Summary

Foams are used to control and improve the injection profile in secondary or tertiary gas-injection processes and reduce gas mobility far from the injectors. Over the last decade, significant progress has been made in the understanding of the complex foam processes in porous rocks. The goal of this paper is to review consistently, albeit somewhat subjectively, several important field tests of foams, compare their performance, and critically evaluate the economic benefits from foam injection. It is shown that early, transient, and usually small oil production responses to surfactant injection are real, and depend critically on the reservoir architecture and gas flood implementation. These early production responses are related to the improvements of gas-injection profile by foam and are often quite profitable. The delayed, but bigger oil production responses are caused by foam propagation into the reservoir and could be very profitable, depending on the injection policy. An outline of an ideal future foam pilot is presented, and important advances in rigorous modeling of foam processes are discussed.

Introduction

Although there are eyewitness reports19 of surfactant and air injection in Russia and China in the early sixties, the first documented field test of air foam was carried out in the Siggins field8 (Oct. 1964 to March 1966). The first steam-foam test was conducted in a Kern River steamflood pilot3 (Oct. 1976 to March 1977). Since then, there have been about 25 significant steam-foam projects (most of them listed and referenced in Ref. 19, Tables 1 through 4) and dozens of smaller commercial applications.2 As of this writing, reports of nine CO2 foam projects appeared in the literature.19 Noncondensible gas foams are being used worldwide by service companies for well cleanup and stimulation.19 A successful foam fracturing of a well in a pressure-depleted (subhydrostatic) reservoir has been reported in India.19 Use of areated (foamed) drilling muds in areas of subhydrostatic reservoir pressures has also been reported.19 In China, a noncondensible gas foam mixed with cement has been used successfully to block thief zones near water injectors.19

It is well known that the average bubble size in a foam controls its mobility.7 Because of the limited space here and excellent reviews elsewhere,7,10 I will omit the fundamental research on foam mechanisms and transport. Instead, I will compare several recent field tests of foams and point out their similarities, important differences, and lessons learned from them. I will try to avoid repeating already published materials. This paper is also a critique of the historical pilots and will end with a wish list for an "ideal" pilot in the future. Finally, I will summarize the current advances in modeling of foam flow in porous media.

Steam Foam

Background.

Dozens of steam-foam projects have been implemented to do the following.

  • Limit gravity override in flat and moderately thick reservoirs.

  • Improve vertical sweep in massive and dipping reservoirs.

  • Improve areal sweep and improve areal steam distribution.

  • Inject steam into lower zones, otherwise unheated in thin and/or layered reservoirs.

  • Improve injection profile, block thief zones, and reduce steam channeling.

  • Improve heat distribution in steam soaks.

  • Reduce residual oil saturation to steam.

Performance of these steam-foam projects was measured by the following.

  • Mass of surfactant injected per unit volume of incremental oil.

  • An increase of steam injection pressure.

  • A redistribution of steam injection profile.

  • A reduction of steam production rate and casing pressure.

  • An improvement of vertical and areal sweep (reservoir desaturation and heating).

There are two approaches to enhancing heavy oil recovery with steam foam. Historically, the first one was to inject surfactant continuously for several years to propagate foam into the reservoir as far as possible and to displace oil. Good examples of this approach are the four-pattern Shell Mecca and Bishop pilots in Kern River17 and Unocal Dome-Tumbador pilot in Midway Sunset (MWSS).12 The two Kern River pilots were conducted in a moderately thick (27 to 30 m gross), slightly dipping (3°), and relatively uniform reservoir that suffered from a severe gravity override. Steam was injected either at the base or full interval. The Dome-Tumbador pilot was conducted in a massive (137 m gross), coarsely layered (three major zones), and dipping (16°) reservoir that also suffered from a severe gravity override. Steam foam was injected at the base of the bottom zone and steam leaked somewhat into the overlying two zones.

Another approach was to use foam to improve the steam injection profile. The best examples here are the Mobil Tulare pilot4 and Chevron Sec. 26, Pattern 68BW, MWSS.25 The Tulare pilot was conducted in Zones Band C of a finely layered, thick (61 m), and dipping (6°) reservoir that suffered from the nonuniform injection profile when steam was injected full interval. Finally, the Sec. 26 MWSS pilot was conducted in a thick (107 m gross), finely layered and dipping (20°) reservoir, with full-interval steam injection into 12 separately perforated intervals that spanned 70 m. Reservoir permeabilities varied greatly, but were as high as 2 µm2 [2 darcy] in all the pilots.

The injector-producer distances varied from 46 m at Dome-Tumbador to 142 m in Tulare. The average rate of steam injection [in m3/d of cold water equivalent (CWE) per injector] was 40 in Kern River, 87 in Dome-Tumbador, III in Tulare, and 80 in Sec. 26. Nominal steam quality was about 50% by mass at the wellhead. Tables 1 through 4 in Ref. 19 reveal other details.

