The 10th Annual Meeting and Jubilee celebration will be held at the New Orleans Ernest N. Morial Convention Center, in walking distance from the historic French Quarter and the Riverwalk. Conference hotels will be adjacent to Convention Center in the Warehouse and Arts Districts. Plan to attend, engage in cutting edge research on porous media, and enjoy one of the most historic cities in America.
Conference technical program, important dates, accommodation details will be available shortly on the InterPore site.
The local organizing committee looks forward to welcoming you to New Orleans in 2018.
Topics and applications
|
|
Event Management |
The injection and storage of supercritical CO2 (scCO2) have been conducted in fractured sandstone reservoirs at In Salah, Algeria and Snøhvit, Norway, and planned in fractured sandstone, carbonate, and dolomite reservoirs at Longyearbyen, Norway, Hontomin, Spain, and Kevin Dome, USA, respectively, with matrix permeability varying from 0.01 to 60 md. For densely fractured reservoirs with low matrix permeability (e.g., at Longyearbyen, Norway), injected scCO2 can dissolve into the resident brine at fracture-matrix interfaces and the dissolved CO2 (dsCO2) can diffuse into the rock matrix making solubility trapping the dominant trapping mechanism. For fractured reservoirs with intermediate matrix permeability (e.g., at In Salah, Algeria), the storage of scCO2 in the rock matrix dominates with strong fracture-matrix interactions observed through field monitoring at In Salah. We developed a comprehensive conceptual model for enhanced CO2 storage to account for global migration of scCO2 in the fracture continuum, local storage of scCO2 and dsCO2 in the matrix continuum, driving forces for scCO2 invasion and dsCO2 diffusion from fractures, and brine outflow through connected matrix blocks.
For the dominant matrix scCO2 storage, we developed high-resolution fracture-matrix models for individual matrix blocks, homogeneous columns of fractures and matrix blocks, and heterogeneous REVs consisting of multiple columns of matrix blocks with varying flow properties and sizes. The multiscale modeling results show that the equilibrium efficiency of local scCO2 storage strongly depends on matrix entry capillary pressure, matrix-matrix connectivity, and reservoir thickness, while dynamic efficiency and transfer function are also sensitive to fracture spacing and matrix flow properties. The transfer functions calculated for various REVs were used along with reservoir-scale dynamics of scCO2 plume flow in fractures, showing that the preferential migration of scCO2 through fractures is coupled with bulk dsCO2 storage in the rock matrix that in turn retards the scCO2 fracture plume. The bulk matrix storage is mainly driven by buoyancy between fracture scCO2 and matrix brine and facilitated by matrix-matrix connectivity that allows displaced brine to outflow, enabling the rock matrix to act like an open system. Conventional dual-continuum models cannot capture these processes because they model isolated matrix blocks with no capillary continuity, thereby underestimating storage efficiency.
For the dominant matrix dsCO2 storage, we developed the unified-form equations of diffusive flux of dsCO2 into brine-bearing matrix blocks of varying shapes (i.e., spheres, cylinders, slabs, squares, cubes, rectangles, and rectangular parallelepipeds) and sizes (Zhou et al., 2017a, b). We then applied the flux equations to a fractured reservoir with various scenarios of matrix blocks by assuming 1-D and 2-D radial scCO2 flow in fractures and by using diffusion of dsCO2 from fracture-matrix interfaces into matrix blocks as the sink for scCO2 in fractures. For each scenario, the dynamic dsCO2 plume with different mass fraction was produced analytically, showing that solubility trapping is significant in fractured reservoirs with low matrix permeability and small fracture spacing.
Injection of CO2 deep underground into porous rocks, such as saline aquifers, appears to be a promising tool for reducing CO2 emissions and the consequent climate change. During this process CO2 displaces brine from individual pores and the sequence in which this happens determines the efficiency with which the rock is filled with CO2 at the large scale. The aim of this work is to better understand the impact of different flow regimes, during immiscible two-phase flow, on the displacement and storage efficiency of CO2 deep in saline aquifers. Using multi-GPU free energy Lattice Boltzmann simulations we directly solve the hydrodynamic equations of motion on a three dimensional geometry reconstructed from micro-CT images of Ketton limestone and consider fluid flows in a range of capillary numbers Ca and viscosity ratios. We first verify the existence of the three typical fluid displacement patterns, namely viscous fingering, capillary fingering and stable displacement [1]. We examine how these distinctively different flow regimes can affect the displacement efficiency, defined here as the fraction of the defending wetting fluid that has been displaced from the pore matrix when the injected non-wetting phase reached the outlet of the domain. Continuing the injection beyond this point we establish the maximum displacement efficiency or storage capacity. Our results indicate that the maximum displacement efficiency decreases with decreasing Ca. As capillary fingering becomes the dominant displacement process at low Ca, storage efficiency converges to a limiting value irrespective of the viscosity ratio.
Particular focus is given to the low Ca flow regime, where displacements at the pore scale typically happen by sudden jumps in the position of the interface between brine and CO2, Haines jumps. We demonstrate that the method reproduces the expected features of the jumps, i.e. sharp increase in the non-wetting phase velocity, abrupt drop in the pressure signal and significant fluid rearrangement. We quantify the degree of fluid redistribution associated with these sharp events by identifying each event from the pressure signal. Preliminary results from this analysis suggest that pressure fluctuations and waiting times between the jumps follow an exponential distribution, in agreement with theoretical predictions, while the same also applies for the event filling volumes probably due to the extensive fluid redistribution. More importantly a significant decrease in storage efficiency is observed, irrespective of the direction of the jump relative to the overall flow direction, contrary to the arguments by Yamabe et al. [2]. This is due to irreversible fluid rearrangement during Haines jumps that alters the displacement pathways and renders regions of the porous rock inaccessible to the injected non-wetting fluid. This has important implications in the context of geological sequestration of CO2, as Haines jumps become a limiting factor in the storage process.
We examine the linear and weakly nonlinear stability analyses of the dissolution-driven convection induced by the sequestration of carbon dioxide in a geological formation. The mathematical model consists of Darcy's equation, the conservation of mass and the conservation of solute equations. The model accounts for anisotropy in both carbon diffusion and permeability which is modeled by a decaying exponential function of depth. The presence of a first order reaction between the carbon-rich brine and host mineralogy is also included. We prescribe either Neumann or Dirichlet boundary condition for the concentration of carbon dioxide at the rigid upper and lower walls that bound a layer of infinite horizontal extent. We consider a Rayleigh-Taylor-like base state consisting of a carbon-rich heavy layer overlying a carbon-free lighter layer and seek the critical thickness at which this configuration becomes unstable. With this approach, standard mathematical methods that were successfully used in the study of Rayleigh-Benard convection can be applied to this problem. We quantify the influence of carbon diffusion anisotropy, permeability dependence on depth and the presence of the chemical reaction on the threshold instability conditions and associated flow patterns using the classical normal modes approach. The critical Rayleigh number and corresponding wavenumber are found to be independent of the depth of the formation. The weakly nonlinear analysis is performed using long wavelength asymptotics, the validity of which is limited to small Damk\"{o}hler numbers. We derive analytical expressions for the solute flux at the interface, the location of which corresponds to the minimum depth of the boundary layer at which instability sets in. We show that the interface acts as a sink leading to the formation of a self-organized exchange between descending carbon-rich brine and ascending carbon free brine. Plots of the high order perturbation terms for the concentration successfully reproduce the fingering pattern that is typically observed in experiments and full numerical simulations. Using the derived interface flux conditions, we put forth differential equations for the time evolution of the upward migration of the interface as the dissolution process progresses. We solve for the terminal time when the interface reaches the top boundary thereby quantifying the time it takes for an initial amount of injected super-critical Carbon dioxide to be completely dissolved. We also consider the case where the interface migration is accompanied by interface deformations that conform to the convection pattern.
Injection of CO2 into a saline aquifer leads to a two-phase flow system, including a supercritical CO2 phase and a brine phase. Various modeling approaches, including fully three-dimensional (3D) models and vertical-equilibrium (VE) models, have been used to study the system in unfractured formations. Three-dimensional models solve the governing flow equations in three spatial dimensions and are applicable to generic geological formations. VE models assume rapid and complete buoyant segregation of the two fluid phases, resulting in vertical pressure equilibrium and allowing integration of the governing equations in the vertical dimension. This reduction in dimensionality makes VE models computationally much more efficient, but the associated assumptions restrict the applicability of VE model to formations with moderate to high permeability.
In this presentation, we extend the VE and 3D models to simulate CO2 injection in fractured aquifers. This is done in the context of dual-continuum modeling, where the fractured formation is modeled as an overlap of two continuous domains, one representing the fractures and the other representing the rock matrix. Both domains are treated as porous media continua and, as such, can be modeled by either a VE or a 3D formulation. The transfer of fluid mass between fractures and rock matrix is represented by a mass transfer function connecting the two domains. Because the fracture domain is usually much more permeable than the matrix domain, we apply VE modeling to the fracture domain but not the matrix domain. We refer to the resulting model as a hybrid VE-3D model, with the VE model applied to the highly permeable fractures and the 3D model in the less permeable rock matrix.
Our hybrid VE-3D model includes both dual-porosity and dual-permeability types. The dual-porosity model conceptualizes the rock matrix as sugar-cubes that are isolated uniformly by vertical and horizontal fractures, or as match-sticks that are isolated by vertical fractures through the entire thickness of the aquifer. In contrast, the dual-permeability model explicitly represents the 3D flow dynamics in the rock matrix. We derive mass transfer functions that couple the VE model in the fracture to the different models in the rock matrix. We then apply the hybrid VE-3D model to simulate CO2 migration in fractured saline aquifers and compare with 3D-3D models where both the fracture and rock matrix are modeled in 3D. The hybrid VE-3D models are much more computationally efficient while providing results that are close to those from the 3D-3D models. These vertically-integrated dual-porosity and dual-permeability models provide a range of computationally efficient tools to model CO2 storage in fractured saline aquifers.
Injection of carbon dioxide (CO2) into deep geologic formations has been widely proposed as an effective way for the permanent storage of CO2. Modification of the interfacial properties of CO2 in minerals by using surfactant has been proposed aiming on increasing the mobility of CO2 through porous media. Surfactants are proven to effectively alter the interfacial tension and wettability in both CO2/water/mineral system, improving the displacement and sweep efficiencies of CO2 in porous media. In the meantime, biosurfactants have been drawing much attention as an alternative to the chemical surfactants for their biodegradability, ecological suitability and low toxicity. However, the question as to the extent of microbial alterations in fluid wettability and interfacial tension under reservoir pressure and temperature conditions still warrants further investigation. Therefore, this study investigated the role of lipopeptide biosurfactant on wettability and interfacial tension alterations in a CO2/brine/mineral system for different CO2 phases during the growth of thermotolerant and barotolerant bacteria, Bacillus subtilis, and the production of lipopeptide biosurfactant, surfactin. Quartz, mica and carbonate substrates were selected and used as representative minerals. While monitoring the changes in the interfacial tension and wettability with pH, fluid samples were acquired from the brine phase, and the concentrations of glucose, nitrate, ammonium and surfactin in the acquired samples were quantitatively assessed using various assays and spectroscopic methods. As a result of surfactin production by B. subtilis, we observed the reductions in interfacial tension and increases in contact angle at all tested cases. The concentration of surfactin and the rate of wettability alteration differed with the experimental conditions. The modification of CO2 wettability was the greatest for liquid CO2 while the least of modification was observed for gaseous CO2. The obtained results allow in-depth assessment of the feasibility of using biosurfactant-producing bacteria for effective geologic carbon storage practices.
The Bunter sandstone formation in the Southern North Sea and the Captain sandstone formation in the Northern North Sea represent two of the largest potential CO$_2$ stores in the UK, with estimated capacities of up to 14 Gt and 1.7 Gt respectively [1, 2]. With current UK CO$_2$ emission totalling ~400 Mt/yr [3], the Bunter and Captain formations alone have the potential to store UK emissions for many years.
In order to determine the long-term fate of the injected CO$_2$ in these systems, accurate characterisation of the multiphase flow behaviour and trapping is needed [4]. Conventionally, multiphase flow functions, namely relative permeability, capillary pressure and trapping, are derived from viscous limit core flood experiments, measured at high flow rates on subsurface rock cores preferentially selected for homogeneity [5], and either used directly in field scale modelling or for further upscaling.
However, for modelling low potential flows characteristic of buoyantly driven CO$_2$ plume migration, it is important to derive properties that capture the impacts of rock heterogeneity. Sub-metre scale capillary pressure heterogeneities will control local fluid distribution, resulting in equivalent relative permeabilties which are dependent on the flow direction, rock heterogeneity, and the capillary number [6]. Modelling studies have estimated that this can have a significant impact on plume migration and trapping from the mm-km scale [7,8]. However, no experimental protocols have been developed to inform the models with appropriate properties measured on heterogeneous rock samples in the laboratory.
To address the impacts of small scale heterogeneity on large scale flow and trapping of CO$_2$, we present a combined experimental and numerical study on rock cores from the Bunter and Captain sandstone formations. We analyse 38 small rock cores covering the entire 100m interval of the Captain D reservoir unit in the Northern North Sea, and a smaller selection of cores taken from the Bunter Sandstone in the Southern North Sea. We use a recently developed characterisation approach [9] to create a 3D numerical model of heterogeneous rock cores, based on laboratory observations. We incorporate hysteresis into the characterisation by building on the recent approach developed by [10]. Once characterised, the numerical cores can accurately predict equivalent relative permeabilties and trapping, dependent on the capillary number and direction of fluid flow.
The numerical models are then used to investigate multiphase flow hysteresis and trapping across the range of conditions estimated to prevail in the CO$_2$ storage reservoirs. Under these conditions, we systematically explore the impact of hysteresis and heterogeneity on flow and trapping at multiple scales. The migration of CO$_2$ may be significantly enhanced by heterogeneity when flow can align with the direction of layers. This situation may arise in gravity currents of plumes underneath a confining caprock layer. In contrast, flow is impeded by heterogeneity when the dominant direction crosses bedding layers, as may occur in predominantly upward buoyantly driven migration. In this case, the lowered mobility results in significant spreading of the plume and residual trapping is also enhanced.
CO$_2$ injection into a saline aquifer leads to a two-phase flow system (supercritical CO$_2$ and brine), which often involves large spatial and temporal scales that require high computational cost. To address the computational challenge, in the past decade, a series of simplified models based on vertical integration of the full multi-dimensional governing equations have been developed. These vertically integrated models either assume a rapid segregation between CO$_2$ and brine due to strong buoyancy (i.e., vertical equilibrium assumption) or solve the one-dimensional vertical two-phase flow dynamics as fine-scale problems on top of the (coarse-scale) vertically integrated equations. The former is ofen referred to as vertical equilibrium (VE) model, while the latter relaxes the VE assumption and is called dynamic reconstruction (DR) model [1,2]. The major computational cost of the VE and DR models comes from solving the coarse-scale vertically integrated equations while the computation associated with the vertical reconstructions (either VE or DR) is minor. As such, they are much more computationally efficient than full multi-dimensional models and have been used to answer many important engineering questions. However, the vertically integrated VE or DR models are often limited to aquifers with homogeneous or layered heterogeneous properties. Thus, for aquifers with strong 3D heterogeneity, the computationally expensive 3D models are to date the only robust option.
In this talk, we present a hybrid multilayer framework to couple full multi-dimensional models with the various vertically integrated models. Such a framework allows us to use full multi-dimensional models in highly heterogeneous layers of an aquifer where full multi-dimensional model is the only robust option, while applying simplified vertically integrated models in layers with homogeneous or layered heterogeneous properties. We develop algorithms to couple the full multi-dimensional model with vertically integrated models (VE or DR), as well as algorithms for the coupling between the VE and DR models. In addition, we develop a local criterion to adaptively switch between VE and DR reconstructors [3], i.e., use VE reconstructor when the two fluid phases are in equilibrium while use DR reconstructor to capture vertical dynamics when the fluids deviate from vertical equilibrium. Comparisons with full multi-dimensional models (MRST [4] is used in our work) show that our adaptive hybrid multilayer model is much more computationally efficient than full multi-dimensional models while providing results with similar accurary, making this hybrid model an attractive tool for modeling of CO$_2$ injection and migration in highly heterogeneous saline aquifers.
Understanding the relation between CO2 saturation/distribution and velocity/attenuation of acoustic wave propagation is fundamental for an accurate and reliable quantitative interpretation of (seismic) CO2 monitoring data.
Quantitative interpretation of geophysical data requires understanding of the relationship between the physical properties of the rock, the microstructure of the rock, and the spatial distribution of the fluids saturating the pore space. A large number of laboratory acoustic measurements on rocks has been carried out in the past 10 years, with indirect assessments of how the saturation distribution is formed under in-situ conditions from ultrasonic range (Mavko et al, 1994) to seismic frequency scale (Tisato et al., 2014). However, there are almost no efforts towards quantitative measurements of how saturation distributions affect the acoustic measurements. In particular, no direct experimental link has been demonstrated between the quantitative distributions of saturations at pore scale and the acoustic properties of the saturated rock.
The relationship between elastic wave velocities and fluid saturations (brine/CO2 and brine/oil/CO2 mixtures) is strongly dependent on the spatial saturation distribution, i.e., whether distribution is heterogeneous (patchy) or homogenous (uniform). The saturation distribution in turn has a strong signature on attenuation and dispersion of propagating waves, even at seismic frequencies, in the porous media due to wave-induced fluid flow mechanisms (Müller et al., 2004 and 2010). Hence, predicting saturation effects on the seismic response requires a fundamental understanding of how attenuation and velocities are affected by fluid distributions. Seismic quantitative interpretation of CO2 monitoring data can be largely flawed if such mesoscopic phenomena are not taken into account, both for amplitude-based methods (due to attenuation) and waveform-based methods (due to attenuation and dispersion) (Dupuy et al., 2017). The correct quantification of saturation and pore pressure levels require understanding the fluid distribution and physical properties (Rubino et al., 2011).
The relationships mentioned above are difficult to predict and highly case-dependent. We present preliminary results on a technique for establishing the relationships by core studies based on acoustic measurements in the ultrasound scale on CO2-brine saturated rocks. The experimental setup is simultaneously imaged in X-ray CT-scanner where the rock, its microstructure, and the fluid distribution at pore scale are visualized. This is a first step in the effort to study in laboratory the time resolved (4D) and spatial relationships between microstructures, physical properties and saturation distribution of reservoir rocks exposed to CO2-bearing fluids.
One gram of soil can contain up to 100 million to 1 billion microrganisms and up to 1 million different species of microorganisms. Despite this fact, geotechnical engineers have, until fairly recently, ignored biological activity in the soil or possible biological amendments that could be introduced. Over the last ten years research has focused on bioaugmentation strategies (i.e. the injection of a single strain of bacteria) to alter hydraulic and mechanical behaviour of porous and fractured media (e.g. microbially induced calcite precipitation). One challenge of bioaugmentation technologies is the transportation of bacteria within the ground. This study investigates for the first time the potential use of fungal networks for ground engineering applications. Fungi produce hyphae, long filamentous structures which collectively are called a mycelium. Mycelium can grow to vast sizes, with individual mycelia (in forest floors) covering areas up to 9km2 in North America. As such there is great potential ‘to grow’ fungal mycelia for earth infrastructure over large areas.
We investigated the hydraulic behaviour of sandy soils treated with fungal mycelia using P. ostreatus (oyster mushroom) in order to: (i) Assess the level of hydrophobicity induced and (ii) understand the influence of P. ostreatus mycelia on water flow through the soil profile. To investigate these, we grew mycelia in petri dishes and conducted water drop penetration tests to ascertain induced hydrophobicity. We also determined the surface water evaporation rates for soils with mycelia and those without, at different starting moisture conditions. Next, we set up a 1-dimensional infiltration column test with mycelia inoculated and incubated to grow overtime throughout the soil profile. The infiltration column was instrumented with tensiometers and Time Domain Reflectometer (TDR) probes for the real-time measurement of suction and water content. Soil infiltration water fronts were obtained for both treated and untreated soils in respective columns. The presence of fungal mycelia resulted in significantly altered hydraulic characteristics of the soils. Mycelia induced extreme hydrophobicity on fine sands and reduced surface water evaporation rates. Infiltration time was slower for fungal-treated soils than untreated soils. These results highlight the potential for fungal mycelia to be used in the creation of semi-permeable or impermeable barriers in a range of ground engineering applications.
Key words: fungal mycelia; soil hydraulics; ground improvement; 1-D infiltration;
Bacterial colonization and the spread of biopolymer, gel-like material, on porous media are known to decrease permeability by several order of magnitude and to cause bioclogging thereby altering the hydraulic flow systems of porous media. Attention to microbial bioclogging has been increasing owing to the increasing demand of microbial soil treatment and soil improvement. Successful microbial bioclogging treatments require geophysical monitoring techniques to provide appropriate spatial and temporal information on bacterial growth and activities in the subsurface; such monitoring datasets can be used to evaluate the status of plugged sections and optimize re-treatment if the plug degrades. Therefore, this study investigated the feasibility of using P- and S-wave velocity and attenuation for monitoring the accumulation of bacterial biopolymers and the permeability variations during bioclogging. In sand-packs, Leconostoc mesenteroides was cultured and stimulated to produce insoluble biopolymer and generate bioclogging. During such bacterial bioclogging, permeability and high-frequency P- and S-wave responses were monitored. P-wave velocity was consistent and S-wave velocity was increased with biopolymer accumulation. Both P-and S-wave attenuation, evaluated by using spectral ratio method, were increased with increasing biopolymer saturation. Increases in seismic attenuation are closely linked to the biopolymer saturation and permeability reduction. Herein, we also presented a theoretical model to correlate biopolymer saturation, permeability, and seismic attenuation by modifying three-phase Biot model and Pride-Berryman double-porosity model.
Bio-mediated ground improvement technologies harness subsurface biological and chemical reactions to improve the engineering properties of soils with reductions in detrimental environmental impacts when compared to conventional methods (Seagren and Aydilek 2010; DeJong et al. 2013). One such technology, Microbially Induced Calcite Precipitation (MICP) or bio-cementation, has received significant recent attention and leverages the biologically-mediated hydrolysis of urea by soil microorganisms to enable the precipitation of calcium carbonate (Ferris et al. 1996). The process can bind soil particles at particle contacts and coat particle surfaces resulting in large increases in soil shear strength and stiffness with simultaneous reductions in hydraulic conductivity and porosity (DeJong et al. 2006; Phillips et al. 2016).
Although MICP has been most commonly performed using the injection of non-native bacteria containing urease enzymes, such as Sporosarcina pasteurii, more recently researchers have shown that bio-stimulation, or the enrichment of native ureolytic microorganisms in-situ, can achieve comparable microbial ureolysis rates and successful completion of the MICP process (Fujita et al. 2000). For example, Gomez et al. (2016) demonstrated that similar improvements in calcite contents and engineering properties could be achieved using either stimulated indigenous or S. pasteurii microorganisms to complete bio-cementation in meter-scale experiments. More recently, Gomez et al. (2018) showed that native ureolytic microorganisms capable of completing the MICP process could be successfully enriched at treatment depths near 12 meters in gravelly sands.
Despite significant advances in stimulated ureolytic MICP, treatment techniques have not rigorously attempted to minimize reagent consumption. Instead, stimulation techniques have targeted fast ureolysis rates indicative of successful enrichment and cementation techniques have largely provided urea in excess of calcium to complete precipitation under calcium-limited conditions. While perhaps effective from an engineering performance standpoint, reductions in process reagent consumption and related environmental impacts are critical if the technology is to become environmentally and financially feasible for field-scale applications.
In this study, a series of soil column experiments were completed to explore the effectiveness of treatment techniques involving significant reductions in urea and calcium consumption. PHREEQC geochemical modelling was used to identify chemically feasible treatment strategies to be investigated experimentally. First, a soil column experiment was completed to explore large reductions in urea concentrations during cementation. Tests considered urea-to-calcium ratios ranging from 1.4, used in previous tests, to 1.0, believed to offer the theoretical best chemical efficiency. Following this investigation, a second column experiment was performed to examine the effect of urea concentrations on microbial enrichment during stimulation. Although previous experiments identified the importance of solution pH, yeast extract, and ammonium on ureolytic enrichment, it was not clear if reductions in urea during stimulation would influence obtained ureolysis rates. Columns were monitored in time using aqueous chemical and non-destructive geophysical measurements, and geotechnical and calcite content measurements were obtained following all treatments. The effect of reagent reductions on obtained engineering properties, microbial ureolytic enrichment, chemical reaction kinetics, and process environmental impacts was evaluated. The results provide critical information needed to optimize treatment techniques and reduce process environmental impacts for future upscaling of stimulated ureolytic MICP.
Microbially Induced Carbonate Precipitation (MICP) through the urea hydrolysis reaction has been extensively studied in the lab and implemented at field-scale several times, most notably for fracture sealing (Cuthbert et al., 2013; Phillips et al., 2016), for erosion control (Gomez et al., 2015), and for ground improvement (van Paassen et al., 2010). Grouting strategies used in industry are commonly based on experience derived from the injection of Ordinary Portland Cement, (e.g. use of the split-spacing method). Field-scale injection strategies for MICP are likely to differ considerably from traditional cement grout injections as:
i. the low viscosity of the injection fluids allows near-surface grouting with minimal risk of ground heave and the potential for larger soil/rock volumes to be treated around each injection point,
ii. strength improvement occurs without complete permeability reduction and multiple injections are required to incrementally reach the desired strength and permeability,
iii. flow velocity (to control bacteria attachment), pH adjustment (to control CaCO3 saturation state), and temperature (to control the rate of ureolysis) may all be used to limit blocking of the injection points, and
iv. abstraction boreholes may be required for the removal of waste ammonium.
We present here a multi-scale micro-continuum MICP model implemented in OpenFOAM and solving fluid flow with the Navier-Stokes equations. The model is intended to inform the choice of injection strategy used in field-scale pilot projects and solves for 1) bacteria injection, velocity dependent attachment, and encapsulation within precipitating CaCO3; 2) re-agent transport, urea hydrolysis and CO3 production with Michaelis-Menten kinetics and 3) CaCO3 precipitation, porosity reduction, and subsequent flow path alteration.
In this model injections can be driven by a constant flow rate, constant pressure, or stepped flow rate and can be planar flow (e.g. for groundwater movement) as well as radial flow from/ to multiple injection/ abstraction wells. The model includes the ability to import 2D and 3D data from image analysis software ImageJ allowing X-ray micro-CT results from lab scale experiments to be modeled (at pore or micro-continuum scale) for validation purposes, and for heterogeneous site conditions to be modelled (at continuum-scale). A parametric sweep function is included to assess sensitivity to the choice of parameter values.
Results show that grouting with MICP is fundamentally different to grouting with cement. Operators may wish to replace sequential injection through a series of boreholes with simultaneous injection through multiple boreholes, or use abstraction boreholes to both collect waste ammonium and direct the transport of MICP re-agents.
Microbially induced desaturation and precipitation (MIDP) via denitrification has the potential to mitigate earthquake-induced liquefaction by two mechanisms: biogenic gas production to desaturate and dampen pore pressure changes in soil and calcium carbonate precipitation to mechanically strengthen soil. Lab-scale tests have demonstrated that both desaturation and precipitation are effective mitigation mechanisms. However, small-scale laboratory column tests at ambient pressure lead to gas pockets and lenses, causing upheaval due to low overburden pressures. Therefore, biogenic gas formation, distribution, and retention need to be evaluated with more realistic over-burden pressures to understand the effectiveness of this treatment mechanism. Centrifuge tests of soil desaturated by MICP treatment are currently being performed to simulate field pressures and stresses. In addition, a numerical model was developed to evaluate the scaling effects on biogenic gas generation between the centrifuge model and prototype scale. The centrifuge tests are conducted within a laminar box on the 1-m radius centrifuge at the University of California, Davis NHERI/CGM centrifuge facility. Desaturation is induced in the laminar box prior to acceleration in the centrifuge by augmenting saturated soil with an enriched culture of denitrifying microorganisms. The models are accelerated to 80 g in stages and measurements of soil moisture content are made over time to see the combined influence of steady-state pore pressure and overburden pressure on the degree of saturation. Upon reaching the final centrifuge acceleration, the models are subjected to strong shaking until either liquefaction is triggered or the capacity of the centrifuge is reached. Test results provide evidence of the capacity for MIDP to mitigate the potential for earthquake-induced soil liquefaction by desaturation. Comparison of modeling results to test data suggest that the numerical model does not consider certain pore-scale influences and the effects of mixing from liquid-gas transfer and transport observed in the centrifuge tests. Thus, future work will add these features to the model.
The biogenic gas behavior in porous media, which includes bubble nucleation and growth, migration, coalescence and trapping is affected by the gas generation rate, distribution of reactive sites and the pore scale characteristics of the sediment. In this study, experiments are performed using a micro-fluidic chip in which different gas bubble behavior mechanisms in the porous media are observed. Secondly, the behavior of biogenic gas is simulated using a pore-network model extracted from the 3D X-ray image of an in-situ sediment. The formation of biogenic gas bubbles is modeled using the classical gas nucleation theory. Several numerical algorithms and criteria developed for the expansion of gas bubbles during the biogenic gas formation, size-dependent rising velocity of gas bubbles, bubble coalescence, slug formation and movement, escaping, and trapping in the pore space. The amount of produced gas bubbles, residual gas saturation and hydraulic conductivity are calculated during the simulations. Results of the simulation are qualitatively compared with the microfluidic chip experiments.
Ancient stone relics and historic buildings are often subject to significant degradation. The protection and restoration of these monuments is extremely urgent. Here, a method of building repair based on microbial induced carbonate precipitation (MICP) has been tested on marble stone. In previous research, microbial mortar (stone powder treated by MICP) was tested as a filling material to repair cracks within stone. In this paper, the effect of microbial treatment on degraded marble consisting of larger particle sizes is studied. In the experiment, we focus on altering the permeability and porosity of crushed marble grains and show that the porosity and the permeability of the sample are notably decreased by carbonate precipitation.
MICP treatment is carried out in a column filled with marble grains with the injection of six batches over six days. A white CaCO3 precipitate is produced which matches the original marble colour and is sufficiently strong to cement the marble sand together from the inlet up to a depth of 150 mm into the column. To understand the micro-scale distribution of the CaCO3 precipitation within the column, and its effect on flow and transport properties, we analyse the MICP-treated column using X-ray CT with a resolution of around 3 microns. The X-ray CT scan data, support the macro-scale observations of a gradient in the degree of cementation along the direction of liquid flow, indicating that producing an evenly solidified sample is a problem that needs to be resolved.
We use the core-scale experimental data to derive cm-scale fluid transport properties using tracer breakthrough curves taken, prior to, and after MICP treatment. The fitted transport properties show that the fraction of pores containing mobile water decreases with increasing cycles of MICP. Pore-scale modelling using the X-ray CT data supports these findings, showing that cementation leads to a change in the pore network structure, with flow increasingly focussed into a smaller number of faster moving open channels.
Our experiments show that CaCO3 precipitation is greatest at the inlet. It is reasoned that this could be avoided by modification to the injection strategy. Prevention of re-agent mixing outside the marble grains, careful choice of marble grain size distribution, and tailored injection flow rates could deliver re-agents deeper into the media and take advantage of the formation of stable flow pathways to maximise seal uniformity. MICP is a promising technique for the restoration of marble structures and monuments.
Worldwide demand for new and sustainable approaches to geotechnical engineering problems has generated novel research opportunities in the emerging field of bio-mediated soil improvement. The most widely researched of these processes is microbially induced calcite precipitation (MICP), which has shown promise for a wide variety of engineering applications. Initially MICP was accomplished by bio-augmentation with a high density of the constitutively ureolytic bacterium, Sporosarcina pasteurii (Stocks Fischer et al., 1999). The amended soil was then supplemented with liquid medium containing calcium salts, urea, and sometimes growth-promoting organic compounds. Bacterial hydrolysis of urea generates a molecule of carbonic acid and two of ammonia. The resulting ammonia, a weak base, equilibrates with water and tends to form ammonium and hydroxide ions. This shifts the carbonic acid-bicarbonate-carbonate equilibrium toward carbonate, which will precipitate as calcium carbonate in the presence of sufficient calcium, ideally in the immediate vicinity of ureolytic bacteria, thereby cementing adjacent soil particles and increasing soil strength and stiffness. More recently bio-stimulated MICP has been fully demonstrated in native sands with prospects for eliminating costs and environmental impacts of propagating and transporting large quantities of bacteria. In our recent column experiments, completed on 14 different sandy soils from different depositional environments -- including several samples obtained from natural deposits as deep as 12 meters -- bio-stimulated MICP was always successful (Gomez et al. 2014; 2017; 2018).
Over 300 bacterial pure cultures were obtained from the most recently bio-stimulated soils and were stored (-80°C) to enable future physiological and genetic studies. A study of the urease kinetics of 8 randomly selected bacteria enriched in a meter-scale stimulated MICP demonstration (Gomez et al., 2017) showed that whole cell rates of urea hydrolysis follow Michaelis-Menten kinetics, with half-maximum values achieved at urea concentrations ranging from 56 to 837 mM and maximum rates varying from strain to strain by 100-fold. In progress 16S rRNA sequencing of the culture collection shows it includes a wide variety of ureolytic strains closely related to, but not identical to, the S. pasteurii (ATCC strain 11859), which has been employed almost universally in bio-augmentation experiments. Certain strains were found repeatedly in all or most of our bio-stimulated MICP experiments. Physiological differences between these strains will be discussed along with their surprisingly high diversity at the end of bio-stimulated MICP treatments. We have also begun to link pure culture physiology with relative abundance of these same strains in bio-stimulation treatments by extracting and amplifying bulk DNA from dilute aqueous bacterial suspension. In progress sequencing of 400 full-length clones is expected to provide a higher resolution but lower density sampling of diversity versus time for a single bio-stimulated MICP column. In parallel, high throughput 16S amplicon sequencing will provide much higher depth but lower resolution snapshots of changes in bacterial diversity throughout this same progression.
The effectiveness of bentonite clay as a wellbore plugging material often depends upon its penetration into near-borehole cracks associated with the drilling process. Here we present research aimed at understanding and maximizing the ability of clay materials to plug near-borehole cracks. A device was constructed such that the borehole is represented by a cylindrical chamber, and a near-borehole crack is represented by a slot adjacent to the center chamber. The experiments consist of placing bentonite clay pellets into the center chamber and filling the entire cavity with distilled water or brine so that the pellets hydrate and swell, thereby intruding into the slot. Results indicate that the bentonite clay pellets do not fully plug the slot. We propose a model where the intrusion length into the slot is limited by: 1) the free swelling potential intrinsic to the system comprised of the bentonite pellets and the hydrating fluid, and 2) the resisting shear force along the walls of the slot. This model accounts for the fact that narrow slots have a smaller volume for the clay to fill than wider slots, but wider slots present less resistive force to clay intrusion. These two limiting mechanisms work against each other, leading to a non-monotonic relationship between slot width and intrusion length. Specifically, the results show a linear increase of the intrusion length with the slot width for narrow slots where the limit is the shear strength resisting intrusion and scaling in proportion to the slot width over which the driving swelling pressure is applied. For wide slots the intrusion length is controlled by available swelling volume and scales inversely with the volume of the slot, and hence also inversely with the slot width. The experiments also show that increasing the salinity of the solution leads to a dramatic decrease in the clay expansion, evidenced here by the substantially smaller intrusion length when the pore fluid contains 10 g/L NaCl. We conclude that there exists a range of length to width ratios of near wellbore cracks that will be effectively plugged, and where a suitable plugging criterion is determined by a competition between availability of volume and shear resistance to clay intrusion.
Cement failure in subterrenean wells may compromize zonal isolation, cause corrosion of steel casing and, in the worst case scenarios result in a catostrphic event of well collapse. Portland cement and its blends are used for cementing the majority of subterrenean wells. Although they proved to provide a robust solution in many cases, they are not durable in aggresive environments with high total dissolved solids, acidic gases and fluids and large thermal and mechanical stresses. Although strong, cements in general are brittle and, as a result, may fractures or break under various stresses. Because of the difficulties in locating and accessing damaged areas in subterranean wells self-healing cements are of particular interest. Cementitious blends with pozzolanic materials may form resilient cements with self-reinforcing properties at longer curing times. This work revisits and updates applications of cementitious blends that include pozzolanic materials for carbonate-rich, strongly acidic high-temperature wells with large thermal and mechanical stresses. Data on strength recoveries and cracks sealing or bond re-formation for samples subjected to a repeated compressive damage followed by 5d 300oC healing periods in water, alkaline carbonate and geothermal brine are presented. Blend cements formation mechanisms, mechanical properties, resistance to mild and strong acids and cement/carbon steel bond characteristics and corrosion protection are discussed. Properties of the class G oil well Portland cement modified with silica for high-temperature applications are presented as a benchmark. Possible healing mechanisms based on XRD, Raman analyses and SEM coupled with EDX data are proposed. The evaluated pozollanic materials include fly ash F, ground granulated blast furnace slag, natural zeolites, and micro glass fibers type E.
All wells will one day need to be permanently plugged and abandoned (P&A'ed), and this last phase of a well's life cycle does not end – but has an eternal perspective. The North Sea is a mature petroleum region, and a heavy focus on P&A activities are planned here for the next decades. Research and technology development needs to keep up with this shift in activity level, and a strong focus has thus recently been put on R&D relevant for P&A in Norway.
This talk will give an overview of P&A challenges and opportunities in the North Sea. First of all, it will point out exactly why today's P&A operations are so time consuming and expensive. Technology gaps will be outlined, together with issues related to standards and regulations in the North Sea countries. Thereafter, emerging P&A solutions will be discussed, and results will be shared from ongoing R&D projects aiming to simplify future P&A operations.
Examples of topics that will be outlined are:
- Making use of shale formations as permanent barriers in wells
- Techniques for simpler removal of steel pipes from wells
- Methods for avoiding steel removal during P&A
- Advice for maximizing long-term integrity of plugged wells
Thoughts about possible value creation from P&A, which is an ever-growing and long-lasting market, will also be shared with the audience. Differences between the North Sea and the Gulf of Mexico when it comes to plugging costs and operations will also be outlined.
Chemical enhanced oil recovery, specially surfactant-polymer flooding, involves porous media flows of simple and complex fluids through highly heterogeneous formations. We will discuss various models based on systems of partial differential equations which pose a variety of mathematical and computational challenges. We will start a dialogue on these challenges with some possible pathways to address these challenges thereby opening doors for new and challenging problems in this area. Our current efforts to develop high performance numerical methods and results based on these models will be discussed. Effects of various chemical agents on oil recovery as well as various fluid mechanical phenomena including fingering and channeling will presented and discussed.
We report a study of heavy oil recovery by combined water flooding and electromagnetic (EM) heating at a frequency of 2.45 GHz used in domestic microwave ovens. A mathematical model describing this process was developed. Simplified model equations are solved analytically and the solution is presented in an integral form for the one dimensional case. Complete model is solved numerically using Finite Difference Schemes. Experiments consisting of water injection into Bentheimer sandstone cores, either fully water-saturated or containing a model heavy oil, were also conducted, with and without EM heating.
Model prediction was found to be in rather good agreement with an experiments indicating that EM heating induces an overall improvement of the mobility ratio between the displacing water and the displaced heavy oil.
The Brazilian Synchrotron Light Laboratory (LNLS) is currently engaged in the construction and development of Sirius, the largest and most complex scientific infrastructure ever built in Brazil and one of the first 4th-generation Synchrotron Light Sources in the World. Its ultra-low emittance (0.28 nm.rad) and high brightness will allow the execution of very competitive experiments, opening new perspectives for research in many different fields, including material science, nanoscience, physics, earth and environmental science.
MOGNO is being designed to be a world-leading micro/nano imaging beamline focused towards multi-scale analysis (resolution ranging from hundreds of nanometers to dozens of micrometers) of the internal 3D structures of different materials and objects. The beamline will be primarily devoted and specialized in zoom-tomography, where a specimen can be studied at low and high-resolution. In parallel, MOGNO competences will be extended to 4D imaging through in-situ experiments, which will allow the researchers to observe and quantify material responses during mechanical, thermal or chemical loadings, in real time. The goal is to obtain full resolution scan times in the order of 1 to 5 seconds. This can be achieved given the extremely high flux provided by a 3.2T bending magnet, the well design optical system and improvements in detection efficiency, as evident in direct detection devices. Overall, all the processes (from robotic arms for sample exchanges, automatic alignment, data post-processing - reconstruction, segmentation and quantitative data analysis), are being optimized to make Mogno a high-throughput beamline.
Pressure-driven flow within the plane of a confined thin porous medium takes place in a number of natural and industrial processes. This includes flow during manufacturing of fibre reinforced polymer composites with liquid moulding processes, passive mixing in microfluidic systems and paper making. The thin porous media considered here is a simplified, well-structured model of a porous media where the solid parts have the shape of vertical cylinders placed on equal interspatial distance from each other in a quadratic pattern. The array of cylinders are confined between two parallel plates, hence the permeability depends on both the diameter and height of the cylinders, as well as their interspatial distance. In order to study the flow tomographic Particle Image Velocimetry is used and the fluid is indexed matched so that a whole volume of the flow can be scrutinized without optical distortion in each measurement. The results reveal that the averaged flow field changes substantially and that the wakes formed behind the cylinders plays a major role.
Numerical simulations of injection of CO2sc or of a mixture (aqueous solution + CO2sc) in sandstone cores were carried out. These preliminary numerical tests have twofold targets: i) helping for designing lab experiments and to anticipate the experimental behaviors of such complex systems and ii) to explore complex conditions (flow-pressure-temperature-salinity) that could not be easily achieved in laboratory experiments.
These 1D simulations were done using the numerical code TOUGH2 (Pruess et al., 1999), with the ECO2N module (Pruess, 2005). This code allows performing multiphasic reactive transport (THC) modelling considering gas-liquid equilibria and the possibility for halite to precipitate (according to thermodynamic equilibrium) with potential feedback on the petrophysical properties (k-phi) of the rock. During these simulations, the rock matrix is supposed to not react significantly (i.e., no dissolution of the sandstone at the scale of the simulation time).
To get closer to the experimental constraints, all THC simulations were done through cores the dimensions of which correspond to the actual samples used in lab tests. The hydrodynamic and petrographic properties are those of a Vosges sandstone sample taken in a quarry in the East of France. The duration of the simulated injections is about 15 hours. In order to characterize the hydrodynamic and salt deposition processes in the core, a very fine mesh was set up: the core was divided into 75 cells of 1 mm thickness and of identical hydrodynamic properties. Initially, the core is saturated with NaCl-bearing brines (with salinity from 3 to 6 molal).
Among the various simulations performed, one corresponds to a low-pressure gradient, i.e., a low gas flow rate. The time needed to dry out the column is extremely long and the risk of clogging is important. Indeed, after the piston effect happens (flush of the most mobile water), water evaporation (i.e., water transfer into the CO2-rich fluid) is the only mechanism able to dry the porous medium. TOUGH2 allows modelling a capillary feedback (opposite to the advective flow of the CO2-rich phase), which maintains a quasi-stationary back-flow of the more concentrated brine toward the core entrance. Thus, the porosity close to the core inlet is continuously fed with brine, the evaporation of which provokes a massive precipitation of salt at the column inlet. Successive cycles of injection/stop further increase these risks by allowing brine flowing back towards the inlet of the core because of the re-distribution of fluids during the stop phases.
In conclusion, although the numerical code is devoted to simulations at the reservoir scale, it is possible to represent processes at small (centimeter) scales helping smart lab experimentations. The advantage of the numerical approach is also to perform a large number of simulations to better determine the key processes and critical parameters and thus define the most relevant experimental tests to perform for more focused locks and mechanisms. The global methodology of these exploratory modeling will be presented at this minisymposium. The experimental results will be discussed and the relation “Flowrate-Pressure-Temperature” will be discussed regarding the injectivity of sandstone saline reservoirs.
CO$_2$ geological storage in deep saline aquifers represents a mediation solution for reducing the anthropogenic CO$_2$ emissions. So far, little is known about both the CO$_2$ storage impact on the underground geochemistry and on the microbial diversity inhabiting deep aquifers. Consequently, this kind of storage required adequate scientific knowledge and tools at the pore scale to evaluate injection scenarios or to estimate reservoir capacity. In this context, porous media designed inside high pressure / high temperature microfluidic reactors (micromodel or geological labs on chip – GloCs [1]) turn out to be excellent tools to complement the classical core-scale experimental approaches to investigate the different mechanisms associated with CO$_2$ geological storage in deep saline aquifers [2].
This talk will first highlight the latest results obtained at ICMCB concerning the application of the GLoCs to study the invasion processes of CO$_2$ in water and brine saturated GLoCs. In particular, direct optical visualization and image treatments allow following the evolution of the CO$_2$/brine phase distribution within the pores, including displacement mechanisms and pore saturation levels [3]. We will then present some ongoing work aiming at integrating in situ spectroscopy techniques in HP microreactors to get information about the dissolution and mineralization trapping. We have developed an experimental set-up to recreate 3D reactive porous media within a microfluidic channel (fixed packed bed of calcium carbonate – CaCO$_3$ microparticles). Thanks to X-ray laminography carried out at the european Synchrotron facility (ESRF), we have observed on reconstructed 2D images, the dissolution phenomena occurring during the successive injection of constant volumes of non-equilibrium solution. This proof of concept has opened new possibilities for using this methodology to acquire kinetic data on 3D reactive front phenomena in porous media.
Eventually, we will introduce the use of GLoCs as a significant tool to mimic the in situ biogeological reservoirs conditions to study CO$_2$ bioconversion (in the frame of the ERC project “Big Mac”). Indeed, beyond CO$_2$ geological storage investigations, the GLoCs could provide new insights into bioremediation process to restore the CO$_2$ as a valuable energy resource (i.e. CH$_4$ via methanogenesis process). These tools could also find wider applications in geological-related studies such as Enhanced Oil Recovery, shale gas recovery or geothermal energy.
The effective mineral surface area is often the least well constrained variable in the prediction of geochemical reaction rates and, therefore, poses a large uncertainty when modelling reactive-transport in porous rocks. The mineral surface area is often estimated using a combination of microscopic and spectroscopic techniques or using disaggregated sediment in dissolution experiments. Both approaches have been shown to overestimate the accessible or effective mineral surface area. In this study we have developed core-flood experiments to determine the effective surface areas of quartz, kaolinite and muscovite. Fluids with a pH of 2 and 12 are injected into a sandstone core plug at high flow rates to prohibit secondary mineral formation. The effective mineral surface area is calculated based on the general reactive-transport equation and uses the steady-state outflow concentrations of Si, Al and K, the fluid residence time and the known mineral dissolution rate constant at the given pH and temperature. This approach directly compares the dissolving mineral surface area in a consolidated sample to the normalised mineral surface area used to derive mineral specific dissolution rate constants. The derived effective mineral surface area therefore represents an intrinsic rock property. Variable injection rates and the associated differences in fluid residence time and cation concentrations in the outflow lead to similar surface area estimates as long as no secondary minerals are formed. The abundance and surface area of carbonate minerals can be estimated from the initial non-steady state core flooding phase using an injection fluid of pH 2 and the temporal changes in Ca, Mg and pH in the outflow by fitting the reactive-transport simulation curve to the experimental results.
