I-

.._

-

-

Tek!n~caPetrolleum Se'Nioo·s ltd: ..

s.u~te.2:5001, S3iQ ~ ,B·thAvenue S.. W. Cal!gl,ary', A~berta.
T2P 358,

June 2001

Enhanced Oil Recovery Course Contents
Chapter I Enhanced recovery methods and screening guildelines 1. Introduction 2. Enhanced Recovery methods 3. Residual Oil Saturation 4. Screening Guildelines

II

1. 2.

Chemical EOR Processes Introduction Mobility Control Processes 2.1 Polymer flooding 2.2 Foam flooding 3. Low-1FT Processes 3.1 Surfactant Flooding 3.2 Alkaline Floods Miscible EOR Processes 1. Introduction 2. Background 3. Phase Behavior -> P-T and P-X Diagrams 4. Terminology 4.1 First Contact Miscibility 4.2 Multiple Contact Misciblity 5. Misciblity Mechanism 5.1 Gas Condensing Mechanism 5.2 Gas Vaporizing mechanism 5.3 Combination of Condensing I Vaproizing Phase Behavior 6. Misciblity Design - Hydrocarbon and C02 6.1 Correlations 6.2 Equation of State 6.3 Laboratory Techniques 6.4 Miscibility Criteria 6.5 Miscibility Equivalence 7. Factors Affecting Displacement Behavior 7.1 Mobility Ratio 7.2 Fluid Dispersion 7.3. Flow Regimes 7.4 Viscous Fingering 7.5 Sweeping Efficiency

III

8. Improvement Sweep Efficiency 8.1 Gravity Stabilized Floods 8.2 Water Injection to Improve Mobility Ratio 8.3 Foam 9. Immiscible Flooding 10. Project Development Considerations 10.1 Simulations 10.2 Economic 10.3 Implementation and Monitoring IV 1. 2. 3. 4. Thermal methods and Application Introduction Hot Water floods Steam floods Combustion

Appendix I Glossary II Details of EOR Projects

Chapter I Enhanced recovery methods and screening guildelines
1. Introduction 1.1 Primary Oil Recovery 1.2 Secondary Oil Recovery 1.3 Efficiency of Conventional Recovery Methods Enhanced Recovery methods 2.1 EaR Principles 2.2 EaR Oil Recovery Factor Residual Oil Saturation EaR Screening Guildelines

2.

3. 4.

1.

Introduction

Oil is found in underground rock formations called reservoirs. The oil resides together with water, and sometimes gas, in very small holes (pore spaces) and fractures. The size, shape, and degree of interconnection of the pores vary considerably from place to place in an individual reservoir. Thus, the anatomy of a reservoir is complex, both micro- and macroscopically. A complete, detailed quantitative description of a reservoir is never possible. The detailed data which can be obtained from wells represents only an infinitesimal fraction of the reservoir volume, even under the best of circumstances. Properties of crude oil and formation water in different parts of an individual reservoir generally vary only slightly, although there are notable exceptions. For different reservoirs, crude properties cover a wide spectrum of differences. Some are thinner than water, while others are thicker than cold molasses. Generally, crude oils are lighter than water. The water in different reservoirs also varies, in salinity and other mineral content. Because of the existence of a wide range of properties of both rock and fluids, reservoirs act differently and must be treated individually.

1.1

Primary Oil Recovery

After discovery, most oil reservoirs undergo a period of production called "primary oil recovery". Primary oil recovery used natural reservoir energy to drive the oil through the complex pore network to producing wells. The driving energy may be derived from one or more of the following: gas that evolved from solution out of the oil; expansion of free gas; influx of natural water; or gravity force. The primary recovery efficiency is generally low, varied from 5 to 20 % OOIP, when gas is the drive agent; much higher recoveries are associated with water drives. Reservoirs with effective gravity drainage can also have higher recoveries. Eventually, the natural drive energy is dissipated. When this occurs, energy must be added to the reservoir to produce any additional oil.

1.2

Secondary Oil Recovery

Secondary oil recovery involved the introduction of energy into a reservoir by injecting gas or water under pressure where the gas is immiscible with oil. Separate wells are usually used for injection and production. The technique was called "pressure

maintenance". The added energy stimulates the movement of oil, providing additional recovery at increased rates. Today, limited use is being made of gas injection because of its low oil displacement effectiveness and the need for gas supplies in the market. When gravity drainage is effective, pressure maintenance by gas injection is effective. Waterflooding is the principal secondary recovery method and accounts for about half the current U.S. daily oil production.

1.3

Efficiency of Conventional Recovery Methods

Ultimately, conventional primary and secondary recovery are expected to produce about 20 to 40 % of the original oil discovered. The recovery from individual reservoirs can range from as low as 5 percent to as high as 80 percent of original oil in place (OOIP). This broad range of recovery efficiency is a result of the variations in the properties of the specific rock and fluids involved from reservoir to reservoir, as well as the kind and level of energy that drives the oil to producing wells, where it is captured.

2.

Enhanced Recovery methods

The term "enhanced oil recovery" refers to any method used to recover more oil from a reservoir than would not be obtained by primary recovery. Since the early 1950's, a significant amount of laboratory research and field testing has been undertaken, and some of the resulting findings have been developed on a commercial scale.

2.1

EOR Principles

The intent of enhanced recovery methods is to • improve sweep efficiency by reducing the mobility ratio between injected and inplace fluids, • eliminate or reduce the capillary and interfacial forces and thus improve displacement efficiency, and • act on both phenomena simultaneously. The basic principles of the most promising EOR methods used are given in the following Table 1. Table 1. Methods of Enhanced Recovery Methods Used Chemical 1. Mobility Control Processes methods Polymer-augmented waterflooding; Foam-COraugmented waterflooding; immiscible CO2 displacement. 2. Low-1FTProcess surfactant flooding; alkaline flooding; Miscible fluid displacement using Miscible CO2 nitrogen, alcohol, LPG or rich methods gas, dry gas. Thermal

Basic Principle Improvement of sweep efficiency; improvement of displacement efficiency.

Improvement of displacement efficiency.

Cyclic steam injection; steam drive; Improvement of both sweep

methods

in situ combustion.

efficiency and displacement efficiency.

Other processes, such as bacterial activity, electrical heating of the reservoir, and so on have been proposed, but their potential for adding to proved oil reserves must be demonstrated. Figures 1 to 5 show schematic different EOR processes being conducted. A briefly introduction for each process is given as follows. 1. Chemical methods of enhanced oil recovery are characterized by the addition of chemicals to water in order to reduce the mobility of displacing agent and/or to lowering the interfacial tension. Polymer flooding, using polyacrylamides or polysaccarides, is conceptually simple and inexpensive, and its commercial use is increasing despite the fact that it raises potential production by only small increments. Surfactant flooding is complex, requiring detailed laboratory testing to support field project design. It is also expensive and is used in few large scale projects. Alkaline flooding has been used only in those reservoirs containing specific types of high-acid-number crude oils. Miscible methods have their greatest potential for enhanced recovery of lowviscosity oils. These processes are mainly in reducing the 1FT to improve displacement efficiency. Among these methods, hydrocarbon gas and CO2 miscible flooding on a large scale is expected to make the greatest contribution to miscible enhanced recovery in the future. Thermal methods are for oil gravity less than 25 degree. These processes provide a driving force and add energy (heat) to the reservoir to reduce oil viscosity and/or vaporize the oil. This makes the oil more mobile, so that it can be more effectively driven to producing wells. Steam injection has been commercially applied in California since the early 1960's and is the most advanced of all EOR methods in terms of field experience. The performance of steam injection can be estimated with less uncertainty than other methods. In situ combustion is normally applied to reservoirs containing low gravity oil, but has been field tested under a wide variety of reservoir conditions. Only a few projects have proven economical enough to advance to a commercial scale. To date, in situ combustion has been most effective for the recovery of viscous oils in moderately thick reservoirs (National Petroleum Council, 1984).

2.

3.

2.2

EOR Oil Recovery Factor

The oil reserves obtained as a result of EOR methods in addition to the primary or conventional reserves may be expressed as the percentage of original oil in place (OOIP). To estimate how much EOR methods can add to oil reserves, the recovery potential of the reservoir has to be known. This is defined by the reservoir's characteristics and prior recovery mechanism.

This solubility depends on pressure level. The flow behavior in a given unit volume of rock is determined by the relative permeability characteristics of the rock. for oil reservoirs where ER 25-40 percent of OOIP obtained by primary or conventional methods. the selection of a certain process aimed at enhancing oil recovery must be made only after detailed investigations of the pertinent oil reservoir data (especially residual oil saturation). possibly attained by the respective recovery mechanism. The oil and water are said to be immiscible. two phases form . = 3. Similarly. oil and natural gas also are immiscible. water permeability decreases to zero at a finite . the relative permeability endpoint residual oil saturation to distinguish it from the material balance or average residual oil saturation left in a reservoir after waterflooding. slightly fractured or heavy oil reservoirs 10-25% Oil reservoirs produced mainly by solution gas drive 25-40% Oil reservoirs producing under partial water drive. Similarly. of OOIP. This trapped oil is called "oil droplet" which is retained in the reservoir by the interfacial tension and viscous forces. gas injection. Residual Oil Saturation Oil and water do not mix. as follows: Rfinal 5-10% Tight oil reservoirs. This limiting saturation is called the residual oil saturation or. although natural gas has some solubility in oil. perhaps more precisely. laboratory tests. For instance. It is this most elusive oil which is the target for "enhanced oil recovery" projects. However. as for example in the displacement of oil by water. and as soon as the solubility limit is exceeded. If these two fluids are poured into a vessel and allowed to settle.a gas phase and a liquid phase. separated by a sharp interface. These values can be used for a quick evaluation of the improved oil reserves and for the selection of methods that should be applied. separated by an interface. the implementation of the EOR methods would increase these values to 35-55 percent of OOIP. two distinct liquid phases are apparent. percentage of OOIP. but oil permeability reduces to zero and oil stops flowing while the oil saturation is finite. the ultimate oil recovery factor of individual reservoirs under primary and/or conventional recovery methods may range from 5 percent of OOIP for the poorest reservoir characteristics for viscous oil. some of the oil being trapped in the pores as shown in Figure 6.For instance. or gravity drainage 40-55% Oil reservoirs produced by conventional waterflood A possible estimation of the additional oil reserves. Oil relative permeability decreases with decreasing oil saturation. When two immiscible phases flow simultaneously in a porous medium. To achieve this desideratum the oil reservoirs are classified by several models according to the average of the ultimate oil recovery Rfinalexpressed as a percentage . that could be recovered as the result of EOR processes is shown in Table 2. to as high as 55 or 60 percent of OOIP for the best reservoir characteristics or for light oil. and field pilots.

"EOR Screening Criteria Revisited. . In summary. Here the residual oil saturation is plotted vs. "Scraping the Barrel. J.. residual oil saturation is insensitive to capillary number. a. this causes earlier water breakthrough in the oil-wet rock and requires that more water be injected to achieve a given reduction in oil saturation than in water-wet rocks. Dafter. EOR Screening Guildelines Reference 1 provides the most recent literature review on EOR screening criteria. reservoir fluid properties . J. Other factors being equal. Taber. It can have a pronounced effect on immiscible displacement. similar behavior is found when a gas immiscibly displaces oil. oil viscosity. 4. reservoir geology . reservoir conditions . factors used in the screening are as following.by several orders of magnitude or more. and the shape of the relative permeability curves between the endpoints determines how much fluid must be injected to reduce the oil saturation in the swept rock to a given level.rock type and depth and permeability and porosity References 1. et ai. 1996. relative permeabilities for oil-wet rocks are more unfavorable for waterflooding than are relative permeabilities for water-wet rocks. Capillary number is an approximate measure of the ratio of viscous to capillary forces. Craig defined wettability as the tendency of one fluid to spread on or to adhere to a solid surface in the presence of other immiscible fluids. is required to achieve a significant reduction in waterflood residual oil saturation. Table 3 was copied from reference 2 for a quick checking usage. 2. Later in the course note." SPE/DOE 35385. Oil and water both flow at saturation between the two endpoints. a more detail guildelines will be discussed for each process. 1980. Over ranges of velocity. presented SPE/DOE Tenth Symposium on Improved Oil Recovery.temperature and pressure b. The effect of oil/water 1FTon residual oil saturation is illustrated in Figure 7. the product of Darcy velocity and oil viscosity divided by 1FT. Although the discussion of this section centered on waterflooding. Ray. capillary number. Figure 7 shows that a drastic reduction in oil/water 1FT. The Worldwide Potential for EOR" Financial Times management Reports. For example. In addition to being affected by rock pore structure. and oil/water 1FT found in conventional waterflooding.value of water saturation called the irreducible water saturation. residual oil saturation and relative permeability are affected by rock wettability and by interfacial tension (1FT) at the oil/water interface.oil viscosity and density and formation water salinity c.

+15% 40%-55% +10% 35%-50% +15% 40%-55% +12% 37%-52% +15% 40%-55% 40-55% ADDmONAL OIL.LING ALTERNATING WATER-GAS INJECTION +30% 40%-55% +20% 30%-45% +15% 40%-55% +10% 35%-50% +15% 40%-55% +2-4% +10% 50%-65% CO2 INJECTION +10% 50%-65% - - +7% +5% +5% IMPROVED CONVENTIONAL ME1HODS WATERFLOODING AFTER GAS INJECTION WATER INJECTION IN SECONDARY GAS CAP PRESSURE PULSING +3-5% +5% +3% +5% +5% GAS INJECTION AND GRAVITATIONAL DRAINAGE (ATIIC OIL) CROSS-FLOODING - . on. +35% 40%-45% +25% 30%-35% %OOIP E_ +30% 40%-55% +20% 30%-45% +25% 35%-50% +10% 50%-65% STEAM INJECTION POLYMER INJECTION DRY ANDRICH GAS ALCOHOL.Table 2 Conventional ER BY PRIMARY AND CONVENTIONAL RECOVERY ADDmONAL OIL. R %OOIP Ear".. DISPLACEMENT INFILL DRll.. % OF OOIP TO FINAL RECOVERY EOR ME1HODS IN SIT1J COMBUSTION and EOR Method Effciency (ER) 5-10% ADDmONAL 10-25% ADDmONAL 25-40% ADDmONAL OIL. LPG - +8% 48-63% SOLVENt INJECTION +25% 35%-50% - SURFACfANT FLOODING IMMISCIBLE DISPLACEMENT MISCIBLE . %OOIP ERfiuJ on... % OOIP Ear".

.Q z EQ) ~ Q) .... c\1 .. V a.~ .~CO0 ...~ Il')'-.::: ~ ..s ...Q CI).E c Q) Cti ..:! s A E 8 ..'0 as W s: 15 ~ oco~a..... "0 o0 C) &l .a... ld ....... I oi:? ::0 0 _Q) ~ ~ ~ .. ~ o ~ 0..2 I ~ 8.. ."O~ ~ I: a.. VI ld . 8 ~ VI o .t..~ ~ -0 0.. :: 0 :E! 0 -0 CO. I:-C Q) Q) en _ . "0 0 a.. VI 0.Q) I: CO s: z =0 i 2 = o o o A === NOll') 000 000 c. .. Nc)"c -g ..~{g.. ~Lri"N VI AlAI ~ N V ~ o N V o ~ VI "0 W' -I Q. Q)U._ . ~~ ZO ...

. ..:: I!? g -iU :::> c: :0)...._... (\J(\J(\J ..... ~ c....._. . It) (\J Bv g VI c.. .....OlOlOl (..:: 0 (. .:: Cl ::l • -§.9 c: c: Q) c: c: ~ C'Il C/) C'Il C/) _= :t::Ln _ V c?1 020 C'Ilo _= 020 _ C'Ilo :!:::It) ZoO t.. (\Jet) COOO "".....)(\J co (..c: &.. c: c: o ~. b o (\J V g (\J V VVVVV .. Q) . V ~I (\J cis > ... .......0 (\J E '0 o (\J A E (5 AI AI 8 . 0_ - .::::E 0. ......). I C'Il 00 Q) >-J2 c: o Q) 0- c: 0.co 0-:....) (..._. ....I en Cl 0 el!?· o C'Il It) (\J 0 -0 C'Il w I!? "0 Q) (\J Q) oct ~C:Q. U A &Cij(.. ~ c:c:c: v888 .. t. et) -J • • QC'IlC'll Q.. g> ~ °iii zE:....."0 '0 E ~ A '0 E o (\J A :......... ...co V 0_ ::l 00 Z b o (\J ~~~~~ OlOlOlOlOl ~~. ......)"0 Z '0 .... c: is 02 Cl"O Q.. --VI VI VI 8 (\J c.... o A o~_ c:Q) -:=2 . v ~~gfG~ ."E J: 8 tr ~ tlj ~ w > a:: > II~ I Q) '0 u.."COOl c. ._..

Q) a.. .OC\lLt) . / 0 00 00 1010 o Q. -. 0 J: I- a:: 0 10 C\I VI I :c as c. ....:= /I. as 0 ..>. 0'-10 0)"0 CI) C Q) s Q) Cl en O ~ J: I- j ~I :c Q) e :::J C ~E 00 ..::::..5 o :i: ~ ZC 0.... ..t::..... C\I V ooo~o ·c VI VI VI VI VI " " "0 z 0 .go: ('t) O:::CQ...J C/) Q) C ~ . ~ :::i <C W i ~I ..c "0 '0 asO ...._ 0'~ CI) o 010 i5 0. AI 3: :c as 0 ..co S il) S il) 0 0 0 ..Q) 1: c !B . ::E W.. a..2l.. o ·c c "0 z :.)(··:::Joc Q):=.:::: C\I /I. ~et5oJ!!" :::J CI) 0 C :::J.. .....Q)1O :i: ~ CI)_.. as C/) as '<:t cb cb . 0 Q) E Q) . as I . 0._ ..... J: z 0 I- o .Q) o .. C\I " 8 V CO I 0. 8 C\I AI ('t) I 0....t::.. => C_ C as Q).I 'CI '0 as ... 8 C\I o VI > 10 0 a: 10 C W § 0..... 0 as J!! ~~" co 0 10 ......_ ..t::..... 0 f}JQ)... oCOCJ)CJ) c CJ)CJ)CJ)CJ)CJ) a..C °co··CI)cE:C ... :..._......2 C • Q) := := o 0) C...:= 0 0 0 10 V v .:::: 10 ..C\IC\IC\I . <C. "O. E8 ~ 0... o :i: Q) CI) ~ ~ . . /I. ... .. w >a:: w > Z <C W W W W !i~ I ~ 11 :::Ju o as >LL -Q) >00 CI) :::J I ~ Q) oe ~~ s: ~ 0) fii ..( W WC\I 0 E ·c o Z 0 :.. as I o I ~I c . E 0:0 "0 co !. ~~ o Q) '2 "0 .0 Q) as 0 zE...~ ..........._... C\I ... .._.... o as Q)..t::..._.. ...J"O L1....:= 00 00 00 /I. :==0 ~ .:=.:::: ~ Q) . Q.t::..... C\lLt) . C CI) o 0 "0 C as as e o V .... 0 .:=........g ....

~ .. No VlO o ~ V . CIS CIS . 1. Q) z Z ° ~ ° § V E I ~ ~ o ~ ~ -~C")I. " CIS cb C» CIS .. .0 § " " o~ 1. o ~ V ~ i 0CIS .I: o ~ o 1..J C) ~§ =E C -(1).I: o o .0..~ c: . CIS ... >a: w tu :!: > ~. (ij 02 0C: (. " 0 "C a. 0 ~.0 Q) ..9 CIS en c: .O CIS :c CIS co Q) I- .') "5 c.0 ::l o~ oS 0 Q) c 0C: ::l c: Q o J: en u: gj .

A Current Appraisal".Meldau. Energy Communications 4(1) 83-106 (1978). Note: figures refer to "screentnq" criteria for selection of EOR candidates. Oklahoma. Surguchev and E.Poettmann. 1977. Society of Petroleum Engineers paper SPE 7546.Halimov.H..S. Socity of Petroleum Engineers paper SPE 5821.M. Cities Service Oil Co. page 36. October 1976. SPE report on Symposium on Improved Oil Recovery.1973.L. Geffen. Getty Oil. Oil & Gas Journal. U. "Enhanced Oil Recovery Potential in the United States". World Oil. Ponca City. "Current Steamflood Field Experience". pages 124-128) and part two (May 1964. Aug. Erofeeve. a-7 Chieh Chu. a-5 Lewin & Assiciates. part one (April 1964. March 1976. Today and Tomorrow". a-s F. a-6 H. a-3 National Petroleum Council.S. "Research and Development in Enhanced Oil Recovery. 1978. Journal of Petroluem Technology. Oklahoma. M. pages 1113-1128. Pergamon. Continental Oil Company. "Enhanced Oil Recovery is an important way t increase oil reserves'" The Future Supply of Nature-Made Petroleum & Gas. page 11-8. ·Polymer Flooding Technology .Yersterday. a-2 Harold EGilliiand.Chang. .REFERENCES FOR TABLE 1 a-1 N. a-10 Office of Technology Assessment. page 24. "Oil Production to Expect from Known Technology". pages 66-76. pages 321-337.M. Part I". January 1978. pages 95-98). May 7. Dec. a-s T.M.Farouq Ali and Robert F. a-4 S. ·A Study on Fireflood Field Projects". Congress. 1978. -Enhanced Oil Recovery". 1976.L. ·In situ Combustion.

. ~~----------~------------~ ii • ~ . o o :z: ~ IU D Ue :I ~ a: o IU > U IU I I I ~:I . c 0 N U -ucn ~ o N~ '0 ... ... -• ~ IU a: o > IlL o II: o • ~ " ..s .! :I -" :.. Q) .- :I U :I ~ i • o• U. c e . " .!. • ~ ~ '0 ~ (U '0 C (U z c i U:I .- . • u c: IlL E~ • 0 0 . ...s . E ~ U ....t e :I IlL c ~ a: .. e Q) e. . ~'i E 0 ..! o• e.5 u• ~...! u c o uin. • NG.. • -• • " ~ ~ E" u 0 'i • ~ ~ • u •• iE ! V" -• • _g • 0 C ...!..... -• :5 ... ~ c D U • " c 0 u E ! C •• ~ .

..IDS OIL AND WATER .Figure 1 C02 Miscible Flooding Process C02 AND WATER OIL AND WATER . D • • II Drive water zone Water/polymer zone Surfactant slug zone Oil and water zone . ~ I • • Injected ~ ~ and water zone D Oil and water zone -crude oil miscible zone Figure 2 Surfactant Flooding Process INJECTION FLU.

water and steam zone Steam zone Oil and water zone ( for weeks or months) D • Figure 4 Steam Drive Process (Steam Flood) STEAM OIL AND WATER . • •.in phase) Huff ( injection phase) STEAM X OIL AND WATER Puff ( production phase) (for days or weeks) • (for days) Hot oil..Figure 3 Cyclic Steam Stimulation Process ~ ( Shut . Hot water zone Steam zone Oil and water zone D .

~ Air and vaporized water zone Combustion zone Steamzone Simplified cross-section of oil field structure III D • Hot water zone Oil and water zone . Injected air and water zone • .Figure 5 In-situ Combustion Process OIL AND WATER AIR AND WATER COMBUSTION GASES • J.

Figure 6 Residual Oil .Oil Droplet Oil Residual Oil Water .

