Gels correct in-depth reservoir permeability variation

Julie E. Smith, James C. Mack
Tiorco Inc.
Denver

This unique in-depth permeability modification process has been applied in 37 field projects over the last 11 years.

In the 29 successful projects, the gels increased oil production at lower water/oil ratios.

To date, the successful projects have recovered an estimated 12.3 million bbl of incremental cumulative oil at an average chemical cost of $1.09/incremental oil bbl. Six projects were unsuccessful for a variety of reasons.

Estimated ultimate oil recoveries in successful treatments exceed 40% original oil-in-place (OOIP), in highly heterogeneous matrix reservoirs with Dykstra-Parsons permeability variations¹ of about 0.8. Chemical costs range from $0.47 to $4.08/bbl of incremental oil recovered.

The colloidal dispersion gel was successfully placed at reservoir temperatures up to 202° F. Oil viscosities were often high, from 10 cps to more than 20 cps at the reservoir bottom hole temperature.

CDG technology

High-permeability variation adversely affects volumetric sweep in many waterfloods. In heterogeneous reservoirs, permeability variation may extend throughout the reservoir; therefore, large gel volumes are needed deep in the formation, not just near the well bore.

If gels are placed only near the well bore, subsequent injection can bypass the gels by vertical crossflow. Two common gel injection methods are bulk and sequential.² In both methods, polymer is dispersed in injection water with a special polymer feeder, then injected into the formation.

With bulk gels, the crosslinker is added downstream of the polymer feeder using a high-pressure liquid chemical injection pump, to form a homogeneous pre-gel solution, or gellant. This then is injected into the formation.

Sequential injection involves injecting separate polymer and crosslinker slugs. This results in polymer and crosslinker layers forming on the pore walls within the rock.

Conventionally, bulk gels use high polymer and crosslinker concentrations to form strong gels near the well bore. The high polymer and crosslinker concentrations commonly employed in conventional bulk-gel treatments make it uneconomical to inject the large gel volumes needed for in-depth problems. Also, faster crosslinking reactions at high concentrations make in-depth placement difficult.

Sequential, or layered, injection in the past has tried to achieve in-depth placement by keeping polymer and crosslinker separated until both are in the formation. One problem in this is that it is hard to control when and how the polymer and crosslinker come into contact to form gel.

CDGs are homogeneous bulk gels developed specifically for in-depth use. The gels have low polymer and crosslinker concentrations that make injecting large gel volumes economical and allow in-depth placement. CDG performance is more predictable, and both crosslinker and polymer efficiency is improved over sequential gels.

Laboratory work

The first bulk aluminum citrate gels, or CDG, trials were in screen factor test units.³ These units are used routinely to evaluate uncrosslinked polymer flow in screen packs.

After 24 hr, the flow of the mixed solution through the screen packs was notably retarded. After 48-72 hr, the bulk gels would not flow at all in the screen factor unit. Although the solutions did block the screens, there was no evidence of crosslinking based on visual examination. The gels looked and poured like uncrosslinked polymer solutions.

These initial results suggested that bulk aluminum citrate gels were more effective at blocking flow in the screen packs than expected and that a tortuous flow test would be needed to evaluate the gels.

The screen factor apparatus was modified to allow testing the gels.⁴ Also, viscosity tests proved unsuitable unless a specialized helipath apparatus was employed. This was considered too time-consuming; therefore screen-pack data were chosen to quantify gel strength.⁵

Gels are now tested and compared using transition pressures. The transition pressure increases with increasing gel strength and represents the differential pressure below which these semifluid gels show markedly improved resistance to flow compared to uncrosslinked polymer in a screen pack.

It was determined that bulk CDG required significantly less aluminum than sequential gels. Polymer-to-aluminum ratios in the range of 20:1 to 100:1 work best. A very pure, partially hydrolyzed polyacrylamide provides the best gels.

Gels with cationic and emulsion anionic polymers were tried but failed. Gels work best with polymers that have higher molecular weights and a larger degree of hydrolysis.

Product purity is of paramount importance. Even minuscule amounts of certain additives in the polymer can adversely affect CDGs. Crosslinking reaction rate is a function of:

• Polymer concentration

• Polymer/aluminum ratio

• Water salinity

• Temperature

• Gel shear regime

• Age of aluminum citrate solution.