Thermally induced cracks can be created when cold fluid contacts a higher-temperature rock. When rock is dry, it first contracts locally near the fluid/rock interface and then fractures in dominantly in tensile mode. However, when rock is saturated or partially saturated with water, the pore water in a rock complicates the behaviors because expanding freezing pore water may interact with mineral skeletons and compete with contracting mineral solids in driving the fracturing. This all occurs in a complex environment of pore and particle structures. When water becomes ice, the volume increases by 9%, which is larger than thermally-induced shrinkage in geomaterials.
Literature regarding freezing saturated porous media investigated slow freezing under low or cyclic thermal gradients occurring in nature. If frozen slowly, water migrates through the porous media by cryogenic suction, leading to ice segregation and lens growth. However, in rapid freezing under extreme thermal gradient using cryogenic fluid, the cryosuction period is postulated to be too brief to allow water migration and the formation of ice lenses because pore water will freeze quickly. This research centers on the effect of pore water on the thermal cracking behavior.
The behavior of water-saturated rock under cryogenic temperature was investigated by laboratory experiments. Concrete, sandstones, and shale are prepared as blocks (8"x8"x8") and cylindrical core specimens (1.5" dia. x 3"). Each specimen type is prepared as dry, partially saturated, and saturated states. Then the specimens are fully submerged in liquid nitrogen. Temperature at the surface of the block is monitored. In microscale (~particle/pore), we investigated how micro fabric is altered/fractured by cryogenic treatment by using SEM and micro-CT. In macroscale (~core/block), we studied macro-scale changes of internal structures caused by the treatment by using micro-CT. In this scale, we also do mechanical testing such as acoustics, permeability, and strength to see effects on macroscale properties.
After the cryogenic treatment of the water-saturated specimens, all major cracks were created near edges during the cryogenic stimulation. In this case, the ice, which is heavily interconnected through pores, was expanding against mineral matrix. Thus, the outer layer of the block exposed to liquid nitrogen experienced water freezing and expanding, whereas the inner block did not. The outer layer expanded laterally, resulting in shear fractures parallel to the exposed surfaces. If cracks were formed from rock contraction (without the effect of ice), more cracks would be located away from and perpendicular to the edges. The block bottom was in direct contact with the cryogen container limiting liquid nitrogen access resulting in the absence of cracks there. We also present micro CT and scanning electron microscopy results, which reveals interesting interaction between expanding ice crystals and mineral skeletons at specimen surfaces as well as internally.
Wettability is a major factor that controls the evolution of interfaces during immiscible fluid displacement from a permeable medium. Three dimensional imaging by ultrafast X-ray tomography allows us to investigate the morphology at different wetting conditions and relate it to the prevalent pore-scale process. In this study we focus on the evolution of the fluid interfaces in random piles of spherical beads with a narrow radius distribution. The contact angle of the invading fluid is varied through surface functionalization and by using different combinations of invading and defending fluids. Three methods are proposed to quantify the morphology of the invading interface from the acquired X-ray images. As governed by the contact angle, two qualitatively different classes of displacement pattern are observed in homogeneous bead packs of uniform wettability [1]. The same contact angle for the cross-over between the two distinct displacement regimes is obtained applying three independent methods to quantify the morphology of the invading fluid. Samples consisting of wettable and non-wettable beads indicate that the displacement pattern is the result of a mixture of different local advancing modes of the interface. These modes lead to either capillary fingering or stable front invasion [1]. The influence of wettability is further delineated by varying the ratio between wettable and non-wettable beads in the packing. Our results demonstrate the strong influence of mixed wet conditions on the displacement process, and provide further insights to the factors that control the formation of the macroscopic fluid flow pattern.
Reference:
[1] K. Singh, H. Scholl, M. Brinkmann, M. Di Michiel, M. Scheel, S. Herminghaus, and R. Seemann. Sci. Rep. 7: 444 (2017).
Enhanced oil recovery testing often relys on either a sand pack filled with sand or use Berea sandstone that doesn’t take into account the full range of petrophysical properties of the actual reservoir. Among the most important reservoir petrophysical properties are the particle sizes and their distribution that serves as the key parameter that influences strongly other key petrophysical parameters such as permeability, porosity, pore throat distribution, capillary pressure, etc. In this work, the authors have devised a new method to synthesize an artificial core in which the porosity, permeability, and particle size distribution nearly match that of a target natural reservoir core: 1. The method entails having the thin section prepared on a sliced target natural reservoir core. 2. The particle size distribution analysis is done on the slice under a microscope. 3. Different sizes of sieved sands have particle size distribution analysis done by a laser particle size analyzer. 4. The particle size analysis of the sands and the target natural reservoir core is superimposed on the same plot and the best-fitting sand is visually picked out. 5. The best-fitting mesh size is either used directly or sieved to cut off a smaller or bigger fraction not found in the particle size distribution of the target natural reservoir core. 6. The sieved sands are compressed into an artificial core with addition of other micron sized sands served as fillers with a specialized adhesive at a set compression time and pressure. 7. The synthesized artificial core is sent for porosity and permeability analysis. 8. If porosity and permeability is not within 5% and 10% of the target respectively, then more fine sand fillers are used to dial down the porosity and permeability on another attempt to synthesizing another core. 9. If porosity and permeability are both matched within acceptable error, the artificial core is sent for thin section particle size distribution analysis. 10. If the medium size of the artificial core’s particle size distribution and the shape of the distribution doesn’t match the target, then another adjustment is made on the sand size and the process is repeated until it does. Through several trial and error attempts using this method, the particle size distribution, permeability, and porosity of the artificial core is closely matched to a set target natural reservoir core. Using this method, a more robust artificial core can be synthesized for use in enhance oil recovery testing.
Unlike conventional reservoirs where pore size distribution has a micrometer scale (Nelson 2009), tight oil and shale gas reservoirs have predominantly mesopores (between 2 and 50 nm) and micropores (below 2 nm). Volume fraction of micropores is not negligible and can be as high as 20% (Kuila et Prasad 2011). As hydrocarbon molecules range between 0.5 and 10 nm (Nelson 2009), interaction forces between confined fluid and pore wall molecules become as significant as inter molecular interactions within the confined fluid. That is why nanofluidic experiments (Wang et al. 2014) and bubble point measurement on hydrocarbon mixture in mesoporous materials (Cho, Bartl et Deo 2017) have demonstrated that confinement considerably changes fluid phase behavior. Consequently the commonly-used equation of state (EOS) such as Peng-Robinson EOS is not able to describe the confined fluid phase behavior. A pore radius dependent EOS is therefore needed in reservoir simulators for accurate large scale tight oil and shale gas production forecast simulations.
The idea of this work is to integrate first the capillary pressure effect into the classical Peng-Robinson EOS and then to calibrate EOS parameters function of pore radius to fit molecular simulations. The capillary pressure which depends on pore radius adds pressure difference between vapor and liquid phase in the equilibrium computation. It is calculated using the Young-Laplace equation. Molecular simulation is performed using Monte Carlo method in the grand canonical and in the NVT Gibbs ensemble with anisotropic volume change in order to calculate equilibrium properties of several pure hydrocarbon components and mixtures in confinement. Kerogen pores are modelled by graphite slit pores and fluid/wall interaction potential is added. For a given pore radius, the critical temperature and pressure are determined for pure components, and liquid and vapor pressures, densities and molar fractions of components are calculated for both pure components and mixtures at different temperatures for calibration. These values are used as reference fitting data for the Peng-Robinson EOS with capillary pressure. The optimization parameters are the Peneloux volume correction constant, the acentric factor and the binary interaction coefficients. The calibration of these parameters allows getting correlations versus pore radius that will be used to model the confined fluid thermodynamic behavior.
The pore radius dependent EOS calibrated with molecular simulation data can therefore be used in reservoir simulators to accurately forecast tight oil and shale gas production. However the grid cells in a dynamic flow simulation is usually in order of several meters to 100m. Such a cell includes a large pore size distribution, the pore radius value used in the EOS is therefore an issue. In order to consider the pore size variability within a simulation cell, an effective radius function of oil saturation is taken. It is determined from the distribution function of pore size volume. Assuming oil is the wetting phase, during a flow simulation, oil is present in small pores and gas appears in larger pores, then the effective pore radius decreases. Oil and gas production simulations with a dual porosity model for a fractured tight-oil reservoir show that this methodology gives more reasonable results than using an average pore radius and of course than a bulk approach.
Natural gas production from shale formations has received extensive attention in recent years. While great progress has been made in understanding the adsorption and transport of single-component gas (usually CH$_4$) inside shales’ nanopores, the adsorption and transport of multicomponent shale gas under more realistic reservoir conditions (e.g., considering CH$_4$/C$_2$H$_6$ mixture) only begun to be studied. In this work, we use molecular simulations to compute the storage of CH$_4$/C$_2$H$_6$ mixtures in single nanopores and their subsequent recovery. We show that, surface adsorption contributes greatly to the storage of CH$_4$ and C$_2$H$_6$ inside the pores, and C$_2$H$_6$ is enriched over CH$_4$. The enrichment of C$_2$H$_6$ is enhanced as the pore is narrowed, but is weakened as the pressure increases. These effects are captured by the ideal adsorbed solution (IAS) theory, but the theory overestimates the adsorption of both gases. We show that the recovery of gas mixtures inside the nanopores toward a gas bath approximately follows the diffusive scaling law. The ratio of the production rate of C$_2$H$_6$ and CH$_4$ is close to their initial mole ratio inside the pore despite that the mobility of pure C$_2$H$_6$ is much smaller than that of pure CH$_4$ inside the pores. By using scale analysis and by computing the Onsager coefficients for the transport of binary CH$_4$/C$_2$H$_6$ mixtures inside the nanopores, we show that the strong coupling between the transport of C$_2$H$_6$ and CH$_4$ is responsible for the effective recovery of C$_2$H$_6$ from the nanopores.
Abstract
Understanding the gas occurrence states under real reservoir conditions is the prerequisite to study the mechanisms of gas flow in shale reservoirs, in which large amounts of nanoscale organic pores exist. Besides, water is inevitable when considering the gas flow in shales. Thus molecular dynamics simulations were performed to study the occurrence states of gas-water mixtures near the organic solid. Results indicate that methane will approach the organic surface spontaneously, accumulate at the solid-liquid interface and form a dense gas region finally. This process has little relationship with gas saturation. Potential of mean force (PMF) was calculated to explain the enrichment of methane. We found that both the wall-gas interaction and the water-gas interaction are beneficial to the adsorption of methane and the former takes the leading role, while the gas-gas interaction impedes the adsorption. In addition, the effects of reservoir temperature, pressure, rock wettability, and carbon dioxide (CO2) on the occurrence states of methane-water mixtures near the solid surface were studied. The temperature, rock wettability and CO2 influence the occurrence states near the surface obviously, while the pressure in the simulating range does not. This study also suggests the potential of thermal exploitation, altering the rock wettability and CO2 injection to enhance shale gas recovery.
Key words shale; gas-water mixtures; organic matter; occurrence state; adsorption; potential of mean force
Gas production from unconventional source rocks, such as ultra-tight shales, has increased significantly over the past decade. However, due to the extremely small pores (~ 1-100 nm) and the strong material heterogeneity, gas transport in shale is still not well understood which poses challenges for predictive field-scale simulations. In recent years, digital rock analysis has been applied to understand shale gas transport at the pore-scale. A widely recognized issue with rock images (e.g., FIB-SEM, nano-/micro-CT images) is the so-called “cutoff length”, i.e., pores and heterogeneities below the resolution cannot be resolved, which leads to two length scales (resolved features and unresolved sub-resolution features) that are challenging for flow simulations. Here we develop a micro-continuum model, modified from the classic Darcy-Brinkman-Stokes framework, that can naturally couple the resolved pores and the unresolved nano-porous regions. Gas flow in the resolved macropores is modeled with Stokes equation. For the unresolved regions where the pore sizes are below the image resolution, we treat them as a continuum and develop an apparent permeability model considering non-Darcy effects at the nanoscale including slip flow, Knudsen diffusion, adsorption/desorption, and surface diffusion. This leads to a micro-continuum pore-scale model that can simulate gas transport in 3D shale images. We present case studies to demonstrate the applicability of the model, where we apply the new micro-continuum model to 3D segmented FIB-SEM shale images that include four material constituents: organic matter, clay, granular minerals, and macropore. We populate the model with experimental measurements (e.g., pore size distribution of the sub-resolution pores) and parameters from the literature, and identify the relative importance of different physics on gas production. Overall, the micro-continuum model provides a novel tool for digital rock analysis of organic-rich shale.
A novel approach is suggested to simulate the gas mixture transport in slit nanopores. The proposed method is based on the modification of the dual control volume grand canonical molecular dynamics (DCV-GCMD) method. The conventional method, DCV-GCMD, describes the gas mixture transport with pre-set constant composition. Due to the selective adsorption in the nanopores, the composition of the produced gas mixture will be affected. The modified method provides the composition in the permeate side. Gas mixtures of CH4/He and CO2/CH4 are investigated in graphene, graphite, and clay slit nanopores. The results show that pore size is the dominant factor in the species separation; the solid surface roughness has pronounced effect on gas separation; the influence of average pressure is not pronounced. The effects of pore length, temperature, pressure gradient, and feed composition are also investigated.
A series of canister desorption tests were carried out on 31 deep (over 3000 m) over-mature Lower Permian-Upper Carboniferous shale cores under atmospheric pressure and at reservoir temperatures of 75 and 80 °C, as well as a higher temperature of 95°C. Organic chemistry and X-ray diffraction were combined to investigate the impact of composition on canister desorption behavior. In order to better understand sorption and emission processes of shale gas, high-pressure methane sorption experiments were conducted at the reservoir temperature and pressures up to 18 MPa . Geochemical measurements show that the total organic carbon (TOC) content ranges from 0.488 wt % to 4.310 wt %. The depositional setting is lagoon and delta. The type of organic matter is mainly Type III. The dominant minerals of the shale samples are clay (25.4-97.0 wt %, average 58.8 wt %) and quartz (1-62.1 wt %, average 33.3 wt %). The content of clay minerals shows a significant negatively correlation with that of quartz with a coefficient of determination, R2, of 0.7260. The results show that an increase from the reservoir temperature to a higher value results in an average of 31% enhancement in desorbed gas volume. The desorbed gas volumes at both temperatures are linearly correlated with total organic carbon (TOC) content, which support the positive relationship with TOC found by the high-pressure methane sorption isotherms. The coefficients of determination, R2, at reservoir temperature and 95°C, are 0.6584 and 0.6444, respectively. The desorbed gas volumes at both temperatures show a slight correlation with clay minerals, indicating that adsorption sites on clay may have an impact on canister desorption testing. In addition, a slight negative correlation between quartz and desorbed gas volume was observed. The shale samples with a lower content of quartz and a higher desorbed gas volume generally have a higher content of clay minerals and TOC, which may indicate that clay minerals and TOC are the dominant contributors to the sorbed gas capacity, even after spontaneous imbibition.
Shale gas has redefined energy landscape[1]. The United States (U.S.) natural gas production is expected to increase every year, and in 2035 the U.S. shale gas production may raise to 50% of the total gas production.
Shale rock consists of micropores and mesopores[2]. It is also composed of inorganic minerals (quartz, clays, calcites, and feldspars, etc.) and organic matter (kerogens and bitumens). The organic matter is mainly composed of kerogens and it is considered as the main gas trapping of methane[3] and shows high capacity for carbon dioxide adsorption trapping[4]. An understanding of shale kerogen adsorption characteristics, under the reservoir condition, is required to successfully exploit the shale formations.
Three different types of kerogen depending on its origin can be distinguished: i) type I from a lacustrine anoxic environment, ii) type II from marine shale and continental planktons, and iii) type III from plants in tertiary and quaternary coals. All these types can be classified according to the elemental ratio of Hydrogen/Carbon (H/C), Oxygen/Carbon (O/C), and Sulfur/Carbon (S/C). The physicochemical properties (structure, adsorption, retention, etc.) of the kerogen strongly depend on its origin and on the burial history of the reservoir where it came from[5].
Molecular simulations allow a detailed picture of the structure, thermodynamics and dynamics of the fluid at the interface. In this work the dry structures of type II immature kerogen with different types of dummy particles are simulated. Then, the kerogen media are used to study the sorption of methane and carbon dioxide in the kerogen matrix and at the kerogen rough slit nanopore surface. Our results are compared with the available experimental data.
Wettability is a paramount factor in multiphase flow through porous media. The preference of the rock mineral surface to oil or brine determines fundamental flow functions in reservoir engineering like irreducible saturations, critical saturations, relative permeability and capillary pressure. Its implications affect all processes in oil and gas recovery, from primary production, waterflooding, to enhanced oil recovery. Except for a few specialized cases (low salinity waterflooding), reservoir modeling is done without considering wettability alteration during these processes.
There is experimental evidence in the literature that the presence of asphaltene can alter rock wettability towards more oil-wet. However, on one hand, modeling wettability of porous rocks is difficult because many complex factors interact in the system, which cannot be considered independently. Even a perfect independent description of rock, and two fluid phases is not sufficient to accurately model the wettability of a rock. On the other hand, asphaltenes (the heaviest, most polar component of crude oils) are a class of complex, not well-defined molecules, with behaviors that affect the water/oil/solid interactions and hence wettability. At the current state of the art, there is no comprehensive model that can predict wettability, not even without the complexity of asphaltene behaviors.
We present here a model based on the DLVO theory with fundamental forces in the equilibrium of a flat solid surface with two fluids (brine/oil, wetting/non-wetting) at nanoscopic scale. The force equilibrium results in a macroscopic contact angle formed by a drop of oil over a thin film of brine over a solid surface. With the inputs of physical parameters, such as asphaltene content, oil/water interfacial tension, electric surface potentials, and brine composition, the developed model is able to predict contact angle alteration as function of changes in any of the inputs.
Droplet impact on porous surfaces is a phenomenon of interest in several applications and is studied here in the context of wind-driven rain impacting building facades. The aim of the this work is understanding and controlling the droplet dynamics and penetration into the porous surface. Despite of the wide applications of this phenomenon, the studies in this field are mostly limited to the impact on impermeable surfaces so far.
During a heavy rain event, more and more droplets hit a porous surface, spread and get adsorbed. At some point, the porous material is saturated at its surface and the subsequent water droplets will lead to a water film to form on the surface [1]. This phenomenon is studied in this work in detail for different porous stones. The rain event is simulated using a train of droplets impacting at the same place, with definite time intervals. The water distribution inside the porous samples is obtained using neutron imaging at the Neutra beamline of the SINQ at Paul Scherer Institute (PSI), Villigen, Switzerland and the droplet impact dynamics is captured using shadowgraphy imaging with high speed camera. These two informations are combined together to understand the relation between the droplet spreading behavior, the moisture content at the surface and the water distribution within the porous stones during the train of impacting droplets.
It is found that the droplet spreading ratio increases with increasing surface moisture content, however this increment is different for each stone, depending on porosity, and pore size distribution, surface tension effects and energy dissipation on the wetted surface. An energy balance model for droplet impact on porous surfaces is proposed and the effect of moisture distribution inside the material on droplet spreading is determined. The model is compared with the experimental results showing good agreement. Moreover, the movement of wetting front inside the material is studied using the neutron images. It is seen that the water penetration into the stones depends mainly on impact velocity, but also a role of stone characteristics is identified. The water distribution inside the material cannot be considered neither as resulting from a uniform source, nor a point source.
An interface of phase transition is a macroscopic surface that separates the domain occupied by a single-phase gas or liquid from another domain occupied by two-phase gas-liquid mixture. In classical fluid dynamics an example of such a surface is the boundary of a cloud or a spray in the air. In porous media, this corresponds to the injection of miscible gas in an oil reservoir, which creates the zone of single-phase gas near the injection well, another zone of single-phase oil near the producing well, and a two-phase zone between them. Two mobile boundaries of the two-phase zone are the surfaces of phase transition. To model such systems, the method of negative saturations has been developed, which enables us to describe the overall fluid in all zones by a uniform system of extended two-phase equations and to solve it by direct numerical simulation without applying any special technique of front tracking. In this method the overall fluid is assumed to be pseudo two-phase. The extended system of equations is obtained from the principle of equivalence between the two-phase and single-phase flow.
Such an extended model of two-phase flow can be reduced to a single equation of nonlinear kinematic waves with respect to the gas saturation. It contains some non-classical terms like the extra-diffusion (the diffusion of phase saturation). This equation is parabolic in the subdomains that correspond to the true single-phase fluid, and hyperbolic in the zone of the true two-phase flow. Then a physical surface of phase transition is equivalent to the mathematical interface of transition between hyperbolicity and parabolicity (HP-transition). It is possible to show that such an interface represents a strong discontinuity of the solution (HP-shocks). We develop the extended Hugoniot-Rankine conditions and the entropy conditions for them. For one-dimensional Riemann problem, a graphical technique has been developed, which enables us to calculate the exact velocity of a HP-shock and the saturation behind and ahead of it. For numerical simulation of the extended two-phase model with HP-transition, two types of regularization are analysed: the introduction of small capillary or fictitious diffusion and the Kirchhoff regularization. Several examples of the solution are shown.
• Theoretical part of this work has just been accepted by Physical Review Letters
Long-term storage is necessary in order for CO2 geological sequestration in to be feasible. Leakage of CO2 by buoyant forces may occur if the trapping of CO2 in porous media is not stable with formation of large CO2 clusters. Ostwald ripening is a well-known phenomenon in two-phase mixtures that may affect bubbles’ stability. During an Ostwald ripening process in an open system, the gas in small bubbles dissolves in the surrounding fluid and diffuses to larger bubbles to grow them. Thus, there is concern about whether a coarsening of CO2 bubbles can occur and lead to such leakage after injection into porous media.
Here we show that Ostwald ripening will not lead to considerable coarsening of CO2 bubbles trapped in porous media. The size evolution of bubbles in a micron-scale porous medium has been shown very different from that in an open system. Unlike the coarsening typically observed in open systems, an initially polydisperse population of bubbles will ultimately become monodisperse and there is egalitarianism in bubble size for sufficient confinement in a homogenous porous medium, with gas from the larger bubbles diffusing to smaller bubbles.
Experiments conducted on a 2.5-D micromodel validate the ripening dynamics models, with the bubble population evolution dynamics well quantified. Our results show that this anti-coarsening effect is driven by the capillary pressure difference, and directed by the micron-scale geometric confinement. A physical model for the evolution dynamics on bubble population is derived from first principles with no empirical parameters, and this model matches our experimental data very well.
Based on our experiments and models, we conclude that, Ostwald ripening can be a positive effect, rather than a negative effect, to improve bubble stability. If bubbles are initially dispersed into small size (similar to or smaller than pore size) during injection, no significant coarsening will happen that lead to gas leakage. In addition, understanding this anti-coarsening effect of bubbles/droplets in porous media is also of great significance for better description and operations in many other applications which involve time scales similar to or longer than several hours, such as the formation and concentration of oil and gas in reservoirs (millions of years), transport of NAPLs in soil (days or even years), foam-based enhanced oil recovery (months to years).
In this work, it is shown that the one-domain approach (Goyeau et al, 2003) can be used to model precisely the average fluid velocity in a channel partially filled with a porous medium. This conclusion is drawn from the comparison of the averages obtained from the solution of the effective transport equations, with position dependent coefficients, and the ones resulting by direct integration of the local velocity values. These were obtained by the solution of the boundary value problem given by Stokes and continuity equations subjected to the no slip condition at the surface of the solid particles and the limiting walls of the channel. It must be mentioned that the comparison at the transition region between the porous media and the fluid is also very good.
Initially, the methodology was derived for porous media formed by parallel plates. In this way, the local velocity, the average velocity, the permeability and the fluid volume fraction could be obtained by analytical expressions (Ochoa-Tapia et al, 2017). However, recently the same kind of results has been obtained using particulate porous media, which for the comparison required the numerical solution of the fully developed Stokes flow problem in the whole system. In principle, the methodology could be extended for porous media of microstructure as complex as required.
Phase separation and the vapor free delivery of liquids is a challenge in a compensated gravity environment. Porous materials are used for liquid and vapor phase separation. They enable the transport (wicking) of liquid and provide a barrier against penetrating gas (bubble point). The wicking process is the imbibition of liquid into porous structures due to capillary forces [1].
To predict the liquid behavior inside porous materials, numerical simulations on the macroscopic level can be performed. The macroscopic parameters – porosity, pore radius and permeability - have to be known to perform macroscopic simulations. For this purpose, a real sample was scanned using X-ray tomography and a 3D model was reconstructed from it. CFD simulations were performed on the pore level using a 3D model and an appropriate representative volume element (REV) to determine the macroscopic parameters.
Wicking as imbibitional flow of a liquid driven by capillary pressure has been an important topic in different areas from simple applications like air refreshers and lightening torches to high-tech one such as Propellant Management Devices (PMD)[1, 2]. In this research, wicking of a liquid into porous wicks made parallel fibers is investigated. In the first step, we developed a model for predicting the height of liquid front during the wicking process as a function of time using Darcy’s law based on the assumption of single-phase flow behind a clear sharp front. Then the theoretically developed model was evaluated by performing mass-gain experiments where it revealed satisfying agreement for the majority of tested wicks. However, the model failed to account for partial saturations in the wicks. Therefore, we tried Richard’s equation to predict liquid saturation. The adapted Richard’s equation for our specific porous wicks is solved numerically in 2-D using COMSOL and analytically in 1-D using Mathematica where we treated the porous wicks as transversely-isotropic porous media. We determined capillary pressure and relative permeability directly from pore-scale simulations in wick microstructure using GeoDict. Also, in order to evaluate the numerical and analytical results for Richard’s equation, we used a new liquid-N2 based freezing technique to determine the saturation level along the wick length experimentally. After including the gravity effect, good agreements between the numerical/analytical predictions and experimental results were achieved in saturation distributions.
The porous pore doublet model that was published in 2008 [1] is presented and discussed. The background to the model is that fabrics used for fiber reinforced composite manufacturing often consist of fibers gathered in bundles.Thus, during manufacturing, the liquid resin impregnates a multiscale porous medium and there is a transport between pores of different scales driven by an applied pressure gradient and capillary action. As a simplified version of a fabric a porous pore-doublet model is studied in order to determine the characteristics of the flow. Experiments, as well as theoretical calculations on this generic geometry will be presented.
Seepage properties of porous media such as flow resistance, permeability, starting pressure gradient, gas flow and diffusion, and imbibitions have been received steady attention for decades in the area of pore-fractured porous media (PFPM) such as oil/water/gas reservoirs, hot dry rocks etc. The pore-fractured porous media, which consist of irregular pores in matrix with embedded fracture networks, widely exist in oil/water/gas reservoirs and hot dry rocks. Fractures in PFPM usually form networks and serve fluid (such as oil/water/gas) pathways, while oil/water/gas are usually stored in irregular pores around the fractured networks. Such pore-fractured porous media are often called dual-porosity media. Study of seepage properties in such media has been a challengeable and hot topic because microstructures of pores and fractures are extremely complicated. Usually, it is very difficult to find the seepage properties in the media analytically based on Euclidean geometry. Fortunately, available observations showed that the microstructures of naturally formed pores and fractures have the fractal characteristics, and the fractal geometry theory has been shown to be powerful in determination of seepage properties of the media. This paper attempts to summarize some progresses in applying the fractal geometry theory for pore-fractured porous media to analyze seepage properties in the media. Finally, some comments are made with respect to the theoretical developments and applications in this area in the future.
Multiphase flows in disordered porous media have a significant influence on many industrial applications, such as unsaturated soil mechanics, carbon geo-sequestration, and oil recovery. In this paper, we discuss two aspects of the problem, namely, (1) wettability at dynamic conditions and (2) disordered microstructures.
First, to simulate surface tension, wetting effects as well as dynamic contact angle at the pore-scale, a smoothed particle hydrodynamics (SPH) model with inter-particle force and modified solid-liquid interface formulation is applied. A newly introduced viscous force at solid-liquid interface is implemented to generate the rate dependent behaviour of contact angle with moving contact line, which can reflect the realistic interaction between different phases based on physical quantities. The dynamic contact angles are simulated under various moving contact line speeds and the correlation between corresponding capillary numbers are further analysed. The results are in good agreement with experimental observations under dynamic loading conditions and the effectiveness of the modified model is demonstrated.
Moreover, to study the effects of disordered pore networks on drainage characteristics, we used both numerical simulations (OpenFOAM) and physical experiments. The simulations covered a wide range of non-dimensional number, considering wettability, viscosity and surface tension. An improved Bond number equation for disordered porous media was proposed to capture the residual saturation as a function of gravity for different pore structures. Furthermore, we redefined the calculation procedure of the disorder parameter Iv, based on Voronoi tessellation, which is found to have a negative and linear correlation with the residual saturation for given gravity-driven drainage conditions. A qualitative agreement has been found between the simulation and experimental observations.
These preliminary results demonstrated the potential of using these proposed methods for accounting the pore-scale effects, including the hysteretic contact angle and disordered pore structures, and for simulating multiphase flows in porous media under various loading conditions.
The heterogeneous nano-scale pore structures have a significant effect on hydrocarbon properties and matrix mechanics in shale. Marine and lacustrine shales may have varying pore structures due to different depositional settings. Understanding the nano-scale pore structures in marine and lacustrine shales can provide important insights into fractal geometry theory. Therefore, we use the low-pressure nitrogen adsorption and new visual method of multi-scale field emission scanning electron microscopy (FESEM) to measure the geochemical and petrological characteristics of marine and lacustrine shale cores. Based on the fractal and multifractal analysis to study pore heterogeneities, we explicitly investigate the shale microstructures, including irregularities and distributions and their relations to transport phenomena.
In this work, we collect 32 lacustrine shale samples from Lower Jurassic Formation and 10 marine shale samples from Longmaxi Formation of Sichuan Basin in China. We carry out the spatial fractal analysis by nitrogen adsorption with Frenkel-Halsey-Hill (FHH) model, which has been widely applied in the characterization of the pore size distributions and properties. And use visible multi-scale FESEM images to investigate pore heterogeneities by multifractal analysis.
The results of low-pressure nitrogen adsorption with FHH model show that all samples, adsorption fractal dimensions can be classified into two regions based on fractal dimensions D1 and D2, corresponding to large pores and small pores, respectively. For both marine and lacustrine shales, D1 is generally larger than D2, indicating the weaker heterogeneity in small pore structures. Meanwhile, The fractal dimension D1 and D2 of lacustrine shale are generally lower than that of marine shale, indicating smaller heterogeneity in lacustrine shale, due to lower maturity and fewer organic pores.
Based on the features of multifractal spectrum curves, the results of the multifractal analysis show that the heterogeneities of lacustrine shale are generally lower than that of marine shale, which is the same as the FHH model. Meanwhile, the three multifractal parameters of multifractal spectrum curves depicted that the heterogeneities closely related to the shale composition, organic matter properties, and geophysical characteristics. The multifractal parameters of marine shale are generally larger than lacustrine shale, indicating that marine shale pore structures are more complicated.
Compared with results of FHH model by nitrogen adsorption, the results of multifractal analysis by multi-scale FESEM could provide more accurate information on the heterogeneity properties of shale samples. The measurements of petrophysical and geochemistry properties are closely associated and multifractal parameters. In addition, the quantitative analysis of multifractal parameters is in agreement with the field research, which is important for the evaluation of oil and gas flow system in shale.
Recent years, tight oil and gas has become a very important and reliable replacement for conventional energy. The porosity in tight reservoir is very small and there is a threshold pressure gradient (TPG) phenomenon and stress sensitivity in it, which make the fluid in tight reservoir hard to move and has a different employment law from the conventional reservoir, so we need a further research. Combined with the characteristics of tight reservoir, Hydraulic fracturing is the key technology for the development of tight reservoir. Traditional flowing model for multi-fractured horizontal wells in tight oil reservoirs are mostly based on classical Euclidean geometry. However, the structure of real tight oil reservoirs is highly heterogeneous and of multiscale, and the fractal theory provides a powerful method to describe the disorder, heterogeneity, uncertainty and complexity of the complex and disordered systems.
The fractal reservoir has been studied in the following aspects by making use of the Mathematical physical method, numerical calculation, computer graphics processing technology, software engineering and theoretical knowledge and application technology of the analysis method are based on the research results:
(1) Fractal geometric theory and method are good approximations to describe the complicity and we can more easily analysis all kinds of the pressure-sensitive deformable double media fractal complex reservoir relatively than others. Based on Warren-Root model, introducing fractal parameter( and ) and deformable coefficient( and ), this thesis constructed all kinds of seepage flow mathematical models to the pressure-sensitive deformable double media fractal reservoir with the defined production or closest outside boundary when the effects caused by permeability and porosity company with the pressure change ware concerned.
(2) Based on three linear flow, we established a well-testing model of fractured horizontal well in fractal reservoir, applied Laplace transform and Stehfest numerical inversion, and obtained the formulas of bottom hole pressure under the constant output. On the basis of model validation, this paper divided flow patterns and analyzed the influence of sensitivity parameters on bottom hole pressure
(3) It is shown that there mainly exists wellbore storage, transition flow, pseudo-linear flow, pseudo-bilinear flow and boundary flow and at the same time fractals dimensions has a great influence on the bottom hole pressure.
A recent experimental study [Herring et al., 2016] shows the potential of enhancing residual trapping of supercritical CO2 (scCO2) via cyclic injections. Two competing mechanisms were identified that impact residual scCO2 trapping: (1) the wettability of solid surfaces is altered due to direct contact with scCO2; (2) different capillary pressure results in different initial states of scCO2 fluid connectivity and topology prior to imbibition. To trap more scCO2 after imbibition, the former mechanism requires higher extent of scCO2 drainage while the latter requires lower extent of scCO2 drainage. Due to experimental limitations, control of local alteration of wettability in real rock is not possible; extensive and strict parametric study of cyclic injections are very expensive and difficult to achieve. Direct numerical simulation can largely overcome the above issues and reveal the relative importance of the two mechanisms. In this work, we employ our in-house developed lattice Boltzmann code to perform pore-scale simulations on micro-CT scans of Bentheimer sandstone to study the scCO2 trapping mechanisms in cyclic injections. Following the experimental procedure, we apply different capillary pressure to achieve different scCO2 configurations after drainage. Wettability is altered in the simulations on solid surfaces that are directly exposed to scCO2. The computation cost of parametric study is very high due to the multiple cycle injections and different combinations of parameters, thus manycore supercomputers are employed to perform the simulations.
Acknowledgements
This work was primarily supported as part of the Center for Geologic Storage of CO2, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science. Q.K. and H.V. acknowledge the support provided by LANL’s LDRD Program and the Department of Energy (DOE) Basic Energy Sciences program (DE-AC52-06NA25396). The supercomputers used in this work including TACC Stampede 2 and PSC Bridges both provided by XSEDE.
Investigating the mechanisms that govern flow of fluids at the pore-scale are the cornerstone of understanding multiphase flow in porous media for a wide range of applications, including hydrocarbon recovery, CO2 sequestration and contaminant hydrology.
Microfluidic devices, coupled with visualization techniques allow us to study pore-scale processes [1, 2]. Glass substrates are often preferred over silicon and polymers for the manufacturing of microfluidic devices because of their high transparency, thermal stability, hardness and chemical resistance. However, conventional manufacturing of glass-based microfluidic devices is time consuming, expensive, complex and multistep [3]. We have recently developed a novel and relatively inexpensive laser-based process that can be used for the fabrication of microfluidic devices using glass substrates [4]. Features generated on the glass surface by using a picosecond laser beam enable more complex shapes than the surface features produced by more conventional photolithography and etching. Although laser-generated micro-channels have a limited aspect ratio (typically < 4:1), much higher surface roughness than etched micro-channels, and angled walls with a rounded bottom, rather than steep, vertical walls with a flat bottom, these characteristics do not limit the ability to closely simulate real porous materials relevant to CO2 sequestration, hydrology or hydrocarbon recovery.
To design and fabricate appropriate micromodels we have coupled micromodel flow experiments and pore-scale numerical simulations to investigate fluid flow behaviour in micro-structures under various experimental conditions. Simulations are particularly useful for guiding the prototyping of micro-models, since the geometry, physical dimensions, and surface properties of micro-channels and pores in the structure have a significant effect on fluid flow dynamics. In this research, we pinpoint critically important parameters that have an impact on the multi-phase flow at pore-scale, which should be considered in the fabrication of micromodels. We investigate the fluid displacement front, saturation distribution, and the influence of the micromodel imperfections on the bulk flow will be evaluated. We utilize the TETHYS computational fluid dynamics (CFD) code, developed by Pacific Northwest National Laboratory and previously used to perform simulations of microfluidics experiments and complex 3D porous media flows [5, 6, 7].
Additionally, a set of dynamic flow tests will be performed on the fabricated micromodels to obtain valuable qualitative (flow images) and quantitative (pressure, flow rate) data. This will allow the comparison between the measured experimental data after performing fluid flow tests on the fabricated micromodels with the results of pore-scale numerical simulations. We will report the impact of different parameters e.g., surface roughness and aspect ratio on fluid flow.
Over the past decade, laboratory based X-ray computed micro-tomography (micro-CT) has given unique insights in the internal structure of complex porous materials in a broad range of applications, improving the understanding of pore scale processes and providing vital information for pore scale modelling. The non-destructive nature of micro-CT imaging, combined with dedicated X-ray transparent in situ equipment (eg. flow cells, tensile stages, heating and cooling stages) make it possible to monitor a changing pore structure over time in 3D. Recent advances in lab-based micro-CT hardware have pushed the temporal resolution from the hours down to seconds, enabling the visualization of fast dynamic processes and real-time imaging (Bultreys et al., 2016). Dynamic acquisitions however generate a vast amount of raw projection data, which needs to be reconstructed and further post processed. It is therefore key to quickly identify the interesting moments in time prior to reconstruction to optimize the amount of data that is generated, but also incorporated the added time dimension in the 3D analysis workflow to improve image quality.
In this work we present challenges and possibilities in dynamic micro-CT imaging for fast real-time acquisitions, reconstruction and analysis. The methodology and dedicated workflow from acquisition to analysis is illustrated using different flow experiments performed in a custom made X-ray transparent flow cell on limestone, sandstone and sintered glass samples. The first experiment, described in Boone et al. (2016), is single phase solute transport, where during continuous acquisition a tracer salt is injected in the pore space of a limestone sample. The capabilities of dynamic reconstruction of this experimental data is shown and analysis of the resulting 3D images enable distribution mapping of solute in the pore space through time. Other studies to be shown consist of two multiphase flow experiments of drainage and imbibition. In the drainage experiment, described in Bultreys et al. (2015), oil is injected in a brine saturated sandstone and the pore filling process can be visualized. By incorporating the temporal information in the 3D analysis of the pore space, individual pore filling events can be automatically identified and the size of these events monitored. For imbibition, on the other hand, water as wetting phase is injected in a sintered glass sample and the growth of water films and speed of pore filling can be analysed by reconstructing images from different time intervals and merging the appropriate temporal and local information for analysis.
Fuel Cell is considered one of the promising technology which can be utilized in many different applications, such as stationary, transportation, and portable usage. One of fuel cell efficiency limitation is water flooding at the cathode gas diffusion layer (GDL) at high current densities or low operating temperatures. There are many different experimental and theoretical studies regarding this issue, yet the cathode GDL design is not settled. In this study, the Buckley-Leverett method is used to predict the transient water percolation throughout the GDL. In order to model the very slow water flow rates expected in a GDL, the Buckley-Leverett method was modified by including capillary pressure effects. This modified Buckley-Leverett method has the advances of predicting the transient response of the water percolation inside the GDL as well as the maximum and minimum level of water saturation. Predictions of water saturation level using the modified Buckley-Leverett method show a good agreement with existing models and ex-situ experimental observations.
Pore-network modelling is an efficient method to simulate pore-scale multi-phase flow. The pore-network consists of a collection of idealized interconnected discrete network elements – pore nodes and pore throats. Capillary-dominated flow is modelled based on invasion-percolation rules. Although pore-network modelling is much less resource-demanding than direct simulation approaches, current implementations of the invasion-percolation algorithm are still time-consuming, in particular its phase clustering component, which identifies and determines phase trapping. Computational efficiency is essential when calculation of representative flow properties requires very large pore-network models of 10-100s millions network elements, for example for carbonates that exhibit multi-scale pore systems, and when performing multiple simulations for uncertainty analysis of, for example, wettability distributions. The present work introduces a new approach, denoted as dynamic phase connectivity, to track the changes in phase clustering after each displacement step. The relative permeability evaluation procedure has been optimised to accommodate for widely accessible multi-core CPU architectures. The combined speed-up factor of the proposed methodology is from two to three orders of magnitude compared to the best conventional pore-network modelling implementations.
While quasi-static pore network models (PNM) have been used to investigate the relative permeability (Kr) behaviour of reservoir rocks since the seminal work by Bakke and Øren (1997), the capacity of these models to capture the appropriate physics and predict experimental data remains contentious (Sorbie and Skauge, 2011; Bondino et al., 2012; Berg et al., 2016). It is generally accepted that PNM have to be calibrated against relatively cheap experimental data in order to predict more complex experiments. However, even after matching the former, it is not clear when and why PNM fail to reproduce experimental trends in the latter. We propose that the main reason for this is that the information in the Pc and Kr-curves is extremely densely encoded: many aspects of the porous medium and the fluid behaviour are lumped into the same parameter. PNM are thus very likely overfitted, making it hard to assess whether model failure is due to inappropriately trained model parameters or due to a fundamentally lacking description of the pore scale flow physics.
To unravel this, there is a need for experimental data which probes the underlying properties that make up a Kr-curve. Thanks to recent innovations in micro-computed tomography (micro-CT), it is now possible to image a rock’s pore scale fluid distribution during flow experiments. We propose that this data can be used to constrain the source of errors in a PNM, if the issue is addressed by investigating three questions in specific order (by decreasing fundamentality):
1. Does the model predict the correct pore scale fluid distributions? This relies on the accuracy of the pore and throat radii, on the assigned contact angles in each pore and throat, and on the physics implemented in the invasion algorithm.
2. Does the model predict the correct saturation for a certain fluid distribution? This depends on the volume partitioning between pores and throats and thus on pore/throat length partitioning, which is arbitrary; as well as on the description of wetting layers.
3. Does the model predict the correct flow rate for a certain fluid distribution? This is chiefly determined by calculation of the pore, throat and layer conductivity in the model.
In this work, we focus on the first question. We present a novel modeling/experimental methodology to compare the pore-by-pore fluid distributions in a PNM to micro-CT based steady-state flow experiments. Using this analysis, we assess the importance of errors in the filling state of pores and throats to the flow, while seeking to split off the influence of volume and conductivity assignment. The influence of the pore scale contact angle distribution is investigated by automatically measuring a large number of contact angles on an in-situ micro-CT image (AlRatrout et al., 2017). Each of these contact angles is assigned to the pore that corresponds with its location in the image. This allows to perform PNM simulations with directly measured contact angle distributions for the first time. We explore how this influences PNM simulation errors by comparison to simulations with contact angles drawn from random distributions.The correctness of an adaptated version of the PNM approach outlined in Dong (2007) and Valvatne and Blunt (2004) is investigated by applying the methodology to a water-wet Bentheimer sandstone. In drainage, the model is found to reproduce the main oil flow paths in the experiment, while however significantly overestimating the corresponding water saturation. In imbibition, tentative results show a higher level of snap-off in the experiment compared to the model.
Acknowledgements:
Steffen Berg and Øve Wilson (Shell) are thanked for valuable comments and discussions on this work. Shell is acknowledged for financial support and for permission to present this abstract.
The capillary entry pressure (Pce) and corresponding pore throat size control the thickness of an oil or gas column that may be sealed beneath a mudrock. Mudrock seals typically have nanometer-scale pore throats, and the Pce often exceeds the minimum horizontal effective stress in these rocks. Mudrock seals can fail through fracturing either by buildup of fluid pressure or during faulting or folding, creating fractures with much smaller Pce than the pore space, allowing hydrocarbon to escape. Understanding the factors at the grain and pore scale that influence Pce is an important component of risk assessment and prospect evaluation.
Prediction of capillary pressure and relative permeability behavior is especially difficult in shaly sands and mudrocks as the sediments are usually heterogeneous mixtures made of disparate grains shapes and sizes. Mudrocks are primarily a mixture composed of silt (coarse) and clay (fine) grains. The porosity for a mudrock varies according to the fractions of these components, and the minimum porosity is achieved when clay particles occupy all the interstitial space between the larger silt particles (Daigle and Reece, 2014). Schneider et al. (2011) showed that silt bridging, in which silt-sized grains are present in an abundance sufficient to create a connected stress chain through the rock matrix, preserves large pore throats, thereby affecting the permeability as well as Pce.
To simulate these effects at grain scale, bidisperse sphere packs were generated by a cooperative rearrangement algorithm for efficient packing with the ratio of larger to smaller spheres as 5:1. We varied the different sphere volume fractions to mimic mudrocks. The generated sphere packs were subsequently converted to pore throat models by employing Delaunay tessellation similar to Mehmani and Prodanovic (2014). Following Mason and Mellor’s (1995) approach, an invasion percolation algorithm was applied to the models to generate drainage and imbibition capillary pressure curves and relative permeability curves.
It was observed that on successive increase of the fraction of larger spheres in a sphere pack, the capillary pressure curves displayed a dual percolation threshold behavior. This can be explained by larger pores being preserved adjacent to the spheres due to the silt-bridging effect as implied by Schneider (2011). These larger pores are responsible for the lower percolation threshold, while smaller pores between smaller grains give rise to the higher capillary threshold. Increasing the larger sphere fraction also decreased the residual water phase saturation, possibly due to the trapped phase escaping through the larger pores. This confirmed that the concentration and radius-ratio of the grains strongly affect the capillary pressure behavior in dual-component systems like mudrocks.
The next step in this project will be exploring the capillary behavior at different sphere-radius ratios and distributions, and for sphere packs generated by sequential addition under the effect of gravity. The principal application of this work is improved assessment of seal capacity from microstructural data in deep-water, subsalt environments. Ultimately this work may affect estimates of reserve capacity of reservoirs and risk management efforts in developing prospects by providing a more accurate understanding of seal quality.