J .. FOSTER LEFEBVRE DU PREY [:::::1 BARDON <.. z >a: ::::> 10-5 « .R... Oi I Saturation .V..J n. Capi Ilary Number versus Residual (Bardon and Longeron) P. b ~ ::t 10-4 CD ~ w a: . « u 10-6 10 20 30 - 40 % 50 60 RESIDUAL OIL SATURATION.V/'! W.

which means that the oil spontaneously spreads over a flat water/gas interface. Finally we :perform a series of gravity drainage experiments in sand ~olumns at high and low oil/water surface tensions. We show that although most oils will spread over water. In both experiments virtually all the oil can drain from the top of the column. Yizika. Blunt. Drainage through .few molecules thick. [1993]. since the oil ganglia will drain undergravity by swelling the oil films. U. which is much lower than the residual oil saturation in the presence of water alone. We compute the equilibrium thickness of a flat \Ioillayer between water and gas considering both van der Waals lind structural forces. SPE Members COpyrIght 1994. only a molecular oil film is present. Telex 163245 SPEUT. Kantzas et al. WrHe I. it has been known that gravity drainage in water wet rock can lead to a high oil recovery. 1988.0. Since the work of Dumore and Schols [1974]. and Dengen Zhou.S. Contents 01 the paper • • ~ted. Further studies on sandstone cores. It was hypothesized by Oren [1992]. the low 'surface tension case gives a lower final recovery. The abstrect should contain conspicuous acknowledgment Ubrarian. Both spreading ~d non-spreading oils can form thin films. Stanford U. illustrationS may not be copied. Box 833836. The film thickness.predicted rate agrees well with the measured values.oil is recovered. We then present results from drainage . they will eventually form a film that is only a . It was suspected that the high oil recovery was due to drainage through thin layers of oil that lie between the water and the gas in the pore space. Papers presented at SPE meellngs are sUbJect to publlcallon review by Editorial Committees 01the SocIety to an abstract of not more than 300 words. Air is ttmtroduced into the top of the column containing waterflood ~ . Society 01 Petroleum Engineers. is determined by the capillary pressure and by ~\vhetheror not the film was formed from thinning a thicker layer for by encroaching over a hitherto clean interface. stabilized by capillary forces. Richardson. it was not clear under what circumstances drainage through an oil layer would occur. OklahOma. 1988a. 371 .ted by air bubbles along its length.O. whereas for the high surface tension fluids..experiments in capillary tubes with a square cross-section. However. bead packs and sand columns [C/uztzis et al. Thts paper was prepared lor presentation at the SPEJDOE Ninth Symposium on Improved Oil Recovery held In Tulsa.. as presented. 1993] and Kalaydjian [1993] that systems with a positive spreading coefficient... Hs 01 Petroleum Engineers. I'BSTRACT We present a theoretical and experimental treatment of three lphase flow in water wet porous media from the molecular level :10the core scale.SPE/DOE 27816 ( What Determines Residual Oil Saturation in Three-Phase Flow? M.H. this film would be far too slow to account for the recoveries seen ~ xperimentally. Thts paper was selected lor presentation by an SPE Program' Committee loIlowing review 01 Inlormation contained In an abstract submitted by the aUlhor(s).A. However. have not been reviewed by the SocIety any position 01the SocIety 01 Petroleum Engineers. [Vizika. P. We predict not only that eventually all the . The gravitational Fving force in the experiment allows oil ganglia to drain by 'pwelling the oil film. but that we can calculate the drainage rate. This film drainage has been observed directly in two dimensional etched glass micromodels by Kantzas et al. [1992]. SPE. would experience film drainage and high recoveries. However.J. 01 Petroleum Engineers and are subject to correction by the aUlhor(s). U.1-5 nm. or members. TX 75083-3836. Dumore and Schols [1974] saw the same t. INTRODUCTION The displacement of oil by gas in the presence of water is an important recovery process in oil fields and in the clean up of contaminants spilled below ground. Inc. 17-20 April 1994.residual oil. In both cases oil is displaced by air. Yizika. which is in the range . whereas non-spreading systems would see lower recoveries.A. since a significant amount of oil remains in the column in the thick oil layers. with residual oil saturations of a few percent in the presence of both gas and water.. Kalaydjian [1992] and Soli et al. The residual oil saturation in the presence of water and gas can be almost zero under favorable circumstances. The . [1988b].S. Oren et al. D. Fenwick. does not necassarlly reflect olfocers. The tube has water wetting the comers and contains oil ganglia '~para. Oren and Pinczewski [1992]. The material. At the end of the low surface tension experiment " there is a thick oil layer (several microns across) between water and gas. Permission to copy ISrestricted 01 where and by whOm the paper IS presented. 1993] confirmed these results.

which is controlled by capillary and intermolecular forces. 1989]. r. Table 5 of [Muskat. the system can be in equilibrium with a macroscopic blob of oil if < 0 where = Y'gw. When tis of the order of molecular width. The effective (5) spreading where rgw. (b) If the oil forms a thin fiim between the gas and water. [1988b] that the recovery could be determined by the stability of the film. It was also implied by Kantzas et al. If Cs> O. Rowlinson and Widom. as shown in Figure 1.w where = rgo + row + Ev(t) + to(t) t = O. On a flat substrate. However.row = Cs. if there is no pressure difference across the film. (a) Three fluids in contact. rgo and row are the gas/water. we imply a loose arrangement of oil molecules at the gas/water interface. 1928. Consider the oil film illustrated in Figure 2(a).rgo . oil layers stabilized by capillary forces are formed as an oil bank moves through the system. or around 0. the contact line between the three phases is unstable and the oil spreads..1 urn. of a film of thickness tand area A is F=A{rgo+row)+E. ~ . 104.row = Cs The equilibrium thickness of the film is given by i}F/iJt or: (3) = o. starting at the molecular scale.rgo .(O) = rgw . (4) c: c: P = O(t) = -oEv/ot H a large drop of oil.by r. representing the free energy per unit area of the oil fiim.. F = ggwA. If rgw > rgo + row the oil wiil spontaneously spread between the gas and water. Gibbs. such that the size is greater than the range of intermolecular forces. F. the interaction is described by van der Waals forces. Ev may have significant contributions from structural forces. This paper investigates the fundamental mechanisms of oil recovery in three phase flow.rgo . is when the capillary pressure equals the disjoining pressure. (2) the thickness of oil layers during drainage. the oil eventually forms a thin layer of approximately molecular thickness. the effective spreading coefficient is zero or negative [Adamson. when t 0 (Figure 2(b» the free energy F rgwA. The water/gas interface r. The principal issues are: (1) the thickness and stability of thin oil films controlled by intermolecular forces. The equilibrium film thickness. E" is the interaction energy (principally van der Waals) of the film and P is the pressure difference between gas and water. = Cs = rgw .row· Cs> 0 indicates an oil that spreads.(t)A+PAt. Most hydrocarbons and crude oil mixtures (see p. In experiments in micromodels and capillary tubes. and Viz i k a [1993] found higher recoveries if the spreading coefficient was positive. and thick oil layers can exist for both spreading and non-spreading oils.rgo - row (1) r.2 WHAT DETERMINES RESIDUAL OIL SATURATION IN THREE PHASE FLOW? SPE 27816 residual oil saturation (approximately 3%) for both a spreading and a non-spreading system.r~ . containing this oil tilm has a much lowered effective surface tension for which. (3) the flow rate in these layers.row and Y'gw is an effective surface tension.w is the ~ energy per unit area of the~as/water interface separated by an oil film of thickness t. where rgw is the gas/water surface tension. (4) the final oil recovery. the initial spreading of the drop depends on the spreading coefficient: where O(t) is called the disjoining pressure. Hence: = = Rgure 1. gas/oil and oil/water surface tensions respectively. rather than the spreading coefficient alone. The free energy of the film of thickness t and area A is DOES OIL SPREAD ON WATER? (a) (b) F=A(rgo+row)+Ev(t)A+PAt (2) Gas Gas Solid surface Solid surface where Ev is the interaction energy of the film per unit area and P is the imposed capillary pressure between the gas and water.w = rgw when coefficient c: is c: = Ev(t)+tO(t) (6) 372 . By definition. is placed on water. Ev(t = 0) = rgw . 1990. whereas Oren et at. (b) When the thickness t = O. (a) The free energy. For oil films much thicker than the molecular width ao. in core floods. This defines E. no oil films are observed for non-spreading systems. For t < QQ. We may derine an effective gas/water surface tension in the presence of an oil film. teq. [1992] • Oren and Pinczewski [1992]. Gas (a) t f Oil Water Gas (b)---Water Figure 2. 1949]) do have a positive spreading coefficient.w of/oA. Define a spreading coefficient Cs = rgw .

5 226. The dielectric response functions are calculated using the Lorentz harmonic oscillator equation: 40.9 173.75 13.3 0. where the interaction across the film is determined solely by the dissipation energy of an electromagnetic wave propagated through the system.2 197.rSPE 27816 ) .0 95. where the constant gj describes the bandwidth of the adsorption frequency Wj.380 0. Table 1 lists the data used in the calculations presented here.38 126. c is the speed of light in a vacuum and T is the absolute temperature.1235 0.47 8.• 1961. In this regime.271 (1014 radls) 0.8 17. Lifshitz.7 0.0 and El. ) = $1 - $1 + PE1/E2 .0923 0.819 sl =~£I/E2 -1+ p2.3. $1 + P PE1/E2 n-eetane (8) infrared ultraviolet 5. Equation 7 is valid in a continuum.05 n=O where: f~~Er2UI J 1 +/2)p2dp (7) infrared ultraviolet 12 = ( 1 ~12~32 exp(2 P~n c t. the following expression for the disjoining pressure between two semiinfinite materials I and 3 separated by medium 2 of thickness t.851 7.2. materials 1 and 3 are water and air respectively and medium 2 is oil.!.7 The first sum represents the Debye microwave relaxation with resonance frequency (J)r.152 0. Water Cr microwave Wj to. K= Cs{I+4n ) 2 ao 373 (11) .319 1.1466 0. E2. The summation is over integral values of n and the prime on the summation means that the first term in the sum should be multiplied by 1/2. Thus . and assuming t » ll().fE2) _1)-1 s3-P s3+P $3 74. Table of constants used to compute the dielectric response. ~12 -. E3 are the dielectric response as functions of the imaginary frequency i~.09978 0.079 0.023 0.057 0.0 fj gj 1.1304 0.156 0.1489 0.425 0.5 23. DARRYL FENWICK AND DENGEN ZHOU 3 ENERGY 'Using quantum field theory.019 0. MARTIN BLUNT. ~32=--. When the film thickness is close to a molecular size.925 ~. e = 1 for air.735 0.143 0.54 187. which we set to 300 K in our calculations.1 151.54 128 136 147 160 176 196 221 252 288 335 0.08132 0.3 21.1 a-pentane infrared ultraviolet 5. structural forces become important.1 38. the disjoining pressure is modeled empirically using an equation proposed by Israelachvili [1991]: n(t) = K cos(21rt/ao) e-t/ao (10) The constant K can be determined by knowing that the interaction energy at t = 0 is equal to the spreading coefficient.0392 0. For our case.228 0.THE INTERACTION .3 281.: u32 = $3 - PE3/E2 + PE3/E2 tetradecane infrared ultraviolet 5.06177 0.36 11. The second sum accounts for the infrared.398 3.).154 0.577 0. The data for water and tetradecane comes from Parseglan (1981) and that for pentane and octane from Hough (1980).8 (1014 radls) 0. where ~n =2nkT liz (n = 1. has been derived [Dzyaloshinskii et al.0928 0.1178 0. (1011 radls) 1. The functions £1 and E2are computed from the adsorption spectra of water and oil respectively..429 0. 1955]: n(t)=~ Table 1.4 31.2 13. k is Boltzmann's constant and Iz is Planck's constant. o o o o o = $1- p. visible and ultraviolet adsorption spectra.: u12 s3 =~E3/£2 _1+p2.038 6.048 1.9 26.. -.1 45.9 32..54 186.03960 5.025 0.

..0 10 10" .. Q) ~ 0 ..:.-J._.....4 nm and Cs = 0. __ -'--_L--..... as ~ Q) (f) (f) 6000 If) If) :::> I!! .2 mNm-l.. ...:...4 WHAT DETERMINES RESIDUAL OIL SATURATION IN THREE PHASE FLOW? SPE 27816 Equations 7 and 10 for the disjoining pressure are not additive [lsraelachvili.. with an imposed capillary pressure difference......__. The maximum value of O(t) (at 1 : is simply K..__.. Equation 7 applies when 1 » ao.. which is mucn higher than typical capillary pressures in a porous medium. Therefore._. Figure 4 shows the disjoining pressure for n-octane between water and air accounting for structural forces.. which means that the film is unstable and tends to spread further. ...6 ·0. .. . much larger thar maximum values of 0(1) in Figure 3..... c: c: a............:""' ..J. .......... · ....... Structural effects give an oscillatory force on the film .__. as :::J Q) "(f) (f) IIIIIII ' ::::::: : : :: ::: : : : : : : :... .....--L-L. · ........0 Q) ~ 0 1111111111~\~_.. i~ g>-O.._... considering van der Waals forces alone. ::::: ::::: :::: c....... . -2000 . ...E!::.':"':'........... · . ..._ ._. · ._ c: 0 0> c:: .:. Structural forces become significant at around 4 nm. 2000 thickness (nm) a. : .... 1 10 thickness (nm) Figure 3. At larger separations.L..... The only exception would be for oil on a flat plane of water. Figure 3 shows the disjoining pressure as a function of film thickness for pentane.. which is Equation 4. The disjoining pressure for water/n-octane/air computed using van der Waals forces (Lifshitz Equation) and structural forces.. · .. .. I!! OJ 0 -4000 :::J 4000 c '2 '0 'iii' -6000 . .... the disjoining pressure has both negative and positive values as the fjlthickness approaches zero. the equilibrium film thickness is that for which 0(1) = P....._. .. . .. 0> 1....05 Nm-I and 10000 : : : : ::: ~ 8000 6000 r= a..L-J'-'-'.. P is in the range of 500 Pa... or approximately 10 molecular diameters.... The structural forces were calculated from Equations 10 and II with ao = 0.... between water and gas....!a _ -2000 : :: :::: ::: ............... Because of the oscillatory nature of the structural force......6 4000 2000 10000 a..' ..._... ............... octane and tetradecane... while Equation lOis valid for film thicknesses of a few molecular diameters...........!a 0 -4000 -6000 -8000 I I I I -10000 L--_---'_-IJ-JL...... 8 and 9 and the data in Table 1... ..L__. For small film thicknesses...... P....:L....... · . The disjoining pressure for water/n-alkane/air systems... RESULTS FOR THIN OIL FILMS l r = 10--4 m. · ..4 02 0..... .....~ . the disjoining pressure is negative. Lifshitz Equat... the disjoining pressure becomes positive.:.L. ... which is approximately 2 x 107 Pa.... 374 ..... .2 ~ ~. ._. .0 ... Structural Force is ~. . ...:...._. ... If we assume typical values of 0.2 -0. ..I.... .......... there is no stable film thickness in a porous medium for the hydrocarbons studied here.. If we consider oil spreading in a porous medium.. .. .... -10000 L.0 -0...8 ·1......... .. . 0..........8 0..6 0. .. 1991]... : ::: ::: : ::: ::: : : : : ::: : : : :::: ::: ~ _lh_ic_k_ne_S_s_(n_m_) lo"-t 8000 c...J. · · . .... n ... . :.4 -0. __._..... using the Lifshitz Equations 7...-_--' _ 1 10 ' 102 thickness (nm) n-octane n-pentane tetradecane Figure 4..._.. . .. .. ...~~'~.. Characteristic capillary pressures are of the order r / r where r is the surface tension and r is a representative pore radius........4 iii -0.. where an oil film tens of nanometers thick may be stable if P is extremely small (or zero)... ...

stable oil layers are possible for nonspreading oils at the trailing edge of an oil bank in sand column and core flood experiments. such as at the trailing edge of an oil bank in gas flooding. If there is a film of oil that connects the two blobs. it is stabilized by capillary forces. there will be pressure continuity in the oil phase which allows the upper ganglion to drain into the . of molecular -_ /thickness.Ia. since a film is formed from t = O. Oren et al. If the thickness of this layer is outside the range of intermolecular forces. DARRYL FENWICK AND DENGEN ZHOU 5 Therefore. which may be one or more molecular diameters across. where the gas/oil interfacial curvature is zero. 375 A molecular film provides pressure continuity for the oil phase. . The position of equilibrium then jumps to around 3. If P « K. increasing P may cause the equilibrium thickness to decrease abruptly. Moreover. The effective spreading coefficient. Systems with a positive initial spreading coefficient always form oil films. Water is considered to be the wetting phase. is negative and given by -21tC_se-1I4. Most oils spread on water and form a film. Two blobs of oil are separated by a height h. Hirasaki [1988] has shown that a stable film has P = n(l) and dfI{t)/dt <0. This explains why oil films are not observed in micromodel and capillary tube experiments when Cs < O. or t = aoln(KlP). from Equation 6. On the other hand. == P. In this case. the flow through them is minimal and they are unlikely on their own to account for the high recoveries seen in gravity drainage experiments. Thus we see that there will be several stable film thicknesses for a given P. where for P less than approximately 500 Pa. Van der Waals forces on their own do not lead to the formation of thick films (5 nm or larger) in porous media. nW(tw) represents the disjoining pressure of the water film adhered to the solid surface. the thickness decreases suddenly again. teq is the equilibrium thickness of the molecular oil film. 1988]. The pressure differences between the phases are given by the augmented Young-Laplace Equation: Pcgo = Pg -Po rgo =-+n(l) rgo (12) and Pcow=Po-Pw= row+nW(tw) row (13) where llrow and l/rgo are the total curvatures of the oiUwater and gas/oil interfaces respectively. a stable film may form for both positive and negative spreading coefficients. as it finds a new position of equilibrium at a thickness approximately one molecular diameter smaller than before. the solution with the largest value of 1 is chosen. structural forces do allow a stable film. This situation arises when a film is formed as the oil saturation decreases. On the basis of these calculations. The effective spreading coefficient may be positive or negative depending on the exact position of equilibrium. this is unlikely to be the case. previously isolated oil ganglia become connected. we ignore the oscillato~ factor in Equation 10 to find a film of thickness t where Ke . Figure 5 shows the distribution of three phases in a single pore throat of square cross-section.6 nm. stabilized by structural forces.500 Pa. as we increase the pressure beyond 500 Pa. DRAINAGE RATES IN A SINGLE THROAT Solid Figure 5. If the spreading coefficient is positive and \ P « K this will occur when cos(21ttlao) is small or t == ao/4 and the effective spreading coefficient. The thickness depends on the way in which the layer was formed and is a function of capillary pressure. the solution with the smallest value of t is chosen. If the film is formed from the thinning of a thicker layer. This means that during gas injection. If the film is being created on a previously clean interface. We will ignore the disjoining pressure in the analysis below. which may be several molecular widths. Consider the arrangement of fluid illustrated in Figure 6. A layer of oil will exist in the pore space if rgo > row. the addition of oil to the gas/water interface is energetically disfavored and no film will form. nine molecular widths. and the disjoining pressure in Equations 12 and 13 is negligible. oil is intermediate wet and the gas is non-wetting. The flow rate through these swollen layers is much greater than through a molecular film and can account for the good recovery and drainage rates observed experimentally. [1992] suggested that the : equilibrium film thicknesses of the order of 100 nm observed in air/soltroUwater systems were due to van der Waals forces. Drainage can then proceed by swelling the oil layers. If the initial spreading coefficient is negative. using Equation 10. and hence the contact angle between the gas and water phases separated by the film. the equilibrium thickness is around 4 nm. or 10 molecular widths. Since the molecular films are so thin.SPE27816 MARTIN BLUNT. The film thickness will depend on how the film is formed. At P = 1. Nonspreading systems only form oil films if the film is generated from thinning a thicker layer. This Jcan be seen in Figure 4. An oil layer in a square crevice of the porous medium. The same phenomenon has been described for wetting films of water [Hirasaki. are also dependent on the capillary pressure and displacement history.

difference between the oil and water phases is: The pressure 376 .Pg 2 Ygo/r. . Since there is very little pressure difference in the air bubble. To account for flow in the oil phase and to compute the flow rate due to film drainage. If the curvature does not vary with height. Fluid configuration in a square capillary tube at the beginning of drainage. the pressure in the upper and lower ganglia must be approximately the same and the oil can drain downwards under gravity. The gas/oil radius of curvature is approximately constant. The upper ganglion can drain if there Is pressure continuity In the oil phase.. at least. Thus the total curvature of the gas/oil interface IIr go is equal to 2fr. row is: (16) (17) air where: 110 = 2yow rgpw t (18) If we estimate the oil layer thickness by then: = (. From Equation 14 we can write: = We assume that the water and gas are stationary and in capillary equilibrium and take the water pressure at Z 0 to be zero. During drainage. The water pressure as a function of height is: = Po -Pw = Yow = Po +(hJ row +h2 +z)Pwg (20) Equation 20 is used to find row and then the thickness of the oil layer is: (21) where: (22) and X= = (14) and the gas pressure is: (15) 2yow r (23) We can write a Darcy like equation for the oil flow: (24) assuming that the gas density is negligible.1)( rgo - row) • Figure 6.fi . we must derive a more accurate expression for the oil pressure. and this will force the film thickness to be greater than zero for all values of h2 and z. Two oil ganglia are separated by a gas bubble of height h. When the oil flows under gravity. the oil layer will swell under the influence of the raised pressure in the oil phase until it is of the order of microns thick.6 WHAT DETERMINES RESIDUAL OIL SATURATION IN THREE PHASE FLOW? SPE 27816 lower one. the pressure in the oil phase will be larger than that implied by Equation 16. be a molecular oil film present Initially. and the radius of curvature. Hence Po . which means that there must. The capillary tube is square with cross-section 2r and inscribed radius r. while the water curvature becomes progressively smaller with height The oil layer thickness is zero when z = h2 0 and increases with z. then the curvature of the oil/water interface at hI must also be 2fr.

DARRYL FENWICK AND DENGEN ZHOU 7 .1 mNm-I. we controlled the height of the gas bubble.5 mNm-I. if the density difference between the oil and water phases is small.8 mNm-1 38. reaching an asymptotic value where m = Pwg = Pog when h is very large. We used a square capillary tube of side 2r 0.48 X 10-3 kgm-1s-1 ) wherefh andfo are the values offatz Low surface tension system: water and oil phases formed from a mixture of water. Q increases with h. iso-octane and iso-propanol. = (. Po = 690 kgm-3 = h and z = o respectively. The last step was to allow some oil into the tube. row= r80 = 21. This was estimated to match the experimental results . Jlo= 0. the second term in Equation 25 is always larger than the first. row= 4. where Q is the flow rate and the conductance of the oil layer is . The solid and dotted curves are our predicted times. For runs without a = 3n . The fluid properties are shown in Table 2. Densities: Pw = 1000 kgm-3. However. Table of fluid properties.49 X 10-3 kgm-Is-I fh = x+(Pw -Po)gh2 +Pwgh (30) Figure 7 shows the results for the system with a high interfacial tension (IFT). We repeated the experiment for various values of nz and h. High surface tension system: water/iso-octane/air We cannot solve Equation 25 analytically. Water was subsequently drained out of the opposite end to allow air to enter the tube. above the gas. Figure 8 shows the results from the low IFf system in the same capillary tube. By controlling the amount of water that drained out. Jlo = 0. We can find the flow rate from Equation 26: (31) where m is given by Equation 28.41 mNm-I. we may write: (27) which has the solution: Susface tensions: rgw = 72./2" -1) 3 cr3 (26) Table 2. In each experiment the volume of oil V above the gas was 0. (29) Viscosities: J1.only one parameter was used to match both curves. c is a geometrical factor that is also affected by the boundary condition at the oil/water interface and is thus influenced by both the oil and water viscosities. The tube was then placed vertically and we recorded the time for all the oil in the upper ganglion to drain. We first filled the tube with water. The drainage times are longer because the viscosities of the water and oil phases are higher than for the high IFf fluids. directly. The only unknown parameter in the equations is c. Again the agreement between experiment and theory is good.15 mm. and so to a good approximation. We conducted experiments for high and low oil/water surface tensions. The triangles are the drainage times for different values of h when there is no initial oil bank (h2 = 0) and the squares are the drainage times for the same fluids for various h and h2 4 cm. which assumes that we have film drainage with a . The agreement between experiment and theory is good. free gas/oil interface. Equation 32. h. the conductance constant for the oil film. Written as ct3.w= 10-3 kgm-1s-I.225 rnm-'.8 X 10-3 kgm-Is-I. EXPERIMENTAL RESULTS IN CAPILLARY TUBES We tested our prediction for the drainage time. Then oil was introduced into the tube at one end to form the lower oil blob. (28) Surface tensions: Densities: rgw = 36 mNm-I. Increasing h reduces the drainage time. Ygo= 20 mNm-1 Pw = 975 kgm-3. The time td taken to drain a ganglion of volume V is: td=V Q (32) = with Q given by Equations 28 to 31.fSPE 27816 ~\ ) MARTIN BLUNT.w = 2. Po = 690 kgm-3 Viscosities: J1. Substituting Equation 21 into Equation 24 yields: (25) where: m= lower oil ganglion (h2 0) this step was omitted.

Porous media contain crevices and grooves that would allow a bulk oil layer to form. A bulk oil layer could no be established and the oil was forced to drain through a mue thinner film than before. Our calulations show that the maximum oil laye~ thickness is not a few nanometers.5.. i·. Comparison of measured and predicted drainage times in a capillary tube of square cross-section for the system with a low oillwater surface tension. film provides a pathway through which the oil ganglier. water and gas in vertical eqUilibrium. water + oil gas/oil contact oil/water contac water o ~~~~~~~~~~~~~~~~~~ 5 o 10 h (cm) 15 20 Figure 8.. but around 15 urn (froIIJ Equation 19 for large z). Q Transition zone Gas. _~ drain. Therefore.Ocm measured for h2=4. Figure 9.Ocm To test what might occur in reservoir rock. !~ The drainage times for a single ganglion are of the order o~ minutes.Ocm Measured for ~-O.. .-~ --. We will attempt to estimate how much oil is left in tbl transition region above the oil bank. Predicted for h2=O. Experiments in circular tubes artificially supress the natural flow in roughness and corners expected in more realistic media.Ocm measured for h2=O. In this case drainage times were greater than a day.shows a schematic of the final fluid configuration in vertica equilibrium. A schematic graph of the configuration of oil. as predicted by th9 intermolecular force calculations. At the oil/water contact we define the oil/ware capillary pressure to be zero and similarly P ego 0 at the gas/oi contact.. or approximately I. Figtn ----.:: CD ~ c CD • 8 6 4 2 Predicted for h~. 378 . which is comparable to the width oftbe tube. Comparison of measured and predicted drainage times in a capillary tube of square cross-section for the system with a high oillwater surface tension.:: ~\ m ca '! Q c ..- :.. 3 ' Figure 7. = E E m ca f! • .. a film of oil is rapidly established between the gas and water.B WHAT DETERMINES RESIDUAL OIL SATURATION IN THREE PHASE FLOW? SPE 2781€ 10. I We performed similar experiments with the high 1FT flui~ using a capillary tube of circular cross-section. ".Ocm lCD E E CD 5 til 8 6 4 2 t c :.' in the pore space.Ocm predicted for h2=4. • 15 o ~~~~~~~~~~~~~~~~ 5 10 o h (cm) 20 Both fluid systems studied had a POSItIve spreading coefficient. This means that in field situations. oil drainage may bel rapid if there is a well connected pathway of water wet crevi.' times longer than in the square tube. It does this by swelling the oil film to form a bulk layer that occupies most of the comer of the tube. THE EQUILIBRIUM POROUS MEDIUM ARRANGEMENT OF FLUID IN A 10 . an analysis drainage under gravity in sand columns was performed. When the tube is held vertical' . • .

Regardless of the value of a. The final oil saturation was 0. J is equal to total curvature of the fluid interface. the initial oil saturation was 0. The parameter a is determined by the fluid properties. apart from a molecular film. A more comprehensive series of experiments on unconsolidated columns was performed by Kantzas et al. water.n the transition region.SPE 27816 MARTIN BLUNT. This exercise was continued until there was no additional production of either oil or water. If a < 1. Thus at a given depth we expect J(Sw + So) < J(Sw) that indicates: The larger the value of a. Substituting Equations 33 to 37 into Equation 38 gives us: z+hob"+ ( ) a< n z+~---:-- Pw-Po g (39) (Po -Pg)g where we define: a row(po -Pg) rgo(pw -Po) (40) This is a generalization of an expression first derived by Kantzas et al. The phase pressures are: . then their disjoining pressures will be negligible and can be ignored in the equation above. the low IFf system gives the lower recovery.2 (and Cs > 0) and was 0. being 1. If both oil and water form thick layers. From Equations 12 and 13.89 for the high 1FT system and 0. we find that a 3.2 m long and 2 em in diameter. we may ignore the disjoining pressures to find: h .23 when a 2. J should monotonically decrease with increases in saturation. For the low 1FT run. They found higher recoveries from controlled Equation 38 is equivalent to saying that for a connected oil phase 'go> 'ow in Figure 5. = Pego-n(t) rgo < Peow-nw(tw) row (38) SAND COLUMN EXPERIMENTS We tested our prediction that a system with a < 1 should give a lower recovery in gravity drainage than one with a > 1 by performing a series of gravity drainage experiments in unconsolidated sand columns. and this layer may retain a significant quantity of oil. [1988a]. who found high recoveries similar to ours from gravity drainage. there is some finite height. is quite general and makes no explicit assumption about the microscopic fluid ~onfiguration. The final oil saturation was 0. he where the oil phase becomes disconnected.04 and the water saturation was 0.278 and a permeability of 48 Darcys. If a < 1 the inequality holds for all values of z .8 (and Cs < 0). The sand had a porosity of 0.between the other two phases. A similar set of experiments of gravity drainage in sand columns was performed by Yizlka [1993]. Our capillary tube experiments demonstrated that this is not the case. there is a stable oil layer at all heights above the oil bank.105 and 0. however. although both values of the final oil saturation are higher than ours. DARRYL FENWICK AND DENGEN ZHOU 9 At the end of drainage. we pumped air onto the system periodically to remove the liquids accumulated at the bottom of the column. [1988b]. The columns used were cylindrical. Again lower values of a give lower recoveries. .06. I are only interested in bulk layers of oil. and then using Equations 12 and 13 to derive the inequality. We established a residual oil saturation in water before allowing any air into the system. This analysis. The oil will remain in a stable layer . bulk layers in the pore space above the oil bank. If we consider the fluid properties listed in Table 2.the oil is connected at all heights above the oil bank.11 in a system with a 4. (33) (34) (35) \ and we have allowed for a finite gas density. The J function is a property of the porous medium. If a > 1. Equation 39 gives us the condition for the oil phase to remain connected above the oil bank. we may write: (36) (37) where So and Sware the oil and water saturations. where a stable oil layer was considered a necessary prerequisite for film drainage. If we = = 379 . Starting from the second day.19. If the capillary pressures can be scaled by a capillary pressure function J. We used the same fluid systems as in the capillary tube experiments. This conclusion is different from that implied by Kantzas et al. As we expect. the final configuration of fluid implies that more oil will be left from gravity drainage of a system with a < 1. the smaller he.. [1988b].49 for the low 1FT case. with final saturations of 0. the columns were allowed to drain freely under gravity with the oil and water collected from the bottom of the column. . which means that less oil is left in stable layers above the oil bank and the final oil recovery is higher. For the high oil/water 1FT we had an initial oil saturation of 0. since more oil will remain in thick.03 for the oil and water phases respectively. For the first day of production. oil ganglia in both experiments drained by swelling the oil layer in an attempt to reach capillary equilibrium.17. However.hob Ca-I (41) . oil and gas are all connected _--.

