FIGURE 1. A 1,000-cp polymer flood result for Sor reduction for one layer, (top) is compared to a 1,000-cp polymer flood result for two layers, free crossflow (bottom), in this North Slope case study. (Images courtesy of NETL)
The US Department of Energy’s (DOE) Office of Fossil Energy has historically supported a large number of laboratory and field tests in an effort to improve oil recovery processes. Field-scale surfactant and polymer floods were launched in the 1970s, often as “demonstration” projects with R&D support. DOE also supported the 1984 National Petroleum Council (NPC) study on EOR. The NPC study evaluated the potential of three major EOR methods: thermal, miscible, and chemical flooding.
An extensive reservoir database that represented about 70% of total US original oil in place at that time was used. The chemical EOR section of the NPC study considered polymer, surfactant, and alkaline flooding. Screening criteria developed based on previous chemical EOR projects were used in the selection of reservoirs amenable to each process. Simplified predictive models for each process were used to estimate that nearly 40% of the US EOR potential is in chemical flooding. That fact supported a continued effort to develop effective surfactant/polymer systems. Studies also concluded that new surfactants and polymers were needed for a wide range of temperature and salinity conditions where available products were ineffective.
As a result of these ongoing initiatives, in the period between 2008 and 2010 the DOE’s National Energy Technology Laboratory (NETL) awarded a number of projects to further the next generation of chemical EOR, encouraging applicants to pursue small field or pilot testing. Two of the awarded research projects are related to newly developed surfactants and surfactant polymer combinations to reduce interfacial tension and surfactant adsorption in porous media. Two further projects were awarded to promote polymer applications in viscous oil, and another project was awarded to develop advanced reservoir characterization for implementation of alkaline surfactant polymer floods.
Other EOR research activities supported by NETL include focusing on the development of mobility control agents using surfactants injected with COrather than with water for COEOR in heterogeneous carbonate and sandstone reservoirs, field-testing gels for conformance control, developing advanced computer simulation and visualization capabilities, designing and testing an electromagnetic monitoring system to track the COflood front, and a field case study of an existing COflood targeting the residual oil zone.
FIGURE 2. Viscosity vs. shear rate and polymer concentration are plotted for both polymers.
This three-part series focuses on recently completed chemical EOR projects conducted by DOE, the theoretical and experimental results obtained, and the resulting field implementation.
Using water-soluble polymers to recover viscous oil
Several projects are the focus of the current study.
Fractional flow. This project examined the potential of polymer flooding for recovering viscous oils when the polymer is able to reduce the residual oil saturation (Sor) to a value less than that of a waterflood. Experiments and fractional-flow calculations were used to examine this question for conditions in North Slope reservoirs with viscous oils. Variables considered included oil viscosity, water/polymer viscosity, relative permeability characteristics, connate water and residual oil saturations, formation layering, presence/absence of crossflow, and pore volume throughput. It was found that induced changes in Sor could make a significant difference in recovery efficiency.
As expected, the impact of Sor reduction by a polymer flood on oil recovery is more pronounced in reservoirs where residual oil saturations are high at the start of polymer flooding. The impact of Sor reduction diminishes with increasing degrees of heterogeneity.
An important finding was that at low mobility ratios the two-layer free-crossflow recovery curves could approach those for one homogeneous layer. Figure 1 demonstrates this finding for 1,000-cp polymer displacing 1,000-cp oil. A polymer flood can be effective for recovery of viscous oils even if the reservoir is extensively waterflooded before application of the polymer flood. A reduction in Sor was beneficial for all water-flood scenarios that were examined.
FIGURE 3. G’, G”, and complex viscosity behavior were similar for the two polymers.
Rheology. The project also examined the rheology of a new sulfonic associative polymer in porous media. Associative polymers have been investigated as a possible substitute for hydrolyzed poly-acrylamide (HPAM) polymers in EOR applications. A new hydrophobically associative polymer, Superpusher DP/C1205, is an anionic-polyacrylamide-based tetra-polymer that has associative properties. Typically, the hydrophobic monomer content ranges from 0.025 mol% to 0.25 mol%. Molecular weights (Mw) range from 12 million g/mol to 17 million g/mol, and the total anionic content is between 15 mol% and 25 mol%. Less than 8 mol% sulfonic monomer is present. A process derived from micellar polymerization made the associative polymer, but the “hydrophobic” monomer used was amphiphilic and dissolved very well in water. A slight amount of surfactant was used as a process aid, but the final amount of remaining surfactant did not exceed 0.25% of the final polymer product and had a concentration in brine of less than 10 ppm, well below any critical micelle concentration.
Throughout the project, the brine contained 2.52% total dissolved solids (TDS), specifically including 2.3% sodium choloride and 0.22% sodium bicarbonate. The studies were performed at 25°C (77?F). This brine and temperature are representative of those associated with a large polymer flood in Canada. Comparisons have been made with the performance of a conventional HPAM, SNF Flopaam 3830S (Lot X 1899). Viscosity vs. shear rate and polymer concentration were plotted in Figure 2 for both polymers. For polymer concentrations of 500 ppm, 900 ppm, 1,500 ppm, and 2,500 ppm in 2.52% TDS brine, viscosity vs. shear rate was quite similar for the HPAM and associative polymers. Figure 3 demonstrates that the elastic modulus (G’), loss modulus (G”), and complex viscosity behavior also was very similar for the two polymers.
