EOR TEST DETERMINES WAYS TO HALT DOWNDIP STEAM MIGRATION

Lynelle S. Bautista
Chevron USA Production Co.

Eric A. Waninger
Chevron Chemical Co.
Bakersfield, Calif.

A one-pattern steam foam field trial determined a number of recommendations to prevent the onset of downdip steam migration in new steam floods, and addressed downdip steam migration once it occurs.

The trial was conducted in the Amnicola sand in the steam flood of Section 26W, Cymric field, near Bakersfield.

The pattern was in an area of downdip steam migration, soon after steam injection began, as a result of well spacing (which skewed injectors closer to downdip producers) and channeling (because of vertical permeability stratification).

The primary objective of the trial was to economically minimize the downdip steam migration, and consequently improve the sweep efficiency of the steam flood.

Chevron Chaser SD1020 was continuously injected with steam and nitrogen for about 7 months. Reservoir conditions and production were monitored and analyzed to determine the effectiveness of steam foam as a mobility control agent.

The held trial also identified nonfoam alternatives effective in reducing downdip steam migration. Operating strategies employed in the steam flood to accelerate early oil production were analyzed to understand their effect on the reservoir, given skewed well spacing and vertical permeability variation.

The accompanying box summarizes the objectives and recommendations from the trial.

PROJECT DESCRIPTION

The 26W steam flood targets the Tulare and Amnicola sands with a three-row line drive configuration (Fig. 1). The 34-acre, project lies on a 15-45 dipping, northeastern flank of a northwest-southeast trending antictine that continues through the Cymric field. Rock and fluid properties and typical performance characteristics are similar to those throughout the Cymric field (Table 1).

The flooded sands are distinct and continuous throughout the property. The Tulare reservoir is characterized as fluvial/deltaic in deposition, with generally high, and vertically uniform permeability within each of its five distinct sands (Fig. 2).

In contrast, the Amnicola reservoir is comprised of a single lucustrine sand and is characterized by coarsening upward grain size. Core data indicate higher permeability and oil saturation in the top 10-20 ft of the 40-60 ft thick Amnicola sand.

The Tulare and Amnicola are structurally similar with 15-45 dips within the drive area.

PROJECT HISTORY

The area was first developed in 1973 as a successful 14 well cyclic steam project. Because the cyclic response declined over time, continuous steam injection was started in August 1989.

Well spacing was based on studies recommending that to increase areal sweep in a steeply dipping reservoir, injectors should be closer to downdip producers. 1 2

In May 1990 after 10 months of steam injection, sudden increases in flow line temperatures, static temperature profiles, and casing pressures indicated steam migration to downdip producing wells 3-10B and 4-10B. Static temperature profile surveys for 3-1OB and 4-10B taken in May 1990 indicated over 320 F. and 390 F., respectively , in the Amnicola sand (Fig. 3a).

Static fluid level measurements indicated Amnicola reservoir pressure of 380 psi, corresponding to a saturated steam temperature of 440 F. Although static temperature measurements did not indicate steam breakthrough, they did suggest direct communication between injectors and downdip producers.

Additional data pinpointed steam migration in the Amnicola. Contact temperature data taken in downdip observation well 701E showed heat response in only the Amnicola sand (Fig. 3b). In addition, wells 3-10A and 4-1OA, which are completed in only the upper Tulare sands, did not show high temperatures or casing pressures.

In response to high temperatures, pumping fluid levels were increased in all downdip producers. While this operating strategy decelerated oil production, it reduced pressure drawdown into the producer while cooling the immediate well bore to prevent damage to the casing and tubing (Fig. 3a).

Subsequent static temperature surveys indicated that increasing pumping fluid levels significantly reduced maximum temperatures, but did not modify the profile shape. In addition, casing pressure dropped when well bore pressure was increased.

Simulation studies later indicated that increasing well bore pressure should eventually redirect steam migration updip and improve areal sweep efficiencies. 3 4

STEAM MIGRATION

It was initially suspected that cyclic steaming prior to steam flooding may have created preferential steam paths.5 Analysis, however, of cyclic steam history indicated that the downdip area was not actively steamed in the Amnicola sand. The original downdip producers were not drilled through the Amnicola.