. Society' of Petroleum Engineers. and J. where they succeeded in removing the oil bank completely from the system (and some of the transition zone). Soc. Lifshitz. Texas. Proceedings of the SPE Rocky Mountain Regional Meeting. of the SPE. A. and F. 1992. such as at the trailing edge of an oil bank. J. 1928... edited by J. SPE 24881.-C. H. Oren. 8 of International Series of Monographs on Chemistry.. CONCLUSIONS Most hydrocarbons spread on water to form a film a few nanometers across.. Proceedings of the 67th Ann. Mechanisms of capillary displacement of residual oil by gravity-assisted inert gas injection. Washington. 1993.-M. 1981. Oren. 0.. of the SPE. Green. Oklahoma. M. E. and Exhib. The effect of wettability and spreading coefficients on the recovery of waterflood residual oil by miscible gasflooding. The predictions agreed well with experimental measurements. Proceedings of the 67th Ann. 1989. Water Resour. 285-289. F. Proceedings of the SPEIDOE Symposium on Enhanced Oil Recovery. 29(9).C. Lifshitz. 1992. P. a stable film will be present for systems that have both negative and positive initial spreading coefficients. Society of Petroleum Engineers... presented before the Division of Petroleum Chemistry. American Chemical Society. J. J.1993. A. Halpern and J. 2963-2974. V. The calculation of Hamaker constants from Lifshitz theory with applications to wetting· phenomena. White. 81. Society of Petroleum Engineers. Inc. 29. Vol. Systems with a large value of a give the best final recoveries. I. Houston. SPE 24878. Rowlinson. and Exhib. J. J. (November). 437-444. Chatzis. J. Advances in Physics. 1992. Vizika. than free drainage. . I. A. 1ntermolecular and Surface Forces.. Spectroscopic parameters for computation of van der Waals forces. K. E. and G. Kantzas. S. A. Parsegian. E. Oil remains near the bottom in an oil bank or in bulk layers in the pore space.. Kantzas. D. R. D. Widom. 1988a. Tech. Dullien. and W. Dzyaloshinskii. 1955. V. 1988b. Proceedings of the 68th Annual Technical Conference and Exhibition of the SPE. and L. L. Muskat. Dullien. REFERENCES Hough. Chatzis. 1949. Eval. Form. Advances in Colloid and Interface Science. Oxford. Rowlinson. when all the transition zone was left in the column. Soc. Proceedings of the SPEIDOE Enhanced Oil Recovery Symposium.1961. W. and W. Society of Petroleum Engineers. SPE 26671. L. This prediction was confirmed by experiments on sand columns. A. S.. M. P. The Collected Works of J. Pitaevskii. Mobilization of waterflood residual oil by gas injection for water-wet conditions.. Proceedings of the 63rd Ann. A low oil/water surface tension ensures that these layers are always stable and results in a lower recovery than from high IFf fluids. We predicted the drainage rate in a square capillary tube. Society of Petroleum Engineers. Israelachvili. Pinczewski. Vol. J. Celia. L. 1990. Kantzas.. Drainage capillary-pressure functions and the influence of connate water. SPE 18284.. Wettability: fundamentals and surface forces. New York. The oil film ensures pressure continuity in the oil phase.C. G. 1991. Performance and analysis of three-phase capillary pressure curves for drainage and imbibition in porous media.165-209. and L. and P. and computerassisted tomography. 1993. Denver.. Vizika. and R. Physical Chemistry of Surfaces. Billiotte. New York. Chatzis. Soli. and Exhib. Boston... Willard Gibbs.. Adamson. 1974. SPE 17379.. stabilized by structural forces. M. 10f Longmans. A. F. that allows oil ganglia to drain. 10. Tulsa. and B.. 1980. Houston. E. Eng.. Oklahoma. Society of Petroleum Engineers. P. These experiments demonstrate that the final amount of oil left in the system depends on the presence of oil in bulk layers in the ppre space. Physical Principles of Oil Production. A. L. 1nc. A. Kalaydjian. 1988. Tech. Eng. Pinczewski. E. Clarendon Press.. Virtually all the oil can be recovered in gravity drainage from the top of a long water wet column. SPE 17506. S. W. Res. J. The general theory of van der Waals forces. Texas. Hirasaki. 1988. Tulsa.. Weiss. Tech. 2nd ed. Enhanced oil recovery by inert gas injection. B. Washington D. 5th ed. Journal of Colloid and Interface Science. M. Gibbs. On the investigation of gravity-assisted inert gas injection using micromodels. the thickness of this film depends on the capillary pressure and the displacement history. Schols.. J.. Conf. John Wiley & Sons. London.. long Berea Sandstone cores. of the SPE.. McGraw Hill. Conf.. Wilson. Molecular Theory of Capillarity. In a porous medium.-M. If the oil film is formed as a thicker layer of oil recedes. Journal of Experimental and Theoretical Physics. A. Proceedings of the Symposium on Enhanced Oil Recovery. E. Effect of the spreading coefficient on the efficiency of oil recovery with gravity drainage. Academic Press. O. Dumore. Wyoming. M. J. I. Pet. Conf. Kalaydjian. Munkerud. 14. SPE 17367. V. Dullien.10 WHAT DETERMINES RESIDUAL OIL SATURATION IN THREE PHASE FLOW? SPE 27816 drainage. The drainage is facilitated by the swelling of the oil layers in crevices in the pore space to thicknesses comparable to the radius of the pore. and F. Moulu.. I. Micromodel studies of three-fluid porous media systems: pore-scale processes relating to capillary pressure-saturation relationships. Three-phase flow in water-wet porous media: determination of gas/oil relative permeabilities under various spreading conditions. Pet... 70-78. 3-41. (March). A. Casper. and F. 94. Colorado.

SPEIDOE 35385

EOR Screening Criteria Revisited
Joseph J. Taber, SPE, F. David Martin, SPE, and R.S. Seright, SPE, New Mexico Petroleum Recovery Research Center Introduction Oil production from enhanced oil recovery (EOR) projects continues to supply an increasing percentage of the world's oil. About three percent of the worldwide production now comes from EOR. Even though EOR production in the U.S. appeared to peak in 1992, Fig. 1 shows that the EOR percentage of the U.S. production is larger than ever; therefore, the importance of choosing the "best" recovery method becomes increasingly important to petroleum engineers. About 100 years ago, oil producers injected gas to restore pressure to their dying oil wells. I Since air was cheaper than gas, air was often injected to -increase production from the older fields. For many years, operators had the choice of air or gas, and sometimes they injected both into the same reservoir.' Naturally, there were safety and other problems with air. However, not until about 1928 did natural gas become the injectant of choice for pressure maintenance.' Water injection was legalized in Pennsylvania in 1921 (it was done secretly before that)." The choice of injectants has widened considerably since those early days, but the petroleum engineer still must choose an injection fluid and an overall process to try to recover the maximum amount of oil from the reservoir while still making a profit. Screening criteria have evolved through the years to help the petroleum engineer make these decisions.':" Some of the early work in this field was done by Geffen':" before there was much field experience with most EOR methods. Many of his criteria have stood the test of time. Perhaps the best known, and most widely used, screening criteria appeared in the 1976 and 1984 NPC reports." We will comment later on some of the predictions based on these criteria. One paper that we are "revisiting" (see title) is SPE 12069 by two of'us." Although we retain the format of some of the tables in SPE 12069, all have been revised. We are basing our criteria in this paper on the results of much more field and laboratory information that has become available. In recent years, computer technology has improved the application of screening criteria through the use of artificial intelligence techniques, but the value of these programs depends on the accuracy of the input data used.":" In this paper, we present screening criteria based on a combination of the reservoir and oil characteristics of successful projects plus our understanding of the optimum conditions needed for good oil displacement by the different EOR fluids. One goal is to provide realistic parameters that can be used in the newer computer-assisted tools for reservoir management.
Copyright 1996. Society of Petroleum Engineers. Inc. This paper was prepared for presentation at the SPEIDOE Tenth Symposium on Improved Oil Recovery held in Tulsa, OK. 21-24 April. 1996 This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been rviewed by the Society of Petroleum Engineers and ere subject to oorrection by the euthor(s). The material, as presented. does not necessarily reflect any position of the Society of Petroleum Engineers. Hs officers. or members. Papers presented at SPE meetings are subject to J)I.bIication review by EdHoriai Committees of the Society of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspiaJOUs acknowledgment of where and by whom the paper is presented. Write Librarian, SPE. P.O. Box 833836. Richardson, TX 75083-3836, U.S.A Telex, 163245 SPEUT.

Abstract Screening criteria have been proposed for all EOR methods. Data from EOR projects around the world have been examined and the optimum reservoir/oil characteristics for successful projects have been noted. The API gravity ranges of the oils of current EOR methods have been compiled and the results are presented 'graphically. The proposed screening criteria are based on both .fIeld results and oil recovery mechanisms. The current state of the art for all methods is presented briefly, and relationships between them are described. Steamflooding is still the dominant EOR method. All chemical flooding has been declining, but polymers and gels are being used successfully for sweep improvement and water shutoff. Only CO2 flooding activity has increased continuously. Screening criteria are useful for cursory examination of many candidate reservoirs before expensive reservoir descriptions and economic evaluations are done. For example, we have used our CO2 screening criteria to estimate the total quantity of CO2 that might be needed for the oil reservoirs of the world. If only depth and API gravity are considered, it appears that about 80% of the world's reservoirs could qualify for some type of CO2 injection. Since the decisions on future EOR projects are based more on economics than on screening criteria, future oil prices are important. Therefore, we examined the impact of oil prices on EOR activities by comparing the actual EOR oil production to that predicted by earlier NPC reports. Although the lower prices since 1986 have reduced the number ofEOR projects, the actual incremental production has been very close to that predicted for $20lbbl in the 1984 NPC report. Incremental oil production from CO2 flooding continues to increase, and now actually exceeds the predictions made for $20 oil in the NPC report, even though oil prices have been below that level for some time.

387

2

J.J. TABER, F.D. MARTIN, AND R.S. SERIGHT

SPE 35385

EOR-IOR-ASR-Reservoir Management, etc. In the past few years, the term lOR (improved oil recovery), has been used instead of the traditional EOR, or the more restrictive "tertiary recovery." Most petroleum engineers understand the meaning of all of the words and phrases, but our technical communications are improved if we use the terms with their intended technical meanings. We certainly endorse the wider use ofIOR (and note that our paper is part of this conference's session on "lOR by Reservoir Management"), but we will cling to the technical mfanings of EOR and tertiary recovery. Successful enhanced recovery projects are being conducted as tertiary, secondary, and even enhanced primary operations. The terms should continue to be used with their evolved historic meanings. Tertiary should not be used as a synonym for EOR because some EOR methods to work quite well as either secondary or tertiary projects (e.g. CO2 flooding) while others, such as steam or polymer flooding, are most effective as enhanced secondary operations. To us, EOR simply means that something other than plain water or brine is being injected into the reservoir. We will use the terms "enhanced secondary" or tertiary when necessary for clarity. Others may use the phrase Advanced Secondary Recovery (ASR)'7-21 for EOR in the secondary mode. We are convinced that engineers should consider this improved (enhanced or advanced) secondary option much more often in the future. Classification of EOR Methods. Table I lists more than 20 EOR methods that experienced intensive laboratory and, in most cases, significant field testing. The methods utilize about 15 different substances (or specific mixtures) that must be purchased and injected into the reservoir, always at costs somewhat greater than for the injection of water. The economics ofEOR will be discussed more later, but experience shows that the best profits come only from those methods wherein several barrels of fluid (liquid or gas at reservoir pressure) can be injected per barrel of incremental oil produced.f-" This limits the main methods to either water (including heated, as steam, or as a dilute chemical solution) or one of the inexpensive gases. For some methods (e.g., micellar/polymer) there have been some technical successes but relatively few economic successes. These methods are included in our screening criteria because they are still being studied and applied in the field. If oil prices rise significantly, there is hope that these methods might become more profitable. We provide screening criteria for the nine methods that are either the most important, or still have some promise. These nine methods are shown in bold type in Table I, along with the number of the table for those eight methods that are examined in more detail. These "current" EOR or lOR methods include the three gas (nitrogen, hydrocarbon, CO2), three water (micellar/polymer plus alkaline/surfactant/polymer; polymer flooding; gel treatments) and the three thermal-mechanical (combustion, steam, surface mining) methods. A convenient way to show these methods is to arrange them by API gravity as shown in Fig. 2. This "at a glance" display also provides approximate API gravity ranges for the field projects now underway. The size of the type is intended to show the relative importance of each of the EOR methods in terms of

current incremental oil production. When examining the rationale for some of the screening parameters, it is instructive to consider the oil displacement mechanisms for the EOR methods. Table 2 shows that there are three main mechanisms for displacing additional oil with an injected fluid: (I) solvent extraction to achieve (or approach) miscibility, (2) interfacial tension reduction, and (3) viscosity change of either the oil or water, and/or plus additional pressure added to the injection fluid. There is overlap of the mechanisms. For example, interfacial tension is lowered as miscibility is approached in the "solvent" methods. The reservoir and injection conditions should be chosen to optimize the displacing mechanisms wherever possible (e.g., use a high enough pressure to achieve miscibility in solvent flooding, and look for shallow reservoirs to reduce wellbore heat losses in steamflooding). Note that we have added "enhanced gravity drainage" by gas inj ection to Table 2. Although not shown as a separate method in Table I, it is covered in Table 3 because it is a part of each of the three gas injection methods.

OiVReservoir Characteristics of Successful Projects
The depth, and the corresponding API gravity, of most of the EOR projects in the world are shown in Figs. 3 and 4. We have included projects for which 'data are available from a recent Oil and Gas Journal survey." We are more familiar with the U.S. projects (Fig. 3.) than those in other parts of the world (Fig. 4). In addition to the very broad distribution of the EOR projects, Fig. 3 shows the general trend, ranging from the many steam projects for the heavy oils at shallow depths in California to the / very deep projects for the lightest oils that can be miscibly . displaced by dry gas or nitrogen at high pressures. The waterbased methods utilize oils in the mid-gravity range, while the CO2 projects cover a fairly broad range of oil gravities between 30 and 45° API. Fig. 3 confirms that all CO2-miscible projects are at depths greater than 2,000 ft. Fig. 4 shows that the nonU.S., world distribution of projects is similar, but there are more hydrocarbon and fewer CO2 projects than in the U.S. The incremental oil production from each EOR project is shown in Figs. 5 and 6. The dominance of steamflooding stands out clearly in these figures. Not only are there far more steamfloods, but the oil produced by steamflooding far exceeds that from all the other methods combined. Note that the largest EOR projects (in terms of oil production), are steamfloods, with the "off-scale" (Fig. 6) Duri steamflood in Indonesia producing more than twice as much oil (245,000 bid) as any other project in the world.

Suggested Criteria for EOR Methods
Oil and reservoir characteristics for successful EOR methods are given in Table 3. The table was compiled from field data for the projects shown in Figs. 3-6, and from the known oil-displacement mechanisms for each of the methods. Very brief descriptions of these mechanism are given in the "thumbnail sketches" of the methods in Tables 4-11.

388

SPE 35385

EOR SCREENING

CRITERIA

REVISITED

"3

Note that we have avoided notations such as ~ (equal to or greater than) because we want to emphasize that the suggested '\arameters are never absolute. They are intended to show '~pproximate ranges of the criteria for good projects. In most cases, when we show >x or <y, etc, there is not a specific upper (or lower) boundary to the parameter except for the limits of the oil and reservoir characteristics, as found in nature. For example, we show that nitrogen floods are recommended for oils lighter than 35 API, but this does not mean that the probability of doing miscible nitrogen floods drops to zero at 34 API. This obvious shortcoming of most screening criteria tables has been noted by ~uthors who use artificial intelligence (AI) methods to select EOR processes for specific reservoirs. I I To overcome the problems that arise by using rigid boundaries in their "crisp" expert systems, some AI workers have utilized "fuzzy logic" methods to obtain much more realistic results 12. In Table 3, we attempt to show that, for a given parameter, if >x is feasible, »x may be even better for a given process. By underlining a value, we indicate the average or mean of the parameter for that EOR method. For example, for the oil gravity in miscible nitrogen floods, >35 ..... .....means that the process 48 should work with oils greater than 35 API (if other criteria are met) and that higher gravity oils (..... are better, and that the ) approximate mean or average of current miscible nitrogen projects is 480 API. The ascending arrow is meant to indicate that higher gravity oils may be better yet. In general, the upper and lower values in Table 3 (> or <) have come from process-mechanism understanding (laboratory ,~xperirnents), and they also include parameters of successful .field projects. For example, even though we are unaware of any miscible CO2 projects in reservoirs with oils of less than 29 API, we list 22 API as the lower limit because extensive laboratory work shows that the required pressure (i.e., MMP, see Table 6) can be met in typical West Texas reservoirs with oils of that gravity. Also, we have lowered the oil gravity requirement to > 12 0 API for immiscible CO2 floods to include a successful project in Turkey (see Fig. 6).
0 0 0 0 0

Gas Injection Methods. Gas injection, the oldest EOR method, is a bright spot in EOR technology. Although most EOR production comes from steamflooding, Figs. 5 and 6 show that gas injection methods are next in importance and appear to be growing throughout the world. Oil production from CO2 flooding is the only EOR method that has continued to increase (Fig. 1) in the U.S. in spite of various declines in oil prices through the years, and more projects are planned. Hydrocarbon gas injection is second to steamflooding for the entire world. Thanks to efforts to reduce gas flaring, gas injection should continue to grow in importance as worldwide oil production expands. After years of extensive laboratory and field experience, the gas EOR methods are now well understood, and screening criteria can be recommended with more confidence than before. Although studied most extensively for CO2, the concept of minimum miscibility pressure (MMP) explains the efficient oil displacements by N2, hydrocarbons, and CO2, As long as this MMP can be achieved in the reservoir, good oil recovery (greater than 90% OOIP in the

region swept) should result, although CO2 displacements are usually more efficient than N2 or CH4. Even though the oil gravity-pressure-depth (MMP) requirements are different for the three gases, Table 3 shows that there is overlap of the criteria for the three methods. Thus, any of the methods will work in a high percentage of the deeper reservoirs, and the fmal choice will often depend on the local availability and cost of the gas to be injected. Nitrogen (and Flue Gas) Injection. Other than compressed air, nitrogen and flue gas are the cheapest gases (especially in terms of volumes at reservoir temperatures and pressures) that can be injected. They are considered together because the pressures required (MMP) for good displacement are similar," and it appears that they can be used interchangeably for oil recovery. Indeed, at least three of the current nitrogen projects (Block 31, TX; E. Binger, OK; Hawkins, TX) 24were operated successfully for years as flue gas injection projects.23•26 However, corrosion was a problem (especially for flue gas from internal combustion engines), and all have switched to nitrogen injection with good results. In addition to its low cost and widespread availability, nitrogen is the most inert of all injection gases. Unfortunately, it has the highest MMP, so miscible displacement is possible only in deep reservoirs with light oils. Hydrocarbon Injection. As one of the oldest EOR methods, hydrocarbon injection was practiced for years before the MMP concept was, well understood. When a surplus of a lowmolecular-weight hydrocarbon existed in some fields, they were often injected to improve oil recovery. The three different methods were described by Stalkupf and are summarized very briefly in Table 5, including tirst-contact-miscible (LPG solvent), condensing (or enriched) gas drive and the vaporizing (or highpressure) gas drive. In terms of the pressure required for efficient miscible displacement, we rank the hydrocarbon gases between the very high pressures required for nitrogen and the more modest range of pressures for CO2 (see Table 6 for the reservoir depth requirement for.different gravity oils). This ranking is correct for methane. However, if a shallower reservoir depth requires a lower pressure, it can be achieved by adding more enriching hydrocarbons (usually C2-C4 if the economics permits).":" This fine-tuning method is practiced most in Canada where cheap CO2 is in short supply and hydrocarbon gases are available. CO2 Flooding. There may be more optimism for CO2 flooding in the U.S. than for any of the other EOR methods. As noted before, it is the only method that has had a continuous increase in production since CO2 flooding started more than 35 years ago. The technical and economic reasons for the success of CO2 flooding have been explained before." In the West TexasNew Mexico Permian basin, a large pipeline supply of natural CO2 is available at a low cost compared to methane, and the pipelines are being extended to more fields.":" The screening criteria in Tables 3 and 6 show that a fairly wide range of crude oils and reservoir depths can meet the requirements for miscible CO2 flooding. The solubility of CO2 in oil decreases with temperature, so the MMP required for a given oil must also increase with

389

4

J.J. TABER, F.D. MARTIN, AND R.S. SERIGHT

SPE 35385

temperature." Since the reservoir temperature normally increases with depth, the MMP must also increase with depth, as shown in Fig. 7 for a 40 API oil in typical West Texas reservoirs. Fortunately, the pressure required to fracture reservoirs increases faster than temperature with depth. Therefore, there is an MMP ''window of opportunity," as shown in Fig. 7. The correlations in Fig. 7 and Table 6 come from many sources and are reviewed briefly in Refs. 29, 32, and 33. Most of the relationships among temperature, oil composition and pressure come from extensive work by various workers, primarily on oils from fields in the U.S.34-]7The MMP screening criteria in Fig. 7 should work well for oils that have hydrocarbon distributions similar to the average mid-gravity crude oils of the U.S., especially those from the Permian basin of West Texas and Southeast New Mexico. However, if the oil differs significantly from the types of crudes for which the correlation was developed, additional laboratory tests may be required. Hagadorn and Orr" have shown that a high percentage of multi-ring aromatics will raise the MMP significantly because they are extracted so poorly by the CO2 phase.
0

Chemical and Polymer Flooding and Gel Treatment Methods. Figs. 3~ show that there are relatively few chemical flooding projects (shown as polymer or micellar/polymer) in the world and that these projects contribute little to worldwide EOR production when compared to steamflooding and gas injection. For our screening criteria, we want to concentrate most on current technology that can be applied profitably today. Therefore, we have limited our criteria in Table 3 to these broad methods that are often included in the general term "chemical flooding." We are not aware of any pure alkaline floods at present. There are alkaline/surfactant/polymer (ASP) projects that, hopefully, are a low-cost improvement over micellar/polymer or surfactant flooding. Therefore, we have dropped the separate alkaline flooding category and combined it with the two main surfactant (interfacial-tension lowering) methods as shown in Tables 3 and 7: micellar/polymer; alkaline/surfactant polymer (ASP) and alkaline flooding. There is stilI some excellent chemical flooding research and development work underway in laboratories around the world. The polymer injection projects (especially in the U.S., see Fig. 3) far outnumber the other chemical flooding methods. However, there has been some confusion between polymer flooding for enhanced oil recovery and the injection of gelling polymers for water shutoff in either injection or production wells. Therefore, they are considered separately in Table 3 and in Tables 8 and 9. Wettability is another area of importance to waterflooding, and significant progress on understanding the influence of wettability on oil recovery in being made.":" However, it would be premature to try to include wettability in our screening criteria at this time. Micellar/Po/ymer, AlkalinelSurjactanVPo/ymer (ASP) and Alkaline Flooding. The goal of the chemical methods is to reduce the interfacial tension between oil and water, generally to displace discontinuous trapped oil (Sor) that remains after a

waterflood. Since it is approximately ten times more difficult tOI replace trapped oil than continuous oil," the surfactant slugs for. these tertiary processes must be very efficient. The oil displacement mechanics are well understood and many formulations have been devised to give very high recoveries in laboratory experiments with actual reservoir rocks and fluids. There have been some technical successes in the field.42•4] However, there have been fewer economic successes because the cost of the injectant is too high. Therefore, there has been an effort to lower the injectant cost by adding more alkali and less surfactant or cosolvent to the formulations during the past few years.23•44•4S These mixtures are often termed alkaline/surfactant/polymer (ASP) processes, and very large "slugs" can be injected since the cost is so low compared to the classical micellar/polymer formations. The alkali costs much less than the surfactant or cosolvent and it helps to lower the interfacial tension and reduce adsorption of the surfactant on the rock.46•47 In one case, workers were able to reduce surfactant concentration by ten times by adding low-cost alkali, and the formulation still provided very good oil recovery" The ASP process has also been tested in the field." A recent fieldwide project in Wyoming reports costs of$1.60-$3.50Ibbl of incremental oil produced." Polymer Floods and GeiTreatments. In the past, polymer floods and gel treatments were often lumped together as a single technology." However, these processes have very different technical objectives, so here, they will be considered separately. The distinction between a mobility-control process (e.g., a I polymer flood) and a blocking treatment (e.g., involving cross- ., . linked polymers or other gels) is an important concept to understand. For polymer floods and other mobility-control processes, the mobility-control agent should sweep evenly through the reservoir. In other weirds, the polymer should penetrate as far as possible, into the low-permeability zones because that action provides the driving force for displacing and producing unswept oil. , In contrast, for gel treatments, gel penetration should be minimized in less-permeable, oil-productive zones.. Any gel that forms' in the oil-productive zones reduces the oil-displacement efficiency and retards oil produc-

'I

tion."
For existing gels and gelants that are used as blocking agents, the following behavior is observed during flow through porous media. S3-SS First, before gel aggregates grow to approach the size of pore throats, gelants flow through porous media like solutions without crosslinkers. Second, after gelation (or after gel aggregates grow to the size of pore throats), gel movement through porous rock is negligible. Third, in porous rock, the transition from a freely flowing gelant to an immobile gel occurs abruptly. After gel formation, crosslinked polymers, gels, gel aggregates, and the so-called "colloidal-dispersion gels" do not flow through porous rock like viscous polymer solutions. 54 Also, they do not enter and block the most-permeable strata first and then sequentially enter and block progressively less-permeable zones. Gelants and polymer solutions enter all zones simultaneously. S2 (Of course, the distance of polymer or gelant penetration depends directly on the permeability.) Understanding these concepts is particularly important for projects that were designed

390

SPE 35385

EOR SCREENING

CRITERIA

REVISITED

5

as polymer floods but that used HP AM crosslinked with aluminum citrate (i.e., the "colloidal-dispersion gels")." For these .projects, an important question is, "Would the field response j have been better ifHPAM had been injected without aluminum

citratej?"
Polymer Flooding. Over the past 35 years, a large number of polymer floods have been applied over a remarkably wide range of conditions. 57.58Reservoir temperatures ranged from 46 to 235°F; average reservoir penneabilities ranged from 0.6 to 15,000 md; oil viscosities ranged from 0.01 to 1,494 cp; net pay \ ranged from 4 to 432 ft; and resident brine salinities ranged from 0.3 to 21.3% total dissolved solids (IDS). At project startup, the percent of original oil in place (%OOIP) ranged from 36% to 97.1%, and the producing water/oil ratio ranged from 0 to 100. During the 1980s, polymer floods were applied in sand or sandstone reservoirs about four times more frequently than in carbonate reservoirs. 58 In concept, a polymer flood could improve sweep efficiency during any waterflood. However, a number of technical and economic factors have limited the application of successful polymer floods. The cost-effectiveness of polymers (i.e., the mobility reduction or viscosity provided per unit cost of polymer) is the main economic limitation. For example, if the cost ofacrylamide/acrylate copolymers (HPAM) and xanthan polymers were substantially lower, higher polymer concentrations and larger polymer-bank sizes could be afforded in a given application. This, in turn, would lead to greater oilrecovery efficiencies, higher profits, and a wider range of potential applications. \ Cost-effectiveness also impacts the permeability constraints ) for polymer flooding. For a given polymer, chemical retention increases and the rate of polymer propagation decreases with decreasing rock permeability. Current high-molecular-weight polymers often experience high retention and low propagation rates for rock penneabilities below 100 md." This permeability constraint can be relaxed by using polymers with lower molecular weights. However, the viscosity provided by a polymer decreases with decreasing molecular weight, so more polymer (and a higher cost per viscosity unit) is needed as the rock permeability and the maximum allowable polymer molecular weight decrease. An important issue related to reservoir permeability is that of injectivity (injection rate per pressure drop). In wells that are not fractured, injection of viscous polymer solutions will necessarily decrease injectivity. In order to maintain the waterflood injection rates, the selected polymer-injection wells must allow higher injection pressures. This requirement becomes increasingly difficult to fulfill as the formation penneability decreases, unless the wells are fractured. If injectors are fractured, one must ask, "Will the increased injectivity from fracturing outweigh the increased risk of channeling?" (In a later section, we suggest that horizontal injection wells may alleviate injectivity limitations in some cases.) Cost-effectiveness also affects the temperature constraints for polymer flooding. More than 95% of previous polymer floods were applied in reservoirs with temperatures below .200°F.58 This fact reflects wide-spread doubt that HPAM and xanthan polymers are sufficiently stable at elevated temperatures.