Some gel strength is lost if sufficient shear is applied, but the gels also tend to form more slowly if presheared. This is an advantage in the field.

Colloidal dispersion gels are more economical in fresher waters, requiring less polymer and crosslinker.

Transition pressures were related to resistance-factor data by testing gels with known transition pressures in a 15 Darcy sandpack. Fig. 1 [9243 bytes] shows that resistance factors were higher than transition pressures by several orders of magnitude.

Fig. 2 [32371 bytes] shows that a gel with a transition pressure of about 10 psi can divert flow from a high-permeability sandpack into a lower-permeability sandpack. In a parallel test, gel was injected simultaneously into two sandpacks, 19 and 1.7 Darcy, respectively. The upper plot shows fluid diversion as volume fraction of total flow into each sandpack. Uncrosslinked polymer injection caused diversion of about 20% of the total flow into the low-permeability sandpack.

Following CDG injection, diversion was complete, with 100% of the injected fluid entering the low-permeability sandpack. The lower plot, which traces the differential pressure across the sandpacks, shows that the residual resistance following colloidal dispersion gel injection was higher than that following straight polymer.

Initially, the pressure was about 0.5 psi. After uncrosslinked polymer, the pressure stabilized at 0.5-1 psi, and after the gel the pressure stabilized at about 3 psi. Flow diversion was maintained after returning to water injection.

Gels

Colloidal dispersion gels consist of low polymer and crosslinker concentrations. In the work described in this article, the crosslinker is aluminum citrate; however, other metals also give colloidal dispersion gels. Polymer concentrations normally range from 100 to 1,200 ppm.

In this concentration regime, there is not enough polymer to form a continuous gel network, so that a conventional bulk type gel cannot form. Instead, a solution of separate gel bundles forms, in which a mixture of predominantly intramolecular and minimal intermolecular crosslinks connect relatively small numbers of polymer molecules.

By contrast, in a conventional bulk gel the crosslinks form a continuous network of polymer molecules, through predominantly intermolecular crosslinks. The difference in the two types of gels is illustrated in Fig. 3 [23707 bytes]. Gel particle bundles have been described in other work, with chromium as the crosslinker, and referred to as aggregates.⁶

Field use

From a laboratory standpoint, the characteristics of colloidal dispersion gels provide a potential advantage in the field by positively affecting fluid flow distribution in heterogeneous reservoirs. The low polymer and crosslinker concentrations give the combination of slow reaction rate and large volumes needed for a successful in-depth fluid diversion process.

The low chemical concentrations allow economical injection of large volumes. The gel shear-thinning behavior is also an advantage.

Near an injection well, the gels behave like uncrosslinked polymer, and have no plugging effect. This is partly because the crosslinking reaction has not yet taken place and partly because the colloids will not swell until the differential pressure they are exposed to is decreased below their transition pressure.5

Deeper in the formation, when the gels have had time to react and the pressure differential has decreased, the colloids swell and restrict high-permeability channels.

Field development

Laboratory colloidal dispersion gel research was encouraging. The next step was to take the process to the field.

Two primary concerns still existed. Although basic laboratory data indicated the gels should propagate, field data were needed to show that the gels would travel in-depth in the formation. Also, concern existed relative to gel stability and longevity, which would ensure sweep improvement for the life of a waterflood.

Potential benefits included:

• CDGs improved and simplified field design over the sequential aluminum process. Development of a simple, rapid laboratory test, the transition pressure, allowed a better understanding of gel strengths with various polyacrylamide and crosslinker concentrations in field water. Changes in the field process could be made with some knowledge based on this laboratory testing. The sequential aluminum citrate process required expensive core work to quantitatively define the benefits of the process.

• CDGs provided better control in the field, allowing program changes if needed. The sequential process required long-term polyacrylamide injection, followed by a short-term slug of aluminum citrate, finally followed by more long-term polymer injection. An operator could not determine the benefits until finishing all three sequences, and by then it was too late to make changes. CDGs allowed continuous monitoring of injectivity and resistance build-up, so that changes could be made if needed.

