Article

Solvent Diffusion and Dispersion in Partially Saturated Porous Media: An Experimental and Numerical Pore-level Study

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Recent research on unlocking the solvent dispersion as a physical blending process in porous media has been mainly focused on core-scale observations. Pore-level studies of multiphase displacements can help to develop models that correlate rock macro-scale characteristics with its small-scale features, particularly for unsaturated rocks where the intricate fluid arrangements cause dynamic events to be localized in preferential pathways. In this work, to draw new insights into the physics of pore-level interfaces and crystallize the role of small-scale phenomena on the efficiency of solvent-aided bitumen recovery processes, numerical simulations coupled with experiments are conducted in pore-level domains. We simulate fluid flow and transport phenomena through millimeter-sized three-dimensional slabs of consolidated and unconsolidated packings of grains representing the geological rock types of the McMurray formation. We propose a robust numerical workflow for simulation of miscible-floods in unsaturated porous media and investigate the impact of matrix heterogeneity, connate water, cementation, and injection velocity on the longitudinal dispersion coefficient. In particular, primary drainage is simulated at low capillary numbers resulting in two-phase fluid occupancies through pore space domains. Finite element simulations are then carried out in order to solve the mixing advection-diffusion equations within the water-free pore space, and lastly, the effluent history is analyzed to predict the dispersion coefficient in both fully- and partially-saturated conditions and evaluate the efficiency of miscible displacements in the presence of microheterogeneities. A new analytical model for calculation of dispersivity, together with the numerical simulation results, is utilized to adjust a general model for the prediction of longitudinal dispersion coefficient in unsaturated sandy porous media of either uniform or non-uniform grains. Moreover, miscible-flood experiments at low to high injection velocities are conducted in a transparent glass micromodel. According to the results, rock micro-heterogeneities and immobile water both increase the longitudinal dispersion coefficient, and two-phase equilibria control the velocity field by creating connected regions of brine and low resistance oil-filled channels and consequently influence the solute transport and mixing processes. The effect of viscosity contrast on the longitudinal dispersion coefficient is also noteworthy, as the viscous fingering at unfavorable viscosity ratios widens the mixing zone.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... [14] Recent studies have focused mainly on diffusion measurements and applications. [15][16][17] Yassin and coworkers', [18] for instance, used hydroxypropyl methylcellulose (HPMC) as a model system to implement and develop the current understanding of the role of diffusion in swelling polymers. Using terahertz pulsed imaging, a tool combining high contrast and high spatial resolution between different phases through changing refractive indices and fast timeresolved analysis, they investigated the growth kinetics of the process as well as followed the water diffusion front as it propagated through the porous tablets of HPMC. ...
... In this respect, porosity is used to describe the reduction in the cross-sectional area and quantify the tortuosity. [15] The diffusion rate of the UP resin was measured in the compressed EG tablets, opposing the gravity direction and by vacuum as a driving force (Fig. 3). The diffusion rates of the UP-LV and UP-HV resins through the EG tablets are shown in Fig. 4. The EG1, EG2, and EG3 particles had the densities 0.0038-0.0049 ...
Article
Polymer/expanded graphite (EG) nanocomposites have great importance in many industrial applications mainly due to their high electrical/thermal conductivity or flame retardancy. However, to fully employ the benefits of polymer/EG nanocomposites one must consider the high degree of porosity of EG. The high degree of porosity of EG can deteriorate the composites’ mechanical properties if the polymer chains cannot diffuse completely into the EG pores. In this article, an insight is given into the diffusion of unsaturated isophthalic polyester (UP) resin, consisting of a combination of maleic anhydride and isophthalic anhydride in the resin backbone, with two viscosities, into the pores of the EG particles of various degrees of porosity. The diffusion experiments were carried out on compressed EG tablets with the same density but different porosity due to the different porosity of the EG particles. The results showed that the diffusion rate of the UP resin with higher viscosity slightly decreased when the EG porosity decreased but, in the opposite way, it strongly increased for the low viscosity UP resin. The EG nanocomposites samples were molded at varying pressures. The micrographs of the fractured surfaces of the EG nanocomposites showed that the EG pores were not filled with resin, thus the EG nanocomposites had residual pores. It was found that composites containing EGs with higher expansion ratio and larger particles and pores showed larger residual pores. Furthermore, the composites prepared with the more viscous UP resin showed more residual pores. By applying a pressure of 10 bar instead of 1 bar, a reduction of 7–20% in the residual pores of the nanocomposites was observed which led to improved mechanical properties by up to 20% in flexural strength for the EG with the highest expansion ratio.
... More recently, X-ray CT images have been applied to model flow and transport processes using direct numerical simulation methods (Bultreys et al. 2016). One example of numerical solutions is those of the Stokes equations, which as being linear, can be solved for a stationary flow condition using Finite Difference Method (FDM, Wiegmann 2007), Finite Elements Method (FEM) (Mohammadmoradi et al. 2018), Finite Volume Method (FVM, Guibert et al. 2015), Lattice Boltzmann Method (LBM, Khan et al. 2012) or Smooth Particle Hydrodynamics (SPH, Dal Ferro et al. 2015. The average flow velocity could then be selected by a simple scaling of the velocity field. ...
... This software uses an analytical solution of the 1D macroADE to solve the inverse problem for D h by performing a least-squares minimization. This method is similar to the one proposed by Mohammadmoradi et al. (2018). Each numerical column experiment was analyzed considering individual BTCs obtained at L 1 , L 2 , L 3 to compute dispersion coefficients D h 1 , D h 2 , D h 3 , respectively. ...
Article
Full-text available
X-ray micro-computed tomography (\(\mu\)CT) can produce realistic 3D-images of the pore structure of a material. Extracting its geometry enables the computation of effective properties of the material—such as the permeability (k) and the hydrodynamic dispersion coefficient (\(D_h\))—, through the solutions of the Stokes equation (SE) and Advection-Diffusion equation (ADE), respectively. In this study, the effect of the image resolution on these properties is discussed. For such purpose, four different resolutions are evaluated for both a real sample of Fontainebleau sand and a numerically generated sample created by degrading the Fontainebleau image with highest resolution. The SE was computed using the commercial software GeoDict. To solve the ADE, a Finite Volume software was developed which includes a high order total variation diminishing scheme for advection. The analysis of dispersion was based on numerical breakthrough curves. Our model was tested in a large range of Peclet numbers (Pe) and travel distances, accurately describing the transition between diffusion and advection dominated regimes of dispersion. The \(D_h\) exhibits a linear increase with travel distance for Pe \(> 10\). This classical effect increases with increasing Pe. The percentage change on k and \(D_h\) increases with decreasing resolution in agreement with the corresponding behavior of porosity, specific surface and pore size distributions. The images directly scaled with the \(\mu\)CT showed more discrepancy than the numerically scaled images. The criteria to estimate the quality of permeability from the pore size distribution proposed on our previous study remains valid. The \(D_h\) is less sensitive to resolution than k.
... Hence, experiments cannot be considered as a general approach to this issue. However, unavoidable details must be considered in experimental procedures, and interested readers are referred to Nilsson and Stenstrom [12], Mohammadmoradi et al. [13], Jia et al. [14], Hibi and Kashihara [15], Chan et al. [16], Unsworth et al. [17]. On the other hand, analytical relations are obtained by simplified assumptions for calculating the ratio of effective mass diffusion coefficient to the binary gas diffusion coefficient in porous medium and have been presented in the literature by Maxwell [18], Nam and Kaviany [19], Andisheh-Tadbir et al. [20], Xiao et al. [21], Xu et al. [22]. ...
Article
Pore structure has a significant effect on transport phenomena inside porous media. This effect can be considered in simulations by using suitable transport coefficients. Some correlations are reported in the literature, which can be applied for simple pore geometries. In the present study, a pore-scale simulation approach is presented to determine mass diffusion coefficient considering molecular and Knudsen diffusion in a fibrous porous medium as a complex porous geometry. The methodology is employed for species in catalytic combustion of Methane inside a fibrous porous structure. The effect of Solid Volume Fraction (SVF), fibers orientation, and diameter are discussed in different temperatures. It is found that SVF plays the dominant role in mass diffusion, specifically above 600K. Mass diffusion coefficients obtained in the present study and flow permeability, conduction and radiation heat transfer coefficients from the previous study are used to simulate methane combustion inside the fibrous structure on the macro scale. An experimental setup is developed for validation. The results indicated that the simulation could well predict the temperature distribution, also 6.4% error in estimating the Methane conversion rate was observed. Due to the low Peclet number, the concentration of CH4 and O2 decreased unexpectedly before entering the catalytic zone.
... (2) Miscible flooding model Different from immiscible and near-miscible flooding, the interfacial tension between CO2 and oil disappears under the miscible condition [3,19], assuming the following: 1) CO2 can be fully mixed with oil, but the resultant oil swelling, oil extraction, and chemical reaction process can be ignored [27]; 2) CO2 and oil can be considered Newtonian fluids [28]; 3) the change in the diffusion coefficient is small and can be ignored [27]; 4) classical Fick's law could be valid for cases with ideal binary mixtures [29]; then, the continuity, fluid flow, and convection-diffusion equations under the miscible condition can be written as follows [30]: ...
Article
Full-text available
Miscible and near-miscible flooding are used to improve the performance of carbon-dioxide-enhanced oil recovery in heterogeneous porous media. However, knowledge of the effects of heterogeneous pore structure on CO2/oil flow behavior under these two flooding conditions is insufficient. In this study, we construct pore-scale CO2/oil flooding models for various flooding methods and comparatively analyze CO2/oil flow behavior and oil recovery efficiency in heterogeneous porous media. The simulation results indicate that compared to immiscible flooding, near-miscible flooding can increase the CO2 sweep area to some extent, but it is still inefficient to displace oil in small pore throats. For miscible flooding, although CO2 still preferentially displaces oil through big throats, it may subsequently invade small pore throats. In order to substantially increase oil recovery efficiency, miscible flooding is the priority choice; however, the increase of CO2 diffusivity has little effect on oil recovery enhancement. For immiscible and near-miscible flooding, CO2 injection velocity needs to be optimized. High CO2 injection velocity can speed up the oil recovery process while maintaining equivalent oil recovery efficiency for immiscible flooding, and low CO2 injection velocity may be beneficial to further enhancing oil recovery efficiency under near-miscible conditions.