Literature reports 59question whether these polymers are stable for field applications above 175°F. More stable polymers (e.g., scleroglucan and acrylamide/ AMPS/NVP copolymers and terpolymers) are available for high-temperature use, but the cost and cost-effectiveness of these polymers have limited their application to date." Of course, significantly higher oil prices and/or breakthroughs in reducing polymer production costs could change this situation. For many years, water salinity has been an important issue in polymer flooding. 59 In the range from 0 to 1% TDS, the viscosities of HPAM solutions decrease substantially with increased salinity. Thus, high-salinity HPAM solutions are relatively ineffective during polymer flooding. Differences of opinion existed concerning the viability of injecting low-salinity HPAM solutions into reservoirs with high-salinity waters. An important paper that addressed this issue was presented by Maitin." In a well-documented field study, Maitin demonstrated the conditions needed for low-salinity HPAM solutions to be effective in high-salinity reservoirs. In reviewing literature reports of polymer floods, we often noted considerable uncertainty in assessing the benefits after a given project was completed. Most previous polymer floods used relatively small quantities of polymer (both in terms of polymer concentration and bank size)." Consequently, relatively small lOR values (1 to 5% OOIP) were often projected that resulted in small alterations of the oil-production decline curves and the WOR curves. Commonly, these small alterations were difficult to discern when comparing the actual polymer-flood response with the projected waterflood response. In contrast, several polymer floods stand out that showed definitive responses, such as at the Marmul," Oerrel,":" Courtenay.Pand Daqing" fields. Properties of these successful polymer floods are listed in Table 8 along with median values for all polymer floods that were applied during the 1980s. The four. successful floods listed in Table 8 had a number of features in common. These characteristics may be useful as screening criteria for .today's economic environment. First, the floods were applied in high-permeability (>0.87 darcy) sands and low-temperature (86°F to 136°F) reservoirs. High oil saturations (71 to 92 % OOIP) were present at project startup, and the oil/water viscosity ratios (15 to 114) at reservoir temperature were relatively high. The injected polymer solutions contained relatively high HPAM concentrations (900 to 1,500 ppm) in low-salinity waters, and large quantities of polymer (162 to 520 lbs polymer/ac-ft) were injected. Finally, the incremental oil recoveries (11 to 30 % original oil in place or 155 to 499 bbl oil/ac-ft) were high. Gel Treatments. Gel treatments have been applied under conditions as diverse as those listed previously for polymer floods.58.65As mentioned earlier, the technical objective ofa gel treatment should be very different from that of a polymer flood. In most cases, the objective of a gel treatment is to prevent channeling of fluid (usually water) without damaging hydrocarbon productivity. After extensive discussions with experts from the oil and service companies.Y" we developed the criteria listed in Table 9 for selection of gel-treatment candidates for injection and production wells.

391

6

J.J. TABER, F.D. MARTIN, AND A.S. SERIGHT

SPE 35385

For both injection and production wells, the first criterion indicates that a sufficient target of mobile oil must be present, and realistic calculations should indicate that the oil can be recovered economically. In Table 9A, several qualifications were added to clarify what is meant by "low sweep efficiency." We felt that this clarification was necessary because different operators interpret this phrase in radically different ways. Cases exist where gel treatments were implemented even though more than 50% OOIP had been recovered before the treatment." Point c in the first criterion is particularly important for applications in gray areas. One criticism of this point might be that, often, insufficient manpower or reservoir description is available to adequately predict the benefits of a gel treatment. If that is the case and if the second or third criterion in Table 9A cannot be met, then we feel that a gel treatment in an injection well is unlikely to be successful. Recovery calculations were also felt to be very important in the selection of production-well candidates (Table 9B). An essential element of improving the success rates for gel treatments is adequate identification of the source and nature of the channeling problem. Determination of the importance of fractures as channels is a first priority. For applications in injection wells or waterflood production wells, interwell tracer studies can be the most effective tool in identifying very severe channeling between injector-producer pairs." Very rapid tracer transit times (less than one week) probably indicate that the channel is a fracture or a formation part. In addition to diagnosing the severity of the channeling problem, tracer studies can be very useful in designing the volume of gelant to be injected and assessing the ultimate effectiveness of the treatment. 68 Table 9 indicates that high well injectivities or productivities are important for candidate selection. These requirements recognize that a gel treatment will generally reduce the flow capacity of a well. For a given pressure limit, the injection or production rates will be lower after a treatment than before a treatment. By having excess injection or production capacity before the treatment, the operator will be more likely to tolerate the reductions in flow capacity that result from a gel treatment. Simple injectivity or productivity calculations can aid in establishing the nature of a channeling problem." For unfractured injection wells, Table 9A lists three important conditions that must be present before candidate selection: (1) poor injection profiles must correlate from well to well, (2) effective barriers to crossflow must exist, and (3) gels can be placed in the offending channel without damaging oil zones. To prevent damage to oil productivity, zones usually must be mechanically isolated during gelant placement in unfractured injection wells." This requirement also applies in unfractured production wells. If zones are not isolated during gelant placement in unfractured wells, low-permeability zones can be seriously damaged even in extremely heterogeneous reservoirs (e.g., Dykstra-Parsons coefficient ofO.9).5l Most vendors and operators feel that the candidate well should be in good mechanical condition for applications in injection wells," but this requirement may be less critical for production-well applications. Not all gel applications were directed at in-depth channeling

problems. Often, gels are used as a substitute for cement in fixing casing leaks or flow behind pipe. The main advantage of gels over cements is a superior ability of gelants to penetrate into constricted spaces, such as narrow channels behind pipe, small casing leaks, and fractures. Also, because gelants can penetrate into porous rock, gels can sometimes form a better pipe-formation seal than can be obtained using cement. 67-69 In addition to gels, many different water-shutoff methods are available, including cement, sand plugs, calcium-carbonate plugs, resins, foams, particulates, emulsions, precipitates, microorganisms, mechanical devices (e.g., packers and bridge plugs), well abandonmentlinfill drilling, pattern flow control, and horizontal wells. Each of these methods will work very well in solving some types of problems but not others. The proper choice of water-shutoff material and method depends critically on the cause of the problem. Gel treatments work best for those problems identified in Table 9. Thermal-Mechanical Methods for Heavier Oils and Tar Sands. Thermal methods account for the biggest share of the world's enhanced oil production. The largest EOR operations in many countries (e.g. Canada, Columbia, Germany, Indonesia, Trinidad, U.S. and Venezuela) are either steamfloods or surface mining operations. In the past, the production of bitumen from tar sands has not normally been included in EOR screening. criteria or surveys.perhaps because the mining operations are not considered a part of reservoir engineering. However, the resource is so important that hydrocarbon recovery from tar sands should be included in listings ofEOR or lOR processes. There is a very strong effort to try to recover these extremely viscous oils by in- ' situ methods 70 to avoid the cost of surface mining and to open up vast deeper reserves. One method that shows promise uses horizontal wells in a variation of steamflooding known as Steam Assisted Gravity Drainage or SAGD7I-73 This mechanism is akin to the enhanced gravity drainage by immiscible gas injection mentioned previously and for which screening criteria are given in Table 3. In general, the screening criteria for SAGO would be similar except that the depth, viscosity and API gravity ranges should be extended to include the tar sands. Thermal EOR projects have been successful for more than 30 years, and the methods have been described in many publications and a few books.74-76 We comment here on only a few aspects that relate to screening criteria. In general, .thermal methods have been used for those heavy oil reservoirs that cannot be produced in any other way because the oil is too viscous to flow without the application of heat and pressure. In order to be produced at profitable rates, the sands must have a high permeability and oil saturations must be high at the start of the process. Therefore, the successful projects are almost always enhanced secondary (or even enhanced primary, since primary production was essentially nil in many fields). In-Situ Combustion. In situ combustion seems like an ideal EOR method because: It utilizes the two cheapest and most plentiful of all EOR injectants: air and water.

392

" Turta also described benefits of horizontal wells that have shown promising results in two Canadian combustion projects. 88 As for screening criteria. the observations in our earlier paper still apply. at a recent symposium on in situ combustion. the oil displacement mechanisms are well understood. However.SPE 35385 EOR SCREENING CRITERIA REVISITED 7 . Farouk Ali-claimed that "insitu combustion remains the most tantalizing EOR method. Although not normally listed with EOR screening criteria. it burns about 10% of the least desirable fraction of the oil." As the world's largest EOR project." Normally" the steam is injected into the upper well of two parallel horizontal wells. 91..14 Deeper. the decision is not if. we left a question mark for the upper limit to the API gravity for steamflooding a medium gravity oil that could be waterflooded as well as steamflooded. where it is pumped to the surface. there is an increased effort to producethese viscous hydrocarbons by in-situ methods such as the SAGD process.It works over a wider range of field conditions than steam. It does appear that light-oil steamfloods should always be planned as enhanced secondary operations. The steamflood should produce much more oil. At the other end of the oil-gravity-steamflooding spectrum are the aforementioned Steam Assisted Gravity Drainage (SAGD) projects in heavy-oil or tar sands. good projects require thick. especially since the heavy oil has less value than higher gravity crudes. and the resulting combustion front moves down-dip to displace the oil either miscibly or immiscibly by the flue gas produced from the combustion. Sarathi and Olsen showed that only one of eight cost-shared projects was an economic success.. .20W API inclusive. i. The criteria are also useful for surveys of a large number of fields to determine if specific gases or liquids could be utilized for oil recovery if an injectant was available at a low cost. Most of the other light-oil steamfloods had much lower oil saturations.. . the tar melts and flows by gravity to the lower well. the cogeneration of steam and electric power has been very beneficial to both the economics and environmental problems. almost all of these require one or more horizontal wells to inject the steam and withdraw the melted bitumen. Air injection has been underway since 1981. It will take a careful economic analysis of each potential light-oil steamflood to determine if the additional oil will pay for the additional cost of the steamflood. so the combustion front can propagate down dip. and may upgrade the rest. Steamflooding is the oldest commercial EOR method. the mining method is used only when the oil is so viscous that it cannot be recovered by any other method because the mining and upgrading of the bitumen are more costly than in-situ recovery methods. This complicated process has been studied extensively" and tried in many different types of reservoirs. However. These problems and their solutions were described in a recent review. air injection must start in the uppermost part of the reservoir. Much of the current emphasis is on improving the economics through better reservoir management. but a cursory examination with the technical criteria is helpful to rule out the less-likely candidates. engineers consider the various IOR-EOR options much earlier in the productive life ofa field. For example. light-oil reservoirs with significant dip are also targets for a new method of in situ -ombustion that might be considered another variation of )manced gravity drainage by nitrogen or flue gas." In times of low oil prices. Mining and Extraction.":" However. it is a complicated method with safety and corrosion problems that always need attention. In Table 3.tion. Although different techniques are under development. but when to inject something. For this reason./ooding." Steam flooding. so economic success was more difficult. With the application of hot steam and pressure. its starting oil saturation of 63% is near the average of the successful steamfloods in the world. the Duri steamflood is certainly successful (see Fig. horizontal wells are being planned to improve light-oil. we include surface mining because the tar sands are such an important hydrocarbon resource. but it provided valuable information on how to engineer a successful project." References indicate that their use and other advanced engineering methods should make it possible to extend steamflooding to both lighter and heavier oils." In another application. especially in deep reservoirs. iave also been conducted in light-oil reservoirs and a few have been successful. 0 Utilization of Screening Guides With the reservoir management practices of today." Several field tests . As mentioned in the previous section. but an effective waterflood will be cheaper.10 At the same symposium. air is injected in the formation.e. the economics are very tight. The operator hopes that horizontal wells will increase the recovery from the current 20-30% to 50% OOIP.. the tar sands must have a high oil (bitumen) saturation. preferably with a line drive well configura. For many fields. and the ratio of overburden to tar sand must be low as shown in the screening criteria of Table 3.For fuel. IS In this process. economics always play the major role in "go-no go" decisions for expensive injection projects. Obviously.92 The Duri project in Indonesia is sometimes referred to as a light-oil project because its 22 API oil is outside Unitar's defmition of heavy oil: 10Y2. Newer materials and technology should help solve some of the field problems. 6). and the production of synthetic crude from recovered bitumen keeps increasing.":" Steamflooding was probably the first EOR method to take advantage of the benefits of horizontal wells.11 ' According to Turta. This application of the CO2 screening criteria is described in the next section. In recent years." Laboratory tests show that steamflooding is an -fficient mechanism for displacing light oils.":" In general. oxygenenriched fireflooding continues to look promising for reservoirs that require large volumes of gas at high flow rates where oxygen is cheaper than air. in-situ combustion projects (310 -4r API) in North and South Dakota. 393 ." Combustion continues to have great promise for a much wider range of fields than the original heavy oil targets. especially in deeper reservoirs. Efforts are continuing to improve the combustion process and to apply it to different types of fields. shallow deposits with high oil saturations and good permeabilities.

394 . However. and 10% recovery was assumed for the immiscible projects. The only price common to both NPC / reports is for $201bbl. this potential for underground CO2 storage would be increased significantly. However. Figs. AND R. in case that becomes necessary to avert global warming.000 oil fields was estimated by using the depth and oil gravity data published in a recent Oil and Gas Journal survey." The basis for the assumed incremental oil recovery percentage and CO2 utilization factors and other details are given in Ref. one of the most efficient CO2 disposal/storage systems would be the combined injection of CO2 into oil reservoirs and into any aquifers in the same or nearby fields." Although this is a very large amount of CO2. 9. However. Table 13 shows that there was an increase in the percentage of projects that were profitable. Therefore. the scale by 1990 if plotted on the 1984 graph of Fig. profitable. The impact of the lower oil prices since 1986 was fmally felt in 1992 when EOR production (except for CO2 flooding) dropped for the first time due to fewer projects. For most EOR methods. as part of their ongoing search for ways to store or dispose of very large amounts of CO2. 33. 8 and 9 show that total EOR production did increase in the early eighties when oil prices were high. it is possible to compare the NPC predictions with actual oil production to date. supercritical CO2 was available (presumably by pipeline from power plants) for each of the fields and/or regions of the world.e.S. the EOR production did not decline until 1992. but they were not nearly as optimistic as those made $ in 1976.33•1°O By including aquifers. It was just assumed that pure.":" However. less than $20 per barrel. Economics were not a part of this initial hypothetical estimate. There is a time lag before the effect is noted. the NPC predictions have merit. Assuming that one-half of the potential new miscible projects would be carried out as more-efficient enhanced secondary operations. our estimate agrees well with other estimates of the quantity of CO2 that could be stored (or disposed ot) in oil reservoirs. Since oil prices were at or below $201bbl for much of the period since 1986. This is in spite of the fact that oil prices were near or less than $201bbl for much of the time since 1986. the $201bbl prediction of 1984 is close to the $1 Olbbl value of 1976. Table 13 shows that the profits from EOR projects did not decline during the recent years of low oil prices." These extensive studies used as much laboratory and field information as possible to predict the EOR production in the future for different ranges of oil prices. The country by country results and other details (including separate sections on the costs of CO2 flooding) are given in Ref.P'" Figure 10 shows that (after the long "incubation" period) CO2 flooding has now exceeded the NPC prediction for oil prices of$20lbbl. SERIGHT SPE 35385 Estimation of the Worldwide Quantity of COl that Could Be Utilized for Oil Recovery. Therefore. The 1976 $201bbl prediction would be off '. Impact of Oil Prices on EOR Major new EOR projects will be started only if they appear. supercritical CO2 has been flowing into the Permian Basin ever since. England. the 1984 predictions benefited from experience gained from the field projects conducted in the interim." We have extended these graphical comparisons and reproduced them here as Figs. and all of the regions were totaled in Table 12 to provide the world totals. TABER. The potential for either miscible or immiscible CO2 flooding for almost 1.S. The pipelines are being extended. the figures confirm that EOR production increases when prices increase and EOR production declines when prices fall. 33. This estimated oil recovery for each country was then totaled by region.D. The large natural CO2 sources were developed. . .i The optimism that came from the much higher oil prices in the late seventies and early eighties was probably very fortunate for the CO2 flooding industry in the U. A CO2 utilization factor of 6 mcf per incremental barrel was assumed for all estimates. The miscible and immiscible screening criteria for CO2 flooding in Table 6 were used to make a rough estimate of the total quantity of CO2 that would be needed to recover oil from qualified oil reservoirs throughout the world. Now. Although not much better than an educated guess with many qualifying numbers. and pipelines were built during this period. i. In general. when the CO2 demand is spread over the several decades that would be required for the hypothetical CO2 flooding projects. from the viewpoint of overall net cost. more open-ended CO2 disposal methods (such as the more-costly deep ocean disposal) will probably be needed if the complex general circulation models of the atmosphere ever prove conclusively that global warming from excess CO2 is underway. but not to the extent predicted. and the quantity sequestered could have a significant impact on reducing the atmospheric CO2 emissions from the world's power plants. This was in response to an increase in the number of projects during this period when prices of even $SOlbbl or more were predicted." The percent of the fields in each country that met the criteria in Table 6 for either miscible or immiscible CO2 flooding was determined and combined with that country's oil reserves to estimate the incremental oil recovery and CO2 requirements. The 1984 predictions were made while oil prices were high (::: 30Ibbl). The inexpensive. and more projects are being started as CO2 flooding efficiencies continue to increase.J. 8-10. Although the rate of increase slowed in 1986 when oil prices dropped precipitously. an average recovery factor of22% OOIP was used. F. This depends on the perception of future oil price. it would reduce the worldwide power plant CO2 emissions into the atmosphere by only a few percent per year. Note that the actual oil production does track predictions of $IOlbbI for 1976 and $20 for 1984 in Figs. The estimate was made for the lEA Greenhouse Gas R & D Programme. The number' of EOR projects has been declining steadily since 1986. presumably because the less-efficient projects were discontinued. Cheltingham. when oil prices were lower. the year that oil prices fell. after several years of low oil prices. the relationship between future oil prices and EOR was a major thrust of the two aforementioned NPC reports. These comparisons were made recently to see how oil prices might affect oil recovery from future CO2 projects. Table 12 shows that about 61 billion metric tons (61 gigatonnes-see conversion factors) of CO2 would be required to produce 206 billion barrels of additional oil. MARTIN. 8 and 9.8 J.

" presented at AICHE Summer Meeting. the results show that there is a wide choice of effective methods that range from miscible recovery of the lightest oil by nitrogen injection to steamflooding and surface mining for heavy oil and tar sands.. 20.: Physical Principles of Oil Production. Ray. Screening criteria and brief descriptions are presented for the major EOR methods. Richardson. W. 16." World Oil (Oct. Beecher.S.P. (1976)." Oil & Gas i. National Petroleum Council. These and other technological advances are expected to improve the process efficiency and cost effectiveness of EOR methods in the future. 1986. Oct. 4. July 11-14. San Francisco. the actual incremental production has been very close to that predicted for $20lbbl in the 1984 NPC report. DOE (Oct. (May 1983) 66--76.: "An Evaluation of Known Remaining Oil Resources in the State of California: Project on Advanced Oil Recovery and the States.K. Acknowledgments The authors express their appreciation to Tommy Morris for assistance in screening many reservoirs for CO2 flooding potential. 2). G.: "Ranking Reservoirs for Carbon Dioxide Flooding Process. 2. 0. Muskat. National Petroleum Council.S.M. U. R. 19. 13..~21990.: Waterflooding. Basnieva." DOElBC/I443 I-I (Vol. and (5) foam for mobility control. 5-8. and those used in lieu of more costly infill drilling. there are many technology advances that should continue to improve the outlook for EOR and lOR.SPE 35385 EOR SCREENING CRITERIA REVISITED 9 References . 2. Conclusions I. it appears that about 80% of the world's reservoirs could quality for some type of CO2 injection to produce incremental oils.M. Taber. 5. Contract No. 5. San Diego CA. and oil production from hundreds of projects are displayed in graphs to show the wide distribution and relative importance of the methods. Contract No.. and Martin.S. C.1. 1994). TX (1994) 299-305." paper SPE 12069 presented at the 1983 Society of Petroleum Engineers Annual Technical Conference and Exhibition. 29-31." paper SPE 28002 presented at the 1994 Society of Petroleum Engineers Centennial Petroleum Engineering Symposium. 9.D. Ray.:Expected Even with the low oil prices. et al. Department of the Interior. Parkinson.M. 1994).l et al. 77.. etc.M. (2) use of horizontal injection as well as production wells. F. especially in CO2 flooding. M.J. D. New York (1928). 8. R.. DE-FG22-89BCI443 I. The impact of oil prices on EOR production in the U. DOE (Oct. in old as well as new fields. LA. These include: (1) 3D seismic-to learn where the target oil is located.: "Increasing the Recovery of Oil by Repressuring. et aI.Future Technical and Economic Improvements .D. Ibid. Polymer flooding continues to show promise. S..". oil gravity. R. Secretary of Energy. Society of Petroleum Engineers. Embid. 3. M. 3-5. et al.: Enhanced Oil Recovery. 1. /3.95-103." DOEIBC/1443 I-I (Vol. However.. AIME. The criteria are based on oil displacement mechanisms and the results ofEOR field projects. there >isoften a wide overlap in choices. was considered by comparing the recent EOR production to that predicted by the NPC reports for various oil prices. New Orleans. II. 17. 1983) 84-88. "Improved Oil Recovery Could Help Ease Energy Shortage. W. especially if projects are started at high oil saturations." Proc. Contract 4. June 23. H. If only depth and API gravity are considered. McGrawHill. Although lower oil prices since 1986 have reduced the number of EOR projects. Ray. No." DOEIBC/1443 I-I (Vol. 18. Washington. .O. and ElmtaIab.: "Technical Screening Guides for the Enhanced Recovery of Oil.: "Lab Evaluation Requires Appropriate Techiques. 1991. 12. DE-FG2289BC14431.D. I). Parkinson. Aug. Industry Advisory Committee to the U. Haynes.C. Steamflooding continues to be the dominant method. R. T. Tulsa. 47-54.. short radius. L: "A Practical Artificial Intelligence Application in EOR Projects. CA.. I. Geffen. Willhite." Discussion in Chapter Viii. et al. 10. DE-FG2289BC14431. Oil & GasJ." SPE Advanced Technology Series (year?) 2." (3) cheaper horizontal injection wells with multilaterals. T.: Enhanced Oil Recovery. to Mailin Seldal and Steve Anderson for help on the figures and to Liz Bustamante for valuable assistance in the \preparation of this manuscript. Tulsa OK. F. et al. U.: "An Evaluation of Known Remaining Oil Resources in the States of New Mexico and Wyoming: Project on Advanced Oil Recovery and the States. "Petroleum Engineering Problems. 7. Geffen. and Bolivar.: "Screening EOR Methods with Fuzzy Logic.O.: "An Evaluation of Known Remaining Oil Resources in the States of Louisiana and Texas: Project on Advanced Oil Recovery and the States. (1984). Aug. Goodlett.: "An Evaluation of Known Remaining Oil Resources in the United States: Project on Advanced Oil Recovery and the States. Industry Advisory Council to the U. The CO2 screening criteria were used to estimate the capacity of the world's oil reservoirs for the storage/disposal of CO2.J. Elemo. The criteria for the effective use of gels to control water flow in both injection and production wells are presented. Rivas. 15.. 715-724. The depth. et al. 4). New York (1949) 709-10. G. Contract No.S. 3).S. Ray.M. If only oil gravity is considered. there is less chemical flooding of the intermediate-gravity oils that are normally waterflooded. Arnold. 395 . "Oil Production to Expect from Known Technology.: "Screening Criteria Decide Best Process for Enhanced Oil Recovery.' Petroleum Development and Technology in 1927. 379-82. Lubbock. R. Bailey." DOEIBC/14431-1 (Vol. Nov. OK..S. 1. 41st Annual Southwestern Petroleum Short Course Association Inc. and now exceeds the predictions for $20 oil in the NPC report." presented at Third International Reservoir Characterization Conference.M. Incremental oil production from CO2 flooding has increased continuously. With low oil prices. 6. DOE (Oct. U. 1.: "AnExpert System for Screening Enhanced Oil Recovery Methods. 19. but hydrocarbon injection and CO2 flooding are increasing. TX (1986). (4) more efficient reservoir (simulation methods.M. 1994).. 14. RE. " paper SPE 26248 presented at the 1993 Society of Petroleum Engineers Computer Conference.: "Comparative Analysis of Successful Application of EOR in Russia and CIS..

