|Novel CO2 Foam Concepts and Injection Schemes for Improving CO2 Sweep Efficiency in Sandstone and Carbonate Hydrocarbon Formations
||Last Reviewed 6/10/2014
The goal of this project is to develop mobility control agents using surfactants injected with carbon dioxide (CO2) rather than with water for CO2 enhanced oil recovery (CO2-EOR) in heterogeneous carbonate and sandstone reservoirs. Objectives are to (1) reduce CO2 mobility in CO2 swept portions of the reservoir—but not at the light-hydrocarbon enriched CO2 displacement front; (2) reduce mobility more in higher permeability than in lower permeability intervals; (3) improve distribution of injected fluids in natural fracture networks; (4) reduce the amount of surfactant required to achieve successful mobility control; and (5) achieve displacement efficiency well below the MMP by using surfactants with ultra-low water/oil interfacial tension. These attributes will divert CO2 from swept regions to regions with unswept oil.
The University of Texas at Austin, Austin (UT), TX 78712-1500
Rice University, Houston, TX
Carbon dioxide flooding has become a routine technology for enhanced oil recovery worldwide. In the United States, the number of CO2-EOR projects increased from 20 in 1980 to 86 in 2008. Over 600 million tons of CO2(11 trillion standard cubic feet) transported through 3,500 miles of high-pressure CO2 pipelines were injected into over 13,000 CO2 EOR wells. The oil production rate from CO2 projects in 2010 was 273,000 barrels of oil per day.
Carbon dioxide EOR has exhibited strong growth in the past 30 years and has expanded despite fluctuations in oil price. However, there are several issues challenging the oil recovery, economic efficiency, and applicability of the process. The oil recovery efficiency is low and the CO2 utilization rate—the amount of CO2 injected to recover an incremental barrel of oil—is high. This is due to the low viscosity of CO2 compared to oil and water and the resulting unfavorable mobility ratio. This, combined with reservoir heterogeneities, leads to poor sweep efficiency and bypass of un-contacted oil. This problem is evident by cycling of CO2 through swept intervals. Also, many shallow reservoirs are below the minimum miscibility pressure (MMP) for efficient CO2 displacement, and this oil is typically not targeted for CO2 flooding.
The inability to accommodate these needs across reservoirs has previously limited CO2 foam application. The proposed approach differs from previous practice primarily because (a) the process design is based on phase behavior and foam morphology/stability associated with colloid and interfacial science and thermodynamics; (b) the surfactant (in the case of nonionic surfactants) is injected with CO2 rather than in water; (c) CO2 foam with surfactant that generates low interfacial tension between oil and water can have good displacement efficiency at pressures well below the MMP; and (d) these processes can be used in the paleo-residual oil zone as well as in the water flooded residual oil zone. Injecting surfactant with CO2 will place more of the surfactant in the thief zone compared to injecting surfactant with water.
The method to be developed in this project will improve the mobility ratio of CO2 flooding and increase recovery efficiency, which in turn will reduce the cost of purchased and recycled CO2 and broaden the application of CO2flooding to more heterogeneous reservoirs. Combining surfactant flooding and CO2 foam mobility control will make shallow reservoirs with pressures below the MMP candidates for a combined CO2/surfactant EOR process. A revised national resource assessment for CO2-EOR (July 2011) prepared for DOE by Advanced Resources International indicated that "Next Generation" CO2-EOR can provide 137 billion barrels of additional technically recoverable domestic oil, with about half (67 billion barrels) economically recoverable at an oil price of $85 per barrel. The proposed work will contribute to such advanced technology.
Improving the efficiency of CO2-EOR not only increases the supply of domestic oil but, by making the process more broadly applicable, also increases the demand for CO2 in more geographic locations and thus promotes its sequestration.
Nonionic and anionic ethoxylated surfactants were investigated in micromodels and in a foam generator in series with a capillary flow viscometer. Highly viscous CO2/water foams with viscosities orders of magnitude higher than pure CO2 were formed between 24 and 70°C. These experiments demonstrated that foams are formed over a wide range of salinities, with very low surfactant concentrations, for a secondary alcohol ethoxylate. They also show that Dow Tergitol 15-S-20 forms a maximum viscosity foam at ~95 percent quality. Creation of strong foams with small amounts of water is an important milestone. The foam still forms with only 0.01 wt.% of this surfactant in the aqueous phase, which greatly benefits the economics of EOR. To further demonstrate this concept, the UT team showed that carbon dioxide foams break in the presence of injected oil as a function of oil concentration. Thus, this “smart surfactant” satisfied a major goal of forming foam and lowering mobility in the absence of oil, while selectively not forming foams in the presence of residual oil. Furthermore, foamability, CO2solubility, and dodecane/bine emulsion stability were investigated for two nonionic – cationic switchable, ethoxylated (5EO) coco and tallow amines obtained from Akzo Nobel. Both surfactants were tested at South Hobbs field conditions (40°C, 1500–1700 psia) and salinity (5.98 g/L NaCl, 2.03 g/L MgCl2•6H2O, and 1.18 g/L CaCl2•2H2O). At 1wt. percent surfactant concentration in brine, both surfactants formed CO2/water foam that was more than one hundred times more viscous than pure CO2. The tested ethoxylated amine surfactants are CO2 soluble and generate unstable dodecane/brine emulsions.