... The diffusion and dispersion coefficients were used to adjust the simulation as closely as possible to the observed degradation results. These parameters were chosen as they have been shown to alter the transport of species in porous media [62]. According to our results, the best decolorization results are achieved for combinations of filters with immobilized enzymes and nanocomposites prepared in the presence of acetone. ...
Article
Full-text available
Removal of azo and diazo dye content from textile industry wastewaters is crucial due to their environmental impact. Here, we report on the use of the fungal laccase from Pycnoporus sanguineus CS43 immobilized on silica nanoparticles and entrapped in textile-based filters for the degradation of Congo Red. Laccase immobilization and synthesis of the nanocomposites were carried out by two different methods, one in the presence of acetone and the second using water as solvent. This led to a change in the hydrophobicity of the obtained biofilters. Successful preparation of the nanocomposites was confirmed via FTIR spectroscopy. Changes in the secondary structure of the enzyme were inspected through the second derivative of the FTIR spectra. Six different types of filter were fabricated and tested in a continuous flow bioreactor in terms of their decolorization capabilities of Congo Red. The results indicate removal efficiencies that approached 40% for enzymes immobilized on the more hydrophobic supports. Backscattered electron (BSE) images of the different filters were obtained before and after the decolorization process. Percentage of decolorization and activity loss were determined as a function of time until a plateau in decolorization activity was reached. Experimental data was used to recreate the decolorization process in COMSOL Multiphysics ® (Stockholm, Sweden). These simulations were used to determine the proper combination of parameters to maximize decolorization. Our findings suggest that the treatment of textile-based filters with immobilized laccase in conjunction with hydrophobic nanocomposites provides a suitable avenue to achieve more efficient laccase dye decolorization (39%) than that obtained with similar filters treated only with free laccase (8%). Filters treated with silica-based nanocomposites and immobilized laccases showed an increase in their decolorization capability, probably due to changes in their wetting phenomena.
... This methodology has proved useful for porous media data treatment. 40 In this way we calculated the number of pixels of the gas-liquid interfacial perimeter. Since the height of the micro device is constant (70 μm), a was calculated as the ratio of interfacial perimeter per bed area. ...
Article
We describe the co-current flow pattern of gas and liquid through micro-fabricated beds of solid and pillars under variable (i) capillary number, (ii) contact angle or wettability and (iii) pillar arrangement, i.e. modifying the distance between pillars or their size and comparing regular with more chaotic systems. Laser-induced fluorescent microscopy and image analysis are used to study the hydrodynamic interactions in terms of dynamics, liquid hold-up, and gas-liquid interfacial area per reactor volume. Those parameters provide insights into the multiphase flow patterns in these systems, how to control them, maximize mass transfer rate and unlock the potental of microreactors to reveal further intrinsic information.
... Solvent extraction is one of the most popular methods potentially for the industrialization of treating oil sands, due to its high extraction efficiency. However, the extraction of traditional solvents [7][8][9]has many problems, such as difficulty in recovery, volatility and so on [10][11][12]. ...
Article
In this article, the composite switchable water was prepared with hydrophobic tertiary amine and switchable water, which provides a new method for washing oil sands. Firstly, pH and ¹HNMR were used to demonstrate that N,N-diethylethanolamine (DEEA) with the excellent CO2 response could be treated as potential switchable water. In addition, N,N-dimethylaniline (DMA) was chosen as a suitable hydrophobic tertiary amine, because DMA is a more steady molecular state than DEEA at the same pH. Then, the composite switchable water (DMA + DEEA(aq)) was prepared with DMA and DEEA solutions. There was the possible presence of inductive effect between DMA and DEEA detected by ¹HNRM and it was also found that DEEA reduced the surface tension. Next, the CO2-triggered switchable behavior of the composite switchable water has been demonstrated by pH and conductivity. Finally, a practical application of washing oil sands with the composite switchable water was accomplished. The results showed that the residual oil content of oil sands was reduced to 0.872 wt% from the original 4.64 wt% heavy hydrocarbons (the ratio of DMA in gram to DEEA solutions in mL was 4:15) and both DMA and DEEA solutions also were recycled during the process. In addition, a process for the extraction and separation of simulated oil (kerosene) with the composite switchable water was depicted to further demonstrate the behavior of dissolving oil.
Article
Spontaneous coal combustion was proved to cause methane accumulation and methane explosions in coal mine goafs. To avoid possible methane explosions, the empirical engineering measure, ventilation dilution, is proposed in coal mines though its disaster prevention mechanism has not been well understood. Through experimental and numerical simulations, the superposition effect of the air leakage and coal combustion-induced chimney effect was studied to reveal disaster prevention effect of ventilation dilution. Research results show that the high temperature area of coal combustion is steady and can provide continuous buoyancy force to form upward airflow even under ventilation dilution; the drifting methane accumulation is observed under the superposition effect of air leakage and upward airflow; ventilation dilution can weaken and even eliminate methane accumulation by overcoming chimney effect, but an increase in coal combustion temperature will enhance the upward airflow of chimney effect to cause methane accumulation again. The competitive relationship between the coal combustion-induced chimney effect and air leakage provides a new insight to study methane migration for the disaster formation and prevention mechanism.
Article
Full-text available
The viscosity of extra-heavy oils including bitumen can be reduced significantly by adding solvent such as toluene to enhance extraction, production and transportation. Thus, prediction of viscosity and/or rheology of bitumen-solvent mixtures has become necessary. More so, selecting a suitable rheological model for simulation of flow in porous media has an important role to play in engineering design of production and processing systems. While several mixing rules have been applied to calculate the viscosity of bitumen-solvent mixtures, rheological model to describe the flow characteristics has rarely been published. Thus, in this investigation, rheological behaviour of bitumen and bitumen-toluene mixtures (weight fractions of bitumen WB = 0, 0.25, 0.5, 0.6, 0.75, and 1 w/w) have been studied at the flow temperature (75 °C) of the bitumen and in the range of shear rates between 0.001 and 1000 s−1. The data was fitted using different rheological models including the Power Law, Cross Model, Carreau-Yasuda Model, and the newly introduced ones herein named as Cross-Logistic and Logistic models. Then, a computational fluid dynamics (CFD) model was built using a scanning electron image (SEM) of rock sample (representing a realistic porous geometry) to simulate pore scale flow characteristics. The observations revealed that the original bitumen exhibits a Newtonian behaviour within the low shear rate region (0.001 to 100 s−1) and shows a non-Newtonian (pseudoplastic) behaviour at the higher shear rate region (100 to 1000 s−1). Conversely, the bitumen-toluene mixtures show shear thinning (pseudoplastic) behaviour at low shear rate region (0.001 to 0.01), which appears to become less significant within 0.01 to 0.1 s−1, and exhibit shear independent Newtonian behaviour within 0.1 and 1000 s−1 shear rates. Moreover, except for the original bitumen, statistical error analysis of prediction ability of the tested rheological models as well as the results from the pore scale flow parameters suggested that the Power Law might not be suitable for predicting the flow characteristics of the bitumen-toluene mixtures compared to the other models.
Article
A methane explosion can take place when a methane concentration is within the explosive concentration range. Therefore, the methane migration around a coal combustion zone is important for assessing the disaster risk and revealing the disaster formation process. The thermal buoyancy effect has seldom been considered before, even though it could influence methane movement in coal mined-out areas. In this study, an integrated investigation using experimentation and modeling was conducted to explore the disaster formation process. For the experiment, a coal mine gob was heated locally to produce the buoyancy effect, and the consequent methane accumulation was observed. To explain this observation, a gas flow model reflecting the buoyancy effect was developed and verified against the experimental observations. Through this analysis, the formation processes for a methane explosion disaster in a coal mined-out area were revealed as follows: (1) coal combustion decreases the gas density to produce the buoyancy effect and create negative pressure in the combustion area; (2) this negative pressure leads to methane inflow, and the buoyancy effect causes upward methane movement; and (3) the local methane inflow and the local upward movement develop methane accumulation in the coal combustion area to form a methane explosion.
Article
Based on the law of conservation of mass and the law of linear infiltration, the theoretical model of slurry density in porous media under the combined influences of convection–diffusion–infiltration is established after applying the Laplace transform. Using a laboratory grouting test, the variation of cement slurry density for different grouting parameters and different times is studied. In addition, the differences between the experimental and theoretical values are analyzed. The results demonstrate that the experimental values of slurry density are positively correlated with time, while the average growth rates of each group are different. The experimental and theoretical values of slurry density are close in some test groups; the deviation rate between the experimental and theoretical values varies between − 12.9% and 30.1%. The average growth rate of the experimental value is smaller than that of the theoretical value. The results have some practical value and can provide a reference for construction design.
Article
Full-text available
Dispersion exists in many scientific and engineering applications, especially for CO2 enhanced gas recovery, which is a vital factor for controlling the contamination of remaining natural gas and gas recovery. In this paper, an in situ method for the dispersion coefficient measurement of liquid/supercritical CO2–CH4 in a sandpack using CT was proposed. The dispersion coefficient in the sandpack was obtained directly from the CO2 mole fraction profiles translated from a CT greyvalue image, which eliminate the deviation caused by the entry/exit effect. The finite difference method, Crank–Nicolson method, was applied to solve the advection dispersion equation for obtaining the dispersion coefficient. The breakthrough profile of the effluent gas was also analyzed and the apparent dispersion coefficient containing the entry/exit effect was measured using the dynamic column breakthrough method. The entry/exit effect enlarged the dispersion coefficient in the range of 14–23% under a water-free experiment according to the deviation between the two methods. And the dispersion coefficient with the sandpack containing residual water was smaller than that of the water-free condition, which was probably caused by the dissolution of CO2 in the displacing frontier into residual water. The dissolution stabilized the dispersion in the displacing frontier and resulted in the reduction of the dispersion coefficient.
