timescale of NHF development. In low-permeability rocks, fractures have more stable initiation and much longer propagation timing compared to those in high-permeability rocks. Another aspect of great interest is the influence of fluid flow on fracture spacing and pattern development for multiple NHFs propagation in a poroelastic environment. My new poroelastic geomechanial model combines the natural hydraulic fracturing mechanism with the mechanical interactions between fractures. The numerical results show that as host rock permeability decreases, more fractures can propagate and a much smaller spacing is reached for a given fracture set. The low permeability slows down the propagation of long fractures and prevents them from dominating the fracture pattern. As a result, more fractures are able to grow at a similar speed and a more closely spaced fracture pattern is achieved for either regularly spaced or randomly distributed multiple fractures in low-permeability rocks. Investigation is also conducted in analyzing the distributions of fracture attributes (length, aperture and spacing) in low- and high-permeability rocks. For shales with high subcritical index, low permeability helps the fractures propagate more closely spaced instead of clustering. Meanwhile, in low-permeability rocks, factures have relatively smaller apertures, which lead to a slower fracture opening rate. The competition between the slow fracture opening rate and quartz precipitation rate will affect the effective permeability and porosity of the naturally fractured reservoir. However, the competition is trivial in high-permeability rocks. Other factors, such as reservoir boundary condition, layer thickness, subcritical index and pattern development stage, all have considerable impact on fracture pattern development and attribute distribution in a poroelastic media. / text

京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第21580号 / 理博第4487号 / 新制||理||1644(附属図書館) / 京都大学大学院理学研究科地球惑星科学専攻 / (主査)准教授 宮﨑 真一, 教授 福田 洋一, 教授 橋本 学 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DGAM

The purpose of this research was to study perfusion pressure-flow relationships in self-oscillating synthetic poroelastic vocal fold (VF) models before, during, and after vibration. This was accomplished by developing a custom ultra-soft poroelastic material, incorporating the poroelastic material as the cover layer in a synthetic VF model, and studying the model vibratory response and the flow rate of fluid perfused through the cover layer while undergoing flow-induced vibration. The custom ultra-soft poroelastic material was developed using the method of direct templating with sucrose spheres as the sacrificial template and silicone as the infiltration medium. The average modulus of elasticity of the poroelastic material was found to be 3.30 kPa, which represented an 84% decrease compared to the same non-porous silicone. Porosities between 62.8% and 67.2% were estimated. The fabrication process of the poroelastic VF model is presented in detail, including steps to prepare the model for vibration. The apparatus for measuring perfusion pressure flow-relationships in the VF model is described. Vibratory characteristics of subglottal onset pressure, frequency, glottal area, and glottal width are presented and compared to those of the human VF and other published VF models for varying perfusion pressures. The effects of vibration on perfusion flow rate and permeability are reported. The poroelastic VF models had an average onset pressure of 1.01 kPa while vibrating at an average frequency of 117 Hz and with a glottal width of 1.40 mm. Perfusion flow rate decreased between 15% and 22% from rest to vibration and increased between 29% and 33% after vibration ceased. Permeability followed the same trend of decreasing with vibration and increasing after vibration, with measured values on the order of 10^(-11) m^2 to 10^(-9) m^2. It is anticipated that this poroelastic material and model will form the basis for future studies of perfused flow through human VFs, engineered VF tissues and biomaterials, and VF models.

Cementation is known to significantly influence the mechanical and hydrologic properties of granular porous media by increasing the stiffness of the elastic response to stress and reducing permeability. The relationship between the changes in cementation and changes in permeability are well documented in literature. However, limited quantitative data exists on the relationship between changes in the amount of cementation and changes in the mechanical response of granular media. The goal of this research is to quantify the effects of cementation on the mechanical properties of granular porous media at the meso-scale and investigate the influence of the competing roles of mechanical and hydrologic properties on fluid flow and deformation at the macro-scale. To accomplish this goal, we developed a multiple scale approach that utilizes the parameterization control of meso-scale Discrete Element Method (DEM) models and the ability to couple fluid flow and solid deformation physics with macro-scale poro-elasticity models. At the meso-scale, a series of DEM models are designed to simulate biaxial tests of variably cemented sandstone in order to investigate the effects of cementation on the elastic and inelastic response of the porous media. The amount of cementation in the DEM model is quantified using a bond to grain ratio (BGR). The BGR is the number of bonds (the bonds represent the cement) divided by the number of grains in each model. The BGRs of the DEM models correlate to BGRs of natural samples and allow constraint of the percent cementation in the DEM models. A decrease in BGR from 2.25 to 1.00 results in a two fold decrease in shear modulus. The resulting shear moduli from the DEM models are used as input properties into two dimensional, axial symmetric poroelastic models of an isotropic confined aquifer. The poroelastic models address the implications of changes in mechanical properties and hydrologic properties on large scale fluid removal and deformation as well as address the importance of the competing roles of hydrologic and mechanical properties.

A validated model of water transport in the cerebral environment is both an ambitious and timely task; many brain diseases relate to imbalances in water regulation. From tumours to strokes, chronic or acute, transport of fluid in the brain plays a crucial role. The importance and complexity of the brain, together with the range of unmet clinical needs that are connected with this organ,make the current research a high-priority. One of the most paradoxical cerebral conditions, hydrocephalus, serves as an excellent metric for judging the success of anymodel developed. In particular, normal pressure hydrocephalus (NPH) is a paradoxical condition with no known cure and existing treatments display unacceptably high failure rates. NPH is considered to be a disease of old age, and like many such diseases, it is related to a change in the transport of fluid in the cerebral environment. This complex system ranges from organ-level transport to cellular membrane channels such as aquaporins; through integrating it in a novel mathematical framework, we suggest that the underlying logic of treatment methods may be misleading. By modelling the transport of cerebrospinal fluid (CSF) between the ventricular system, cerebral tissue and blood networks, we find that changes to the biophysical properties of the brain (rather than structural changes such as aqueduct obstruction) are capable of producing clinically relevant ventriculomegaly in the absence of any obstruction to CSF flowthrough the ventricular system. Specifically, the combination of increased leakiness and compliance of the capillary bed leads to the development of enlarged ventricles with a normal ventricular pressure, replicating clinical features of the presentation of NPH. These results, while needing experimental validation, imply that treatment methods like shunting, that are focussed on structural manipulation, may continue to fail at unacceptably high rates.