We study immiscible fluid-fluid displacement in rough-walled fractures with a focus on the combined effect of wettability, the viscous contrast between the two fluids, and fracture surface topography on drainage patterns and interface growth. We have developed a model to simulate the dynamic displacement of one fluid by another immiscible one in a rough geological fracture; the model takes both capillary and viscous forces into account. Capillary pressures at the fluid-fluid interface are calculated based on the Young-Laplace equation using the two principal curvatures (aperture-induced curvature and in-plane curvature), while viscous forces are calculated by continuously solving the fluid pressure field in the fracture. The aperture field of a fracture is represented by a spatially correlated random field, with a power spectral density of the fracture wall topographies scaling as a power law, and a cutoff wave-length above which the Fourier modes of the two walls are identical. Results show that the model is able to produce displacement patterns of compact displacement, capillary fingering, and viscous fingering, as well as the transitions between them. Both reducing the aperture variability and increasing the contact angle (from drainage to weak imbibition) can stabilize the displacement due to the influence of the in-plane curvature, an effect analogous to that of the cooperative pore filling in porous media. Based on scaling analysis, we derive a relation between a dimensionless interface-smoothing parameter ($N_{Sm}$), defined by wettability and aperture variability, and the commonly used capillary number ($N_{Ca}$) and mobility ratio ($M$). This relation gives a surface in the three-dimensional ($N_{Ca}$–$M$–$N_{Sm}$) parameter space, which predicts the separation of the stable and unstable displacement regimes.
The behavior of CO2 inside a reservoir (i.e., two-phase flow in CCS, or three-phase flow in CCUS) is influenced by interfacial tension, pore structure, wettability and other reservoir parameters (e.g., pressure gradient), which vary significantly from one reservoir to the next. Therefore, understanding multi-phase flow under various reservoir conditions is crucial to estimating CO2 storage capacity, leakage risk, and storage efficiency. In this study, we calculated two-phase or three-phase fluid displacements in natural digital rocks using lattice Boltzmann (LB) simulation, and characterized the influence of reservoir conditions (e.g., interfacial tension, pressure gradient, and wettability) upon CO2 behavior. By mapping the CO2 saturation on the diagram of capillary number and viscosity ratio of the two fluids, we could identify the suitable environments for effective CO2 storage. We further calculated two-phase and three-phase relative permeability of digital rocks under various conditions. Porous flow simulation also contributes to geophysical monitoring of CO2 behavior in reservoirs. For example, we calculated geophysical properties (e.g., seismic velocity) of the digital rocks with injected CO2 under various reservoir conditions. Using the relationship between seismic velocity and CO2 saturation parameterized by reservoir conditions, we could quantify in situ CO2 saturation in reservoir from geophysical monitoring data (seismic velocity). In this presentation, we would like to show how porous flow simulations practically contribute to the safe and effective CO2 storage.
The behavior of CO2 inside a reservoir (i.e., two-phase flow in CCS, or three-phase flow in CCUS) is influenced by interfacial tension, pore structure, wettability and other reservoir parameters (e.g., pressure gradient), which vary significantly from one reservoir to the next. Therefore, understanding multi-phase flow under various reservoir conditions is crucial to estimating CO2 storage capacity, leakage risk, and storage efficiency. In this study, we calculated two-phase or three-phase fluid displacements in natural digital rocks using lattice Boltzmann (LB) simulation, and characterized the influence of reservoir conditions (e.g., interfacial tension, pressure gradient, and wettability) upon CO2 behavior. By mapping the CO2 saturation on the diagram of capillary number and viscosity ratio of the two fluids, we could identify the suitable environments for effective CO2 storage. We further calculated two-phase and three-phase relative permeability of digital rocks under various conditions. Porous flow simulation also contributes to geophysical monitoring of CO2 behavior in reservoirs. For example, we calculated geophysical properties (e.g., seismic velocity) of the digital rocks with injected CO2 under various reservoir conditions. Using the relationship between seismic velocity and CO2 saturation parameterized by reservoir conditions, we could quantify in situ CO2 saturation in reservoir from geophysical monitoring data (seismic velocity). In this presentation, we would like to show how porous flow simulations practically contribute to the safe and effective CO2 storage.
We developed a sharp-interface level-set method for immiscible pore-scale two-phase flow with a thin wetting film on the solid surface. The lubrication approximation is used to model the thin-film equation efficiently. The incompressible Navier–Stokes, level-set, and thin-film evolution equations are coupled sequentially. Hamilton–Jacobi level-set reinitialization is employed to construct the signed-distance function, which takes into account the thin film on the solid surface. The level-set simulation method is validated and shown to match the augmented Young–Laplace equation for a meniscus in a capillary tube. Viscous bending of an advancing interface over a precursor film is captured by the level-set method and agrees with the Cox–Voinov theory. We model the evolution of an advancing bubble surrounded by a wetting film. The predicted film thickness compares well with both theory and experiments. We then demonstrate that the multiscale level-set approach can model immiscible two-phase flow with a capillary number as low as 10−6.
We developed a sharp-interface level-set method for immiscible pore-scale two-phase flow with a thin wetting film on the solid surface. The lubrication approximation is used to model the thin-film equation efficiently. The incompressible Navier–Stokes, level-set, and thin-film evolution equations are coupled sequentially. Hamilton–Jacobi level-set reinitialization is employed to construct the signed-distance function, which takes into account the thin film on the solid surface. The level-set simulation method is validated and shown to match the augmented Young–Laplace equation for a meniscus in a capillary tube. Viscous bending of an advancing interface over a precursor film is captured by the level-set method and agrees with the Cox–Voinov theory. We model the evolution of an advancing bubble surrounded by a wetting film. The predicted film thickness compares well with both theory and experiments. We then demonstrate that the multiscale level-set approach can model immiscible two-phase flow with a capillary number as low as 10−6.
The complex fluid flow processes occurring during EOR methods are difficult to implement in numerical modeling. Therefore, this presentation aims to share some light on experimental results from microfluidics studies conducted for various multi-phase fluid displacement processes with a focus on chemical EOR methods using brine, polymer-, alkaline- and surfactant-solutions. A wide range of important fluid-fluid and fluid-grain interactions can be observed and characterized using the visual outcomes of the experiments. Upscaling to continuum-scale reservoir simulation still poses challenges, but the microvisual studies can support these.
Applications ranging from water injection all the way to ultra-low interfacial tension surfactant displacement processes will be shown including some numerical simulation results of the experiments.
Considering the paradigmatic case of random piles of spheres, fluid front morphologies emerging during slow immiscible displacement with a global front velocity of 3 µm/s are investigated in real time by X-ray micro–tomography and quantitatively compared with model predictions. Controlled by the wettability of the bead matrix two distinct displacement patterns are found with a transition region of about ≈ 30° separating both regimes. Within each regime the displacement behavior is fairly insensitive to the exact contact angle [1].
A compact front morphology emerges if the invading fluid wets the beads while a fingered morphology is found for non–wetting invading fluids, causing the residual amount of defending fluid to differ by one order of magnitude. The corresponding crossover between these two regimes in terms of the advancing contact angle is governed by an interplay of wettability and pore geometry and can be predicted on the basis of a purely quasi–static consideration of local instabilities that control the local progression of the invading interface as similarly introduced by Cieplack and Robbins for 2D systems [2], [3]. In particular the absence or the appearance of ‘Burst-Instabilities’ where the local instability occurs when the pressure in a throat exceeds the required filling pressure can be used to distinguish the transition between both wetting regimes. For the non-wetting system where the local front progression occurs mainly by ‘Burst-Instabilities’ an interconnected and very extended network of invading and defending phase develops at later times. The interconnected network of the defending phase is slowly drained by gutter flow leading to an increased residual saturation which is reached only after a substantial flush of the invading phase. If time allows, also the situation of bead and grain packs consisting of heterogeneous bead sizes, grain shapes or heterogeneous wettability will be discussed in brief.
Solute transport in porous media is important for several industrial applications, i.e.: hydrology, building stone performance and waste management. Spreading and mixing during solute transport is significantly impacted by the pore scale heterogeneity found in natural porous media, which complicates upscaling (Dentz et al., 2011). Therefore, simulations and experiments which investigate the evolution of pore scale solute concentration fields in such materials are very valuable. However, direct visualisation of these concentration fields at the micron-scale in rocks is complicated by the high spatial and time resolutions that are required. Bultreys et al. (2016) and Boone et al. (2016) present first tests on imaging solute transport in a carbonate rock using fast laboratory-based micro-CT. In this study, we extend this work by attempting to quantify micro-CT concentration fields, in order to investigate spreading and mixing under different flow conditions in porous materials with different degrees of heterogeneity.
A significant part of this work is aimed at the methodological challenge of performing in-situ micro-CT scans of solute transport with imaging times on the order of seconds. We use the EMCT scanner of UGCT (www.ugct.ugent.be), a micro-CT system specially designed for in-situ imaging, with a rotating X-ray tube and detector in a horizontal plane (Dierick et al., 2014) and investigate the quantitative correctness when imaging the concentration of a dissolved tracer salt (0 wt%, 2 wt%, 4 wt%, 6 wt%, 8 wt% and 10 wt% CsCl) in porous sintered glass at 12 seconds per scan, with a voxel size of 13 micron. The CsCl-concentration increases the X-ray attenuation coefficient of the fluid, which causes an increase in grey values observed in the reconstructed micro-CT datasets. The high temporal resolution at which the micro-CT images are taken, is inherently linked with a limited signal-to-noise ratio. Despite this drawback, the first experimental results suggest a linear relationship between the grey values of the tracer-solution in the fast scans and the tracer concentration.
Results from the presented experiments can be used to investigate flow structures at the pore scale and to validate pore scale solute transport simulations. Further development of the methodology could also lead to valuable insights in multi-phase solute transport and reactive transport.
We consider the problem of advection, matrix-diffusion and bimolecular reactions in fracture-matrix systems, with two example applications: (i) Weathering reactions in fractured bedrock and (ii) in-situ chemical oxidation (ISCO) for remediation of fractured rock. In both cases, a reagent (a weathering agent such as H+ or dissolved oxygen, or permanganate in the case of ISCO) are supplied through a fracture, and react with a second species (immobile mineral species in the case of weathering reactions, or TCE/PCE in the case of ISCO) initially contained in the fracture-matrix system. In both cases, moving reaction fronts form and propagate along the fracture and into the rock matrix. The propagation of these reaction fronts is strongly influenced by the heterogeneity/discontinuity across the fracture-matrix interface (advective transport dominates in the fractures, while diffusive transport dominates in the rock matrix). We present analytical solutions for the concentrations of the oxidant/weathering agent, weathering mineral or TCE and natural organic matter; and the propagation of the reaction fronts in a fracture-matrix system. Our approximate analytical solutions assume advection and reaction dominate over diffusion/dispersion in the fracture and neglect the latter. In the ISCO problem, the behavior of the reaction-diffusion equations in the rock matrix is posed as a Stefan problem where the supplied oxidant reacts with both diffusing (TCE) and immobile (natural organic matter) reductants. Our analytical solutions establish that the reaction fronts propagate diffusively (i.e. as the square root of time) in both the matrix and the fracture. Our analytical solutions agree very well with numerical simulations for the case of uniform advection in the fracture. In the context of the ISCO problem, we also present extensions of our analytical solutions to non-uniform flows in the fracture by invoking a travel-time transformation. These non-uniform flow solutions are relevant to field applications of ISCO, which employ forced-gradient flow systems. Our approximate analytical solutions are relevant to a broad class of reactive transport problems in fracture-matrix systems where moving reaction fronts occur, and may be generalized further to consider multiple interacting species.
We study the Lagrangian dynamics of steady three-dimensional (3D) Stokes flow over granular media consisting of simple cubic (SC) and body-centered cubic (BCC) lattices of closed-packed spheres, and uncover the mechanisms governing chaotic fluid advection. Due to the cusp-shaped sphere contacts, the topology of the skin friction field is fundamentally different from that of continuous (non-granular) media (e.g. open pore networks), with significant implications for fluid advection. Weak symmetry breaking of the flow orientation with respect to the lattice symmetries imparts a transition from regular advection to strong chaotic advection in the BCC lattice, whereas the SC lattice only exhibits weak advective mixing. Using a numerical simulation of the flow at various flow orientations, we quantify the strength of chaotic mixing from the Lyapunov exponent, and examine how it is distributed over the parameter space of mean flow orientation [1]. We furthermore analyze the flow topology and show that the occurrence of chaotic advective mixing is controlled by the existence within the flow of transverse intersections between stable and unstable manifolds originating from the spheres [1]. These insights are used to develop accurate predictions of the Lyapunov exponent distribution over possible flow orientations [2]. The difference of behavior observed for the SC and BCC lattices, which share the same symmetry point group, results from their different space group symmetries: a glide symmetry of the BCC lattice allows the occurrence of chaotic advection [1]. These results point to a general theory of advective mixing and dispersion based upon the inherent symmetries of arbitrary crystalline structures.
Few current bioreactive transport solvers currently provide a comprehensive mechanistic description of biogeochemical cycles and allow easy integration of all the involved processes, including flow in variably saturated media, solute transport and kinetic/equilibrium biochemical reactions. The parameterization of these processes is particularly challenging since our knowledge of model parameters and of their spatial heterogeneity is typically incomplete. Therefore, it is crucial to study the impact of each process and related parameters uncertainty in the outputs of interest.
In this communication, we consider reaction networks describing biochemical degradation of herbicides, atrazine and glyphosate (la Cecilia and Maggi, 2017a, 2017b). In particular, we quantify the degradation potential and accumulation of toxic substances and we study the biomass and ecological structure dynamics in soil and groundwater. Our discussion encompasses the assessment of model outputs sensitivity on the identified model structure and related parameters. We focus on the effects of local mixing, which entails the full characterization of spatial and temporal fluctuations of solutes concentration within the time-space domain. We frame these analyses in terms of dimensionless parameters describing the integrated processes of interest, i.e., fluid flow, solute transport, and biogeochemical reactions and we identify global trends and statistical indicators to characterize the system at steady state. This work will address system nonlinearities linked to the interplay of diverse processes with the aim of increasing our understanding of dominant mechanisms involved in agrochemicals bioreactive transport.
Semi-Analytical Particle Tracking Scheme For Advective/Diffusive Transport in Porous Media
The particle tracking scheme of David W. Pollock [Ground Water 26(6), 1988] provides a computationally efficient and mass-conservative method for Lagrangian transport in the absence of diffusion. In this work, a generalization of Pollock's scheme that allows for the inclusion of diffusion is presented. The new scheme is based on a semi-analytical representation of the advective/diffusive motion. The scheme does not require stepping at sub-grid-cell time or length-scales and thus is computationally efficient. It is formulated in such a way that it becomes exact for Pe going to zero and infinity, and provides an accurate numerical approximation in the intermediate Pe number range. Application examples dealing with Darcy flow in heterogeneous porous media and Stokes flow in resolved pore-space geometries document the capabilities of our new scheme.
Residual trapping is one of the key trapping mechanisms for CO2 geological storage, yet difficult to determine in-situ. The present study addressed determination of residual trapping over the entire range of scales from pore to core to field scale, based on data from Heletz, Israel [1] pilot CO2 injection site. During 2016-2017 two dedicated push-pull experiments have been carried out at the site for the specific purpose of quantifying the residual trapping in-situ, in a well-characterized reservoir layer at 1.6 km depth. The field experiments use a combination of hydraulic, thermal and/or tracer tests before and after creating the residually trapped zone of CO2 and the difference in the responses of these tests is used to estimate the residual trapping of CO2 in-situ. The first experiment is based on hydraulic withdrawal tests before and after the creation of the residually trapped zone. In this experiment, the residually trapped zone was also created by fluid withdrawal, by first injecting CO2, then withdrawing fluids until CO2 was at residual saturation. In the second test, the main characterization method is injection/withdrawal of water and partitioning tracers, whose recovery with and without residually trapped CO2 in the formation is compared. In the second experiment the residually trapped zone is created by first injecting CO2 and then injecting water saturated with CO2 in order to push away the mobile CO2. The experimental field results have been modelled both with simplified analytical models for guidance and with ‘full-physics’ TOUGH2 [2] simulators, to match the observations and to obtain values for in-situ residual trapping. The resulting estimates are discussed as well as compared to results from laboratory measurements on rock cores, including their modeling with pore network models. The pore network modeling has been based on laboratory data on rock samples from the site. In the first set of pore network modeling, data such as throat size distribution, permeability and characteristic two-phase flow functions [3] were used to calibrate the model, while the second analysis was based on actual scanned pore space data [4] along with measured hydraulic values.
Reactive transport in river corridors can be greatly complicated by fluctuations in the boundaries, which may cause changes from gaining to losing over time. The seasonal influxes of water cause wetting and drying of soils near the river, rainfall causes distributed periodic inputs to the surface, and chemical and physical heterogeneity affect the possible reaction sites. The combination of these behaviors can be computationally challenging to simulate because there is an apparent hysteresis in the reaction rates when the system is wetting versus drying. Most studies that have sought to model these behaviors in the past have used a strictly Eulerian approach. Here a hybrid Eulerian-Lagrangian method is used to investigate the differences in upscaled reaction rates under transient conditions. The simulation framework uses the Eulerian integrated model ParFlow to simulate a transient, variably saturated flow field that represents an aquifer adjacent to a fluctuating river with sparse rainfall. The pressure field is used to compute seepage velocities for a colocation based, reactive, random walk particle tracking algorithm. Complex reactions on particles are simulated by assigning multiple component masses to each particle, allowing nearby particles to exchange mass of the components, evaluating multi-component reactions on each particle independently, then moving the particles. Boundary conditions, spatially variable rate laws, interactions with solids, exposure time based reactions, and other rules are easily incorporated, allowing a high level of realism with regard to the physical processes. The approach is also unique because it separates spreading and mixing processes, which have fundamentally different impacts on reactive transport. We show that the framework is able to simulate the reaction hysteresis of wetting and drying conditions and discuss the potential applications, advantages, and limitations of this hybrid approach.
The recent developments of microfluidic is offering a new and efficient tool to visualize transport processes of bacterial fluids from microscopic to macroscopic scales and assess the influence of well controlled environments. We explored a situation where the motility and pore geometry are the dominant ingredients influencing the hydrodynamic dispersion of a bacterial fluid.
To this aim, we designed a microfluidic channel mimicking natural pore structures and choose a motile strain that does not stick to the surfaces. Doing so we were able to track bacteria from the pore scale and up to 15 pores. The statistical information obtained at that scales are then upscaled using continuous time random walk. In this presentation, we will show that the coupling between motility and flow induces a motility induces trapping effect and we will show its influence on the large scale transport.
Nonaqueous phase liquids (NAPLs) are still a major challenge for all traditional groundwater treatment technologies. NAPLs often contaminate the subsurface following an accidental spill or due to a defect in the oil storage tank. These pollutants remain trapped in the form of droplets and / or immiscible clusters within the aquifer, thus constituting a persistent source of pollution that is difficult to decontaminate. Predicting the fate of this pollutant requires characterizing all the mechanisms involved and in particular the biodegradation, which can occur in the vicinity of the pollutant source or further, to the dissolved plume. If a significant research effort has been put into investigating the transport and biodegradation of dissolved contaminants, comparatively very few works (e.g., Bahar et al., 2016) are focused on the study of such processes in multiphase conditions (oil/water/biofilm systems).
In this study, we give an attempt to address this open issue from an experimental and numerical perspective. First, we illustrate impact of bacteria on dissolution of pure organic phase from micromodel experiments. The experimental set-up is made of a micromodel (i.e. 2D transparent flowcell) used to study the dissolution of oil phase. Changes in toluene saturation are directly monitored from recorded two-dimensional images and dissolved concentrations at the outlet are measured by gas chromatograph. Results of toluene dissolution and biodegradation by a toluene-degrading strain (Pseudomonas putida F1) are compared with experiments in abiotic conditions.
In parallel, we present a two-dimensional pore-scale numerical model (Benioug et al., 2017) to investigate the main mechanisms governing biofilm growth and NAPL dissolution in porous media. Fluid flow is simulated with an immersed boundary–lattice Boltzmann model while solute transport is described with an interface reconstruction finite volume approach (Benioug et al., 2015). A cellular automaton algorithm combined with the immersed boundary method was developed to describe the spreading and distribution of biomass. Different conditions are considered (spatial distribution of biofilm, reaction kinetics, biosurfactant production, NAPL toxicity) and their impacts on the dissolution process are analyzed.
Geophysical monitoring of bacterial activities in subsurface has drawn significant interest in various civil engineering, hydrocarbon recovery, soil remediation practices. This study explored the feasibility of use of complex electrical responses to monitor bacterial biofilm formation in soils. Two runs of column experiments were conducted, in which the model bacteria Shewanella oneidensis MR-1 were cultured in a sand-pack and stimulated to form biofilms. During the bacterial growth and biofilm formation, the variations in complex electrical conductivity were monitored at a frequency range of 0.01–1000 Hz. As a result of the bacterial growth and biofilm formation, it was observed that the imaginary conductivity significantly increased by more than 500% and the real conductivity was reduced by more than 13% in the both runs. However, we observed the spatial variations in the complex conductivity values, showing the greatest variations near the nutrient injection port occurred and the least variations near the outlet fluid port. It appeared that the imaginary conductivity effectively captured bacterial growth and biofilm formation in porous media, while the real conductivity was heavily affected by porosity reduction as well as pore fluid conductivity. The obtained results suggest that complex conductivity can be effectively used to capture the bacterial growth and biofilm formation in soils.
The injection of substrates, e.g. hydrogen with the purpose of energy storage, into subsurface structures could stimulate the growth of all present microbial species which are able to use this substrate for their metabolism. The linkage between transport, the growth of microorganisms, substrate availability and biodegradation results in a strongly coupled dynamic system. The difficulty in the development of a general model is the inclusion of processes which appear on different length and time scales. In this work, a flexible numerical model was developed which uses effective representations of the processes on Darcy scale. The mass exchange between two phases (gas and water) is treated instantaneously by using an equilibrium law. Multiple metabolic reactions can be included kinetically by defining the stoichiometry and kinetic coefficients. Different mathematical models can be selected to describe the substrate-limited microbial growth. Microorganisms can be considered as an immobile biofilm or as partially mobile within the water phase. The jump-like appearance of usual growth models and the strong coupling to the reactive transport equations results in a very stiff equation system. The numerical instability was overcome by a proper adaptive time step selection and a check for the physical possibility of each solution before it is continued to the next one. Example simulations are shown for a near wellbore and a field scale study of an underground hydrogen storage.
Life in porous media, as soil bacteria, are used since more than 40 years ago as bio-fertilizer contributing to the development of a sustainable agronomy. Even though they are extensively used due to their low cost, such biotechnology is still far from being efficient and many challenges are opened for basic research in porous media science.
Our microbiological system of study are the Bradyrhizobium diazoefficiens, bi-flagellar bacteria. One of the opened question regarding its efficiency is which is the aim of its bi-flagellum system developed by this specie and not developed in general in Bradyrhizobium. It may be an adaptive trade-off between energetic costs and ecological benefits among different species. We work interdisciplinary on bacteria diffusion in porous media, numerically and experimentally, imitating the complex and structured soil with artificial micro-fluidics devices. With a better visualization in transparent devices, easy to manipulate in a laboratory we aim to understand and control the transport properties of the system for further realistic applications. In this work we were able to report numerically their recent reported strategies to swim [Quelas et at, Sci.Rep. 2016]. Further we simulate their dynamics with those realistic parameters, under a broad spectrum of micro-confinement environments, contributing to micro-fabricate different geometries of porous media. All these studies contribute to understand their diffusion properties versus their flagellar systems (different motility) and versus the porous arrangement. These in vitro contributions hopefully will be useful for further development in a sustainable agronomy.
Shakil A. Masuma*, Hywel R. Thomas
Geoenvironmental Research Centre, School of Engineering, Cardiff University, Cardiff, United Kingdom
*masumsa1@cardiff.ac.uk (corresponding author’s E-mail)
Sequestration of anthropogenic carbon dioxide in deep geological formations, such as, saline aquifers, un-mineable coal seams is a plausible way to reduce global greenhouse gas effects. Safety and performance efficiency of storage reservoirs are of utmost importance. Long-term containment of sequestrated CO2 can be achieved by preventing leakage and by ensuring further entrapment such as solubility-trapping and mineral-trapping. These processes can be enhanced by involving subsurface microbial community that restrict flows by forming biofilms and/ favours biomineralization.
Subsurface flows and reactions are complex and often involve multiple phases, chemicals and minerals as well as pressure and thermal gradients. To study such complex interactions, a numerical model has been developed under a coupled Thermal-Hydraulic-Chemical-Mechanical (THCM) framework including subsurface microbial processes and associated bio-geochemical reactions. The model deals with liquid flow, multicomponent gas flows, dissolved chemicals and suspended microbes flows in liquid phase, heat flow, biofilms and minerals growths, mechanical deformations and geochemical/bio-geochemical reactions. The model considers multiple species of biofilms (attached to solid phase) and suspended microbes (in liquid phase). The model predicts microbial growth by both respiration and fermentation. Microbes decay due to endogenous cell death, fluid shear, biocide decay etc. Attachment/detachment of cells to/from biofilm phase to suspended cells also alter their respective quantities. At high pH conditions, growth is favoured while at low pH it slows down. The presence of multiple chemicals, minerals and gas influence transport and availability of nutrients, substrates to microbes as well as facilitates various geochemical and bio-geochemical reactions.
In this presentation, applications of the model to simulate microbial growth in multiphase condition are presented. Flow of substrate in a solution containing multiple dissolved chemicals are considered. Therefore, the effects of microbial activities on solution composition and minerals (and vice-versa) in the system are envisaged. Overall impacts of microbial processes on thermal condition and flow properties of the medium are discussed.
With the increasing demands for sustainable and eco-friendly soil improvement methods, utilization of microbial activities in subsurface has received increasing attention as an way to modify and control the mechanical and hydraulic properties of soils. Many bacteria can produce biofilms, which are matrices of organic materials consisting of microbial cells and extracellular polymeric substances (EPS). The formation of biofilms in soil can cause pore clogging, hence reduce the permeability by several orders of magnitude. Therefore, stimulating the biofilm formation has been proposed as a method to seal cracks and leakage in earth structures, such as embankments, dams, and levees. However, because of the possibility of biofilm removal or degradation over time, the durability of biofilms over a long period has been questioned, and this has been hampering the implementation of biofilms in field-scale engineering practices.
Herein, we explored the feasibility of using submicron fine particles to maintain or even enhance the clogging and permeability reduction in coarse sands. It was hypothesized that biofilms could trap and retain fine particles in their matrices, and this is expected to increase the durability of the induced clogging. To test this hypothesis, we performed the control test and the bentonite slurry injection test. Bacillus subtilis were chosen as model bacteria. In the control test, B. sutilis were cultured and stimulated to form biofilms formation in a clean sand-pack while monitoring the permeability. The permeability was reduced by ~50% after 1–2 weeks of experiments, and de-ionized water was injected to test the possible degradation and removal of biofilms. In the slurry injection test, upon the biofilm formation and ~50% permeability reduction following the control test, 1% bentonite slurry was injected. It was found that the injected bentonite particles were effectively attached or retained in biofilm matrices, resulting in slight but additional reductions in permeability. Upon the completion of slurry injection, de-ionized water was again injected to examine the durability of the clogging. This study showed that the biofilms is effective in retaining sub-micron fine solid particles in an aqueous phase, and this provides a valuable basis in biofilm utilization in many civil engineering practices.
One of the most economical and viable methods of soil improvement is dynamic compaction. However, dynamic compaction can only be applied on deposits where the degree of saturation is low and the permeability of the soil mass is high to allow for good drainage. The technique does not work very well on soils having a large content of fines. Also dynamic compaction produces lateral ground vibrations which can travel far from the construction site, which can discomfort the people living close to these areas. The current research aims to develop a new technology to desaturate saturated soils using the process of MIDP – Microbially Induced Desaturation and Precipitation through nitrate reduction, which extend the use of dynamic compaction as a ground improvement technique to saturated soil conditions and soils with higher fines content. To evaluate the feasibility of this technology an experimental program has been performed, in which soil columns have been treated with MIDP. In this process, indigenous micro-organisms are stimulated to convert nitrate into nitrogen gas, which desaturates the soil. After the process has finished, dynamic loading is applied to the soil columns. During the MIDP process and during the dynamic loading, the deformation and pore pressures area measured for various amounts of compactive effort and compared with the non-treated soil. Based on the experimental results the feasibility of the proposed technology is discussed.
Urbanization in coastal areas can be a significant source of pathogenic microorganisms, such as viruses and fecal indicator bacteria. Recently, coastal water contamination is becoming an important issue due to global warming [1]. Viruses can migrate long distances though porous media, such as beaches and coarse sediments, because of their biological characteristics, i.e., their size and motility. The attachment process of viruses onto soil grains can retard the virus transport, significantly. But, it can be reduced by the large salinity of the coastal water.
Here, a study that explores the effect of salinity on the instability of the virus front during transport in saturated porous media is presented. One dimensional (1D) transport model was developed following Cao et al. (2010) [2]. The model consists of two mass conservation equations of the virus and the salt concentration coupled through the constitutive equations of attachment/detachment mechanism. We observed that in the presence of hydrodynamic dispersion an instability develops at the virus front due to the formation of a mixing zone where the attachment is negligible. This instability develops in a pulse that travels at the speed of the average flow velocity. The magnitude of the pulse increases with the decreasing flow velocity.
The oil displacement effect of polymer flooding is mainly governed by the rheological property of polymer in the deep reservoir, while the mechanical degradation is one of the key factors affecting its rheological properties [1]. Polymer molecular chains can be mechanically degraded in shear flow and extensional flow, which both involves in the flow through a porous medium. Mechanical degradation includes shear degradation and stretch degradation. The questions are which one dominates, the shear or the stretch, how much it degrades, and how to simulate.
Maerker [2] conducted degradation of partially hydrolyzed polyacrylamide (HPAM) through consolidated sandstone plugs and reported that the degradation of HPAM through porous media was caused by large viscoelastic normal stress in elongational flow field and larger flow rate, longer flow distance and lower permeability induced more severe degradation. Dupas et al. [3] carried out HPAM degradation experiments through an API (American Petroleum Institute) capillary system, indicated that after polymer was degraded at velocity 8m/s, shear viscosity drops by 10%, while extension viscosity decreases up to 60%.
The objective of this work is to develop a device to simulate the mechanical degradation of polymer in micro pore throat, identify the shear effect and elongational effect, examine the viscosity loss of polymer solution through different pore throat model and present the relationship between viscosity loss and extensional rate.
Mechanical degradation is induced by forcing fresh polymer solution through a device which consists of a series of pore throat models with 4 kinds of throat length (15mm,35mm,70mm,100mm)and 3 kinds of diameter(100μm,300μm,500μm). Viscosity measurement of native and degraded polymer solution is made with Haake RS6000 rheometer at 25℃. The throat flow rate varies from 6.37 m/d to 3821.66 m/d. Polymer is HPAM with intrinsic viscosity 2510 dL/g. The concentration of polymer is 250mg/L, prepared in brine with salinity 32868 mg/L. Viscosity loss caused by extensional degradation can be obtained by calculating the viscosity loss of polymer solution degraded through pore throat model with throat length 0 mm. Accordingly, viscosity loss caused by the shear degradation through pore throat model with different throat length is calculated.
The results show that there is no obvious correlation between viscosity loss of HPAM in the pore throat model and throat length. More than 95% viscosity loss is caused by stretching degradation in pore throat, which is the main mechanism of mechanical degradation. The relationship between viscosity loss in pore throat and extensional rate (ε) has two characteristic values (ε0 and εL).ε0 is the critical extensional rate, εL is the ultimate extensional rate. When ε <ε0, the viscosity of polymer solution decreases slowly; whenε0 <ε < εL, the viscosity drops sharply; when ε>εL, viscosity loss keeps unchanged.ε0 and εL can be used as index parameters to evaluate anti-mechanical degradation ability of oil displacement polymer.
We describe recent advances in pore scale dynamics direct numerical simulation. The Volume of Fluid method associaed with well-balanced surface tension methods allows for the simulation of low capillary numbers. Further progress should involve thin film and corner flow formation, contact line dynamics and efficient combinations of parallelism and grid adaptation.
This ensemble of approaches is expected to particularly interesting for enhanced oil recovery modelling.
Among the known EOR techniques which have been investigated and tested during last decades the solvent injection proved to reach one of the best recovery factors. From the other side the solvent retention problem makes it relatively expensive and thus less attractive. So the successful practical application of solvent having only an active EOR agent has not been frequently reported.
A good combination for solvent became intriguing especially after the collapse of VAPEX technology tests which most probably resulted from disastrous choice of the application framework. In any case the recent solvent injection history has incorporated heating within the framework of either an additive to steam injection technique, hybrid thermal recovery with resistive (cable) and radio-frequency (RF) heating or finally, hot solvent injection. Undoubtedly the solvent story has not yet finished and simply is waiting for a good moment to get back.
In our current review we summarize, present and analyze the recent simulation results of different solvent-assisted applications for heavy (HO) and viscous oil (VO) recovery through the comparison of their performance and efficiency factors. First we revisited the steam-solvent co-injection analysis in order to resume the principal results in terms of recovery efficiency indicators selected for this study. In the same manner we analyzed the recovery dynamics for most known alkane solvents, this time in combination with resistive cable and radio-frequency antenna heating. In particular, different but always considerably improved thermal efficiency has been demonstrated for all types of heating including SAGD cases.
On the contrary the production dynamics didn't show the same tendency. More intricate subject was also the solvent recovery factor together with positive and/or sometimes negative synergy effect of combined oil displacement. Finally, these difficulties were partially overcome during hot solvent injection which turned out to be advantageous for certain reservoir conditions depending, for instance, on type of available solvent.
The discussion on the relative importance of various physical mechanisms associated to solvent-assisted oil recovery is provided together with practical observations on the computational model performance and efficiency.
Unconventional plays like the Bakken petroleum system (BPS) are the main reason behind the U.S. oil and gas industry renaissance during the last decade. In that period, more than 10,000 wells have been drilled in the Bakken alone, most of them targeting low or ultralow permeability strata. While the hydrocarbon in-place estimates are in the order of hundreds of billions of barrels, most recovery factor estimations range in the single digits. Thus enhanced oil recovery (EOR) has emerged as an area of interest attracting attention from governmental agencies, operators, and academic institutions.
Recently, the Energy & Environmental Research Center (EERC) conducted a comprehensive set of Bakken-centered EOR research activities, including lab experiments, reservoir characterization studies, modeling and simulation exercises, and a field test. Key findings indicate that tight oil formations such as the Bakken may be suitable targets for CO2 EOR opportunities. The use of CO2 serves a twofold purpose: CO2 is an excellent hydrocarbon extraction solvent, and CO2 sequestration contributes to mitigating greenhouse emissions.
This work presents a systematic modeling and simulation study that incorporates freshly acquired laboratory and field data sets from the Bakken Formation. The goal of these efforts was to better understand the implications of injecting CO2, CO2 storage efficiency, oil mobilization and sweep efficiency, and the potential for incremental oil recovery through various schemes.
Core plug-scale measurements were used to calibrate physicochemical parameters of the organic shale members of the BPS. Simulations replicating hydrocarbon extraction experiments using supercritical CO2 allowed the assessment of mass transfer mechanisms at work. The results indicated molecular diffusion and CO2 adsorption had significant effects on fluid flowing behavior in these tight rocks.
Operational observations were used to inform drill spacing unit models. Field data included petrophysical properties, reservoir pressure, temperature, fluid saturations, fluid composition, and primary production records. 3D heterogeneous models were built to investigate different injection strategies with two contiguous, hydraulically fractured, horizontal wells. Sensitivity studies were performed to quantify the effects of key parameters. Several scenarios were examined in detail, including varied well configurations (vertical or horizontal), well schedules, and targeted injection/production rates. Simulation results obtained with the geocellular models revealed natural fracture networks could result in more favorable CO2 storage and oil sweep efficiency in tight oil reservoirs. The natural fractures may significantly increase the contact area between the formation and the (artificially) stimulated region, leading to more favorable conditions for the recovery process. Consequently, reservoir characterization emerged as a critical element to understand the effectiveness of CO2 storage and enhanced recovery for tight oil formations.
This work improves the understanding of the physical and chemical mechanisms occurring in tight oil reservoirs undergoing CO2 EOR. The simulation models and results provide critical assistance for planning and optimizing both oil production and CO2 storage in future efforts.
A new technology is proposed, which consists of injecting H2, CO2 and bacteria into aquifers or depleted gas/oil reservoirs in order to convert them into methane. The conversion occurs by means of bacteria, which use hydrogen and carbon dioxide for their respiratory metabolism. The product of this bio-chemical activity is methane. Thus, we deal with the creation of the artificial reservoirs of natural gas. This also resolves two other fundamental environmental and energy problems: (i) reducing CO2 emissions into the atmosphere by converting CO2 to methane; and store excessively generated electricity from wind and sun in the form of hydrogen (this excess electricity can be converted to hydrogen).
The coupled bio-chemical and hydrodynamic model of the system is developed. The particular attention was payed to the adequate description of the bacterial kinetics and bio-chemical reactions. The mathematical analysis of this model revealed the existence of the phenomena of self-organization caused by the Hopf-Andronov bifurcation. This leads to the appearance of space oscillatory waves of concentration having multi-scale structure. Depending on the structure of these oscillations, they may be favourable or not for the efficiency of the methane production. Thus, the criteria of the efficiency of the underground methanation are strictly related to the mathematical criteria of the existence of self-organization regimes. Such exact criteria of the appearance of the auto-waves have been obtained. They depend on the injection rate, on the composition of the injected gas, and on the bacterial kinetics. Due to the analytical and numerical simulations, we have obtained the estimations for essential parameters of this technique, as the optimal composition of the injected gas, the characteristic time of conversion, and the evolution of the composition of the resulting gas in time.
Along with significant incentives to extract oil and gas from deep and unconventional resources, significant challenges exist in developing advanced technologies while being environmentally responsible. In this context, wellbore leakage and plugging is one of the key concerns that can affect both performance and environment. A similar leakage issue also relates to underground carbon storage sites. Inspired by the delicate architecture of natural materials, herein we present a bottom-up synthesis and design of a strong, tough, and self-healing composite using simple but universal spherical building blocks. Our product consists of inorganic porous nanoparticle loaded with desired sealant material which, upon exposure to heat or other stimuli, releases into the natural gas and fluid channels that are contained in the wellbore structures, effectively creating a barrier for escaping gas/liquid. The porous nanoparticles have unprecedented monodispersity on particle size, particle shape, and pore size, which facilitate their packing and performance including effective loading and unloading. This work paves the path towards fabricating a novel class of biomimetic composites using low-cost spherical building blocks, potentially impacting several areas including remedial and primary well cementing.
Extensive Study of cement additive Gilsonite and its applicability to plug and abandonment application in the gulf of Mexico
Deepwater horizon/Macondo event, which occurred on April 20, 2010, was one of the most catastrophic scenario in the United States deep waters. After that situation, regulations for drilling and completion projects became more rigorous and worst-case discharge calculations are mandatory as part of an Oil Spill Response Plan.
Drilling projects in deep water offshore is a huge investment for a company, and they are exposed to some risks. While drilling, an uncontrolled wellbore flow event could happen with a high discharge of liquid hydrocarbons into the environment. It is known as the worst-case discharge (WCD) scenario. Knowing in advance the total of potential volumes discharged will allow interceding from the surface and drill a relief well.
Some reservoir parameters might impact more than others during a blowout. The impact of multiple stacked reservoir/non-reservoir formation on oil spill volumes/rates is investigated. Preliminary results suggest that although oil spill occurs from the hydrocarbon reservoirs, the geologic sequence impacts the amount of oil spilled.
Setting a formation to formation cement plug in a well needing to be abandoned is a necessity, and the number of wells where this operation will have to be carried out increases each year in the North Sea. Paradoxically, where the mechanical and hydraulic integrity of this barrier matters most is also where these plugs are potentially weakest with today's placement methods. This is in the shallower parts of the well to be plugged, serving as the last barrier for hindering a leakage to sensitive layers near the sea bottom. These plugs tend to vary in cement strength and stiffness, both radially out and from top to bottom. This is due to the feeble confinement and restricted spacer fluid density use at shallow depths, combined with potential damage during scraping of top of plug for verification of set cement.
This talk will outline the benefits of adopting rapidly evolving technology from 3D printing of materials and in particular concrete in the building industry. The potential advantages of this technology are:
- Constructing the plug layer by layer with identical properties for each layer
- Remove the need for spacer fluid flushes, since only small volumes will be cemented in each layer
- Continually controlling the compressive stress at which each plug layer is setting
- Perfect plug dimensions and geometry can be pre-programmed
- Hydration time, conditions and temperature can be controlled continually at the local scale.
Cost reductions can be envisaged by certifying quality and reproducibility of such printed cement plugs, thus removing the need for lengthy plugs and reducing the redundancy in the number of required plugs.
Studying the two phase flow model is of a great importance. It occurs in many fields of engineering such as oil recovery, the storage of nuclear wastes and the remedy of the groundwater infected naps. The mathematical formulation of this problem includes nonlinear partial differential equations. Moreover, various numerical investigations have been the object of the approximation of these equations. Yet, in the case of anisotropic and heterogeneous media, most of the considered contributions fail to satisfy some relevant properties, such as the maximum principal, that the approximate solutions must fulfill. These properties are remarkable in order to ensure the efficiency of the method and to establish the convergence of the scheme. As a consequence, we propose a control volume finite element scheme for the discretisation of the two phase flow model involving these kind of media. This approximation turns out to be efficient and robust since it combines between the feature of finite volume methods, which is the local conservation of the fluxes, and that of finite element, which provides a simple discretisation of the gradient. In addition, we prove that discrete solution converges toward that of the continuous problem. Finally, as to sustain the obtained results, we give some numerical experiments that illustrate the behavior of the scheme in the presence of the anisotropy.
In areas contaminated by the petroleum industry, persistent compounds such as polyaromatic hydrocarbons (PAH) are often accumulated in the porous matrixes of sediments and soils (S&S), implicating risks to ecosystems and human health since these contaminants are released over time to interstitial and surrounding water. Pore size distributions (PSD) and PAH binding strengths to sorption sites on S&S are characteristics that affect such accumulation and release. Sorption, desorption and diffusion are among the critical processes that control the availability of PAH, and it is therefore crucial to evaluate these processes in order to understand and predict transport and fate of these contaminants in S&S, and for selecting effective remediation procedures.
Four S&S samples were obtained from previously contaminated sites, air-dried, and sieved (mesh 200), and organic matter (OM) was reduced in subsamples by hydrogen peroxide treatment, resulting in eight different porous media. Surface areas and PSD were determined (Autosorb IQ2MP, Florida), and OM were estimated by calcination (ASTM, 1993). Benzo(a)pyrene (BaP) was selected as study PAH and adsorption experiments were carried out in the dark with S&S suspended in NaCl (I=0.047 M) solutions, by adding five different concentrations of BaP with 7-14C-BaP as radioactive tracer, between 2.59 and 12.6x10-4 mmolBaP/gS&S, according to TG 106 Guideline (OECD, 2000). Reactors were fed with CO2-free air to keep suspensions oxygenated and allow carrying 14CO2 and stable CO2, produced by mineralization of BaP and OM, respectively, to alkaline traps, where CO2 production was measured by changes in electrical conductivity. Supernatant aliquots in the reactors and alkaline traps were obtained and 14C was measured in a liquid scintillation counter (Beckman Counter LS6500). At the end of the BaP adsorption experiments (14 d), 6.5 cm2 polyoxymethylene (POM) was added as a passive sampler, desorbing BaP from S&S. POM-accumulated BaP was extracted with acetone and sonication (EPA, 2007). 14C was quantified in the extracts as described above.
In the case of samples with complete organic matter, sorption at 14 days varies between 91.5 and 95.8%, while in the samples with reduced organic matter a variation between 78.7 and 89.9% was observed. Desorption velocities were much slower than adsorption rates, finding values between 2.0 and 6.0 % in 14 days and between 2.2 and 11.1% in 43 days. Using these values it is estimated that the time required to desorb BaP varies between 644 and 8,290 days. This difference between rates of adsorption and desorption should be considered when planning remediation actions for contaminated sites.
Two different time dependencies were observed in adsorption and desorption kinetics: a fast step considered to be due to diffusion of the BaP to the external surface, macro and mesopores, and a slow step, considered as diffusion of BaP into the micropores capillaries. Sorption equilibrium constant (KL), sorption sites and kinetic constants were experimentally obtained and a two-step conceptual model that describes the effect of the dynamics in porous media on the reactive transport of BaP in S&S with different PSD was developed.
The relative humidity at which a salt absorbs water vapor from the atmosphere spontaneously to form a saturated solution is called the deliquescence relative humidity (DRH) of the salt. In comparison to the DRH of bulk salt crystals, the DRH of levitated salt nanocrystals increases with decreasing crystal size[1,2]. Even in case of salts confined in pores smaller than 100 nm the DRH is no more equivalent to that of bulk crystals. It is strongly affected by the curvature of the liquid-vapor interface, when the pore size of the porous material gets smaller than 100 nm. To our knowledge, experimental investigations about the influence of confinement below 100 nm in diameter on the deliquescence humidity of a salt are not yet carried out.
The experimental work focuses on the water uptake behavior of different salt-silica-composites with pores smaller than 100 nm carried out to investigate the deliquescence behavior of confined salt. Porous materials like mesoporous SBA-15 and controlled porous glasses are impregnated with NaCl- and KNO3-solutions preventing over-filling of the available pore volumes. The experimental results of the water sorption measurements of bulk salts, pure host materials and the composites reveal a significant decrease of the deliquescence humidity (DRH) of the salt with decreasing pore size. A thermodynamic model approach, based on combined use of the Young-Laplace equation, the Kelvin equation and the Pitzer ion-interaction model, is used to predict the sorption behavior of salts confined in the respective pore sizes. The calculations match well with the experimental results that the main reason for the decrease of the relative vapor pressure over the salt solution in the unsaturated pore is the concave curvature of the liquid-vapor interface. Similarly, this model approach show, that the increase in DRH of levitated nano-sized salt crystals[1,2] is due to the convex curvature of the liquid-vapor interface, whereas the effect of increase in solubility with decreasing crystal size is low. Further experimental work focuses on the water uptake behavior of porous materials containing salts, which form higher hydrated phases before they deliquesce.
Crystallization pressure is the best-known player in the damage of natural and manufactured porous rocks at the Earth surface conditions. In the underground, the combination of pore pressure and confining pressure controls the stress arising in the solid matrix of porous reservoirs, which can lead to different types of damages. We turned our interest towards the role of capillary tension, possibly reaching the negative pressure domain, to deform and even fracture a rigid skeleton limiting the water-hosting pores. The main peculiarity is here to focus on rigid matrix and not only granular stacks with which the compaction effects of capillary bridging are largely referenced.