Article
Full-text available
This paper presents experimental observations that delineate co-optimization of CO2 EOR and storage. Pure supercritical CO2 was injected into a homogeneous, outcrop sandstone under various miscibility conditions. A mixture of hexane and decane was used for the oil phase. Three experiments were run at 70°C and different pressures (1,300, 1,700 and 2,100 psi). Each pressure was determined using a PVT simulator to create immiscible, near-miscible and miscible displacements. Oil recovery, differential pressure and compositions were recorded during experiments. A co-optimization function for CO2 storage and incremental oil was defined and calculated using the measured data for each experiment. A compositional simulator was then used to examine gravity effects on displacements and derive relative permeabilities. Experimental observations demonstrate that an almost similar oil recovery was achieved during miscible and near-miscible displacements whereas about 18% less recovery was recorded in the immiscible displacement. More decane was recovered in the miscible and near-miscible displacements than the immiscible displacement. The co-optimization function suggests that the CO2 storage efficiency was highest in the near-miscible displacement and that the near-miscible displacement displayed the best performance for coupling CO2 EOR and storage. Numerical simulations show that, even on the laboratory scale, there were significant gravity effects in the near-miscible and miscible displacements. It was revealed that the near-miscible and miscible recoveries depend strongly on the end-point CO2 relative permeability which increased with miscibility.
Article
Full-text available
Subsurface gaseous-phase transport is governed by three gas transport parameters: the air permeability coefficient (k(a)), gas diffusion coefficient (D-p), and gas dispersion coefficient (D-H). Among these, D-H is the least understood due to hitherto limited research into the relationship between gas dispersion and soil physical characteristics. In this study, a series of advection dispersion experiments was performed on granular porous media to identify the effects of soil column dimensions (length and diameter), particle size and shape, dry bulk density, and moisture content on the magnitude of gas dispersion. Glass beads and various sands of different shapes (angular and rounded) with mean particle diameters (d(50)) ranging from 0.19 to 1.51 mm at both air-dry and variable moisture contents were used as granular porous media. Gas dispersion coefficients and gas dispersivities alpha = D-H/v, where v is the pore-gas velocity) were determined by fitting the advection-dispersion equation to the measured breakthrough curves. For all test conditions, D-H increased linearly with v. The test results showed that neither soil column length nor diameter had significant effect on gas dispersivity. Under air-dry conditions, higher gas dispersivities were observed for media with wider particle size distribution and higher dry bulk density. The minor effect of particle shape on gas dispersivity was found under both air-dry and wet conditions. Under wet conditions, the variations in gas dispersivity were mainly controlled by the air-filled porosity. An empirical model was also proposed for the prediction of gas dispersivity in granular, unsaturated porous media.
Article
Full-text available
Groundwater plumes originating from continuously emitting sources are typically controlled by transverse mixing between the plume and reactants in the ambient solution. In two-dimensional domains, heterogeneity causes only weak enhancement of transverse mixing in steady-state flows. In three-dimensional domains, more complex flow patterns are possible because streamlines can twist. In particular, spatially varying orientation of anisotropy can cause steady-state groundwater whirls. We analyze steady-state solute transport in three-dimensional locally isotropic heterogeneous porous media with block-wise anisotropic correlation structure, in which the principal directions of anisotropy differ from block to block. For this purpose, we propose a transport scheme that relies on advective transport along streamlines and transverse-dispersive mass exchange between them based on Voronoi tessellation. We compare flow and transport results obtained for a non-stationary anisotropic log-hydraulic conductivity field to an equivalent stationary field with identical mean, variance, and two-point correlation function disregarding the non-stationarity. The non-stationary anisotropic field is affected by mean secondary motion and causes neighboring streamlines to strongly diverge, which can be quantified by the two-particle semivariogram of lateral advective displacements. An equivalent kinematic descriptor of the flow field is the advective folding of plumes, which is more relevant as precursor of mixing than stretching. The separation of neighboring streamlines enhances transverse mixing when considering local dispersion. We quantify mixing by the flux-related dilution index, which is substantially larger for the non-stationary anisotropic conductivity field than for the stationary one. We conclude that non-stationary anisotropy in the correlation structure has a significant impact on transverse plume deformation and mixing. In natural sediments, contaminant plumes most likely mix more effectively in the transverse directions than predicted by models that neglect the non-stationarity of anisotropy. This article is protected by copyright. All rights reserved.
Article
Full-text available
The dispersion coefficient of displacing fluids within porous media is an important parameter to measure accurately for a range of applications. For example, enhanced gas recovery is an emerging technology where carbon dioxide is injected into natural gas reservoirs as a means of maintaining the pressure of the reservoir and to hence enhance the natural gas recovery, with the additional benefit of carbon dioxide sequestration. Research is currently being done to measure the dispersion coefficient of carbon dioxide into methane because carbon dioxide is a contaminant that reduces the value of the natural gas. Such dispersion experiments are difficult to perform in the laboratory using conventional rock core flooding equipment as erroneous contributions to the dispersion process from entry and exit effects (into and out of the rock core respectively) are impossible to eliminate completely. Previously, we have estimated the resultant error to be of the order of 25 % (Hughes et al., Int J Greenh Gas Control 9:457–468, 2012). Here we effectively deploy spatially resolved breakthrough curves which are characteristic of the dispersion process for carbon dioxide replacing methane. These are obtained from time-resolved 1D magnetic resonance imaging (MRI) profiles. By analysis of the additional dispersion between these time-resolved profiles, we are able to eliminate the entry/exit effects and obtain more accurate values for the dispersion coefficient. We demonstrate this using a model porous medium (mono-dispersed glass beads) in a cell capable of holding gas pressures up to 80 bar in a low-field MRI rock core analyser. Via simultaneous conventional IR analysis of the carbon dioxide effluent to determine the dispersion coefficient, we are also able to directly quantify the magnitude of the error due to these entry/exit effects.
Article
Full-text available
Dilution of solute plumes in groundwater strongly depends on transverse mixing. Thus, the correct parameterization of transverse dispersion is of critical importance for the quantitative description of solute transport. In this study we perform flow-through laboratory experiments to investigate the influence of transport dimensionality on transverse mixing. We present a high-resolution experimental setup to study solute dilution and transverse dispersion in three-dimensional porous media. We conduct multi-tracer experiments in the new 3-D setup and compare the results with the outcomes of analogous tracer experiments performed in a quasi 2-D system. We work under steady-state flow and transport conditions and consider a range of velocities relevant for groundwater flow (0.5-8 m/day). Transverse dispersion coefficients are determined from high-resolution concentration profiles at the outlet of the flow-through chambers (7 × 7 ports in the 3-D setup and 7 ports in the quasi 2-D system), considering conservative tracers with significantly different aqueous diffusion coefficients, namely fluorescein and dissolved oxygen. To quantify dilution in the 2-D and 3-D systems, we experimentally determine the flux-related dilution index using the flow rates and the concentrations measured at the inlet and outlet ports, and we propose semi-analytical expressions to predict its evolution with travel distance in uniform groundwater flow. The experimental results in the quasi 2-D and 3-D flow-through systems are consistent and show a compound-specific behavior of the transverse dispersion coefficient and its non-linear dependence on the seepage velocity in both setups. The degree of dilution and the compound-specific effects of transverse dispersion are considerably more pronounced in 3-D than in quasi 2-D transport systems.
Article
Full-text available
Steam-Assisted Gravity Drainage is effective at extracting oil from heavy oil and bitumen reservoirs. However, steam generation by natural gas combustion can render the process uneconomic. It has been suggested that addition of injected solvents improves oil rates or at least maintains similar oil rates with reduced steam. This is the basis of the Expanding Solvent -Steam Assisted Gravity Drainage process. The central idea is that steam plus solvent is better than steam alone to recover heavy oil and bitumen. In this research, the steam and solvent injection strategy is designed by optimizing the cumulative steam-oil ratio by using simulated annealing and a commercial reservoir simulator.
Article
Full-text available
In this work, the constitutive relationship between capillary pressure (Pc), saturation (Sw), and fluid-fluid interfacial area per volume (IFA) is characterized using computed microtomography for drainage and imbibition experiments consisting of a nonaqueous phase liquid and water. The experimentally measured relationship was compared to a thermodynamic model that relates the area under the Pc - Sw curve to the total IFA, an, and the capillary-associated IFA, anw. Surfaces were fit to the experimental and modeled Pc - Sw - an and Pc - Sw - anw data in order to characterize the relationship in three dimensions (3D). For the experimental system, it was shown that the Pc - Sw - an relationship does not exhibit hysteresis. The model is found to provide a reasonable approximation of the magnitude of the 3D surfaces for an and anw, with a mean absolute percent error of 26% and 15%, respectively. The relatively high mean absolute percent errors are primarily the result of discrepancies observed at the wetting- and nonwetting-phase residual saturation values. Differences in the shapes of the surfaces are noted, particularly in the curvature (arising from the addition of scanning curves and presence of an - Sw hysteresis in the predicted results) and endpoints (particularly the inherent nature of thermodynamic models to predict significant anw associated with residual nonwetting-phase saturation). Overall, the thermodynamic model is shown to be a practical, inexpensive tool for predicting the Pc - Sw - an and Pc - Sw - anw surfaces from Pc - Sw data.