et al." presented at the Symposium on Field Studies. 1991) 6.: "Alkaline-Surfactant-Polymer Technology Potential of the Minnelusa Trend. 33. 32." JPT (Dec. L. et al. 1994).G. LA.S. MARTIN.: "Detailed Evaluation of the West Kiehl Alkaline-Surfactant-Polymer Field Project and Its Application to Mature Minnelusa Waterfloods. Tulsa. H." paper SPE 29565 presented at the 1995 SPE Rocky Mountain RegionILowPermeability Reservoir Symposium." Oil & Gas J. Maitin." paper SPE 24118 presented at the 1992 SPEIDOE Symposium on Improved Oil Recovery.s. R. Sydney. J.: "Effect of Oil Composition on Minimum Miscibility Pressure-Part 2: Correlation. Powder River Basin. (1994). and Taber.D. C. ..: "Surfactant Flood Process Design for Loudon. OK.: "The Alkaline-Surfactant-Polymer Combination Flooding and Application to Oilfield for EOR.: "Determination and Prediction of CO2 Minimum Miscibility Pressures..: "Guidelines Help Select Reservoirs for NaHC03 EOR.: "Nitrogen Management at the East Binger Unit Using an Integrated Cryogenic Process. Dallas.K.. Oct.S. 40. 1990. T. J. 52. SPE 35385 21. 1995) 51-? Heller.R. 30. and Metcalfe. Moritis. et al. S. Orr." annual report (January to December 1993).M. (Sept. S. 39. Y. lr. 1987. 5-8. 479-91. Chatzis. and Peru." JPT (April 1992) 396-400." SPEPF (Nov. K.: Polymer-Improved Oil Recovery.3. et al. . l 43. DOE.S. AND R. 345-404. 41. 1.1. DOE (Dec.. 1988. TABER.1. Petrol. J. 45. Koinis. 37. New Orleans. Dallas. 1.27773 presented at the 1994 SPEIDOE Symposium on Improved Oil Recovery. L.: "Wetting Alteration by Brine and Crude Oil: from Contact Angles to Cores. 5). Buckley.S. U. and Hongfu. SERIGHT No.S. Vienna (1995) 1. (April 14.S.S. 187-193. 14.C.: "EOR Dips in U.: "Correlation of Capillary Number Relationships for Sandstone.R." SPEI (April 1983) Maerker. 1991.: "Improved EOR by Use of Chemicals in Combination.S.F.: "Gel Placement in Fractured Systems.: "Performance Analysis of Several Polyacrylamide Floods in North German Oil Fields. F. Seright. Manning. 1990) 1476-84. II." JPT(lan. Tulsa. Tulsa." first annual report.: "Placement of Gels to Modify Injection Profiles." SPERE (Nov.S. 42. W. 56. 1.: "Laboratory Investigation of Combination of Alkali/Surfactant/Polymer Technology for Daqing EOR. Novosad. No. 1980) 160-68. Gibbons. 60. 1983). (Sept. International Energy Agency Collaborative Research Program on EOR. DE-AC22-92BCI4880. Contract No.R.L.W. 1993) 2-72.. DOElBC/14860-5. Contract No.J.1.M. DOE (Oct.3. C-F. 47. Bousseau.S. Holm.1. 22-25. Reppert. 1995). Sept. 38. 29. M. Illinois." International Energy Agency Collaborative Research Program on EOR. 27.S. 1990..O. V. April 22-24.B. Hsu. Fielding. F. C..: "In-Depth Drive Fluid Diversion Using and Evolution of Colloidal Dispersion Gels and New Bulk Gels: An Operational Case History of North Rainbow Ranch Unit.A." paper SPE 30765 presented at the 1995 SPE Annual Technical Conference and Exhibition. Sorbie. W.S. U. DOElBC/14880-5. Oil Recovery Europe Symp. Morrow.S." SPERE (August 1995) 10. (July 17." DOE report DOElBC/I0327-19. Wyatt.P. lEA Greenhouse Gas R&D Programme. U. March 2-5. U. 1994) 51-79.: "Influence of Reservoir Depth on Enhanced Oil Recovery by CO2 Flooding. 57.S. OK.. 31. New Orleans. Copenhagen. 48." paper SPEIDOE 20218 presented at the 7th Joint Symposium for Enhanced Oil Recovery. I. 241-248. April 17-20. Cheltenham. 53.: "Environmental Improvements and Better Economics in EOR Operations. F.A: "Effect of Oil Composition on Miscible-Type Displacement by Carbon Dioxide." MS Thesis. 23. and Gale. 187-193. Reimer. Midland." ." Presented at the Symposium "EOR by Gas Injection. 55." paper SPEIDOE 20219 presented at the 7th Joint Symposium for Enhanced Oil Recovery. Batycky. Seright. 1987) 1. lr. 1." Oil & Gas J. D." paper SPE . and Legrand. April 17-20. Science (May II.: "A Technical Survey of Polymer Flooding Projects.: "The Application of New Technology in the Judy Creek Miscible Project. Hagedorn.E. W. and Taber." paper SPE 25169 presented at the SPE International Symposium on Oilfield Chemistry. Taber.W. M. 8th EAPG Impr. April 22-25.R.S." paper SPE 15591 presented at the 1986 SPE Annual Technical Conference and Exhibition. 1. and Liu. and Taber. 44. King Fahd Univ. N." Oil & Gas J. F. 1. 148." SPERE (August 1995) 10. New Mexico. U. et al. F. and Costain. Inc.: "The Use of Flue Gas for the Enhanced Recovery of Oil. R. T. R. R. DE-FG22-89BC 14431.D. 26.P. R.. 28. 1982) 87-98. 1989) 53-57. R. and Oklahoma: Project on Advanced Oil Recovery and the States. Stalkup. Blackie. M.: "A Study of Technical Feasibility for the Utilization of CO2 for Enhanced Oil Recovery. 36. 34. V.R. B. Orr. P. 1993.: Miscible Displacement. Baviere et al. Sibbald. Yellig.: "Wettability and Its Effect on Oil Recovery. Z. DOE (March. R.1. DE-FG22-89BCI443 I. (Oct.M. Experience Speed CO2 Flood Design.D." SPEI (Feb. And Mundorf. 1995) 26-29. 26.: "Permeability Reduction by a XanthanlCr(III) System in Porous Media. March 13-14. F. March 20-22." DOEIBC/1443 I-I (Vol. 134-204. G.: "Technology. 49. 1994).E. CRC Press. Monograph Series. P. Denver.P. Pitts. DOE (Nov. 24. Seright." SPERE (Aug.1." In Situ (1990) 14 (4). D. 59. 1994). 23. and Josendal. Min. 58.M.. L. et al. 1. U. Oct. Yang.: "Design of AlkalineiSurfactant/Polymer Enhanced Oil Recovery Scheme for a Saudi Arabian Limestone Reservoir.: "Improved Techniques for Fluid Diversion in Oil Recovery Processes. 373. lr..3." second annual report (DOElBC/14880-1 0)." Proc. R. and Fox." Oil & GasJ. The Utilization of Carbon Dioxide from Fossel Fuel Fired Power Stations.K..: "Increased Rate of EOR Brightens Outlook. L.M. and Nagra. R. Taber. Texas. Lorenz. 22. 25.: "Methodology for the Specificiation of Solvent Blends for Miscible Enriched-Gas Drives. Leonard. Australia. No.83-114.: "Component Partitioning in COiCrude Oil Systems: Effects of Oil Compositions on f:0 Displacement Performance. 54. TX. 1993) 299-304. England (1994) Appendix B. 50..: "Improved Techniques for Fluid Diversion in Oil Recovery. G. "C02 Activity Picks up in W.: "Second Ripley Surfactant Flood Pilot Test. C. Martin. K..61-64." SPERE (Nov. F. Society of Petroleum Engineers.H. Surkalo. Tulsa. 35. F. 1.10 J.: "Carbon Dioxide Flooding.1.K. Ray. 51. DOE (Oct." paper SPEIDOE 17332 presented at the 1988 SPEIDOE Enhanced Oil Recovery Symposium.: "An Evaluation of Known Remaining Oil Resources in the States of Kansas. New York (1983). 183-190. 1984) 563-69.. and Orr. Eme. 1995) 5164.R. 30. (Sept.: "Use of Carbon Dioxide in Enhanced Oil Recovery. April 22-25." Oil & Gas J.." paper SPE 15001 presented at the 1986 Permian Basin Oil and Gas Recovery Conference. 46. And Morrow.. Huabin. Taber. N. 396 . Tulsa. No. K. Sept. Emmons. and Silva. 1986)71-101. Shutang. Seright. Denmark. Hejri. but Remains a Significant Factor.

96. (Jan. Pet." Oil & Gas J." Oil & Gas J.C. A.E. April 21-22. Tulsa.. 8. (Oct.: "Steamflooding as an Alternative Eor Process for Light Oil Reservoirs. M.H. and Pedron. China. 95. 97. (Jan. 4-7. OK (1994). Sanders. P. 50. and Olsen. 2-5.S.V. SPE. 25-28. Farouq Ali. Chu.: "A Study of Fireflood Field Projects.: "Evaluation ofa Pilot Polymer Flood in the Mannul Field." Oil & Gas J.J.. San Antonio.: "Enhanced Recovery Engineering-combustion Processes." paper SPE 26991 presented at the 1994 SPE III Latin American & Caribbean Petroleum Engineering Conference. 80. R. 138. Tulsa.: "Light-Oil Steamflooding: an Emerging Technology. Present and Future Application.A. B. Blevins. TX. 67. 62. "Thermal System Ups Heavy Oil Flow. D. 86.: "Horizontal Injection and Production Wells for EOR or Waterflooding. 1. 18.: "Gel Conformance Treatments Increase Oil Production in Wyoming. 1977) 111-119.K. Tulsa. and Moore. 71. 1992) 40-45.: "Commercial Test of Polymer Flooding in Daqing Oil Field. 6. Venezuela." J. D. No.S. 1996) 23." Oil & Gas J. Part II: Prospective Study towards the Year 2020 and Beyond.C.R and Yannimaras. Argentina. 1995) 21-22. "Suncor Set to Try SAGD at Burnt Lake Heavy Oil Property.: "Operational Advances from Field Application of Short-Radius Horizontal Drilling in the Yates Field Unit. P. A. and Seright. 1." SPERE (Nov. 79. Say Experts. E. 47. Putt." paper SPE 10319 presented at the 1981 SPE Annual T~chnical Conference and Exhibition. B.S." Science (Dec. 12." paper SPE 23952 presented at the 1992 SPEIDOE Symposium on Enhanced Oil Recovery. OK. Caracas. Ribeiro. Penwell. OK (1983). R.: "Numerical Simulation for Planning and Evaluation of Polymer Flood Process: A Field Performance Analysis. (April 1972) 477-485. Palmgren. 1996) 271. April 22-24.. 84. Oct. 1995) 4. V. 98." J.M.: Thermal Recovery Methods. Pennwell. European Commission Joint Research Centre. A. 11. Tulsa. (Dec. Gangle. B. Tulsa." Science. (Feb. 1991) 406-412. New Orleans. April 17-20. April 27-29. 15. Seright. and Heemskerk. 14-17. B. "Royalty Incentives Spawn Alberta Oilsands Plans. 74. 88. Lighter Crudes Eligible?" Improved Recovery Week (Dec. 64. C and Renard G.R.473 (ISC 4) presented at the 1994 DOEINIPER Symposium on In Situ 82. J. S.H. 8th EAPG Impr.O. Oman." SPE paper 29902 presented at the 1995 SPE International Meeting on Petroleum Engineering..470 (ISC 1) presented at the 1994 DOFlNIPER Symposium on In Situ Combustion Practices-Past." paper SPE 24612 presented at the 1992 SPE Annual Technical Conference and Exhibition. Present and Future Application. Italy (Jan. H. 8. Beijing. Koning. Shahani. 87. 63." Oil & GasJ.K.: "In Situ Combustion-from Pilot to Commercial Application. 92. Summers. Tulsa. Elk Hills.256-264.G.. 1. April 21-22. 1994 Update. Tech. Petzet." J." Proc.1565-1567." SPE paper 17631 presented at the 1988 SPE International Meeting on Petroleum Engineering. "Worldwide production falls as market plays its wild cards. D. "Redeeming Features of In Situ Combustion.C.S. 1995). S. Sept. and Curtis. 100. Pet.472 (ISC 3) presented at the 1994 DOEINIPER Symposium on In Situ Combustion Practices-Past. 48. Richardson.E.: Thermal Recovery. 85. 1992) 315-320. Ispra. E. OK." paper SPE 18092 presented at the 1988 SPE Annual Technical Conference and Exhibition. 1979) 1332-1342. L. D.: "Screening Criteria for the Application of Steam Injection and Horizontal Wells. 4. V. Maitin. Sarathi. Tech. and Gunardson. Alberta.: "Light Oil Steamflooding: A Laboratory Study of Oil Recovery from Oil-Wet and Water-Wet Porous Media from a 2-D Physical Model. The Canadian Heavy Oil Association. 68. 66. 73. Present and Future Application.. R. Duerksen. Taber. and Ault.: "Oilfield Application of Colloidal Silica Gel..L. E.J.471 (ISC 2) presented at the 1994 DOEINIPER Symposium on In Situ Combustion Practices-Past..E.: "A Current Appraisal of In Situ Combustion Field Tests. S." paper SPEIDOE 27792 presented at the 1994 SPEIDOE Symposium on Improved Oil Recovery.M. G. S.H.J. 27. Houston. 20. Sydansk. 83. and Bennett. Chambers.. Monograph Series. 25. 25-28. Washington. 77. K. Chu. Canada (1991). Nov. RH. "Light Oil Air Injection to Revolutionize lOR. Fassihi. and Lane. and Sohn. Oct. Combustion Practices-Past.T. K. 91. Farouq Ali." J. Farouq Ali. Daboul. 5-7. J." paper SPE 28502 presented at the 1994 SPE Annual Technical Conference and Exhibition. 94." Improved Recovery Week (Dec. 1. II LAPEC. et al. and Meldau. Tech.G. Duncan. M. March 8-11. New Orleans. 93. TX (1982). April 21-22. 1982) 19-36. 15. Hanzlik.M.J.: "Case History and Appraisal of the Medicine Pole Hills Unit Air Injection Project. 78. et a1.: "State-of-the-Art Review of Fireflood Field Projects. California. "It's Official: First Glimmer of Greenhouse Warming Seen. 81. "Current Steamflood Technology. Odorisio. (Oct. 1995) 270. Mentzer. C.: "Successful Gas Shutoff with Polymer Gel Using Temperature Modeling and Selective Placement in the Prudhoe Bay Field. J. "Syncrude Operation Claims Production Mark from Oil Sands. and Liang.: "Horizontal Projects Buoy Williston Recovery.J. No.W. Wang Demin et al. Turta. 397 . No.: "A Survey ofField Applications of Gel Treatments for Water Shutoff. D. TX." Institute for Prospective Technological Studies. RF.K.: "DOE Cost-Shared In Situ Combustion Projects Revisited. Jurinak. Khalbad." paper DOEINIPER 609. 76. And Moss. G." paper DOFlNIPER 609. Pet. Olsen. Oil Recovery Europe Symp.6.: Steamflood Reservoir Management.: "Commercial Polymer Injection in the Courtenay Field. Vienna (1995) 2." SPERE (Aug.D." Improved Recovery Week (Dec. (Jan. 137. and Henry." paper DOEINIPER 609." paper SPE 23692 presented at the 1992 SPE Latin American Petroleum Engineering Conference." Oil & Gas J (Jan." JPT (July 1984) 1115. Worldwide Production Report: OGJ Special (Dec. F.M. China. White. 9. G. 89. The Canadian Heavy Oil Reservoir Handbook. Tulsa.. "Syncrude Production Keeps Climbing." paper DOEINIPER 609. and Stemler. 90. Nov.. April 21-22. Calgary. G. Tianjin. 72. M.J. 1-4.. "1995 the Wannest Year? Yes and No.. T. Prats. Oct. Bazin. 1. 75." World Oil (Jan. (Jan. 1993) 37-87. 1995) 4. Kumar. 70.. Sept. Hong. Present and Future Application.. J. Pet. 99. 65. Tech. 1. W. C." SPE paper 28601 presented at the 1994 SPE Annual Technical Conference and Exhibition.:" Light-Oil Steamdrive Pilot Test at NPR-l. 1995) 30.: "Carbon Dioxide Disposal and Storage Technologies: Part I: Current Status.SPE 35385 EOR SCREENING CRITERIA REVISITED 11 61. Buenos Aires. 69. Tulsa. 1995) 46. "Oxygen Enriched Fireflooding. 1966) 65. P. 1995) 4.

MARTIN.609344* mile x 6.290304* sq ft x 9.0283 m' 1000 ftl (mscf or mcf or Mcf or MCF) = 28.144* yard *conversion factor is exact.863640 scf 6. oil industry = 1 atm and 60°F (1.5/(131. 14.150 lbs) = 0.S.8 E+03 E+03 E + 05 E . = 0.2 lbs) 1 ton U. gallons = 0.001868 glcc or 1.S. 1 Gt (gigatonne) = 1 billion metric tons 1 barrel oil (35° API) = 0. = 907 kg (1 kg '" 2.S.000 = lb. = 6 mcflbbl = 1.950 kg/m') CO2 utilization factor used for Table 12 tonne of CO/tonne of incremental oil.J. AND R.1166 lb/ff 17.7 psia) (may vary in some U.01 E-03 E+Ol E-02 E .3 m' Standard conditions in U.448222 x 4.048* ft 2. TABER.5 + °API) Pa-s em' m3 °C m cm N kg 3.8 .4516* sq in x x 9.013250* 1.907 metric ton or tonne 1 tonne COl 18.01 E+OO E+OO E .831685 CF . m' 141.54* in x Ibf x 4. = 2.0l3 bar.233489 1.01 m2 m3 g/cm' Pa . F. = 0.6° C and 1 atm.87 ftl of CO2 at 60°F (1 atm) = (weighs) 1 ton U.535924 Ibm x 1.589873 1.894757 psi x 2.31324 tonne CO/tonne oil (35° API) (Crude oil density typically ranges from 0.S. (2. states) COl density at standard conditions kg/m' = 0.S. SERIGHT SPE 35385 Metric Conversion Factors acre acre-ft °API abn bbl cp cu in cu ft of x x x x x x x x x 4.S.12 J.638706 2.16895 tonne 1 mscflbbl = 0.32)/1.S.0* 1.S.904 scf@ 15.D.000 1 ton U.16982 ton U.879 398 .01 E+OO E+OO E-02 E+OO E-02 E .046873 1.95 g/cm' or 800 .0.159 m' 1 cubic foot = 0. km kPa std cm2 m2 m '" Additional Conversion Factors Useful for Reading COl Flooding Literature (Some factors are rounded for convenience and quick estimates) 1 barrel (bbl) = 42 U.

Table Number 4 (4) 5 6 Improved Waterflooding Methods alcohol miscible solvent flooding micellar/polymer alkaline flooding alkaline/surfactant/polymer polymer flooding gels for water shutoff microbial injection (ASP) flooding (surfactant) flooding 7 7 7 low interfacial waterflooding 8 9 Thermal Methods in situ combustion standard forward combustion wet combustion 02-enriched combustion reverse combustion steam and hot water injection hot water flooding steam stimulation steamflooding II 10 Surface Mining and Extraction 399 .SPE 35385 EOR SCREENING CRITERIA REVISITED 13 Table I. Current and Past Enhanced Recovery Methods Gas (and Hydrocarbon Solvent) Methods "inert" gas injection nitrogen injection flue gas injection hydrocarbon gas (and liquid) injection including: high pressure gas drive enriched gas drive miscible solvent (LPG or propane) flooding carbon dioxide flooding Detailed .

oil-bearing ore Interracial Tension Reduction Processes micellar/polymer flooding (sometimes included in miscible-type flooding above) alkaline/surfactant/polymer flooding Viscosity Reduction (Or Oil) or Viscosity Increase (Or Driving Fluid) plus Pressure steamflooding fireflooding polymer flooding enhanced gravity drainage by gas or steam injection 400 . Classification of Current Enhanced Recovery Methods (Excluding Gel Treatments) by the Main Mechanism of Oil Displacement Solvent Extraction and/or "Miscible-Type" Processes nitrogen and flue gas hydrocarbon miscible methods carbon dioxide flooding "solvent" extraction of mined. MARTIN.J.14 J. AND R. SERIGHT SPE 35385 Table 2. TABER.D. F.S.

.!= 0 0 '0 o.a i :i ~c -1 J s g.c:: U . ...ii...a ~ E.. . &." " " " " . .~ ...So 8.=tJ ii... 2i .. ii.an ~"O ~ . .~ t: - ~ ~ ~I f"'I ~ I:: ~ "='" 0 0 c II eo.~ :> 0 Q.D . ~ ~ ~ 0 -8 ~ii Z -ss ! II a u cj :i cj :i G)"O =G) > Ecj u ! II DII " 0 t G) 0 N "0 " 0 " ~ G) G) j~ e 'CI :i a a ~[ ~ ~c2 8] 3t: ~c2 3t: ] e~ a <11- ~8 . 0 0 an 0 0 "l.a II 8. o cj ~~ ~~ "" eo..'f.. 0 0 0 ~~ 1 0 0 0 M 0 0 an 1 " V cj :i :i 3 =. ~z ·i~ 0 = 0 .5 l·5 ... ~-I .0 ~[ ~ = . I'Il:aS :i cj :i ~ a e Q. o cue. "0 QI) _ . " o '0 ~I .~G) " e- ~ . QI) [<1 ::EQ.0 " ~ z S.1 .= cj 0!8.8 '8 ~a ~~ iu iu OG) G)tU == OG) G)tU G)tU == 30 == ....~ cj :i cj :i :i oG) 0- cj :i cj ~I 0 0 1 ~I 1 0 0 k Q..!!"o 8. r.~ "'= ~l'Ills ~ 1 ~ ~·u:Sl " y~ 0·'. an .. 0 0 <II <II ·c ~ y 0 fn= ~ .a S·~]C ]o. 3--. ...eo.. S 0 U <II . cj - <II cj " 0 0 00..9 8.!.~I ~I .-ct:: l5 ei]' G)G) e- .i ii G)<II <11- .!. .~ ~8 .a a :i r..It! WI 1·1 1-[ ~~ ~~ ·ll·e j . :f!....t t z"" i. .s: -g ~ "0 e C "ti ::E 0 UJ ." . .§~ ~ l·5!S.<10-1"" r-- oQ. 1 ~-e"O ":":]E! if on ~~ 0 " 0 an 8 ~ cj :i u "" I'll ef- ~ = .5 Q.:: I =:a =-8 ....9> :::~Q...fa y.!! f"'I ~ ~B "I'll 0 ~ an ~ ~ ~ I'll G) G) 0 oJ.G) QI) QI) .... _.9 G)- cj ~ ~ :i <II ~QI) cj :i <II ~QI) cj :i :i QI) • ~I .0 19 e~ .. .. ._ . " i II . :E a . an I'll I'll I'll ! 3.:-a -.J ..N ·-U ·-U ::c ::c ~y ii.. " M'G)f " ani '0 \} \} 0 0 ~I .. 8.U.It! ~ ~ Q yG) :a .. !:EU "1 0 " ~ II f"'I 0 .5 e §. <II <II G) .D ~ ~ U 0 0 ~ ::E G) I'll "='" an '0 00 CI'I I 401 I ...] G) "'~ . -I ~ ~ ~ 1 f"'I ~I 1 V '0 1 1 0 0 0 ~ V an " V 0 ~ 0 an 0 0 ! V ~ vi 0 0 ~ = 0 y :s! ~ \} CI'I C'o e- '0- ~G) ~ an f"'I i N ~ ..B..:. o o . 0 0 " "='" o.9 a WI .r. f"'I eo..Cj .. ~ i: G) <II > <G) Q.-.. 00.iI II ii. " "l...! ..SPE35385 EOR SCREENING CRITERIA REVISITED 15 ~ e- G) f- S- 8..!! _g ~~z Q 0• .:: o u ~~ t 8 :f!..~ .. ~Oy ~ -e ~ .i ::C<ll ii... OG) .5 s !:E~ "'! " u 1 . .