Isochoric cooling of liquids trapped in closed cavities put them under tensile strength up to the stability limits (e.g., Shmulovich et al., 2009; Qiu et al., 2016). In situ Raman spectroscopy can be used to record the frequency shift of two quartz peaks as a direct function of the growing negative pressure in water occluded inside the same quartz. The procedure uses a classic coupling between thermal cycles (microthermometry) and Raman micro-spectroscopy, of large use in fluid inclusions studies. We demonstrated how the solid rapidly react to in-pore water negative pressure, and may lead to local fracturing, with possible feedbacks towards permeability. Further modelling in terms of elastic solid mechanics, confirms the trends and enlarges the genericity of the measurements, giving also some insights into the role of different parameters (size of pore, values of tension, shape ration of the water-hosting cavity, etc.). One dataset gave contradictory values with respect to the models outputs, which demonstrate that certain parameters are still missing or wrongly parameterized.
These findings illustrate the potentially complex strain-stress relationships in a single pore submitted to tensile forces, like the ones developing during drying processes. The best example of such situation in nature is the aquifers in which supercritical (drying) CO2 is injected to mitigate atmospheric CO2 burden. We demonstrated that capillary tension arisen in crystalline or cemented materials can strongly modify their strength, and even may provoke their fracturing. The mechanical balance all around an injection well, and therefore the integrity and injectivity properties through time, should include full consideration of the capillary forces reigning there.
Experimental measurements of electrical resistivity of synthetic rocks with confining pressures up to 1700 bars and 200°C were performed in a high pressure cell to evaluate how the correlation between size pore distribution and brine composition is [1]. The porous media were manufactured with homogeneous glass beads and were saturated with different salt concentrations brines. According to our results the electrical resistivity decrease in synthetic rocks with high porosity [2]. In this talk we explain in what conditions this electrochemical effect produce a decay in the resistivity and how is related to porous media. Our results indicate that the fluid phase is the principal responsible of the electrical resistivity. To prove this result we test on dry synthetic porous media and we compare how the wettability modifications on the solid phase modify the behavior when the rock is saturated. The aim of this work is to describe the electrical resistivity under reservoir conditions in synthetic porous media saturated with brine [3}. Our results can be useful to predict the behavior in some real rocks.
Fine-grained sedimentary rocks, such as mudstones and shales, contain abundant nanometer- to micrometer-sized pores. These narrow pores create intense fluid-rock interaction that may lead to complicated fluid storage and transport process. Concerns about the accurate evaluation of gas content and diffusion kinetics have led to many experimental studies about gas sorption on shales. However, data on high-temperature high-pressure sorption isotherms of shales are still scare. In particular, the burial depth of the Paleozoic shales in the Upper Yangtze region of China is mostly in a range of 2000–4000 m, which indicates that the temperature and pressure of shale reservoirs are in the range of 60–120 °C and 20–40 MPa. Experimental techniques employed in obtaining sorption data have to be optimized and at the same time the measuring conditions have to be extended to in-situ conditions of deep shales while many published sorption data are limited to moderate pressures and temperatures.
In this work, high-temperature high-pressure sorption data for methane on shales from Sichuan Basin have been obtained at 30–120°C and pressures up to 25 MPa using a specially designed two-temperature-zone manometric setup. Dubinin-Polanyi potential theory was modified to extend to supercritical gas sorption over wide temperature and pressure ranges. A modified adsorption potential method was proposed to calculate the characteristic curves for supercritical gas sorption, and then a rigorous function from the supercritical Dubinin-Astakhov equation was also developed to describe the modified characteristic curve. Furthermore, the physical meaning of characteristic curve has been elucidated by comparing characteristic curves of different kinds of shales and clay minerals.
The measured excess sorption isotherms of shales follow the physisorption trend of decreasing amounts of methane adsorbed with increasing temperature. Characteristic curves of methane on shales at 30–120°C were calculated using a new expression of adsorption potential. It is found that if the thermal expansion of adsorbed phase is considered, these modified characteristic curves are temperature-invariant. The characteristic curve equation is capable to predict methane sorption at other temperatures based on the easily tested isotherm at room temperature. Using the sorption isotherm and characteristic curve equation at 30°C, the predicted isotherms at 120°C agree well with experimental data. The modified characteristic curves comprehensively characterize the available pore space for sorption and the affinity of methane molecules. The later stage of the modified characteristic curves (limited adsorption volume) is mainly controlled by the available pore space provided by organic matter and clay minerals. The limiting adsorption volume of shales in the gas window is larger than shales in the oil window with the same TOC content. The initial stage of the characteristic curves reflects the affinity of methane molecules for sorption on organic matter. According to the characteristic curves, shales in the gas window show higher affinity than shale in the oil window and clay minerals, though the clay minerals may provide comparable adsorption volume. The sorption characteristic energy shows a parabolic-like shape with a minimum approximately around Req =1.1%, which are related with the evolution of porosity of shales.
Phase change at the nanoscale is critical to many industrial applications including rapidly emerging unconventional oil and gas production from nanoporous shale reservoirs. The thermodynamic behaviour of hydrocarbons confined to these nanopores is expected to deviate significantly from bulk properties and there is little experimental data to validate theories. This research aims to visually observe the evaporation of hydrocarbons in a nanofluidic chip that accurately represents the geometric dimensions and the pressure/temperature conditions observed in shale. The chip consists of a nanoporous network of two-dimensional (2D) nano pores with dimensions down to 8 nm. Using an experimental procedure that mimics pressure drawdown during shale oil/gas production, our results show that evaporation of pure propane takes place at pressures lower than predictions from the Kelvin equation (maximum deviation of 11%). We probe evaporation dynamics as a function of superheat and find that vapor transport resistance dominates evaporation rate. For the transport resistance in the sub-10 nm nanoporous media, the contributions of the Knudsen flow and the viscous flow are found to be approximately equivalent. We also observe a phenomenon in sub-10 nm confinement wherein lower initial liquid saturation pressures trigger discontinuous evaporation resulting in faster evaporation rates. Additionally, we also extend this work to study evaporation and cavitation dynamics in nanofluidic devices with (a) mixture of pore sizes coupled with (b) mixture of hydrocarbons. Collectively, the results presented will aid in increasing the efficiency of shale production and will inform modelling and simulation of shale gas production.
Objectives:
CO2 injection, as one of the effective techniques for enhancing recovery of shale gas, has been widely used and proved economically available. In shale, clay minerals play an important role on methane adsorption due to its large volume of micropores. So far, however, a few attentions have been paid on competitive adsorption of CO2/CH4 Mixtures on clay minerals. In this study, we conduct molecular simulations of CO2/CH4 mixtures to provide a better understanding of competitive adsorption behaviors on clay minerals with the grand canonical Monte Carlo (GCMC) simulation.
Methods Procedures/Process:
We conduct GCMC simulations of CO2/CH4 mixtures adsorption in various clay minerals. Based on the actual conditions of shale gas reservoir, the competitive adsorptions of CO2/CH4 mixtures are investigated at various temperatures of 303.15K, 333.15K, and 363.15K with the pressure range of 0-35Mpa. For comprehensive comparison, the effects caused by pore size, mole fraction of CO2/CH4, and different clay minerals on competitive adsorption are processed. The competitive adsorption behaviors are characterized by selectivity and such key parameter are employed to evaluate the density profiles of CO2 and CH4, the characteristics of CO2 adsorption over CH4, and the timing of CO2 Injection.
Results/Observes/Conclusions:
Due to strong quadrupole moment and higher van der Waals interactions, CO2 possess a stronger affinity for clay minerals than that of methane, which is nonpolar. CH4 has the characteristic of single-layered adsorption, while the CO2 is able to form multi-layers adsorption with higher adsorption amount. Molecular simulations results show that the selectivity of CO2 in competitive adsorption decrease with decreasing of pressure. In addition, the selectivity of CO2 is independent on temperature. Because of negatively charged silicate layers, CO2 adsorption in nanopores of illite and montmorillonite are stronger than that of in kaolinite. As a result, the selectivity of CO2 in kaolinite models is less than that of in illite and montmorillonite models. When pressure is higher than 10Mpa, however, the selectivity of CO2 in three different clay minerals are similar. Due to the more adsorption sites occupied by CO2, the selectivity increases with mole fraction of CO2. By comparison with different pore size adsorption, the selectivity is insensitive to the mole fraction in micropores, while it increases with the increasing of mole fraction of CO2 in mesopores.
Novelty:
This work is a study on CO2/CH4 Mixtures adsorption behaviors on shale clay minerals. The molecular simulations with GCMC are proposed to give an insight in competitive adsorption mechanism, which is expected to provide a more accurate understanding of CO2 injection for enhancing recovery of shale gas.
Understanding how pressure fronts propagate (diffuse) in unconventional reservoirs is fundamental to transient flow analysis as well as reservoir drainage volume estimation. We have developed an alternative approach to the solution of the 3-D diffusivity equation by directly solving the propagation equation for the “pressure front” of the transient solution. The pressure front equation is an Eikonal equation, which is obtained from the high frequency asymptotic limit of the diffusivity equation in heterogeneous reservoirs and whose properties are well developed in the literature. Most importantly, the Eikonal equation can be solved very efficiently by a class of solutions called the Fast Marching Methods for a “diffusive time of flight” that governs the propagation of the pressure front in the reservoir. The diffusive time of flight can be used as a spatial coordinate to reduce the 3-D diffusivity equation into an equivalent 1-D formulation, leading to a simplified method for rapid reservoir modeling.
Based on this theory, we may further introduce a novel data-driven approach for production analysis of unconventional reservoirs without the traditional rate transient and pressure transient (RTA/PTA) assumptions of specific flow regimes. Our approach uses a transient generalization of the Matthews-Brons-Hazebroek method for the PSS drainage volume which relies on a w(τ) function to characterize the flow geometry from the transient drainage volume. Together with a calculated instantaneous recovery ratio, it has been successfully used to rank refracturing candidates and to obtain optimal fracture spacing. Given well pressure and flow rate data, we can calculate the transient well drainage volume with time. The time evolution of the drainage volume can be inverted to derive the w(τ) function which then provides a high resolution diagnostic plot that can be used for quantitative analysis to obtain fracture surface area, matrix properties, stimulated reservoir volume (SRV), and additional reservoir and fracture characteristics that are not apparent in the usual rate and pressure transient analysis techniques.
We have applied our methodology to field examples from the Montney and Eagleford shales. The comparison to standard RTA/PTA shows linear flow and fracture interference features more clearly than conventional RTA/PTA. It also provides detailed characterization of complex non-planar hydraulic fracture geometry, partial completion effects, the development and growth of the SRV, leading to the estimation of future decline rate and ultimate recovery.
The major advantage of the proposed approach is the data-driven model-free analysis of production data without the presumption of specific flow regimes. It provides a simple and intuitive understanding of the transient drainage volume and instantaneous recovery efficiency, irrespective of the complexity of the geometry of the reservoir depletion.
Chemicals in the form of nanoparticles or surfactants provide opportunities to improve oil displacement from rocks. They increase the rate of hydrocarbon recovery by breaking down the oil trapped in by-passed zones and separating the residual oil from rock surfaces in the form of tiny droplets suspended in the water phase. In this study, a series of heavy oil displacement experiments are conducted by flowing a series of aqueous solutions through an oil-wet and transparent network of microfluidic devices. Micromodels are fabricated by soft lithography techniques on a silicon wafer and replicated with Polydimethylsiloxane (PDMS) polymer. The effect of silica nanoparticles and three different types of surfactants (SDS, Tween 20, and Silwet) on the displacement of heavy oil, removal of oil films, and mobilization of trapped oil droplets are investigated. Furthermore, the patterns of residual oil and final oil recovery factors are explored. Also, the synergism effect between nanoparticles and different types of surfactants are reported.
3D Confocal microscopy coupled with fast speed fluorescent imaging of the displacement process reveals the effect of each chemical additive on oil mobilization. Silica particles show the tendency to reduce or remove the remaining oil film thickness while the surfactants break up the oil phase in the by-passed channels into tiny clusters that can be transported by the displacing fluid. The results demonstrate that the addition of the silica nanoparticles increases the rate of oil recovery up to 60% resulted from the wettability alteration and oil film removal. Moreover, the recovery factors increase upon adding the silica nanoparticles to a constant concentration of Tween 20 and SDS. The silicon-based surfactants improve the oil recovery up to 80 % where the recovery improvement by the addition of nanoparticles is negligible.
The developed microfluidic-based model is a powerful mimetic prototype of real porous media which can clarify the mechanisms underlying the process of chemical-based flooding for oil recovery. Considering the time-consuming and expensive nature of core-flood experiments, the proposed microfluidic approach provides an attractive alternative for rapid and low-cost enhanced oil recovery (EOR) screening studies.
Channel fracturing, as a novel technology, has reveived increasing attention in recent years due to its great advantage in promoting the fracture conducitivity as well as reducing consumtion of water and proppant. The open channels created by heterogeneous distribution of proppant are the priority path for oil or gas to pass instead of the pores exist in proppant pack. Since the flow pattern has been changed from Darcy flow to the flow mode dominated by Navier-stokes percolation by this technique, traditional method to calculate and predict its permeabity may not that accurate and appropriate. So this full paper presents a new model to measure the flow capacity of channel fracturing combined with a power law embedment model. The new approach is based on Navier-Stokes equation instead of Darcy-law to compute fracture permeability in channel fracturing. Besides, four influencing, namely, fracture height, propant pillar size, proppant area friction and embedment, are investigated to analyse their effect on permeability. The proposed analytical model is found to in good agreement with the experimental data, which verifies the precision and the feasibility of this model. Based on this model, this paper can provide a theoretical basis for channel fracturing design to evaluate the key factor governing the conductivity(permeability), which is helpful to providing a reference for proper pulse time and pumping rate optimization.
The storage and flow mechanisms in shales depend largely on their microstructure. We use two parameters to characterize microstructures, namely specific surface area (SSA) and pore-size distribution (PSD). We use N$_2$ adsorption at 77K to quantify SSA and PSD of nanopores. There are two limitations of the N$_2$ adsorption method due to (1) uncertainties in molecular area due to the quadrupole moment of N$_2$ molecules result in 20% uncertainty in calculated BET SSA, and (2) kinetic restriction of N$_2$ molecules prevent it to access narrow pores (< 0.7 nm). To circumvent these limitations, we also used other adsorptives, such as CO$_2$ and Ar, for the measurements.
We present results from adsorption measurements of CO$_2$ at 273 K and Ar at 77 K on shales and compare them to N$_2$ adsorption at 77 K. Adsorption measurements with CO$_2$ at 273 K allows for detailed characterization of ultramicropores (< 0.7 nm), which are inaccessible to N$_2$ molecules. Our results from CO$_2$ adsorption reveal significantly larger micropore SSA in comparison to N$_2$ probed SSA. Ar molecules do not have quadrupole moment and resolve the uncertainties of molecular area for BET calculation.
Liquid foams are a proposed solution to overcoming conformance issues while increasing apparent viscosity over gas/water only injections in petroleum reservoirs. However due to the diversity of flow behaviour foam flow in porous media providing a sufficiently accurate and low uncertainty model for industrial use can represent a significant challenge. Understanding the specifics of foam flow dynamics in terms of elements of the porous media microstructure could alleviate some of the task involved with describing foam displacements.
In our study we make use of a microfluidic experimental setup including a high speed camera (800 fps obtained in one configuration), and precise fluid injection rates and measured pressure drop over a 2D micromodels of varying structure through which we observe foam flow. A significant increase in viscosity is observed as the pressure drop rises until obtaining a smooth pressure profile interpreted as the steady state. To obtain quantified data regarding foam bubble population and transport, an image processing workflow was developed that involves binarizing the images through a thresholding method based around the foam films, then successively eliminating the objects representing the solid grains impeding flow. The final step of the workflow involves tracking individual bubbles through the image series, and averaging the results over multiple image series. The dynamic tracking of a high number of objects (approx. 5 000 000 velocity data points per model heatmap) allows us to establish of steady state velocity and flux fields within the medium. This data is furthermore combined with the bubble sizes. In this way the flow behaviour can be observed in terms of either flow velocity per bubble size or inversely.
Initial results of velocity and flux fields demonstrate that significant differences exist between flow of differing bubble sizes. Whether it be through self-segregation or bubble size adaptation, the flow of largest bubbles is exclusively found in a few paths that are seen to be on average wider and oriented in parallel to the flow. These paths are also more likely to carry higher velocities. On the contrary, the smallest bubbles populate transverse small pathways perpendicular to the flow direction. For a means of comparison, we perform 2D Lattice Boltzmann simulations of simple Stokes flow on a digitized template of the micromodel, and compare the obtained foam fluxes with the monophasic numerical analogue. When compared to the monophasic flow simulations, the amount of transverse flux in paths perpendicular to the flow direction is much smaller in all the foam examples.
Through a more quantitative analysis flow in terms of microstructural parameters we explore relationships (or lack thereof) between foam flow and local porous media properties such as pore/constriction sizes, coordination and immediate neighbourhood properties. Foam flow is only weakly correlated to pore scale properties as in a steady state situation the entirety of a flow path over the medium has to be taken into account as a single flowing object and flow preference needs to be evaluated in terms of the complete path properties.
Foam injection is believed to be a promising technique to enhance oil recovery. One of the key characteristics of the foam and its rheology in porous media is its texture which describes the spatial distribution and size of the gas bubbles. Yet, the description of the texture of a foam confined in a real porous medium is a challenging issue because conventional methods do not have adequate spatial and temporal resolution or are not adapted to opaque media. The foam texture has been extensively investigated in 1D and 2D micromodel experiments or from bubble size measurements outside the core. Until now, in-situ measurements of foam’s texture during foam flow in porous media has not been experimentally investigated.
Small Angle Neutron Scattering (SANS) is a powerful technique to probe the microstructure of Bulk foams (from the nanometer scale to the micrometer scale). We propose here to extend this technique to characterize the foam in a real 3D granular media by using contrast matching conditions. In this study, SANS acquisitions are made in a specific cell allowing pressure drop measurements and control of flow rate injections. The foam made of Sodium Dodecyl Sulfate (SDS) and Nitrogen is generated in-situ by co-injection of gas and the surfactant solution. The porous media are made of fused silica grains, prepared and sieved according to specific targeted grain distributions. The geometrical characteristics of the pore network are extracted by image analysis from X-ray micro-tomography and compared to the calculated bubble size. Foam texture is measured as a function of foam quality, interstitial velocity and grain size distribution.
In parallel, we run experiments on a Bentheimer core using a CT X-ray scanner in which pressure drop measurements and saturations are measured for different foam qualities and interstitial velocities. We used a methodology to trace back to the in-situ texture. The two experimental approaches are compared in terms of foam generation, flow and in-situ texture.
Foam can improve sweep efficiency in gas-injection enhanced oil recovery. Surfactant-alternating-gas (SAG) is a favored method of injection, in part because of excellent injectivity during gas injection. However, liquid injectivity is usually very poor in a SAG process, and fracturing of the well can occur. We report a coreflood study of liquid mobility under conditions like those near an injection well in SAG application in the field: i.e., after a prolonged period of gas injection following foam.
We inject foam, gas (nitrogen) and surfactant solution into a 17-cm-long Berea core at elevated temperature (90 ℃) with 40 bar back-pressure to minimize gas-expansion effects. Pressure differences are measured separately across five sections of the core and supplemented with CT scans to relate water saturation to mobilities. From these data we estimate the velocities, saturation changes, and mobilities of the various banks that determine liquid injectivity. We examine liquid injection directly following a period of foam injection, as in previous studies, and then following prolonged periods of gas injection following foam, to reflect injectivity near the well in a SAG process.
Liquid injectivity directly following foam is very poor, as shown in previous studies. Liquid first fingers through the trapped foam. It then dissolves gas trapped within the liquid fingers, and the overall mobility rises sharply. During prolonged gas injection following foam, however, a region forms near the inlet and slowly propagates downstream in which gas mobility is much greater. The abrupt rise in gas mobility appears to reflect the decline in water saturation below about 0.2 in our experiments. This decline in liquid saturation reflects in part liquid evaporation, and also pressure-driven flow and capillary effects on the core scale. In the region of lower liquid saturation, subsequent liquid mobility is much greater than downstream, and liquid sweeps the entire core cross-section rather than a single finger. Mobility in the region of liquid fingering is insensitive to the quality of foam injected before gas and the duration of the period of gas injection. These results suggest that there is a small region very near the well, crucial to overall injectivity, in which liquid mobility is much greater than that further from the well. These conditions are not described by current foam models. The results can inform a model for liquid injectivity based on radial propagation of the various banks seen in the experiments.
Steam injection has been implemented with residual oil saturation in swept zones being as low as 10% in the case of Steam Assisted Gravity Drainage (SAGD) processes. The primary difficulty with a gas injection of any kind is conformance due to low density and viscosity of the gas. Therefore, gases tend to form channels and pathways through oil rather than displacing the oil. Traditionally foam-forming additives such as surfactants have been added to reduce gas mobility. However, the surfactants ability to perform under harsh conditions (e.g. temperature and salinity) are limited. Nanoparticles can be a promising enhancer/stabilizer for foam additives at reservoir condition. This is promising for EOR processes with steam (i.e. SAGD) since the increased mobility control and the transport stability, can improve the conformance control and consequently sweep efficiency.
In this study, the foam height tests (Nitrogen as the gas phase) were performed with the surfactants, nanoparticles, and combinations of the two with the help of high-pressure high-temperature (HPHT) visual cell. Foams generated with surfactant were treated as a baseline and effect of nanoparticle addition on foam behavior was studied. Dynamic light scattering was used to characterize all nanofluid solutions before and after foam height tests to ensure nanoparticle-surfactant compatibility and stability in solution at HPHT condition. After screening the various foaming agents, the dynamic flow experiments were performed with a specially configured core flood apparatus designed for steam foam flooding. Propagation of the nitrogen foam was measured using four differential pressure taps located along the length of the porous media (100 ºC and 500 psig). In addition, steam foam flow experiments were also performed in the presence of nitrogen as a non-condensable gas (500 psig and 250 ºC). Effluent samples were analyzed for nanoparticle and surfactant effective concentrations using high accuracy inductively coupled plasma triple quadrupole mass spectrometry that yielded retention and adsorption.
In initial foam height tests (500 psig and 100 ºC) it was observed that the combination of a studied coated silica nanoparticle and sodium dodecylbenzenesulfonate surfactant produced foam with a better texture, and a longer retention of foam height compared to foam generated with the only surfactant. In addition, the foam produced with a proprietary nanoparticle alone outlasted the surfactant stabilized foam at HPHT condition. The increase in foam half-life was also correlated with higher mobility reduction and faster foam propagation in nitrogen and steam foam flow experiments. Effluent composition profiles and tracer composition profiles developed by ICP-QQQ analysis yielded mass balance data in addition to providing detailed information about the throughput of injection fluid. Dynamic flow experiments revealed the synergy between nanoparticle and surfactant in steam foam application (i.e. mobility control) compare to that of the steam-surfactant system.
The results obtained through this study yielded data regarding multiphase transport of nanostructured foams through porous media at high temperatures, such as propagation through porous media, nanoparticle stability in harsh conditions, retention and adsorption of surfactant, nanoparticles, and their combinations. The knowledge obtained from this study significantly improves our understanding of nanoparticles as an additive in steam foam application.
The foam efficiency in the oil displacement processes is governed by foam stability which is generally reduced with the presence of oil. In this study, we investigate the effect of oil type and saturation on foam strength in Berea sandstone cores using coreflooding and nuclear magnetic resonance (NMR) imaging. Foam quality scan in the presence of remaining hexadecane showed higher apparent viscosities compared to the oil-free case except at very high foam qualities. We analyzed the foam-induced oil displacements mechanisms from the saturation profiles, measured by NMR, and determined the relative significance between the increased capillary number and the micellar solubilization. Furthermore, we carried out foam-oil co-injection tests with hexadecane, octane, and reservoir crude oil to correlate foam apparent viscosity with the oil saturation. Under our experimental condition, it was determined that with the increase in oil saturation, foam apparent viscosity first decreases because of foam weakening with oil; and then increases due to oil emulsification with the surfactant solution. The observed trend was similar for octane and hexadecane. However, octane resulted in higher foam destabilization and lower emulsification.
Mechanisms of pore gas transport and exchange across the porous medium-atmosphere interface in wind-exposed porous media was investigated for a range of porous media under different near-surface wind conditions.
Four dry porous media with mean particle diameters of 1.6, 2.0, 4.2 and 10.5 mm were used. These relatively coarse materials were selected, to facilitate easier identification of the parameters governing wind-induced porous medium gas transport and exchange with the atmosphere. Cylindrical porous medium samples 25 cm in diameter and thickness ranging between 15 and 35 cm were used.
Experiments were carried out under controlled conditions for 10 different wind conditions with respect to wind direction, wind speed magnitude and wind speed variability yielding 40 combinations of porous medium and wind condition. Average resultant wind speed as measured 4-12 cm above the porous medium surface ranged between 0.5 and 3 m/s.
Tracer (air and CO2) gas breakthrough curves were measured at five depths within each sample. A total of 400 individual tracer gas breakthrough curves (including replicates) were produced.
Porous medium gas transport and exchange with the atmosphere was approximated as a one-dimensional dispersive/diffusive process with a depth-dependent dispersion coefficient (Dw). Two models for describing the depth – Dw relationship using two and three empirical fitting parameters, respectively, were evaluated. For each combination of porous medium and wind condition, fitting parameters were determined by numerically solving the dispersion equation (with Dw described by either of the two models), while fitting the solution to all five tracer breakthrough curves simultaneously. In all cases both models yielded close fits to measured breakthrough data.
Results showed that in addition to porous medium depth (or distance to the wind-exposed surface), Dw (and thus, gas transport and exchange) was strongly dependent on both mean wind speed, and wind speed power spectrum characteristics, indicating that not only mean wind speed, but also wind gustiness is important. Results further showed that porous medium particle size and to a lesser degree also hydraulic conductivity and air permeability correlated significantly with gas transport and exchange (Dw).
We present recent results relevant for an application of level set methods to track moving interfaces and free boundaries. We cover the most important tasks typically required in level set methods like an advection of level set function, a preservation of signed distance property, and an extrapolation of missing data in the normal direction to the interface.
Typically a movement of the interface is driven by some additional processes that require a coupled modelling on evolving domains. The level set methods result in implicitly defined computational domains when some interface conditions must be approximated without an explicit reconstruction of the interface. For that purpose the immersed interface methods are developed for convenient unfitted grids. The methods are well covered for elliptic and parabolic problems, but not so much for a hyperbolic type of problems. The reason is so called cut cell problem due to a CFL time step restriction for numerical schemes with an explicit in time discretization that are typical for hyperbolic problems. We are interested in semi-implicit methods that are unconditionally stable with respect to the choice of time steps and that have no CFL restriction for cut cells.
A brief illustration of the level set tools will be given for a flow in porous media with an evolving part of a boundary (a moving groundwater table).
Accurate numerical simulations for density-dependent flow and transport model is one of the crucial keys for successful water resources management in coastal areas and on islands. However, traditional modeling approaches without special treatment may not be able to resolve accurate sharp moving fronts and corresponding groundwater flow velocities due to the numerical instabilities.
In this presentation, we employ the enriched Galerkin finite element methods (EG), which enriches a classical continuous Galerkin finite element methods with piecewise constant functions to ensure local and global mass conservation. EG has the same bilinear forms as the discontinuous Galerkin (DG) finite element methods but EG has fewer degrees of freedom in comparison with DG. Moreover, dynamic mesh adaptivity approaches are employed to save computational cost for realistic large-scale problems and an efficient Krylov solver with preconditioner is provided. We will present initial numerical results for existing benchmark problems to show efficiency and effectiveness of the proposed method in density-driven flow modeling.
Kamal S. Bisht, Michael E. Dreyer, University of Bremen, Faculty of Production Engineering, Department of Fluid Mechanics, ZARM, University of Bremen, Germany
In propellant management devices (PMD), porous media are widely used for the phase separation process and delivery of vapor free propellant. However, instability of the free surface flows in open capillary channels limits the flow rate capacity and affects the effectiveness of the porous media. In low-gravity conditions, capillary pressure only balances the pressure difference across the liquid-gas interface. At the critical flow rate, the maximum capillary pressure is exceeded; the free surface collapses and gas bubbles are ingested into the liquid. The presence of gas degrades the quality of the propellant and severely affects the engine efficiency. In this project, a setup with a metallic porous screen covering a rectangular groove channel is investigated to obtain a higher flow rate in capillary channels without collapse of the free surface. The saturated porous screen permits liquid to pass through but acts as a barrier to the gas breakthrough until the differential pressure across the screen exceeds to the bubble point pressure.
This feature is governed by porous media properties such as permeability, porosity, wettability and pore diameter. A theoretical study has been done and is currently under investigation using numerical tools and ground tests. The computations are performed with Matlab and the computational fluid dynamics program Ansys Fluent. The setup for the experimental test facility is defined which shall be tested next year in summer during drop tower test. The theoretical, experimental and numerical results will be presented in the talk.
Wicking is the capillary phenomena by which a preferentially wetting liquid is drawn into a porous medium due to surface tension forces. The Washburn equation, governing the spontaneous invasion of a wetting fluid in a capillary was developed in the early 1900’s. The equation can also explain a wide variety of impregnation processes in homogeneous porous media effectively. However, there are cases where the porous media is heterogeneous, like in the geological layers of oil reservoirs, where applicability of the Washburn equation is limited at large scale. Most of the heterogeneities in geological layers is due to various factors such as, wettability, geometry of the pores, tortuosity of the medium etc., In recent studies, it is observed that the wicking phenomena is significantly affected by the grain sizes in a hydrodynamically interacting layers of porous medium. Experiments in heterogeneous bi-layered porous medium show that the wetting fluid imbibes faster in the fine pores and slower in the larger pores. However, in three layered porous medium, wicking is dependent on permeability ratio and capillary pressure ratio of the layers. In addition, the arrangement of the layers with respect to each other also impacts the wicking rate. It is still unknown that how the viscosity of the resident fluid impacts the flow behavior in the layered porous medium.
In the present work we explore the wicking behaviour in the layered porous medium, when a viscous non-wetting fluid is already present in the porous medium. We explain the wicking phenomena by using a quasi one-dimensional lubrication approximation model including the insights from the simulations in analogous system of pores. Further, the numerical model explaining the wicking phenomena is supported by the anticipated pressure gradients in the porous medium followed by experimental observations of the same.
In developing the numerical model for spontaneous imbibition, we consider the possible cases of wicking in a two layered porous medium, i.e., (i) wicking is faster in the fine pores, (ii) wicking is faster in the large pores. For each of the possible cases, we develop the numerical model and verify the same with the anticipated pressure gradient graphs. Our analysis of the model shows that the case (ii) is never possible. Experimentally, we show a similar wicking pattern as predicted by our quasi 1-D, Wasburn like model for a layered porous media. Wicking in fine pores is observed to occur at a faster rate than the wicking in large pores as predicted by the one-dimensional model. We compare the results of the experiments with the numerical model and find that the proposed one-dimensional model sufficiently explains the spontaneous imbibition in a two layered porous medium. We have also extended our model for three layered porous medium, where the arrangement of the layers alo impact the wicking behavior.
This work provides insights into the physics of wicking in layered porous media and will be applicable to fields like oil recovery from fractured reservoirs, ground water remediation and paper wicking.
Unsaturated flow through thin layers of porous media are encountered in many industrial applications, including the liquid-absorbing hygiene products such as wipes, paper towels, and diapers [1-4]. These consumer products demand specific absorbent properties with storage of liquid playing a significant role. Understanding fluid flow and deformation processes in thin swelling porous media is critical for the development and design of these products.
We use an averaging approach [5-17] for modeling a system consisting of multiple layers of thin, absorbing swelling porous media as the layer-wise 2D interacting continua, to rigorously derive a 2D averaged macroscopic mass-balance model for each layer and to develop the required constitutive relationships for a system of thin porous layers made of one liquid (water) and two deformable solid phases (fiber and hydrogel). The developed model consists of a set of partial differential equations that keep track the time dependent behavior of variables such as piezometric head, saturation, porosity, and layer thickness, as the liquid moves throughout the multi-layered porous medium. Hence, this model can be used to describe the absorbency process [18], to predict and understand the flow and storage of a liquid in conjunction with the deformation of layers in multilayered thin porous media that is absorbing the liquid, and swelling during deformation. This model will enormously improve the computational speeds, allowing one to develop a fast and reasonably accurate simulation of the unsaturated flow.
The numerical simulations are carried out with the flow parameters and geometries for a few representative cases such as wicking into dry horizontal and inclined porous plates. The simulation predictions, which predict detailed 2-D flow fields, are found to be in good agreement with the experimental and 3D computational results.
This paper investigates through reservoir simulations the set-up of a foam EOR process in a real sector of a low permeability reservoir produced by waterflooding. The objectives are: to demonstrate our capability to simulate foam injection at field scale; to investigate the technical feasibility of foam injection in terms of injectivity, foam stability and improved oil recovery; and to analyze its economic feasibility as a function of technical uncertainties and market prices variations.
The study is conducted using a reservoir simulator combined with uncertainty and optimization capabilities. Our foam model modifies the gas relative permeability according to a multi-parameter mobility reduction function. Several sensitivity analyses and optimizations were performed to design the injection sequence that optimizes the foam displacement and the resulting sweep efficiency, and to determine the influence on the oil recovery and the net present value of the foam quality, its maximum gas mobility reduction, its stability and the surfactant adsorption. By applying Monte-Carlo methods on surface responses, the process robustness against surfactant adsorption and hydrocarbon prices uncertainties was assessed.
The optimization of foam injection at pilot scale requires a specific study because that process is driven by many parameters with opposite effects. Specifically we found that foam viscosity may hinder injectivity, potentially leading to a decrease in the oil production, while foam may greatly enhance displacement profiles at the same time. Moreover numerical simulations show that early gas breakthrough can occur even for highly viscous foams because the apparent foam viscosity decreases in the near wellbore due to higher velocities. This can be mitigated by performing a 3-4 months surfactant water injection pre-flush. It is then possible to find an optimal foam injection sequence that generates a high financial profit for realistic foam parameters. The foam quality turns out to be a key parameter to control the foam EOR process, with quite different values found to maximize either oil production or net present value. For the case under consideration that involves a surfactant adsorption of 250 µg/g and an optimized 5-year injection of a foam of quality 0.45 following a surfactant aqueous pre-flush, we show that the foam process is profitable and robust with respect to market prices variations and technical uncertainties.
This innovative optimization methodology shows that foam processes could be eligible for this pilot. After designing properly the injection sequence, including a surfactant pre-flush prior to foam injection, reservoir simulations predict that this EOR process may highly improve the exploitation profitability. Moreover profitability is robust to economic uncertainties, with a breakeven price of 30 $ per barrel, as well as technical uncertainties, of which adsorption up to 600 µg/g.
Turbulence in porous media is a phenomena that is more prevalent than commonly thought. Such flows happen in packed bed reactors (e.g., oxidation of petrochemicals), in pebble-bed nuclear reactors (where Reynolds numbers can be on the order of 100,000), and at the fluid-solid interface of rivers to name a few examples. One of the characteristics of turbulent flows is the notable deviation from Darcy's law; the resulting flows are nonlinear, and are described by more-or-less empirical formulations referred to here as the Darcy-Forchheimer-Ergun equation.
In this work, we take a unique approach to the problem by combining conventional volume averaging theory with direct numerical simulation (DNS). Direct numerical simulation seeks to eliminate the closure models conventionally used in turbulence theory with simulations of all relevant time and length scales all the way down to the scale of viscous dissipation. DNS simulations are computationally intensive, requiring potentially millions of degrees of freedom to simulate even simple systems containing a small number of grains. The problem of resolution becomes significantly more difficult as the Reynolds number increases.
For the work we report, we provide (1) a revised examination of the upscaling problem as originally examined by Whitaker (1996), (2) development of a novel closure scheme for relating the microscale velocity deviations to the effective macroscale momentum equation, and (3) the results of a sequence of high-Reynolds-number flows, with Reynolds numbers up to 1000. Our DNS results validate our proposed closure scheme; comparison of the upscaled and DNS momentum balances show that the predicted macroscale flow parameters are able to reproduce the integrated results of the DNS simulations with high fidelity.
Multiphase flows in porous media play an important role in many natural and industrial processes, such as transport mechanisms in the vadose zone, CO2 sequestration in saline aquifers or oil recovery in petroleum applications. The traditional picture for such flows is one at low Reynolds number where the distribution and flow of the different phases is controlled by interfacial energies of fluid/fluid and fluid/solid interfaces, with a major influence of wettability and capillarity \cite{Muskat1946,Whitaker1986}. While this is accurate for creeping flows in relatively low permeability porous media, highly permeable porous structures --such as those found in trickle bed reactors, fuel bundles in nuclear cores or distillation columns used in chemical engineering applications-- challenge the relevance of this represention. In those, the relative importance of interfacial energies may be reduced, with much larger inertial effects and exchanges of momentum between fluid phases.
Here, we discuss whether the continuum models used for flows in low permeability porous media, such as generalized Darcy's laws, are adequate for highly permeable porous structures (high Re, Bo and We numbers). We first propose an alternative representation for mass and momentum transport of two-phase immiscible flow at the continuum-scale, which is based on a multiscale analysis starting from the flow problem at the pore-scale. Compared to the generalized Darcy's law, our representation contains additional drag and cross terms that account for inertial effects and exchanges between fluid phases. We then go on to determine constitutive relationships for the effective parameters using experimental data on co- and counter-current flows from recent water/air experimental data \cite{ClavierChikhiFichotEtAl2017,Wang_thesis}. Results show that the macroscale model allows us to capture important physical aspects of the flow that the generalized Darcy's law fails to describe. We further find that the impact of the cross and inertial terms increases with the Reynolds numbers of the phases.
Shear-thinning fluids flow in the rough fracture is encountered in numerous industrial applications such as hydraulic fracturing fluids flow in rough hydraulic and natural fractures to carry the proppants, polymer gel extrusion through rough fracture to reduce excessive water production in naturally fractured reservoirs, CO2 sequestration and leakage through rough fractures, etc. We investigate both the macroscopic flow behavior and microscale flow details of shear-thinning fluids through a realistic rough fracture for both linear and inertial flow regimes.
For the first time, a fully-3D flow simulation of Cross power-law shear-thinning fluids through a rough fracture is conducted. The flow domain is extracted and refined from a computed microtomography image of a fractured Berea sandstone. The modified Navier-Stokes equation incorporating Cross power-law fluid rheology model is solved. The critical Reynolds number above which the linear Darcy’s law is no longer applicable is evaluated for both Newtonian and shear-thinning fluids.
First, the Newtonian fluid flow simulations are conducted, and the hydraulic/equivalent aperture and inertial coefficient in Forchheimer’s equation are fitted. Second, based on the simulation results of shear-thinning fluid flow, both the shift factor and inertial coefficient are obtained. Specifically, the shift factor is a critical parameter in the definition of “in-situ shear rate” which could be used to evaluate the “equivalent” viscosity based on the “bulk” viscosity. Our results show that the shift factor is dependent on both the fracture geometry and fluid rheology properties while both the inertial coefficient and critical Reynolds number are only functions of fracture geometry, which is consistent with the recent experimental results. Third, to explain the above phenomena, a detailed analysis of microscopic flow patterns is conducted.
Based on the analysis of a large number of simulations, we propose a correlation for shift factor which is quantified by the product of one fluid property parameter and tortuosity. The fluid property parameter is obtained from the analytical/semi-analytical solutions of the same shear-thinning fluids flow in a smooth slit. Two approaches are provided for tortuosity quantification. One is based on the detailed pore-scale velocity field and produces a very accurate shift factor. The other is the geometric tortuosity obtained by image analysis without doing any pore-scale simulations, which provides an approximate value for shift factor. The Forchheimer’s equation with our newly improved shift factor correlation can be used in the higher-level macro-scale simulators to direct the relevant industrial applications.
Below 2.17 K, helium no longer behaves as a classical fluid: it has almost no viscosity and a high effective thermal conductivity that is used to cool superconducting devices such as the Large Hadron Collider’s magnets. Beyond a critical velocity, quantum turbulence arises and complex flow patterns appear.
We developed a numerical tool to simulate helium superfluid flow in porous media at the pore-scale. We were able to reproduce and explain, for the first time, interesting and unexpected experimental observations about thermal counterflow of He II past a cylinder reported ten years ago. Eddies are generated through a complex transient process that involves the friction of the normal fluid species with the solid walls and the mutual friction between the superfluid and normal species. The vortices remain in single pores and eddies spanning over several pores are not observed suggesting that a Darcy-Forchheimer type law can be used to model quantum turbulence in porous media.
Hard pellet model and capillary model as two types of physical models for porous media, can’t accurately describe the microscopic pore structure features of porous media. Shale rock porosity is dominated by nanometer and micron scales [1], the traditional models and methods can not accurately describe the pore structure of shale reservoirs. Numerous studies have shown that real porous media in nature have fractal features over a range of scales [2]. Based on fractal theory, this paper provides new ideas and methods for the characterization of pore structure in shale reservoirs. SEM scanning electron microscopy, nanometer CT and focused ion beam (FIB) and other techniques used to describe the micro-pore structure of shale reservoirs [3], Based on the imaging technology, remarkable achievements have been made in the analysis of shale pore formation, pore type classification and organic matter nanometer pore characterization. The fractal dimension can quantitatively describe the complexity of shale pore structure [4-5].
Fractal theory can be applied to predict the permeability of porous media. Domestic and foreign scholars have made some progress in predicting the permeability of porous media. Domestic and foreign scholars have made some progress in predicting the permeability of porous media. Based on the fractal theory, fractal characteristics of shale reservoir pores were studied by image analysis, and the expression of shale permeability was derived. The influence of maximum pore radius, minimum pore radius and fractal dimension on shale permeability was analyzed [6]. In this study, the pore of shale reservoir was taken as the research object. Based on the shale reservoir skeleton, the Sierpinski carpet model was used to study the distribution of shale porosity. Through the statistics and analysis of the pores in shale raw binary image, the relationship between the number of pores larger than the pore size and the pore size in the double logarithmic coordinate system. The negative of the slope of the stable straight line in the data is the fractal dimension of the porosity. Fractal dimension calculation of pore structure should be based on the distribution of data points and pore size range of regression, which can not only ensure data integrity, but also a true reflection of pore distribution.
Shale permeability is a parameter that characterizes shale seepage capacity, and its size is related to porosity, geometry of the pores in the liquid infiltration direction, particle size and arrangement. Because the fractal dimension of porosity is a parameter that reflects the change of the number of pores under different pore sizes, the fractal dimension of porosity is negatively correlated with the porosity. Therefore, the larger the fractal dimension of porosity and the smaller the porosity, the more the shale permeability small. The permeability of shale decreases with the increasing fractal dimension of the pores because when the fractal dimension of tortuosity increases, the pore channels become more tortuous and the permeability decreases. The maximum pore radius, inflection point radius, fractal radius fractal dimension and tortuosity fractal dimension all have a great influence on the permeability of shale. Shale permeability has a significant positive correlation with the maximum pore radius and inflection point radius, and the pore radius Fractal dimensions and tortuosity fractal dimension showed a significant negative correlation, and pore radius fractal dimension of the correlation is small.
Abstract: This paper further studied the pore throat model, by fractal equation with Forchheimer equation, obtained the analytical expressions of Darcy permeability, non-Darcy equivalent permeability, and the ratio of non-Darcy equivalent permeability and Darcy permeability. Darcy permeability is the function of porosity, average particle size, fractal dimension, is not the function of Reynolds number, the conclusion is consistent with the permeability definition that permeability is related to medium skeleton and has nothing to do with the fluid flow in the medium. Non-Darcy equivalent permeability is a function of the Reynolds number and decrease with the increase of Reynolds number; it is the same with dimensionless results of Barree and Conway's. Using dimensionless form for non-Darcy equivalent permeability, get the ratio of non-Darcy equivalent permeability and Darcy permeability.
Keywords:Darcy and non-Darcy flow, fractal equation, Forchheimer equation, Reynolds number
The petrophysical properties of rocks, such as thermal conductivity, electrical conductivity, and fluid transport, have been studied based on fractal geometry theory in many areas over several decades and determination of fractal dimensions is always the focus of researches and applications by means of fractal-based methods. In this work, a new method for calculating pore space fractal dimension and tortuosity fractal dimension of porous media is derived based on fractal tortuous capillary model assumption. The presented work establishes relationship between fractal dimensions and pore size distribution, which can be directly used to calculate the fractal dimensions. The published pore size distribution data for 8 sandstone samples are used to calculate the fractal dimensions and simultaneously compared with prediction results from analytical expression. The proposed fractal dimension method is also tested through Micro-CT images of three sandstone cores, and are compared with fractal dimensions by other algorithm. There is a self-similar fractal range in sandstone when excluding smaller pores and the fractal dimensions are intimate related to the microstructures of porous media.
The size of the fractures and vugs ranges from micron scale to centimeter scale in frac-vuggy reservoir. And there is almost no flow in the rock matrix. Due to the multiscale of media, inertial coefficient is a key parameter to predict the correct production performances and behavior of frac-vuggy reservoirs. This paper introduced the process of making multiscale frac-vuggy media and will study the inertial coefficient of Forchheimer equation and its effect on oil-water two-phase flow in the media.
The experimental results of flow law showed that if flow rate is constant, the existence of non-linear flows for single water phase is determined by the fracture width and filling degree. And the effect of the vug can be ignored. However, for oil-water two phase flow, the fracture and vug both play an important role. Meanwhile, based on Rescaled Range Analysis(R/S), a mathematical model of judging non-linear flow is proposed. The Receiver Operating Characteristic (ROC) curve showed that it can accurately determine the flow law for oil-water two phase flow.
Through the analysis of the experimental data of non-linear flow, this paper proposed a modified Geertsma's empirical expression of inertial coefficient, which is a function of wetting phase saturation, fracture width, vug diameter, fracture porosity, vug porosity and total permeability. It’s more suitable for multiscale frac-vuggy media than previous literatures reports.
The experimental results of oil-water relative flow capacity showed that non-linear flow seriously affected water and oil relative permeability curves. When the flow law transforms from linear to non-linear, the irreducible water saturation will increase, the range of water saturation where oil-water two-phase can flow together will decrease and the same relative permeability point will decrease. When the inertia increases, it will be more serious.