It was found that downdip steam migration may be due to one, or a combination of:

• The proximity of the injectors and producer

• The vertical stratification within the Amnicola sand.

Based on previous studies, the injectors in the 26W steam flood were skewed downdip to approximately one third of the pattern distance from the downdip producers, placing the injector at an average 160 ft from downdip producers. 1 2

Because of high reservoir pressures and initial oil viscosities, injectivity into the Amnicola was poor, and it was difficult for steam to migrate into the cold updip area. Concurrently, both updip and downdip producers were cyclically steamed and pumped off.

The downdip area, however, was more evenly heated because of closer well spacing and because of larger steam cycle volumes. This improved the heating of the area between the downdip producers and injector, while also creating a pressure drawdown. These factors could create an environment in which viscous forces dominate gravity.

Amnicola grain size is characteristically coarsening upward, with injectivity highest in the top 10-20 ft of the sand. Initial injection rates, however, were based on processing the total net oil sand.

The total steam injection, therefore, may be channeling at high rates through a small, highly permeable stringer at the top of the Amnicola. This channeling may have been aggravated because of excessively high steam injection rate factors of 2.0 b/d/acre-ft used in the Amnicola, relative to the 1.5 b/d/acre-ft used in the Tulare. With such high injection rates, the effect of gravity is reduced, and piston-like displacement occurs. 6

Steam migrates concentrically away from the injector and will first be observed in the closest well, which in this case, because of offset well spacing, is the downdip producer.

Simulation studies support the sensitivity of steam migration to well spacing dip, and vertical stratification. 2-4 7

Overall a real sweep efficiency and recovery may still be higher with skewed well spacing despite consequent problems due to downdip steam migration. A different operating strategy, such as maintaining high well bore pressures or the use of steam foam, can help to minimize or prevent downdip steam migration to realize the full benefit of improved areal sweep with skewed well spacing.

FIELD TRIAL

The two objectives of this trial were to determine the effectiveness of steam foam in:

1. Economically reducing downdip steam migration from injector to downdip producer

2. Improving areal steam distribution in reservoirs with downdip steam migration.

In designing the field trial and monitoring program, it was important to identify specific parameters that determine the success or failure of meeting these objectives.

Objective No. 1 would be met by reduced reservoir temperature or casing pressure, or by increased oil production without an increase in breakthrough conditions in downdip producers.

Objective No. 2 would be fulfilled when the pattern with surfactant injection showed evidence of improved areal steam distribution relative to base line conditions.

CONTROL ANALYSIS

To quantify the incremental affect on production and reservoir conditions because of steam foam alone, surfactant and nitrogen were injected into a single pattern. The pattern, designated as the test pattern, had downdip steam migration.

An adjacent pattern also experiencing downdip steam migration was designated as a control pattern to monitor relative changes in production, temperatures, and casing pressures in the downdip area without steam foam application (Fig. 1).

All available data show similarity in geologic and reservoir characteristics, and in production and operating histories between the test (4-10W) and control (3-10W) patterns. Any operational changes, such as well pulling and pumping speeds, were imposed on both the control and the test patterns.

SURFACE SETUP

A trailer-mounted system was used for nitrogen generation and chemical injection. A pressure swing adsorption unit and compressor provided on site compressed nitrogen for the field trial.

A 10,000 gal storage tank, a 500 gal tank, and a variable-stroke, positive-displacement pump were used to inject surfactant at the injector wellhead (Fig. 4).

The steam-injection rate into test injector 4-10W was regulated with a variable orifice-size choke. To maintain critical flow and control of the steam rate, the orifice

size was increased as the bottom hole injection pressure increased with surfactant injection.

The steam rate and quality were monitored with a portable steam testing vessel to design and maintain chemical concentration and nitrogen injection rates. To maintain an adequate liquid-volume-fraction of steam, a soft water injection line was connected upstream of the surfactant and nitrogen injection points on the steam line leading to the wellhead of injector 4-10W.

The test pattern injector, well 4-10W, is dually completed, isolating Tulare injection (short string) from Amnicola injection (long string) with a packer. Because there was no evidence of downdip steam migration in the Tulare sands, it was decided to inject surfactant into the Amnicola only.

Insulated tubing between the dual packers minimized heat transfer between the short and long string. There was no evidence of excessive scale formation in the long string.