The non-hydrocarbon gases must be separated from the saleable produced gas. For miscible.4 cp High percentage of light hydrocarbons Recommended Range of Current Projects 38-54 0 \ API (miscible) 0. F. Problems Viscous fingering results in poor vertical and horizontal sweep efficiency. SERIGHT --_. or immiscible enhanced gravity drainage. AND R. Because of their low cost. MARTIN._------- SPE 35385 Table 4. Mechanisms Nitrogen and flue gas flooding recovery oil by: -vaporizing the lighter components of the crude oil and generating miscibility if the pressure is high enough -providing a gas drive where a significant portion of the reservoir volume is filled with low-cost gases --enhancing gravity drainage in dipping reservoirs (miscible or immiscible). Nitrogen and Flue Gas Flooding Description Nitrogen and flue gas are oil recovery methods which use these inexpensive non-hydrocarbon gases to displace oil in systems that may be either miscible or immiscible depending on the pressure and oil composition (see Table 3 for immiscible criteria). a dipping reservoir may be crucial to the success of the project.3 402 . TABER. Injection of flue gas has caused corrosion problems in the past. TECHNICAL SCREENING GUIDES Crude Oil Gravity Viscosity Composition Reservoir Oil Saturation Type of Formation Net Thickness Average Permeability Depth Temperature Limitations Developed miscibility can only be achieved with light oils and at very high pressures. therefore.07-0.16 J. large volumes of these gases may be injected. A steeply dipping reservoir is desired to permit gravity stabilization of the displacement which has an unfavorable mobility ratio.500 ft 59-80% VP <0.S. >40%PV Sandstone or carbonate with few fractures and high permeability streaks Relatively thin unless formation is dipping Not critical >6000 ft Not critical 10. nitrogen is being injected into large successful projects that formerly used flue gas. Nitrogen or flue gas are also considered for use as chase gases in hydrocarbon-miscible and CO2 floods.J.000-18. deep reservoirs are needed. At present.D.

mostly methane) and possibly water. called Enriched (Condensing) Gas Drive.900 ft 30-98% VP 24-54 °API (miscible) 0. called High Pressure (Vaporizing) Gas Drive. consists of injecting a 10-20% PV slug of natural gas that is enriched with ethane through hexane (C2 to C6). and most common method. Immiscible criteria are given in Table 3. the viscosity of the oil. A steeply dipping formation is very desirable to permit some gravity stabilization of the displacement which normally has an unfavorable mobility ratio. consists of injecting lean gas at high pressure to vaporize C2--C6 components from the crude oil being displaced. followed by natural gas or gas and water. Crude Oil Gravity Viscosity Composition Reservoir Oil Saturation Type of Formation TECHNICAL SCREENING GUIDES Range of Current Projects Recommended >23°API <3 cp High percentage of light hydrocarbons >30%PV Sandstone or carbonate with a minimum of fractures and high permeability streaks Relatively thin unless formation is dipping Not critical if uniform >4000 ft Not critical 4.000-5. Three different methods have been used. Problems Viscous fingering results in poor vertical and horizontal sweep efficiency.000 psi for the High Pressure Gas Drive. gas displacement. followed by lean gas (dry.SPE 35385 EOR SCREENING CRITERIA REVISITED 17 Table 5. 403 . Mechanisms Hydrocarbon -generating -increasing -decreasing -immiscible miscible flooding recovers crude oil by: miscibility (in the condensing and vaporizing gas drive). Hydrocarbon Miscible Flooding Description Hydrocarbon miscible flooding consists of injecting light hydrocarbons through the reservoir to form a miscible flood. The required pressure ranges from about 1200 psi for the LPG process to 4.3 cp Net Thickness Average Permeability Depth Temperature Limitations The minimum depth is set by the pressure needed to maintain the generated miscibility. The third. Solvent may be trapped and not recovered in the LPG method. Large quantities of expensive products are required. depending on the oil. The enriching components are transferred from the gas to the oil.040-15. especially enhanced gravity drainage with the right reservoir conditions.04-2. The first-contact miscible method used about 5% PV slug of liquefied petroleum gas (LPG) such as propane. the oil volume (swelling). A second method.

Although CO2 is not first-contact miscible with the crude oil.S.9° <13° 1.. Carbon Dioxide Flooding Description Carbon dioxide flooding is carried out by injecting large quantities of CO2 (30% or more of the hydrocarbon PV) into the reservoir.9° Flooding: Depth must be greater Feet 2. -lowering the viscosity of the oil (much more effectively than N2 or CH4).219 <2ZO Fails miscible..300 4. Depth vs Oil Gravity Screening Criteria I.006 1.800 3._----_. for C02 Flooding Meters 762 853 1.9° 28-31.18 --_---------- . and. -lowering the interfacial tension between the oil and the CO2 -oil phase in the near-miscible regions. For Immiscible COz Flooding (Lower Oil Recovery) 13-21. and about 20 bar above the CO2 critical pressure for the immiscible floods at the shallow depths. SERIGHT ----. \ Mechanisms CO2 recovers crude oil by: -swelling the crude oil (C02 is very soluble in high-gravity oils).000 than: '.500 2..----------- ---_--. F.J. if the pressure is high enough.------_ SPE 35385 Table 6. Immiscible displacements are less effective.D. TABER. 404 . the CO2 extracts the light-to-intermediate components from the oil. -generation of miscibility when pressure is high enough (see below). For COz-Miscible Oil Gravity °API >40° 32-39.9° 22-27._--_ . screen for immiscible 2.) -See Table 3 for other screening criteria. See text. MARTIN. develops miscibility to displace the crude oil from the reservoir._-- J. Reservoir temperature is assumed from depth.800 549 All oil reservoirs fail at any depth <1800 feet (549 meters}-all reservoirs fail screening criteria for either miscible or immiscible flooding with supercritical CO2_ (The arbitrary criteria have been selected to provide a safety margin of approximately 500 feet (152 meters) above typical reservoir fracture depth for the required miscibility (MMP) pressures. AND R. but they recover oil better than waterflooding (see below and Table 3 for immiscible criteria).._----- .

High amounts of anhydrite. The polymer concentration often ranges from 500-2000 mg/L. -solubilization of oil in some micellar systems. Alkaline Surfactant Polymer (ASP) and Alkaline Flooding Description Classical Micellar/Polymer Flooding consists of injecting a slug that contains water. Mechanisms All surfactant and alkaline flooding methods recover oil by: -lowering the interfacial tension between oil and water. The surfactant slug is followed by polymer-thickened water. High adsorption of surfactant. For Alkaline Flooding much of the injection water was treated with low concentrations of the alkaline agent and the surfactants were generated February 1. >35%PV Sandstones preferred Not critical >10md <about 9000 ft (see Temperature) <200°F Recommended >20° API <35 cp Light intermediates are desirable for micellar/polymer. Alkaline/Surfactant/Polymer (ASP) Flooding is similar except that much of the surfactant is replaced by low cost alkali so the slugs can \ be much larger but the overall cost is lower and polymer is usually incorporated in the larger. Micellar/polymer.SPE 35385 EOR SCREENING CRITERIA REVISITED 19 Table 7. TECHNICAL SCREENING GUIDES Crude Oil Gravity Viscosity Composition Reservoir Oil Saturation Type of Formation Net Thickness Average Permeability Depth Temperature Limitations An areal sweep of more than 50% on waterflood is desired. especially in the alkaline methods. The size of the slug is often 5-15% PV for a high surfactant concentration system and 15-50% PV for low concentrations. Problems Complex and expensive systems. dilute slug. Available systems provide optimum behavior over a narrow set of conditions. and the volume of polymer solution injected may be 50% PV or more. 405 . surfactant. -wettability alteration (in the alkaline methods). or clays are undesirable. 1996 in situ by interaction with oil and rock. and possibly a hydrocarbon (oil). Degradation of chemicals at high temperature. usually a cosolvent (alcohol). gypsum. -mobility enhancement. formation water chlorides should be <20. Relatively homogeneous formation is preferred. Interactions between surfactant and polymer.000 ppm and divalent ions (Ca ++ and Mg") <500 ppm. -emulsification of oil and water. electrolyte (salt). With commercially available surfactants. Possibility of chromatographic separation of chemicals in reservoir. Organic acids needed to achieve lower interfacial tensions with alkaline methods.

Polymer Flooding Description The objective of polymer flooding is to provide better displacement and volumetric sweep efficiencies during a waterflood. . SERIGHT SPE 35385 Table 8. we recommend that the reservoir characteristics and polymer-flood features be closer to those of the four successful projects listed below. When identifying polymer-flooding candidates. In polymer flooding. properly sized treatments may require 25 to 60% reservoir PV.1 0. MARTIN.87 71 10 1.ec ts te I Marmul 114 115 15 -92 1 1.J. P roperties Property Oil/water viscosity ratio at reservoir temperature Reservoir temperature.500 162 -13 -1. OF Permeability.4 120 0.9 1. For the others. TABER.D. F. *IOR over pnmary production for this case only.000 373 25* 1.20 J.000 271 II 1980s median (171 projects) 9.57 155 -230. Polymer concentrations from 250 to 2000 mgIL are used. certain high-molecular-weight polymers (typically polyacrylamide or xanthan) are dissolved in the injection water to decrease water mobility. Projected bbl oil/ac-ft 27 461 499 . incremental over waterflooding .075 76 3 460 25 4. lOR IS 406 . % OOIP Projected bbl oil/lb-polymer 0fP oiymer FI00dimg F· Id P ro . TECHNICAL SCREENING GUIDES A very broad screening guide is given in Table 3.5 4 1.96 Daqing 15 113 0.S.2 Oerrel 39 136 2 81. darcys % OOIP present at startup WOR at startup HPAM concentration.4 Courtenay 50 86 2 78 8 900 520 30 0. AND R. ppm Ibs polymer/ac-ft Projected lOR.

using zone isolation). In un fractured wells. 3. 5. 3.e. Water breakthrough occurs much earlier than expected (i.g. 2. Reduced injectivity (caused by the gel) can be tolerated.SPE 35385 EOR SCREENING CRITERIA REVISITED 21 Table 9. c. 4. b. Treatments can be effective in both injection and production wells. Recovery calculations indicate that considerable mobile oil remains that could be recovered more cost-effectively if a blocking agent could be realistically placed in the proper location.. High WOR values are observed. B. c. The source of the excess water production is identified (e. Selection Criteria for Gel Treatments in Production Wells 1. Recovery calculations indicate that considerable mobile oil remains that could be recovered more cost-effectively if a blocking agent could be realistically placed in the proper location. from standard calculations or simulations or from comparison with the performance of other patterns in the field). A. Poor injection profiles must be correlatable from well to well.g. logs. Gel can be placed in the offending channel without damaging oil zones (e. or tracers). The well to be treated is in good mechanical condition.. 5. b. If barriers to crossflow do not exist. They can also be effective in blocking water zones in unfractured wells where impermeable barriers prevent crossflow. no vertical fractures). WOR values at offset producers are much higher than expected. a. a. Gel Treatments Description Gel treatments are applied most effectively when fractures cause severe channeling or excess water or gas production. Effective barriers to crossflow must exist (very low kjkh' no flow behind pipe. 407 . The gelant can be placed without damaging oil zones (e. The candidate well exhibits high productivity. 4.g.. using zone isolation). Selection Criteria for Gel Treatments in Injection Wells 1. Reservoir and production data indicates low sweep efficiency during waterflooding.. Gel treatments are also useful as an alternative to cement squeezes when plugging narrow channels behind pipe. then interwell tracers must show very rapid transit times (probably indicating that fractures or formation parting cause the channeling problem). but only if hydrocarbon-productive zones are protected during gel placement. using profiles. 2.

22 J. One of the variations of this technique is a combination of'forward combustion and waterflooding (COFCA W). Produced flue gases can present environmental problems. F. paraffinic oils. TABER. >50%PV Sand or sandstone with high porosity >10 ft >50md >11. Mechanisms In-situ combustion recovers crude oil by: -the application of heat which is transferred downstream by conduction and convection.500 ft >1000 F 85-4. and pipe failures in the producing wells as a result of the very high temperatures. no successful field trials have been completed for reverse combustion.000 cp <500 cp Some asphaltic components to aid coke deposition 408 . serious oil-water emulsions.S. is difficult to control. however. requiring large capital investment. If excessive coke is deposited. the rate of advance of the combustion zone will be slow. MARTIN.J. Problems Adverse mobility ratio.300 ft 100--221 of 62-94 Recommended 10--27 API 0 Range of Current Projects 10-400 API (miscible) 6-5. The most common techniques is forward combustion in which the reservoir is ignited in an injection well. and air injection is then switched to adjacent wells.000md 400--11. Oil saturation and porosity must be high to minimize heat loss to rock. In-Situ Combustion Description In-situ combustion or fireflooding involves starting a fire in the reservoir and injecting air to sustain the burning of some of the crude oil. and the quantity of air required to sustain combustion will be high. and air is injected to propagate the combustion front away from the well. Complex process. SERIGHT SPE 35385 Table 10. thus lowering the viscosity of the crude -the products of steam distillation and thermal crackling which are carried forward to mix with and upgrade the crude -burning coke that is produced from the heavy ends of the crude oil -the pressure supplied to the reservoir by injected air TECHNICAL SCREENING GUIDES Crude Oil Gravity Viscosity Composition Reservoir Oil Saturation Type of Formation Net Thickness Average Permeability Depth Temperature Limitations If sufficient coke is not deposited from the oil being burned. Operational problems such as severe corrosion caused by low pH hot water. increased sand production. deposition of carbon or wax. This prevents the application for high-gravity. Process tends to sweep through upper part of reservoir so that sweep efficiency is poor in thick formation. A second technique is reverse combustion in which a fire is started in a well that will eventually become a producing well. the combustion process will not be sustained.D. AND R.

Mechanisms Steam recovers crude oil by: -heating the crude oil and reducing its viscosity -supplying the pressure to drive oil to the producing well TECHNICAL SCREENING GUIDES Crude Oil Gravity Viscosity Composition Reservoir Oil Saturation Type of Formation Net Thickness Average Permeability Transmissibility Depth Temperature Limitations Oil saturations must be quite high and the pay zone should be more than 20 feet thick to minimize heat losses to adjacent formations. 409 . Recommended 10-25° API <50. high-permeability sandstones or unconsolidated sands. Because of excess heat losses in the wellbore. steamflooded reservoirs should be as shallow as possible as long as pressure for sufficient injection rates can be maintained. Steamflooding is not normally used in carbonate reservoirs. the cost per incremental barrel of oil is high.000 cp 35-90% VP 63-10. Normal practice is to precede and accompany the steam drive by a cyclic steam stimulation of the producing wells (called huff and puff).000 md 150-4.500 ft 60-280 of Lighter.SPE 35385 EOR SCREENING CRITERIA REVISITED 23 Table 11. Steamflooding Description The steam drive process or steamflooding involves the continuous injection of about 80% quality steam to displace crude oil towards producing wells.000 cp Not critical but some light ends for steam distillation will help >40% V Sand or sandstone with high porosity and permeability preferred >20 ft >200 md (See Transmissibility) >100 md ftlcp <5. Since about one-third of the additional oil recovered is consumed to generate the required steam. less viscous crude oils can be steamflooded but normally will not be if the reservoir will respond to an ordinary waterflood. A low percentage of water-sensitive clays is desired for good injectivity. Steamflooding is primarily applicable to viscous oils in massive.000 ft not critical Range of Current Projects 8-27°API 10-20.

04 28.86 2. Percent Reported Method Steam Combustion Hot Water CO2 Hydrocarbon Nitrogen Flue Gas Polymer MicellarlPolymer Alkaline or Alkaline/Surfactant as Profitable 1982 86 65 1988 95 78 89 66 100 100 100 92 0 1990 96 88 78 81 100 100 100 86 0 1994 96 80 100 81 100 100 -- 21 50 100 100 72 0 40 -0 100 100 100 (1 success) Table updated from Ref.18 10.78 13.52 205.59 34.45 2. SERIGHT SPE 35385 Table 12.29 60. Profitability ofEOR Projects in the U.84 4.45 0.23 1.59 1.J. 33 Table 13.14 4.78 • Net reduction in world total.43 Urgency. TABER. Timing.80 Total CO2 Required to Produce Incremental Oil (from Tables in Ref. AND R. 410 .45 2.19 3.80 9.12 65. 23 and 24.S. 33) billions of barrels Middle East Western Hemisphere Africa Eastern Europe and CIS Asia-Pacific Western Europe World Totals 141.24 J. % -From r» Ref.73 1.98 3. 33) (G tonnes) 44.85 8.05 3.59 3.D.83 1. Summary of Estimated Worldwide CO2 DemandlUtilization Potential Oil Recovery and Oil-Producing gion Re- Potential Oil Production by CO2 Injection (from Tables in Ref.42 10. MARTIN. F.S.96 Gtonnes 23. or Regional Adjustment (%) Potential CO2 Utilisation (G tonnes) -12 +10 -5 0 -5 +15 _~-7]· 39.

_. 2 1114 1114 1111 .S. 2-API gravity range of oil that is most effective for EOR methods. 10 EORas% of U. (data from Ref._ '"2 1114 lHe 200. .S.-_-_-_-_-_-_~ 45 50 I 55 60 I c::::::::::: Polymer Floods C Gel Treatments -- A1kalinelSurfaclanllPo!ymer _ . ------------.. I-EOR production in the U. < _--=:::=========::::::: ._.. 411 .Miscf~-f~-_-_-_-_. Relative production (bid) is shown by size of type.000 1110 . . o 5 10 I 15 Oil Gravity GAPI 20 25 30 35 40 I -<::]mmiscible Gas -_-_-_-_-_-_-_-_-_-_-_-_-...._ '"2 1114 lHe Fig. 2 1114 1111 1111 .: _-: _ _. -------- < N.-_-:': . & Flue gas ... 24). Production 8 2~~--~~~~~~--~~~~ 1110 ..SPE 35385 EOR SCREENING CRITERIA REVISITED 25 ) 12 r----------------._ & Micellar yd rocarbon-----~~~~ Pi~r_. ----_ Fig. ._ CO2 .

..000 •• • • • IQ) ~ 0 ~0 0 0 Steamftoodlng 0 0 8§P cPo 0 Hat Water • + Of!] § 0 [J] 00 ~ o 0 o 00000 0 0 Combultkln o 0 [J:> 0 § CO2 • 0 0 g 0"".000 "'" + 20.000 0 10 20 30 40 50 60 0 MlcellarPolymer X AIIiaIlnei Surfactant ·API Fig.-- 1--..___.S.I...__ .. (data from Ref._--'_ ._-.I.J.000 • * 0 "'" + 0 HC Miscible & IlI1IT1iacjble + "'" "'" ~ i Nitrogen Miscible & Immiscible "'" + + Polymer 15. 4-Depth Ref..000 l- HC Miscible & Immiscible 15.D.) • and API gravity of producing EOR projects in the U..000 ....S.. _..._-.. F.26 0 J.S... o ~--------------------------------------.. _ _j.000 Io o • • o g Q II t 10. t Q II 10.. AND A.. MARTIN.. 412 .000 r- Polymer "'" o Mlcenar-Polymer X 20. Alkaline! o 10 20 30 40 50 60 Surfactant API Fig.. I.. TABER. 24. SERIGHT SPE 3531 5.I--I. 3-Depth 24..--_...) and API gravity of producing EOR projects outside the • (data from u.~--------~ o Steamflooding Hot Water Cbmbustion C02 • 5..__-.

000 245. Nitrogen l 011Gravity [ • API] Fig.000 ~ I: ~ :::II "CI o 60. 24).*.__ 80.000 1 Ill: 40..000 EOR SCREENING CRITERIA REVISITED 27 80.. 6--EOR production vs oil gravity outside the U. (data from Ref..000 ~ ~ :I D- s 60. 100.000 Stum ~ . 24).000 l ~ 40.OD Combustion Polymer & MlcellllrIPoIymer Hydrocarbon .000 . (data from Ref. 413 .000 fa 20.000 Hydrocarbon 011 Gravity [ • API] Fig. 5-EOR production vs oil gravity in the U.000 w 20.~PE 35385 100.41 J JAil .S.S.

./ /' ./ .._- '/ 17 '0 c ~ n. data from Refs./ "/ .-""T""-"r"--r-~-.000 .. AND R. EOR production vs 1976 NPC predictions. 32)........000 C) z-: /' .-..' ~ ~ ~ 3./ 't:J C o . . 33.S.. . 10...... ::l 't:J ~ 1980 ~ ~ ~ ~ ~ ~ ~ ... _ $151 Barrel_ _ _ ---..._ _ ..--.J..../ ...000 ~----------------------------------------------------------~ i... ... ...ooo_--r-../ .... F. $ 51 Barrel o ~----~----~--~----~----~--~----~----~--~----~ 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 Fig.. -.. 7-Increase in MMP and fracture pressure with depth for Permian Basin reservoirs (from Ref.... (Extended from Ref..../ I::..."..----------:...000 Actual EOR Production ..~ ~ ~ . SERIGHT SPE 3538( II ..) U..c ~ ~ C'G 2.. MARTIN...000 1976 Constant Dollars '0 (/) ...... IIJ :::l IIJ a: 1000 o o 2··3 4 5 6 7 8 9' 10 II 12 DEPTH IN THOUSANDS OF FEET Fig.. -_.. TABER. - $10 I Barrel - --. .. a..28 J...._.-..--r-....... 8-Actual 7 and 24.. a: en en Q........ F .. .. 414 . ....r-.S. " 4.D. 0:: 1.

. 8 and 24.. 33.: .. _.~a~~\-.... 100Sensitivity of U. . " _. data from Refs. 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Fig.--- ----- $ 20/ Barrel . __ __ ... $~O_lBarrel ..SPE 35385 EOR SCREENING CRITERIA REVISITED 29 _ 2.. _..'_"- ' ./ /" ---~ $30 I Barrel Actual EOR Produgti"op/ /" /" /' /_. data from Refs. .) 415 .) C UJ (IJ 1.--/ . EOR production to the crude price predicted in 1984 NPC report. (Extended from Ref. 9-Sensitivity of U..000 ..~/ .. ./7.' - / -_ $401 Barrel -" _--_ -. $30 I Barrel ----. :J 'U 1S 0 c a..500 1983 Constant Dollars 10% ROR / $5Q_/...- /"' (IJ / .-/" . 33._ 0::: (IJ :J 0 / /'.. CO2 production to crude oil price predicted by 1984 NPC report.-_.._. _ . .000 >- m '0 'U (IJ (IJ ~ 800 1983 Constant Dollars 10% ROR / ___ _ _ .s (IJ - .'- . / .. ...s s: . - -_ ..' _. 8 and 24. : / --. _--.._ 0::: / / / 1S :J 0 C 200 / e a. _.. -----. 'U ~<~:~:~'_....' ..... 500 s> .$SO/Ba rrel --$40 I Barrel / / / / / / / / / / / / UJ c UJ 600 :J 0 Actual CO2 Production 400 ~ ..S.- ../"' 'U C (IJ 1.. UJ ~ m '0 UJ UJ ~ 1.. (Extended from Ref..000 C Q) ...S. 0. e 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Fig...

This process includes polymer flooding and foam flooding.1 Polymer Flooding .CHAPTER II CHEMICAL RECOVERY PROCESSES Contents 1. 2. / IIp Mw-o Mw/ Mo = Mobility control processes inject a low-mobility displacing agent to increase volumetric and displacement sweep efficiency. This chapter is divided into two sections corresponding to above two processes. Mobili~yControl Processes The mobility is defined as the ratio of permeability to the viscosity and the mobility ratio is defined as the mobility of displacing phase (water) to the mobility of the displaced phase (oil).2 Surfactant Flooding Alkaline Flooding 4.2 3.1 3. 2.1 2. Conclusions 1. 2. and/or (2) lowering the oil/water interfacial tension (lOW-1FT process). Introduction Chemical flooding process is to recover oil by (1) reducing the mobility of displacing agent (mobility control process). Polymer Flooding Foam Flooding Low-1FT Processes 3. Mp = k. Introduction Mobility Control Processes 2.

polyacrylamides reduce water mobility mainly by . copolymers of acrylic acid and acryamide.1. water banks thickened with polymer are used to provide mobility control in pushing miscible orsurfactant banks in tertiary recovery projects. Polyacrylamides are synthetic chemical products. The mobility ratio of a waterflood can be effectively reduced by about a factor of 10 by adding polymer to the water. The effect of mobility ratio on vertical invasion efficiency and on pattern coverage efficiency was covered in the Waterflooding section. the volumetric coverage efficiency can be improved by "thickening" the injected water to improve the mobility ratio. carboxymethyhydroxyethylcellulose (CMHEC). the most attractive oil reservoirs. However. Generally.volumetric coverage efficiency and the residual oil saturation. glucan. polyacrylamide (PAM).1 Types 'of Polymer Several polymers have been considered for polymer flooding: xanthan gum (XCbiopolymer). Only two basic types of polymer . at least one-half of the original oil remains after water flooding. decreases. copolymers of acrylamide and 2-acrylamide-2-methyl propane sulfant (AM/AMPS). dextran polyethyleneoxide (PEa). polyacrylic acide. hydroxyethyl cellulose (HEC).XC-biopolymer and partially hydrolyzed polyacrylamides are widely used in field recovery projects. XC-biopolymer reduces the mobility ratio simply by increasing water viscosity. XC-biopolymer is a natural polysaccharide produced by a microbial fermentation process. Under current technology. As we will see later. 2.biopolymer to the injected water. Under some conditions. most light-oil reservoirs are waterflooded unless a strong natural water drive or an effective gas cap drive is already taking place.Polymer flooding is a process that a small amount of polymer is added to thicken brine (water) to reduce the water mobility. hydrolyzed polymer (HPAM). reservoirs for polymer flooding will fall in the 20-200 cp viscosity range. The oil recovery that can be expected by water flooding a given reservoir is controlled by two factors . and polyvinyl alcohol. Both of these efficiencies increase as the mobility ratio The main incentive for polymer is in flooding medium viscosity oil As a result. The mobility ratio of a waterflood can be reduced by adding a chemical such as polyacrylamide or XC.

. Viscosity Relations and Non-NewtonianEffects A fluid whose viscosity decreases with increasing shear rate (e) is shear thinning. The behavior of the polymer solution is caused by the uncoiling and unsnagging of polymer chains when they are placed in a shear flow. By contrast. retention. Obviously. c. This permeability reduction is relatively permanent and is not removed by a brine flush. the XC-biopolymer reduces water mobility in a core much less than does the polyacrylamide.1. d. qualitative trends. for a concentration of 600 ppm. or mechanical degradation. Figure 1A shows the viscosity increase as a function of polymer concentration. 2.2 Polymer Properties Because of the complexity of the subject. as shown in Figure 1B. the polyacrylamide reduces the permeability of the core to water. and chemical. the original permeability is soon restored.reducing the formation permeability to water. as illustrated in Figure 1B. a. a. lip = K (e) n-1 where K and n are the power law coefficient and exponent. when the XC-biopolymer is followed by a brine flush. For a given concentration. Another useful relationship is Meter model. a few quantitative relations and representative data are presented on these properties: viscosity relations and non-Newtonian effects. biological. a comprehensive study of polymer properties is not possible. respectively. However. However. The polymer solution viscosity 1 shear rate relationship may be described by a power law model. permeability reduction. lip where When = Ilpoo + (Ilpo -lipOO) I (1 + (e I e1l2t-1) e112= 1/2 x (Ilpo + Ilpoo) applied to permeable media flow these general trends and equation continue to apply is usually called "apparent" viscosity and the effective shear rate is based on capillary tube concepts. The effect of each type of polymer is illustrated in Figure 1. b. XC-biopolymer results in a much greater increase in viscosity.

causes the loss of polymer from solution. and residual resistance factor (FRr). Permeability reduction depends on the type of polymer. which can cause mobility control effect to be destroyed. the permeability to water behind the bank is greatly increased. the mobility of water-rich phase after polymer injection Awa d. The to water caused by the polymer is a big advantage since a reduction in permeability relatively small polymer bank can be followed by brine without losing mobility control. biological and mechanical degradation . FR FRr where = (kw / f. The bank must be driven by thickened water to compensate for this relative permeability increase. c. Retention also causes a delay in the rate of polymer propagation.where Uw = superficial flux of polymer-rich (water) phase kw = permeability = porosity to polymer-rich phase o b. This phenomenon is measured by the following factors: the resistance factor (FR). Field measurement values of retention range from 20 to 400 Ibm polymer / acreRetention ft bulk volume with desirable retention level being less than 50 Ibm / acre-ft. Chemical. Polymer Retention Polymer retention is due to solid surface adoption or mechanical trapping within porous media.lp ) Awa AWb = = the mobility of water-rich phase before polymer injection. Mobility control is required for processes using miscible or surfactant banks. Permeability Reduction Polymer retention reduces the apparent permeability of the rock. the pore size distribution. and the average size of the polymer relative to the pores in the rock. Because these banks remove the residual oil. the amount of polymer retained.lw ) / (kp = Awb / / f.