Micromodels made with polydimethylsiloxane (PDMS) have been constructed to observe foam in heterogeneous pore systems. A paper on how wettability can be adjusted in micromodels was submitted for publication. The Rice team acquired a glass micromodel from Shell to observe CO2 foam in a pressurized visual cell.
Hobbs reservoir, operated by Oxy, has been identified as a new carbonate reservoir candidate in addition to the currently investigated sandstone reservoir (Steamboat). The injection and production historical and well log data from Steamboat reservoir have been analyzed to characterize fluid mobility control issues. A detailed characterization of fracture distributions and orientations with respect to lithology and across the structure has been conducted. The effects of fracture characteristics on the performance of Tensleep reservoir have also been evaluated. In parallel, geologic and reservoir fluid models have been developed to history match the field and foam process forecasting.
A critical literature review on the experiments, modeling, and significance of gas trapping during foam flood has been conducted. A new model of gas trapping has been developed based on a pore-network modeling approach in conjunction with a dynamic percolation approach. Different foam models have been tested and compared with well-known models for non-Newtonian fluids that exhibit similar rheological behaviors, such as Bingham plastic and Herschel-Bulkley fluid type. One striking finding is that the gas trapping fraction in both regular and irregular pore networks can be as high as 80 percent and is strongly dependent on local pressure gradient and pore-throat distribution. This finding is consistent with gas trapping experiments reported in the literature.
Simulations of foam processes on a field scale were conducted to provide a guideline for surfactant screening. One of the significant findings was that a higher surfactant partition coefficient results in a lower CO2 production rate over a relatively short period of time owing to deeper surfactant and foam propagation into the reservoir. However, this early-time production behavior dramatically changes at a later time. An increase in the surfactant partition coefficient leads to more spreading of the surfactant concentration distribution toward the producer, enhancing the significance of the surfactant concentration effect. Therefore, surfactants with relatively lower partition coefficients can improve vertical sweep efficiency and well injectivity. Substantial differences in optimal injection strategies for conventional foam and foam with CO2-soluble surfactants have been identified.
Foam corefloods with reference surfactants have been conducted to evaluate existing foam simulation models. Among these models, the CMG/STARS foam model has been successfully used to history match both corefloods and field foam trials.
Current Status (June 2014)
Research continues for developing classes of surfactants to design high-quality foams with very low surfactant concentrations. The selectivity for stable C/W foams in the absence of oil that break in the presence of oil and solubility in CO2 (to simplify injectivity) is a very important benchmark for advancing this technology. The distribution coefficients of the leading candidate surfactants currently being studied in corefloods will be determined in the next six months. Nonionic surfactants including Surfonic L24-22 , Neodol 25-9, and Neodol 25-12 have been tested in glass bead packs. Stable, viscous (up to 350 cP) CO2/South Hobbs Brine foam at 40 °C, 1700 psia, and superficial velocities approaching reservoir conditions (less than 10 ft./day) was formed. Experiments to determine the effects of salinity, temperature, and selectivity for CO2/water foams without forming emulsions of oil and water will continue. Coreflooding experiments with injection of oil as well as CO2 and water will continue. In addition, the partitioning of the surfactants in oil/water and CO2/water mixtures will be studied to determine their selectivity. These experiments will be closely integrated with the corefloods at larger scale in the Nguyen lab and adsorption experiments in the Hirasaki lab.
Adsorption of candidate surfactants on formation rock materials and pure minerals is being evaluated. Anionic surfactants are generally used in sandstone formations because of their lower adsorption rates compared to nonionic and cationic surfactants. Since the adsorption of anionic surfactants is high in carbonate formations, nonionic and ethoxylated cationic surfactants are being evaluated for carbonate formations. A nonionic surfactant (Tergitol 15-S-30) was found to adsorb less on dolomite than on any of the anionic, cationic, and zwitterionic surfactants that were evaluated. Ethoxylated cationic surfactants have low adsorption on pure calcite, but adsorption can be high in the presence of significant silica and/or clays.
The East Seminole field (part of the San Andres reservoir) operated by Tabula Rasa Energy was identified as a possible site for CO2 foam injection. The field is in a design phase for tertiary oil recovery by CO2 injection. Tabula Rasa will conduct a field test to perform CO2 foam injection for mobility control. Oil, brine, and cores from East Seminole field have been shipped to UT Austin for lab work. The Principal Investigator, Dr. Nguyen, will support Tabula Rasa on establishing a monitoring program for tracking CO2 breakthrough in the pattern and obtaining more information about fluids movement and heterogeneity.
Models for fluid properties such as equation-of-state and surfactant transport properties will be developed and verified against the PVT data and the static and dynamic adsorption experiments. These models will be used in conjunction with the geologic model to simulate the foam process on a field scale using CMG STARS.
Simulations of various injection strategies for field application have begun and will provide a guideline for surfactant screening and testing of foam in outcrop and reservoir cores under reservoir conditions. Existing foam models and current foam modeling approaches will continue to be evaluated and improved, respectively, based on coreflood results. A methodology for better estimating the foam parameters in the CMG STARS has been developed.
Project Start: January 12, 2011
Project End: December 31, 2014
DOE Contribution: $1,548,772
Performer Contribution: $425,113
NETL – Sinisha (Jay) Jikich (firstname.lastname@example.org or 304-285-4320)
University of Texas at Austin – Quoc Nguyen (email@example.com or 512-471-1204)
If you are unable to reach the above personnel, please contact the content manager.