TRIAL DESIGN

The trial began in May 1991. Chaser SD 1020 was injected at 0.32% by weight surfactant into 256 bbl of steam/day CWE (cold water equivalent) at 82% quality (84 MMBTU/day heat injection). A liquid-volume-fraction of 0.55% was maintained. Over the 7 month life of the project, approximately 18,750 gal of surfactant and an average of 2.2 Mcfd of nitrogen were injected into test injector 4-1OW.

STEAM INJECTION

Initial increases in bottom hole injection pressure data indicated that the original steam properties, surfactant concentration, and nitrogen supply were adequate for foam generation. It was originally decided, therefore, to maintain, through the trial period, initial target steam rates of 250-300 b/d and 240 b/d in the test and control patterns, respectively, both at 80% quality.

During the first half of the field trial the steam supply to the 26W steam drive was erratic in quality and total rates due to steam plant modifications. As a result, the bottom hole injection pressure data did not conclusively indicate the generation of foam. Therefore, in September 1991 the liquid-volume-fraction in the test pattern was increased from 0.55 to 2.1%, decreasing the steam quality from 82 to 56% in the Amnicola (long) string with make-up water to ensure foam generation.

The total steam injection rate was increased to 350 b/d, resulting in a slightly higher heat-injection rate from 84 MMBTU/day to 94 MMBTU/day.

Although the chemical injection rate was increased, the surfactant concentration was maintained at 0.32% by weight.

These steam, surfactant, and nitrogen design rates were unchanged through the remaining 3 months of the field trial. Steam quality and rates were returned to pretrial conditions at the termination of surfactant injection.

Despite significant modification to the test pattern, similar adjustments were not made in the control pattern because of difficulty in determining the proper adjustment for analogous modification.

Because of the slightly lower Amnicola reserves the control pattern initial target steam rate of 240 b/d at 80% quality (79 MMBTU/day injected heat) was already lower than the heat injected in the test pattern at the start of the foam trial in May 1991 (84 MMBTU/day) and in September 1991 (94 MMBTU/day).

Moreover, uncontrollable and frequent chances in the steam supply made it difficult to establish base line conditions in the control pattern. Because of high costs, the extensive quality and rate testing on the test injector 4-10W were not acquired for control in ector 3-10W.

FOAM INJECTION

Seven months of continuous surfactant injection began in May 1991 and lasted through December 1991. in designing the field trial, preliminary studies indicated that 6 months of surfactant injection should be adequate to substantially reduce downdip steam migration. 3 4 Temperature, bottom hole-injection pressures, injection quality and rates, casing pressures, and pumping conditions were monitored regularly during the 7-month trial and continued for the 6-month anaivsis following the termination of surfactant injection.

Prior to surfactant injection, both producers 3-10B and 4-10B were carrving approximately 500 ft of fluid in the well bore to maintain high well bore pressures, and consequently to minimize steam migration, as discussed previously.

After 3 months of surfactant injection, the pumping speeds on both 3-10B and 4-10B were increased, decreasing well bore pressures, and increasing displacement. Pumping speeds were repeatedly increased throughout the remaining 3 months of surfactant injection.

FOAM GENERATION

As discussed previously, bottom hole injection pressures did not conclusively indicate foam generation in the first four months of surfactant injection (Fig. 5).

Following the adjustment to liquid-volume-fraction in September 1991, the bottom hole injection pressures increased by 40 psi, indicating adequate foam generation. The termination of surfactant injection in December 1991 resulted in a surface injection pressure drop from 570 to 410 psig, providing additional evidence of foam generation during the trial.

TEMPERATURE, PRESSURE

During surfactant injection, maximum Amnicola temperatures from static temperature surveys decreased by approximately 70 F. in the test pattern (Fig. 3a). Through the trial period, static temperature surveys in the control pattern show a monotonical increase of 50 F.

Static fluid level surveys showed concurrent a decrease in Amnicola reservoir pressure of 70 psi in both the test and control patterns (Fig. 6a). Casing pressure increased by 30 psi in the test pattern after surfactant injection was terminated, while casing pressures remained relatively constant in the control pattern.

During the test and analysis period, the reservoir pressure dropped 70 psi, corresponding to a decrease in the saturated steam temperature of 20 F. To normalize the effect of an overall decrease in saturated steam temperatures of 20 F., the ratios of maximum measured Amnicola temperature to saturated steam temperature through time were compared (Fig. 6b).