Water quality must be much higher for a polymer test than for ordinary water injection.g. A combination of water quality and polymer quality caused these injectivity problems. viscosity. surface equipments (valves. orifices. It was found that the Kelco biopolymer used in these projects contained about 11 weight percent cellular debris which caused much of the plugging.1. formaldehyde or isopropyl alcohol) Mechanical degradation occurs when polymer solution are exposed to high velocity flow.3 Problems with Polymers Field tests have exposed severe operational problems with both types of polymer.g.The average polymer molecular weight can be decreased to the detrimental of the overall process by chemical. Injectivity problems were encountered with XC-biopolymer water in most field pilots. The affect a water salinity on viscosity increase for both types of polymer is illustrated in Figure 2. High shear stresses can be encountered in surface pumps. and meters as well as at the point where fluids heave the wellbore and enter the formation. The major operational problem with polyacrylamide is that the material can be mechanically degraded when subjected to high shear stresses. Chemical degradation ca be minimized by restricting polymer usage to low temperature reservoirs and adding oxygen scavengers (e. problem with the polyacrylamides is related to water salinity. The retention of polymer in the pores of the rock is a very serious problem for either type of polymer. a greater concentration of polyacrylamides is needed to achieve a target XC-polymer is relatively unaffected by water salinity.. sodium sulfate or sodium sulfite) to the polymer solution. Another As water salinity increases. Biological degradation can be eliminated by adding oxygen scavengers and biocides (e. The molecules are stretched and ruptured by shear stresses and their ability to reduce mobility is permanently reduced. Lost polymer is absorbed on the walls of main flow channels and it . pumps. or tubing) and downhole conditions (perforations or screens). Biopolymer marketed by Pfizer has proven to be much better from the injectivity standpoint. and mechanical degradation. 2.. valves. biological.

.1. 2. although in most cases polymer resulted in reduced producing water-oil ratios. 4. Permeability should be greater than 20 md. as well as theoretical considerations. 2.2 FoamFlooding . The best results were obtained in projects where polymer was used early in the waterflood life. If the reservoir is being waterflooded. the greater retention of polyacrylamide results in a more permanent decrease in reservoir permeability. 3. The table emphasizes oil recovery and screening parameters used for polymer flooding.300 to 400 pounds By comparison. pounds per . acre-foot. less than 0. 100 field polymer Very few of the field projects were successful from the economic standpoint. the producing water-oil ration should be These guidelines resulted from experience obtained in approximately flood projects. The average oil recovery (AOR) is about 3. Oil viscosity should be in the range 20-200 cp.2.1.56 % SOIW and 2.4 Screening Candidatesfor Polymer Flooding should not be considered for any reservoir unless the following Polymer flooding guidelines are met: 1.5 Field Results Table 1 shows a polymer floods statics for more than 250 polymer floods from USA by Manning et al.may completely plug off narrow pore channels.69 STB The large variations of of improved oil recovery (lOR) per pound of polymer injected. As Polymer retention is more serious with per acre-foot in consolidated are retained XC-biopolymer pointed out retention will range from 150 to 300 previously. Currently very few large-scale thickened waterflood projects are in operation. Reservoir temperature should be less than 200°F. AOR and lOR reflects the nature of polymer flooding. polyacrylamide sandstones. 2. 2.

2. and have been both field tested as mobility control agents in miscible and steam floods. If surfactants are added to the liquid. cationics. Foams are dispersions of gas bubble in liquids.e. they exhibit surfactant properties simply by electronegativeity contrasts bewteen their constituents. A class of surfactants are those that contain aspects or two or Even small more of the previous classes. These surfactants do not form ionic bonds.2. (e. relatively resistant to retention.1 Foam Physical Properties ) Foams are characterized by three measures. Amphoterics. c. b. Nonionics. stability is improved greatly so that some foams can persist indefinitely. Most of the discussion applies to MP surfactant as well. Anionics.i. having strong affinity for oil) and a polar portion (hydrophile .Gas/Liquid foams offer an alternative to polymers for providing mobility control in chemical floods. strong affinity for water). In this case the surfactant molecule contains a cationic hydorphile and an inorganiC anion to balance the charge. texture. Nonionics are much less sensitive to high salinities than anionics and cationics. Anionics surfactants are the most common in EOR because they are good surfactants. and amphoterics. variations can change surfactant properties drastically. To understand the foam properties requires some discussion of surfactants and their classifications. stable. and bubble size range.e.i. Gas/liquid dispersions are normally unstable and usually will break in less than 1 second.. and relative inexpensive. however. a. . but when dissolved in aqueous solutions. Cationics. Table 2 classifies the surfactant into four groups: anionics. sulfates tend to be less thermally stable than sulfonate). The anionics surfactant molecule is uncharged with an inorganic metal cation (usually sodium) associated with the monomer. A typical surfactant monomer is composed of a nonpolar portion (lipophile .. d. Within anyone class there is a huge variety of possible surfactants.g. the entire monomer is an amphiphile (has affinity for both oil and water). quality.. nonionics.

3. Bubble size range.( a. The quality increases with increasing temperature and decreasing pressure both because (1) the gas volume increases.2 Foam Stability The liquid film The stability of a foam can be illustrated as shown in Figure 4. local heating.2. A foam with quality greater than 90 % is a dry foam.3 Field Results ( Fields tests of foam injection alone have been scarce. Thinning can be caused spontaneously by diffusion of gas from small to large bubbles and by gravity drainage. Foams with a large distribution range are more likely to be unstable because of gas diffusion from large to small bubbles. The fOllowing three conditions are worth to mention to foam stability. Texture is a measure of average bubble size. c. Very low surface tension would not be favorable. or low viscosity if in motion. The surface tension at the gas/liquid interfaces plays an important role in film stability. a.2. Quality is defined as the ratio of foam (gas) volume to the total volume. and (2) gas dissolved in liquid phase evolve from solution. The foam has been sucessfully applied to steam and miscible flood for mobility control. If the average bubble size is larger than the average pore diameter. This condition is most nearly realized in permeable flow for high foam quality. or contact with a hydrophobic surface. The hydrophiles of surfactant are oriented into the interior of film and the lipophiles toward the bulk gas phase. 2. b. External effects that may cause the foam to collapse are the presence of a foam breaker (oil or a high electrolyte concentration). separates two gas bubbles in cross section. 2. Low-1FT Process . the foam flows as a progression of films spreading individual gas bubbles. b. c. The pressure in the gas phase is assumed constant because of its low density if the foam is static. The texture determines how the foam flow through a porous media.

If the surfactant concentration then is increased. An idealized version of a MP flooding sequence consists of 5 stages: preflush.1 SurfactantFlooding If anionic surfactant is dissolved in brine.1. microemulsion. 3. the surfactant disassociates into a cation and a monmer. defined as the ratio of viscosity force to the capillary force as Nc = lIw J. CMC's are quite small (about 10. soluble oil. surfactant. residual oil saturation.3 to 10-4moll L). and chemical flooding. The curve begins at the origin. then levels off the critical micelle concentration (CMC)._ \ J .and ultimately. MP slug.1 Surfactant/Oil/Brine Phase Behavior .Low-1FTmethods rely on injecting or forming in-situ a surface-active agent (surfactant) which lowers oil/water 1FT. Processes that inject the surfactant are called "MP" floods because of the tendency for surfactant to form micelles in aqueous solutions and inevitable need to drive the micellar solution with polymer. increases with unit slope. and chase water.lw / awo The CDC is a nearly horizontal plateau at small Nc until a critical value above which both residual phase saturation decrease. low-tension. J 3. since these processes rely on reactions with acidic components of crude to generate the surfactant. The basic tool for illustrating how lowering 1FTreduces residual oil saturation (ROS) is the capillary distribution curve (CDC) shown in Figure 5. MPflooding has appeared in the technical literature under many names: detergent. Typically. fresh buffer. Figure 6 shows a curve of the surfactant monomer concentration vs total surfactant concentration. mobility buffer. The capillary number is . At the second critical Nc the residual phase saturations are zero and complete recovery of original trapped phase. the lipophilic portions of the surfactant begin to associate among themselves to form aggregates or micelles containing several monomers each. High-pH or alkaline processes produce a surfactant in-situ.

The overall composition within the two phase region now split into an excess brine phase and an o/w microemulsion. PR.usually is located quite close to the oil apex as shown in Figure 7. shows the entire progression of phase environment from Type 11(-)o 11(+). An overall composition within three phase region separates into excess oil and excess brine and microemulsion phase whose composition is represented by an invariant point. a typical MP surfactant will exhibit good aqueous (water-rich) phase solubility and poor oleic (oil-rich) phase solubility. there is a continuous change between Type II{-) and 11(+) nd a a third surfactant rich phase is formed. The right plait point ins such system. MP phase behavior is affected strongly by the salinity of the brine pseudocomponent.. The phase i \ environment is a Windsor Type 11(+) ystem. A second critical tie line also splits at Ceu (high effective salinity limit for Type III phase environment).The Type III region forms through the splitting of a critical tie line that t lies close to the brine/oil boundary as the salinity increases to Cel (low effective salinity limit for Type III phase environment). The Plait point PL is now close to brine s apex as shown in Figure 8. For an optimal salinity. This environment is called a Windsor type III as shown in Figure 9. At low brine salinity. For high brine salinities. therefore ternary diagrams are always used to represent the micellar solution phase behabior as shown in Figures 7 to 9. II means that no more than two phases can form. This type of phase environment is called a Windsor Type I system or Type II system. Thus an overall composition near brine/oil boundary of ternary will split into two phases: an excess oil phase that is eseential pure oil and a (water external) microemulsion (o/w) phase that contains brine surfactant and some solubilized oil. (-) means that the tie lines have negative slope. The micellar solution operates misbly with oil and water without phase separation and it differs from microemulsions due to the microscopic size of discontinuous phase. Figure 10. electrostatic forces drastically decrease the surfactant's solubility in the aqueous phase. a prism diagram. The phase behavior was presented originally by Windsor. .

. respectively: Assuming cos . 'J a equals 1. They are defined as Fmo = v. solubilization parameters between the microemulsion oleic phase. we can write equations for the capillary pressure (oil phase pressure minus water phase pressure) at the constricted downstream end (r1)and at the upstream end of the bubble (r2). 3..2 Phase Behavior and 1FT To investigate the relation between 1FT'sand phase behavior. these capillary pressures are p and cl = 200w r1 where: Oow oil-water interfacial tension .~.and the microemulsion water phase. Fmw = v. 'l% . ~.t»._.Jv. Figure 13 shows the corresponding behavior of the soulbilization parameters and 1FT's. The residual oil is in a discontinuous phase.1. . Fmo. With the aid of the capillary pressure equations..dynes/cm = r1 0000 The difference in capillary pressure over the length of the droplet is: = radius at constricted downstream end-ern r2 = radius at upstream end-ern .0. The trapped oil droplet shown in Figure 3 has been distorted in trying to leave the pore. and (}mw will be defined from Cel to high salinities.. Fmw will be used. (}mo will be defined from low salinity up to Ceu.3. we need to again look at an oil trapped in a pore of the rock and analyze the capillary forces that prevent the droplet from moving out.3 Fluid/Rock Interaction To understand the theory of surfactant flooding.1.

.29(106) L dynes X 1. The pressure gradient calculated in this example is at least 3 orders of magnitude This is why the ( greater than pressure gradients that are normally attained in the field. Problem Calculate the pressure gradient required . r2 where L = droplet length in direction of flow-cm.48em (em" x em) dyne / em" ft ~ = 571psi / ft.> ) Me = 1. Me L or = 2(30) 4(10-2) (1 9(10-4) 4(10- 1) 3 = 1.45(10-6) psi x30. (1A) The following example will show the magnitude of pressure gradient required to move out a droplet of residual oil in a typical water-wet sandstone reservoir. form of the capillary pressure gradient Equation (1) can be written in the Me = 28 ow (_!_ _ _!_) L L r. the water phase pressure must drop by an amount equal to ~Pc over the length of the droplet. r1 =9x10-4cm Use the following data: r2 Oow = 4 x 10-3 em = 30 dynes/cm droplet length = 4 x 10-2 em Solution Pressure gradient is calculated by equation (1A).29(106) dyne / em' / em '.to move a residual oil droplet out of a pore in a typical water-wet consolidated sandstone.(1) In order to move the bubble out of the pore.

Reservoir water salinity should be less than 50. However. and surfactant plus some other components such as alcohol that tend to stabilize the system.1. Divalent ion concentration should be less than 500 ppm. 8. 5. The droplets of the dispersed phase are so small that the emulsion appears transparent and is very stable. however the surfactant bank size is much smaller when the microemulsion is used. Waterflood residual oil saturation should be greater than about 30 percent. having a shelf life of several years.oil droplets will remain trapped in a field waterflood regardless of how much water passes through the reservoir. No surfactant project can be started until a detailed reservoir study including an economic analysis has been . A typical microemulsion might contain 45% oil. Homogeneous sandstone reservoir.03 dynes/em. 3. Oil viscosity should be less than about 20 cp. 7. The numbers quoted above are not absolute limits. This can be done by contacting the residual oil droplets with a surfactant that has been designed for this purpose. A reservoir cannot be ruled out just because a few of the above guidelines are not met. Permeability should be at least 100 md.3 Screening Guideline ) } Following is a list of reservoir and fluid properties that are considered to be desirable for a surfactant flood. oil. 3.000 ppm. and 10% surfactant. 45% water. A microemulsion is a stable solution of brine. 2. 4. Surfactants can be introduced to the reservoir by a simple solution of surfactant in water or by using a microemulsion.5 psi/ft. Guideline values are given for certain reservoir and fluid properties. The solution generally carries a much lower concentration of surfactant (2-3%) than the microemulsion (10%). completed. but serve as guidelines. 1. 6. Reservoir temperature should be less than 160°F. Sand should be at least 10 feet thick. In order to move the droplet under field gradients which are of the order of 0. the interfacial tension must be reduced from 30 dynes/cm to a value less than 0. a good prospect will meet most of these guidelines. Almost the same amount of surfactant is used with either method.

such as sodium carbonates.2 High-pH or Alkaline Flooding High-pH or alkaline flooding is an improved waterflood process in which the pH of the injected water is modified for increasing the oil recovery. such as sodium hydroxide (NaOH). since the { . b. High-pH's indicate large concentrations of hydroxide anions (OH} solution is defined as The pH of an ideal where the concentration of hydrogen ions. d. in moll L.4 Field Results At least 10 field tests of surfactant flooding processes have been done. These are: a. 3. c.3.2. Field experience to date suggests that the high-concentration micellar solution are more 3. Wettability alternation. . the concentration of H+ decreases. two concentrations are related through the dissociation of water. As the concentration of OH.1 Mechanisms There are four mechanisms believed to be involved in making alkaline flooding work. Emulsification and flow.is increased. effective than lo-concentrations of surfactant. and Rigid interfacial films. Interfacial tension alternation. CH+ . These considerations suggest two means for introducing high-pH's into formation: dissociation of hydroxyl-containing chemicals. Kw = (COH-) (CH+) I CH20+ and the water concentration is constant.1. or adding chemicals that preferentially bind hydrogen ions.

Therefore it is assumed that the lowering of 1FT due to reaction that takes place between organic acids and the alkaline solution also leads to the formation of emulsion. Crude oils form insoluble films at crude oil-water interfaces upon contact. As the reverse wetting agent arrives at the pore and the concentration of the agent begins to increase the surface of pore becomes progressively more oil-wet. By decreasing the 1FT. c. o is 1FT. Interfacial tension alternation Figure 13 shows an example of interfacial tension vs pH. Capillary forces now favor movement of oil droplet into the smaller parts and tight places within the pore. Figure 15 shows an oil droplet trapped under water wet conditions. The emulsion could be OIW or W/O or both depending upon oil. This forming material was from the family of resins and asphaltenes in the crude oil and that once . d. emulsification and entrainment. Nc must be increased to 10 -4. Rigid interfacial films. Rather there are may be an optimum concentration for which the lowest tensions are obtained as shown in Figure 14. Emulsion and flow In the laboratory studies.and u Darcy velocity.a. the produced oil from the alkaline flooding are fine droplets. Increasing the concentration of NaOH would not necessary decreases the 1FTcontinuous. one must reduce the capillary force to improve oil recovery. Recall that the capillary pressure defined as Pc = 2 (j Ir Changing in the pH activates the surface-active materials to some crudes and brings about gross wettability change.capillary number (Nc) is increased. Wettability alternation Two mechanisms in the emulsification process have been proposed to explain the incremental oil recovery: In waterflooding. b. Recall where Jl is the viscosity. To be effective.

the addition of NaOH or other alkaline materials to the water neutralizes the acid constituents in the interfacial films thereby solubilizing these films. more salinity resistant. which translates into recoveries. the development of more cost effective chemicals such as surfactants and polymers that b. d.2 Field Results Table 2 summarizes the high-pH flood statics. 4. to obtain good mobility control. removing technical limitations on the use water-soluble polymers. 3. indicating a declining oil rate. and more reliable prediction technique. The suqqestlons for future chemical flooding development are: a. The oil production rate declines as the total fluid production increases. Therefore these films reduce oil recovery. Conclusions As the oil viscosity increases. as they are temperature sensitivity. development of more efficient design procedure such as applying MP technology to high-pH flooding deSign. ranges from 0. c.0006 to 8. can be manufactured in-site or can be recycle. There are many features in these data that are common to the response of other chemical flooding process. Note the wide range in reservoir and oil characteristics and in oil saturation at the start of the flood. lOR expressed as a fraction of PV. . Hence increases oil recovery. more of the respective chemical is required Chemical flooding usually is limited to moderate to low oil viscosities because of economics. have minimal retention.2. expressed as a fraction of the OIP at the project start that are comparable with but slightly smaller than those reported from polymer floods.formed the film would dissolve but very slowly into oil phase.

..... o 1.... L- w >.O.... t- C 000000 q~~ C\I ."OC L-O "O~ ro·c·- 000 ro _0>> M<OI. ...O~<OOOO M .J U. zo... ..0 ~ :E o E·o Q.. . en o o o . CO W ..J m oCt t- ....DO> 0>0 f/) OOCOCOC\lM I.0 ...... o M co M ..... o 1.....0 o <0 CO 1. o en E :J E ~ <OM en oCt t= t= .. :J I .J ~ L- O· .C\I ..OO>C\I <OCO<O.....OCOCO.<OI........ I.. .

_ E ._._ C C c. -"'O""'~O Q) ~CN"'O_ Q)._Q.c o E E (5 .1a o I .) E 0 0- ....._ 0 :. E :._:.~ C ._ C.. 0._ «S ._ "'" «S "'0 Q) .- c z « en z t- o en « t« LL LL ::l a: o Z o t« o en en o « ...:.L:.EEoo ~ .C :... 000~ . c o "§ z o . s. .cn<.~ 00 . 'E...0... o C E :.c t- '" ...en .. '"5 Q._ "'0 C I C1) 0 :.J N LL + ~..J w 00 I « x w :E 0.

.

. <D 0 0 0 0 0 LO 0 en W _J ~ ~ 0 E·-..<OE"".. LOCO ) ....._- C\l0 MN 00 LOC\I C\I 0 C\I 0 . 0 ..._ a.... +:.. "0 Q) C ~ en () C 0 cr:LL Q)....c o .0 N- <0::> U ('I') 0 0 a> u: J: e.. ~ cu ..- .. C\I ('I') 0 0 ... u~o·E !:..oE~ -Q).: C\I..c ~ o u '..o~ c_ o . t Q) Q) ..- 0 C\I o 0 ~ ~ ~ J: LL J: 0 0 . LO 0 0 .0 0 uE~ <::JC) C) cE C') ·0 C\I_ ~....- CO C\I 0 0 \ 00::> C)~ ::Jo ('I') 0 0 m~ ...- 00- <t > a: c -........- C\I ~ 0 en I M W _J o 0 <t ~ m ..c u.....o .: ~o o- <D 0 -~...._. . W - LOLO LOLO C\l0 <D 0 . Q)Q)Q)..c 0 0 I ..Q) . ('I') 0 0 0 ('I') OC\l ·0 CO • 0 0 0 .c 0 0 I .: UcE: .(J)~()~ >-. ~ ~ ·in .c ('I') 0 0 ('I')C\I o~ .s.. Q) E:O U·-u ('I') ~ . -oQ):C ....0 Q) a.. OUa.0 Q).2 c ~ ...oECi>E . U~ >-c . I ....0 o·0 0 0 ('I') 0 0 0 .~ t z .-> > ... ~...~ Q)EQ) 0 O:C o ()O C') z ~ N Z ~z N~ 00 :c:c ~~ zz 0 :c ~ ·in ..- .

U 0 ~ u 0 0 CIJ cJ 0 U J..0 U g ~ CIJ J.-4 0 P4 0 l ~ ~ ~ ~ o . ) s: r:::: ~ 0 ~ (1) o CIJ "0. I..4 J..UOllnlos . CIJ cJ CIJ "-4 0 ~ ~ ~ oM orf.4 8 ~ '" P4 :> I 0 til ~ ""' =' ~ 0 ~ ..[amAIOa }O ~l190~81A .... ""• 0 0 uOllnlos ~amAIOa JO ~llllqoR aAllela~ \.j ).-4 < • j d~ .4 Cd ~ u cJ J.-4 o Q. u . ~ ~ :c 0 :E r:::: 'C ~ '" 0 ~ as . ~ ...... 0 0 . 0 0 . :> ~ 0 ! w 0r- '0 0 .-4 0 ~ . e ::s C) ~ 0 U "-4 '" . ~ ~ 0. 0 cJ ~ '" D._ E >"0 0 . a 0. i Cd t' cJ !:4 0 Po.....s 0 N .4 Cd e c:s 0 CIJ -u...... CD CD >. 1? .fJl on ~ J.4 0 U a:s ~ In 0 <J ~ • s In ~ 0 ~ ... .. 0 ) on ~ 0 0 r:I ~ ~ P4 ~ ! .

- B·rine (20. ( '6-1 -- . 0 ~/ .t.. / "..t 0 CJ .....::s ..- 0 C1.u . . . . ~ d '20 [ 10 8 6 4 polya~de 2 ~ ".ppm Figure 2 Effect of Salinity on Water Viscosity Increase .t..{.. .f.QOO ppm) Fresh Wa·ter ~.... til .&oJ :>. 100 80 60 _..... i .. 0 .' # XC-Biopolymer 6 40 CJ I 0 ... ... ·co co .. /' :> 1----------------------------~ 1500 1000 o 500 Polymer Concentration .

a. '0 +i ::J t- e c o .e 5 c "0 CD Co Co Q.c ~ is .

- __ ---- ADSORPTION / / / / / .t- :) 20 <:>.a: --- ~.M. PHASE /WETTING PHASE (/)W -0: O::. Contraction Larger surface tension. S.a: ~(/) _. I- Sorw o 30J---=:""'_--. ./ . CONC.."./ ./ .C./ . Less adsorption./ C.' NORMAL RANGE WATERFLOODS 0_ 00:: 0:: Z _. Figure 4 Surface Tension and Adsorption of a surfactant vs Concnetration t..a: 10 ~~ WI- NONWETTING Figure 5 Schematic Capillary Distribution Curves .w -.SURFACE TENSION \ / . Thinner.

SURfACTANT MONOMER CONCENTRATION TOTAL SURfACTANT CONClNTRATION Figure 6 Schematic of the CMC .

. . . / s WAT~R EXTERNAL MIC ROEMULSION EXCESS OIL B COMPOSITION OVERALL o Figure 7 Schematic of low-salinity/surfactantlbrine/oil phase behavior ( ...

No+ 5 EXCESS BHINE SINGLE OIL EXTERNAL MICROEMULSION PHASE B OVERALL COMPOSITION o Figure 8 Schematic of high-salinity/surfactantlbrine/oil phase behavior ) .1 No1BRINE Na+ .SWOLLEN MICELlE .

.( BRINE ( EXCESS BRINE s EXCESS OIL MIOOLE PHASE MICROEMlJLSION I I NVARIENT ~ OVERALL COMPOSITION 0 phase behavior Figure 9 Schematic of the optimal-salinity/surfactantlbrine/oil { ~.

.11: ~ '" en .c c._ 'a.+oJ c C .0 ._ e en "C E fn .c fn ~ E (5 ~ as ._ C"' CD fn '~ 7 . 'I- 0 CD CJ C CD ::s en c .LL e ::s 'i .fn +oJ +oJ e C CD E . ~ 0 + z t._ c > c CD CD fn e as .

.. : tZ f \ o o 0. % NaCI Figure 111FT and solubilization ratios .6 SALINITY. &AI :E z 10-1 >o Q CI) z - (J""mo z ~ . o .J.8 1...10-' ~---------------------------------------.4 0..2 1. o 0:: &AI 10-·.

J9
c
0 0

E

CI)

c
:;:J
o

w
(!)

e0

0:

... <0

:r '"

U

c t)

ca ..., ca

'C

tn

)

~

.C.)

..... 0

tn

:s

z
Q~
~:I
(!)

en·
0:

i) 0:

z

~o ~~
I.L.I.L.

tJ)
0r-

.c C.)

10 E CI)

N

o
en
o

'"

W

o

wt: UQ.

..JZ ..JO

s~ °en
'" 0
..J

e :s .LL

C)

g3
..J

'" en
Q.

'"

f \,

2B.-----------------------------------------------------------------~

24

<,

E E

20

z
z
0

16

en

w

::z:

I.J <!


\

I-

a: w
::z:

<! u,

0

12

e

\
4

\

\

\

\

pH

Figure 131FT vs pH for a Crude Oil
f

40.0

1\
10.0 ~
~ ~ ~
I-

\

u

en
Z
0

"-

5.0~

\

e

·ccr. TI.,,-,,,,, - - _. -_ .. _ - --PHloSl

""\;r~·..Js

""o( tt

D"STILLEO "";ECTJI)t,.