Characterizing dissolved chemical migration in porous media through the Advection Dispersion Equation requires the knowledge of the fluid velocity field and of dispersivity values associated with diverse geomaterials which can make up the internal architecture of the system. Several studies have focused on the assessment of the impact on solute concentration dynamics of an incomplete knowledge of the fluid velocity field, the latter being typically due to uncertainty of hydraulic properties of the hosting media (e.g., permeability). Limited attention has been devoted to analyze the way uncertainty about spatial distribution of dispersivity values can propagate to uncertainty of solute concentration fields. Here, we address this issue by focusing on a simple one-dimensional domain filled with two distinct porous media and subject to a pulse injection of a tracer. We derive and solve numerically the equation governing the expected value and associated variance of solute concentration by considering uncertain dispersivity values and conceptualizing the domain as a random composite medium (where the location of the interface between the two materials can be uncertain). The solutions of such moment equations compare well against corresponding moments evaluated through a numerical Monte Carlo analysis. Our results suggest that in the investigated set-up (i) solute concentration variance exhibits a three peaks behavior, even in the presence of conditioning on a given location of the interface between the two materials and (iii) the actual sequence of the materials traveled by the solute impacts spatial distributions of expected value and variance of concentrations.
Urbanization in coastal areas has been on an increasing trend during the last century. In some coastal regions, groundwater is one of the major source of potable water for the population, the industry, and the agriculture with an average demand of 30 m3/s [1,2]. Sea-level rise has been recorded to be approximately 40 mm/yr [3] with the potential consequence to favor significant intrusion of seawater into potable coastal aquifers [4] and groundwater flooding of urban areas [5] and infrastructure. Despite some indicators of the interaction between infrastructures with urban coastal aquifers, few studies have been dedicated to develop methods and models to quantify this interaction [6,7].
Here, we report a study to investigate the interaction of a coastal urban aquifer with a sewer network. The area (Hoboken, NJ) is located in the North-East of the United States at the Hudson river estuary within the metropolitan area of New York city. The work was motivated by large concentration of fecal indicator bacteria in the river during dry weather suggesting groundwater inundation of the sewer. The watershed was implemented in MODFLOW with a geology determined by applying geostatistics on few localized geological data. Boundary data consisted of variable aquifer head and tidal river level. The hydraulic conductivity and the recharge were estimated through stochastic inverse modeling on the hydraulic head measurements within the domain of interest. By knowing the location of the sewer pipes and compared with the estimated groundwater table with uncertainty, the potential of groundwater inundation of the sewer was assigned to each part of the network.
Model improvement by conditioning on data collected at multiple scales remains a challenge in complex settings. We employ an information-theoretic approach that allows for seamless integration of multi-resolution data into multi-scale simulations to upscale conductivity of heterogeneous formations. Fine-scale information is summarized into a coarse scale representation by setting a probabilistic equivalence between the fine and the coarse scale, with parameters that are determined via minimization of observables error and mutual information across scales.
The highly compressible nature of some aquitards leads to nonlinear consolidation where the groundwater flow parameters are stress-dependent. The case is further complicated by the heterogeneity of the hydrogeologic and geotechnical properties of the aquitards. To adequately model land subsidence in these systems, we develop a modeling approach to couple a nonlinear 1-D groundwater flow and consolidation model with a data assimilation scheme based on ensemble Kalman filter. This modeling approach allows to estimate the ensemble mean distribution of state variables and stress-dependent parameters, such as hydraulic conductivity (K), pore-pressure and total settlement. Zapata-Norberto et al. (2017) have shown that in randomly heterogeneous highly compressible aquitards under 1-D vertical flow, the parameter with largest impact on ensemble total settlement and its variance is K. We therefore consider the case where only K is randomly heterogeneous. We consider cases where pore-pressure and/or K measurements are available at given time intervals. We test our approach by solving the nonlinear flow and consolidation problem on a generated 1-D realization of lnK with exponential spatial correlation. These results are taken as our “true” solution. We take pore-pressure and/or lnK “measurements” at different times from this “true” solution. The ensemble Kalman filter method is then employed to estimate ensemble mean distribution of lnK, pore-pressure and total settlement based on the sequential assimilation of those measurements. The ensemble-mean estimates from this procedure closely approximate those from the “true” solution. This procedure can be easily extended to other random variables such as compression index and void ratio.
Chemical flooding is one of the most promising EOR technique in both laboratory research and field trials. It has been applied in conglomerate reservoir as well as sandstone reservoir. To full understand the displacement mechanisms of chemical flooding in reservoirs with different lithology, it is essential to recognize the residual oil displacement in pore scale.
We selected three cores with similar permeability, including outcrop sandstone, sandy conglomerate and inequigranular conglomerate to perform unclear magnetic resonance experiments. Deuteroxide was used to replace water as aqueous to distinguish signals of water phase and oil phase. The threshold values of different pore sizes were established from the relationship between mercury injection curves and NMR T2 spectrums. The distribution and migration of residual oil in different flooding processes was evaluated by quantitatively analyzing the change of the relaxation time. The oil displaced from pores of different sizes after the water flood, polymer flood, and the surfactant/polymer (SP) flood was calculated, respectively.
Compared with sandstone, the diagenesis of conglomerates normally takes place in a shallower depth and then possesses more tortuous pore structures. The sequence of ultimate oil recovery of chemical flooding from high to low is outcrop sandstone, sandy conglomerate and inequigranular conglomerate. On contrary to sandstone rocks, more residue oil is observed after water flood in conglomerated rocks, especially in pores which size is less than 5um. Although both polymer and SP flood can mobilize residual oil in which pore size is larger than 1μm, SP flood has much higher oil recovery than polymer flood, especially for the conglomerate core. The residual oil in medium pores (5μm to 15μm) contributed the most to the incremental oil recovery for the SP flood. It was nearly no trapped oil in large pores which sized larger than 15μm after SP flood. Therefore, the residual oil in small and medium pores were the main target for EOR after the SP flood in both sandstone and conglomerate reservoir.
Since there are many different lithology in conglomerate reservoir, it is necessary to investigate the effect of lithology on residual oil displacement. Our studies elucidated the pore-scale oil recovery mechanisms of different lithology during three flood processes. This research result can help ensuring the distribution of residue oil based on well log result and offer a good guidance to field production.
Fractures are ubiquitous in the subsurface, and provide primary pathways for fluids traveling underground. The roughness and wettability of the fractures in the subsurface cause a major impact on multiphase flow behavior. Nevertheless, published analytical solutions for multiphase flow properties of fractures fail to account for the complexity of the surface mineralogy heterogeneity and its effect on production from fractured reservoirs. Since analytical solutions for fracture surfaces with heterogeneous wetting properties are very limited, we propose a direct simulation approach. The purpose of this work is to correlate the reduction of relative permeability in multiphase flow through fractures with different mineralogy and surface roughness.
Utilizing the Shan-Chen multiphase model of the lattice-Boltzmann method (LBM) we are able to simulate oil and water displacement in 3D fractures. Simulation domains were extracted from 3D micro-CT scans and digitally synthesized. A surface interaction parameter is adjusted to mimic the time-dependent microscale wettability of the different minerals present in fractures. We then map the mineralogy of surfaces obtained by SEM imaging and integrate them with contact angle measurements on individual minerals to provide input for the simulations. We also account for the possible change in contact angle over time due to wettability alteration. These simulations were carried out at the Texas Advanced Computing Center.
In this work, we quantify the effect of different mineralogy arrangements and wetting states on the relative permeability. From these measurements, we derive correlations based on fracture aperture, surface roughness, and the spatial distribution of minerals.
The seismoelectric and self-potential methods are showing promises to characterize both the vadose zone of the Earth, hydrocarbon reservoirs and CO2 sequestration. That said, the dependence of a key parameter, the streaming potential coupling coefficient, with the saturation remains highly debated. We explore here the relationship between the streaming potential coupling coefficient, the water-gas saturation and the salinity in saturated and partially saturated carbonate rocks characterized by distinct textures. All the samples are saturated with NaCl brines, from 2.10-3 Mol L-1 to 2 Mol L-1. The magnitude of the coupling coefficient increases when the brine salinity decreases. Moreover, the streaming potential coupling coefficient seems independent of the nature of the rock in the range 2 – 600 mD. The core samples are characterized in terms of their porosity and intrinsic formation factor. A new core flooding system is used to measure simultaneously both the relative permeability, the resistivity index and the streaming potential coupling coefficient in steady-state two-phase flow conditions as a function of the saturation with CO2 or N2. The results are compared with a recently developed theoretical model, which can accommodate either the Brooks and Corey model. This model is predicting a set of relationships between the streaming potential coupling coefficient, the relative permeability and the second Archie’s exponent. We found a good agreement between the model based on the Brooks and Corey approach and experimental data.
We consider log hydraulic conductivity (Y) as uncertain and predict steady-state groundwater head (h) through three different, independent approaches. The first two of them are based on the ensemble Kalman filter (EnKF), their difference being in the way statistical moments (SM) of state variables and parameters are estimated numerically before the Kalman filter is applied. Whereas in the first approach Monte Carlo simulations are used for this aim, in the second approach the required SM are obtained by solving nonlocal stochastic moment equations (ME) of steady state flow. In the third approach, the ME are subjected to a geostatistical stochastic inversion using a pilot point parameterization. Additionally, a less computationally demanding version for each of the first two approaches was obtained by modifying its algorithm to estimate Y geostatistically (i.e., using generalized kriging) at the pilot points of the third method instead of estimating over the entire grid as its original algorithm does. The three methods and their second versions were applied separately in a synthetic 2D steady-state groundwater flow domain to compare their performances. Our results show that all methods and associated versions were effective in estimating h, reaching at least 81% of predictive coverage. For log hydraulic conductivity similar accuracies were obtained in terms of the mean absolute value error for all methods. In terms of numerical performance, we found that coupling the EnKF methods with kriging reduces the CPU time required for data assimilation while both estimation accuracy and uncertainty do not deteriorate significantly.
A numeric model of non-reactive flow through naturally stratified sandstone samples is presented. This work is based on laboratory experiments in which it was established that solute migration in saturated stratified porous media was dominated by stratification. The experiment results strongly suggest that the effect of the stratification is dominant for flow parallel to the lamination in these sandstones. For flow perpendicular to stratification, the behavior is the expected for a homogeneous medium. In this work, an artificial porous media is built based on the parameters obtained from the experiment samples (sample size, grain size, and porosity). The Lattice Boltzmann method is used for modeling the flow through the samples. The aim is to use this method to compare the results from the model with the ones obtained from the laboratory experiments on real media. The sandstone modeled, consisted of a repetition of layers of i) medium-grained sand and ii) fine-grained sand. The physical properties of this structure were used as parameters to build the artificial porous media. The results from the LBM flow simulation show that when the flow is parallel to lamination, conditions arise for preferential flow, i.e., fingered flow. This instability is associated with the arrangement of grains of different sizes found in the layers. On the other hand, the model predicts uniform flow when the flow is perpendicular to stratification. It is caused by the fine-grained layers, which diffuse the fluid momentum when it moves through these layers.
In the middle-high permeability sandstone reservoir with strong heterogeneity and low permeability reservoir, invalid recycling of the injected water is formed along the natural high permeability belt and interwell fracture at high water cut stage, which results in reduction of reservoir recovery degree. According to the field tests, the reservoir recovery degree can be improved by coupling between injection and production. However, mechanism research on coupling between injection and production is rare at present. This study conducted mechanism research in terms of capillary inhibition and streamline field, and established numerical model. This model can describe inhibition and absorptive percolation phenomena caused by capillary force and additional channeling flow phenomena caused by elastic force. At last, effects evaluation was implemented combining field tests in Sheng Li oil field, coupling between injection and production may increase recovery efficiency about over 1% in water-wet oil reservoirs. At the same time, reasonable coupling cycle and ratio of injection and production is provided for reference in different types of reservoir.
We report on an extensive investigation of solute mixing and spreading in reservoir rocks, including Bentheimer Sandstone (BS), Ketton Limestone (KL), Edward Brown (EB) carbonate and Indiana Limestone (IL), as well as unconsolidated bead pack (BP) as control material. We observe that the selected rock samples possess distinct strength of subcore-scale heterogeneity and present characteristic features, such as uniform pore structure (BS), a significant degree of microporosity (KL, IL) as well as vuggy porosity (EB). Helium pycnometry, mercury intrusion porosimetry and micro-CT image analysis were applied on small sub-sets (plugs) of each rock samples to provide a distribution of baseline microscopic properties, such as skeletal density, porosity, pore- and grain-size distribution. Unidirectional pulse-tracer tests have been carried out on each rock core over a range of Peclet numbers (Pe = 20-400) and by simultaneously measuring breakthrough curves to provide estimates of the solute longitudinal dispersivity. The unique aspect of this study is that the tracer tests are supplemented by the combination of two imaging techniques: X-ray Computed Tomography (CT) is used to quantify subcore-scale heterogeneities in terms of porosity and permeability contrasts at a spatial resolution of approximately 10 mm3, while Positron Emission Tomography (PET) is applied to image the spatial and temporal evolution of the full tracer plume non-invasively. The latter provides unprecedented insight on the transport mechanisms inside the porous sample at a resolution of a few mm (Figure 1 - image attached). Most significantly, PET imaging enables computing macroscopic measure of mixing as a function time, such as spatial moments, the dilution index and the scalar dissipation rate, which in turn serve as quantitative metrics to compare observations for the different rock samples. Different models have been successfully applied to match the observed breakthrough curves, including the classic Advection-Dispersion Equation, the Multi-Rate Mass Transfer model and their combination within a streamtube framework. The validity of each model was assessed by evaluating its capability to predict the internal tracer concentration profiles measured by PET. We observe that the effects of macrodispersive spreading can overcome those of local dispersion for heterogeneous rocks (particularly carbonates). In this context, the use of PET in combination with X-ray CT provides significant opportunities to advance our understanding of miscible displacements in consolidated porous media, thus including those involving additional phenomena, such as adsorption, chemical reactions and capillary effects.
Structural trapping is the ultimate barrier for reducing the risk of leaks at CO2 storage sites. Small pores in high specific surface clay-rich caprocks give rise to high capillary entry pressures and high viscous drag that hinder the migration of buoyant carbon dioxide CO2.
In this work we show measurements of the CO2 breakthrough pressure and ensuing CO2 permeability through sediment plugs prepared with sand, silt, kaolinite and smectite. Our experiments and data from the literature demonstrate that the breakthrough pressure can reach ∼6.2 MPa in argillaceous formations, and 11.2 MPa in evaporites. The CO2 relative permeability after breakthrough increases up to a maximum of ~0.2. Our parametric study highlights the inverse relationship between breakthrough pressure and pore size, as anticipated by Laplace’s equation. In terms of macro-scale parameters, the breakthrough pressure increases as the sediment specific surface increases and the porosity decreases.
In addition, we introduce two dimensionless numbers that help pre-asses the safety of storativity condtions in-situ. The “sealing number” and the “stability number” combine the initial fluid pressure, the buoyant pressure caused by the CO2 plume, the capillary breakthrough pressure of the caprock, and the stress conditions at the reservoir depth; these two numbers provide a rapid assessment of potential storage sites.
In a geological carbon storage (GCS) project, it is critical to predict the extent of injected CO2. However, it is not practical to quantify the uncertainty in the CO2 plume extent by conducting full physics flow simulations for hundreds of geological models representing high geological uncertainty. In this study, a computationally efficient surrogate model is introduced to quickly approximate CO2 plume migrations in a 3-dimensional heterogeneous reservoir during an injection period. CO2 plume migrations are approximated based on connectivities between a CO2 injector and other locations, which are computed using rock and fluid properties. The connectivity-based surrogate model saves about 90% of the computational cost in quantifying the uncertainty in the extent of CO2 plume compared to a full physics flow simulator.
Injection of supercritical carbon dioxide (CO2) into geological formations is used for both atmospheric greenhouse gas reduction (climate change mitigation) and enhanced oil recovery. In an effort to fully understand CO2 trapping efficiency, the capillary trapping behaviors that immobilize subsurface fluids were analyzed at the pore-scale using pairs of proxy fluids representing the range of in situ (supercritical) nonwetting and wetting fluids. The pairs of fluids were cycled through imbibition and drainage processes using a flow cell apparatus containing a sintered glass bead column. Computed x-ray microtomography (microCT) was used to identify immobilized nonwetting fluid volumes after imbibition and drainage events.
From the images, the trapped residual (post-secondary imbibition) nonwetting phase was spatially correlated to both the original (post-primary imbibition) and the initial (post-primary drainage) nonwetting phase; relationships referred to as the original saturation dependence (So-dependence) and initial saturation dependence (Si-dependence), respectively. Statistically significant trends of decreasing So- and Si-dependence with increasing wetting and nonwetting fluid phase viscosities were observed. This finding implies that the amount of CO2 injected and ultimately trapped is dependent on the nonwetting phase (e.g. oil or gas) already present in the formation, as well as on the manner in which supercritical CO2 is initially injected.
To quantify in-situ CO2 residual trapping for CO2 geological storage, dedicated push-pull experiments have been carried out at the Heletz, Israel pilot CO2 injection site. The site is well characterized and instrumented for CO2 injection and sophisticated sampling and monitoring (Niemi et al., 2016) and residual trapping experiments have been carried out during 2016-2017. The objective of the present work is to develop a simulation model capturing the CO2 transport and trapping behavior consistent with the recorded pressure and temperature data at the injection well, with special focus on the coupled wellbore-reservoir flow. For this purpose, the simulation of the CO2 push-pull (injection-withdrawal) experiment is carried out with the numerical simulator T2Well/ECO2N (Pan et al., 2011) to account for the role of wellbore-reservoir coupling. Of particular interest in this work is to accurately model the period when the well is self-producing fluids and to analyze what conditions are causing the observed gas-release behavior. Comparison of numerical model simulations and the measured data suggests that the gas saturation in the reservoir at the onset of the self-release period is only slightly above the residual gas saturation of the formation. In addition, the results indicate that the effective permeability in the reservoir is small enough to be the controlling factor for the gas inflow rate into the wellbore. This detailed modeling of the well self-release behavior allows a more reliable overall estimation of the in-situ residual trapping at the site.
Understanding the influence of CO2 injection on rock stress is one of the key elements to analyze CO2 Enhanced gas recovery and long term CO2-storage in tight sand gas reservoirs. Producing natural gas from reservoir and injecting CO2 to the tight reservoir causes a change in pore pressure, which in turn, changes the three dimensional effective stress state. The stress path followed by the reservoir controls the evolution of the effective stress state, and with it the changes in deviatoric stresses which cause reservoir rock deformation and permeability changes. In order to depict these effects, a new experiment and modeling method-CO2 flooding considering the Interaction between CO2 and Rock - are required. This paper presented a series of stress-sensitive experiments during CO2 flooding. A total of twenty-eight sandstone rock samples, which include sand stones with natural fissures, shear fissures and artificial sanding crack fissures, were selected to study stress-sensitivity during gas production and CO2 flooding. Core flooding experiments were conducted in these rocks. The flow regime used was first depleted follow by injecting CO2. After the experiments, the numerical simulations coupled with nonlinear geo-mechanic model with fluid flow were performed. The simulation results give a detailed understanding of the experimental geo-mechanic system. It is concluded that the experimental and simulation methods can be used in combination to evaluate the potential for stress changes during CO2 flooding in tight gas reservoir. The results show that the evolution of the stress state, captured as a stress path, significantly affects the gas production profile and CO2 storage capability. This research of spatial and temporal changes in stress state laid the basis for studying CO2 storage and enhanced gas recovery.
To understand how does the injected CO2 migration could help increase the available storage capacity in geologic formations, this paper reports a series of experiments of core flooding. To examine the effects of CO2migration pathways in geologic formations, our team have developed a core flooding test of displacing water in porous media with CO2. The samples were obtained from the Ordos Basin, all formations being used for carbon capture and storage(CCS) pilot—Shenhua CCS demonstration project, with a capacity of 100,000 tones CO2 per year, which is the first fully process of CCS in saline reservoir in China. All flooding tests in this paper were performed in the industry CT scanner, medical CT scanner and micro CT scanner.
Experiments were performed over several weeks by injecting CO2-saturated brine through samples. At the same time the samples were scanned with a computed tomography (CT) scanner at regular intervals (0.5mm) during the course of the experiments. Injection flow rates and temperature of the system were varied for each experiment. For the first test of every sample, the helium gas as the flow constant pressure 100 psi at different flow rates (0.72ml/min, 1.48 ml/min, 2.96 ml/min ) was test. Then the brine and CO2 as the flow constant pressure 2000 psi at different flow rates (1.72ml/min, 3.43ml/min, 5.13 ml/min) was test. The constant injection pressure resulted in unstable flow patterns. For the subsequent tests a constant injection rate was set with the Isco pumps and with additional software controls to ensure the pore pressure did not exceed the confining pressure (2500 psi). As long as the injection pressure was less than the confining pressure the flow rate was constant for a constant delta pressure. When the injection pressure increased to a value close to the confining pressure the flow rate was decreased to ensure safe operations.
A review of the findings in common among the studied is presented in the final sections of this paper.
In the context of geological carbon sequestration (GCS), carbon dioxide (CO2) is often injected into deep formations saturated with a brine that may contain dissolved light hydrocarbons such as methane (CH4). In this multicomponent multiphase displacement process, CO2 competes with CH4 in terms of dissolution, and CH4 tends to exsolve from the aqueous into a gaseous phase. Because CH4 has a lower viscosity than injected CO2, CH4 is swept up into a ‘bank’ of CH4-rich gas ahead of the CO2 displacement front. On the one hand, this may provide a useful tracer signal of an approaching CO2 front. On the other hand, the emergence of gaseous CH4 is undesirable because it poses a leakage risk of a far more potent greenhouse gas than CO2 if the cap rock is compromised. Open fractures or faults and wells could result in CH4-contamination of overlying groundwater aquifers as well as surface emissions. We investigate this process through detailed numerical simulations for a large-scale GCS pilot project (near Cranfield, Mississippi) for which a rich set of field data is available. An accurate cubic-plus-association (CPA) equation-of-state (EOS) is used to describe the non-linear phase behavior of multiphase brine-CH4-CO2 mixtures, and breakthrough curves in two observation wells are used to constrain transport processes. Both field data and simulations indeed show the development of an extensive plume of CH4-rich (up to 90 mol%) gas as a consequence of CO2 injection, with important implications for the risk assessment of future GCS projects.
The progression of reactions in systems where mixing occurs has been the subject of investigation for decades; however, there is still much that is unknown about such systems. One area of particular interest to us is the influence of the initial configuration of a system as it evolves in time. Many (if not most) investigations of mixing are formulated at the long time limit, which requires that a certain amount of relaxation of the system has occurred. Although this is an interesting regime that has relevant to, for example, reactions occurring in the subsurface, it does not describe the initial phases of mixing well.
In this work, we examine how the initial configuration of a system can influence the mixing and reaction process evolution. Such early time conditions have relevance to many systems. For example, in industrial processes the basic design constraints of mixing facilities essentially requires that the mixing regime is dominated by early time behavior (e.g., the effective reaction rate for multi-component injections into tubular, packed bed, or fluidized bed reactors would all generally depend strongly on the initial configuration of the chemicals introduced).
Because mixing is exceptionally complex, we have chosen to examine a two-component mixing and reaction process within a tubular reactor. This choice is motivated in part by the simplifications that this geometry allows, and in part because of recent successes we have had in better understanding the early time dispersion process (pre-asymptotic Taylor dispersion) in such systems. The results of this work will focus primarily upon (1) development of the effective mass transport equations (resulting in an explicit representation of convection, effective dispersion, and the effective reaction rate), and (2) presentation of a closure scheme for this problem, and an prediction of the effective rate of reaction via numerical computations. In particular, we will discuss the need to search for effective empirical dynamical scaling laws even in the presence of a fully-predictive theory. The need for empirical models arises because, in this particular case, the process of upscaling does not reduce the complexity of the problem to the extent that would be the most useful for applications.
Carbon dioxide huff-n-puff process can be used as an effective geological storage approach for oil reservoirs. In this study, the experiment method of carbon dioxide huff and puff was established by high-pressure physical simulation system for large scale outcrops. The corresponding development effectiveness and influence factors of the huff-n-puff process were analyzed. Such as time of huff-n-puff process, injection pore volume multiple, time of shutting in, fracture density, fracture length, fracture spacing, fracture shape and fracture flow capacity. The results show that in consequence of the parameter of huff-n-puff process and fracture, the carbon dioxide huff-n-puff process in staged fracturing horizontal wells is one of the effective enhanced recovery techniques and the cumulative recovery percent of reservoir rises from 12.37% to 17.35%. Optimized parameters of the huff-n-puff process and fracturing parameters can provide powerful technical support for geological storage of carbon dioxide.
In EOR research, the use of natural reservoir core is faced with four main problems: 1. The repeatability of the petrophysical properties in natural reservoir core is poor, making it unsuitable for contrast and repeat experiments to examine the influence of individual petrophysical factor under specified physical conditions. 2. The physical properties of the natural reservoir core are uncontrollable and cannot be tailored to provide the macroscopic material properties according to the research needs. Thus natural reservoir core is unsuitable to be used to carry out controlled experimental studies, such as the research on mechanism by which permeability and wettability influence seepage and oil displacement. 3. The natural reservoir core has complex mineral composition and physical properties. The influence on seepage and oil displacement in natural reservoir core is thus a combined result of many complex factors, with quantitative analysis of contribution from each individual factor and interpretation of the experimental result difficult. 4. Natural reservoir core are typically cored from a localized zone in the reservoir and thus cannot model the macroscopic heterogeneity of the reservoir. Additionally, the size of natural reservoir core is generally rather small with large end effect, and thus cannot be used for two-dimensional experiment nor experiment to investigate distributive properties (e.g. pressure distribution, etc). These shortcomings of natural reservoir core highlight the importance and irreplaceability of artificial core. The authors introduce an artificial core synthesized with a specialized adhesive served as binding in this study that addressed the mentioned shortcoming of the natural reservoir cores in two ways. First, permeability of the artificial core introduced in this work can be tailored to be in the range of 10-5000 mD through adjustment of the ratio of differently sized sand particle, optimization of sand compression pressure, and optimization of sand compression time. This is accomplished by first finding out the relationship between sand particle size and resultant porosity/permeability of the synthesized artificial core. Second, through this established relationship between sand particle size and the porosity/permeability of the synthesized artificial core, authors were able to accurately synthesize artificial core with comparable porosity and permeability within a tight tolerance to the target natural reservoir core . Through this study, the adjustment of the ratio of differently sized sand particle can achieve synthesized artificial core with different permeability and porosity. The method introduced in this work can match closely the permeability and porosity to the natural reservoir core, but the work load is large to find the correct ratio of each type of sand. Also, the accuracy of the other petrophysical properties of the synthesized artificial core is not guaranteed, such as the particle size distribution and pore throat distribution.
Density-driven convection can accelerate the rate of CO2 solubility trapping during geological CO2 storage in deep saline aquifers. We present a bench-scale experimental method based on refractive light transmission (RLT) in an analogue system that enables comprehensive study of solutally induced density-driven convection in saturated porous media. In an analogue system, we investigate density-driven convective mixing under conditions relevant to geological CO2 storage. A range of Ra values relevant to potential storage sites are investigated by varying the grain size and density contrast in the laboratory setup. We show that the method accurately determines the solute concentration in the system with high spatial and temporal resolution. We can thereby quantify the onset time of convection (t_c), mass flux (F) and flow dynamics for the different Ra values tested. Based on our findings, we present a scaling law for t_c. The resulting dependence of t_c on Ra, indicates that t_c is more sensitive to large Ra than previously thought. Our findings can also show why F is described equally well by a Ra-dependent or a Ra-independent scaling law. The new method and findings can serve to improve the understanding of convective mixing processes in saturated porous media, and aid the assessment of CO2 solubility trapping, including potential for trapping under given field conditions.
Dimensional analysis applied to bacterial chemotaxis towards NAPL contaminants
Xiaopu Wang 1§, Beibei Gao 2§, Shuaiwei Gu 3, Wei Zhong 1, Kenneth S. Kihaule 1, Roseanne M. Ford 2*
1. National Engineering Laboratory for Subsea Equipment Testing and Detection Technology, China University of Petroleum (East China), Qingdao 266555, China
2. Department of Chemical Engineering, University of Virginia, Charlottesville, VA 22904, United States
3. Shandong Provincial Key Laboratory of Oil & Gas Storage and Transportation Security, China University of Petroleum (East China), Qingdao 266555, China
The use of chemotactic bacteria in bioremediation may improve the efficiency and decrease the cost of restoration, which means it has the potential to address environmental problems caused by oil spills. However, most previous studies were focused at the laboratory-scale and there lacks a formalism that can use these laboratory-scale results as input to evaluate the relative importance of chemotaxis at the field scale. Thus, dimensional analysis was adopted to solve this problem. First, the main influential factors were extracted according to the previous researches on bacterial chemotaxis and a set of dimensionless numbers were obtained according to Buckingham theory. After collecting basic parameter values from previous studies, we formulated a dimensionless equation shown as BR=1.987(P1^-0.0179)(P2^0.3235)(P3^0.0319), where BR (bacterial ratio) is the ratio of maximum bacteria concentration to its original value, and P1, P2 and P3 are combinations of the derived dimensionless numbers. For BR greater than one, the bioremediation strategy based on chemotaxis is expected to be effective in relative contaminated groundwater system, and chemotactic bacteria are expected to accumulate around non-aqueous phase liquid (NAPL) contaminant sources efficiently.
Keywords: Bioremediation; Bacterial chemotaxis; Numerical simulation; Dimensionless analysis
*Corresponding auther.
E-mail address: rmf3f@virginia.edu (R.M.Ford)
§ These authors contributed equally to this work.
Microbially Induced Calcite Precipitation (MICP), or bio-cementation, has shown significant promise as an environmentally-conscious alternative to traditional geotechnical ground improvement technologies, which oftentimes rely on hazardous grouting chemicals, high mechanical energy, and energy-intensive materials to improve the engineering properties of soils (DeJong et al. 2013). In the urea hydrolysis driven process, soil microorganisms containing urease enzymes catalyze a hydrolysis reaction that degrades urea, producing total ammonium, dissolved inorganic carbon, and hydroxide ions (Stocks-Fisher et al. 1999). In the presence of soluble calcium from treatment solutions or groundwater, the process can supersaturate solutions with respect to calcium carbonate and initiate mineral precipitation on soil particle surfaces and contacts. Bio-cementation can transform the mechanical properties of granular soils through large increases in shear stiffness and strength, with reductions in hydraulic conductivity and porosity (DeJong et al. 2006; Montoya & DeJong 2015; Gomez & DeJong 2017).
Despite many recent advances with respect to MICP, environmental concerns regarding the fate of produced nitrogen by-products have remained largely unaddressed. Although ammonium is a commonly encountered source of inorganic nitrogen in soil systems, high aqueous ammonium concentrations produced following MICP may present serious environmental and human health concerns if left untreated in subsurface soils. In order for MICP to become an environmentally-conscious technology, it is clear that scalable methods to manage, remediate, and/or remove nitrogen by-products following bio-cementation must be investigated.
In this study, a series of soil column experiments were performed to examine the transport, removal, and transformation of nitrogen by-products following MICP treatments. In all columns, bio-stimulation treatment techniques following previous experiments (Gomez et al. 2018) were used to enrich native ureolytic soil microorganisms and enable bio-cementation. Columns were 7.6 cm in diameter, 17.8 cm long, and contained a poorly-graded sand material prepared to an initial porosity of 35%. Prior to all treatments, columns were saturated with de-ionized water and an ammonium bromide solution was applied to investigate the transport of ammonium relative to a passive tracer. Soil columns received identical treatment solution injections over 14 days targeting a final post-treatment calcite content of 4 to 5% by mass. During treatments, non-destructive shear wave (Vs) and compression wave (Vp) velocity measurements were completed to monitor changes in soil skeleton shear stiffness and potential changes in pore fluid compressibility resulting from bio-gas production and carbonate degradation. Aqueous samples were obtained from all columns in time to examine changes in solution chemistry related to microbial urea hydrolysis, calcite precipitation, and ammonium sorption and transport. Following all cementation treatments, soil columns received daily rinse solution injections with differing ionic strengths and pH values to investigate the effect of rinse solution chemistry on by-product removal and the potential for ammonium leaching during solution retention periods. Following all treatments, the effect of rinses on by-product removal and bio-cementation integrity was evaluated. Results will guide efforts to model the transport and removal of ammonium by-products following bio-cementation.
Biologically mediated processes are being developed as an alternative approach to traditional ground improvement techniques. Denitrification has been investigated as a potential ground improvement process towards liquefaction hazard mitigation. During denitrification, microorganisms reduce nitrate to dinitrogen gas and facilitate calcium carbonate precipitation as a by-product under adequate environmental conditions. The formation of dinitrogen gas desaturates soils and allows for potential pore pressures dampening during earthquake events. While, precipitation of calcium carbonate can improve the mechanical properties by filling the voids and cementing soil particles. As a result of small changes in gas and mineral phases, the mechanical properties of soils can be significantly affected. Prior research has primarily focused on quantitative analysis of overall residual calcium carbonate mineral and biogenic gas products in lab-scale porous media. However, the distribution of these products at the pore-scale has not been well-investigated. In this research, denitrification is activated in a microfluidic channel simulating a homogeneous pore structure. The denitrification process is monitored by sequential image capture, where changes in the gas and mineral phase are evaluated by image processing. The results from the experimental study are compared to the results of two-dimensional simulation model which involves the relevant biochemical reactions, diffusion, and convection.
Denitrification is one of the key microbial reactions for sandy soils to induce desaturation and calcium carbonate precipitation. As the replacement of urea hydrolysis for microbially induced carbonate precipitation (MICP), the effect by denitrification has been evaluated. Calcium carbonate precipitation and biomass production occur in soil through the reaction process and some of these accumulate in the pore space or on the surface of the soil particles. Due to the accumulation, reducing the porosity, permeability of the soil possibly reduces. Well understanding of pore structure alteration through the reaction makes possible to control permeability efficiently.
This study evaluated pore diameter distributions of three sand samples before and after MICP treatment via denitrification by air intrusion method. By measuring air flow rate applying controlled air pressure into water saturated specimens of the sands, air permeability and pore diameter were calculated. Comparing non-treatment sands, the higher air pressure was required to push pore water against capillary pressure out of pore throats of sands after the treatment. The pore diameter distribution curves were slightly shifted to the smaller pore size range after the treatment. These results indicate that pore water retention ability of the samples was altered by the treatment.
The Center for Biofilm Engineering (CBE) at Montana State University has a long, successful history of investigating biofilm and mineral precipitation processes in subsurface environments. This poster summarizes many of the experimental approaches the CBE has taken to develop field-suitable technologies. There are numerous applications for engineered biomineralization. The CBE has largely focused on the sealing of leakage pathways, water remediation, soil stabilization, dust suppression and enhanced resource recovery. The experimental systems to examine these engineering applications have been designed to interrogate biofilm and mineral precipitation processes in a variety of environments, conditions, geometries and chemistries. This poster will highlight experimental systems from low pressure to high pressure, bench scale to field scale and low temperature to high temperature.
To date, soil bio-cementation via Microbially Induced Carbonate Precipitation (MICP) has been extensively studied as a promising alternative technique for ground improvement to address the growing environmental concerns of traditional chemical cementing agents. This paper presents a new one-phase injection method of biocementation using an acidified all-in-one biocementation solution (i.e., a mixture of bacterial culture, urea, and CaCl2). The key feature of this method is to generate a lag period of the MICP process, which can be controlled by adding acidic pH buffer to the biocementation solution, so that the formation of bio-flocs and CaCO3 crystals is significantly delayed. This feature allows the low-pH all-in-one biocementation solution to be fully injected into the sand column before the biocementation occurs, hence serious surface clogging can be avoided. The duration of the lag phase was evaluated using different amount of HCl or acidic buffer. The performance of biocementation using the acidified all-in-one solution was tested for short and long sand columns bio-stabilization, showing a significant improvement of uniformity. The results of this study show strong potential to scale up the proposed approach to filed applications.
Soil is a complex environment in which the presence of several phases creates numerous interfaces (solid-liquid, liquid-gas and solid-gas). Understanding the local hydrodynamics in soil pores and the biogeochemical processes such as nutrient cycling has been of growing importance in the field of bioremediation and ecology. Besides the coexistence of two immiscible phases (air and water) in the pore space, microorganisms, especially bacteria, are often found in large numbers in natural soil environments. The complex spatial distribution of air and water results in the development of a mosaic of regions of very low water velocity, including areas where water or air is trapped and of preferential channels of high velocity. This landscape of conditions enables microorganisms to live in the free-swimming phase and to form surface attached communities known as biofilms.
At the same time, the biofilms’ structure influences pore geometries resulting in altered hydrodynamics, affecting biofilm development and therefore mass transport. To study influences of soil conditions on biofilms and vice versa, we have studied two soil-born microorganisms, Pseudomonas and Bacillus, at the pore scale using microfluidic devices. We have explored the biofilm forming behavior under different physical conditions such as varied water saturation and flowrate. Carefully designed channel geometries coupled with automated video microscopy allowed us a zoomed-in view on specific interactions while controlling the water saturation by varying the gas flow into the channel. The simplified geometries of the devices resulted in a varied biofilm growth caused by the presence of an immiscible phase.
Microbial dynamics in porous media are drivers for a number of applications in subsurface pollutant remediation. Biofilms are communities of microorganisms that are attached to interfaces (pores-grains), and embedded within a matrix of extracellular polymeric substances (EPS) that they have produced. Growing biofilms have a very small effect on porosity, but a very significant effect on the hydraulic conductivity, that reduces well beyond the value that would be obtained from the Kozeny-Carman updating formula, and further results in order(s) of magnitude increase in the estimated dispersion coefficient. We present a simplified conceptual model that is capable of providing practical expressions for the variations in conductivity and porosity. The advantage of the expressions is that they are written in terms of observables that are relatively easy to measure in the lab or the field, contrarily to most existing expressions. We then tested our simplified expressions in a number of reported experiments. Finally, we see how the simplified model captures the most significant processes of a global multi-compartment mechanistic model recently presented.
After production, all wells need to be permanently plugged and abandoned (P&A'ed). Long-term well integrity will then rely on the integrity of cement, which is the material typically used for permanent well plugging and for filling the annular spaces between casing/rock. The cement is pumped into the well as a slurry, and hardens to form mechanical and hydraulic seals. Cement has proven to be a robust material for subsurface constructions, but concerns are directed towards the interfaces between cement/rock and cement/casing. These are prone to delamination (commonly referred to as "debonding") and can thus act as leakage paths for formation fluids (oil, water, gas) along the well. In order to reduce the long-term environmental impact of oil and gas production activities, it is thus necessary to improve this "weak link" of today's well construction.
Bonding between cement and steel is controlled by the material structure and packing in the interfacial transition zone (ITZ), a thin (50-100 micron wide) zone located in cement near the steel surface. Applying positive potential on the steel wall improves the bonding and thus could be used to improve long-term well integrity in the field. In downhole conditions, however, cements are subject to elevated pressure and temperature. Moreover, composition of well cements is usually designed so as to stabilize the slurry and thus to reduce attractive forces between the particles. These operational factors may adversely affect the efficiency of such eletrophoresis-induced bonding enhancement.
Experiments are performed in order to investigate the effect of different operational factors (ionic strength, zeta-potential, particle size distribution, additives in cement) on the electrophoresis-induced bonding enhancement. A mesoscopic particle-based model is constructed and used to study the effect of operational factors on the process. The model is based on the discrete-element method (DEM) where particle interactions are introduced via the lubrication force. The lubrication accounts for, in general, non-Newtonian rheology of the carrier fluid in the cement slurry. Stokes flow is assumed. The model enables two-way coupling between the electric field and the particles. It also accounts for the effect of electrolysis near anode upon the particles.
Hot solvent injection is an in-situ technology which uses heated solvent for efficient and sustainable viscous oil (VO) recovery (cf. 93 kg per barrel less GHG emission than SAGD technology). The process reduces the oil viscosity via mass and heat transfer so that the combined effect of heating and solvent dilution yields a better result than in steam (SAGD) or cold solvent injection (VAPEX) cases.
In this case, the injection of the solvent in gaseous state reduces the amount of solvent required per unit volume of produced oil, and takes advantage of higher rate of solvent diffusion. The solvent will not remain in gaseous state but will condense downstream and release a latent heat making easier its mixing with the VO in-place due to the local temperature increase. Recently the hot solvent injection technology (known also under the name NSolv) had its first successful pilot project designed to recover bitumen from oil sands.
Physically speaking the VO displacement in the solvent-based process is a complex combination of dynamic energy and mass transport and phase transformation phenomena. The rapidly emerging experimental technique of fluid dynamical measurements and observations on micromodels (MM) is proved to be a powerful mean providing quantitative information on multiphysical processes in porous media.
The numerical simulation of solvent injection for VO displacement in a MM setup has demonstrated its feasibility and usefulness both for model design and experimental results analysis. This includes first a pore-scale imaged-based study of micromodel transport properties. Then the Darcy-scale model has been developed and applied for dynamic displacement study. It
has been shown that although being not capable to reproduce in detail the fluid- and thermodynamic diversity of the displacement (especially at pore scale), the developed numerical model has indicated the process key parameters and offered the framework for their quantitative determination.
Finally the dedicated study of process dynamics and corresponding adaptation of the numerical model parameters has been presented and discussed.
Shale rocks play an essential role in petroleum exploration and production because they can occur either as caprocks for subsurface storage in conventional reservoirs or as unconventional reservoir rocks for hydrocarbon extraction via hydraulic fracturing. The ability to produce gas from rocks previously only considered caprocks is an unprecedented and innovative feat, but does not come without an environmental impact and costly issues with permeability reduction of engineered fracture systems. The quantities of water required for hydraulic fracturing and developing these formations for production have been large, and the amounts of flowback and produced water after the hydraulic fracturing processes have been astronomical. These volumes make it imperative that a water recycling solution be found and applied to the development of these fields.
In this study, a batch reaction was conducted with Marcellus shale. Both outcrop and reservoir cores (from different points along wellbore trajectory) were exposed to de-ionized water and a synthetic hydraulic fracturing mixture at reservoir temperature for up to four weeks at a high fluid to rock volume ratio. The chemistry of the created simulated flowback and produced water were analyzed using an ICP. In addition, microstructural analyses were performed in order to establish mineralogical and structural properties, as well as presence of microfractures. Furthermore, indentation tests were conducted at both micro and nanometer level to link the geochemistry and geomechanics of shale rocks, through mechanical properties mapping, the volumetric proportions of each phase can be estimated based on the differential mechanical properties.
The key findings include an analysis of the variation of the simulated flowback water from surface down to cores at a depth of 6420ft, with a focus on heavier mineralogical elements and metals. Less than 1% of the fluid used in these tests consisted of hydraulic fracturing fluid additives, however, even with this small volume of additives used, a significant difference in mineral dissolution compared to the samples treated with water only was observed. The carbonates in the rock samples showed a high level of dissolution, which can cause an increase in permeability, but can also precipitate causing fracture bridging as well as scale buildup in wellbore structure/pipes. The concentration of Pb was found to be significant in the water comparison, posing a potential environmental issue. The indentation results showed a significant difference in mechanical properties as the result of the alteration in microstructures and mineralogical composition during the batch reaction, and the change of microstructure causing by dissolution/precipitation of individual phase were correlated with the alteration in bulk response of the rock.These findings are preliminary and would require and extensive study that would include numerous samples for different location within Marcellus Shale as well comparison to other producing shale formations.
In order for a deep-water wellbore to uphold its integrity under high pressure - high temperature conditions, the wellbore must possess complete zonal isolation while surrounded in an extreme environment. Highly variable temperature and pressure ranges, shallow flow zones, as well as potentially corrosive fluids and gasses all present unique challenges to the job of the cement which maintains that zonal isolation. As such, alternative options to mainstream choices often present themselves as attractive avenues of discovery.
As it is of utmost importance to maintain structural integrity under HPHT conditions, cement slurries are pumped downhole to provide zonal isolation and structural support to offshore wells. The wellbore system potentially faces a variety of temperature and pressure fluctuations from the immediate onset. These fluctuations may affect the hydration properties of the cement. It is also important to consider the chemical interactions that the cement may have at the rock-cement interface where potential degradation or annulus gaps may occur further risking a decrease in zonal isolation. This presentation intends to review some of the important issues regarding zonal isolation in HPHT conditions and to highlight critical knowledge gaps in order to generate important research questions.
3d printing in the oil and gas industry is in its infancy. The ability to use 3d printing to not only produce tools and equipment, that could not be manufactured in traditional manners, is only the beginning. There are several possible applications for 3d printing to facilitate this project:
• Controlled deposition of unique and varied barrier materials
• Equipment development for purpose
• Integration of data sensors within casing materials
• Production of planning models for analysis and review
With 3d printing it is possible to “print” in a controlled manner the desired material or material blends to get unique characteristics that would not be possible any other way. The ability to create and blend materials in a planned structure to accommodate the downhole environment would be a step forward in the evolution of hydraulic barriers. 3d printing will allow us to rethink the way oil and gas downhole tools and materials are created and the functions that they perform. The power of 3d printing may allow us to implant long term data gathering sensors right in the barrier materials themselves using technology derived from muti-material deposition from standard FDM 3d printer technology. 3d printing also has many uses in the physical 3d modeling of formation structures and flow analysis for peer review. 3d printings unique layering technology can recreate the details of complex natural geological formations, bringing another dimension to project planning. In 3d printing the user of the tool limits the possibilities of this technology more than the technology of 3d printing itself.
We present here our efforts to characterize wellbore interfaces via chemical and mechanical characterization methods. These methods including mechanical “push-out” tests to measure interfacial bond strengths between cement, host media, and polymeric seal repair materials, 3D geochemical modelling of the wellbore environment during subsurface operations, and “mock-wellbore” experimental test setup for simulating downhole stresses and strains and measuring the ensuing in situ permeability. We will show applications of these techniques to case studies, as well as illustrate how a combined experimental and computational approach can be implemented to better understand the multi-scale, coupled processes that are critical to the design and prediction of subsurface seal performance.
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525. SAND2017-13699 A
• Akerke Mukhamediarova (Institut Elie Cartan, Université de Lorraine)
• Mikhail Panfilov (Institut Elie Cartan – Université de Lorraine ; and
Institut Jean le Rond d’Alembert, Sorbonne Universités)
Oil displacement by water which contains microorganisms able to produce bio-surfactants is one of the most promising methods of oil recovery. The bio-surfactant significantly reduces the surface tension and weakens the negative role of capillary oil trapping. The second effect caused by surfactants is inversion of wetting, which is even more important for oil recovery, since it allows separating oil from pore walls, making it non-wetting (in carbonate reservoirs). We develop the mathematical model of this process, which takes into account both mentioned effects. The model of wetting alternation is its key point. On the macroscale this effect leads to the modification of the relative permeability curves, which may be modeled by special kinetic relationships. The closure relationships for the characteristic time of wetting inversion has been obtained by modelling this process at the pore-scale. The numerical method of diffuse interface was applied to system water-surfactant-oil separated by a meniscus on a solid surface.