Normalized temperatures also show decreasing temperatures of 13% during surfactant injection and increasing temperatures of 20% after surfactant injection was discontinued. In contrast, Amnicola temperatures monotonically increased by 15% in the control pattern.

Reservoir temperature and casing pressures are comparable in both the test and control patterns after termination of surfactant injection.

Fig. 6b also shows that Amnicola temperatures did not dramatically increase in the control pattern as a result of decreasing well bore pressures in August 1991. Such a long-term effect on Amnicola conditions indicates that the imposed increase in well bore pressure since June 1990, when downdip breakthrough was first identified in the project, was effective in improving areal sweep and altering steam migration.

Reservoir pressure continued to decrease throughout the trial and analysis period, possibly as a result of ongoing steam communication from injectors to downdip producers in both the test and control patterns.

As typically observed in steam flood operations, reservoir pressure increases as the steam front moves toward the producer, then decreases dramatically as the steam zone continues to develop.

The similarity in reservoir pressure data in the test and control patterns suggest that the lower steam quality in the test pattern did not significantly affect the test results; in the lower quality cooled the reservoir and quenched the steam zone, the reservoir pressure in the test pattern would have dropped much more drastically than in the control pattern.

Because of a significant reduction in reservoir temperature in the test pattern, it appears that steam foam was effective in redirecting a portion of the steam which had been channeling directly downdip. Reservoir pressure data, however, suggest that because reservoir pressures did not increase in the test pattern, steam foam did not completely eliminate downdip steam migration or collapse the existing steam zone.

The immediate effect of terminating surfactant injection on reservoir temperature and casing pressure further supports that steam foam did not drastically, or permanently affect the downdip steam zone. Such rapid degradation of foam could be, in part, a result of returning to the pretrial steam injection properties, and consequently a lower liquid-volume-ratio, at the termination of surfactant injection.8

Although monitoring data suggest that steam was no longer channeling exclusively downdip during the trial period, the new direction of steam migration was not revealed. Observation wells 701E (downdip of test injector) and 702E (on-strike with test injector) showed no significant change in Amnicola reservoir temperature trends.

Moreover, updip producers in the control and test patterns were affected by injectors in adjacent updip patterns, and consequently could not indicate any changes due to steam foam.

SF6 tracer tests were inconclusive in determining either the direction or magnitude of steam migration due to steam foam.

Significant reduction maximum Amnicola reservoir temperature in the test pattern, and no evidence of areal redistribution of steam, could suggest a vertical redistribution.

Simulation studies showed that steam foam should be effective in temporarily improving vertical sweep, although areal steam redistribution would not occur.4 However, because of limitations in monitoring data, vertical redistribution could not be substantiated.

Perforations in the Amnicola were too closely spaced for accurate steam injection profile surveys. Temperature surveys in nearby observation wells showed no change in vertical profiles, although redistribution because of steam foam could have been misinterpreted as conductive heat. Moreover, static temperature profiles in producers could not show the detail needed to verify improvement of vertical sweep in a 50-ft sand.

Because the limitation in monitoring ability was understood prior to implementing the field trial, improvement in vertical sweep was not included as a primary objective of the study.

INFILL PRODUCERS

To accelerate Amnicola production in the downdip area of the 26W steam drive, four infill producers were drilled in January 1991. Because of mounting evidence of downdip steam migration, the infill wells remained inactive until the long-term effects of downdip steam migration were analyzed.

To complement the steam foam field trial, infill wells 3-1OC, 4-10C, and 5-10C were opened to production in August 1991, concurrent with decreasing pumping well bore pressure in downdip producers 3-10B and 4-10B.

The effect of pressure drawdown because of the production of the infill wells could influence reservoir forces and redirect steam. It is assumed that such effects would be similar in both the control and the test patterns.

Relative to control pattern infill wells 3-10C and 4-10C, test pattern infill well 5-10C shows delayed temperature response in the Amnicola (Fig. 7). While surfactant injection may have been adequate to reduce downdip steam migration during the trial period, it was not sufficient to permanently alter the balance of gravity over viscous forces.