Frac-pack design is still done on conventional hydraulic fracturing models that employ linear elastic fracture mechanics. However it has become evident that the traditional models of fracture growth are not applicable to soft rocks/unconsolidated formations due to elastoplastic material behavior and strong coupling between flow and stress model. Conventional hydraulic fracture models do not explain the very high net fracturing pressures reported in field and experiments and predict smaller fracture widths than expected. The key observations from past experimental work are that the fracture propagation in poorly consolidated sands is a strong function of fluid rheology and leak off and is accompanied by large inelastic deformation and shear failure leading to higher net fracturing pressures. In this thesis a numerical model is formulated to better understand the mechanisms governing fracture propagation in poorly consolidated sands under different conditions. The key issues to be accounted for are the low shear strength of soft rocks/unconsolidated sands making them susceptible to shear failure and the high permeabilities and subsequently high leakoff in these formations causing substantial pore pressure changes in the near wellbore region. The pore pressure changes cause poroelastic stress changes resulting in a strong fluid/solid coupling. Also, the formation of internal and external filtercakes due to plugging by particles present in the injected fluids can have a major impact on the failure mechanism and observed fracturing pressures. In the presented model the fracture propagation mechanism is different from the linear elastic fracture mechanics approach. Elastoplastic material behavior and poroelastic stress effects are accounted for. Shear failure takes place at the tip due to fluid invasion and pore pressure increase. Subsequently the tip may fail in tension and the fracture propagates. The model also accounts for reduction in porosity and permeability due to plugging by particles in the injected fluids. The key influence of pore pressure gradients, fluid leakoff and the elastic and strength properties of rock on the failure mechanisms in sands have been demonstrated and found to be consistent with experimental observations. / text

Large amounts of hydrocarbon reserves are trapped in fractured reservoirs where fluid flux is far more rapid along fractures than through the porous matrix, even though the volume of the pore space may be a hundred times greater than the volume of the fractures. These are considered extremely challenging in terms of accurate recovery prediction because of their complexity and heterogeneity. Conventional reservoir simulators are generally not suited to naturally fractured reservoirs’ production history simulation, especially when production processes are associated with large pressure and temperature changes that lead to large redistribution of effective stresses, causing natural fracture aperture alterations. In this case, all the effective processes, i.e. hydraulic, thermal and geomechanical, should be considered simultaneously to explain and evaluate the behavior of stress-sensitive reservoirs over the production period. This is called thermo-hydro-mechanical (THM) coupling. In this study, a fully coupled thermo-hydro-mechanical approach is developed to simulate the physical behavior of fractures in a plane strain thermo-poroelastic medium. A hybrid numerical method, which implements both the finite difference method (FDM) and the displacement discontinuity method (DDM), is established to study the pressure, temperature, deformation and stress variations of fractures and surrounding rocks during production processes. This method is straightforward and can be implemented in conventional reservoir simulators to update fracture conductivity as it uses the same grid block as the reservoir grids and requires only discretization of fractures. The hybrid model is then verified with couple of analytical solutions for the fracture aperture variation under different conditions. This model is implemented for some examples to present the behavior of fracture network as well as its surrounding rock under thermal injection and production. The results of this work clearly show the importance of rate, aspect ratio (i.e. geometry) and the coupling effects among fracture flow rate and aperture changes arising from coupled stress, pressure and temperature changes. The outcomes of this approach can be used to study the behavior of hydraulic injection for induced fracturing and promoting of shearing such as hydraulic fracturing of shale gas or shale oil reservoirs as well as massive waste disposal in the porous carbonate rocks. Furthermore, implementation of this technique should be able to lead to a better understanding of induced seismicity in injection projects of all kinds, whether it is for waste water disposal, or for the extraction of geothermal energy.

abstract: Hydraulic fracturing is an effective technique used in well stimulation to increase petroleum well production. A combination of multi-stage hydraulic fracturing and horizontal drilling has led to the recent boom in shale gas production which has changed the energy landscape of North America. During the fracking process, highly pressurized mixture of water and proppants (sand and chemicals) is injected into to a crack, which fractures the surrounding rock structure and proppants help in keeping the fracture open. Over a longer period, however, these fractures tend to close due to the difference between the compressive stress exerted by the reservoir on the fracture and the fluid pressure inside the fracture. During production, fluid pressure inside the fracture is reduced further which can accelerate the closure of a fracture. In this thesis, we study the stress distribution around a hydraulic fracture caused by fluid production. It is shown that fluid flow can induce a very high hoop stress near the fracture tip. As the pressure gradient increases stress concentration increases. If a fracture is very thin, the flow induced stress along the fracture decreases, but the stress concentration at the fracture tip increases and become unbounded for an infinitely thin fracture. The result from the present study can be used for studying the fracture closure problem, and ultimately this in turn can lead to the development of better proppants so that prolific well production can be sustained for a long period of time. / Dissertation/Thesis / Masters Thesis Mechanical Engineering 2014