For the kinetics of bacterial population grow and decay, we suggest new nonlinear relationships, which enables to model various physiological stages, including the lag stage.
The results of modeling showed the appearance of regimes of self-organization manifested in the form of auto-oscillatory waves in time and in space.
Surfactants can drastically reduce the water/oil interfacial tension (IFT) to mobilize residual oil. However, surfactant flooding is viscously unstable inherently because of its large mobility ratio. Alkaline/surfactant/polymer (ASP) flooding has been serving as a conventional solution, where polymer plays a vital role in increasing viscosity for a more stable oil bank. Recently, a gravity-stable surfactant (GSS) flooding was proposed as a promising alternative to stabilize the displacement front without mobility control.
In comparison with the sandpack/coreflood experiments, glass-etching model provides a powerful visualization means to examine both static occurrence and flow dynamics of the fluids from a micro perspective. Based on pore network patterns, conceptual models were etched to study the basic mechanism; while actual water-wet models were etched to investigate the effectiveness in heterogeneous scenarios. After the model was vacuumed and NaCl brine-saturated, we took advantage of the gravity, therefore, oil injection (from the top), initial brine injection and surfactant flooding (from the bottom) were vertically performed in turn. Using computer image preprocessing technique, microscopic residual oil in the model is automatically recognized and extracted.
After injecting surfactant solution, microemulsion phase was generated between the oil bank and surfactant slug. Their movement and interaction was visualized and recorded. Firstly, we compared the results among brine at under-optimum, optimum, and over-optimum salinities. Besides, under the condition of optimum salinity, formulation was tuned to reduce microemulsion viscosity to obtain higher critical velocity. Results show that a significant enhanced oil recovery was achieved; and the displacement front was sufficiently maintained with collected oil bank moving ahead under high critical velocity. This validates the modified stability theory and corresponding reservoir simulation. Specifically, according to topological structures, we classify the residual oil into five types for a detailed discussion: membranous flow, droplet flow, columnar flow, multi-porous flow, and clustered flow. And their ratio were calculated through shape recognition. During water flooding, the clustered flow (continuous phase) was mainly transformed into multi-porous flow and columnar flow (non-continuous phase). After the surfactant flooding, yet some reversion occurred due to the in situ emulsification. Moreover, in the follow-up experiments, a less surfactant slug size with a brine post-flush could also realize the desired result, saving more chemical cost.
The novelty of the paper lies in 1) applying microfluidics to the newly designed GSS flooding for microscopic investigation; and 2) quantitatively charactering the residual oil during the flooding. This approach illustrates the phenomena and analyzes the mechanism of the GSS flooding at pore scale, which offers guidance for further effective optimization and economical implementation of the flooding.
Poultice technology is currently mainly used for the desalination of masonry structures in the field of architectural heritage conservation [1]. Wet poultices are coated on the porous material to be treated, and kept in place before being removed when dry. The efficiency of the process basically depends on the drying behavior of the system poultice/substrate, but so far little is known concerning drying of systems composed by soft materials applied to substrate surface. Here we show that, in contrast with various other materials (polymers, smectite clay, silica gel, latex, cellulose, etc) kaolin clay pastes have unique absorption properties as they may extract almost all the liquid (and the suspended elements) from the porous medium at a high rate.
We followed drying of different poultice systems coated on a substrate (glass microbeads with two different pore size), with Nuclear Magnetic Resonance (NMR). Different NMR sequences were used which provided the evolution of key internal characteristics in time: 1) the water saturation distribution; 2) transverse and longitudinal relaxation spectrum; 3) the poultice shape (shrinkage). From this information we identified the different drying regimes of poultice/substrate systems in relation with their microstructure evolution.
The remarkable properties of kaolin pastes appear as follow. In a first period, the paste shrinks axially (along the direction of evaporation) without fracturing, so that it maintains a full coverage of the porous medium surface. At the end of this period the material has formed a solid porous structure which, interestingly, will resist capillary pressure during the next steps of the process. When the pore size of this system is smaller than that of the substrate the liquid is progressively extracted from the substrate, while the poultice apparently remains (on average) saturated. In fact, air necessarily penetrates the kaolin paste in the form of transient paths reaching the substrate then closing back. This period lasts until almost 90% of water initially in the substrate is evaporated. A third period starts once capillary forces inside the poultice and the substrate are balanced. The water saturation now decreases almost homogeneously throughout poultice and substrate. Surprisingly, despite these complex processes, the drying rate in both the first and second period remains constant (i.e. at its initial, possibly high, level). This implies that almost all the water contained in substrate moves through the poultice to be evaporated at its free surface under approximately constant conditions [2]. This also implies that in general the elements suspended in the liquid will be transported and finally accumulated inside the poultice, which means that such a material is remarkably efficient.
Foam fluid is a gas-liquid dispersion system whose range of application covers various fields due to its excellent properties, especially in oil and gas field development including enhanced oil recovery, matrix acidizing, gas breakthrough control, profile control, plugging removal, etc. However, factors such as dependence on natural N2 sources, breakthrough of N2 to production wells, transportation of N2 and field equipment, environmental problems and safety have restrained broad use of the N2 foam injection technique in offshore reservoir.
To overcome these issues and extend the application of foam fluid, a new method, namely, underground foaming method, should be used to replace the conventional foaming method and the foam should carry heat. In this paper, we used a self-generated heat system (SGHF) to achieve this goal. The technique is based on the injection of two aqueous solutions i.e. gas-forming and gas-yielding solutions of certain concentrations into the reservoir. In this research, SGHF was obtained by a chemical reaction between gas forming ammonium chloride and sodium nitrite and the reaction products are very friendly, which are N2 gas, H2O, NaCl, and heat. This system has been studied and used in petroleum engineering.
In this work, we firstly perform surfactant evaluation experiments to identify the temperature resistance and salt resistance surfactant suitable for application conditions of offshore reservoirs. Then, we carried out sandpack flooding experiment including single sandpack flooding experiment and parallel sandpacks flooding experiment, as well as the 2 dimension visualization model experiment of three different research systems (SGHF systems, self-generate heat (SGH) system and conventional foam system) to study and analyze the enhancing oil recovery mechanisms of SGHF system.
After analyzing and discussing the experimental results including trap gas volume analysis, calculation of internal chemical reaction degree in sandpack, temperature distribution inside sandpack and temperature variation coefficient, we draw the conclusion that the main mechanisms of enhancing oil recovery are:(1) SGHF system can generate more uniform and even foam than the conventional foam, which not only eliminate the need of gas resource and gas transport, but also provide better gas fluidity control and adjust the production profile of heterogeneous reservoirs. (2) Foam can adjust the temperature distribution. Temperature distribution inside the sandpack is relatively uniform and high for SGHF system which benefits to reduce the viscosity of heavy oil in a larger scale, and we think the reasons are mainly including: Firstly, the presence of foam and the properties of trap gas cause the gas to accumulate near the inlet area, the effective permeability at the inlet decreases, therefore, the reaction heat production is performed evenly in the front and rear parts. Secondly, the presence of a large amount of trapped gas in the sandpack reduce the overall thermal conductivity inside the sandpack and delays the loss of heat.
In areas contaminated by the petroleum industry, persistent compounds such as polyaromatic hydrocarbons (PAH) are often accumulated in the porous matrixes of sediments and soils (S&S), implicating risks to ecosystems and human health since these contaminants are released over time to interstitial and surrounding water. Pore size distributions (PSD) and PAH binding strengths to sorption sites on S&S are characteristics that affect such accumulation and release. Sorption, desorption and diffusion are among the critical processes that control the availability of PAH, and it is therefore crucial to evaluate these processes in order to understand and predict transport and fate of these contaminants in S&S, and for selecting effective remediation procedures.
Four S&S samples were obtained from previously contaminated sites, air-dried, and sieved (mesh 200), and organic matter (OM) was reduced in subsamples by hydrogen peroxide treatment, resulting in eight different porous media. Surface areas and PSD were determined (Autosorb IQ2MP, Florida), and OM were estimated by calcination (ASTM, 1993). Benzo(a)pyrene (BaP) was selected as study PAH and adsorption experiments were carried out in the dark with S&S suspended in NaCl (I=0.047 M) solutions, by adding five different concentrations of BaP with 7-14C-BaP as radioactive tracer, between 2.59 and 12.6x10-4 mmolBaP/gS&S, according to TG 106 Guideline (OECD, 2000). Reactors were fed with CO2-free air to keep suspensions oxygenated and allow carrying 14CO2 and stable CO2, produced by mineralization of BaP and OM, respectively, to alkaline traps, where CO2 production was measured by changes in electrical conductivity. Supernatant aliquots in the reactors and alkaline traps were obtained and 14C was measured in a liquid scintillation counter (Beckman Counter LS6500). At the end of the BaP adsorption experiments (14 d), 6.5 cm2 polyoxymethylene (POM) was added as a passive sampler, desorbing BaP from S&S. POM-accumulated BaP was extracted with acetone and sonication (EPA, 2007). 14C was quantified in the extracts as described above.
In the case of samples with complete organic matter, sorption at 14 days varies between 91.5 and 95.8%, while in the samples with reduced organic matter a variation between 78.7 and 89.9% was observed. Desorption velocities were much slower than adsorption rates, finding values between 2.0 and 6.0 % in 14 days and between 2.2 and 11.1% in 43 days. Using these values it is estimated that the time required to desorb BaP varies between 644 and 8,290 days. This difference between rates of adsorption and desorption should be considered when planning remediation actions for contaminated sites.
Two different time dependencies were observed in adsorption and desorption kinetics: a fast step considered to be due to diffusion of the BaP to the external surface, macro and mesopores, and a slow step, considered as diffusion of BaP into the micropores capillaries. Sorption equilibrium constant (KL), sorption sites and kinetic constants were experimentally obtained and a two-step conceptual model that describes the effect of the dynamics in porous media on the reactive transport of BaP in S&S with different PSD was developed.
Flows in packed beds are encountered in many engineering applications, such as solar thermal energy storages, chemical catalytic reactors, petroleum and civil engineering, magnetic refrigerators, biological tissues, and pebble-bed nuclear reactors.
Critical challenge of designing packed beds involves understanding the total pressure loss, complex flow fields, heat and mass transfer phenomena occurring within the interstitial regions. Unfortunately, complex geometries and randomly connected void spaces within packed beds have hindered efforts to characterize the underlying transport phenomena.
Geometrical complexity inside of a randomly packed bed represents a challenge to experimental and computational efforts in order to construct transport models that have been previously built upon volume averages of micro-scale parameters, however, should accurately capture the flow behaviors.
Fully leveraging the advantages of this type of packed beds requires a fundamental understanding of flow topology within the randomly packed sphere beds. Multiple points or full-field measurements of flow characteristics at a high level of spatial and temporal resolutions are needed to fully map the complex flow patterns and to provide data at high spatial density to permit accurate volume averaging in the pebble bed.
Texas A&M University is conducting isothermal measurements of pressure drops, flow measurements in a randomly packed spheres experimental facility to support the research on advanced nuclear reactors sponsored by Department of Energy (DOE). The main purpose of these tests is to perform high spatial and temporal resolution measurements, and use the obtained results for code validation and model development.
In this poster, we present our experimental results from pressure drop and non-intrusive velocity measurements in different facilities of randomly packed spheres at various porosities and Reynolds numbers.
Pressure drops across the axial length of a versatile facility were measured by high accuracy pressure transducers at various modified Reynolds numbers, and friction factors were accordingly computed. The obtained experimental results were compared and in a good agreement with previous studies available in literature.
High-fidelity velocity measurements at the pore scales and near the wall boundary in a facility of packed spheres were performed by featuring a combined approach of matching-refractive-index (MRI) and laser-diagnostics, such as Time-resolved Particle Image Velocimetry (TR-PIV) and Time-resolved Stereoscopic PIV (TR-SPIV). This approach allows us to non-invasively probe the flow within packed spheres at the microscopic scales with high temporal and spatial resolutions. Statistical results including mean velocity, root-mean-square velocity and Reynolds stress, computed from the TR-PIV and TR-SPIV measurements are illustrated. Effects of wall enclosures and Reynolds numbers to the flow patterns are investigated. Finally, we applied Proper Orthogonal Decomposition analysis to extract coherent flow structures in the near-wall and far-wall regions of the packed beds.
The effective visco-elastic properties of reservoir rocks are strongly dependent on characteristics of the pore geometry and of the inherent viscous pore fluids, saturation degree and excitation frequency. Constraining these dependencies is important for the interpretation of seismic data from geothermal or oil and gas reservoirs. Thus, experimental studies are needed that focus on effective visco-elastic properties in the frequency range of seismic field data. For this purpose, some laboratory apparatuses were developed in the past, using the forced oscillation method, in which stress and strain of a harmonically loaded sample are recorded and analyzed. In most of these studies, Young's modulus and seismic wave attenuation were measured at frequencies below 100 Hz (e.g., [1-4]).However, to interpret complete seismic surveys, laboratory data are required up to even higher frequencies. Therefore, we developed a forced oscillation setup aiming at measuring the effective visco-elastic properties of partially and fully saturated rock samples in the seismic frequency range up to 1 kHz. Cylindrical samples are excited to axial oscillations during which the axial force and the axial as well as the lateral strain of the sample are measured to derive Young's modulus and Poisson's ratio and the two corresponding attenuation coefficients by a sliding-window fast Fourier transformation.
We present forced-oscillation experiments with varying amplitude (amplitude sweep) or frequency (frequency sweep) on different materials at ambient conditions. Calibration experiments were performed on visco-elastic polymethyl methacrylate (PMMA), commonly called Plexiglas, and elastic AlCuMgPb-alloy. These well characterized standard materials with contrasting behavior were already used to test similar apparatuses (e.g., [1-4]). The determined visco-elastic properties do not vary with amplitude in the investigated axial strain range of 4e-06 to 5e-05 (PMMA) or 3e-07 to 5e-06 (AlCuMgPb-alloy), but show a frequency dependence for PMMA. For example, the Poisson's ratio of PMMA decreases continuously with increasing frequency, in agreement with previously reported trends [2]. Currently, we perform forced-oscillation experiments on samples of Berea sandstone with a porosity of approximately 18 %. The axial strain range, in which the four mentioned visco-elastic properties show no amplitude dependence, is identified in amplitude sweeps and used in the subsequent frequency sweeps.
Most of the water-flooding fields in the eastern part of our country have now entered the stage of high-water-cut mining, their actual recovery rates are generally low. Because of this, the research on the remaining oil distribution in the reservoir is urgent. In order to reflect the influence of the pore structure parameters of the core on remaining oil from the microscopic scale, this article starts with sandstone core. The displacement experiment of the selected rock samples and the micro-grayscale images of the cores were obtained after CT scanning. Then images are processed by non-local uniform filtering and watershed segmentation algorithm. Finally, some of the pore structure units were extracted from the remaining oil distribution obtained and a pore network model was constructed. Experimental results show: Rock pore radius is proportional to the degree of enrichment of remaining oil and inversely proportional to the water-flooding effect. The coordination number and shape factor of rock are inversely proportional to the degree of enrichment of remaining oil and proportional to the water-flooding effect. What’s more, the pore-throat ratio value is more intermediate, the lower the enrichment of remaining oil, the better the water-flooding effect. This study is of guiding significance for the design of the remaining oil in the water-flooding oilfield.
Geomaterial pore networks are highly tortuous with intricate geometries and varying surface roughness. It is reported in literature that both pore geometry and surface roughness influence flow through porous media (Ketcham and Carlson, 2001; Noiriel et al., 2016; Lv et al., 2017). Surface roughness is quantified by the deviations in the direction of flow perpendicular to the real surface. Simplified pore networks with known geometric shapes and the quantified surface roughness affords the opportunity to back-calculate internal forces and begin to quantify the effect on contact angles. 3D printed models printed using acrylonitrile butadiene styrene were designed with an internal structure of void geometries to represent a flow path with different geometric interfaces. To look at surface roughness, different techniques were used to add surface roughness to the models. The models were exposed to chemicals that reacted with the material surface to add microscopic surface roughness and macroscopic roughness was added via design and printing techniques. Each model was placed in a core flooding setup and exposed to a series of CO2-saturated brine and scCO2 injections to mimic underground conditions. Once at residual conditions, the core-flooding setup was set to shut-in conditions and scanned used X-Ray micro-computed tomography. 3D reconstructions contain information to measure contact angles, analyze forces, and correlate each to the geometries and surface roughness of each model. Analysis of the local impact to scCO2-brine contact angles within pores with varying surface roughness will be presented.
Salt are a major cause of destruction by crystallization of porous media. Salt will in general enter a porous medium by advection with moisture or diffusion within the moisture. A special situation which occurs often in marine environments in which case there is a permanent supply of sea water at one side of a porous material such as a concrete structure. At the other side, the structure is exposed to continuous drying in the open air. In such a case there will be a mixed situation for describing the transport. Evaporation from the air exposed side provides a continuous moisture sink which is compensated by capillary suction, i.e., wicking, of the sea water. As a result there will be a continuous flux of NaCl ions to the surface. As a result of this combined process the NaCl concentration at the drying surface will increase slowly until crystallization starts. Using a specially designed Nuclear Magnetic Resonance (NMR) setup, the 1H, 23Na and 35Cl content can measured quasi-simultaneously. Hence this NMR method gives us the possibility for real-time monitoring of transport processes of the ions during experiments. We have measured the transport of the both Na and Cl during this combined process for both a sandstone and concrete. The results can be described by a simple analytic model which indicates the concentration profiles can be described an exponential decay.
Direct measurement of shale gas adsorption isotherms at high pressures and high temperatures (HPHT) is intricate and requires expensive apparatuses. Most of the documented studies only report shale gas adsorption data at pressures below 12 MPa, which is much smaller than the reservoir pressure, e.g., up to 36 MPa in Eagle Ford shale. Recent studies also suggest that the excess adsorption isotherm of shale gas exhibits distinct features from that observed at low pressures. Therefore, predicting gas adsorption isotherms at reservoir conditions may be useful.
On the basis of the simplified local density (SLD) theory, we developed a novel dual-site adsorption model for shale gas. Shale matrix composed of both organic matter and inorganic minerals and the pores located within kerogen can be smaller than inorganic pores. Our grand canonical Monte Caro (GCMC) simulations also confirm that the adsorption capacity of organic matter is much greater than those of inorganic minerals. Therefore, the model for shale gas adsorption isotherms should distinguishes methane adsorption in kerogen surface from that of the inorganic substrates. Our proposed dual-site SLD model takes into account the different pore sizes and fluid-solid interaction energy parameters of organic matter and inorganic minerals.
We first used conventional SLD model to match the excess adsorption isotherms of CH4 in graphene and montmorillonite slit (pressure: 0-40 MPa). Excellent agreements are observed, which manifest that the SLD model is able to describe gas adsorption in a slit at both subcritical and supercritical states. Then we examined the validity of our proposed dual-site model using high-pressure CH4 adsorption isotherms on shale reported in the literatures. These experimental data were measured at pressures up to 25 MPa and temperatures up to 150 ℃. Our proposed model fit these adsorption isotherms very well. If we use the experimental results measured at low pressures (<12 MPa) to make the fit, the high-pressure isotherms predicted using the fitted parameters are very close to the measured data, which demonstrated the validation of our model.
We also probed the differences of original gas-in-place (OGIP) and production performance estimated using low-pressure adsorption isotherms (always characterized using the Langmuir adsorption isotherms) and high-pressure adsorption isotherms. The great derivations suggest that reliable adsorption isotherms of shale gas under reservoir conditions are very essential. Our proposed dual-site SLD model provides an alternative method to predict shale gas adsorption isotherms under reservoir conditions using low-pressure experimental measurement.
In hydraulic fracturing of unconventional reservoirs, the stimulation fluid is injected at a different temperature than initial reservoir temperature. The dynamic temperature profile of stimulation fluid during the treatment can provide critical information for fracturing design. In this work, an analytical solution to model the stimulation fluid temperature profile during hydraulic fracturing is presented. This analytical model is derived from the energy balance equation for fracture system coupled with a fracture propagation and fluid leak-off model. The procedure to obtain this analytical solution from the governing equation involve Method of Characteristics with valid assumptions. Several important features of the treatment have been preserved, including dynamic fluid leak-off and stimulation fluid velocity inside the fracture. Simple procedures to apply this solution are presented, which provide a convenient way to relate the warm-back temperature profile to the fracture, reservoir, and fluid properties.
The results of the analytical model are presented in terms of the temporal temperature variation inside the fracture. These results are compared and validated in multiple cases with numerical results from commercially available simulation software, as well as simulation results reported in the literature. We identify and analyze the major mechanisms contributing to the temperature signal, which involves the conduction to surrounding stimulated reservoir volume and convection associated with fluid leak-off and varying fluid velocity inside the fracture. The dynamic temperature profile for individual fracture and associated stimulated zone are strong functions of fracture and fluid properties, which include the leak-off coefficient, fracture pore volume, heat transfer coefficient between fracture and stimulated reservoir volume, and density and specific heat of the stimulation fluid. The effect of these factors on temperature distribution is investigated in the sensitivity analysis, which produces several dimensional parameters from this analytical solution. Various types of fracturing treatment design are applied to the developed solution to show its feasibility.
Despite many previous numerical studies on the same issue, this analytical solution brings direct insight into physics behind the stimulation fluid temperature profiles including fracture propagation and fluid leak-off. Besides multiple applications mentioned above, this work can be used as a theoretical basis for a potential analytical approach to address the proceeding warm-back temperature analysis after hydraulic fracturing.
Unconventional shale reservoirs with high organic content and swelling clays may have a high affinity for uptake of carbon dioxide (CO2). The pore space and mineral surfaces that sorb/contain petroleum are also potential sorption sites for CO2 and could become available for CO2 uptake once the reservoir is produced and depressurized. Understanding how shales interact with CO2 is important for enhanced resource recovery in the near term, and potential geologic carbon sequestration in the long term.
A series of experiments are discussed where fractured Bakken and Marcellus shale samples were exposed to CO2 at in-situ conditions for extended periods, a computed tomography scanner was used to visualize changes in structure, and simultaneously the fracture permeability was recorded. These measurements allowed for the correlation of hydro-mechanical changes in the fracture which are inferred to be a result of matrix swell and aperture closure.
A detailed examination studies in the literature to extend these results beyond two small samples is then presented. From the experimental CO2/shale interaction results and the growing body of literature on this topic several salient recommendations are presented to unify shale interaction study results into results that can be expanded beyond individual studies. This includes more rigorous characterization of sample constituents, maintaining micro-fabric of the samples, and enhanced control of initial shale water content.
Shale gas reservoirs are typically characterized by nanometer pore throats and very low permeability matrix requiring hydraulic fracturing stimulation of horizontal wells. Water is the main fluid used in hydraulic fracturing and a variety of chemicals are mixed with the water each for a different pur¬pose. Given the very low permeability of shales, very high pressure gradients are experienced in order to achieve economic production rates. The pressure drop results in vaporization of the water by the producing gas. The water being vaporized may be the formation water or a mixture of formation water with the stimulation fluids leaked off into the reservoir. The more pressure drop, the more water is vaporized into the gas and, as a result, the more salt dries out. With the nano-meter size pores in shale gas reservoirs and significant pressure gradients required for economic production, the shale gas reservoir permeability can be significantly damaged due to salt deposition. Salt dry-out can significantly reduce the productivity. Significant adverse impact of salt precipitation on well performance were observed in Marcellus shale play where brine salinities are relatively high.
In this paper, we model a shale gas reservoir having gas-water thermodynamic equilibrium, and investigate the effects of pressure decline induced by production on salt deposition and its consequences on permeability reduction. Reservoir properties representative of Marcellus shale gas play are considered for this study. Peng-Robinson EOS is used for reservoir fluids (methane and water) modeling. A horizontal well is placed in the middle of the reservoir and hydraulic fracturing is considered to create fractures around the well. Conducting hydraulic fracturing, the fractures are invaded by stimulation fluid. Gas is then produced through the fractures under constant rate constraint and water is allowed to be vaporized. Gas-water flash calculations are performed to evaluate the amount of vaporized water. Water vaporization causes the dissolved salt to be dried-out. Change in porosity and then permeability due to salt dry-out is calculated using Kozeny-Carman formula. Saturation of dried-out salt is evaluated and correlated as a time-dependent skin around the wellbore. The method is verified by modeling salt precipitation during CO2 injection into an aquifer and comparing the results with 1D analytical solutions found in the literature. Verifying the method, controlling the gas production from shale gas reservoirs by salt dry-out is elucidated.
As a kind of clean and potential energy resource, large quantities of gas hydrates have been proved to exist widely in the permafrost and in deep marine environments with high pressure and low temperature conditions favorable for their formation. Recently, how to develop and exploit the natural gas hydrate reservoir efficiently is one of the critical issues in the energy resource R&D in 21st century. However, gas hydrate dissociation is an endothermic process and the researches on the heat analysis of dissociation process is not enough and the results are often various.
Base on the research findings about the mechanism of gas hydrate dissociation, this study develops the mathematical model including mass conservation, energy conservation, chemical reaction kinetics and geo-mechanic equation and conducts the numerical simulation of gas hydrate dissociation in porous media. Three cases have been simulated in a cylindrical reactor adopting a vertical well in the center with the production well bottom pressure Pin=3.1, 2.5, 2.1 Mpa, respectively. All these cases are well matched with the experiment on the gas recovery factor of methane hydrate in porous of the Konno et al(2014) using the unique apparatus referred to as High-pressure Giant Unit for Methane-hydrate Analyses (HiGUMMA). Both of those Cases (1 to 3) are not dissociated completely in 200 mins because of the higher initial gas hydrate in the porous media. The experiment and the numerical simulation indicate obviously the existence of a “freezing stage” in case of Pin=2.1(Case 3). Four phases of gas, liquid, ice, and hydrate are observed to coexist in the reactor in this freezing stage. Compared with other two cases, the hydrate dissociation in Case 3 is promoted by three kinds of heat: the reservoir sensible heat Qr-sen, the conducted heat from the boundaries Qsur , and the latent heat from ice transition Qi-lat. Meanwhile, the discussions of various heat consumption in Case 3 are represented in detail. It could be conclude that Qr-sen and Qi-lat act as an essential role in inducing the rapid gas production rate in two peaks, which also means the faster dissociation rate of natural gas hydrate in porous media. Furthermore, the entropy of dissociation process is calculated specially according to the second law of thermodynamics. Based on above analysis, the rate of pressure reduction is optimized and the energy efficiency is enhanced. Although the formed ice may cause flow blockage for the gas and water, the released latent heat is still remarkably attractive for the fast hydrate dissociation. In this work, the basic reservoir unit of conceptual model is established and developed, which provides theoretical guidance for the exploitation of the actual hydrate reservoir in the future. Hereafter, the various reservoir conditions would be conducted to explore the transformation patterns of the energy exploitation and entropy.
Natural Gas Hydrate (NGH) widely distributed in marine sediments and permafrost areas has attracted global attentions as potential energy resources. Permeability is a critical parameter that influences the gas production potential from hydrate reservoirs. The hydrate saturation affects the characteristics of the porous media, which is also the key factor determining the permeability. In this study, the absolute permeability and the relative permeability of water were experimentally measured at varying hydrate saturations (0-0.347) in porous medium made of quartz sands and the porosity was 0.3. During permeability measurement, the steady flow and stable differential pressure were obtained under certain water injection rate. Hydrate saturations were controlled and calculated precisely based on the amount of injected and produced gas/water, the system pressure and temperature.
The result indicated that the water relative permeability reduced exponentially with the increase of methane hydrate saturation, and the reduction exponent value of 7.9 was obtained in the Masuda’s permeability model. In addition, hydrate with different saturation in porous media is stored with different forms, which exerts considerable influence on permeability as well. Therefore, a new permeability model based on the weighted combination of pro-filling and grain-coating model was proposed. The weight of pro-filling model is defined as the N-th power of hydrate saturation, and the weight of grain-coating model is defined as the N-th power of non-hydrate saturation. In this work, the calculated index N was 5.6. Compared with the Masuda’s model, the new model not only shows the relationship between permeability and hydrate saturation, but also reflects the aggregate performance of hydrate in porous media.
Hydraulic fracturing fluids (HFF’s) have been used for several decades to control mechanical, hydraulic, and geochemical behavior in unconventional reservoirs during stimulation. The interactions that occur in these environments during stimulation (hydrofracturing) are designed to prevent scaling, improve production, and prevent damage to formations. However, there is still uncertainty with regards to near fracture geochemical reactions and the evolution of the fluids temporally as they interact with the reservoir rock1.
Rock cores taken from the Marcellus shale, both in outcrop and from a production well, were exposed to simulated HFF’s and simulated formation brines (SFB). The fluids were designed based on regional averages obtained from operators. The tests were conducted at pressures and temperatures representative of regional reservoir conditions, 19.3 MPa and 71°C respectively, and with all fluids under a nitrogen atmosphere to limit free oxygen. Core samples were fractured and loaded with 40/70 US Silica White™ quartz proppant. Tests lasted approximately 96 hours with low flow rates to represent a shut-in period. Geochemical samples were taken daily and analyzed using Inductively Coupled Plasma Mass Spectrometry and Ion-Chromatography. Imaging of the cores was done before exposure using Computed Tomography (CT) scanning and after exposure using CT, Scanning Electron Microscopy (SEM), and Raman Spectrometry.
Control samples, using deionized water as the flow medium, exhibited little to no change in the rock core or in the effluent of the system, dominated by Ca and SO4. Experiments using only SFB showed minor increases in major elemental chemistry in the effluent, consistent with minor dissolution or entrainment of free particles, and minor precipitation of barite/calcite on the fracture surface observed with SEM. HFF chemicals, without SFB, resulted in increases in most elemental constituents in the effluent, with few exceptions including barium which showed an increase followed by a major decrease in concentration. The rock core exhibited significant reaction in CT images and only minor traces of barite/calcite precipitation. The combination of HFF and SFB resulted in large increases in most metal constituents in the effluent during the reaction with the core and significant alteration of the rock matrix adjacent to the fracture. However, this mixture resulted a continual decrease in Ba and SO4 during the experiment, signaling potential deposition of these constituents.
In general, experiments indicated minor pyrite oxidation/dissolution, dissolution of carbonates, and minor precipitation of barite. The degree of precipitation was not of the magnitude observed in Paukert et al. (2017), but there is evidence that carrier fluid composition (>90% of the total volume) is an important consideration in precipitation within unconventional systems and may provide the nucleation surfaces for precipitation. Work continues analyzing precipitates, fracture surfaces, and base-fluid importance in precipitation.
A multi-relaxation-time lattice Boltzmann (LB) model for nanoscale liquid flow is developed to investigate the liquid flow characteristics in nanoporous media. The slip length and effective viscosity obtained from molecular dynamics (MD) simulations are adopted to account for the nanoscale effect. First, the LB model for water flow in nanopores is built and water flow characteristics in nanoporous media are investigated. The results show that: (1) the nanoscale effect can either increase or decrease the water flux in nanoporous media, depending on the fluid-solid interaction force; (2) the nanoscale effect impacts the velocity distribution in porous media, making it more uniform in hydrophobic porous media while more heterogeneous in hydrophilic porous media; (3) the end effect caused by the bending of streamlines plays a significant role in water flow in nanoporous media, and neglecting the end effect can greatly overestimate liquid flow ability; and (4) the pore structure also has significant influence on water flow in nanoporous media. With the increase of specific interfacial length, the nanoscale effect increases. In addition, the LB model for oil (octane) flow in quartz nanopores is also established by incorporating the MD simulation results [1]. Oil flow simulation in quartz nanoporous media shows that the conclusions obtained for water flow are also applicable for oil flow.
Heat conduction in granular porous media is a phenomenon that is relevant to a broad spectrum of problems in science and engineering disciplines including physical, earth, and biological sciences, to name a few. Effective thermal conductivity in granular porous media is a function of morphological features of the medium such as grain shape, grain size, and geometrical structure. Thermal contact resistance can also affect heat conduction due to topological features such as the surface profile of the grain. Furthermore, the compressive pressure and presence of different fluids in the pore space along with partial saturation can also dictate the nature of the effective thermal properties.
To study the effect of all these factors on the effective thermal conductivity of granular porous medium, we simulate heat conduction by developing a two-dimensional, parallel and thermal Lattice Boltzmann Method (T-LBM) simulator using existing open source libraries. We use this simulator on a digitally reconstructed, two-dimensional granular porous medium that is generated with an existing packing algorithm. We then conduct a progressive investigation by first, introducing thermal contrast resistance as surface roughness on the grains and study its effect on thermal conductivity. Second, we introduce thermal anisotropy in the system by inclusion of elliptical grain packing in the medium. Third, we investigate the effect of partial saturation of water and air in pore space. We use an LBM single component multiphase model to simulate phase segregation in the pore space. We also incorporate elastic deformation of grains based on an existing model, which depicts the surface topology of grains as a self-affine fractal function. This elastic deformation is a function of Young's modulus of the grains and the external compressing pressure.
Based on our investigation, we observe that thermal contact resistance due to the surface roughness of grains reduces effective thermal conductivity. Elliptical packing of grains, manifest thermal anisotropy in the system and causes local heat flux deviations especially when the grain orientation angle changes. External compressive pressures cause elastic deformation of the grain surface and enhance the thermal conductivity of grains with lower Young’s modulus. Introducing partial saturations of water and air in the pore space offsets the effective contribution in heat conduction from the grains as well as the effect of compressing pressure. All of these observations are further accentuated if the thermal contrast ratio of the granular porous medium is changed. We also compare results for selected observations for consistency. A qualitative agreement is obtained with the existing experimental data.
Proppants are small, granular additives used in hydraulic fracturing to keep induced fractures open and permeable after the reservoir pressure is lowered; typically sand is used. These materials are designed to resist the closure force across a fracture face and allow fluid to migrate out of the system. While the simple mechanical support of the proppant keeping a fracture open is well understood, the interplay between hydraulic fracturing additives, rock strength, and proppants is still lacking.
Experiments were conducted on Marcellus Shale samples that were cored and saw cut to create an artificial, uniform fracture. 40/70 US Silica White™ proppant was loaded into the fracture. Cores were then subjected to a confining pressure of 20.7 MPa at a temperature of 71°C. Each core was then exposed to either deionized water, simulated hydraulic fracturing fluids, or air and allow to react for a period of 5 days. Each core was scanned using computed tomography (CT) before and after the test to evaluate for proppant embedment. Scanning Electron Microscopy (SEM) was used to take high resolution images of the fracture surfaces after the experiment to evaluate embedment features that occur below the CT scan resolution.
Fractured cores that were filled with air or water did not exhibit mechanical signs of change in the CT or SEM. Rock integrity was not compromised and the proppants performed as designed with very little in the way of crushed proppant or rock material. When the hydraulic fracturing components were added, there were clear signs of indentation on the fracture surface.
Natural gas hydrate is an ice like crystalline compound with a cage structure under high pressure and low temperature. The hydrate will decompose when the pressure is lower than the hydrate equilibrium pressure, so the pressure propagation rule is different from the porous medium without hydrate. Based on the theoretical analysis method of fluid mechanics in porous medium and considering the influence of hydrate existence on the law of influent in porous media, a seepage flow model of hydrate in porous media is established. Analytical solution of pressure distribution in porous media when the hydrates are decomposed can be obtained. Using the numerical simulation software CMG to simulate Gas Production from methane hydrate reservoir by depressurization in Shenhu Area . The law of pressure propagation in the process of decomposition by depressurization mining is obtained, and the results obtained by theoretical analysis are verified.
A domain decomposition algorithm is introduced to couple non isothermal compositional gas liquid Darcy and free gas flow and transport. At each time step, our algorithm solves iteratively the nonlinear system coupling the nonisothermal compositional Darcy flow in the porous medium, the RANS gas flow in the free-flow domain, and the transport of the species and of energy in the free-flow domain.In order to speed up the convergence of the algorithm, the transmission conditions at the interface are replaced by Robin type boundary conditions.The Robin coefficients are obtained from a diagonal approximation of the Dirichlet to Neumann operator related to a simplified model in the neighbouring subdomain.The efficiency of our domain decomposition algorithm is assessed on several test cases focusing on the modeling of the mass and energy exchanges at the interface between the geological formation and the ventilation galleries of geological radioactive waste disposal.
Objectives/Scope:
Foam can improve sweep efficiency in gas-injection enhanced oil recovery. Surfactant-alternating-gas (SAG) is a favored method of foam injection due to injectivity and operational considerations. Laboratory data indicate that foam can be non-Newtonian in the high-quality regime, and therefore during gas injection in a SAG process. We investigate the implications of this finding for mobility control and injectivity, by extending fractional-flow theory to gas injection in a non-Newtonian SAG process in radial flow.
Methods, Procedures, Process:
Non-Newtonian behavior in the high-quality regime means the limiting water saturation for foam stability varies as superficial velocity decreases with radial distance from the well. We discretize the domain radially and perform Buckley-Leverett analysis on each ring; solution characteristics are of constant foam quality. For the first time, we show the implications of this behavior for mobility control at the displacement front as well as injectivity. We base the foam-model parameters and the extent of non-Newtonian behavior on laboratory data in the absence of oil. We compare results to mobilities determined by conventional simulation, where grid resolution is limited.
Results, Observations, Conclusions:
For shear-thinning foam, mobility control improves as the foam front propagates from the well, but injectivity declines somewhat with time. The change of mobility ratio at the front can be considerable, given the huge velocity difference between the wellbore and further out. This change is not simply that measured at steady state at fixed foam quality in the laboratory, however, because the foam front in a non-Newtonian SAG process does not propagate at fixed foam quality. Injectivity benefits from the increased mobility of shear-thinning foam near the well. The foam front, which maintains a constant dimensionless velocity for Newtonian foam, decelerates somewhat with time for shear-thinning foam. For shear-thickening foam, mobility control deteriorates as the foam front advances, though injectivity improves somewhat with time. Overall, however, injectivity suffers from reduced foam mobility at high superficial velocity near the well. The foam front accelerates somewhat with time. Overall, injectivity is a complex result of changing saturations and varying superficial velocities very near the well. Conventional simulators cannot adequately represent these effects, or estimate injectivity accurately, in the absence of exceptional grid resolution near the injection well.
Novel/Additive Information:
For the first time we extrapolate laboratory steady-state foam data for non-Newtonian foam to investigate the implications for injectivity and mobility control in gas injection in SAG in the field.
Foam injection into the subsurface is generally performed to improve gas mobility control during enhanced-oil recovery (EOR) and contaminated site remediation (Lake et al., 1989; Hirasaki et al., 2000; Mulligan et al., 2006). Several experiments have been conducted to study the foam generation mechanism at both the pore and continuum scales (Kovscek et al., 1994; Kam et al., 2003; Gauteplass et al., 2015; Prigiobbe et al., 2016). Pore scale experiments allow to understand the mechanism of bubble formation with a potential to help formulating constitutive equations for foam flow models improving their accuracy of prediction. However, pore scale studies have not been used to formulate foam generation rate, yet. Here, we present an experimental and modeling work on foam generation mechanism with a porous medium chip. Systematic tests at different flow conditions were performed using various chemicals to stabilize the foam, such as the surfactant, nanoparticles, and a combination of them. The pressure drop and the foam texture were monitored continuously using a pressure transducer and a high-speed high-resolution camera. We observed that to generate a foam in the presence of nanoparticles requires larger energy than when the surfactant is used to stabilize the lamellae. Possibility due to the larger critical capillary pressure for bubble rupture (Pc*) that can be reached in the presence of nanoparticles. Upon image processing, the results show that the generation rate and, therefore, the total number of bubbles increase with the injection rate, creating a more uniform bubble size distribution. We observed that nearby the gas injection the controlling mechanism of the bubble formation is snap-off, while afar from that lamella division dominates.
Although foams are known for effectively reducing gas mobility and enhancing oil recovery in many field applications, it is still not clear how far the injected fine-textured foams will propagate into the reservoirs. Lacking such a knowledge makes the design of foam field treatments difficult and often unreliable. The purpose of this study is to investigate CO2 foam propagation distance as a function of injection foam quality and injection total rate by using bubble population balance model. This study is believed to cover the steps needed from the pore-scale to field-scale events.
In order to meet the purpose, this study performs the following tasks: (i) fitting bubble population balance model to lab coreflood experiments and determining model parameters; (ii) establishing the mathematical framework to determine foam propagation distance during EOR processes; and (iii) characterizing foam propagation distance at different injection strategies. The laboratory data consists of three foam states (weak-foam, strong-foam, and intermediate states) as well as two different flow regimes (high-quality and low-quality regimes) of the strong-foam state.
The mobilization pressure gradient is one of the key model parameters to distinguish gaseous CO2 foams from supercritical CO2 foams. It is because, the mobilization pressure gradient being proportional to the interfacial tension, supercritical (or dense) CO2 foams exhibit much lower mobilization pressure gradient compared to gaseous CO2 foams, often with a couple of orders of magnitude difference.
The results show that the presence of three different foam states as well as two different strong-foam flow regimes (high-quality and low-quality regimes) plays a key role in model fit and field-scale propagation prediction. More specifically, this study finds that supercritical CO2 foams can propagate a few hundreds of feet easily, which is a few orders of magnitude higher than gaseous CO2 foams. For dry foams (or, strong foams in the high-quality regime), higher injection gas fractions result in shorter foam propagation distance, while for wet foams (or, strong foams in the low-quality regime) the propagation distance is not really sensitive to injection gas fractions. In addition, the higher injection rates (or pressures), the farther foams propagate – such an effect is shown to be much more pronounced for dry foams.
Abstract
Random fractures widely exist in water/oil reservoirs, soils etc. Study of the permeability of the fractured networks has been one of focuses in the area of mass transfer in the past decades. Generally, the fractures in scale reservoirs distribute randomly and have statistical self-similarity and fractal characteristic. In this paper, the permeability model for gas flow in the fractured networks in shale reservoirs is derived based on the fractal geometry theory with gaseous slip flow included. The validity of the proposed model is verified by comparisons between the model predictions and experimental data, and the parametric study is also performed in detail. The present results show that the proposed permeability model can reveal more mechanisms of seepage characteristics in the media than the traditional models.
Pore structure of large scale porous limestone reservoir with strong heterogeneity is very complex,so it is difficult to evaluate its pore structure of Mishrif Formation of W oilfield in Iraq. Based on thin section observation,porosity and permeability test and mercury injection capillary pressure test,fractal theory was applied to quantitative pore structure evaluation,and the pore fractal dimension criterion for reservoir type classification was established. There are two types of reservoir pore structure fractal characteristics. Some samples called“single segment”perform obvious fractal character overall. Others called“multiple segments”have distinct large pore throat system and small pore throat system which perform unique fractal characters respectively while have no uniform fractal character overall. The complexity and heterogeneity of pore structure of porous limestone can be reflected by fractal dimension,the greater the fractal dimension,the more complex pore structure,and the more conspicuous segmental character in the relationship between capillary pressure and water saturation,the stronger the heterogeneity. The samples were classified based on the fractal dimension combined with porosity and permeability distribution of the samples. The majority of type Ⅰ-Ⅱ and type Ⅲ-Ⅳrespectively corresponded to“multiple segments”and“single segment”fractal characteristics. It has an important guiding significance for the quantitative evaluation of pore structure to similar carbonate reservoir.
Fractal model of gas diffusion coefficient is derived for porous nanofibers, which are assumed to be composed of a bundle of tortuous capillaries whose pore size distribution and roughness of wall surfaces of capillaries follow the fractal scaling laws. The analytical expression for gas relative diffusion coefficient is a function of the relative roughness, fiber radius and microstructural parameters (porosity, the fractal dimension for pore size distribution and tortuosity, the maximum and minimum pore diameter and the characteristic length). The proposed fractal model is validated by comparison with available experimental data and correlations. At the same time, the effect of microstructural parameters of porous nanofibers on gas diffusion has been studied in detailed. The results show that roughness of wall surfaces of capillaries in porous nanofibers should not be neglected. It is believed that the current work can reveal gas diffusion mechanism in porous nanofibers and may be applied in other porous materials.
Zhang Yongfeng, Jiang Yongxu, Lu Guoqiang
Exploration and Development Research Institute of Daqing Oilfield Company Ltd.,Daqing,China
ABSTRACT: This papers identity the coal structure in Daqing exploration area, and to discuss types of pore in different basins. Coal is a complicated porous medium. Adsorbability and permeability of its pore structure for coalbed methane (CBM) has drawn extensive attention increasingly. Porosity of coal enables the coal reservoir to store gas and allows CBM to desorb, diffuse and percolate. For this reason, it is of significance to study the pore structure characteristics for exploration and development of CBM as well as evaluation of its mineability. In the three basins of Daqing exploration area, the coal basins include Hailar basin in the west and Jixi and Higang basins in the Sanjiang region in the east as representatives. The coal from the Huhehu depression belongs to lignite. A great deal of intact plant tissue pores can be observed in the SEM. The cell cavities deform to different extents because of compaction effect, but they arranges in a uniform direction with similar shapes, which indicates feature of plant tissues. The fine stratification and fissures can be seen in locality. Its fractures occur in interlaminations, but the fractures connecting the pores are rare, which contribute less to the permeability of coal. So, this is to the disadvantage of migration and deposit of CBM. The Huhehu depression is the one where original texture coal mainly develops. The coal in Jixi and Hegang Basins develops better between gas coal and coking coal. Though parts of plant tissue pores remain in the coal, the pores are filled nearly with minerals. A great many blowholes and emposieu exist in the coal with lithification. Some microstructures such as friction surfaces occur with anabatic deformation. The original texture is destroyed in the tectonic coal so that the pore structure becomes complicated. A large number of micro-fractures and shrinkage joints form connected bridges between pores, which improves seepage capability among coal pores to some extent. There are three types of low temperature nitrogen adsorption cures in Daqing exploration area, which represents different types of pore structure. There is type I in Huhehu drepression. The type I indicates that the coal reservoir has a great many open breathing holes and a few of non-air holes whose one end is closed. The pore size is distributed in twin peaks. Type II and type III can be find in Jixi and Hegang basins. Type II indicates their pores are in composite with multi-pore form and contain ink bottle pores and non- breathing pores whose one end is closed. Type III adsorption and desorption curves have distinct hysteresis loop. The pore volume exists two peaks, but pores at a size of 3—4nm have greatest specific surface area. The micro-pore becomes a greatest contributor to specific surface area. Occurrence of a large number of ink bottle micro-pores is a major reason for difference of adsorption capacity.