Article
Full-text available
It is known that dispersion in porous media results from an interaction between convective spreading and diffusion. However, the nature and implications of these interactions are not well understood. Dispersion coefficients obtained from averaged cup-mixing concentration histories have contributions of convective spreading and diffusion lumped together. We decouple the contributions of convective spreading and diffusion in core-scale dispersion and systematically investigate interaction between the two in detail. We explain phenomena giving rise to important experimental observations such as Fickian behavior of core-scale dispersion and power-law dependence of dispersion coefficient on Péclet number. We track movement of a swarm of solute particles through a physically representative network model. A physically representative network model preserves the geometry and topology of the pore space and spatial correlation in flow properties. We developed deterministic rules to trace paths of solute particles through the network. These rules yield flow streamlines through the network comparable to those obtained from a full solution of Stokes’ equation. Paths of all solute particles are deterministically known in the absence of diffusion. Thus, we can explicitly investigate purely convective spreading by tracking the movement of solute particles on these streamlines. Then, we superimpose diffusion and study dispersion in terms of interaction between convective spreading and diffusion for a wide range of Péclet numbers. This approach invokes no arbitrary parameters, enabling a rigorous validation of the physical origin of core-scale dispersion. In this way, we obtain an unequivocal, quantitative assessment of the roles of convective spreading and diffusion in hydrodynamic dispersion in flow through porous media. Convective spreading has two components: stream splitting and velocity gradient in pore throats in the direction transverse to flow. We show that, if plug flow occurs in the pore throats (accounting only for stream splitting), all solute particles can encounter a wide range of independent velocities because of velocity differences between pore throats and randomness of pore structure. Consequently, plug flow leads to a purely convective spreading that is asymptotically Fickian. Diffusion superimposed on plug flow acts independently of convective spreading (in this case, only stream splitting), and, consequently, dispersion is simply the sum of convective spreading and diffusion. In plug flow, hydrodynamic dispersion varies linearly with the pore-scale Péclet number when diffusion is small in magnitude compared to convective spreading. For a more realistic parabolic velocity profile in pore throats, particles near the solid surface of the medium do not have independent velocities. Now, purely convective spreading (caused by a combination of stream splitting and variation in flow velocity in the transverse direction) is non-Fickian. When diffusion is nonzero, solute particles in the low-velocity region near the solid surface can move into the main flow stream. They subsequently undergo a wide range of independent velocities because of stream splitting, and, consequently, dispersion becomes asymptotically Fickian. In this case, dispersion is a result of an interaction between convection and diffusion. This interaction results in a weak nonlinear dependence of dispersion on Péclet number. The dispersion coefficients predicted by particle tracking through the network are in excellent agreement with the literature experimental data for a broad range of Péclet numbers. Thus, the essential phenomena giving rise to hydrodynamic dispersion observed in porous media are (1) stream splitting of the solute front at every pore, causing independence of particle velocities purely by convection; (2) velocity gradient in pore throats in the direction transverse to flow; and (3) diffusion. Taylor's dispersion in a capillary tube accounts only for the second and third of these phenomena, yielding a quadratic dependence of dispersion on Péclet number. Plug flow in the bonds of a physically representative network accounts only for the first and third phenomena, resulting in a linear dependence of dispersion on Péclet number. When all the three phenomena are accounted for, we can explain effectively the weak nonlinear dependence of dispersion on Péclet number.
Article
Full-text available
Stable and unstable displacement experiments were performed in millstone and limestone cores. Concentration histories at ten locations along the core samples were obtained by acoustic measurements. Particle-tracking simulations of the displacements were also made utilizing permeability distributions measured with a permeameter. The combination of experimental observations and simulations indicate that superstable ([ital M][lt]1) displacements suppress the influence of heterogeneity; this suppression was reflected in smaller apparent dispersivities as the mobility ratio decreased below unity. In the millstone, which exhibited random heterogeneity, two-dimensional particle-tracking simulations reproduce with reasonable accuracy the growth of the fingered region in unstable displacements. In homogeneous porous media, concentration histories obtained in three-dimensional simulations did not differ significantly from their two-dimensional counterparts. In the more heterogeneous limestone, unstable displacements accentuated the influence of heterogeneity leading to longer transition zones. Two distinct flow regimes were observed in unstable displacements: (1) an initial period of rapid transition zone growth and (2) a subsequent period in which leading and trailing edges of the transition zone travel at nearly constant velocities.
Article
Full-text available
Pulsed gradient spin echo nuclear magnetic resonance (PGSE NMR) has been used to examine the approach to asymptotic dispersion for flow of water in a randomly packed array of 500 mum latex spheres in a cylindrical tube of inner diameter 10 mm. Measurements have been performed using single pulse pair and double pulse pair PGSE NMR in directions both transverse to and co-linear with the flow direction, Péclet numbers (Pe) covering a range from 2.8×103 to 8.4×103. In the case of longitudinal flow we find evidence for both reversible and irreversible dispersion over observation times of up to 10tauv where tauv is the correlation time associated with flow across one bead diameter. The fluctuating component of the velocity field appears to obey a simple Ornstein-Uhlenbeck correlation function with a single correlation time in close agreement with tauv. In transverse dispersion, no stationary velocity distribution is evident and the correlation time for fluctuation is much smaller, over the entire range of Pe values.
Article
Full-text available
Gas phase miscible displacement experiments were conducted to quantitatively investigate the advective and dispersive contributions to gas phase transport in unsaturated porous media over a range of soil water contents. Furthermore, the independence of measured dispersivity values was evaluated through comparison of nonreactive and reactive tracer transport. Methane was used as a nonreactive tracer, while difluoromethane (DFM) and trichloroethene (TCE) were used as reactive tracers. At soil water contents below 17%, measured dispersivity values are observed to be independent of the tracer compound and of the soil water content. Conversely, the dispersivities are tracer dependent at the highest soil water contents, wherein the values for DFM and TCE are 3 and 6 times larger than that of methane's, respectively. The significantly larger dispersivity values obtained for DFM and TCE are most likely due to rate-limited mass transfer of these compounds between the gas phase and soil water, which is not observed for methane because of its minimal water partitioning. The dispersivity values obtained here range between 0.3 and 3 cm and are similar to those reported in the literature. The results indicate that within a given ``ideal transport'' range, dispersivities measured at one soil water content with a given tracer may be representative of the same porous media system at other soil water contents and for other compounds.
Article
Full-text available
Macroscopic dispersion is the mixing, on the scale of several hundreds of grain diameters, at a point in a permeable medium that is free of boundary effects. Megascopic dispersion is the one-dimensional (1D) dispersion derived by averaging across an entire cross section. This work investigates how both dispersions vary with heterogeneity, aspect ratio, diffusion coefficient, and autocorrelation. The theoretical results are compared to existing field and laboratory data and to existing theories for limiting cases. The degree of autocorrelation in the medium determines whether or not megascopic dispersivity (dispersion coefficient divided by velocity) is uniquely defined. Large correlation distances (with respect to the medium dimensions) imply a dispersivity that grows with distance traveled. Small correlation distances imply a dispersivity that is eventually stabilized at some constant value. This value is related to the heterogeneity of the medium. On the field scale, diffusion is insignificant, but on a laboratory scale, it can stabilize the dispersivity even if the medium is correlated. Macroscopic dispersivity is sensitive to diffusion in both the laboratory and field scale. It is smaller than or equal to megascopic dispersivity, also in conformance with experimental data, and comparable to laboratory-measured dispersivity.
Article
Taking advantage of the analogy between hydraulic and electrical flows to facilitate the prediction of porous media characteristics is a longstanding practice in engineering and science applications. The relationships between hydraulic and electrical properties are widely used in the characterization of transport phenomena relying on the strong correlation between electric and hydraulic flow conductance. However, due to the lack of small-scale investigations the similarity among their pathways/tortuosities is still unclear. Here a series of direct finite element numerical simulations are conducted within pore-level microstructures to extract and compare the streamlines of both electric and fluid flow currents and examine the accuracy of the hydraulic-electric analogy by predicting the petrophysical characteristics of the case studies. The fluid flow and electric transports are simulated through low-porosity digital unconsolidated packings of polydispersed grains representing the Athabasca oil sands deposit as the second largest oil reserve in the world. The formation factor, porosity, and absolute permeability of the media under consideration are predicted, and consequently, the streamlines of both electric and hydraulic currents are extracted and compared in terms of length, shape, and pathways. According to the results, the fluid flow streamlines are longer and pose differently than the homogeneous electric current streamlines. The average ratio between the hydraulic and electric tortuosities is 1.15 and a local extremum occurs at the porosity of eighteen percent. Pedotransfer functions for tortuosities, dimensionless permeability, and formation factor are proposed underpinning the rigorous relationships between transport processes in porous media.
Article
A stable spreading film between water and gas can extract a significant amount of bypassed non-aqueous phase liquid (NAPL) through immiscible three-phase gas/water injection cycles. In this study, the pore-scale displacement mechanisms by which NAPL is mobilized are incorporated into a three-dimensional pore morphology-based model under water-wet and capillary equilibrium conditions. The approach is pixel-based and the sequence of invasions is determined by the fluids’ connectivity and the threshold capillary pressure of the advancing interfaces. In addition to the determination of three-phase spatial saturation profiles, residuals, and capillary pressure curves, dynamic finite element simulations are utilized to predict the effective permeabilities of the rock microtomographic images as reasonable representations of the geological formations under study. All the influential features during immiscible fluid flow in pore-level domains including wetting and spreading films, saturation hysteresis, capillary trapping, connectivity, and interface development strategies are taken into account. The capabilities of the model are demonstrated by the successful prediction of saturation functions for Berea sandstone and the accurate reconstruction of three-phase fluid occupancies through a micromodel.