• CDGs reduced aluminum citrate requirement. Resistance to water flow in the reservoir required considerably less aluminum citrate than in the sequential process.

• CDGs reduced overall program cost, not only by reducing the aluminum citrate requirements, but also by improving efficiency and control.

• Finally, CDGs improved waterflood performance by increasing waterflood efficiency and producing more oil and less water.

Initial field results

The Edsel Minnelusa Unit,⁷ in the Powder River Basin of Northeast Wyoming, was one of the first to employ CDG. A volumetric sweep improvement program was already in place. This program used the sequential aluminum citrate process for building resistance in high-permeability zones within the reservoir.

The Hall plot⁸ for the injection wells monitored the build-up of flow resistance. Fig. 4 [18239 bytes] shows a Hall plot for an Edsel injector.

Prior to beginning the CDG process, the Hall slope did not increase, but was instead fairly stable. Once the CDG process was started, the slope increased slowly while injecting 200,000 bbl of gel. This indicated greater flow resistance and gel propagation.

In the Tyler sand of Central Montana,⁹ the North Melstone Unit also employed this process about the same time. Similar Hall slope changes on injection wells verified that greater flow resistance was being built than with the sequential process. Large injection volumes suggested the process was placed in-depth in the reservoir.

Shortly after the experiences at Edsel and North Melstone, the process was tried at the Pfieler Lease,7 a Minnelusa reservoir in Northeast Wyoming. Again, the sequential aluminum citrate process was used first with limited results. Once the CDG system was implemented, the operator decided to reduce the injection rate.

To avoid the expense of changing pumps or the pump setup, the operator reduced the rate by placing the injection plant on a time clock. The plant was shut down twice a day, with the first shutdown lasting about 3 hr and the second about 9 hr.

Table 1 shows the pressure data for daily shut-ins over 6 months. Initially, the well, during the second shut-in, would go on a vacuum shortly after the shut-in. In February, the well went on vacuum after only 4 hr of downtime. In March, it went on a vacuum after 81/2 hr of downtime and, subsequently, the well maintained wellhead pressure that increased to 250 psi in about 6 months.

Although this is not a complete fall-off test, it does show that flow resistance was being built over a long period and suggests in-depth penetration of the colloidal dispersion gel. The pressure fall-off information did not indicate any skin damage.

Field projects

In the past 11 years, colloidal dispersion gels have been used in 37 reservoirs to improve volumetric sweep and oil recovery. Tables 2, 3, and 4 list the projects. The projects are broken into unsuccessful, marginal, and successful projects.

Unsuccessful projects

Of the 37 projects, 6 were considered unsuccessful. Much was learned about applying the technology for improving oil recovery from these unsuccessful projects.

Several projects were started too late. Once a waterflood matures, it becomes more difficult to improve volumetric sweep and overcome the water recycling and flow patterns in the reservoir. This was true at both the South Little Wall9 and Bishop Ranch projects.

The water cut at South Little Wall was greater than 90% when the process was initiated. During the program, changes in water/oil ratio and flow resistance at the injection wells were encouraging, but the results were too little, too late.

A similar situation existed at Bishop Ranch, in which case produced water volume was so great that it became difficult to inject the process in the freshwater.

Produced water was recycled for proper disposal. Freshwater for gel injection was restricted because of the produced water disposal requirement. Consequently, the process was in a catch-up mode throughout and never did achieve significant results.

The process potential may be limited by out of zone injection. Swartz Draw, Long Tree, and Ottie Draw were all cases of injection out of zone. Both Swartz Draw and Long Tree did exhibit oil response to injection, but the injection efficiency was extremely low, which reduced the economics and the effectiveness.

No response to the program at Ottie Draw occurred from the first injection well. The process was then changed to a second injection well where limited response, much like Swartz Draw and Long Tree, was noted at the offset producers. Because of out of zone losses, economics did not justify continuing these programs.

Poor well bore completions led to problems at the Wolf Draw Unit. A combination of near well bore damage and low-permeability rock restricted injection of uncrosslinked polymer, and certainly of the CDGs. The process was stopped after a serious reduction in injectivity, which hampered the overall waterflood economics.