KEYWORD: Daqing exploration area; Pores characteristics; Low temperature nitrogen adsorption method; Pore structure
Dual-porosity media widely exist in natural reservoirs and have been received much attention in heat and mass transfer. Due to multiplicative cascade effects, the microstructure might be disordered and complicated, with fractures/pores scale-invariantly distributed. In this study, we briefly introduce the concept of General Fractal Topography proposed recently which not only reduces modeling complexity significantly but also admits scaling objects and fractal behaviors as complex as possible. And then, we developed an algorithm to model fractal fracture-pore porous media according to the scaling-invariant topography based on Voronoi tessellations. The original complexity of the fractures and pores distribution is wrapped in the determined phase of scaling object, while the behavior complexity is defined by the fractal topography. Our investigation provides an open framework to unify the definition and modeling of pore, fracture network, and dual-porosity media.
Fluid flow though geologic fractured/porous media tends to become non-Dacian as a result of the competition between viscous and inertial forces and the effect of pore geometry variation. The Forchheimer equation has been widely shown to apply in these situations, in which the coefficient of viscous permeability (kv) is largely predictable, but this is not so for the coefficient of inertial permeability (ki). Synthesizing thousands of pore-scale flow models and field and laboratory observations, we show that ki can be predicted from kv via the equation ki~(kv)^(3/2) across twelve and sixteen orders of magnitude in ki and kv, respectively. kv is thus sufficient for predicting flow across viscous-to-inertial regimes for most geologic media.
Subsurface flow processes involving non-Newtonian fluids play a major role in many engineering applications, from in-situ remediation to enhanced oil recovery. The fluids of interest in such applications (f.e., polymers in remediation) often present shear-thinning properties, i.e., their viscosity decreases as a function of the local shear rate. We investigate how fracture wall roughness impacts the flow of a shear-thinning fluid. Numerical simulations of flow in 3D geological fractures are carried out by solving a modified Navier-Stokes equation incorporating the Carreau viscous-shear model. The numerical fractures consist of two isotropic self-affine surfaces which are correlated with each other above a characteristic scale (Meheust et al, 2003). Perfect plastic closing is assumed when the surfaces are in contact. The statistical parameters describing a fracture are the standard deviation of the wall roughness, the mean aperture, the correlation length, and the fracture length, the Hurst exponent being fixed (equal to 0.8). The objective is to investigate how varying the correlation length impacts the flow behavior, for different degrees of closure, and how this behavior diverges from what is known for Newtonian fluids. The results from the 3D simulations are also compared to 2D simulations based on the lubrication theory, which we have developed as an extension of the Reynolds equation for Newtonian fluids. These 2D simulations run orders of magnitude faster, which allows considering a significant statistics oSf fractures of identical statistical parameters, and therefore draw general conclusions despite the large stochasticity of the media.
Modeling DNAPL source zone plume evolution using traditional flow and transport models is a computationally intensive process that requires specification of a large number of material properties and hydrologic/chemical parameters. Given its computational burden, Monte Carlo simulation using such models is particularly ill-suited for uncertainty assessment and/or subsurface sampling optimization in real field applications. In this work, we present an innovative approach that couples machine learning, adjoint sensitivity theory, and statistical analysis to optimize borehole sampling for quantification of the evolution of down gradient flux-averaged concentration.
Probabilistic models, based on discriminative random fields (DRF), are first employed to synthesize stochastic realizations of a subsurface source zone consistent with known, limited, site characterization data. Using a suite of full scale simulations as training data, a statistical model is developed to predict the spatial distribution and uncertainty associated with key features (i.e. permeability and sequestered contamination [aqueous, sorbed, and NAPL]) that control plume evolution and persistence. Given an initial spatial distribution of contaminant mass, conditioned on measured field data, the adjoint state sensitivity method is then employed to quantify the importance of local system properties on down gradient flux-averaged concentration. The optimal sampling design problem is then addressed using first-order second-moment uncertainty analysis. In the decision process, the costs of additional measurements are justified by a sufficient decrease in the uncertainty, selecting measurements associated with the highest expected worth.
The utility of this probabilistic statistical modeling approach is demonstrated using numerically generated, two-dimensional, heterogeneous DNAPL source zones. Results reveal that down gradient flux-averaged concentration sensitivities to initial contaminant mass compartments are strongly affected by local permeability values. In addition, initial aqueous and sorbed concentrations and their corresponding variances have a major impact on down gradient flux-averaged concentration at early times, while the influence of initial NAPL saturation persists for a longer period. This innovative sampling strategy, coupling sensitivity analysis and uncertainty quantification, shows promise for enhancement of our ability to guide characterization of source zones under realistic field conditions.
The inverse problem of parameter identification consists in the optimal determination of model parameters using water-level observations. We are concerned with the estimation of the transmissivity and storativity in a confined aquifer in transient conditions. One of the approach used to solve this problem, is called the Differential System (DS) method. It is based on the solution of a Cauchy problem. Because of the numerical derivatives need to be calculated, the issue with this method is when noisy data are used. In this work, we propose an improvement to the DS method by using a predictor-corrector scheme. For the predictor part, we consider an estimator obtained from the Bayesian formulation problem. The study case presented here is a synthetic but realistic aquifer; the model was chosen in order to check how the estimation is made in complicated conditions.
We present a theoretical investigation on the processes underpinning the reduced longitudinal spreading documented in stable variable density flows, as opposed to constant density settings, within heterogeneous porous media. We do so by decomposing velocity and pressure in terms of stationary and dynamic components. The former corresponds to the solution of the constant density flow problem, while the latter accounts for the effects induced by density variability. We focus on a stable flow configuration and analyze the longitudinal spread of saltwater injected from the bottom of a column formed by a heterogeneous porous medium initially fully saturated by freshwater. We adopt a perturbation expansion approach and derive the equations satisfied by section-averaged concentrations and their ensemble mean values. These formulations are respectively characterized by a single realization and an ensemble dispersive flux, which we determine through appropriate closure equations. The latter are solved via semi-analytical and numerical approaches. Our formulations and associated results enable us to discriminate the relative impact on the density-driven solute displacement of (a) covariance of the permeability of the porous medium, (b) cross-covariance between permeability and concentration, which is in turn linked to the coupling of flow and transport problems, and (c) cross-covariance between the dynamic and stationary velocities.
We present a numerical analysis of fluid phase distributions and relative permeabilities obtained from direct pore-scale simulations of two-phase flow through a real pore space with diverse conditions of wettability. Exploring the effects of wettability on the fluid behaviors within porous media is of fundamental relevance for a variety of engineering as well as environmental applications, including, e.g., conventional and unconventional hydrocarbon extraction, along with CO2 storage in subsurface systems. The pore space investigated is a three-dimensional limestone rock reconstructed via X-ray micro-tomography at 2 microns voxel size resolution. The one millimeter-scale rock sample is associated to a connected porosity Phi = 0.128 and an absolute permeability k = 357 mD. We employ a simulation procedure mimicking steady-state protocols implemented to estimate multiphase relative permeability by laboratory-scale experiments. We first simulate a primary-drainage (oil injection) process until steady-state conditions are reached. Then, we simulate a series of secondary-imbibition (water injection) processes by injecting diverse amount of water/oil volumes until steady-state conditions are achieved. The ratios between oil and water densities and viscosities are set equal to 0.78 and 2.87, respectively. In all our simulations the capillary number is constant and equal to 1e-5. Flow field and fluid phase distributions are calculated within the explicit three-dimensional pore space using a finite volume-based solver and implementing the so-called Volume-Of-Fluid technique. Capillary end effects that may arise at the inlet and outlet boundaries are removed by implementing ad-hoc periodic conditions (see ref [1]). In this context, adjustment of fluid-phase saturations is achieved through local injection of one of the two fluids. The total fluid flow rate is indirectly set through a pressure jump across the periodic inlet and outlet boundaries. Simulations were achieved using HPC resources at CINECA computing center [2], within the LISA-Prod Project HPL13P71U0.
In order to assess the impact of wettability conditions on the resulting two-phase relative permeability curves, multiphase flow simulations are performed by using diverse contact angle values at the solid-fluid boundaries resulting in water-, neutral- and oil-wet conditions. The numerical errors associated with both geometry and pore space discretization are analyzed by performing a sensitivity analysis reconstructing the pore-space with diverse spatial resolution scales. We observe that errors in the geometrical representation have an impact on the averaged capillary pressure as well as on residual fluid saturations, while errors originating from the pore-space discretization impact water and oil relative permeabilities. Our results document that varying the contact angle has a strong influence on residual oil and water saturations, relative permeability curves and spatial distribution of fluid phases observed at the end of the primary-drainage and secondary-imbibition processes.
Flow and transport in porous media is encountered in many industrial and hydrogeological applications, such as hydrogen fuel cells, inkjet printing, hydrocarbon exploration, and subsurface remediation. The relevant study domain can cross multiple scales from a few nanometers to hundreds of kilometers. Therefore, in porous media research, the upscaling and multiscale techniques have been widely used to bridge the hierarchical scales. Recently, with the increase of powerful computational resources and the improvement of characterization methods of pore structures, image-based modeling is attracting much attention. It sheds light on fundaments of pore-scale flow dynamics, and assists in the upscaling of porous media models. In this work, we will discuss the current challenges of image-based modeling of single-phase reactive transport and two-phase flow in various porous materials (e.g., paper and shale). Then, we will present our PDMS-based micromodel benchmark experiments of two-phase flow in porous media. The obtained data will be used to calibrate and validate pore-scale models.
The pore blocking caused by solid particles migration is the major reason to formation damage. In order to further describe the solid particles blocking process, the realistic pore network model is established based on the results of micro-CT scanning. At the same time, the granularity distribution model is generated according to the solid phase particles size distribution. Then the “blocking volume” is introduced to judge whether or not pore blocking happen. Based on the conception of “blocking volume”, pore element is generated and all these characteristic parameters of pore element are calculated such as the flow distribution, pore size, particles granularity, blocking volume. According to the judgement standard, the intrusion ratio of solid particles, the sediment ratio of solid particles, the blocking ratio of solid particles, the sediment depth and the blocking depth are all obtained. In addition to this, the parameter sensitivity analyses of influence factors are taken. Through the study of pore blocking simulation, the aim to predict the probable formation damage caused by solid particles based on the structure of core and particles parameters have been achieved. It provides
The phenomena of tight gas reservoirs with super-normally saturated water, which exhibit the unique combination of very high initial water saturation combined with low to very low permeability (permeability of less than 0.1 mD) exist extensively in a number of regional sedimentary basins. Traditionally, multiphase flow of natural gas and connate water is evaluated in laboratory by conducting flow experiments. However, due to the lack of sufficient insight vision, this traditional method only reflects the macroscopic phenomena but fails to reveal some important microscopic behaviors inside the cores of different type of reservoir during multiphase flow process. In this paper, a new method, which combined the multiphase flow experiment and nuclear magnetic resonance experiment, was introduced to effectively observe and characterize the water flow to various multiphase flow scenarios in different types of rocks. First of all, a series of core displacement experiments were carried out. Then, based on the displacement experiments, the pore size distribution of rock samples is converted by NMR T2 spectrum by T2 spectrum experiment. The distribution of the remaining water under different water saturation of different cores was analyzed and the water production mechanisms of tight high water saturation reservoir were revealed. Finally, we prove the relationship between gas & water production and pore structure using effective flow formula. Through theoretical analysis and calculation, the physical significance and the value of two main parameters in gas production formula are made clear. This experiment and evaluation methods provided a valuable tool in the evaluation of whether given pay zones in an ultra-tight gas reservoir situation are worthy of completion and what kind of pressure should be provided for production. It also provides information on how this new method can be used in the exploitation of this ever increasing area of tight gas reservoir production.
Physics of two-phase flows in heterogeneous natural rocks plays an important role in many applications, such as carbon sequestration in deep saline reservoirs and recovery of oil from hydrocarbon reservoirs. Although pore-scale models are used to compute macroscopic average properties required in field-scale simulators, most work is limited to small sample size. There is a need for pore-scale models that can accurately represent the 3D complex pore structure and heterogeneity of real media. Pore network modeling (PNM) simplifies the geometry and flow equations at pore-scale, but can provide characteristic curves in capillary-dominated systems on fairly large samples with huge saving on computational costs compare to direct numerical simulation methods. However, there are limitations for attaining a large representative pore network for heterogeneous cores such as technical limits on scanning size to discern void space and computational limits on pore network extraction methods. To address this issue, we propose a novel pore network stitching method to provide large-enough representative pore network for a core.
In this study, we use industrial (as core-scale) and micro-CT (as pore-scale) scans of actual reservoir rock samples to characterize the pore structure of a core. Few signature parts of the core are selected from industrial scans, and their micro-CT scans are taken. Equivalent 3D pore network of each signature part is extracted by applying maximal ball pore network extraction algorithm. The space between signature networks is filled by using stochastic random network generator that uses statistics (radius, shape factor, connection number, and length of pore elements) of all signature networks and a layered stitching method that glues network pieces. The outcome is a large pore network that can be used in a fast quasi-static PNM solver to obtain absolute permeability, relative permeability and capillary pressure curves. We have tested the developed method on various generated and extracted networks by 1D stitching direction, and it will be improved and extended to 2D and 3D stitching directions.
This work is primarily supported as part of the Center for Geologic Storage of CO2, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science and partially supported by the International Institute for Carbon-Neutral Energy Research (WPI-I2CNER) based at Kyushu University, Japan.
Target formations for large-scale CO2 sequestration are often saturated with brines that contain dissolved methane and other light hydrocarbons. When CO2 is injected in such deep formation, non-trivial phase behavior may result in methane exsolving from the brine and forming a free gas phase. Because methane has a lower viscosity than the (generally supercritical) CO2, this methane-rich gas is swept up ahead of the CO2 front. On the one hand, such a `bank’ of methane may provide an early tracer warning of the approaching CO2 plume, e.g. in observation wells. On the other hand, the emergence of gaseous methane poses risks of releasing a potent (greenhouse) gas contaminant in overlying groundwater and potentially the atmosphere if the formation integrity is compromised by (open) fractures, faults, or leaky wells.
The transport of methane in water-saturated formations is also important in the context of natural-gas production from deep reservoirs. If a producing well is compromised, e.g., at shallower depths, leaking stray natural-gas may contaminate groundwater resources. Whether this contamination occurs in a small radius around the well (due to buoyancy) or travels significant distances laterally before contaminating groundwater wells depends strongly on the formation heterogeneity, notably fractures.
The modeling of these important processes is complicated by 1) strong heterogeneity in fluvial target formations for CO2 storage and fractured groundwater aquifers, and 2) the complicated phase behavior of mixtures of water and hydrocarbons. The latter can be modeled accurately by the cubic-plus-association (CPA) equation of state (EOS), which takes into account the polar nature of water molecules, its self-association, and the polar-induced cross-association between water, CO2, and methane molecules (as well as compressibility of the aqueous phase). While accurate, the CPA EOS is highly non-linear and computationally expensive. In this work, we develop new efficient algorithms to adopt the CPA EOS for large-scale simulations. Flow and transport are modeled by the mixed hybrid and discontinuous Galerkin methods, respectively, and discrete fractures are incorporated through a cross-flow equilibrium approach.
Simulation results are presented for 1) the Cranfield large-volume CO2 storage pilot project, and 2) for lateral migration of stray methane leaking from a compromised natural-gas well into shallow fractured groundwater aquifers for conditions representative of those overlying the Barnett formation in Texas.
Accurate modeling and robust computation of the phase behavior is essential for optimal design and cost-effective operations in petroleum reservoirs as well as in petroleum processing plants, where we need to understand the fluid flow of partially miscible multi-component multi-phase mixture in free spaces or in porous media. Phase behavior calculation of fluid mixture consists of stability analysis and flash equilibrium calculation. The goal of the NVT stability analysis is to determine whether a phase is stable at specified volume, temperature, and mole numbers. If it is not stable, the NVT flash equilibrium calculation is to establish the composition and amount of each stable phase after phase splitting. Conventional algorithms for stability analysis and flash calculation are based on fixed-point iteration, Newton-type iteration, or their combination, and the convergence has never been guaranteed. In this work, we propose an energy-stable iterative method for NVT stability analysis and NVT flash equilibrium calculation. We consider fluid mixture modeled by the Peng-Robinson equation of state, and our proposed algorithm is an iterative algorithm motivated from the dynamics of two-phase fluid system with Fick's law of diffusion for multi-component fluids. The proposed iterative procedure is proven to be energy stable under certain conditions. Numerical examples are tested to demonstrate efficiency and robustness of the proposed method. We also discuss the extension of the algorithm to NVT flash in multiple spatial dimensions, which can be used to model the interface of nonzero thickness between two phases.
Barometric pressure variations are often one of the main drivers of gas transport in fractured rock, a process that is referred to as barometric pumping. Barometric pressure variations are complex, multi-frequency signals influenced by latitude, weather, elevation, lunar phase, time of year, and diurnal and semi-diurnal earth tides. However, our results indicate that it is often a subset of the pressure frequencies that lead to the vast majority of transport while the majority of frequencies result in minor or even insignificant transport. Identifying the dominant pressure frequencies for transport allow us to more simply and effectively characterize the potential for gas transport to the surface at different geographic locations. We will present barometric pressure decomposition analyses on gas transport in fractured rock.
A deep geologic repository (DGR) for low- and intermediate-level radioactive waste has been proposed at the Bruce nuclear complex on the eastern flank of the Michigan Basin in southeastern Ontario, Canada. The proposed location for the repository is at a depth of ~680 m, in the middle of a ~450 m-thick sequence of Ordovician-aged shale and limestone with extremely low porosity and permeability, which makes fluid flow and mass transport processes very slow. Significant underpressure exists in these rocks, and questions have been raised about whether gas phase methane is present and how it relates to the generation and persistence of the underpressure here, as well as those in numerous other shale- and gas-rich sedimentary basins around the world. Multiphase flow simulations have suggested that water can become underpressured in the presence of gas phase due to transient glacial loading cycles, and a previous modeling study of the Bruce site, in which the presence of gas phase was approximated using ad hoc adjustments to single-phase flow parameters, showed that underpressures can persist for geologically significant periods of time. However, while multiphase interaction and migration processes have been studied extensively for conventional petroleum and environmental engineering applications, they are relatively poorly understood in low-permeability argillaceous rocks such as those at the DGR. The goals of this study are to: (1) determine which rock and fluid parameters are most critical for understanding the multiphase flow processes that may have occurred in the low-permeability formations at the Bruce site through geologic time, (2) assess uncertainty in our understanding of those parameters, and (3) investigate, using the multiphase flow simulator iTOUGH2-EOS7C, whether the presence of gas phase methane could have generated or contributed to the underpressure. Results suggest that the presence of gas phase methane does not by itself fully explain the underpressure.
The goal of this study is to optimize a dual-fiber filter media by increasing the dust holding capacity (DHC), while maintaining the initial pressure drop and initial filter efficiency. Three main parameters define the performance of a filter, namely the DHC, filter efficiency and pressure drop. The DHC defines the quantity of solid particles which a filter media can trap and hold before the maximum allowable pressure drop is reached. The key idea is the use of micro structure models to optimize the filter media. To generate various filter media, the FiberGeo module of GeoDict® software is used, while the FlowDict and FilterDict modules simulate the flow behavior and the particle filtration behavior of the media, respectively.
Three different dual-fiber filter media are modelled. They all consist of the same two types of glass fibers (bi-modal diameter distribution). The distribution of the coarser fibers is uniform for all three models. They provide stiffness to the media and support the finer fibers. However, the finer fibers are distributed differently over the through-direction of the filter media [1]. The three models are called homogeneous model, linear increasing model, and exponentially increasing model. The naming scheme is based on the distribution of the finer fibers in through-direction of the filter media. In order to make the three models comparable, they were designed to have the same pressure drop in the clean state and to have the same initial filter efficiency. To ensure this, FlowDict and FilterDict simulations were used to guide the choice of geometric model parameters.
The first simulation step is to model the filter media with different spatial distributions of the two types of fibers. The next step is to simulate the clean fluid flow through the filter media. The final step is to simulate the particulate flow and particle deposition.
Filtration simulations on the three different models were done using the FilterDict module of GeoDict®. Filter life time simulations were carried out using the multi pass mode. In a multi pass simulation, fluids move in a circuit through the system, and the particle size distribution and concentration in front of the filter change over time. Outputs of these simulations are pressure drop, filtration efficiency and count/mass of the deposited dust as a function of time. The results show that by changing the distribution of the thin fiber type over the through-direction, the DHC of the filter media changes as well. In this way, the structure of the dual-fiber filter media can be optimized to achieve higher DHC, while keeping the initial pressure drop and initial filter efficiency the same.
Reverse Osmosis Membrane (ROM) filtration systems are widely utilized in waste-water recovery, seawater desalination, landfill water treatment, etc. During filtration, the system performance is dramatically affected by membrane fouling which causes a significant decrease in permeate flux as well as an increase in the energy input required to operate the system. Design and optimization of ROM filtration systems aim at reducing membrane fouling by studying the coupling between membrane structure, local flow field and foulant adsorption patterns. Yet, current studies focus exclusively on oversimplified steady-state models that ignore any dynamic coupling between fluid flow and transport through the membrane. In this work, we develop a customized solver (SUMs) under OpenFOAM to solve the transient equations. The simulation results not only predict macroscopic quantities (e.g. permeate flux, pressure drop, etc.) but also show an excellent agreement with the fouling patterns observed in experiments. It is observed that foulant deposition is strongly controlled by the local shear stress on the membrane, and channel morphology or membrane topology can be modified to control the shear stress distribution and reduce fouling. We demonstrate how channel morphology and membrane topology can be jointly optimized in order to increase the efficiency of the system. Finally, we identify optimal regimes for morphological and topological modifications in different operation conditions.
In order to improve the performance of the exhaust after treatment system and keep reasonable complexity, the number of the used devices is reduced by enhancing wall flow particulate filters with a catalytic functionality, like selective catalytic reduction in diesel or three way catalysis in gasoline vehicles. In this case the solid matrix of the filtering media consists from inert grains and active grains. For the simulation of such devices, effective 1D models are regularly used, because they are relatively simple and fast in comparison to higher dimensional models, while reproducing most of the experimentally observed phenomena with sufficient accuracy. The reduction of the complexity in the modeling, however, is achieved by introducing many simplifying assumptions.
As it will be seen in this presentation, 1D models do not always reproduce detailed 3D simulations. Here we compare a standard 1D homogenous wall model with a 3D pore-scale model. The latter describes convection and diffusion in the pores and in washcoat grains, and absorption in the washcoat grains. In fact, washcoat particles are nanoporous and surface reaction (adsorption) occurs at this scale. Softare tool called PoreChem [1] has been used for simulating the reactive flow at pore scale. A wall segment of a real particulate filter was used in the simulation. The three dimensional structure of the wall segment was obtained by X-ray microtomography, in which we resolved the different materials: Pores, (inert) Substrate and (active) Washcoat. A first order reaction was studied to examine if 1D simulations based on effective (Darcy scale) model can describe the behavior of a catalytic filter wall sufficiently well. In simulations the adsorption rate constant varied while the other properties, like temperature and flow speed remained constant. The conversion (adsorbed amount) was computed for different reaction rate. The discrepancy between the results obtained with the 3D and 1D models increase with higher reaction rates. They are caused by the inhomogeneous flow and washcoat distribution in the 3D system that cannot be described well with a homogenized 1D model. 3D models are therefore useful to optimize the microstructure of catalytic particulate filters.
In this work, we study the efficiency of the filter media at different operational regimes. In particular, we focus on how the media microstructure influences the efficiency performance, which is quantified by macroscopic parameters. We investigate the microscale characteristics since the filtration is an intrinsically multiscale process. On one hand, contaminant adsorption onto the fibre surface occurs at the pore-scale of the non-woven filter media, i.e., it is a microscale process. On the other hand, we are interested in the overall performance of the filter media, which is a macroscale characteristic, and how the performance changes for different macroscopic characteristics, such as porosity and thickness of the filter media and operational conditions. Filtration processes at both scales are fully coupled. Therefore, to model this problem we use the method of multiple scales, which is an upscaling technique that models variations at macroscale while accounting for filtration processes at microscale. The advantage of this multiscale method is that the coupled micro- and macroscale behaviour is captured without the computational expense of globally resolving the microscale filtration problem. We account for: a single-phase fluid flow through the filter medium, the contaminant transport with convection, diffusion and adsorption and the evolution of the filter medium microstructure due to the contaminant adsorption. Using developed simulation framework, we perform extended numerical study to investigate the influence of microstructure on the filtration performance at different filtration regimes.
Cold Heavy Oil Production with Sand (CHOPS) is widely used as a primary non-thermal production technique in thin heavy oil reservoirs. Development of the complex wormhole networks (i.e., high-permeability channels caused by sand production) renders the scalability of post-CHOPS solvent-aided processes to field applications challenging. It is widely accepted that configuration of wormhole networks and foamy oil flow are key characteristics pertinent to these processes.
First, a series of mechanistic compositional simulation models at the lab scale is constructed to model a cyclic solvent injection scheme (CSI). These models are calibrated against experimental measurements of solvent diffusion measured in porous media. Next, a set of detailed high-resolution (fine-scale) simulation models, where both matrix and high-permeability wormholes (modeled as fractal networks) are represented explicitly in the computational domain, is constructed to model how the solvent propagates away from the wormholes and into the bypassed matrix. Finally, a statistical scale-up procedure is developed to assign parameters of the equivalent dual-permeability model (e.g., dispersivities, shape factors) in accordance to the grid size and wormhole intensity within the grid block. The novelty of this scheme is that the bivariate distributions between effective dispersivities and wormhole intensity at the coarse scale are calibrated from detailed fine-scale simulations.
In the end, field-scale simulations are constructed using average petrophysical and fluid properties extracted from several existing CHOPS reservoirs. Wormhole development and the end state of CHOPS are modeled using the concept of critical pressure gradient. Multiple field injection scenarios (i.e., of different number of cycles and durations of the soaking period) are analyzed. As expected, extended soaking period is more beneficial in terms of ultimate oil recovery, but it also reduces the early production rate. Interestingly, when an economic limit (i.e., minimum oil producing rate) is imposed, the optimal soaking time is not necessarily the longest one. It depends on the trade-off between extracting additional oil recovery at late times versus producing at a higher rate at early times. Our results also support the strategy of injecting all the solvent in one single consolidated cycle, with an extended soaking period, rather than performing shorter consecutive cycles.
Field-scale flow simulations are often performed to approximate the reservoir response and to optimize operating strategies. However, grid block sizes in field-scale models are generally much larger than the wormhole scale, and numerical analysis is often performed by arbitrary adjustment of dispersivity. This work, however, offers a statistical scale-up workflow that facilitates the construction of coarse-scale dual-permeability models, whose shape factors, fracture spacing, fracture porosity, and effective dispersivities are assigned based on calibration against simulation results of detailed fine-scale wormhole network models. The proposed method serves as a starting point for formulating a systematic workflow that can be integrated with commercial reservoir simulators to effectively simulate solvent processes in wormhole networks that span over multiple scales.
Water is the most predominant component in steam injection processes, such as steam-assisted gravity drainage (SAGD). The main hypothesis in this research is that in-situ oil transport can be substantially enhanced by generating oil-in-water emulsion, where the water-continuous phase acts as an effective oil carrier. The objective of this paper is to evaluate the capability of oil-in-water emulsions to transport bitumen in porous media within a temperature range of 350 – 470 K at 3.5 bar.
Athabasca bitumen and two organic alkalis of different chemical structures were used in this research. An alkali concentration of 0.5 wt% was found optimal to form oil-in-water emulsions with bitumen and NaCl brine at a range of salinity. In-line density measurements were conducted to confirm that there was single-phase emulsion in a 200-ml accumulator. Then, flow experiments through glass beads packs were carried out to estimate effective viscosities of oil-in-water emulsions at typical shear rates in oil sands under SAGD (e.g., 1 to 10 [1/sec]). The bitumen concentration in each emulsion sample was quantified directly by using an emulsion breaker. Finally, emulsion molar flow rates were calculated with an analytical equation of gravity drainage.
Emulsion viscosity measurements showed shear-thinning behavior with much lower viscosity than the original bitumen. At an estimated shear rate of 1.0 per second, for example, the viscosity of oil-in-water emulsion for 0.5 wt% alkali was 12 cp at 350 K, which is much lower than the original bitumen viscosity, 190 cp, at the same temperature.
The obtained experimental data, such as effective emulsion viscosity, bitumen content in emulsion, and properties of the porous media used, were then used to quantify the bitumen molar flow under gravity drainage. Results showed that oil-in-water emulsion can enhance in-situ bitumen transport by a factor of 4 in comparison with SAGD, and by a factor of 2 in comparison with hexane-SAGD.
This research demonstrate that the clear advantage of bitumen transport via water-external emulsion can be obtained by adding only 0.5 wt% of alkali in brine. The mobility of the bitumen-containing phase is substantially enhanced because the oil-in-water emulsion flows as a single phase with a much lower viscosity. In contrast, conventional solvents, such as n-alkane mixtures, reduce bitumen viscosity by dilution using a substantial amount of those expensive solvents.
A difference between the two alkalis used is in their ability to dissociate in water; one of the alkalis has three more dissociation sites than the other. This led to different phase behaviors and rheological properties observed in our experiments.
This is the first time the potential of organic alkalis to improve bitumen transport is quantified experimentally. The experiments were newly designed to evaluate the molar flow rate of oil-in-water emulsions at SAGD operating conditions. The alkaline concentration required for effective bitumen transport is even lower than 1 wt% in the brine. Recovery mechanisms of alkali-SAGD are explained in detail, and compared to the flow characteristics in SAGD and other solvent-assisted SAGD.
An important and largely unsolved problem is accurate prediction of the onset of precipitation of heavy organic compounds, such as asphalts, on the surface of the pores of a porous medium, such as oil reservoir. It was suggested some time ago that an effective method of identifying the onset may be through measuring the electrical conductivity of the solution that contains the heavy organic compounds. We test this hypothesis by carrying out computer simulation studies of asphalt flocculation and precipitation processes, utilizing two models of asphalt
formation that have been suggested in the past, namely, those that are based on diffusion-limited and reaction-limited aggregation processes. The simulations model the evolution of the asphalt aggregates as the precipitation agent is added to the solution, and computes the electrical conductivity of the solution as a function of time and the concentration of the asphalt aggregates.
The results indicate that, regardless of the aggregation model, under certain conditions the points at which the aggregates begin flocculating and precipitating induce discernable changes in the conductivity of the solution at particular points. The same changes are detectable in the number of the asphalt aggregates in the solution. However, the changes happen only if the concentration of the asphalt in the solution is relatively high. Otherwise, the electrical
conductivity of the asphalt-containing solution exhibits no special feature and, therefore, its measurement is not an accurate method for identifying the onset of asphalt precipitation.
Steam assisted gravity drainage (SAGD) is the main technologically and economically feasible method for in situ bitumen extraction. While SAGD is effective, the steam generation process is a major source of CO2 emissions, and many strategies are currently under test to improve both the environmental and economic performance. Solvent-based processes – either pure solvent injection or coinjection of solvent with steam – have proposed to improve SAGD performance. The addition of solvent, however, complicates an already complex multicomponent thermal-chemical process. Microfluidics is well suited to quantify the pore-scale of solvent-based recovery processes as well as rapidly determine the fluid properties with a tight control over experimental parameters. Here, we developed a suite of microfluidic platforms with relevant reservoir and fluid characteristics combined with optical-thermal imaging to (1) study pore-scale transport in solvent-based processes (e.g., recovery rate, emulsification, and precipitation) using both pure solvents and industrial diluents; and (2) measure fluid properties relevant to solvent-bitumen systems (e.g., solvent solubility and diffusivity). With strong potential to inform oil sand operators, microfluidics is faster (minutes instead of hours/days) and significantly cheaper than conventional methods; all stemming from small fluid sample required for experiments.
From Images to Rock Properties
In the Oil and Gas industry, digital images of rock samples are being collected and utilized for reservoir formation characterization more frequently than ever. This is in part due to the fact that imaging tools such as X-ray CT scanners and SEM’s have become more prevalent, and also due to the challenges and time-consuming processes in traditional core analysis of unconventional and complicated conventional reservoirs. Rock images (e.g. optical images of thin section) have traditionally been used for qualitative and to a limited extent quantitative analysis of rocks, often using laborious point-counting methods. With the advent of computerized image analysis techniques and image-based simulation models, the information that can be obtained from rock images has been tremendously improved both in quantity and quality. Nowadays, one can use rock images to characterize the pore-space of a rock sample to calculate porosity and then simulate the fluid flow in the digitally characterized pore-space to calculate permeability and even multi-phase flow properties. The accuracy of the image-based results would depend on the quality of the images, the resolution at which the images are acquired, and the algorithms behind the image analysis and image-based modeling tools. Having high-quality imagery data at multiple scales and multiple resolutions, and consistent and robust image analysis workflows is the key for achieving the best results.
PerGeos, the Digitial Rock Image Analysis and Modeling Suite has been developed for large and multi-modal data analysis. Using this software suite, one can analyze data from multiple sources and at multiple scales. Combined with the workflow automation capability, the analysis tools in PerGeos are used to generate effective workflows for various types of data and analyses. An overview of the digital rock technology will be presented in this talk. Examples of micro-CT, whole core CT, SEM, and FIB-SEM data analysis of rock samples will be shown.
Three-dimensional (3D) analyses of the pore structure of building materials are becoming progressively more important in recent years in order to get more accurate interpretations and simulations of moisture and heat transfer properties. These characteristics are a major determinant for the durability and sustainability of structures as well as for the health and comfort of the building occupants. Building materials are characterized by a large variety in pore radii (e.g., nanometer to millimeter scale) that influence these characteristics. The problem of constructing a pore level model at the representative scale, must be solved by including information from each length scale of this multi-scale system.
Two different imaging techniques are used to visualize the pore structure at different scales. On a micrometer scale computed tomography (CT) has proven to be an excellent and versatile tool to perform these analyses non-destructively. To visualize even smaller pore structures of building materials on the nanoscale, scanning electron microscopy combined with focused ion beam (FIB-SEM) is used. Post processing of the 2 dimensional FIB-SEM images results in a reconstruction of the 3D pore space. Both techniques allow calculating relevant parameters such as pore size, shape and orientation in 3D. Additionally, the pore network is also indirectly characterized by mercury intrusion porosimetry (MIP), resulting in a validation of the results of the direct imaging techniques with a multi scale MIP analysis.
Moreover, building materials are often non-granular in nature, resulting in pore networks comprising of complex pore shapes. Hence, these materials are ideal test cases for pore-shape analysis. We will look at different shape descriptors using the ratios of the longest, intermediate, shortest dimensions and compactness of the pore shapes, based on an approximating ellipsoid, in order to obtain a thorough and objective description of pore shapes as well as their orientation.
Also the study of the Representative Elementary Volume (REV) of these different parameters is nessecary to assess the quality of the model. Because the used datasets transient different length scales, the results of the REV analysis will be compared to the observations of Norris et al. (1991) and Nordahl and Ringrose (2008) for geological samples. They both suggested the existence of different REV sizes at different scales which has major implications when determining a relevant upscaling strategy.
References:
Nordahl, K. & Ringrose, P. S. 2008. Identifying the Representative Elementary Volume for permeability in heterolithic deposits using numerical rock models. Mathematical Geosciences, 40, 753-771.
Norris, R., Lewis, J. & Heriot-Watt, U. 1991. The geological modeling of effective permeability in complex heterolithic facies. Paper SPE 22692 presented at the 66th Annual SPE Technical Conference and Exhibition, Dallas, TX, 1991. 6-9.
Calcium carbonate (CaCO3) precipitation is a frequently-occurring natural subsurface process, in which supersaturated CaCO3 in brine can precipitate in subsurface environments. This phenomenon can naturally occur as part of diagenesis of rocks and can also be utilized for soil improvement. This study explored abiotic carbonate precipitation in coarse sands using X-ray microtomography (X-ray μCT) and examined its effect on permeability. CaCO3 was abiotically precipitated in a sand column, where brine supersaturated with CaCO3 was flowed into the column while controlling pH. During precipitation, the variations in porosity and pore saturation of precipitated carbonate were monitored by measuring the bulk mass of the test column. At the same time, the variation in permeability of the sand column was also measured using the falling-head permeability test. The pore-scale morphological pattern of carbonate precipitation was also examined by using X-ray μCT, imaging the interior structure of the cemented sand. Reductions in porosity and permeability and an increase in CaCO3 saturation were confirmed from the bulk mass measurement. The acquired X-ray images also showed that the precipitated CaCO3 usually coated sand grains. The porosity and CaCO3 saturation values were computed from the obtained X-ray images, and these values were validated against the bulk mass measurement. Further, the measured permeability was compared with the predicted values by Kozeny-Carman model and Kozeny-grain coating model. It indicated that the local porosity reduction, internal porosity within carbonate minerals, and the specific surface area had pronounced effects on the permeability reduction. Meanwhile, existence of of sub-voxel size carbonate crystals is found to act as limitation in the voxel-scale flow simulation.
In this presentation, we explore applications of the maximal inscribed sphere (MIS) map to characterize porous media and show connections with other laboratory measurements. Three-dimensional maps can be computed from x-ray micro-tomography images of porous rocks and have been commonly used to simulate mercury injection capillary pressure (MICP) curves. We present additional applications of MIS maps to porous media characterization, that can be particularly useful in cases of heterogeneous multiscale pore systems.
A frequency distribution of the 3D MIS map can be used also as a definition of pore-size distribution, herein PSD-MIS. We discuss how drainage and imbibition hysteresis curves relate to the PSD-MIS curves. We also show a relationship between NMR and PSD-MIS; by tracking the actual location of the diffusing particles in NMR simulations, we found a clear correspondence between the T2 relaxation times classification and the PSD-MIS classification.
A relationship between 2D MIS-PSD and 3D MIS-PSD is presented, which enables the use of 2D images to estimate 3D MIS-PSD, allowing us, for instance, to take advantage of current SEM imaging technology that produces high resolution very large BSEM stitched images, and combine multiple scales of images to extend the size range from Nano scale to millimeter scale in the computed PSD-MIS. We show how the MIS tool complements MICP lab data for sizes above 100 microns, and gas adsorption lab data for sizes above 300 nanometers.
Finally, we introduce an adjustable 2D particle-particle separation method based on MIS. We apply this method to estimate grain size distributions that are compared with similar results from laser particle analysis. Another application addresses the estimation of MICP curves from 2D images.
Based on the diversity of applications of MIS maps presented here, we conclude by discussing generalized MIS-based models as alternative multiscale approaches for porous media characterization.
Numerous porous materials are made of intricate clustering of polydisperse nanoparticles. The particle organization on a length-scale ranging from nanometers to some micrometers is a cornerstone to properly understand transport properties (diffusion-permeation). A strong need to a bottom_up approach mixing SAXS, SANS and 2D -3D imagery technics is highly suitable for these types of multiscale complex systems. This multimodal structural analysis offers the possibility to use 3D reconstructions and to build constrained models mimicking observed geometrical features [1]. These models can then be used to compute transport properties allowing comparison with experimental determinations.
In this talk, we focus on a long and slow dynamics inside nanoporous porous materials essentially driven by molecular diffusion. This mechanism plays a major role to estimate the sustainability over a long period of time of numerous manufactured porous materials. We analyze the interplay between pore structure, adsorption and diffusion and how these different processes act in long term molecular transport. We show how intermittent dynamics [2] involving adsorption, surface diffusion and relocation inside the pore space induces a large part of the strong reduction of molecular diffusion inside pore network.
Self-assembly of surfactants in confined geometries plays an important role in environmental, chemical and pharmaceutical technology. Adsorption of surfactants at metal oxide and other polar/charged surfaces depends primarily on the nature of their head groups: cationic surfactants exhibit high-affinity adsorption due to the interaction of the positively charged head groups with negative surface charges. Nonionic surfactants, on the other hand, exhibit low affinity adsorption isotherms due to the weak effective interaction between the hydrated head groups and hydrated surface. This difference in head group interaction with the surface causes a different evolution of surfactant aggregate morphologies at the surface. We are using small-angle neutron scattering (SANS) to study the influence of confinement on the aggregate structure of nonionic and cationic surfactants. Ordered mesoporous silica materials such as SBA-15 have favourable properties for such studies, due to the uniform size and shape of their primary pores. Studies of the aggregate morphologies of nonionic surfactants in SBA-15 have been published [1-3]. Here we will present results for the cationic surfactants DPCl and CPCl and discuss the similarities and differences in their aggregate morphologies in comparison to nonionic surfactants of the CnEm type.
The behavior of microemulsions in porous substrates is of relevance for a variety of processes, from tertiary oil recovery to soil decontamination. We will present preliminary results of a SANS study of droplet microemulsions imbibed in SBA-15, focussing on the questions if droplets of diameter greater than the pore size can enter the pores and to what extent the droplets become attached to the pore wall. Both kinds of information can be obtained by SANS measurements using a H2O/D2O solvent mixture that matches the scattering length density of the silica matrix, so that the microemulsion droplets appear against a uniform scattering background.
The confinement of liquids in porous media greatly influences their physical properties, in particular, when the pore size approaches the molecular length scale. Several mechanisms, such as the pure geometrical restriction and the liquid-solid interaction at the interface contribute to the confinement effects, however, their roles for the drastic changes in the thermodynamic and dynamic behaviors of the liquids are not clearly understood. Especially, water molecules adsorbed on the surface and restricted within the pores are interesting in the scope of biochemistry, catalysis, and energy storage. The influence of the pore size on the melting and freezing points of confined water has been studied extensively. In contrast, fewer studies have been made concerning the effects of the surface polarity on the properties of spatially confined water, which may be due to the lack of highly defined porous substances. Periodic mesoporous organosilicas (PMOs) fill this gap perfectly because they combine the highly ordered pore structure of the well-established M41S-phases with the variety of surface chemistry within the pores. PMOs are synthesized using bis-silylated precursors of the form (R'O)3Si-R-Si(OR')3 where R is an organic bridging group which can be altered according to the desired surface properties [1]. Furthermore, PMOs with an aromatic bridging group may exhibit a molecular-scale periodicity within the pore walls. This allows a periodically alternating surface chemistry along the pore channel, caused by arrays of silica and aromatic organic groups. It is assumed that water will adsorb differently at the diverse areas of the pore wall surfaces owing to the varying hydrophilic or hydrophobic properties. Because of the endless possibilities for the organic bridging function, the surface chemistry of PMOs can be fine-tuned. For example, when going from a benzene bridging group to a biphenyl bridge, the organic hydrocarbon part becomes larger and thus the overall pore wall becomes more hydrophobic. For a divinylaniline bridge, the amino function offers the possibility of forming hydrogen bonds with water, thus making the material considerably more hydrophilic. Here, we show that the molecular mobility of water confined in periodic mesoporous organosilicas (PMOs) is influenced by the polarity of the organic moiety. Multidimensional solid-state NMR spectroscopy directly probes the spatial arrangement of water inside the pores, showing that water interacts either with only the silicate layer or with both silicate and organic layers depending on the alternating surface polarity. A modulated and a uniform pore filling mode are proposed for different types of PMOs [2]. Our study gives a molecular-level picture of the adsorbate-surface interaction, which helps understanding various confinement effects and provides a new design concept of the pore structures with the desired properties.
At the nanoscale the positions of coexistence lines on the phase diagrams are shifted and their new locations depend mainly on the size and shape of the nano-confinement, the structure of the confining walls, and their interaction with the confined substance. Here we show that it is possible to induce structural transformations in a confined system by simply varying the number of molecules adsorbed in the pore. We found that the mechanism of these novel, adsorption-induced structural transformation in nano-pores differs from the capillary condensation. First, the structure of the confined gas is determined by a competition between adsorption sites attractive forces and intermolecular interaction. Second, at low temperature, the transformation is discontinuous because it is defined by limited number of adsorption sites [1,2].
The confined, equilibrium structures are not characterized by mean positions of molecules but rather by a probability distribution of molecular positions around adsorption centres. This distribution changes when the number of molecules in the pore increases. The character of transformation is temperature dependent: strongly discontinuous at low temperature, it evolves into a continuous transition when the temperature increases. The mechanism of the transformation is also modified when the size of the gas molecules and types of interaction change. In particular, we report the existence of the intermediate phase, observed only above a critical strength of the attractive interactions.
Non-uniform fluid displacements in heterogeneous porous media are commonly observed, while are unfavorable in oil/gas recovery processes. Recently, the deformable polymer particle gels such as preformed particle gel (PPG) [1] and soft micro-gel (SMG) [2] were successfully applied to improve the sweep efficiency under such bad conditions. Although experimental studies presented qualitative observations of their in-depth plugging and flowing diversion capabilities, deeper numerical mechanisms on the deformable polymer micro-gels for enhanced oil recovery (EOR) are still rare [3]. The difficulties mainly lie in the following reasons: (1) the migration of deformable polymer particle gels is non-continuous, invaliding the classical continuum based theories; (2) viscoelastic behaviors of the gel particles should be considered and (3) the problem is confined in the complex porous geometries.
In this paper, an Immersed Boundary-Lattice Boltzmann (IB-LBM) framework is established to capture the motion of viscoelastic polymer micro-gels in porous media. The LB method is used to solve the Navier-Stokes equations in a Eulerian coordinate for the main flow field. The fluid–particle interaction is carried out by coupling with the IB scheme within a Lagrangian coordinate. The viscoelastic deformation and behaviors of the particles are recovered by the capsule model containing a spring-network model and by setting another kind of fluid property.
Primarily, the critical size of a polymer particle gel to deform and pass through one single throat is studied. We find that the different geometry (particle-throat diameter ratio, the length of the throat), pressure gradient and the viscosity ratio are key factors. Thence we investigated the transport of the polymer micro-gels in the cross-shaped channel with heterogeneity. Once the gel is placed in the higher permeable zone, the resistance in this zone will increase. Thus the following injected water will be diverted into the lower permeable zone and help the residual oil displacement there. Our results improve the understandings of the mechanisms of the deformable polymer particle gel for enhancing oil recovery in heterogeneous porous media.