Article
Permeability, k, is a property of a porous medium that measures the capacity and ability of the medium to transmit fluids. The permeability of a porous medium (porous rock, soil, packing) can be predicted using empirical relationships, capillary models, statistical models, and hydraulic radius theories. It generally has been recognized that grain size is the fundamental independent variable that controls permeability in unconsolidated sediments [Shepherd, R. G. 1989]. Sub-pore scale modeling is used in this study to evaluate the relation between grain size distribution and permeability of the porous media. In this method the governing equations are applied directly on the porous medium images. The meshing algorithm is applied directly on the pore and grain sections of the images, and the studied mechanism is simulated at the level of pores and grains. Virtual porous media are generated using regular packings of uniforms spheres. The permeability of each packing is calculated computationally and then compared to the predictions of the Kozeny-Carman equation. An in-house developed pattern generator is used to generate spherical grains packing with pre-specified particle size distributions. A sub-pore scale modeling approach is used to calculate permeabilities of generated virtual porous media. The equivalent diameters that provide match from the Kozeny-Carman equation are then predicted.
Article
Enhanced Gas Recovery (EGR) involves CO2 injection into natural gas reservoirs to both increase gas recovery and trap CO2. EGR viability can be determined by reservoir simulations; however, these require a description of fluid dispersion (mixing) between the supercritical CO2 and natural gas. Here, this dispersivity (α) in sandstone rock plugs as a function of residual water fraction is quantified. To ensure the accuracy of such data, a novel core flooding experimental protocol that ensured an even spatial distribution of water was designed, minimized erroneous entry/exit contributions to mixing, and minimized dissolution of the CO2 into the water phase. Dispersivity was found to increase significantly with water content, although the differences in α between sandstones were eliminated upon the inclusion of residual water. This enabled development of a correlation between α and water content and, hence, between the dispersion coefficient and Peclet number that is readily incorporable into reservoir simulations.
Article
Determination of interfacial areas is crucial for accurate identification of multiphase subsurface flow processes, e.g. in situ water treatment, contaminant transport, phase change, convection/ diffusion, and colloid adsorption/ desorption/ migration. Pore-scale imaging and simulation techniques provide an appealing opportunity to explicitly predict such geometrical characteristics of saturated and unsaturated porous media. Here, synthetic unconsolidated sand packs in a wide range of grain arrangements and size distributions are reconstructed to extensively capture all possible heterogeneous pore-level geometries in the McMurray formation, the primary bitumen formation in the Athabasca oil sands deposit. Multi-phase fluid occupancies throughout the partially saturated media during drainage and imbibition are predicted applying a direct pixel-wised pore morphological model, incorporating wetting phase layer covering rock/solid surfaces. The post-processing results are verified using two-phase experimental data points and images, demonstrating a remarkable variation of the interfacial area as a function of saturation profile, rock configuration, and displacement scenarios. Empirical models are proposed to predict the bulk and meniscus areas using average particle diameter and porosity as input parameters.
Article
Assessment of the viability of enhanced gas recovery (EGR), in which CO2 is injected into natural gas reservoirs, requires accurate and appropriate reservoir simulations. These necessitate provision of parameters describing dispersion between the fluids. Here we systematically measure fluid dispersion in various rock cores (sandstones and carbonates), both dry and at irreducible water saturation, at reservoir conditions. In this manner we evaluate the impact of the irreducible water on the miscible displacement processes. As such this represents the first measurement of dispersion as a function of water saturation for supercritical gases in consolidated media. Complementary measurements of water spatial distribution along the rock axis, as well as the pore size distribution occupied by the water were performed using magnetic resonance techniques. Irreducible water was found to increase dispersivity by a factor of up to 7.3. The dispersion coefficient (K) was measured as a function of velocity and the data for both dry and water-containing samples were successfully combined on a K-Péclet number (Pe) plot, enabling ready future inclusion into EGR reservoir models. The power-law dependence of K upon Pe produced an exponent of 1.2 for dry and water-saturated sandstones and 1.4 for dry and water-saturated carbonates, consistent with literature results (Bijeljic et al., 2011; Honari et al., 2015).
Article
Although the previous static experiments provided critical information as to the existence of a critical temperature range that yields the maximum heavy-oil recovery during steam/solvent injection, dynamic experiment are needed to account for the relationship between the solvent introduced into the system and heavy-oil recovery. We conducted a series of dynamic experiments in which liquid (heptane) solvent was injected into a heavy oil saturated rock matrix, surrounded by a fracture with and without pre-thermal injection. Water-wet rock matrix (sandstones) was saturated with heavy oil and placed inside a core holder. Next, the system was placed into an oven and maintained at constant temperature conditions. Then, either hot solvent (superheated to be in vapor phase) or cold solvent was introduced into the system through the fracture at a constant rate. Pressure and temperature was continuously monitored along the core and the properties of oil and liquid condensate from gas produced were measured and analyzed. This scheme was repeated for a wide range of temperature conditions. The first requirement for a successful application is that the solvent should diffuse into matrix effectively before it breaks through and improves gravity drainage of oil by dilution. The second requirement is solvent retrieval. The retrieval of the solvent during solvent injection phase and post-thermal method (steam or hot-water) injection performed at the near-boiling point temperature of the solvent was monitored. Our results and observations indicate that there exists a critical temperature and injection rate that yields a maximized oil recovery and solvent retrieval.
Article
The method used in an earlier paper (Aris 1956) to discuss the dispersion of a solute in a fluid flowing through a tube is applied here to the case in which the solute can also pass into another fluid phase flowing in an annular region around the first. The apparent diffusion coefficient is the sum of the molecular and Taylor diffusion coefficients in the two phases and a term due to the finite rate of partition between them. It is shown how the Taylor diffusion coefficients depend on the ratio of amounts of solute held in the two phases and how this gives a connexion between the coefficient a$^{2}$U$^{2}$/48D found by Taylor (1953) for viscous flow in a circular tube and the 11 a$^{2}$U$^{2}$/48D found by Westhaver (1942) in his analysis of the distillation column. The use of these apparent diffusion coefficients in distillation and partition chromatography is illustrated.
Conference Paper
We simulate flow and transport directly on pore-space images obtained by micro-CT scanning of rock cores. An efficient Stokes solver is used to simulate low-Reynolds number flows. The flow simulator uses a finite-difference method along with a standard predictor-corrector procedure to decouple pressure and velocity. An algebraic multigrid technique solves the linear systems of equations. We then predict permeability and the results are compared with lattice Boltzmann numerical results and available experimental data. For solute transport we apply a streamline-based algorithm that is similar to the Pollock algorithm common in field-scale reservoir simulation, but which employs a novel semi-analytic formulation near solid boundaries to capture, with sub-grid resolution, the variation in velocity near the grains. A random walk method accounts for molecular diffusion. The streamlinebased algorithm is validated by comparison with published results for Taylor-Aris dispersion in a single capillary with a square cross-section. We then accurately predict available experimental data in the literature for longitudinal dispersion coefficient as a function of Peclet number. We introduce a characteristic length based on ratio of volume to pore/grain surface area that can be used for consolidated porous media to calculate Peclet number.
Article
Porous media in which different fluid phases coexist are common in nature (e.g., vadose zone and gas-oil reservoirs). In partially saturated porous media, the intricate spatial distributions of the wetting and nonwetting phases causes their flow to be focused onto preferential paths. Using a novel 2-D experimental setup allowing pore-scale measurement of concentration fields in a controlled unsaturated flow, we highlight mechanisms by which mixing of an invading fluid with the resident fluid is significantly enhanced when decreasing saturation. The mean scalar dissipation rate is observed to decrease slowly in time, while under saturated conditions it decays rapidly. This slow decrease is due to sustained longitudinal solute fingering, which causes concentration gradients to remain predominantly transverse to the average flow. Consequently, the effective reactivity is found to be much larger than under saturated conditions. These results provide new insights into the role that multiphase flows play on mixing/reaction in porous media.