The CDG process has an optimum window of application, in terms of gel strength. In some reservoirs, the process developed inadequate gel strength to provide the fluid diversion required to improve oil recovery. This was particularly true at both South Little Wall and Bishop Ranch.

The problem at South Little Wall was more severe than initially envisioned. Smaller volumes of stronger conventional bulk gels ¹⁰ might better be employed in this type of reservoir.

Poor water quality also leads to process problems. For CDGs, water quality is important in terms of both suitability for waterflooding and salinity. Similar to waterfloods, polymer-augmented projects should have water free of insoluble scale or corrosion products, oil carryover, and bacterial activity. The aluminum citrate colloidal dispersion gel process is best in fairly fresh water.

At South Bishop Ranch, the produced water was too saline for this process; therefore, CDG placement was restricted to freshwater injection. Because of the necessary produced water recycling, the program could only be implemented in a small portion of the total injection flow.

Water quality must be monitored and be in excellent shape for success with this process.

Marginal projects

Two projects, shown in Table 3, fall into the marginal category. These projects were successful from a technical standpoint, showing positive results on both the injection and production side, but overall economics were marginal.

Stewart Ranch, a large Minnelusa¹¹ reservoir in Northeast Wyoming, implemented a volumetric sweep improvement process starting in the early 1980s.The program initially used uncrosslinked anionic polyacrylamide to improve mobility ratio and provide minimum resistance to water flow. The program was changed to the sequential aluminum citrate process to provide better control of permeability variation in the reservoir and distribute water into the low-permeability zones. This proved effective in several areas of the field, but did not, in general, give positive results throughout the reservoir.

One injection well, toward the end of the program, utilized the colloidal dispersion gel process, which proved successful.

The Stewart Ranch reservoir geology is quite complicated, and as many as four separate reservoirs exist. Most of the injection wells were fractured; therefore, several injectors showed out of zone injection and had no effect on waterflood response.

Because the mechanics of the injection facilities were such that all wells were treated, the out of zone injection led to poor economics. Benefits of the volumetric sweep program were limited by the well completion methods, mechanics of the injection facility, and complex reservoir geology.

The other marginal program, the Soaphole lease, showed extreme channeling, and polymer broke through at the producers soon after CDG initiation. The project was temporarily shutdown to allow CDGs to set up better in the reservoir. Production was resumed recently, and time will tell if the strategy was successful.

Successful projects

Table 4 shows the 29 successful projects starting in 1985. The projects, show incremental oil recoveries ranging from 0.6 to 26.2% OOIP. The chemical cost per incremental barrel ranges from $0.47 to $4.08/bbl. The results can be summarized by improved waterflood efficiency, recovering more oil quicker while producing at lower water/oil ratios.

All projects have high Dykstra-Parsons permeability variations ranging from 0.6 to 0.9 and exhibit fairly heterogeneous behavior. The purpose of using the colloidal dispersion gel technology was to treat the heterogeneities within the reservoir, building up in-depth flow resistance in the high permeability watered-out zones and forcing water flow into the lower permeability rock.

The Townsend Newcastle Sand Unit (TNSU)¹² and the East Burke Ranch Unit were programs strictly designed for a permeability variation problem. Both projects were successful and both exhibit favorable mobility ratios. At TNSU, CDG injection at two injectors restricted water breakthrough at offset producers and improved waterflood efficiency.

The East Burke Ranch field is located adjacent to Sage Spring Creek,¹³ which in the late 1970s employed the layered aluminum citrate sequential process to improve volumetric sweep. The program at Sage Spring Creek proved effective, but CDG benefits over the sequential process were more apparent at the East Burke Ranch. The process was designed more accurately, with better control.

Positive results from utilizing colloidal dispersion gels occurred after premature water breakthrough at the North Rainbow Ranch Minnelusa Unit¹⁴ in Northeast Wyoming. Fig. 5 [20003 bytes] shows the time rate curve for the North Rainbow Ranch field. Fig. 6 [12055 bytes] shows the water/oil ratio-vs.-cumulative oil recovery curve for the total field.

When the process was implemented, the North Rainbow Ranch field had one injection well and two producing wells. Water had broken through at both of the producers, with water/oil ratio rapidly rising. The CDG program was implemented over about 2 years.