~:'T[R ""AHA

l.&J

>-

~

\

r.

o ""Oc.uc£o
""toTE Ji

CO~I"'''l[(J

0 Z
(f)

z

.
0.50-

,_

\

\

1.0 ;:- .
~
lI-

l.&J
_.J

....

\

\

.... z
0.05

0:: l.&J

< u < lL.

...
~

~

\

\

\

'\

0.10~ .
~ ~ ~

Q.

<,

I-

-----

_./
J
0.4

/

/

o

i

o

I 0.1

I .O.~

I 0.3

f 0.5

CONCENTRATLON OF No OH.· WEIGHT

PERCENT

Figure 141FT vs NaOH Concentration for a crude at 24 oC.

(
(0)
(b)
~.

-.

~.~

WATER WET
~(d)

OIL WET

~

.
EARLY OESORPTION

~
WATER WET

Figure 15 Mechanism of Trapped Oil Removal by Reverse Wetting

(
\

SPE/DOE 35354

A METHOD FOR SIMPLIFYING FIELD APPLICATION OF ASP FLOODING
Troy R. French, SPE, 80M-Oklahoma,
Copyright This paper 1996,

Inc.

Socie1y of Petroleum Engineers, Inc.
at the SPEIDOE Tenth Symposium on Improved Oil

R8OOII8ry held in Tulsa, OK. 21-24 April, 1996.

was prepared for presentation

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(sl. Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correcbon by the author(s). The material, as presented, doss not necessarily reftect any position of the Society of Petrol8um Engineers, its offioers, or members. Papers presented at SPE meetings are subject to publication review by Ed~orial Committee of the Society of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of where and by whom the paper is presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A Telex, 163245 SPEUT.

Abstract
An injection strategy that should simplify field application of the ASP (alkali/surfactant/polymer) process was developed. Additionally, this procedure should be useful for increasing oil recovery from on going polymer floods. Process design was accomplished with computer simulations which were )upported with laboratory experiments. Small slugs of concentrated alkali and surfactant, which were diluted by in situ mixing, produced suitable 'conditions for ultra-low 1FT (interfacial tension) at the leading edge of a larger polymer slug. Computer simulation was used to determine diffusion/dispersion parameters so that the degree of in situ reservoir mixing of small chemical slugs could be predicted. A three-dimensional, three-phase, reservoir simulator was used for simulation. The derived diffusion/dispersion coefficients were used for chemical flooding simulations, and the simulation predictions were verified with laboratory core floods. Coreflood effluent compositions matched computer simulation predictions. When followed with polymer, injection of small, concentrated, slugs of alkali and surfactant mobilized as much oil as injection of larger, more dilute, slugs of alkali/surfactant. This procedure should simplify field application of ASP flooding because very small slugs of alkali and surfactant can be easily handled and injected in an oil field. It was also shown that injection of small slugs of alkali and surfactant during an on going polymer flood should improve oil recovery significantly. An additional 23% PV of oil was recovered when small alkali and surfactant slugs were injected during polymer flooding.

Alkaline flooding is recognized as a tk>tentially cost-effective chemical flooding method. However,

Introduction 7ield Experiences.

early field projects conducted with strong alkalis, such as sodium hydroxide or sodium orthosilicate, were usually disappointing. Problems encountered in the field were typically associated with high injection pressures and silicate scale which was precipitated in production wells. 1 In the mid-1980s, rock-alkali reactions were identified as the cause of most problems associated with alkaline flooding projects. It was discovered that the extent of rock-alkali reactions could be reduced by using weaker alkalis such as sodium bicarbonate, sodium carbonate, and buffered mixtures. By 1989, it was also known that the interfacial properties of weak alkalis could be further enhanced by addition of small amounts of synthetic surfactants. Patents on this process are held by Shell Oil Company- and NIPER.3 The field viability of surfactant-enhanced alkaline flooding has been shown by successful field experiments performed in the United States by Shell4 and Surtek.5-7 The process is also reported to have been successfully applied in at least one other country, but there is apparently no official report available for that project at this time. Injection Strategies. Laboratory coreflood tests conducted at this laboratory with relatively large chemical slugs, 0.25 to 0.75 pore volume (PV) in size, yielded the following observations about injection sequence: • Oil recovery was highest when alkaline surfactant was followed by polymer. • Oil recovery was only slightly decreased when the alkali, surfactant, and polymer were simultaneously injected (mixed a in single slug), and oil production occurred faster. • Oil recovery was significantly reduced when surfactant was followed with alkaline polymer. • Oil recovery was lowest when surfactant was followed by polymer, and there was no alkali in any of the injected mixtures. When alkaline preflushes were injected before injection of the main chemical slug, it was observed that: • Oil production increased as the preflush volume was increased from 0 to 0.25 PV. • The alkaline preflush increased oil production in all cases, but the effect was most pronounced when the cores were saturated with brine that contained divalent ions. These observations were derived from corefloods conducted with relatively large injection slugs of alkali and surfactant,

97

2

A METHOD FOR SIMPLIFYING FIELD APPLICATION OF ASP FLOODING

SPE35354

and another observation was that, generally, oil recovery decreased when the sizes of the dilute chemical slugs were decreased.8 Since large chemical slugs are expensive to handle in the oil field, injection of very small slugs could simplify the method. Since oil recovery was shown to be decreased by injection of very small dilute chemical slugs, researchers here decided to test an alternate strategy-injection of very small concentrated slugs of alkali and surfactant. The hypothesis was that in situ mixing of small concentrated slugs of alkali and surfactant could produce optimal conditions for low 1FT and adequate mobility control when injected ahead of a larger slug of polymer. As in situ mixing occurs, process conditions should become similar to propagation of a small micellar slug ahead of polymer.9-IO This hypothesis was verified with computer simulations and laboratory corefloods. Computer Software
An injection strategy that may simplify field application of the

apparent for surfactant concentration, where the experimental data were affected by surfactant adsorption. The simulation values of 0.02 and 0.01 ft2/day for alkali and surfactant, respectively, were used for the computer simulations of simplified ASP injection strategies. Polymer concentration was derived from a correlation between effluent viscosity and polymer concentration, and the calculated diffusion/dispersion coefficient for polymer was 0.01 ft2/day. Coreflood Procedure Laboratory corefloods were conducted at 46° C in 25.4 em long, high-permeability Berea cores. The cores were first saturated with 0.6 wt% NaCl, then oil-flooded with crude oil to residual water saturation (Soi). Following oil-flooding, cores were flooded to waterflood residual oil saturation (Sow!) with 2.0 PV of 0.6 wt% NaCI. The chemical systems containing surfactant and alkali were then injected, followed by polymer. Frontal advance rate for chemical floods was 1ft/day. The chemical floods were conducted, after waterflooding,by injecting small, separate, slugs of chemical formulations that contained surfactant or alkali. For comparison, additional corefloods were conducted with larger volume slugs of the same chemicals. The total injected mass of surfactant, alkaIi, and polymer was maintained constant for the series of corefloods. Coreflood effluents were analyzed and the results compared to the computer simulations. Chemicals and Crude Oil A crude oil and an ASP chemical system was selected for / coreflood tests of the. Simplified injection strategy. For the process to be successful, dilution of small slugs of alkali and surfactant by in situ mixing must produce suitable conditions so that very low 1FT will occur in the zone near the leading edge of the larger polymer slug. The coreflood experiments were conducted with crude oil from the L. Wendt lease in Government Wells North (TX) field and mixtures that contained 3.75 wt% Na2C03 and 2.5 wt% active Chevron Cbaser™ CF-IOO anionic surfactant dispersed in 0.6 wt% NaCl. Very low 1FT values were measured between this oil and the chemical system up to 64 fold dilution with 0.6 wt% NaCI, and comparative corefloods were conducted with larger chemical slugs that contained 0.75% Na2C03 and 0.5 wt% surtactant.l+ The mobility control buffer contained Pflzer Flopaam™ 3330S polyacrylamide polymer in 0.6 wt% NaCI. During corefloods, polymer was followed with chase brine that contained 0.6 wt% NaCl. Analytical Procedures Coreflood effluent samples were collected in 5-mL increments and centrifuged before the oil and brine phases were separated. Aqueous phase alkaIinity was measured by titration to pH 4.0 with 0.1 N HCl, and surfactant concentration was measured by two-phase titration as described by Rosen and Goldsmith.15 A Brookfield Model LVT-CIP 1.565° cone and plate viscometer was used for measuring effluent viscosity, and interfacial tension between crude oil and aqueous mixtures was measured at 46° C with a Model 300 University of Texas spinning drop , interfacial tensiometer. -,

ASP chemical process is to inject very small slugs of alkali and surfactant ahead of the polymer flood. Computer simulation was used to determine diffusion parameters so that the degree of in situ reservoir mixing of small alkali and surfactant slugs could be predicted. For the process to be successful, dilution of small slugs of chemicals by in situ mixing must produce suitable conditions so that low 1FT will occur in a zone near the leading edge of the larger polymer slug. Diffusion/dispersion coefficients for surfactant, alkali, and polymer were required for input to the simulation program. PC-GEL is a three-dimensional, three-phase (oil, water, and gas), five-component reservoir simulator developed by incorporating a chemical transport model into a black oil computer simulation program, called BOAST. I I The simulator features include component diffusion/dispersion in addition to adsorption, convection, and in situ gelation parameters. The in situ gelation model and simulator were validated against data reported in the literature. The PC-GEL simulator is a useful tool for predicting the amount of in situ mixing that will occur during transport through porous media. Coreflooding diffusion was simulated in one dimension on laboratory scale. The 25.4 em core length was divided into 25 grids in the model with one injection and one production "well," each at the opposite ends. Both the injection and production speeds were 1 ftlday during the simulations. The core porosity and permeability were 0.2075 and 855 md, respectively. Preflush, sample injection, and water flooding procedures were simulated, and changes in component concentrations versus time and in situ distance were simulated Diffusion/Dispersion Coefficients The Diffusion/dispersion coefficients that were used for input to the PC-Gel numerical simulator were derived from data generated during coreflood experiments that were described in a previous report.12 Diffusion/dispersion coefficients were derived from polymer, alkaIinity, and surfactant data generated r1uringthose coreflood experiments. The simulator derivations are symmetrical (normal distribution) curves, which do not .xactly match the experimental data.13 This is particularly

98

e. which helps to confirm the potential effectiveness of the process. in 0. additional -orenoods were conducted with larger volume slugs of the he chemicals. after waterflooding. Figures 8 and 9 show that both alkali and surfactant persisted in the effluent longer for the coreflood where alkali injection preceded surfactant injection. and that a significant amount of mixing of polymer and surfactant should occur near a wellbore.05 PV of a mixture containing 2.75 wt% Na2C03 was injected.60 PV polymer. (See Table 1. The figure shows that small slugs of surfactant and alkali began to mix immediately. except that a large volume of polymer was injected before injection of the mixtures that contained alkali and surfactant. Figure 4 shows that very little separation of surfactant and alkali occurs as the chemical slug is propagated through the core. Figure 3 shows that alkali and surfactant should be wellmixed in the coreflood effluent.25 PV of a mixture that contained both anionic surfactant and alkali (Na2C03) in the same chemical slug. Laboratory Corefloods Chemical floods were conducted. For comparison. Also. The chemical systems and injection strategies were exactly the same as for the previously described floods. Additional corefloods (corefloods 5 and 6 in Table 1) were conducted to confum the effectiveness of injecting small alkali and surfactant slugs during an ongoing polymer flood.6 wt% NaCI. Comparison of Figure 7 with Figure 6 shows that polymer was detected in the core effluent at the same time that tertiary oil production began. for mobility control. In the mixing zone. Changes in core oil saturations are shown in Figure 12. an under optimum salinity may occur at the back edge of the slug. A~S Alkali injection preceded surfactant injection.500 ppm Pfizer Flopaam™ 3330S polyacrylamide polymer. Oil cuts for the two floods are shown in Figure 11.60 PV of 1. Core oil saturations after waterflood were about 44% PV. Polymer flooding.6 wt% NaCI. Figure 1 shows concentration profiles in the core for each of the three chemical components after 0. the polymer slug was followed with chase brine.6 wt% NaCI. The total injected mass of surfactant. In this case. after 99 . are shown in Figure 10.10 PV of an anionic surfactant. The chemical system simulated in Figures 1 through 3 consisted of injection of 0.6 wt% NaCI. Examination of Figure 6 shows that oil recovery was lowest when alkali and surfactant were injected simultaneously. A~S. Changes in core oil saturation for coreflood 2 described above. The alkali-surfactant mixture was followed with 0. and highest when a small slug of alkali preceded surfactant. Alkali and surfactant are also shown to be mixed with polymer in the zone near the leading edge of the polymer slug. intermediate when a small slug of surfactant preceded alkali. The shapes and positions of the effluent concentration curves shown in Figure 5 are only slightly different than for injection of small.16 Figure 2 shows that mixing of )rfactant and alkali continue during transport through the core. A&S Alkali and surfactant injected simultaneously. Coreflood 2 was conducted with the same injection strategy as core flood 1.60 PV of 1. separate. The figure legend should be interpreted as follows: S~A Surfactant injection preceded alkali injection. i. Core effluent analyses for the three corefloods are compared in Figures 7 through 9. surfactant and alkali concentrations are initially optimum.6 wt% NaCI. A simulation for injection of small chemical slugs of surfactant and alkali followed with a larger slug of polymer is illustrated in Figures 1 through 3. Results from the two corefloods are almost identical. an initially over-optimum salinity condition should be approaching more optimum concentration levels. ~hich are in separate slugs. the maximum concentration of surfactant and alkali is about 50% of the original surfactant and alkali concentrations. This effect should reduce adverse surfactant solubility conditions that could cause precipitation of surfactant by contact with concentrated alkali.5 wt% active Chevron Chaser™ CF-l00 anionic surfactant. in 0. (The simulator does not take into account alkali-surfactant interaction that could occur. separate surfactant and alkali slugs that become mixed in situ. polymer appeared in core effluent at the same time in all three corefloods. and the polymer solution was followed with chase brine that contained 0. a chemical system is simulated in Figures 4 and 5 that consisted of injection of 0. the alkali and surfactant concentrations are lower than for the previous simulation. FRENCH 3 Simulation of ASP Injection Strategies 'liffusion/dispersion was simulated in the PC-GEL oneJimensional model on laboratory scale.5 wt% active Chevron Chaser'" CF-l00 anionic surfactant in 0.05 PV of 3. and polymer was equal to that of the two previous corefloods. Polymer was followed with chase brine. and as dilution with formation water and polymer occurs. slugs of chemical formulations that contained surfactant or alkali.10 PV of alkali (Na2C03). Therefore. The polymer solution was followed with chase brine that also contained 0. using the diffusion/dispersion coefficients that were described above. also shown in Figure 10. and 0. Injection of surfactant/alkali was followed with injection of 0.) Coreflood effluent fractions were analyzed and the results compared with the simulations described above. for mobility control. A~S.75 wt% Na2C03 in 0. by injecting small. which could be an artifact of core permeability distribution.500 ppm Pfizer Flopaam™ 3330S polyacrylamide polymer. For comparison. followed with 0.SPE35354 TROY R.21 PV of fluid injection.. was conducted with higher polymer concentration to confirm that adequate mobility ratio was being used for the series of corefloods. followed with injection of 0.6 wt% NaCI. Coreflood 3 was conducted with a relatively large-volume chemical slug that contained both alkali and surfactant. A total of 0. A nearly identical flood (coreflood 4 in Table 1). Coreflood 1 was conducted with small volume chemical slugs: 0. 0.6 wt% NaCI was injected. The surfactant and alkali slugs were followed with 0.60 PV polymer. Oil recoveries from the three corefloods are shown in Figure 6. The injection of surfactant and alkali was followed with injection of 0.) The well-documented interaction that may occur between polymer and surfactant should be reduced due to the initial separation of surfactant and polymer.25 PV of a mixture containing 0. In each case. alkali. except that the alkali solution was injected before the surfactant was injected. which was shown in Figure 3.

502." JPT v. 1995) 104-29. T. 5. (Oct. The chemical cost per barrel of incremental oil was at minimum value for this amount of surfactant. 14. Patent 4. Patent 4. Special thanks go to C. "Design and Field Implementation of Alkaline-Surfactant Polymer Chemical Enhanced Oil Recovery. Chang. Thigpen." JPT v. as well as being simpler to apply in an oil field 4. This is 'nearly identical to the amount of residual oil recovered by the similar alkali/surfactant chemical flood (A--+S shown in Figure 6) conducted in a core immediately after waterflood. Wyatt." paper SPE 17538 presented at the 1988 SPE Rocky Mountain Regional Meeting. Thigpen. DOE Report No. FY 1995 Annual Technical Progress Report. K. Nomenclature So. M.S. r. B. (Oct. thereby producing suitable conditions for ultra-low interfacial tension (IFf) at the leading edge of a larger polymer slug. Vol. Josephson for his assistance in all phases of the experimental work and to M. J. and S. Pitts. 8.S. Three-Phase Black Oil Applied Simulation Tool (Version 1. R. Meyers. NY (1972) 42324. Goldsmith. % PV Sow! = Core oil saturation after waterflood. 1. but with decreased surfactant concentrations. J. (Apr. DE87001240 (May 1987).. French. "User's Guide and Documentation Manual for 'PC-GEL' Simulation. Staged PreformedSurfactant-Optimized Aqueous Alkaline Flood. (Feb. core oil saturations were reduced to about 18% PV. Pitts.05 PV each). H. For this oil and chemical system. J. A. B. and C. Conclusions 1. 1973) 77-79. Wyatt.817." U. WY. = Initial core oil saturation (after oil flood). Nelson. This corresponds to recovery of 26% PV of core oil after waterflood.. 100 .S. 1989). "Alkaline Flooding Injection Strategy. Casper. T. 12. D. 1985). 24 (Aug. which corresponds to recovery of only 3% PV of oil by polymer. W. Holm.S. M.S. (Oct. diluted by in situ mixing. DOE Report No.S. A. R. Clark. OOEIBCIl0033. An'. 1982). Western. Harpole. 3 . M. v. 6. Easterly." U. and C. and D. Corefloods demonstrated that the Simplified ASP procedure has the potential to be economic in terms of chemical costs. S.541 (Mar. L. H. Evaluation of the Alkaline Waterflooding Demonstration Project. Management and Operating Contract for the Department of Energy's National Oil and Related Programs. C. May 11-13.M. NIPER-705. B. and K. R. Systematic Analysis of Surface Active Agents. % PV So = Instantaneous core oil saturation. and S. NIPER-635. Surfactant-Enhanced Alkaline Flooding Formulations. 15. all conditions are identical except for changing the amount of surfactant injected. This procedure utilizes small slugs of alkali and surfactants. This illustrates the well-known fact that polymer alone does not mobilize much residual oil after a homogeneous core is waterflooded. 1992). Smith. U. 4. Ranger Zone. For the series of floods shown in Figure 13. Rosen. and H." U. 10. M. "The Effect of PolymerSurfactant Interaction on the Rheological Properties ! . Surkalo. Acknowledgment The author wishes to thank R. C. and K." U. "Surfactant-Enhanced Alkaline Flooding Field Project. 2. R... additional corefloods ( corefloods 7 through 11 in Table 1) were conducted with the same injection strategy described for the second coreflood (a small slug of alkali preceding a small slug of surfactant: A--+S). DOE contract DE-AC22-94PC91008.1). R. NIPER-563. 16. French. These corefloods demonstrate conclusively that injecting small alkali and surfactant slugs during an ongoing polymer flood could greatly improve the oil recovery from the polymer flood. 25 (Jan. R. M. Examination of Figure 12 shows that after injection of the small slugs of alkali and surfactant (0. NIPERlBDM-OI93. OK. 11. as well as being simpler to apply in an oil field. Bryant for her guidance and advice in conducting this study.: J. W. 1993).715 (Apr. Corefloods showed that injecting small alkali and surfactant slugs during an ongoing polymer flood should greatly improve the oil recovery from the polymer flood. and H. 1992). "Comments on Finding the Most Profitable Slug Size. Bujnowski. ft2/daY References Dauben. to about 41% PV. OOE Report No. Franchi. 22-24. Pitts. Lawson. Gao. "BOAST: A' Three-Dimensional. the amount of incremental oil remained relatively constant until the amount of injected surfactant was reduced below 48 mg. Funding was provided by Fossil Energy I U. T. (Sept. W. J. % PV D = Diffusion/dispersion coefficient. "A Field Test of Cosurfactant-Enhanced Alkaline Flooding. 22-24. 1. Deborah A." U. (Sept. 9. (Sept. "Finding the Most Profitable Slug Size. OK. An injection strategy that should simplify field application of the ASP (alkalilsurfactant/polymer) process was further refined. M. 7. L. 1995) 795-803.: OOE Report No. I: Technical Description and Fortran Code. "AlkalineSurfactant-Polymer Flood of the West Kiehl..B. 3. reduced the oil saturations. and M. J. 1992). NIPER-569. Wilmington Field. Peru. Aqueous Flooding Methods for Tertiary Oil Recovery. Jones. To further refine and demonstrate the process. Josephson. 1972) 993-94. DOE Report No. A. Wiley-Interscience. This demonstrates that the simplified ASP procedure has the potential to be economic in terms of chemical costs. French. S. Falls. Chang for performing the computer simulations. R. D. OOElBC/10830-5 NTIS Order No. "Design and Application of an Alkaline-Surfactant-Polymer Recovery System to the West Kiehl Field. Computer simulations were in good agreement with the analytical results from coreflood tests. and R." paper SPE/DOE 24117 presented at the SPEIDOE Eighth Symposium on Enhanced Oil Recovery held in Tulsa. 1991). J. 1991).S. Josephson.S. 2. Minnelusa Unit. U. M." paper SPEIDOE 24144 presented at the SPEIDOE Eighth Symposium onEnhanced Oil Recovery held in Tulsa." Proceedings of the Sixth UNIT~ International Conference on Heavy Crude and Tar Sands. M. (Apr.. CA. 13. R.4 A METHOD FOR SIMPLIFYING FIELD APPLICATION OF ASP FLOODING SPE35354 waterflooding.

7 2 3 2.i 37.05 PV surfactant. 0.05 PV surfactant.6 PV polymer 2. Ma~.5 1.0 PV waterflood.SPE35354 TROY R.8 Coref!ood Ici!i!Cti<m stratoov 2.6 PV polymer Surfacta[)t :1S :1S 5 AIkaI. 0.5 37.5 1.5 1.0 PV waterflood.0 PV waterflood.4 432 6) 48 15 6 31 101 . 0.05 PV surfactant. 0. 0.25 PValkaii/surfactant mixture. 0. FRENCH 5 Table 1 .05 PV alkali. 0.4 172 23.7 41.5 1.0 PV waterflood. 0. 1.0 43.8 242 31.5 43.5 46.Corefloods Ch!i!rnical !OQWlctm1i2c.5 1.05 PV surfactant.!mL.0 PV polymer 0.5 2. 0. 0. 0.05 PV surfactant.0 PV waterflood. 0.5 1.05 PV alkali.6 PV polymer 4 5 5a 6 6a 7 8 9 10 11 25 0 :1S 0 25 20 15 5 2 10 44.05 PV alkali.6 PV polymer 2.5 1.5 37.5 f2.7 16. ms SurfactaDt 138 85 S3 73 0 85 0 70 Qil !i!atYm1i2c. 0.5 37.6 PV polymer 2.05 PV alkali.6 402 44.05 PV alkali.5 0 37.6 PV polymer 2. 0.05 PV alkali. 0. 0.5 202 18.05 PV surfactant.5 1.05 PV alkali.0 PV waterflood.05 PV surfactant. rng. 0.05 PV alkali.5 1.6 PV polymer 2.1 41.6 PV polymer 2.8 16.0 PV polymer 0. 0. 0.05 PV ~rfactant. 0.6 45.1 42.0 PV waterflood. 0.1 19.5 Ch!i!rnical 24. 0.~qE~ W$rflooc! 47.5 37. 0.4 46.5 45.5 37.5 0 37. 0.5 37.0 PV waterflood.6 PV polymer 2.6 PV polymer 2.05 PV surfactant.5 1.0 PV waterflood.05 PV surfactant.05 PV alkali. 0.OOIw 1.0 PV waterflood.5 7. 0. 0.6 PV polymer 2.05 PV alkali 0. 1. 0.5 37.0 PV waterflood. 0.8 1.8 21.0 42.

1.4 _J -c a: 0 0.2 FRACTIONAL DISTANCE Ag.60 PVof polymer. 102 .4 \ a: 0 Z 0.60 PV of polymer.21 PV of fluid InJection.10 PV of alkali.10 PV of alkali. \ \ --e.• Polymer IZ Z o 0.0 ---o------c..• 0.6 A METHOD FOR SIMPLIFYING FIELD APPLICATION OF ASP FLOODING SPE35354 1.6 0 0 W o ~ ~ 0.8 0. after 0. Z W IZ -c a: I- 0 .60 PV of fluid InJection.10 PVof surfactant followed by 0. 2-COrefloodlng simulation for Injection of 0.Surfactant _Alkali \ \ \ \ 0. 1-COrefloodlng simulation for Injection of 0. and 0. after 0.10 PV of surfactant followed by 0. and 0.8 Polymer ' Z o 0 0.0 Z - Surfactant t= ~ W 0 -Alkali 0.6 \ \ o 0 W N \ \ _. ~ c( 0.2 Z FRACTIONAL DISTANCE Ag. .

25 PV of a surfactant/alkali by 0.2 \ \ \ \ \ 0.0 0.J \ \ \ \ \ \ \ \ \ Polymer :2 Z -c I \ \ a: 0 0.8 I I I I I I I I o 0 Cl 0. FRENCH 7 1.40 PV of fluid InjecUon..6 0 o C W N 0..6 o ~ 0.60 PV of polymer.- 0.8 1.4 '0... 3-SlmulaUon of coreflood effluent for InJecUon of 0.. -0_ I I I \ \ \ \ \ Z s tZ Z W W 0 t- -Surfactant _Alkali 0. PV FIg.0 .10 PV of surfactant followed by 0. and 0.0 2.0 EFFLUENT VOLUME.2 \ \ \ \ b__ 0.0 \ Z b \ -\ Surfactant 0.0 0.8 Polymer U Z 0. 1. 4-Corefloodlng simulation for InJecUon of 0.2 0.4 I .60 PV of polymer.SPE35354 TROY R.10 PV of alkali.0 FRACTIONAL DISTANCE Rg. 1. followed with chase brine.. after 0. mixture followed 103 .6 0.0 3.- 0 tZ W -Alkali - I- a: « 0.4 :i :2 Z « a: 0 0..

Sign up to vote on this title
UsefulNot useful