Article
When oil is displaced from a horizontal formation by another fluid of lower density, the latter tends to override the former in the shape of a tongue owing to gravity segregation. This gravity tongue has an adverse influence on the oil recovery. If the fluids are miscible, diffusion (mixing) takes place at the interfacial boundary of the gravity tongue. This mixing should have a favorable effect on oil recovery. The report describes a laboratory study of the magnitude of the mixing zone under various conditions, so as to assess the effect of diffusion on oil recovery both in laboratory experiments and under actual field conditions. The technique used enables visual observation and measurement of the size of the mixing zone in transparent glass-powder packs. The results show that in experiments in small models and cores the width of the mixing zone may well be of the same order of magnitude as the height of the model. In such cases oil recovery is favorably affected by mixing. It can further be concluded that, under conditions prevailing in the field, mixing of the injected fluid with reservoir oil is equal to that caused by molecular diffusion alone, eddy-mixing not taking place to any appreciable extent. A simple calculation, then, shows that molecular diffusion is too small for a beneficial effect to be expected from the injection of miscible fluids in horizontal or nearly horizontal reservoirs unless pay zones are thin. Introduction This paper gives results of experiments made on the mixing which occurs when a miscible displacement is carried out in a horizontal reservoir. This mixing takes place at the interfacial boundary of the gravity tongue formed when the lighter injected fluid overrides the oil present in the reservoir. The object of the experiments was to simulate the field case where, for example, propane with a viscosity of 0.075 cp under reservoir conditions displaces an oil of viscosity 0.6 cp. In the experiments the lighter fluid had a viscosity of about 1 cp and the heavier one a viscosity of about 8 cp, so as to obtain the same viscosity ratio. In order to enable the results to be compared with those published in the literature, a set of experiments with viscosity ratio equal to one was also performed. EXPERIMENTAL TECHNIQUE The technique developed for the purpose enables the width of the mixing zone to be studied as a function of time and place. A glass-powder pack is saturated with a water-glycerine mixture of suitable viscosity, which represents the reservoir fluid. The pack is then rendered transparent by dissolving sufficient ammonium thiocyanate in the mixture to obtain a solution with a refractive index matching that of the glass powder. Alkaline water to which phenolphthalein has been added is used as the lighter and less-viscous displacing fluid. In those places where mixing or diffusion of the two liquids occurs, the slightly acidic ammonium-thiocyanate solution neutralizes the alkali in the water, and the glass pack shows up white against the red-colored invading water. If a black cloth is hung over the back of the apparatus, the transparent part of the glass pack appears black. In this way the position of the two phases and the transition zone between them is clearly visible, as shown in Fig. 1 (where the red-colored invading water is the grey-shaded zone lying uppermost).In all experiments the amount of alkali in the water was chosen such that the upper contour corresponded to a concentration of 5 per cent of the dense liquid. The lower contour is determined by the size of the glass grains and the thickness of the pack and, consequently, varies from experiment to experiment. SPEJ P. 317
Article
Methods for miscible flooding have been researched and field tested since the early 1950's. This paper reviews the technical state of the art and field behavior to date for the major miscible flood processes: first-contact miscible, rich-gas drive, vaporizing-gas drive, and carbon dioxide flooding. Important technological areas selected for review include phase behavior and miscibility, sweepout, unit displacement efficiency, and process design variations. Carbon dioxide flood technology is emphasized, and several technical issues are identified that still need to be resolved. Rules of thumb and ranges of conditions are discussed for applicability of each process. A comparison is made of the incremental recovery and solvent slug effectiveness observed in field trials of the different processes. From the limited data available, processes. From the limited data available, there is no clear-cut evidence that field results on average and for a given slug size have been appreciably better or poorer for one process compared with another. Introduction The search for an effective and economical solvent along with development and field testing of miscible-flood processes has continued since the early processes has continued since the early 1950's. Early focus was on hydrocarbon solvents, and three types of hydrocarbon-miscible processes were developed: the first-contact miscible process; the vaporizing-gas drive process, often called high-pressure gas drive; and the rich-gas drive process, often called condensing-gas drive. First-contact miscible solvents mix directly with reservoir oils in all proportions and their mixtures always remain proportions and their mixtures always remain single phase. Other solvents are not directly miscible with reservoir oils, but under appropriate conditions of pressure and solvent composition these solvents can achieve miscibility in-situ by mass transfer of oil and solvent components through repeated contact with the reservoir oil. miscibility achieved in this manner is termed multiple-contact or dynamic miscibility. The vaporizing-gas drive process achieves dynamic miscibility by in-situ vaporization of intermediate molecular weight hydrocarbons from the reservoir oil into the injected gas. Dynamic miscibility is achieved in the rich-gas drive process by in-situ transfer of intermediate molecular weight hydrocarbons from the injected gas into the reservoir oil. Propane or LPG mixtures typically were the solvents used in first-contact hydrocarbon miscible flooding, whereas natural gas at high pressure and natural gas with appreciable concentrations of intermediate molecular weight hydrocarbons were injection fluids in vaporizing-gas drive and rich-gas drive floods. The high cost of propane, LPG, or rich hydrocarbon gas propane, LPG, or rich hydrocarbon gas dictated that these solvents be injected as slugs which usually were driven with natural gas. Flue gas and nitrogen also have been found to achieve dynamic miscibility at high pressures with some oils by the vaporizing-gas drive mechanism. Hydrocarbon miscible processes have received extensive field testing since the 1950's, primarily in the United States and Canada. Over 100 projects were initiated during this time period. The majority were small-scale pilot tests involving one or at most a few injection wells; however, a number of large projects were undertaken involving several thousand acres or more (more than 4 000 000 m2). A few projects tested flue gas injection.
Article
Fluid mixing plays a fundamental role in many natural and engineered processes, including groundwater flows in porous media, enhanced oil recovery, and microfluidic lab-on-a-chip systems. Recent developments have explored the effect of viscosity contrast on mixing, suggesting that the unstable displacement of fluids with different viscosities, or viscous fingering, provides a powerful mechanism to increase fluid–fluid interfacial area and enhance mixing. However, existing studies have not incorporated the effect of medium heterogeneity on the mixing rate. Here, we characterize the evolution of mixing between two fluids of different viscosity in heterogeneous porous media. We focus on a practical scenario of divergent–convergent flow in a quarter five-spot geometry prototypical of well-driven groundwater flows. We study by means of numerical simulations the impact of permeability heterogeneity and viscosity contrast on the breakthrough curves and mixing efficiency, and we rationalize the nontrivial mixing behavior that emerges from the competition between the creation of fluid-fluid interfacial area and channeling. This article is protected by copyright. All rights reserved.
Article
Heavy oil and bitumen are important hydrocarbon resources that are destined to play an increasingly significant role in the oil supply of the world. The total heavy oil resources are estimated to be about 10 trillion barrels, nearly three times the conventional oil in place in the world. The efficient and economic recovery of heavy oil and bitumen reserves is a crucial technical challenge. Most of these reserves lie deep inside the earth's crust and are not easily recoverable owing to their very high viscosity. Hydrocarbon solvent-based processes are thought to be effective EOR technologies for heavy oil and bitumen production. An important feature of the miscible fluid displacement is the mixing of the solvent and oil. This mixing occurs on the microscopic scale and result from the diffusion and dispersion process. Mixing effect has positive and negative effects on the miscible process. It could exerts a considerable damping effect on the growth of viscous and gravity fingers as an advantage of the dispersion and diffusion and on the other hand mixing of the solvent with oil in a reservoir decreases the effective strength of the solvent, which can have annoying effect on the miscibility and recovery efficiency. So, correct treatment of this mixing effect within the simulation of the miscible recovery process is very important. Mixing of the solvent and oil is governed by the diffusion and dispersion coefficients. For measuring these coefficients, we must depend largely on experimental measurements of them, because no universal theory permits their accurate a priori calculation. There are many influencing parameters on the mixing process during solvent injection into porous media. Designing an experiment to investigate all the effective parameters is extremely difficult (if not impossible) and time consuming. So, based on the sub-pore scale modeling method, a special program has been developed to investigate the mass transfer process in any porous medium, considering all medium properties. A realistic image of the porous medium without any simplification is directly used in the program. Developed program used the porous medium image directly, so, heterogeneity effect could be investigated. Pore and throat size distributions, as an important heterogeneous parameter, has a unique effect on the diffusion and dispersion coefficients. A comprehensive study of the pore and throat size distribution effect on the mixing coefficients will be presented in this paper.
Article
Thermal and miscible methods are commonly used for in-situ recovery of heavy oil and bitumen. Both techniques have their own limitations and associated shortcomings, often times yielding an inefficient process. The most common thermal method is steam injection, which is highly energy intensive. Steam generation costs and water production affect the economics of the thermal technique adversely. On the other hand, miscible methods are energy effective but their economics depends on the solvent retrieval. Various combinations of these two techniques such as co-or alternate injection of steam and solvent have been proposed as a solution, but no optimum method has yet been developed. Thermal and miscible methods can be combined by co-injecting solvent with steam or injecting solvent into a pre-heated reservoir. Current work was undertaken to study the performance of solvents at higher temperatures for heavy oil/bitumen recovery. Glass bead packs and Berea sandstone cores were used in the experiments to represent different types of pore structures, porosity and permeability. After saturating with heavy oil, the samples were exposed to the vapor of paraffinic solvents (propane and butane) at a temperature above the boiling point of the solvent, and a constant pressure of 1500 kPa. A mechanical convection oven was used to maintain constant temperature across the setup. The setup was designed in such a way that a reasonably long sample (up to 30 cm) can be tested to analyze the gravity effect. The oil recovered from each of these experiments was collected using a specifically designed collection system and analyzed for composition, viscosity and asphaltene content. The amount of oil recovered in each case was also analyzed and the quantity and nature of asphaltene precipitated with each of the tested solvents under the prevailing temperature and pressure of the experiment was reported. Optimal conditions for each solvent type were identified for the highest ultimate recovery. It was observed that recovery decreased with increasing temperature and pressure of the system. It was also noticed that butane diluted the oil more than propane which resulted in lower asphaltene content and viscosity of oil produced with butane as a solvent.
Article
This pore-scale modeling study in saturated porous media shows that compound-specific effects are important not only at steady-state and for the lateral displacement of solutes with different diffusivities but also for transient transport and solute breakthrough. We performed flow and transport simulations in two-dimensional pore-scale domains with different arrangement of the solid grains leading to distinct characteristics of flow variability and connectivity, representing mildly and highly heterogeneous porous media, respectively. The results obtained for a range of average velocities representative of groundwater flow (0.1-10 m/day), show significant effects of aqueous diffusion on solute breakthrough curves. However, the magnitude of such effects can be masked by the flux-averaging approach used to measure solute breakthrough and can hinder the correct interpretation of the true dilution of different solutes. We propose, as a metric of mixing, a transient flux-related dilution index that allows quantifying the evolution of solute dilution at a given position along the main flow direction. For the different solute transport scenarios we obtained dilution breakthrough curves that complement and add important information to traditional solute breakthrough curves. Such dilution breakthrough curves allow capturing the compound-specific mixing of the different solutes and provide useful insights on the interplay between advective and diffusive processes, mass transfer limitations, and incomplete mixing in the heterogeneous pore-scale domains. The quantification of dilution for conservative solutes is in good agreement with the outcomes of mixing-controlled reactive transport simulations, in which the mass and concentration breakthrough curves of the product of an instantaneous transformation of two initially segregated reactants were used as measures of reactive mixing.