PRODUCTION PERFORMANCE

Relative to base line production rates in test pattern producer 4-10B, oil production increased by 150 bo/d (Fig. 8a). In contrast, the control pattern producer 3-10B increased in oil production by 50 bo/d (Fig. 8b).

Because of the large production increase in the test pattern, it was assumed that the lower-quality steam did not significantly affect the test results. Oil production should have suffered if the lower quality steam significantly cooled the steam zone.

To normalize the effect of slight differences in operating conditions, the productivity indexes of control pattern downdip producer 3-10B and test pattern producer 4-10B were compared (Table 2).

Prior to surfactant injection, both wells 3-10B and 4-10B maintained a 0.34 bo/d/psi productivity index. During and after surfactant injection, the productivity index of well 4-10B increased dramatically to 1.58 bo/d/psi. The productivity index remained at 0.34 bo/d/psi in the control pattern well 3-10B throughout the trial and analysis period (Fig. 9).

Because of this increase in productivity, the estimated impact of steam foam is 75 bo/d, which was maintained during the latter 3 months of surfactant injection. Both the productivity and oil production remained above prefoam base line conditions during the 6 month analysis which followed.

As mentioned previously, the reserves of the control pattern are 22% lower than that of the test pattern. While oil production rates are extremely dependent on reserves, productivity should remain independent. Therefore, because base line productivity indexes were identical in the control and test patterns, it is assumed that the difference in Amnicola pattern reserves is irrelevant to the results of this field trial.

ECONOMICS

The field trial cost was about $110,000 over 7 months of surfactant injection and 6 months of postsurfactant injection monitoring. This included $60,000 for chemical and nitrogen generation, $20,000 for surface equipment, and $30,000 for monitoring.

Incremental oil production from well 4-10B in the test pattern offset total trial costs by January 1992, when surfactant injection was terminated. With sustained increase in oil productivity during the 6 month post-surfactant analysis period, the steam foam field trial was economically successful.

The production impact from updip producers and from infill wells was not included in the economic analysis.

ACKNOWLEDGMENTS

The authors would like to thank Francois Friedmann, Mike Smith, Phil Gauglitz, and Don Shay of Chevron Oil Field Research Co. for their help in designing monitoring, and evaluating this field trial. We would also like to thank Chevron USA for the opportunity to publish the results of this project.

REFERENCES

1. Hong, K.C., "Effects of Gas Cap and Edgewater on Oil Recovery by Steam flooding in a Steeply Dipping Reservoir," Paper No. SPE 20021, California Regional Meeting of SPE, Ventura, Calif., April 1990.

2. Hong, K.C., "Optimum Well Location for Steamflooding Steeply Dipping Reservoirs," Paper No. SPE 21771, Western Regional Meeting, Long Beach, Calif., March 1991.
3. Friedmann, F., Personal Communication, 1991.

4. Friedmann, F., "Strategies to Improve Steam Sweep Efficiency in a Steeply Dipping Reservoir," unpublished report, Chevron Oil Field Research Co. Technical Memorandum, November 1991.

5. Miller, K.A., Stevens, L.G., and Watt, B.J., "Successful Conversion of the Pikes Peak Viscous-Oil Cyclic Steam Project to Steam drive," JPT, December 1991, pp. 1510-17.

6. Van Lookeren, J., "Calculation Methods for Linear and Radial Steam Flow in Oil Reservoirs," Society of Petroleum Engineering Journal, June 1983, pp. 427-39.

7. Abad, B.P., and Hensley, C. A., "The Effect of Stratification, Fractures, and Dip on Steam Sweep Efficiency and Heat Requirements of Steamflood Operations," Paper No. SPE 12748, California Regional Meeting held in Long Beach, Calif., April 1994.

8. Friedmann, F., Smith, M.E., Guice, W.R., Gump, J.M., and Nelson, D.G., "Steam Foam Mechanistic Field Trial in the Midway-Sunset Field," Paper No. SPE 21780, Western Regional Meeting, Long Beach, Calif., March 1991.

9. Hong, K.C., and Stevens, C.E., "Water-Alternating-Steam Process Improves Project Economics at West Coalinga Field," Paper No. CIM/SPE 90-84, Joint CIM/SPE Meeting, Calgary, June 1990.

Copyright 1993 Oil & Gas Journal. All Rights Reserved.