During the treatment, the oil production stabilized and increased following the treatment. At the same time, the water/oil ratio dropped, resulting in better efficiency and more oil.

Acknowledgments

We would like to thank Tiorco's customers, who are the operators of the projects discussed in this work.

References

1. Dykstra, H., and Parsons, R.L., "The Prediction of Oil Recovery By Waterflood," Chapter 12, API Secondary Recovery of Oil in the United States, 2nd Edition.

2. Needham, R.B., Threlkeld, C.B., and Gall, J.W., "Control of Water Mobility Using Polymers and Multivalent Cations," Paper No. SPE4747, Improved Oil Recovery Symposium, Tulsa, Apr. 22-24, 1974.

3. Jennings, R.R., Rogers, J.H., and West, T.J., "Factors Influencing Mobility Control By Polymer Solutions," JPT, March, 1971, pp. 391-401.

4. Smith, J., "Quantitative Evaluation of Polyacrylamide Crosslinked Gels for Use in Enhanced Oil Recovery," International ACS Symposium, Anaheim, Calif., Sept. 9-12, 1986.

5. Smith, J.E., "The Transition Pressure: A Quick Method for Quantifying Polyacrylamide Gel Strength," Paper No. SPE18739, SPE International Symposium on Oilfield Chemistry, Houston, Feb. 8-10, 1989.

6. Hejri S., Jousset, F., Green, D.W., McCool, C.S., and Willhite, G.P., "Permeability Reduction by a Xanthan/Chromium(III) System in Porous Media," SPE Reservoir Engineering, November 1993, pp. 299-304.

7. Hochanadel, S.M., Lunceford, J.L., and Farmer, C.W., "A Comparison of 31 Minnelusa Polymer Floods With 24 Minnelusa Waterfloods," Paper No. SPE/DOE20234, SPE/DOE 7th Symposium on Enhanced Oil Recovery, Tulsa, Apr. 22-25, 1990.

8. Hall, H.N., "How to Analyze Waterflood Injection Well Performance," World Oil, October 1963, pp. 128-30.

9. Mack, J.C., "Volumetric Sweep Improvement in the Heterogeneous Tyler Sand of Central Montana," Paper No. SPE18975, SPE Joint Rocky Mountain Regional/Low Permeability Reservoirs Symposium and Exhibition in Denver, Mar. 6-8, 1989.

10. Sydansk, R.D., and Smith, T.B., "Field Testing of a New Conformance-Improvement-Treatment Chromium (III) Gel Technology," Paper No. SPE/DOE17383, SPE/DOE Enhanced Oil Recovery Symposium, Tulsa, April 17-20, 1988.

11. Mack, J.C., and Duvall, M.L., "Performance and Economics of Minnelusa Polymer Floods," Paper No. SPE12929, Rocky Mountain Regional Meeting, Casper, Wyo., May 21-23, 1984.

12. Hochanadel, S.M., and Townsend, C.L., "Improving Oil Recovery in the Naturally Fractured, Tight, Dirty Sandstone of the Townsend Newcastle Sand Unit-Weston County, Wyoming," Paper No. CIM/SPE90-82, International Technical Meeting, Calgary, June 10-13, 1990.

13. Mack, J.C., and Warren, J., "Performance and Operation of a Crosslinked Polymer Flood at Sage Spring Creek Unit A, Natrona County, Wyoming," JPT, July, 1984, pp. 1145-56.

14. Fielding, R.C. Jr., LeGrand, F., and Gibbons, D.H., "In-Depth Drive Fluid Diversion Using an Evolution of Colloidal Dispersion Gels and New Bulk Gels-An Operational Case History of North Rainbow Ranch Unit," Paper No. SPE27773, SPE/DOE 9th Symposium on Improved Oil Recovery in Tulsa, April 17-20, 1994.

15. King, W.A., "Practical Application of a Reservoir Model to EOR: Lone Cedar (Minnelusa) Unit, Campbell County, Wyoming," Paper No. SPE17539, SPE Rocky Mountain Regional Meeting, Casper, Wyo., May 11-13, 1988.

The Authors