Article
The enhanced recovery of natural gas by the injection and sequestration of CO2 is an attractive scenario for certain prospective field developments if the risks of gas contamination or early CO2 breakthrough can be assessed reliably. Simulations of enhanced gas recovery (EGR) scenarios require accurate dispersion parameters at reservoir conditions to quantify the size of the miscible CO2–CH4 displacement front; several experimental studies using core-flooding equipment aimed at measuring such parameters have been reported over the last decade. However, such measurements are particularly challenging and the data produced are generally afflicted in their repeatability by limited experimental control and in their accuracy by systematic errors such as gravitational and core-entrance/exit effects. We review here the existing experimental data pertaining to EGR by CO2 sequestration and also report new measurements of longitudinal CO2–CH4 dispersion coefficients at temperatures of 40–80 ◦C, pressures of 8–12 MPa and interstitial velocities of 0.05–1.13 mm s−1 [14.2–320 ft day−1] in 5–10 cm long sandstone cores with permeabilities of 12 and 460 mD. The core-floods were conducted in both a horizontal and vertical orientation, with significant gravitational effects observed for low velocity floods in horizontal cores with high permeabilities. We also analyzed the effects of tubing and core entrance/exit effects on the measurements and found that the latter resulted in apparent dispersion coefficients up to 63% larger than would be due to the core alone. Our results indicate that dispersivities for CO2–CH4 at these supercritical conditions are less than 0.001 m, which indicates that excessive mixing will not occur in EGR scenarios in the absence of conformance effects such as heterogeneity coupled with injection well pattern. Inclusion of such conformance effects is essential for detailed reservoir simulation.
Article
Dispersivity affects the displacement and sweep efficiencies and the required slug size of a displacing fluid. Unfortunately, the dispersivity values estimated from field tests are a few orders of magnitude greater than the values obtained in laboratory tests. This investigation studies the effect of small scale (or core scale) heterogeneities on the effective dispersivity value in a typical gridblock size used in a reservoir simulator. The study is restricted to contact miscible displacements with unit mobility ratio. A finite element simulator is used to investigate these effects. Physical dispersion is explicitly included in the simulator. Results show that the effective dispersivity is affected by the degree of heterogeneity, the average length of heterogeneity, the length of the system, and the manner in which the permeability values are spatially distributed. The effect of heterogeneity becomes significant if the coefficient of variance is greater than 0.4. The effect of dimensionless scale length (ratio of average length of heterogeneity to length of the system) on effective dispersivity is insignificant for dimensionless scale values less than .01; above that value dispersivity increases with an increase in the scale length. The effective dispersivity increases almost linearly with an increase in the length of the system for constant dimensionless scale length. This provides an explanation for the similar trend reported for field dispersivity data. A simple correlation is proposed to calculate the effective dispersivity of the porous medium. The effect of the way in which given permeability values are distributed across the medium on the dispersivity cannot be correlated with the parameters investigated, probably indicating that the effective dispersivity is a unique function of the manner in which permeability values are distributed. This effect was not noted previously in the literature; however, it does not affect the general trends. Introduction Dispersion is important in understanding reservoir performance during enhanced oil recovery processes. Large values of dispersivity reduce the displacement efficiency of multiple contact miscible displacements and increase the sweep efficiency. The net result of dispersion may be to increase the required slug size of a displacing fluid. Unfortunately, the values obtained in the laboratory do not coincide with the values obtained in the field. The reasons for the discrepancy between the field and the lab values may include the effects of heterogeneities in reservoir properties and reservoir stratification. This discrepancy should not arise if all the possible heterogeneities in the reservoir, including the small scale heterogeneities, can be described rigorously. However, this type of description may not be practical. Generally, a reservoir simulator gridblock is the smallest unit by which heterogeneities in the reservoir may be represented. This investigation is restricted to study the effects of permeability variations in porous medium on the effective values of dispersivity. In cases studied, the permeability variations are the heterogeneities with length of variation about the same as that of the core length, (for example, several inches to several feet), sometimes called small scale heterogeneities. The effects of these small scale heterogeneities on the effective dispersivity value in a typical block size in a reservoir simulator are investigated. In other words, we are addressing the question of dispersivity scaling from laboratory to a typical block size in a field scale simulator. The studies are restricted to unit mobility ratio and contact miscible displacements. P. 215
Article
The objective of this paper is to determine the effect of a wetting immobile phase (e.g. connate water) on the diffusion and macroscopic dispersion in homogeneous unconsolidated packs. We measured the effective dispersion coefficients of gases as a function of water saturation, pressure, and velocity in a laboratory set-up. The dispersion coefficients were 2 to 3 times higher than in the dry porous medium. This effect is also studied using a pore network model where the immobile phase blocks a fraction of the pores for fluid flow. A random walk through the network and counting cluster sizes both gave approximately the same results, which were in good agreement with the experimental data.
Article
Tracer displacements were run during steadystate, two-phase flow, where tracers were added to both the wetting and nonwetting phases. Such experiments were conducted in the primary drainage, imbibition, and secondary drainage cycles over a wide range of fractional flows. The effluent tracer profiles from these experiments were fit with a four-parameter, capacitance-dispersion model, which divides each phase into three fractions: flowing, dendritic (branching), and isolated. Mixing in the flowing fraction is described by a dispersion coefficient, and communication between the flowing and dendritic fraction is represented by a single mass transfer coefficient. This model is shown to adequately fit tracer effluent data over a wide range of conditions. In some areas the model is nonunique--thus, more than one solution fits the experimental data equally well. Additional experiments were conducted to differentiate mechanisms in these regions. The effects of core wettability were investigated and shown to be important. The results of all of these experiments were used to formulate a mechanistic picture of steady, two-phase flow in strongly wetted porous media.
Article
Experimental measurements of the dispersion of tracer particles in flow through natural porous media are compared with a percolation model. The experiments show that tracer dispersion is a sensitive function of the width of the pore-size distribution as measured by mercury capillary pressure. Measurements of capillary pressure (or electrical conductivity) are used to estimate the geometric correlation length of the dominant flow path in the rock. Percolation theory is used to derive a power-law relationship between the correlation length and the ratio of the dispersivity to the average grain size. The experimental value of the power-law exponent is in agreement with the theoretical prediction. Measurements on samples containing a residual saturation of wetting epoxy show no significant change in dispersion behavior. This result mediates against dispersion models requiring trapping in dead-end pores. Tracer concentration profiles exhibit anomalous long-time tails in two cases. In carbonate rocks, we associate long-time tails with macroscopic permeability heterogeneities. In sandstones, long-time tails occur in samples with a very narrow pore-size distribution. These samples may have permeability heterogeneities as a result of defects in the packing density. In the limit of low flow velocity, the long-time tail disappears, suggesting a convective mechanism associated with flow heterogeneities at a millimeter-or-larger scale.
Article
The effect of anisotropic dispersion on nonlinear viscous fingering in miscible displacements is examined. The formulation admits dispersion coefficient-velocity field couplings (i.e., mechanical dispersivities) appropriate to both porous media and Hele--Shaw cells. A Hartley transform-based scheme is used to numerically simulate unstable miscible displacement. Several nonlinear finger interactions were observed. Shielding, spreading, tip splitting, and pairing of viscous fingers were observed here, as well as in isotropic simulations. Multiple coalescence and fading were observed in simulations with weak lateral dispersion, but not for isotropic dispersion. Transversely and longitudinally averaged one-dimensional concentration histories demonstrate the rate at which the mixing zone broadens and the increase in lateral scale as the fingers evolve when no tip splitting occurs. These properties are insensitive to both the dispersion anisotropy and the Peclet number at high Peclet number and long times. This suggests the dominance of finger interaction mechanisms that are essentially independent of details of the concentration fields and governed fundamentally by pressure fields.
Article
The soil gas diffusion coefficient (D(p)) and its variations with soil air content (epsilon) and soil water matric potential (psi) control vadose zone transport and emissions of volatile organic chemicals and greenhouse gases. This study revisits the 1904 Buckingham power-law model where D(p) is proportional to E with X characterizing the tortuosity and connectivity of air-filled pore space. One hundred years later, most models linking D(p) (epsilon) to Soil water retention and pore size distribution still assume that the pore connectivity factor, X, is a constant for a given soil. We show that X varies strongly with both epsilon and matric potential [given as pF = log(psi, cm H(2)O)] for individual soils ranging from undisturbed sand to aggregated volcanic ash soils (Andisols). For Andisols with bimodal pore size distribution, the X-pF function appears symmetrical. The minimum X value is typically around 2 and was observed close to psi of -1000 cm H(2)O (pF 3) when inter-aggregate voids are drained. To link D(p) with bimodal pore size distribution, we coupled a two-region van Genuchten soil water retention model with the Buckingham D(p) (epsilon) model, assuming X to vary symmetrically around a given pF. The coupled model well described D(p) as a function of both epsilon and psi for both repacked and undisturbed Andisols and for other soil types. By merely using average values of the three constants in the proposed symmetrical X-pF expression, predictions of D(p) were better than with traditional models.
Article
The transport of gaseous compounds in soil takes place by gas diffusion, advection, and dispersion. Gas transport processes are influenced by tic soil-gas diffusion coefficient (D), air permeability (k(a)) and soil-gas dispersion coefficient (D(H)), respectively. Of three gas transport parameters, DH is the least understood especially hov it is correlated to soil type, moisture conditions, and other transport parameters (i.e., D(P) and k(a)). In this study, a unified measurement system (UMS) that enables sequential measurement of D(P), k(a), and D(H) on the same soil core was developed. The experimental sequence is based on a two-chamber measurement of DH and k(a), followed by a one-chamber measurement of L P. Gaseous oxygen concentration and air pressure sensors are located in inlet and outlet chambers as well as at multiple points along the soil column. Using different particle-size fractions of non-aggregated (Toyoura sand) and aggregated (Nishi-Tokyo loam) soils, the effects of soil structure, particle (aggregate) size, and column scale (5-cm i.d. and 30-cm or 60-cm length) on the three gas transport parameters were investigated. For both soils, D(H) linearly increased with increasing pore-air velocity. For Toyoura sand, gas dispersivity (lambda = D(H)/u(0)) decreased with increasing soil-air content, while for Nishi-Tokyo loam, gas dispersivity decreased with increasing soil-air content to a minimum value when inter-aggregate pores were drained and increased again whether pores inside the soil aggregates started to act as tortuous air-filled pathways. In the arterial bore region (corresponding to the total pore volume for Narita sand and the inter-aggregate pore volume for Nishi-Tokyo loam), a linear relation between tortuosity of the air-filled pore network (T, calculated from D(P)) and the gas dispersivity (lambda) was observed.