Improved Injection Profile.

Using krypton and sodium iodide as vapor and liquid tracers, respectively, in steam-foam injector 68BW, Chevron25 confirmed that steam injection profile was improved by foam and that this improvement disappeared rapidly when surfactant injection was stopped. A longer-lasting profile improvement was reported for another injector in MWSS.26 Although not measured directly, steam injection profile was vastly improved in another layered reservoir of comparable gross thickness (Mobil's Tulare), simply because no other plausible mechanism could account for the significant oil response.

For the single and/or partial injection intervals, such as in Shell's Mecca and Bishop and Unocal's Dome-Tumbador, the profile improvement is difficult to quantify and perhaps less important; steam flow in the reservoir is determined more by areal heterogeneities than the injector itself. However, even in a thin sand package (8 m), such as the Guadelupe field, a redistribution of foam injection profiles was noted in four wells.13

p. 79-85

Background.

Dozens of steam-foam projects have been implemented to do the following.

  • Limit gravity override in flat and moderately thick reservoirs.

  • Improve vertical sweep in massive and dipping reservoirs.

  • Improve areal sweep and improve areal steam distribution.

  • Inject steam into lower zones, otherwise unheated in thin and/or layered reservoirs.

  • Improve injection profile, block thief zones, and reduce steam channeling.

  • Improve heat distribution in steam soaks.

  • Reduce residual oil saturation to steam.

Performance of these steam-foam projects was measured by the following.

  • Mass of surfactant injected per unit volume of incremental oil.

  • An increase of steam injection pressure.

  • A redistribution of steam injection profile.

  • A reduction of steam production rate and casing pressure.

  • An improvement of vertical and areal sweep (reservoir desaturation and heating).

There are two approaches to enhancing heavy oil recovery with steam foam. Historically, the first one was to inject surfactant continuously for several years to propagate foam into the reservoir as far as possible and to displace oil. Good examples of this approach are the four-pattern Shell Mecca and Bishop pilots in Kern River17 and Unocal Dome-Tumbador pilot in Midway Sunset (MWSS).12 The two Kern River pilots were conducted in a moderately thick (27 to 30 m gross), slightly dipping (3°), and relatively uniform reservoir that suffered from a severe gravity override. Steam was injected either at the base or full interval. The Dome-Tumbador pilot was conducted in a massive (137 m gross), coarsely layered (three major zones), and dipping (16°) reservoir that also suffered from a severe gravity override. Steam foam was injected at the base of the bottom zone and steam leaked somewhat into the overlying two zones.

Another approach was to use foam to improve the steam injection profile. The best examples here are the Mobil Tulare pilot4 and Chevron Sec. 26, Pattern 68BW, MWSS.25 The Tulare pilot was conducted in Zones Band C of a finely layered, thick (61 m), and dipping (6°) reservoir that suffered from the nonuniform injection profile when steam was injected full interval. Finally, the Sec. 26 MWSS pilot was conducted in a thick (107 m gross), finely layered and dipping (20°) reservoir, with full-interval steam injection into 12 separately perforated intervals that spanned 70 m. Reservoir permeabilities varied greatly, but were as high as 2 µm2 [2 darcy] in all the pilots.

The injector-producer distances varied from 46 m at Dome-Tumbador to 142 m in Tulare. The average rate of steam injection [in m3/d of cold water equivalent (CWE) per injector] was 40 in Kern River, 87 in Dome-Tumbador, III in Tulare, and 80 in Sec. 26. Nominal steam quality was about 50% by mass at the wellhead. Tables 1 through 4 in Ref. 19 reveal other details.

Improved Injection Profile.

Using krypton and sodium iodide as vapor and liquid tracers, respectively, in steam-foam injector 68BW, Chevron25 confirmed that steam injection profile was improved by foam and that this improvement disappeared rapidly when surfactant injection was stopped. A longer-lasting profile improvement was reported for another injector in MWSS.26 Although not measured directly, steam injection profile was vastly improved in another layered reservoir of comparable gross thickness (Mobil's Tulare), simply because no other plausible mechanism could account for the significant oil response.

For the single and/or partial injection intervals, such as in Shell's Mecca and Bishop and Unocal's Dome-Tumbador, the profile improvement is difficult to quantify and perhaps less important; steam flow in the reservoir is determined more by areal heterogeneities than the injector itself. However, even in a thin sand package (8 m), such as the Guadelupe field, a redistribution of foam injection profiles was noted in four wells.13

This content is only available via PDF.
You can access this article if you purchase or spend a download.