Compared with HPAM, the new polymer shows a significantly higher level of shear thinning at low fluxes and a lower degree of shear thickening at high fluxes. The associative polymer appears to contain a species that propagates through porous rock at rates comparable to those for HPAM and a second species that moves much slower and creates much higher resistance factors. This may improve recovery but also may present problems with injectivity and polymer transport.
FIGURE 4. In this figure showing a 500-ppm associative polymer viscosity and resistance factor in Berea sandstone, G’, G”, and complex viscosity behavior also were very similar for the two polymers.
Pore plugging and resistance factor vs. flux and concentration. Figure 4 plots resistance factor as a function of flux in each of the three core sections for injection of 500-ppm associative polymer in a Berea sandstone core. At a given flux, the resistance factors were reasonably consistent in the three core sections. The resistance factors were somewhat higher in the third core section, which argues against plugging of the inlet face.
Consistent with normal HPAM behavior, a strong shear thickening behavior was seen at moderate-to-high flux values, and Newtonian, or a slight shear-thinning behavior, was seen at low flux values.
Figure 4 also plots viscosity vs. shear rate for 500-ppm associative polymer (solid curve). Note that at all flux values, resistance factors were considerably greater than expectations from viscosity measurements. This behavior was noted in both Berea and porous polyethylene for all (fresh) associative polymer concentrations tested.
Residual resistance factors. At the end of associative polymer injection, many pore volumes (PV) of brine were injected to determine residual resistance factors in the Berea and polyethylene cores. After injecting 109 PV of brine into the polyethylene core, residual resistance factors were 1.9, 2.1, and 1.3 in the first, second, and third sections, respectively. This result is consistent with a suggestion of little to no significant pore plugging in the 12,313-md polyethylene core. However, after injecting 170 PV of brine into the Berea core, residual resistance factors were 13, 19.3, and 15.3 in the first, second, and third sections, respectively. This result suggests that some pore plugging or higher polymer retention may have occurred in the 363-md Berea core. In both cores, residual resistance factors were not sensitive to flow rate.
FIGURE 5. Effluent viscosity vs. shear rate and flux is shown at 1,000-ppm C1205 in 0.23% sodium chloride and 0.022% sodium bicarbonate.
Mechanical degradation of an associative polymer vs. HPAM was performed to investigate the effects of reservoir brine salinity and divalent cations on mechanical (shear) degradation of a new hydrophobic associative polymer, SNF C1205. This polymer has a small fraction of hydrophobic monomer groups along the acrylamide/ acrylate backbone. Degradation was induced by forcing polymer solutions through a short core plug using a wide range of rates. The study investigated solutions with salinities from 0.252% TDS to 12.65% TDS and with divalent cation concentrations from 0 to 1.15%. Effluent viscosities and screen factors were measured to determine the degree of mechanical degradation.
For comparison, a conventional HPAM polymer with similar molecular weight was submitted to the same tests. As expected, mechanical degradation of the associative polymer increased with velocity in porous media, and rheology of the associative polymer in the viscometer showed Newtonian behavior at low shear rates and shear thinning at higher shear rates. As with HPAM, calcium exerted a stronger influence than sodium in reducing solution viscosity. However, tests demonstrated that the degree of mechanical degradation of associative polymer solutions does not always increase with ionic strength. This finding differs from previous reports for HPAM.
Effect of injection rate. The rheology of C1205 polymer solutions shows conventional behavior: Newtonian behavior at low shear rates, shear thinning at intermediate shear rates, and near-Newtonian behavior at very high shear rates. For an acrylamide copolymer with 0.75 mol% N-octylacrylamide, as polymer concentrations increased above 3,000 ppm, regions of Newtonian behavior, shear thickening, and shear thinning were observed as shear rate increased. This complex behavior was due to shifting the relative amounts of intermolecular and intramolecular associations. No shear thickening behavior was observed with C1205 in this project because only 1,000-ppm polymer was used or perhaps because C1205 contained significantly less hydrophobe than Bock’s polymer (less than 0.25% vs. 0.75%). Figure 5 describes effluent viscosities for 1,000-ppm C1205 solutions in brine with 0.252% TDS injected at various rates.
Polymer flooding has tremendous potential for enhanced recovery of viscous oil. Field cases now exist where 200-cp to 300-cp polymer solutions are injected without significant reductions in injectivity. Concern about costs is reduced substantially by realization that polymer viscosity increases approximately with the square of polymer concentration. Viscosity can be doubled with only a 40% increase in polymer concentration. Up to a readily definable point, increases in viscosity of the injected polymer solution are directly related to increases in sweep efficiency and oil recovery. Previously published simulation results, suggesting that shear-thinning polymer solutions were detrimental to sweep efficiency, were shown to be unfounded.
Acknowledgments
The work described in the article was supported by awards from DOE/NETL. The author acknowledges permission given by the DOE to write the paper. The paper is based on reports presented by the principal investigators: Randy Seright, New Mexico Tech; Stan McCool, University of Kansas; Jeff Harwell, University of Oklahoma; and Beverly Seyler, Illinois Geological Survey. These and numerous other researchers contributed to the information in the paper.
References available.