Article
Investigations of gas transport and fate processes in packed soil systems require knowledge of the gas diffusion coefficient, D p , as a function of air-filled porosity, e. On the basis of the literature, data from six studies over the porosity range of 0.1 to nearly 1.0, it is reconfirmed that the Marshall (1959) model better predicts D p (e) in completely dry, repacked porous media than do the Penman (1940) and Millington (1959) models. The smaller D p value in wet soil, as compared with dry soil at the same air-filled porosity, is accounted for by introducing a water-induced linear reduction (WLR) term, equal to the ratio of air-filled porosity to total porosity, in the D p (e) model. By adding the WLR term in each of the three D p (e) models for dry porous media, the so-called WLR(Marshall), WLR(Penman), and WLR(Millington) D p (e) models for wet soil are developed. To test the three WLR models, D p was measured at different soil-water contents in six differently textured (6-38% clay) repacked soils. The WLR (Marshall) model accurately and best described D p (e) for all six soils and additional soils from the literature. All three WLR models performed better than previous D p (e) models. This study implies that the smaller D p in a wet soil, which is due to water-induced changes in air-filled pore shape and pore connectivity, can be described by a simple, linear function of relative air-filled porosity. The WLR(Marshall) model represents a conceptual and accurate model to predict D p (e) in sieved, repacked soil.
Article
The paper describes experiments on miscible displacement in various porous media and the results of these experiments. Both glass bead packs and natural cores were used. Bead diameters varied from 0.044 to 0.47 mm, and pack lengths varied from 83 to 678 cm. Natural cores used were Berea and Torpedo sandstone. By taking samples as small as 0.5 cc and using refractive index for analysis, the data on break through curves could be plotted to within ± 0.5 per cent. To plot the data correctly on error function paper, a parameter (Vp - v)/vV was used which allowed for the predicted growth of the front as it moved past the observer. The change in the amount of mixing (length of mixed zone) was studied by varying velocity, length of travel, bead size, viscosity ratio and pack diameter. When the displaced material was less viscous than the displacing material (favorable viscosity ratio), these changes were adequately predicted by theory. When natural cores were used, rather than glass beads, the amount of mixing was greatly increased - also qualitatively predicted by theory. In experiments with favorable viscosity ratios in which the ratio was varied from 0. 175 to 0. 998, it was found that the rate of mixing was changed by a factor of 5. 7. Thus, the rate of mixing is strongly affected by viscosity ratio, even when the theoretical error function relationship for mixing is valid. Experiments using fluids with viscosity ratios near 1.0 showed that the instability effects of even a slightly unfavorable viscosity ratio (1.002) caused disproportionately more elongated breakthrough curves than found with a favorable viscosity ratio (.998). When the viscosity ratio was as high as 5.71 these instability effects were much more pronounced, as evidenced by the shape of the breakthrough curve. The displacements at viscosity ratios above 1.0 no longer followed the theoretical error function curve.
Article
Sir Geoffrey Taylor has recently discussed the dispersion of a solute under the simultaneous action of molecular diffusion and variation of the velocity of the solvent. A new basis for his analysis is presented here which removes the restrictions imposed on some of the parameters at the expense of describing the distribution of solute in terms of its moments in the direction of flow. It is shown that the rate of growth of the variance is proportional to the sum of the molecular diffusion coefficient, D, and the Taylor diffusion coefficient kappa a2U2/D, where U is the mean velocity and a is a dimension characteristic of the cross-section of the tube. An expression for kappa is given in the most general case, and it is shown that a finite distribution of solute tends to become normally distributed.
Article
In this paper, an extensive study is presented on the single-phase flow of xanthan/tracer slugs in a consolidated sandstone. The phenomena studied include polymer/tracer dispersion, excluded/inaccessible-volume effects, polymer adsorption, and viscous fingering. In some floods, there is also evidence of nonequilibrium effects. Macroscopic flow equations are derived that include terms to model all the behaviors listed above. A microscopic approach is also developed that describes certain features of polymer flow in porous media semiquantitatively.
Article
Observations of pore structure in thin-sections are related to the performance of stable, first-contact-miscible displacements in reservoir cores and then to simulations of displacement performance of CO2 corefloods. Results of effluent composition measurements are reported for miscible displacements in seven core samples—three sandstones and four San Andres carbonates from west Texas or eastern New Mexico. Those displacements are interpreted by fitting the measured effluent compositions to the Coats-Smith (C-S) model, which represents the flow as occurring in flowing and stagnant fractions with mass transfer between them. Observations of thin-sections, including measurements of pore-size distributions and a simple measurement of spatial correlation of pore sizes, are also reported. Comparison of displacement results and thin-section data indicates that wide pore-size distributions and preferential flow paths are characterized in the C-S model by high dispersion coefficients and low flowing fractions. Simulations of the interactions of phase behavior and flow in nonuniform pore structures indicate that wide pore-size distributions and preferential flow paths can significantly increase residual oil saturations (ROS’s) in CO2 floods over those for uniform pore structures. Thus, heterogeneities observable at the scale of a thin-section have significant effects in laboratory core but much smaller effects in displacements at field scale. Large-scale heterogeneities present in field floods, however, probably cause similar increases in residual saturation in some fields.
Article
When a soluble substance is introduced into a fluid flowing slowly through a small-bore tube it spreads out under the combined action of molecular diffusion and the variation of velocity over the cross-section. It is shown analytically that the distribution of concentration produced in this way is centred on a point which moves with the mean speed of flow and is symmetrical about it in spite of the asymmetry of the flow. The dispersion along the tube is governed by a virtual coefficient of diffusivity which can be calculated from observed distributions of concentration. Since the analysis relates the longitudinal diffusivity to the coefficient of molecular diffusion, observations of concentration along a tube provide a new method for measuring diffusion coefficients. The coefficient so obtained was found, with potassium permanganate, to agree with that measured in other ways. The results may be useful to physiologists who may wish to know how a soluble salt is dispersed in blood streams.
Article
Inaccurate modeling of reservoir mixing by using large gridblocks in compositional simulation can affect recoveries significantly in miscible gasfloods and lead to inaccurate predictions of recovery performance. Reservoir mixing or dispersion is caused by diffusion of particles across streamlines; mixing can be enhanced significantly if the surface area of contact between the reservoir and injected fluid is increased as fluids propagate through the reservoir. A common way to convert geological models into simulation models is to upscale permeabilities on the basis of reservoir heterogeneity. Upscaling affects the degree of mixing that is modeled, but the importance of reservoir mixing in upscaling is largely ignored. This paper shows how to estimate the level of mixing in a reservoir and how to incorporate mixing into the upscaling procedure. We derive the key scaling groups for first-contact miscible (FCM) flow and show how they have an impact on reservoir mixing. Heterogeneities are assumed to dominate the flow regime so that gravity effects are negligible. We examine only local mixing, not apparent mixing caused by variations in streamline path lengths (convective spreading). Local mixing is important because it affects the strength of the injected fluid and can cause an otherwise multicontact miscible (MCM) flood to become immiscible. More than 1,000 2D numerical simulations are carried out using experimental design to estimate dispersivity as a function of the derived scaling groups. We show that reservoir mixing is enhanced owing to fluid propagation through heterogeneous media. Because mixing is dependent on heterogeneities, upscaling is an iterative process in which the level of mixing in both the longitudinal and transverse directions must be matched from the fine to the coarse scale. The most important groups that affect mixing are the mobility ratio, dispersion number, correlation lengths, and the Dykstra-Parson coefficient. Large dispersion numbers yield greater dispersivities away from the injection well. We show through simulations of both FCM and MCM floods that gridblock size can be increased significantly when reservoir mixing is large. Heterogeneous reservoirs with large longitudinal correlation lengths can be upscaled to larger gridblocks than reservoirs with random permeability fields. This paper shows how to determine a priori the maximum gridblock size allowed in both the x- and z-directions to predict the oil recovery from miscible gasfloods accurately.
Article
The soil gas diffusion coefficient (D(p)) and air permeability (k(a)) and their dependency on soil air content (epsilon) control gas diffusion and advection in soils. This study investigated the effects of average particle size (D(50)) and dry bulk density (rho(b)) on D(p) and k(a) for six sandy soils under variably saturated conditions. Data showed that particle size markedly affects the effective diameter of the drained pores active in leading gas through the sample at -100 cm H(2)O of soil water matric potential (calculated from D(p) and k(a)) as well as the average pore diameter at half saturation (calculated from the water retention curve), both exhibiting similar and exponential relationships with D(50) Under variably saturated conditions, higher D(p) and k(a) in coarser sand (larger D(50)) were observed due to rapid gas diffusion and advection through the less tortuous large-pore networks. In addition, soil compaction (larger rho(b)) simultaneously caused reduced water blockage effects and a reduction of large-pore space, resulting in higher D(p)(epsilon) but lower k(a)(epsilon). Two recent models for D(p)(epsilon) and k(a)(epsilon) were evaluated: the water-induced linear reduction (WLR) model for D(p), and the reference-point power law (RPL) model for k(a), with reference point ka set at -100 cm H(2)O. The performance of both models for the sandy soils (particle size range 0.02-0.9 mm) was improved if the pore connectivity-tortuosity factor and water blockage factors were assumed to be functions of D(50) and rho(b). Water blockage factors, N for the WLR D(p)(epsilon) model and M for the RPL k(a)(epsilon) model, showed a strong nonlinear relationship (R(2) = 0.95) that seems promising for predicting D(p)(epsilon) from the more easily measureable k(a)(epsilon).
Article
Mixing that occurs as on ga displaces another within columns of soil or glass beads was examined by using mathematical model that included mixing by molecular diffusion, mass transport, and gaseous sorption; although apparent diffusion coefficient values were no constant with flow velocity, general behavior of mixing process was adequately described by model.