Solid nanoparticles (NPs) have shown promise to play a major role in novel enhanced oil recovery methods by altering reservoir wettability, reducing interfacial tension and increasing mobility ratio. Solid nanoparticles could be implemented as emulsion stabilizing agents in combination with surfactants and polymers. Stabilized emulsions are a desired state in enhanced oil recovery. In this study, the contribution of NPs in hydrodynamics of multi-phase systems including their effect on interfacial tension and mobility ratio was studied. The free energy Lattice Boltzmann Method was used to solve the Cahn-Hilliard convection-diffusion and the Navier-Stokes equations in a two dimensional Cartesian domain. The NPs were added to the system as point particles with zero volume. A potential function was assumed to represent the chemical potential alteration of the multi-phase system due to the presence of NPs. Attractive and repulsive interaction of the NPS was entered into the model by Morse potential function. The results of spinodal decomposition showed that different types of emulsions (oil in water, water in oil, water in oil in water) could form at the presence of particles with different wettabilities. The effect of the presence of NPs in a capillary tube on contact angle was studied and it was observed that low concentration of NPs does not significantly affect the contact angle. The process of suspended droplets getting coated by neutral-wet NPs was simulated and it was observed that attractive interaction of NPs would result in multiple layers of coating. The Collision and coalescence of two droplets at the presence of different types of NPs was simulated and it was observed that neutral-wet NPs dominate the collision hydrodynamics at high NPs concentrations. It was also observed that the nanoparticles with repulsive forces would stabilize emulsions at lower concentrations comparing to nanoparticles with attractive forces. This work showed that the free energy Lattice Boltzmann method combined with NPs assumed as points is an effective tool to model and simulate the hydrodynamics of multi-phase systems in micro and nano scales. The computation cost of the simulation was also discussed in details.
This work provides a comprehensive study to evaluate and optimize the effectiveness of nanofluids to both prevent fines migration and enhance oil recovery using different utilization approaches: nanofluids co-injection and pre-flush. To do that, 1) a comprehensive review of both laboratory experiments and field cases is adopted to confirm the effectiveness of nanoparticles to control fines migration. 2) A novel model of maximum fines retention concentration is then introduced to find out the physical mechanisms on how nanoparticles control fines migration. 3) Through matching with lab experiments, the physical behaviors of fines migration and attachment with the effects of different types of nanofluids are characterized, including fines attachment and straining rates, and breakthrough time of injected fines. 4) As a new criterion, mitigation index (MI) is defined to find out the more excellent performance of nanofluids pre-treatment that that of nanofluids co-injection. 5) In two-phase oil/water flow, analytical modeling and solutions of nanoparticles to control fines migration is developed, in terms of both enhanced oil recovery and well injectivity. 6) The pros and cons of fines migration on performance of low-salinity water flooding are discussed comprehensively, in this work, and the success of combining nanofluids with low-salinity water flooding is also confirmed to achieve more oil recovery. The outcomes of this work will help extend the applications of nanofluids in reservoirs suffering from problems of fines migration.
Low salinity water flooding is an effective and cost-efficient improved oil recovery method. Wettability alteration is believed by many to be the primary reason for the observations of increased recovery. However, the causes of the wettability alterations and approaches for optimization are not fully understood. We conduct experiments in micromodels by injecting brine at different salinities in these oil-saturated micromodels and observe distinct changes in the contact angle and wettability. Meanwhile, we observe water droplets form and grow with time inside the oil phase when low salinity water was injected. Additionally, significant improvements in oil recovery is observed when high-salinity brine is followed by low-salinity brine, but the response is delayed by hours or even days.
We fabricated glass, water-wet micromodels (homogeneous porous medium and single channel with dead-end) using the method we developed in Ke et al. [2017]. The micromodels are unique in that pores and throats are of different depths, thereby having three-dimensional features that allow oil snap-off behaviors. We conduct dead-end micromodel experiments and homogeneous micromodel experiments to see the oil recovery. For the dead-end micromodel experiments, visualization was conducted under microscope. Micromodels were initially saturated by crude oil (120 cp heavy Middle Eastern oil) and then flooded at very low rate by low salinity water (500 ppm) or high salinity (34000 ppm) water. For the homogeneous micromodel experiments, visualization was conducted under a digital camera. Micromodels were initially saturated by crude oil and then displaced by a high salinity sea water (34000 ppm) at 2 ft/day into micromodels until no additional oil was recovered. Then, low salinity (0-5000 ppm) water was injected at 2 ft/day into the micromodel for at least 2 days.
The dead-end experiments show that the micromodel was initial oil wet and gradually altered to more water wet after the low salinity water injection. We also observed water droplets forming and growing in the crude oil phase and oil swelling. This swelling appears to provide an additional mechanism for recovery.
From the homogeneous micromodel experiments, we observed up to 30% incremental oil recovery using the low-salinity brine after the high-salinity brine. Importantly, a time delay of hours even days was required. For zero-salinity (DI water) case we observed a more than 10 hours of flooding was required before production re-started, and more than 20 hours of delay for 5000 ppm brine.
Our conclusion is that the wettability alteration is a time dependent process caused by the redistribution of negative charged polar compounds within the crude oil. When low salinity water is injected, the polar compounds migrate from the surface to the inside of the oil and form micelles. The lack of polar compounds on oil surface leads to wettability alteration, and the water molecules accumulate around these micelles to form water droplets. The growing of water droplets leads to oil swelling which also contribute to the incremental oil recovery. The wettability alteration and the oil swelling are both consequences of the redistribution of the polar compounds, and both contribute to the additional oil recovery by both wettability alteration and oil swelling.
We propose a novel finite volume method for anisotropic linear elasticity problem. The derivation of the flux approximation method for elasticity problem closely follows our previous work [1] on the nonlinear finite volume methods for diffusion equation featuring positivity and discrete maximum principles. It is based on the extensions of the harmonic point idea of [2] from the scalar to vector equations. We further extend the idea to coupled anisotropic flow and mechanics, featuring full tensors for permeability, Biot coefficient and stiffness tensor. Both methods handle star-shaped polyhedral grids, admit the construction of nonlinear finite volume methods and yield several new research directions.
[1] "Cell-centered nonlinear finite-volume methods for the heterogeneous anisotropic diffusion problem”, KM Terekhov, BT Mallison, HA Tchelepi // Journal of Computational Physics 330, 245-267
[2] "A nine-point finite volume scheme for the simulation of diffusion in heterogeneous media.”, L Agelas, Rt Eymard, and R Herbin // Comptes Rendus Mathematique, 347. 11-12 (2009): 673-676.
We develop higher order multipoint flux mixed finite element (MFMFE) methods for solving elliptic problems on quadrilateral and hexahedral grids that reduce to cell-based pressure systems. The methods are based on a new family of mixed finite elements, which are enhanced Raviart-Thomas spaces with bubbles that are curls of specially chosen polynomials. The velocity degrees of freedom of the new spaces can be associated with the points of tensor-product Gauss-Lobatto quadrature rules, which allows for local velocity elimination and leads to a symmetric and positive definite cell-based system for the pressures. We prove optimal k-th order convergence for the velocity and pressure in their natural norms, as well as (k+1)-st order superconvergence for the pressure at the Gauss points. Moreover, local postprocessing gives a pressure that is superconvergent of order k+1 in the full L2-norm. Numerical results illustrating the validity of our theoretical results are included.
In this work a general model of multiphase flow and multicomponent transport in porous media to simulate, analyze and interpret hydrocarbon recovery processes by injecting low salinity water using open source software is presented. The flow model is multiphase considering capillary pressure and relative permeabilities depending on salinity, while the transport model is multicomponent and includes several relevant physico-chemical phenomena such as advection, diffusion and reactions. To obtain the numerical solution, the finite volume method is applied in space and backward Euler finite difference method in time, resulting in a fully implicit scheme. Its computational implementation was carried out in C++ using the open source software platform DUNE-DUMUX. From the methodological point of view, each stage of the development of the model is described (conceptual, mathematical, numerical and computational). The resulting model is applied to a case of study of low salinity water injection in a core at laboratory conditions.
CO$_2$ capture and storage is an important technology for mitigating climate change. Design of efficient strategies for safe, long-term storage requires the capability to efficiently simulate processes taking place on very different temporal and spatial scales. The physical laws describing CO$_2$ storage are the same as for hydrocarbon recovery, but the characteristic spatial and temporal scales are quite different. Petroleum reservoirs seldom extend more than tens of kilometers and have operational horizons spanning decades. Injected CO$_2$ needs to be safely contained for hundreds or thousands of years, during which it can migrate hundreds or thousands of kilometers. Because of the vast scales involved, conventional 3D reservoir simulation quickly becomes computationally unfeasible. Large density difference between injected CO$_2$ and resident brine means that vertical segregation will take place relatively quickly, and depth-integrated models assuming vertical equilibrium (VE) often represents a better strategy to simulate long-term migration of CO$_2$ in large-scale aquifer systems. VE models have primarily been formulated for relatively simple rock formations and have not been coupled to 3D simulation in a uniform way. In particular, known VE simulations have not been applied to models of realistic geology in which many flow compartments may exist in-between impermeable layers. In this work, we generalize the concept of VE models, formulated in terms of well-proven finite-volume reservoir simulation technology, to complex aquifer systems with multiple layers and regions. The result is a hybrid discretization strategy which couples different governing equations in different regions based on the correct local discretization of gravity and flow terms.
We also introduce novel formulations for multi-layered VE models by use of both direct spill and diffuse leakage between individual layers. This new layered 3D model is then coupled to a state-of-the-art, 3D equation-of-state compositional model. The formulation of the full model is simple and exploits the fact that both models can be written in terms of generalized multiphase flow equations with particular choices of the relative permeabilities and capillary pressure functions. The resulting simulation framework is very versatile and can be used to simulate CO$_2$ storage for (almost) any combination of 3D and VE-descriptions, thereby enabling the governing equations to be tailored to the local structure. We demonstrate the simplicity of the model formulation by extending the standard flow-solvers from the open-source Matlab Reservoir Simulation Toolbox (MRST), allowing immediate access to upscaling tools, complex well modeling, and visualization features. We demonstrate this capability on both conceptual and industry-grade models from a proposed storage formation in the North Sea. While the examples are taken specifically from CO$_2$ storage applications, the framework itself is general and can be applied to many problems in which parts of the domain is dominated by gravity segregation. Such applications include gas storage and hydrocarbon recovery from gas reservoirs with local layering structure.
As length scales grow smaller, the behavior of two fluids flowing through a porous media is more and more strongly affected by capillary pressure. The canonical Young-Laplace equation gives a good estimate of the pressure change across fluid interfaces, but it is technically valid only in equilibrium conditions. It is well-known that for two fluids flowing in capillaries, the interfacial pressure change depends on dynamics, such as flow velocity or, more directly, changes to the interface shape. In this work we investigate ways to represent the dynamical state of pore scale flow in a network model that tracks two-fluid interfaces as they move through a series of connected capillaries. The pore scale model consists of molecular level simulations of two fluids flowing through a single cylindrical capillary. The network model consists of a network of cylindrical capillaries of varying lengths and diameters, for which we combine the equations of motion for two fluids with continuity equations to derive a system of differential algebraic equations, which can be solved for the state of the network in time. To link the two simulations, from the pore scale results we extract the dependence of the capillary pressure on flow speed, which can enter in various ways into the equations of motion for the capillary network.
Wettability of porous media has a remarkable influence on the morphology of invading fronts during fluid-fluid displacement. For example, it has been shown that when invading and defending fluids exhibit an instability-inducing viscosity ratio, the invading phase advances through viscous fingering, and the width of the fingers is dictated by the substrate wettability. When the porous medium has low affinity to the invading liquid (drainage), the width of the fingers is comparable to a pore size. In contrast, when the porous medium has high affinity to the invading liquid (imbibition), the thickness of the fingers is well above the pore size, and the invading phase advances as a more smooth and compact front [1, 2]. Recently, the experimental observations were extended to the strong imbibition regime. It was shown that, in this case, front displacement occurs via corner flow, where the invading fluid advances by coating the posts in a patterned Hele-Shaw cell [3].
Motivated by these experimental observations in patterned microfluidic cells, we build on the work of Cieplak and Robbins [3] to develop a quasi-static pore invasion model for the full range of pore wettabilities, from strong drainage to strong imbibition. We describe the pore geometry as a pore network, and explicitly calculate the critical pressures of pore invasion events to advance the fluid-fluid interface based on these pressure thresholds. This revisited formulation of fluid invasion removes the bias associated with user-dependent choices of pressure increments during the invasion, or conventions for the sequence of interface pore configurations. Our quasi-static simulations show a transition from invasion-percolation to cooperative pore filling to corner flow as the wettability of the medium to the invading fluid increases, in quantitative agreement with the experimental observations on micromodels.
Finally, we extend our model from a quasi-static to a dynamic description by accounting for viscous forces during pore invasion, and buoyancy effects from density difference between the fluids. We apply the new model to investigate impact of wettability on the morphology of unstable flow during secondary oil migration.
Extracting structural information, such as pore networks, from tomographic images is a powerful tool for the study of porous materials. Pore network models are used for predicting different physical properties such as permeability and tortuosity, and simulating chemical process such as reactive transport. Due to its inherent simplifications, pore network modelling requires far less computation cost and time as compared to other pore-scale modelling approaches like lattice Boltzmann and finite volume method allowing researchers to investigate larger volume of porous materials[1,2]. Consequently, they are being increasingly used to study multiphysics in various devices where the porous structure plays a critical role in performance[3], but true pore-scale modeling is computationally infeasible.
In traditional pore network modelling applications, the solid phase network is not usually considered since the only the fluid phase is of interest. In electrochemical devices, however, electron and heat transport through the solid phase of the electrode are equally important. In this work, we report on the concept of extracting dual network models from tomographic images, that includes information for both solid and pore phases and interlinking of these phases with each other. These dual network models provide a new avenue for understanding many critical chemical process like reaction-diffusion in catalysts as well as batteries charging and discharging kinetics in porous electrodes, both of which require understanding the solid phase transport.
The presented algorithm is based on an algorithm recently published by our group[1], and produces output that is compatible with the open-source modeling package OpenPNM[4]
References:
1. Gostick, J. T. Versatile and efficient pore network extraction method usingmarker-based watershed segmentation. 1 Phys. Rev. E 3300, 1–11 (2017).
2. Xiong, Q., Baychev, T. G. & Jivkov, A. P. Review of pore network modelling of porous media : Experimental characterisations , network constructions and applications to reactive transport. J. Contam. Hydrol. 192, 101–117 (2016).
3. Aghighi, M., Hoeh, M. A., Lehnert, W., Merle, G. & Gostick, J. Simulation of a Full Fuel Cell Membrane Electrode Assembly Using Pore Network Modeling. J. Electrochem. Soc. 163, F384–F392 (2016).
4. Gostick, J. et al. OpenPNM: A Pore Network Modeling Package. Comput. Sci. Eng. 18, 60–74 (2016).
We present a major new advance in the interpretation of percolation characteristics, provided by mercury porosimetry, porometry, water retention for soils, and, for nanoscale porosity, Grand Canonical Monte Carlo simulations matched to surface area adsorption measurements. The new method provides a complete analysis of all the void types, providing the user with cumulative distributions that asymptote with size towards the accessible sample porosity. It employs a quasi-Bayesian inverse modelling approach that interprets the percolation characteristic in terms of pores, pore-throats and void clusters, and is shown to be much more accurate than the traditional but incorrect distributions based on the slope of the percolation intrusion characteristic. The advances are applicable to a wide range of micro and mesoporous non-ordered porous materials. Their utility is illustrated by application to three porous materials of current interest: (i) graphite manufactured in Japan for the cores of next generation nuclear reactors, (ii) microporous calcium carbonate being developed for delayed oral drug and flavour delivery, and (iii) naturally occurring hydrophobic soil which will become less fertile and more prone to causing floods under expected climate change conditions. We also describe an improved wetting algorithm, of use not only for soil but also for other important systems such as membranes and fuel cells.
The new void analysis method uses a Boltzmann-annealed amoeboid simplex to inverse model the percolation characteristic. The convergence of the simulated onto the experimental percolation characteristic is automated within the PoreXpert(R) software package [Levy et al. (2015)]. The resulting void structure comprises a `unit cell' with periodic boundary conditions, containing up to 108$\,$000 void features. Since the solution to the inverse problem is non-unique, a series of stochastic solutions is generated, and the most statistically representative structure is chosen. For the graphite, a size-range gap between the derived GCMC and direct mercury intrusion percolation data is reflected in a wider range of stochastic generations due to the uncertainty for voids of around 1$\mu$m [Jones et al, 2018]. The percolation characteristic of the representative structure is then forward modelled. A comparison of the intrusion pressure of each pore with respect to the applied pressure then demonstrates whether each feature is a true pore or a cluster of voids [Matthews et al. 2017].
The working equation for the wetting algorithm, derived from that of Bosanquet, is:
$\dot{x} = \left[ \frac{b \, \left( 1 - \mathrm{e}^{-a ( t_\mathrm{full} - t_\mathrm{entry})} \right)^2}{2 a \left\lbrace t_\mathrm{full} - t_\mathrm{entry} - (1/a) \left(1 - \mathrm{e}^{-a(t_\mathrm{full} - t_\mathrm{entry})} \right) \right\rbrace } \right] ^{1/2}$
$a$ and $b$ are constants defined as
$a = \frac{8 \eta}{r^2\rho} \, , \, b = \frac{P_e}{\rho} + \frac{2 \gamma \mathrm{cos}\theta}{r\rho}$
$x$ is the distance travelled by the liquid front in time $t$ into a cylinder of radius $r$, $\eta$ and $\rho$ are the dynamic viscosity and density of the wetting liquid, $\rho$ is the liquid density and $P_e$ is the external pressure applied at the entrance of the capillary tube. $\gamma$ is the interfacial tension at the meniscus, and $\theta$ the contact angle of the intruding fluid with the solid surface. $t_\mathrm{entry}$is the time when liquid enters a particular pore-throat and $t_\mathrm{full}$ when it becomes completely full. The progress of the many thousands of simultaneous wetting events are found by use of a forward Euler method, which reduces the calculation time from weeks to tens of hours. For the soil, the wetting throughout the network stabilises at a timestep of 10 microseconds, which also then equates to the so-called `Haines jumps' of the wetting front.
With respect to the three materials we answer three specific questions: (i) how does the void structure of the nuclear graphite differ if the manufacturing process is changed to increase the density, (ii) can additional microporosity be generated by packing the microporous calcium carbonate particles, and (iii) how do molecular scale hydrophobicity transitions upscale to cause soil hydrophobicity at core and field scale?
In this work, we investigate CO2 exsolution, transport, trapping and dissolution in shallow subsurface under various conditions.
First, we introduce mathematical model describing the system. For the mass transfer of CO2 the rate limited model is used.
Numerical results obtained using the model are compared to the experimental data obtained from two sets of experiments: 1D column experiments and intermediate scale 2D experiments. In all the experiments water with dissolved CO2 was injected into the tank and the fate of dissolved and gaseous CO2 was observed.
The experiments were conducted under various conditions including different heterogeneity configurations, flow rates and dissolved CO2 concentration.
We investigated the effects of the different conditions in the experiment on the studied processes of exsolution, transport, trapping, and dissolution and addressed these dependencies in the mathematical model.
Soil moisture is closely linked to the near-surface heat and mass transfer that couples the land and atmospheric states. The accurate simulation of the spatiotemporal distribution of soil moisture is constrained by existing knowledge gaps with respect to the mechanisms and processes linking the atmospheric and soil states, their magnitude, and sensitivity to applied soil conditions. In this study, we explore the effects of variations in microclimate conditions on state variable distributions and water balances for different bare soil conditions. A series of evaporative experiments were conducted at the Center for Experimental Study of Subsurface Environmental Processes (CESEP) wind tunnel-porous media test-facility to generate atmospheric and subsurface datasets that were in turn applied in the prescribed soil-atmosphere boundary conditions of a heat and mass transfer numerical model. Experimental results showed that for the length scale and edaphic conditions tested, variations in local soil-atmosphere coupling had a slight impact on the lateral distribution of soil moisture. This localized soil moisture variability could not be reproduced with numerical model. From a water balance perspective however, cumulative water loss could be adequately captured with little loss of fidelity. This demonstrates that heat and mass transfer models may be insensitive to the local microclimate driving bare-soil evaporation but are strongly influenced by local soil properties (i.e., heterogeneity). Together, these findings suggest that greater focus should be given to characterizing subsurface conditions and subsurface constitutive models and heat and mass transfer theory than localized near-surface atmosphere conditions.
We present general formulations of the phase-equilibrium and phase-stability problems for multicomponent mixtures and verify that these formulations generalize the problems of phase-equilibrium and phase-stability at constant volume, temperature, and mole numbers ($VTN$-flash), at constant internal energy, volume, and mole numbers ($UVN$-flash), and at constant pressure, temperature, and mole numbers ($PTN$-flash). Furthermore, we develop a numerical method for solving the general formulation of phase-equilibrium problems. This algorithm is based on the direct minimization of the objective function with respect to the constraints. The algorithm uses a modified Newton-Raphson method, along with a modified Cholesky decomposition of the Hessian matrix to generate a sequence of states with decreasing values of the objective function. The algorithm was implemented in C++ and using generic programing we have a single, portable solver for all three flash formulations. Properties of the algorithm are shown on phase-equilibria problems of multicomponent mixtures in different specifications and with different levels of difficulty. Complexities and numerical performance of the individual flash formulations are discussed.
Emissions of methane and other hydrocarbons from old or abandoned oil and gas wells are a growing environmental and public safety risk. There is growing interest in identifying materials that are inexpensive and can be easily pumped into the wellbore from above in order to create a permanent seal for leaking hydrocarbons. Here we developed a novel bio-clay composite (BCC) material with the capability of swelling in the presence of methane. Conventional bentonite clays swell significantly in water, which would limit the ability of these materials to penetrate deep into the wellbore. In contrast, modified oragnoclay materials swell poorly in the presence of light chain hydrocarbons like methane. BCCs leverage the growth of microbial organisms such as methanotrophs that metabolize the methane and produce biomass. The growth of these microorganism will result in the accumulation of biomass, which in turn will induce the swelling of the BCCs.
In this talk we will report on experimental results carried out over the past year to characterize this approach. The novel BCCs were synthesized by reacting sodium bentonite clay with amino acids using ion exchange process. Conventional organoclays are typically made using quaternary ammonium compounds and other surfactants but these tend to have biocidal properties. The result of this synthesis was a hydrophobic organoclay that was biocompatible. To evaluate the performance of these BCCs we used E. coli as a model microbial organism. Inhibition tests of E. coli growth by amino acids were conducted wherein cells were exposed to different concentrations of amino acids and a positive control, and the bacterial growth was recorded by measuring their optical density. The results of inhibition tests showed that amino acid-based BCCs have insignificant impacts on the growth of E coli. We also conducted swelling tests and hydraulic conductivity tests to understand how biogenic solution will impact the swell index and permeability of BBCs. BBCs were characterized using small angle X-ray scattering (SAXS) to identify the d-spacing of the samples before and after being mixed with biogenic solution. The results demonstrated that the introduction of biomass creates higher swelling index and dramatically reduced hydraulic conductivity for BBCs.
Although two-phase flow in porous media is an established research field since decades, its theoretical background is still incomplete. In particular, while a universal definition of capillary pressure exists at the micro-scale, its upscaling to the macro-scale is still rather vague and a rigorous theory of capillarity at the macro-scale is missing. In this work, a new macroscopic theory of capillarity based on the volume averaging method is presented. The novel feature of the proposed averaging approach is the use of the superficial surface average for upscaling the relevant conservation equations for a surface. This allows for rigorous derivation of the macroscopic momentum balance equation for all the fluid-fluid interfaces contained within the Representative Elementary Volume (REV), thus resolving most of the shortcomings of previous studies, such as the averaging-scale inconsistency and the accounting for the different orientation of interfaces within the averaging volume. This latter aspect is described by an additional parameter arising in the proposed derivation, namely the intrinsic surface average of interface normal vectors $\langle \mathbf{n}_{nw}\rangle^{nw}$. Furthermore, defining the macroscopic capillary pressure as the difference between the intrinsic surface averages of the bulk pressures, it is shown how the capillary pressure-fluid phase saturation curve can be determined in a more consistent manner by upscaling results of pore-scale simulations as oppose to traditional coreflooding experiments. This sets new challenges and opportunities for modelling unsaturated porous materials.
This presentation report recent advances in the framework of the discrete element method (DEM) for multiphase granular media. Computationally efficient methods based on the DEM have been developed for a while for partially saturated materials but they have been generally limited to the pendular regime. In contrast, one hardly avoid expensive direct resolutions of 2-phase fluid dynamics problem for mixed pendular-funicular situations or even saturated regimes. Following previous developments for single-phase flow, a pore-network approach of the coupling problems is described. The geometry and movements of phases and interfaces are described on the basis of a tetrahedrization of the pore space, introducing elementary objects such as bridge, meniscus, pore body and pore throat, together with local rules of evolution [1]. As firmly established local rules are still missing on some aspects (entry capillary pressure and pore-scale pressure-saturation relations, forces on the grains, or kinetics of transfers in mixed situations) a multi-scale numerical framework is introduced, enhancing the pore-network approach with the help of direct simulations [2]. Small subsets of a granular system are extracted, in which multiphase scenario are solved using the Lattice-Boltzman method (LBM). In turns, a global problem is assembled and solved at the network scale, as illustrated by a simulated primary drainage.
Flow and transport processes in fractured rock are strongly influenced by fracture aperture and diffusive interactions between the fractures and rock matrix. The transmissivity of fractures is highly sensitive to fracture aperture (a cubic dependence). Complex thermo-hydrologic-mechanical-chemical (T-H-M-C) coupled processes that alter fracture apertures thus drive feedbacks and pattern formation. For example, dissolution of fracture surfaces increases their aperture and transmissivity, whereas precipitation has the opposite effect. We adopt aperture-integrated approaches for modeling fracture alteration and the resulting evolution of flow, concentration, temperature and aperture fields. I will present examples from previous and ongoing research on fundamental aspects of fluid flow, reactive transport and coupled processes in fractured rock, including comparisons to controlled high-resolution experiments. Behavior in both kinetic regimes (where disequilibrium drives water-rock interaction) and gradient reaction regimes (where solubility gradients drive dissolution-precipitation) are considered. Dissolution of fracture walls by reactive solutions produces highly preferential aperture growth, either by unstable finger propagation in the kinetic regime, or by selective growth in the gradient reaction regime. These phenomena are relevant in the context of acid injection for permeability enhancement in energy recovery systems, water-rock interactions during geological CO2 sequestration, and natural phenomena such as karst and cave formation. Precipitation reactions lead to unusual growth of elongate precipitate bodies perpendicular to the mean flow direction. We also present results from geological scale mountain-hydrologic systems with thermo-hydro-chemical coupling relevant to the formation of hypogene or thermal karst systems. In these systems, both the temperature-dependent (retrograde) solubility of calcite, onset of buoyant convection after a threshold Rayleigh number is reached due to the growth of fracture transmissivity by dissolution, influence the growth patterns that result. Recent applications of our modeling approach to thermo-hydro-chemical coupling in geothermal energy doublet-flow systems reveals useful strategies for sustaining energy production. In silica-dominated geothermal energy reservoirs, we show that injection of undersaturated water leads to long-term sustainable operation of the system. We also show that after an initial period of 1-2 years, sustained energy extraction can be sustained even with recirculation, because silica concentrations do not build up in recirculated water. The large difference in solute versus thermal diffusivity of the rock matrix, however, ensures that thermal energy recovery is sustained. With oversaturated injection, a band of precipitation forms at some distance from the injection well, encapsulating the flow system and limiting heat recovery.
In many subsurface energy activities, such as hydraulic fracturing, geologic carbon storage, deep well disposal, and geothermal energy, the injection or extraction of fluids results in significant mechanical and chemical perturbations. These perturbations pose risks in promoting unwanted fluid leakage pathways, such as faults and fractures, through chemical reactions and mechanical failure. The flow of reactive fluids through fractures has been shown to dissolve reactive minerals and thereby increase permeability. However, it remains unclear how this dissolution influences the frictional properties of fractures and how the altered fractures behave during shear rupture. To properly assess the risks associated with subsurface energy activities, this study experimentally investigates how mineral dissolution on fracture surfaces affect fracture flow and frictional properties under stress.
Experiments were performed using 1.5-inch long saw-cut rock fractures from the Eagle Ford formation, a calcite-rich laminated shale. Experiments were performed in two stages: reactive fluid flow experiments and tri-axial shearing experiments. During the first stage, samples were organized into two sets, where one set was exposed to an acidic brine and the other exposed to a near-neutral brine. After flow, X-ray computed tomographic imaging showed the formation of a highly porous altered layer for samples exposed to the acidic brine, and no significant alteration for samples exposed to the near-neutral brine. This altered layer is created primarily from the dissolution of calcite grains, leaving a porous intact matrix of non-reactive minerals at the fracture surface. During the second stage, samples were sheared in a tri-axial testing apparatus that independently supplied confining pressure and differential pore pressure along the fracture length at prescribed sliding velocities, independently measuring friction and permeability. Samples were confined to 3 MPa effective stress with shear velocity down- and up-steps of 1 um/s and 10 um/s.
X-ray computed tomographic imaging following the shearing experiments revealed that the porous altered fracture surface layers had collapsed into a layer of fine particles that filled the fracture aperture, effectively sealing the fracture. This is confirmed by permeability measurements during the initial compaction in the shearing apparatus, where the permeability of samples with an altered surface layer decreased one order of magnitude lower than the samples without an altered layer. This difference in permeability between the sample sets persisted through the entire shearing experiments. Results from the shearing experiments show that the altered sample set exhibited both lower frictional strength and stability. This is because for the samples exposed to non-reactive brine, the micro-roughness at the fracture interface results in interlocking micro-asperities that both increase the fracture strength and stability. For the altered samples, however, the layer of fine grained particles filling the aperture separate the two fracture walls, preventing the formation of interlocking micro-asperities. As a result, we describe the fracture surfaces as “surfing” on the layer of fine-grained particles.
Dissolution within porous media is a critical process in many environmental and geological settings. The formation and evolution of soils, rocks, and landscapes (Buhmann and Dreybrodt, 1985; Brantley, 2008; Jin et al., 2010), efficiency of carbon capture in geological reservoirs (Matter and Kelemen, 2009), and the weathering of man-made structures are all highly dependent on the rates at which minerals dissolve. However, the factors governing dissolution in geological media are not yet well understood. In general, rock dissolution rates are governed by fluid composition, fluid flow rates, and the surface area of the mineral–fluid interface (Lasaga, 1998). As a surface becomes rougher, the surface area in contact with a reactive fluid increases, and it is therefore often assumed that rough surfaces dissolve more rapidly than smooth surfaces (Fischer and Luttge, 2007). However, a number of studies have shown a complex relationship between roughness and reactivity (e.g., Anbeek, 1992; Gautier et al., 2001). Emmanuel and Levenson (2014) found that erosion rates in fine-grained micritic limestone blocks are as much as two orders of magnitude higher than rates estimated for coarse-grained limestones. AFM imaging suggested this is the result of rapid dissolution along micron- scale grain boundaries, followed by mechanical detachment of tiny particles from the surface. Such chemo-mechanical processes may be the dominant erosional mode for fine-grained carbonate rocks in many regions on Earth. This erosion extends to the micron scale, and grain detachment can be a crucial mechanism controlling denudation rates in carbonate terrains.
In order to better understand this weathering process we undertook a series of experiments looking at the weathering of carbonates. 5/8” x 5/8” cores of four limestones (Netzer, Shiuvta, Carthage Marble, Texas Cream) were exposed to flowing water at 30°C and several pH values to mimic the weathering process. Annular Cd masks of stepped sizes, and small beam stepped locations analyses, were used to analyze the weathering structure by (U)SANS and (U)SAXS as a function of distance from the edge. SEM analysis was also used to look at the pore structure and surface weathering. The results, both in terms of core/rim variations and pore size dependence, were found to be strongly dependent on initial permeability and rock structure, as well as time and pH.
Polymer-electrolyte fuel cell (PEFC) is a promising energy conversion technology with high thermodynamic efficiency, power density and zero-emission. Due to their low cost and material abundancy, PGM-free electrodes are promising candidates for meeting 2020 cost targets set by U.S. Department of Energy (DOE) [1]. However, to compensate the lower volume-specific activity, these catalyst layers are about an order of magnitude thicker compared to conventional Pt-based electrodes. Due to larger thickness, mass-transport and Ohmic losses can be significant within these PGM-free electrodes. Previous modeling efforts, which treat catalyst layer as a porous media with effective properties obtained with imaging methods (FIB-SEM, X-ray tomography), lack in capturing through-thickness morphology inhomogeneity. To better understand the influence of the electrode microstructure on the transport processes, we bridge micro and nano-scales obtained with X-ray computed tomography (CT) to model the transport processes in the PGM-free electrodes.
Two iteration algorithms for modeling scale-bridging between imaged morphology with micro X-ray CT and nano X-ray CT are presented. Micro X-ray CT can capture larger cracks and morphological inhomogeneity. Nano X-ray CT with higher resolution serves as a powerful tool to characterize the meso and nano-scale, which cannot be captured by micro X-ray CT due to imaging resolution limits. In both algorithms, the micro-scale domain is discretized in z-direction into a finite grid, where the effective properties of each grid cell are computed with nano-scale model at each iteration step. Furthermore, the micro- and nano-scale models are linked through boundary conditions. The algorithms work for Poisson’s equations, where two boundary conditions are needed to solve the ODEs. In algorithm 1, micro-scale model generates Dirichlet boundary condition on node i and Neumann boundary condition on node i+1 for nano-scale model to compute the Dirichlet boundary condition for node i+1. Then this Dirichlet boundary condition is fed back to micro-scale model to renew the Neumann boundary condition on node i+1 until solution convergence is reached. In algorithm 2, micro-scale model generates Dirichlet boundary conditions on both sides of the element and passes them to nano-scale model to compute the Neumann boundary condition for node i+1. Then this information is fed back into micro-scale model to update the Dirichlet boundary condition on node i+1. The procedure repeats until the solution converges.
Transport processes in idealized geometries are calculated to assess the algorithms’ effectiveness and to study the convergence rate. Then the two algorithms are applied to study the transport processes in the PGM-free electrodes of the PEFCs. A case study with the reaction rate as a function of local concentration has been studied. The spatial variations of the effective transport coefficient and reaction rate have been shown through the numerical results. The effective diffusivity and tortuosity are proved to be not only determined by the geometry of the material, but also influenced by the reaction rate and boundary conditions in the material. The new algorithms fully consider the influence of microstructure and spatial variations.
References:
[1] A. Kongkanand, M.F. Mathias, The Journal of Physical Chemistry Letters, 7 (2016) 1127-1137.
Mudrocks are the dominant rocks in the Earth crust and are as well noted for their heterogeneity at several orders magnitude of scales. This implies a significant challenge in relating observations at varying scales to one another. A unique attribute that controls petrophysical properties of mudrocks is their microstructure which also controls fluid movement within them. Due to the fine-grained size of mudrocks, high-resolution measurement is required to reveal their microstructural characteristics. High-resolution scanning electron microscopy has advanced knowledge about microstructure of mudrocks but further development of fast and reliable methods to accurately determine the micron-submicron features of mudrocks is still on-going.
In this paper we present several information on mudrocks properties derived from scanning electron microscopy including grain size, grain-orientation, mineralogy and porosity, pore size distribution. The method involves multiple large-area, high-resolution scanning electron microscopy through automated acquisition and stitching of backscattered images (BSE) from polished thin-sections in combination with machine learning segmentation and energy dispersive X-ray analysis (EDX). Preliminary results obtained from the methodology was applied to seven mudrocks from deep-water identified as hemipelagites from the New Zealand Continental Slope (IODP Expedition 317) and Iberian Peninsula (IODP Expedition 339).
Grain size analysis shows that all the samples are within silt-mud class size. Orientation analysis indicates that randomly representative areas within each sample are heterogeneous; displaying a combination of preferred orientation direction and random orientation. The samples were differentiated into clay dominant and calcite dominant based on EDX. Porosity on representative areas for individual samples are heterogeneous reflecting areas that are tightly porous, partially porous and highly porous area. Interestingly, calcite dominated samples showed tighter porosity compared to clay dominated samples. However, there is no significant difference among representative areas per sample in terms of pore size distribution.
Automated image analysis of large area, high-resolution montages presented herein, is fundamental to revealing heterogeneity and deriving plethora of information on mudrock microstructure. The process minimises human subjectivity and bias but the limitations to the workflow are the time involved for individual runs and large amount of computer memory required. In addition, cracks in the sediment samples resulting from drying, preparation of polished thin-section as well as stress relaxation during coring, restrict the area available for high-resolution large-area imaging to that between the cracks. This method is very significant for improved understanding of subsurface mudrocks and their capacity for fluid movement and storage.
It is proposed in this talk to consider multiple resolution observations of dentine materials synthetized at multiple scales with a general inverse problem approach by combining confocal e.g. with SEM observations at different scales. Thus correlative microscopy aims to access a large range of scales for a given region by combining what is impossible by using one instrument alone.
Dentin is the bulk material of the tooth and presents a complex hierarchical structure. At the macroscopic scale, it can be seen as a first approximation as a homogeneous material located in between enamel and pulp cavity. But at the tissue level, it presents a microstructure made of tubules, peritubular dentin and intertubular dentin (entity dimension : a few microns). At a lower scale, the intertubular dentin is a composite made up of collagen fibers and hydroxyapatite crystals (entity dimension : 100nm).
Knowing the influence of the topology at each scale is of utmost importance for dental restoration. It is thus strongly needed to provide a robust and durable link between the restorative material and the sane biomaterial. Local defaults will create stress concentration which will be the source of crack initiation either immediately or later due to cyclic loading and fatigue phenomena. On the other hand low resolution is often required to survey large regions, for example to locate and image sparse features such as critical features, but for which structural information is required at much higher resolution.
The proposed mathematical approach is viewed as an inverse problem: it starts a material structure at a given resolution with a given larger scale apparatus. dispersed topography at diverse point sampled in the structure, find the fine fitted topography at the fine scale everywhere in the sample. To reach this goal we use concepts such as variational image restoration and texture synthesis together with PDE/FEM image analysis.
In this talk, dentin 3D imaging using Focused Ion Beam-Scanning Electron Microscopy (FIB-SEM) and Confocal Laser Scanning Microscopy (CLSM) is performed and used in order to get some physical properties such as permeability and elasticity modulus at different scales.
OBJECTIVES/SCOPE:
Existing rarefied gas flow models cannot accurately unify various flow mechanisms by empirical methods and overlook the van der Waals effect. In this paper, a model for non-ideal rarefied gas flow in nano- and micro-porous media is developed based on the well-recognized Bravo’s conceptual layered model with the rigorous interpretation. The gas transport behavior in nanopores can be simulated by the developed model which can be integrated with hydraulic fracturing models to optimize the production performance of shale gas reservoirs.
METHODS, PROCEDURES, PROCESS:
The cross-section of a nano-capillary can be divided into two zones based on Bravo’s model, i.e., an inner circular zone where the viscous flow behavior mainly exists due to the dominant intermolecular collision and an outer annular zone where non-equilibrium phenomenon exists and the classic constitutive relation breaks down due to a lack of intermolecular collisions. We proposed a virtual boundary between two zones which is determined by Kennard’s collision model, where the radius of inner zone is correlated with the fraction of intermolecular collisions. The convective and diffusive fluxes are rigorously integrated based on the virtual boundary. The mass flux, contributed by different transport mechanisms, thus could be analyzed by varying the Knudsen number. Subsequently, non-ideality property of rarefied gas is characterized by incorporating a compressibility correlation and real gas viscosity function. Physical and numerical experiments show the support of the new formulation and provide approaches to obtain apparent permeability and a generalized Klinkenberg’s parameter which is a function of Knudsen’s number.
RESULTS, OBSERVATIONS, CONCLUSIONS:
The newly proposed model allows for determination of the pressure dependence of the Klinkenberg parameter across the transition flow regime and yields the most accurate prediction compared with five existing models. The apparent permeability does not change obviously when pressure is over 10 MPa and pore size is larger than 100 nm. Although the surface roughness can significantly reduce the apparent permeability, its impact is minor when the pressure is higher than 10 MPa. The molar gas flux declines significantly by incorporating the real gas effect into the model, leading to a decreasing apparent permeability. Sensitivity analysis shows the apparent permeability is found to be strongly dependent on pore size and weakly dependent on roughness. Finally, it is found that Knudsen diffusion dominates the flow performance with a proportion larger than 60% in small pores (e.g., ≤ 50 nm) at the low pressures (e.g., ≤ 0.2 MPa).
NOVELTY:
Instead of the empirical coefficients commonly used in most existing models, the weight coefficient of viscous flow and Knudsen diffusion in the proposed layered model is analytically derived based on Bravo’s layered model. This work also provides approaches to obtain a generalized Klinkenberg’s parameter. In addition, a multi-objective optimization method is adopted to enhance the conveniences of searching local optimal fitting parameters in empirical correlations.
Geologic Carbon Storage is one method available to mitigate excess carbon dioxide produced at point sources. X-Ray micro-computed tomography provides the resolution requirements necessary to image in situ contact angles (θ) at representative conditions; however, experimental data is limited and varies among materials and temperature settings used in literature (Andrew et al., 2014, Lv et al., 2017, Tudek et al., 2017). To further expand the practice of X-ray tomography, experiments were done on Mt. Simon and Nugget sandstone, adding more contact angles to the library of data obtained with this technique. The Mt. Simon core was subject to one imbibition cycle. The Nugget sandstone contained a preferential pathway spanning the length of the core and was further analyzed to correlate the effect of connectivity to the resultant θ measurements. Two separate experiments were completed using the same Nugget core: (1) drainage and two imbibition cycles and (2) drainage and one imbibition cycle. In the first Nugget test, scCO2 remobilized and became trapped in new pores along the flow path after the second imbibition cycle. Detailed analyses were completed for the Nugget sandstone: θ were measured and remeasured after the second imbibition cycle on scCO2 ganglia that remained trapped between brine floods and θ within a pore were measured after each imbibition cycle to determine variation. For comparison with the Mt. Simon core, the sessile drop method was completed on additional Mt. Simon samples.
The θ range was reviewed for both the Nugget and Mt. Simon where θ were measured manually in three different planes throughout each core and summed to an average value. The results were compared to determine any variation between sandstones. An automated algorithm designed to measure θ was developed to verify the sessile drop method results, to check the average θ values for both Nugget and Mt. Simon, and to measure more θ throughout the cores for more representative data. From manual measurements, Mt. Simon sandstone resulted in an average θ of 35° and a range from 5° to 120° while the Nugget sandstone resulted in an average θ of 56° and a range of 5° to 145°. The average contact angles for each classify both cores as weakly water-wet systems while the ranges suggest an intermediate-wet system within some pores. The Nugget average θ was 20° higher than the Mt. Simon sandstone and had a broader θ range by 25°. Different contact angles in two sandstone cores exposed to similar conditions indicate additional factors need to be considered.
Unlike macroscopic objects, any system of nanometric size shows characteristics that strongly depend on its size and geometric form. It is the consequence of the fact that the major part of atoms (or molecules) of nano-object is located at its surface, their cohesive energy is smaller than for the atoms in the bulk.
Here we show that when a fluid is confined in nano-volume, delimited by non-interacting pore walls, its density is heterogeneous, decreases close to the pore wall, and, on average, is smaller than the density of bulk fluid. The heterogeneity of distribution of fluid density, resulting from the nano-confinement, progressively weakens when the pore size increases, and totally disappears for pores larger than 5 nm. On the other side, in the limit of very small pores, the fluid density approaches the ideal gas value. This effect should be distinguished from the well know heterogeneity of density of fluids adsorbed in nanopores, driven by the difference between the strength of fluid-fluid and fluid-pore wall interactions, that varies with the distance from the pore wall.
The reported observation has non-trivial influence on evaluation of excess/total adsorption in nanopores, as these two quantities are calculated assuming the known – and homogeneous – bulk density of gas in the pore. Additionally, the gas density in the pores depends on the definition of the pore volume which is neither straightforward nor unique. The right estimation of both: pore volume and gas density is essential for quantitative interpretation of experimental adsorption isotherms: evaluation of pore size distribution and of the amount of adsorbed gas. We analyze this problem on an example of five gases: H2, CH4, the two intensively studied energy vectors, and N2, Ar, and Kr, commonly used for characterization of porous structures. For H2, the distributions of densities of gas confined in adsorbing and not adsorbing pores are compared and commented.
Clay minerals are ubiquitous in the subsurface: they are found in CO2 sequestration targets (e.g., sandstones) and in the seals above them, and are major constituents of unconventional shale plays considered for natural gas recovery. A significant fraction of the porosity in clay-rich systems is occupied by micro- and meso-pores that provide a large surface area for physical and chemical interactions with the surrounding fluids. Of particular interest to this study is the adsorption behaviour of CO2 and CH4 that leads to the trapping of these gases in the porous structure at liquid-like densities. From a practical perspective, gas adsorption can lead to (i) an increase of storage capacity in reservoirs having larger clay contents, (ii) an advance in storage safety by limiting gas diffusion through cap rock seals, and (iii) an enhancement of gas production from tight shale formations through an adsorption/desorption (CO2/CH4) process.
Supercritical gas adsorption studies on clay and shale samples that address these aspects are found in the literature, but the picture is still far from being complete. The main reason for this is the intrinsic difficulty in performing these experiments at subsurface conditions (high-pressure and temperature), and in their description, because the interactions between the gases and the rock’s constituents (clays, carbonate minerals and organic matter) are quite complex. The lack of a systematic evaluation on the effects of temperature, pore structure and pore chemistry on gas adsorption over the relevant pressure range precludes the development and validation of theoretical models for gas adsorption in rocks that have a sufficient degree of predictive ability. The latter is a necessity if laboratory observations that are inevitably limited in their probed rock volume are to be used to make useful estimates of process-relevant parameters, such as Gas-In-Place and storage capacity.
We report results from a systematic experimental investigation on the adsorption properties of CO2 and CH4 over a wide range of conditions (0-25 MPa and 40-80°C). The systems considered include pure clay minerals (e.g., Na-montmorillonite), shale samples from various (potential) plays (Eagle Ford, Utica and Bowland Shale), as well as reference materials with well-defined surface chemistry and pore structure (micro- and meso-porous zeolites, carbons and silica). Data are interpreted using appropriate quantitative measures, such as the excess adsorption and Henry constants. The measured adsorption isotherms are described using a Lattice Density Functional Theory (LDFT) model that uses as input parameter the pore structure of the material (measured from conventional cryogenic adsorption experiments). As such, the modelling approach is more rigorous, has predictive capability and represents a significant departure from conventional empirical approaches that use Langmuir- or BET-type of models.
Compositional variation in multi-component system caused by adsorption and confinement in organic nanopores leads to capillary condensation and trapping of concentric hydrocarbon liquids. The objective of this paper is to show the presence of capillary condensation in kerogen pores, and argue that pressure depletion/fluid expansion is no longer effective and EOR process is required for these pores. We quantify stripping effects of CO2 on the condensates recovery from the kerogen nanopores and discuss the fundamentals of cyclic-CO2 injection into source rocks. Investigation is carried out using a multi-component hydrocarbon mixture representing fluids produced f